From martinfeck at aol.com Wed Jun 1 08:56:17 2016 From: martinfeck at aol.com (Martin Feck) Date: Wed, 1 Jun 2016 08:56:17 +0200 Subject: [tex-live] tlmgr does not work In-Reply-To: References: Message-ID: Hi there, tlmgr does not work correct ... xelatex martin # tlmgr --version tlmgr revision 39298 (2016-01-07 03:44:29 +0100) tlmgr using installation: /usr/local/texlive/2016 TeX Live (http://tug.org/texlive) version 2016 Message: xelatex martin # tlmgr --gui Lade lokale TeX Live Datenbank (/usr/local/texlive/2016/tlpkg/texlive.tlpdb) Dies kann eine Weile dauern, bitte warten ... Unknown directive ...containerchecksum c59200574a316416a23695c258edf3a32531fbda43ccdc09360ee105c3f07f9fb77df17c4ba4c2ea4f3a5ea6667e064b51e3d8c2fe6c984ba3e71b4e32716955... , please fix it! at /usr/local/texlive/2016/tlpkg/TeXLive/TLPOBJ.pm line 210, line 36. Compilation failed in require at /usr/local/bin/tlmgr line 4654, line 36. Is there a solution? Best regards Martin Feck -------------- next part -------------- A non-text attachment was scrubbed... Name: martinfeck.vcf Type: text/x-vcard Size: 100 bytes Desc: not available URL: From preining at logic.at Wed Jun 1 09:19:30 2016 From: preining at logic.at (Norbert Preining) Date: Wed, 1 Jun 2016 16:19:30 +0900 Subject: [tex-live] tlmgr does not work In-Reply-To: References: Message-ID: <20160601071930.GO965@logic.at> Hi Martin, > tlmgr does not work correct ... no, it *does* work correctly. > Is there a solution? Yes, reading the archives ... reported some time ago. Including the solution: Use tlmgr-update-latest.sh from the tlpretest area. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From preining at logic.at Wed Jun 1 09:32:11 2016 From: preining at logic.at (Norbert Preining) Date: Wed, 1 Jun 2016 16:32:11 +0900 Subject: [tex-live] tlmgr does not work In-Reply-To: References: <20160601071930.GO965@logic.at> Message-ID: <20160601073211.GP965@logic.at> Hi Martin, (please keep list in Cc) > tlmgr-update-latest.sh > from the tlpretest area. Output of running the program? Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From preining at logic.at Wed Jun 1 10:41:59 2016 From: preining at logic.at (Norbert Preining) Date: Wed, 1 Jun 2016 17:41:59 +0900 Subject: [tex-live] tlmgr does not work In-Reply-To: <72d03c8a-b752-e8e5-2d23-972f3166d49b@aol.com> References: <20160601071930.GO965@logic.at> <20160601073211.GP965@logic.at> <72d03c8a-b752-e8e5-2d23-972f3166d49b@aol.com> Message-ID: <7E426BAC-454A-4A33-A1A1-206BF9915B95@logic.at> Hi Martin Again, please keep the mailing list in Cc... >nothing, the window is closed very fast ... i can not read anything. Which window???? You are running Linux, right? Please open a terminal, change the directory to three place where the update-tlmgr-latest.sh is, then type bash update-tlmgr-latest.sh And copy and paste the output into an email. Norbert -- Sent from my mobile device. Please excuse my brevity. From martinfeck at aol.com Wed Jun 1 10:50:40 2016 From: martinfeck at aol.com (Martin Feck) Date: Wed, 1 Jun 2016 10:50:40 +0200 Subject: [tex-live] tlmgr does not work In-Reply-To: <84300cf1-a537-11e3-f86e-ff15bd47f06d@aol.com> References: <20160601071930.GO965@logic.at> <20160601073211.GP965@logic.at> <72d03c8a-b752-e8e5-2d23-972f3166d49b@aol.com> <7E426BAC-454A-4A33-A1A1-206BF9915B95@logic.at> <84300cf1-a537-11e3-f86e-ff15bd47f06d@aol.com> Message-ID: Hi Norbert, this is the output: xelatex Schreibtisch # bash update-tlmgr-latest.sh Verifying archive integrity...Error in MD5 checksums: 990ca70d7aa935a579a1ed0715dd2e25 is different from 3a0eb7c90a643395481f525056ae8335 xelatex Schreibtisch # Martin Feck Am 01.06.2016 um 10:46 schrieb Martin Feck: > Hi Norbert, > > > this is the output ... > > xelatex Schreibtisch # bash update-tlmgr-latest.sh > Verifying archive integrity...Error in MD5 checksums: > 990ca70d7aa935a579a1ed0715dd2e25 is different from > 3a0eb7c90a643395481f525056ae8335 > > Martin Feck > > Am 01.06.2016 um 10:41 schrieb Norbert Preining: >> Hi Martin >> >> Again, please keep the mailing list in Cc... >> >> >> >> >>> nothing, the window is closed very fast ... i can not read anything. >> Which window???? You are running Linux, right? >> >> Please open a terminal, change the directory to three place where the >> update-tlmgr-latest.sh is, then type >> bash update-tlmgr-latest.sh >> And copy and paste the output into an email. >> >> Norbert >> > -------------- next part -------------- A non-text attachment was scrubbed... Name: martinfeck.vcf Type: text/x-vcard Size: 100 bytes Desc: not available URL: From preining at logic.at Wed Jun 1 10:52:16 2016 From: preining at logic.at (Norbert Preining) Date: Wed, 1 Jun 2016 17:52:16 +0900 Subject: [tex-live] tlmgr does not work In-Reply-To: <84300cf1-a537-11e3-f86e-ff15bd47f06d@aol.com> References: <20160601071930.GO965@logic.at> <20160601073211.GP965@logic.at> <72d03c8a-b752-e8e5-2d23-972f3166d49b@aol.com> <7E426BAC-454A-4A33-A1A1-206BF9915B95@logic.at> <84300cf1-a537-11e3-f86e-ff15bd47f06d@aol.com> Message-ID: <76F18E02-608D-47DD-A03D-2886B5CC4AFD@logic.at> Your download is corrupted. Delete the file and redownload it. On June 1, 2016 5:46:42 PM GMT+09:00, Martin Feck wrote: >Hi Norbert, > > >this is the output ... > >xelatex Schreibtisch # bash update-tlmgr-latest.sh >Verifying archive integrity...Error in MD5 checksums: >990ca70d7aa935a579a1ed0715dd2e25 is different from >3a0eb7c90a643395481f525056ae8335 > >Martin Feck > >Am 01.06.2016 um 10:41 schrieb Norbert Preining: >> Hi Martin >> >> Again, please keep the mailing list in Cc... >> >> >> >> >>> nothing, the window is closed very fast ... i can not read anything. >> Which window???? You are running Linux, right? >> >> Please open a terminal, change the directory to three place where the >update-tlmgr-latest.sh is, then type >> bash update-tlmgr-latest.sh >> And copy and paste the output into an email. >> >> Norbert >> -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From martinfeck at aol.com Wed Jun 1 10:58:34 2016 From: martinfeck at aol.com (Martin Feck) Date: Wed, 1 Jun 2016 10:58:34 +0200 Subject: [tex-live] tlmgr does not work In-Reply-To: <76F18E02-608D-47DD-A03D-2886B5CC4AFD@logic.at> References: <20160601071930.GO965@logic.at> <20160601073211.GP965@logic.at> <72d03c8a-b752-e8e5-2d23-972f3166d49b@aol.com> <7E426BAC-454A-4A33-A1A1-206BF9915B95@logic.at> <84300cf1-a537-11e3-f86e-ff15bd47f06d@aol.com> <76F18E02-608D-47DD-A03D-2886B5CC4AFD@logic.at> Message-ID: <3774f6fb-00a4-f376-94fd-9a2106e33a15@aol.com> Hi Norbert, I have already done ten times, without success ... can you send me a working one, please?? Martin Feck Am 01.06.2016 um 10:52 schrieb Norbert Preining: > Your download is corrupted. Delete the file and redownload it. > > On June 1, 2016 5:46:42 PM GMT+09:00, Martin Feck wrote: >> Hi Norbert, >> >> >> this is the output ... >> >> xelatex Schreibtisch # bash update-tlmgr-latest.sh >> Verifying archive integrity...Error in MD5 checksums: >> 990ca70d7aa935a579a1ed0715dd2e25 is different from >> 3a0eb7c90a643395481f525056ae8335 >> >> Martin Feck >> >> Am 01.06.2016 um 10:41 schrieb Norbert Preining: >>> Hi Martin >>> >>> Again, please keep the mailing list in Cc... >>> >>> >>> >>> >>>> nothing, the window is closed very fast ... i can not read anything. >>> Which window???? You are running Linux, right? >>> >>> Please open a terminal, change the directory to three place where the >> update-tlmgr-latest.sh is, then type >>> bash update-tlmgr-latest.sh >>> And copy and paste the output into an email. >>> >>> Norbert >>> > -- > PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info > GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: martinfeck.vcf Type: text/x-vcard Size: 100 bytes Desc: not available URL: From martinfeck at aol.com Wed Jun 1 12:58:57 2016 From: martinfeck at aol.com (Martin Feck) Date: Wed, 1 Jun 2016 12:58:57 +0200 Subject: [tex-live] tex-live Digest, Vol 161, Issue 3 In-Reply-To: References: Message-ID: <3073b347-d11a-4037-e4ad-f2ab419fc51f@aol.com> Hi there, downloading bash file update-tlmgr-latest.sh from: http://mirror.hmc.edu/tlpretest/ solve the problem. tanks a lot Martin Feck Am 01.06.2016 um 12:00 schrieb tex-live-request at tug.org: > Send tex-live mailing list submissions to > tex-live at tug.org > > To subscribe or unsubscribe via the World Wide Web, visit > http://tug.org/mailman/listinfo/tex-live > or, via email, send a message with subject or body 'help' to > tex-live-request at tug.org > > You can reach the person managing the list at > tex-live-owner at tug.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of tex-live digest..." > > > Today's Topics: > > 1. Re: tlmgr does not work (Norbert Preining) > 2. Re: tlmgr does not work (Martin Feck) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Wed, 1 Jun 2016 17:52:16 +0900 > From: Norbert Preining > To: Martin Feck , > Subject: Re: [tex-live] tlmgr does not work > Message-ID: <76F18E02-608D-47DD-A03D-2886B5CC4AFD at logic.at> > Content-Type: text/plain; charset="utf-8" > > Your download is corrupted. Delete the file and redownload it. > > On June 1, 2016 5:46:42 PM GMT+09:00, Martin Feck wrote: >> Hi Norbert, >> >> >> this is the output ... >> >> xelatex Schreibtisch # bash update-tlmgr-latest.sh >> Verifying archive integrity...Error in MD5 checksums: >> 990ca70d7aa935a579a1ed0715dd2e25 is different from >> 3a0eb7c90a643395481f525056ae8335 >> >> Martin Feck >> >> Am 01.06.2016 um 10:41 schrieb Norbert Preining: >>> Hi Martin >>> >>> Again, please keep the mailing list in Cc... >>> >>> >>> >>> >>>> nothing, the window is closed very fast ... i can not read anything. >>> Which window???? You are running Linux, right? >>> >>> Please open a terminal, change the directory to three place where the >> update-tlmgr-latest.sh is, then type >>> bash update-tlmgr-latest.sh >>> And copy and paste the output into an email. >>> >>> Norbert >>> > -- > PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info > GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 > > > ------------------------------ > > Message: 2 > Date: Wed, 1 Jun 2016 10:58:34 +0200 > From: Martin Feck > To: Norbert Preining , tex-live at tug.org > Subject: Re: [tex-live] tlmgr does not work > Message-ID: <3774f6fb-00a4-f376-94fd-9a2106e33a15 at aol.com> > Content-Type: text/plain; charset="utf-8"; Format="flowed" > > Hi Norbert, > > > I have already done ten times, without success ... can you send me a > working one, please?? > > > Martin Feck > > > Am 01.06.2016 um 10:52 schrieb Norbert Preining: >> Your download is corrupted. Delete the file and redownload it. >> >> On June 1, 2016 5:46:42 PM GMT+09:00, Martin Feck wrote: >>> Hi Norbert, >>> >>> >>> this is the output ... >>> >>> xelatex Schreibtisch # bash update-tlmgr-latest.sh >>> Verifying archive integrity...Error in MD5 checksums: >>> 990ca70d7aa935a579a1ed0715dd2e25 is different from >>> 3a0eb7c90a643395481f525056ae8335 >>> >>> Martin Feck >>> >>> Am 01.06.2016 um 10:41 schrieb Norbert Preining: >>>> Hi Martin >>>> >>>> Again, please keep the mailing list in Cc... >>>> >>>> >>>> >>>> >>>>> nothing, the window is closed very fast ... i can not read anything. >>>> Which window???? You are running Linux, right? >>>> >>>> Please open a terminal, change the directory to three place where the >>> update-tlmgr-latest.sh is, then type >>>> bash update-tlmgr-latest.sh >>>> And copy and paste the output into an email. >>>> >>>> Norbert >>>> >> -- >> PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info >> GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 > -------------- next part -------------- > An HTML attachment was scrubbed... > URL: > -------------- next part -------------- > A non-text attachment was scrubbed... > Name: martinfeck.vcf > Type: text/x-vcard > Size: 100 bytes > Desc: not available > URL: > > ------------------------------ > > Subject: Digest Footer > > _______________________________________________ > tex-live mailing list > tex-live at tug.org > http://tug.org/mailman/listinfo/tex-live > > > ------------------------------ > > End of tex-live Digest, Vol 161, Issue 3 > **************************************** -------------- next part -------------- A non-text attachment was scrubbed... Name: martinfeck.vcf Type: text/x-vcard Size: 100 bytes Desc: not available URL: From tex at axelkielhorn.de Wed Jun 1 17:31:06 2016 From: tex at axelkielhorn.de (Axel Kielhorn) Date: Wed, 1 Jun 2016 17:31:06 +0200 Subject: [tex-live] Bug in TL infra In-Reply-To: <201605312142.u4VLgO65015749@freefriends.org> References: <201605312142.u4VLgO65015749@freefriends.org> Message-ID: <2EC444E7-5D8E-4110-806E-A590F8CE15CB@axelkielhorn.de> > Am 31.05.2016 um 23:42 schrieb Karl Berry : > > Hi Axel, > > A space in a filename is a really bad thing. > > We go to great lengths to support spaces in file/directory names in > general. Obviously we fell down when it comes to runme.sh. ?Macintosh HD? is the default name for the first volume on Mac OS since ages. This would normally not be a problem, since runme.sh uses the path: /usr/local/texlive/2016 But I have a dual boot system with TeXLive on one volume and some links on the other. When I run runme.sh on the second system the path will be turned into /Volumen/Macntosh HD/usr local/texlive/2016 Thus an ordinary user will probably not fall into that trap. This also explains why an infra update sometimes success (run on system 1) and sometimes fails (run on system 2). I didn?t realize that this was a problem since everything else works. > I'll > install your fix and work further on it after updates resume. > > - for a in $ROOT/bin/*; do > + for a in "$ROOT/bin/*"; do > > I'm guessing that should be "$ROOT"/bin/* -- want the wildcard > expansion, not a literal "*?. Ooops, you are right. Axel From preining at logic.at Thu Jun 2 05:07:27 2016 From: preining at logic.at (Norbert Preining) Date: Thu, 2 Jun 2016 12:07:27 +0900 Subject: [tex-live] Tlmgr closing takes fairly long In-Reply-To: References: <2f7dbcca-290b-9db0-9be7-4baca67e3802@gmx.net> <20160531002547.GV965@logic.at> Message-ID: <20160602030727.GA22897@logic.at> Hi Bernhard, (and everyone who is running TL on Windows) I have tested this now and I see a similar behaviour, although not as slow as on your side. BUT, I see a very fast shutdown when I do *not* load the remote tlpdb. I am not sure whether this depends on rather minimal installation I am testing on Windows or something else. Can you please start the gui, and then simply exit it again without loading hte remote dbs. Is it a fast or slow finish? And before I hear the answer - here my theory: Loading the 3000 packages and displaying them creates a lot of Tk instances in the table. They need all be destroyed, this happens automatically in the background, but it seems to take quite some time on Windows. ANyway, if those with a *full* TL installation could do the above test I would be happy. Thanks Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From bernhard.kleine at gmx.net Thu Jun 2 11:04:19 2016 From: bernhard.kleine at gmx.net (Bernhard Kleine) Date: Thu, 2 Jun 2016 11:04:19 +0200 Subject: [tex-live] Tlmgr closing takes fairly long In-Reply-To: <20160602030727.GA22897@logic.at> References: <2f7dbcca-290b-9db0-9be7-4baca67e3802@gmx.net> <20160531002547.GV965@logic.at> <20160602030727.GA22897@logic.at> Message-ID: <181af83b-2feb-8e91-8bed-468f06d540b8@gmx.net> Am 02.06.2016 um 05:07 schrieb Norbert Preining: > Hi Bernhard, (and everyone who is running TL on Windows) > > I have tested this now and I see a similar behaviour, although not > as slow as on your side. > > BUT, I see a very fast shutdown when I do *not* load the remote tlpdb. > I am not sure whether this depends on rather minimal installation > I am testing on Windows or something else. > > Can you please start the gui, and then simply exit it again without > loading hte remote dbs. Is it a fast or slow finish? > > And before I hear the answer - here my theory: Loading the 3000 packages > and displaying them creates a lot of Tk instances in the table. They > need all be destroyed, this happens automatically in the background, > but it seems to take quite some time on Windows. > > ANyway, if those with a *full* TL installation could do the above > test I would be happy. > > Thanks > > Norbert > The tlmgr shutdown after solely opening tlmgr without any activity takes 50 sec. Bernhard From josef.kleber at gmx.de Thu Jun 2 15:30:48 2016 From: josef.kleber at gmx.de (Josef Kleber) Date: Thu, 2 Jun 2016 15:30:48 +0200 Subject: [tex-live] Tlmgr closing takes fairly long In-Reply-To: <20160602030727.GA22897@logic.at> References: <2f7dbcca-290b-9db0-9be7-4baca67e3802@gmx.net> <20160531002547.GV965@logic.at> <20160602030727.GA22897@logic.at> Message-ID: <4cc1d355-0e68-9404-0113-c97243118dc2@gmx.de> Am 02.06.2016 um 05:07 schrieb Norbert Preining: > Hi Bernhard, (and everyone who is running TL on Windows) > > I have tested this now and I see a similar behaviour, although not > as slow as on your side. > > BUT, I see a very fast shutdown when I do *not* load the remote tlpdb. > I am not sure whether this depends on rather minimal installation > I am testing on Windows or something else. > > Can you please start the gui, and then simply exit it again without > loading hte remote dbs. Is it a fast or slow finish? Here (Win10, TL2016pretest), it's about 25-30 seconds in both cases. Loading remote tlpdb doesn't matter! Josef -- PGP Public Key: http://kirza.kleberj.de/pgp DAA2 4AA6 5157 AD05 F6B8 043A D1BE 1EEA 86C6 4C5F -------------- next part -------------- A non-text attachment was scrubbed... Name: 0x86C64C5F.asc Type: application/pgp-keys Size: 4725 bytes Desc: not available URL: From josef.kleber at gmx.de Thu Jun 2 15:37:49 2016 From: josef.kleber at gmx.de (Josef Kleber) Date: Thu, 2 Jun 2016 15:37:49 +0200 Subject: [tex-live] Tlmgr closing takes fairly long In-Reply-To: <4cc1d355-0e68-9404-0113-c97243118dc2@gmx.de> References: <2f7dbcca-290b-9db0-9be7-4baca67e3802@gmx.net> <20160531002547.GV965@logic.at> <20160602030727.GA22897@logic.at> <4cc1d355-0e68-9404-0113-c97243118dc2@gmx.de> Message-ID: <4907c2a2-e297-7274-7d74-cd3eba1f612b@gmx.de> Am 02.06.2016 um 15:30 schrieb Josef Kleber: > Here (Win10, TL2016pretest), it's about 25-30 seconds in both cases. > Loading remote tlpdb doesn't matter! BTW, it's the same with TL2015! I don't have older TLs installed. Josef -- PGP Public Key: http://kirza.kleberj.de/pgp DAA2 4AA6 5157 AD05 F6B8 043A D1BE 1EEA 86C6 4C5F -------------- next part -------------- A non-text attachment was scrubbed... Name: 0x86C64C5F.asc Type: application/pgp-keys Size: 4725 bytes Desc: not available URL: From siepo at cybercomm.nl Thu Jun 2 16:05:52 2016 From: siepo at cybercomm.nl (Siep Kroonenberg) Date: Thu, 2 Jun 2016 16:05:52 +0200 Subject: [tex-live] Unable to use install-tl-windows.exe: "Windows XP earliest supported version; aborting..." In-Reply-To: References: Message-ID: <20160602140552.GA6399@bitmuis.thuis.internal> On Tue, May 31, 2016 at 02:06:52PM -0500, Jon Hood wrote: > When trying to run install-tl-windows.exe from > https://www.tug.org/texlive/acquire-netinstall.html I receive an > immediate pop-up stating, "Windows XP earliest supported version; > aborting..." > > My platform: Windows 7 Enterprise Service Pack 1 64-bit > > Current workaround: open install-tl-windows.exe in 7zip and extract > the install-tl-20160405 folder. Run install-tl-windows.bat from the > extracted folder > > Thanks, > Jon I cannot reproduce this with a freshly-downloaded install-tl-windows.exe on Windows 10 64-bit, and have no idea what may have caused this. -- Siep Kroonenberg From ziegenhagen at gmail.com Thu Jun 2 16:13:50 2016 From: ziegenhagen at gmail.com (Uwe Ziegenhagen) Date: Thu, 2 Jun 2016 16:13:50 +0200 Subject: [tex-live] Unable to use install-tl-windows.exe: "Windows XP earliest supported version; aborting..." In-Reply-To: <20160602140552.GA6399@bitmuis.thuis.internal> References: <20160602140552.GA6399@bitmuis.thuis.internal> Message-ID: How does the Script determine the windows version it is running on? I would assume windows 7 Enterprise is using a Different identifier than win7 home Or Professional. Uwe Von meinem iPhone gesendet > Am 02.06.2016 um 16:05 schrieb Siep Kroonenberg : > >> On Tue, May 31, 2016 at 02:06:52PM -0500, Jon Hood wrote: >> When trying to run install-tl-windows.exe from >> https://www.tug.org/texlive/acquire-netinstall.html I receive an >> immediate pop-up stating, "Windows XP earliest supported version; >> aborting..." >> >> My platform: Windows 7 Enterprise Service Pack 1 64-bit >> >> Current workaround: open install-tl-windows.exe in 7zip and extract >> the install-tl-20160405 folder. Run install-tl-windows.bat from the >> extracted folder >> >> Thanks, >> Jon > > I cannot reproduce this with a freshly-downloaded > install-tl-windows.exe on Windows 10 64-bit, and have no idea what > may have caused this. > > -- > Siep Kroonenberg From preining at logic.at Thu Jun 2 17:25:25 2016 From: preining at logic.at (Norbert Preining) Date: Fri, 3 Jun 2016 00:25:25 +0900 Subject: [tex-live] Tlmgr closing takes fairly long In-Reply-To: <4cc1d355-0e68-9404-0113-c97243118dc2@gmx.de> <181af83b-2feb-8e91-8bed-468f06d540b8@gmx.net> Message-ID: <20160602152525.GX965@logic.at> On Thu, 02 Jun 2016, Bernhard Kleine wrote: > The tlmgr shutdown after solely opening tlmgr without any activity takes > 50 sec. On Thu, 02 Jun 2016, Josef Kleber wrote: > Here (Win10, TL2016pretest), it's about 25-30 seconds in both cases. > Loading remote tlpdb doesn't matter! Do both of you have texlive-full installed? This is my guess ... Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From siepo at cybercomm.nl Thu Jun 2 17:27:15 2016 From: siepo at cybercomm.nl (Siep Kroonenberg) Date: Thu, 2 Jun 2016 17:27:15 +0200 Subject: [tex-live] Unable to use install-tl-windows.exe: "Windows XP earliest supported version; aborting..." In-Reply-To: References: <20160602140552.GA6399@bitmuis.thuis.internal> Message-ID: <20160602152715.GB6399@bitmuis.thuis.internal> On Thu, Jun 02, 2016 at 04:13:50PM +0200, Uwe Ziegenhagen wrote: > How does the Script determine the windows version it is running on? I would assume windows 7 Enterprise is using a Different identifier than win7 home Or Professional. The edition (home/pro/enterprise) is independent from the version (xp/vista/w7/w8/w10). The nsis wrapper script loads the WinVer.nsh header file for testing the version, not the edition. This header file does have special code for server versions but that is another story. I just realized that I have remote access to a windows 7 enterprise. On that system, too, I could start up the `real' installer. I am reluctant to spend a lot of time on this: there have been no recent changes in this wrapper, I recall no similar error reports, I cannot reproduce the problem, and the affected user found a workaround for himself (another workaround is downloading the zip version). -- Siep Kroonenberg From josef.kleber at gmx.de Thu Jun 2 17:47:23 2016 From: josef.kleber at gmx.de (Josef Kleber) Date: Thu, 2 Jun 2016 17:47:23 +0200 Subject: [tex-live] Tlmgr closing takes fairly long In-Reply-To: <20160602152525.GX965@logic.at> References: <20160602152525.GX965@logic.at> Message-ID: <4317976b-fc9f-e97f-95fe-bfa0e08e1722@gmx.de> Am 02.06.2016 um 17:25 schrieb Norbert Preining: > On Thu, 02 Jun 2016, Bernhard Kleine wrote: >> The tlmgr shutdown after solely opening tlmgr without any activity takes >> 50 sec. > > On Thu, 02 Jun 2016, Josef Kleber wrote: >> Here (Win10, TL2016pretest), it's about 25-30 seconds in both cases. >> Loading remote tlpdb doesn't matter! > > Do both of you have texlive-full installed? > Yes, I have installed texlive-full! Josef -- PGP Public Key: http://kirza.kleberj.de/pgp DAA2 4AA6 5157 AD05 F6B8 043A D1BE 1EEA 86C6 4C5F -------------- next part -------------- A non-text attachment was scrubbed... Name: 0x86C64C5F.asc Type: application/pgp-keys Size: 4725 bytes Desc: not available URL: From squinky86 at gmail.com Thu Jun 2 23:27:35 2016 From: squinky86 at gmail.com (Jon Hood) Date: Thu, 2 Jun 2016 16:27:35 -0500 Subject: [tex-live] Unable to use install-tl-windows.exe: "Windows XP earliest supported version; aborting..." In-Reply-To: <20160602152715.GB6399@bitmuis.thuis.internal> References: <20160602140552.GA6399@bitmuis.thuis.internal> <20160602152715.GB6399@bitmuis.thuis.internal> Message-ID: Can you point me to the code that does the detection and I can try to debug it on my end? Thanks! Jon On 6/2/2016 10:27 AM, Siep Kroonenberg wrote: > On Thu, Jun 02, 2016 at 04:13:50PM +0200, Uwe Ziegenhagen wrote: >> How does the Script determine the windows version it is running on? I would assume windows 7 Enterprise is using a Different identifier than win7 home Or Professional. > The edition (home/pro/enterprise) is independent from the version > (xp/vista/w7/w8/w10). The nsis wrapper script loads the WinVer.nsh > header file for testing the version, not the edition. This header > file does have special code for server versions but that is another > story. > > I just realized that I have remote access to a windows 7 enterprise. > On that system, too, I could start up the `real' installer. > > I am reluctant to spend a lot of time on this: there have been no > recent changes in this wrapper, I recall no similar error reports, I > cannot reproduce the problem, and the affected user found a > workaround for himself (another workaround is downloading the zip > version). > From siepo at cybercomm.nl Fri Jun 3 11:39:19 2016 From: siepo at cybercomm.nl (Siep Kroonenberg) Date: Fri, 3 Jun 2016 11:39:19 +0200 Subject: [tex-live] Unable to use install-tl-windows.exe: "Windows XP earliest supported version; aborting..." In-Reply-To: References: <20160602140552.GA6399@bitmuis.thuis.internal> <20160602152715.GB6399@bitmuis.thuis.internal> Message-ID: <20160603093919.GA2329@bitmuis.thuis.internal> On Thu, Jun 02, 2016 at 04:27:35PM -0500, Jon Hood wrote: > Can you point me to the code that does the detection and I can try to debug > it on my end? > > Thanks! The nsis script: http://www.tug.org/svn/texlive/trunk/Master/tlpkg/libexec/install-tl.nsi?revision=33535&view=markup Nsis is available as Debian/Ubuntu package. The version on my system is 2.50-1, but what matters is the nsis version on the TUG site which compiles the nsis installer as part of the nightly updates, and which may be a bit older. A Windows version can be downloaded from http://nsis.sourceforge.net/ The header file in question is WinVer.nsh. -- Siep Kroonenberg From karl at freefriends.org Fri Jun 3 23:47:58 2016 From: karl at freefriends.org (Karl Berry) Date: Fri, 3 Jun 2016 21:47:58 +0000 Subject: [tex-live] Unable to use install-tl-windows.exe: "Windows XP earliest supported version; aborting..." In-Reply-To: <20160603093919.GA2329@bitmuis.thuis.internal> Message-ID: <201606032147.u53LlwOU016841@freefriends.org> what matters is the nsis version on the TUG site Which is 2.46, looks like. As I recall, we've never had a reason to update it since we first installed it. Can do so if it'll help ... -karl From d.p.carlisle at gmail.com Sun Jun 5 23:09:29 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Sun, 5 Jun 2016 22:09:29 +0100 Subject: [tex-live] pdftex segfault with \pdflastmatch Message-ID: This plain tex document demonstrates \pdflastmatch failing if used after a \pdfmatch that returned no matches. \pdfmatch{a}{b} \pdflastmatch0 \bye segfaults (for me on cygwin tl2016 pretest) and originally reported by Henri Menke in stackexchange chat on Debian linux TL 2015. he commented > Ohhhh. Could also be a libc bug. Because it works in GDB. David -------------- next part -------------- An HTML attachment was scrubbed... URL: From d.p.carlisle at gmail.com Sun Jun 5 23:20:16 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Sun, 5 Jun 2016 22:20:16 +0100 Subject: [tex-live] pdftex segfault with \pdflastmatch In-Reply-To: References: Message-ID: a gdb trace of the core dump, and valgrind reports are available in chat around http://chat.stackexchange.com/transcript/message/30145736#30145736 David -------------- next part -------------- An HTML attachment was scrubbed... URL: From kakuto at fuk.kindai.ac.jp Mon Jun 6 00:15:42 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Mon, 6 Jun 2016 07:15:42 +0900 Subject: [tex-live] pdftex segfault with \pdflastmatch In-Reply-To: References: Message-ID: <248EACE4D9714086BE38F6AEA223AE07@CJ3001517A> Hi David, > This plain tex document demonstrates \pdflastmatch failing if used after > a \pdfmatch that returned no matches. The crash disappears if I apply an attached utils.c.diff. But I don't know that the patch is right or not. Best, Akira -------------- next part -------------- A non-text attachment was scrubbed... Name: utils.c.diff Type: application/octet-stream Size: 482 bytes Desc: not available URL: From bernhard.kleine at gmx.net Mon Jun 6 09:15:29 2016 From: bernhard.kleine at gmx.net (Bernhard Kleine) Date: Mon, 6 Jun 2016 09:15:29 +0200 Subject: [tex-live] Tlmgr closing takes fairly long In-Reply-To: <20160602152525.GX965@logic.at> References: <20160602152525.GX965@logic.at> Message-ID: Am 02.06.2016 um 17:25 schrieb Norbert Preining: > On Thu, 02 Jun 2016, Bernhard Kleine wrote: >> The tlmgr shutdown after solely opening tlmgr without any activity takes >> 50 sec. > On Thu, 02 Jun 2016, Josef Kleber wrote: >> Here (Win10, TL2016pretest), it's about 25-30 seconds in both cases. >> Loading remote tlpdb doesn't matter! > Do both of you have texlive-full installed? > > This is my guess ... > > Norbert > I have to report that after the full install of texlive2016 closing is down to about 10 sec. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: OpenPGP digital signature URL: From preining at logic.at Mon Jun 6 16:38:10 2016 From: preining at logic.at (Norbert Preining) Date: Mon, 6 Jun 2016 23:38:10 +0900 Subject: [tex-live] Tlmgr closing takes fairly long In-Reply-To: References: <20160602152525.GX965@logic.at> Message-ID: <20160606143810.GC965@logic.at> Hi Bernhard, > >> The tlmgr shutdown after solely opening tlmgr without any activity takes > >> 50 sec. > I have to report that after the full install of texlive2016 closing is > down to about 10 sec. That is very very very *very* strange and I honestly believe there have been other reasons for the long delay. Increasing the number of packages by installing texlive-full can only *increase* the time as more objects/widgets are created and need to be destroyed at the end. So my theory is that you had some background process taking CPU time or IO time (backup service, update service, whatever service) and thus the long delay appeared. But as a non-Windows user (most of the time) I am open to different explanations. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From bernhard.kleine at gmx.net Mon Jun 6 16:39:21 2016 From: bernhard.kleine at gmx.net (Bernhard Kleine) Date: Mon, 6 Jun 2016 16:39:21 +0200 Subject: [tex-live] installation failure Message-ID: <201cf44b-1d12-effe-5fa9-1a3876ee6014@gmx.net> Since Texlive 2016 was announced in the newsgroup I started to install it: Installation failed: F:\Downloads\Software\texlive\install-tl\install-tl-20160405>install-tl-advanced -repository http://mirror.ctan.org/systems/texlive/tlnet PATH=F:\Downloads\Software\texlive\install-tl\install-tl-20160405\tlpkg\tlperl\bin;;C:\Program Files (x86)\AMD APP\bin\x86;C:\ProgramData\Oracle\Java\javapath;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0\;C:\Program Files (x86)\AMD\ATI.ACE\Core-Static;C:\Program Files (x86)\GNU\GnuPG\pub "F:\Downloads\Software\texlive\install-tl\install-tl-20160405\install-tl" -gui perltk -repository http://mirror.ctan.org/systems/texlive/tlnet Loading ftp://ftp.fu-berlin.de/tex/CTAN/systems/texlive/tlnet/tlpkg/texlive.tlpdb Unknown directive ...containerchecksum c59200574a316416a23695c258edf3a32531fbda43ccdc09360ee105c3f07f9fb77df17c4ba4c2ea4f3a5ea6667e064b51e3d8c2fe6c984ba F:/Downloads/Software/texlive/install-tl/install-tl-20160405/tlpkg/TeXLive/TLPOBJ.pm line 210, <$retfh> line 5579. Dr?cken Sie eine beliebige Taste . . . Gnupg installed see Path Regards Bernhard -- -- spitzhalde9 79853 lenzkirch bernhard.kleine at gmx.net www.b-kleine.com, www.urseetal.net - thunderbird mit enigmail GPG schl?ssel: D5257409 fingerprint: 08 B7 F8 70 22 7A FC C1 15 49 CA A6 C7 6F A0 2E D5 25 74 09 -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: OpenPGP digital signature URL: From usenet at siart.de Mon Jun 6 16:51:09 2016 From: usenet at siart.de (Uwe Siart) Date: Mon, 6 Jun 2016 16:51:09 +0200 Subject: [tex-live] installation failure References: <201cf44b-1d12-effe-5fa9-1a3876ee6014@gmx.net> Message-ID: <84wpm2weua.fsf@siart.de> Bernhard Kleine writes: > F:\Downloads\Software\texlive\install-tl\install-tl-20160405>install-tl-advanced Your mirror isn't yet up to date. 20160405 points to the freeze date of TL15. -- Uwe From preining at logic.at Mon Jun 6 16:55:32 2016 From: preining at logic.at (Norbert Preining) Date: Mon, 6 Jun 2016 23:55:32 +0900 Subject: [tex-live] installation failure In-Reply-To: <201cf44b-1d12-effe-5fa9-1a3876ee6014@gmx.net> References: <201cf44b-1d12-effe-5fa9-1a3876ee6014@gmx.net> Message-ID: <20160606145532.GD965@logic.at> On Mon, 06 Jun 2016, Bernhard Kleine wrote: > F:\Downloads\Software\texlive\install-tl\install-tl-20160405>install-tl-advanced Your install-tl is too old. This is not the one released with TL2016. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From preining at logic.at Mon Jun 6 17:34:16 2016 From: preining at logic.at (Norbert Preining) Date: Tue, 7 Jun 2016 00:34:16 +0900 Subject: [tex-live] pdftex segfault with \pdflastmatch In-Reply-To: <248EACE4D9714086BE38F6AEA223AE07@CJ3001517A> Message-ID: <20160606153416.GE965@logic.at> Hi > This plain tex document demonstrates \pdflastmatch failing if used after > a \pdfmatch that returned no matches. I confirm that the same happens on x86_64-linux. > The crash disappears if I apply an attached utils.c.diff. > But I don't know that the patch is right or not. I also confirm that the crash disappears, but for the mentioned document: \pdfmatch{a}{b} \pdflastmatch0 \bye I get a strange output of 0 -1- Not sure if this is supposed to be. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From squinky86 at gmail.com Mon Jun 6 19:07:05 2016 From: squinky86 at gmail.com (Jon Hood) Date: Mon, 6 Jun 2016 12:07:05 -0500 Subject: [tex-live] [NSIS-devel] Fwd: Unable to use install-tl-windows.exe: "Windows XP earliest supported version; aborting..." In-Reply-To: References: <20160602140552.GA6399@bitmuis.thuis.internal> Message-ID: This is occurring on the Army Gold Master image for Windows 7, and thanks to your input, I've been able to track it down: * Force XP compatbility mode in the installers Properties->Compatbility * Rename the installer * Move the installer to C:\ * Use another user (preferably with only English letters in its name) * Try on a different Windows 7 computer Same result on all systems with this image. * Disable your AV * Run the version test suite on your -- https://github.com/kichik/nsis/blob/51e74df6c950cd8176ded182d09fc2553481f1ba/Source/Tests/winver.nsi Bingo - that got it. The issue is with McAfee Host Intrusion Prevention and McAfee VirusScan. Winver copies dll files to %TEMP% including a system.dll file. It then begins executing within a new folder in %TEMP%. McAfee configuration information is available here: https://kc.mcafee.com/resources/sites/MCAFEE/content/live/PRODUCT_DOCUMENTATION/22000/PD22818/en_US/Access_Protection_Rules.pdf p. 6 starting with the section ?Prevent all programs from running files from the Temp folder? And the STIG rule for the fully unclassified side is available at: https://www.stigviewer.com/stig/mcafee_virusscan_8.8_managed_client/2015-06-18/finding/V-59363 So, the moral of the story: NSIS version checking does not work on a DoD-STIG'd system using McAfee VirusScan or HIP. I'm cross-posting this on the texlive list so that their developers know that the issue has been tracked down, and it's an issue with the U.S. DoD + McAfee, not with NSIS. Thanks again, Jon On Mon, Jun 6, 2016 at 11:37 AM, Amir Szekely wrote: > We use relative version comparison to avoid issues like the one you're > describing. There is no list where Windows 7 SP1 is missing. I have > developed this feature on Windows SP1 and have worked with it for the past 3 > years without seeing this issue. Yours is also the first report of such an > issue. I am therefore under the working assumption that this is a local > issue with your computer or that specific installer. > > Here is a list of a few things you can try to pinpoint the issue: > > * Force XP compatbility mode in the installers Properties->Compatbility > * Rename the installer > * Move the installer to C:\ > * Disable your AV > * Use another user (preferably with only English letters in its name) > * Try on a different Windows 7 computer > * Run the version test suite on your -- > https://github.com/kichik/nsis/blob/51e74df6c950cd8176ded182d09fc2553481f1ba/Source/Tests/winver.nsi > > On Mon, Jun 6, 2016 at 7:43 AM, Jon Hood wrote: >> >> I can confirm that the installer is not being run in compatibility mode. >> >> C:\>ver >> Microsoft Windows [Version 6.1.7601] >> >> On Sat, Jun 4, 2016 at 4:10 PM, Amir Szekely wrote: >> > Sounds like you're running your installer in Windows 2000 compatibility >> > mode. Right click the executable, go to the Compatibility tab and make >> > sure >> > everything is unchecked. >> > >> > On Fri, Jun 3, 2016 at 12:14 PM, Jon Hood wrote: >> >> >> >> I had the following exchange with the texlive people who use nsis and >> >> downloaded the latest nsis subversion trunk. The Windows 7 Enterprise >> >> w/ SP 1 version appears to be missing from Include/WinVer.nsh: >> >> >> >> Mine: "Microsoft Windows NT 6.1.7601 Service Pack 1" >> >> >> >> Can you please make sure Windows 7 with Service Pack 1 (version >> >> 6.1.7601) is supported with nsis so that ${AtMostWin2000} evaluates >> >> that I'm using at least Windows XP? (line 72 of >> >> >> >> >> >> http://www.tug.org/svn/texlive/trunk/Master/tlpkg/libexec/install-tl.nsi?revision=33535&view=markup >> >> begins the block that I'm tripping up on). >> >> >> >> Additional reference for Windows 7 version: >> >> http://www.jrsoftware.org/ishelp/index.php?topic=winvernotes >> >> >> >> Thanks! >> >> Jon >> >> >> >> ---------- Forwarded message ---------- >> >> From: Siep Kroonenberg >> >> Date: Thu, Jun 2, 2016 at 9:05 AM >> >> Subject: Re: [tex-live] Unable to use install-tl-windows.exe: "Windows >> >> XP earliest supported version; aborting..." >> >> To: Jon Hood >> >> Cc: tex-live at tug.org >> >> >> >> >> >> On Tue, May 31, 2016 at 02:06:52PM -0500, Jon Hood wrote: >> >> > When trying to run install-tl-windows.exe from >> >> > https://www.tug.org/texlive/acquire-netinstall.html I receive an >> >> > immediate pop-up stating, "Windows XP earliest supported version; >> >> > aborting..." >> >> > >> >> > My platform: Windows 7 Enterprise Service Pack 1 64-bit >> >> > >> >> > Current workaround: open install-tl-windows.exe in 7zip and extract >> >> > the install-tl-20160405 folder. Run install-tl-windows.bat from the >> >> > extracted folder >> >> > >> >> > Thanks, >> >> > Jon >> >> >> >> I cannot reproduce this with a freshly-downloaded >> >> install-tl-windows.exe on Windows 10 64-bit, and have no idea what >> >> may have caused this. >> >> >> >> -- >> >> Siep Kroonenberg >> >> >> >> >> >> >> >> ------------------------------------------------------------------------------ >> >> What NetFlow Analyzer can do for you? Monitors network bandwidth and >> >> traffic >> >> patterns at an interface-level. Reveals which users, apps, and >> >> protocols >> >> are >> >> consuming the most bandwidth. Provides multi-vendor support for >> >> NetFlow, >> >> J-Flow, sFlow and other flows. Make informed decisions using capacity >> >> planning reports. >> >> https://ad.doubleclick.net/ddm/clk/305295220;132659582;e >> >> _______________________________________________ >> >> Nsis-devel mailing list >> >> Nsis-devel at lists.sourceforge.net >> >> https://lists.sourceforge.net/lists/listinfo/nsis-devel >> > >> > > > From thomasgmeyer at online.de Mon Jun 6 19:55:08 2016 From: thomasgmeyer at online.de (Thomas Meyer) Date: Mon, 6 Jun 2016 19:55:08 +0200 Subject: [tex-live] LuaLaTex MacTex 2016 doesn't work Message-ID: <023373fb-d377-334e-f5b9-8b8ab60ca2ef@online.de> Hi, I tried today installed MacTex 2016 with Texshop 3.62 on iMac (El Capitan), but I got only an error message! Example: \documentclass[12pt,a4paper]{scrartcl} \begin{document} Hello \end{document} Message: This is LuaTeX, Version 0.95.0 (TeX Live 2016) restricted system commands enabled. (./HelloTest.tex LaTeX2e <2016/03/31> Babel <3.9r> and hyphenation patterns for 1 language(s) loaded. (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/scrartcl.cls Document Class: scrartcl 2016/05/10 v3.20 KOMA-Script document class (article) (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/scrkbase.sty (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/scrbase.sty (/usr/local/texlive/2016/texmf-dist/tex/latex/graphics/keyval.sty) (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/scrlfile.sty Package scrlfile, 2016/05/10 v3.20 KOMA-Script package (loading files) Copyright (C) Markus Kohm ))) (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/tocbasic.sty) (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/scrsize12pt.clo) (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/typearea.sty Package typearea, 2016/05/10 v3.20 KOMA-Script package (type area) Copyright (C) Frank Neukam, 1992-1994 Copyright (C) Markus Kohm, 1994- )) (./HelloTest.aux) ./HelloTest.tex:3: Missing number, treated as zero. p l.3 \begin{document} ? x 443 words of node memory still in use: 2 hlist, 1 rule, 1 local_par, 1 dir, 5 glue, 10 glyph, 44 glue_spec, 1 temp, 11 if_stack, 1 write nodes avail lists: 2:9,3:1,4:1,5:4,6:18,7:2,8:1,9:2,11:3 warning (pdf backend): no pages of output. Transcript written on HelloTest.log. What's going wrong? Greetings Thomas -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 841 bytes Desc: OpenPGP digital signature URL: From joseph.wright at morningstar2.co.uk Mon Jun 6 20:05:28 2016 From: joseph.wright at morningstar2.co.uk (Joseph Wright) Date: Mon, 6 Jun 2016 19:05:28 +0100 Subject: [tex-live] LuaLaTex MacTex 2016 doesn't work In-Reply-To: <023373fb-d377-334e-f5b9-8b8ab60ca2ef@online.de> References: <023373fb-d377-334e-f5b9-8b8ab60ca2ef@online.de> Message-ID: <8490f4b1-4010-fd33-b360-e412125aaea3@morningstar2.co.uk> On 06/06/2016 18:55, Thomas Meyer wrote: > Hi, > > I tried today installed MacTex 2016 with Texshop 3.62 on iMac (El > Capitan), but I got only an error message! > > > Example: > > \documentclass[12pt,a4paper]{scrartcl} > \begin{document} > Hello > \end{document} > > Message: > > This is LuaTeX, Version 0.95.0 (TeX Live 2016) > restricted system commands enabled. > (./HelloTest.tex > LaTeX2e <2016/03/31> > Babel <3.9r> and hyphenation patterns for 1 language(s) loaded. > (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/scrartcl.cls > Document Class: scrartcl 2016/05/10 v3.20 KOMA-Script document class > (article) > (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/scrkbase.sty > (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/scrbase.sty > (/usr/local/texlive/2016/texmf-dist/tex/latex/graphics/keyval.sty) > (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/scrlfile.sty > Package scrlfile, 2016/05/10 v3.20 KOMA-Script package (loading files) > Copyright (C) Markus Kohm > > ))) (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/tocbasic.sty) > (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/scrsize12pt.clo) > (/usr/local/texlive/2016/texmf-dist/tex/latex/koma-script/typearea.sty > Package typearea, 2016/05/10 v3.20 KOMA-Script package (type area) > Copyright (C) Frank Neukam, 1992-1994 > Copyright (C) Markus Kohm, 1994- > > )) (./HelloTest.aux) > ./HelloTest.tex:3: Missing number, treated as zero. > > p > l.3 \begin{document} > > ? x > 443 words of node memory still in use: > 2 hlist, 1 rule, 1 local_par, 1 dir, 5 glue, 10 glyph, 44 glue_spec, > 1 temp, > 11 if_stack, 1 write nodes > avail lists: 2:9,3:1,4:1,5:4,6:18,7:2,8:1,9:2,11:3 > > warning (pdf backend): no pages of output. > Transcript written on HelloTest.log. > > What's going wrong? > > Greetings > Thomas > > Various changes in LuaTeX mean other material has to catch up. For the present \RequirePackage{luatex85} right at the start should address things. Joseph From komascript at gmx.info Mon Jun 6 20:09:40 2016 From: komascript at gmx.info (Markus Kohm) Date: Mon, 6 Jun 2016 20:09:40 +0200 Subject: [tex-live] LuaLaTex MacTex 2016 doesn't work In-Reply-To: <023373fb-d377-334e-f5b9-8b8ab60ca2ef@online.de> References: <023373fb-d377-334e-f5b9-8b8ab60ca2ef@online.de> Message-ID: <2382684.l2dEdiHFUk@x61s.fritz.box> Am Montag, 6. Juni 2016, 19:55:08 schrieb Thomas Meyer: > What's going wrong? This is not a TeX Live bug, it's a known KOMA-Script bug: ? Markus --- Infos zu LaTeX: Infos zu KOMA-Script: Aktuelles KOMA-Script: KOMA-Script-Anleitung: -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 181 bytes Desc: This is a digitally signed message part. URL: From d.p.carlisle at gmail.com Mon Jun 6 21:43:04 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Mon, 6 Jun 2016 20:43:04 +0100 Subject: [tex-live] pdftex segfault with \pdflastmatch In-Reply-To: <20160606153416.GE965@logic.at> References: <248EACE4D9714086BE38F6AEA223AE07@CJ3001517A> <20160606153416.GE965@logic.at> Message-ID: On 6 June 2016 at 16:34, Norbert Preining wrote: > Hi > > > This plain tex document demonstrates \pdflastmatch failing if used after > > a \pdfmatch that returned no matches. > > I confirm that the same happens on x86_64-linux. > > > The crash disappears if I apply an attached utils.c.diff. > > But I don't know that the patch is right or not. > > I also confirm that the crash disappears, but for the mentioned > document: > \pdfmatch{a}{b} > \pdflastmatch0 > \bye > I get a strange output of > 0 -1- > > Not sure if this is supposed to be. > > Norbert > that is I think the correct output the -> comes out as an inverted question mark unless you add \tt somewhere because cmr10 encoding is entertaining, so that is 0 from \pdfmatch (no match) -1 from \pdflastmatch0 (non existing index of start of requested match) -> the non existing string matched. David -------------- next part -------------- An HTML attachment was scrubbed... URL: From uli.laube at kabelmail.de Mon Jun 6 21:22:20 2016 From: uli.laube at kabelmail.de (Uli Laube) Date: Mon, 6 Jun 2016 21:22:20 +0200 Subject: [tex-live] TL2016 algorithm2e.sty encoding Message-ID: Hi! I just gave TL2016 a try. One of my documents uses algorithm2e.sty. Running it through LuaLaTex gives ! String contains an invalid utf-8 sequence. l.1311 \renewcommand{\algorithmcflinename}{ ?izizgi}% ! String contains an invalid utf-8 sequence. l.1313 \renewcommand{\@algocf at procname}{Prosed r}% ! String contains an invalid utf-8 sequence. l.1315 \renewcommand{\procedureautorefname}{prosed r}% ) Checking with $ file algorithm2e.sty algorithm2e.sty: ISO-8859 text hints indeed an encoding problem. Converting it to utf-8 and placing it in texmf-local with a run of mktexlsr, fixes the messages above. Kind regards Uli Laube From alan at alphabyte.co.nz Mon Jun 6 22:47:18 2016 From: alan at alphabyte.co.nz (Alan Litchfield) Date: Tue, 7 Jun 2016 08:47:18 +1200 Subject: [tex-live] TL2016 algorithm2e.sty encoding In-Reply-To: References: Message-ID: Hmm. When running a file in LaTeX we got the same error. We already had \usepackage[utf8]{inputenc} Probably not associated with TeXLive but a problem with the package. Have added the package maintainer to the thread. Regards Alan On 7/06/16 7:22 am, Uli Laube wrote: > Hi! > > I just gave TL2016 a try. One of my documents uses algorithm2e.sty. > Running it through LuaLaTex gives > > > ! String contains an invalid utf-8 sequence. > l.1311 \renewcommand{\algorithmcflinename}{ > ?izizgi}% > ! String contains an invalid utf-8 sequence. > l.1313 \renewcommand{\@algocf at procname}{Prosed > r}% > ! String contains an invalid utf-8 sequence. > l.1315 \renewcommand{\procedureautorefname}{prosed > r}% > ) > > > Checking with > > $ file algorithm2e.sty > algorithm2e.sty: ISO-8859 text > > hints indeed an encoding problem. > Converting it to utf-8 and placing it in texmf-local with a run of > mktexlsr, fixes the messages above. > > > Kind regards > Uli Laube -- Dr Alan Litchfield AlphaByte PO Box 1941 Auckland, New Zealand 1140 From alan at alphabyte.co.nz Mon Jun 6 22:51:25 2016 From: alan at alphabyte.co.nz (Alan Litchfield) Date: Tue, 7 Jun 2016 08:51:25 +1200 Subject: [tex-live] TL2016 algorithm2e.sty encoding In-Reply-To: References: Message-ID: <0227fc61-01c5-483c-565c-7453469d70c4@alphabyte.co.nz> Package maintainer? Or maybe not... The following message to was undeliverable. The reason for the problem: 5.1.0 - Unknown address error 550-'5.1.1 : Recipient address rejected: User unknown' Reporting-MTA: dns; mx1.univ-montp1.fr Final-Recipient: rfc822;cfiorio at umontpellier.fr Action: failed Status: 5.0.0 (permanent failure) Remote-MTA: dns; [193.51.152.128] Diagnostic-Code: smtp; 5.1.0 - Unknown address error 550-'5.1.1 : Recipient address rejected: User unknown' (delivery attempts: 0) On 7/06/16 8:47 am, Alan Litchfield wrote: > Hmm. > > When running a file in LaTeX we got the same error. We already had > > \usepackage[utf8]{inputenc} > > Probably not associated with TeXLive but a problem with the package. > Have added the package maintainer to the thread. > > Regards > Alan > > On 7/06/16 7:22 am, Uli Laube wrote: >> Hi! >> >> I just gave TL2016 a try. One of my documents uses algorithm2e.sty. >> Running it through LuaLaTex gives >> >> >> ! String contains an invalid utf-8 sequence. >> l.1311 \renewcommand{\algorithmcflinename}{ >> ?izizgi}% >> ! String contains an invalid utf-8 sequence. >> l.1313 \renewcommand{\@algocf at procname}{Prosed >> r}% >> ! String contains an invalid utf-8 sequence. >> l.1315 \renewcommand{\procedureautorefname}{prosed >> r}% >> ) >> >> >> Checking with >> >> $ file algorithm2e.sty >> algorithm2e.sty: ISO-8859 text >> >> hints indeed an encoding problem. >> Converting it to utf-8 and placing it in texmf-local with a run of >> mktexlsr, fixes the messages above. >> >> >> Kind regards >> Uli Laube > -- Dr Alan Litchfield AlphaByte PO Box 1941 Auckland, New Zealand 1140 From d.p.carlisle at gmail.com Mon Jun 6 22:52:29 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Mon, 6 Jun 2016 21:52:29 +0100 Subject: [tex-live] TL2016 algorithm2e.sty encoding In-Reply-To: References: Message-ID: Note this is a package bug, so can't be fixed by texlive, this has come up a few times (I reported it to the algrothm2e author and the algorithm2e-discussion at lirmm.fr list back in May, and I think others have reported it too) the fix is to use markup such as \c{c} in the package so it works with all encodings, if you just re-encode to UTF-8 it will only work in UTF8 and will give a similar error for documents using a 1-byte encoding. David -------------- next part -------------- An HTML attachment was scrubbed... URL: From d.p.carlisle at gmail.com Mon Jun 6 23:05:06 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Mon, 6 Jun 2016 22:05:06 +0100 Subject: [tex-live] TL2016 algorithm2e.sty encoding In-Reply-To: <0227fc61-01c5-483c-565c-7453469d70c4@alphabyte.co.nz> References: <0227fc61-01c5-483c-565c-7453469d70c4@alphabyte.co.nz> Message-ID: There is a probably newer address here https://www.lirmm.fr/lirmm_eng/users/utilisateurs-lirmm/christophe-fiorio David -------------- next part -------------- An HTML attachment was scrubbed... URL: From karl at freefriends.org Mon Jun 6 23:09:15 2016 From: karl at freefriends.org (Karl Berry) Date: Mon, 06 Jun 2016 21:09:15 -0000 Subject: [tex-live] TeX Live 2016 release available Message-ID: <201606062109.u56L96k4020573@freefriends.org> For the sake of anyone who doesn't follow ctan-ann ... TeX Live 2016 has been released. It will soon (but not this instant) be available from CTAN sites (http://mirror.ctan.org/systems/texlive). Package updates will resume in another day or two. As usual, we recommend doing a network install in one way or another instead of downloading the large (3+gb) .iso -- see http://tug.org/texlive/acquire.html for a discussion of the various alternatives. If anyone notices stale or missing information on the TL web pages, or problems with the release, please report as usual. MacTeX (http://tug.org/mactex) has also been released and are available from CTAN. The TeX Collection DVD (http://tug.org/texcollection) is going to production shortly and should be available in due course. I don't have an ETA. Happy TeXing, Karl P.S. Bug reports about packages suddenly not working, especially with luatex, should go to the package maintainers, not TL. From tljack at forallx.net Mon Jun 6 23:24:31 2016 From: tljack at forallx.net (tljack at forallx.net) Date: Mon, 06 Jun 2016 15:24:31 -0600 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none Message-ID: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> I just got TL2016 installed. There is a LaTeX document that with TL2015 typesets (pdfLaTeX) fine with no errors or warnings, but when I typset it with pdfLaTeX from 2016 I see: /tl/texmf-dist/tex/latex/graphics/graphics.sty:94: Package graphics Error: No driver specified. I don't know why it's saying this or why it's asking for a driver. I don't have a funny document or a funny preamble. All it does is this: \usepackage{enumitem,geometry,array,graphicx,amsmath,amssymb} I see this even if the entire textual content between the \{begin,end}{document} the body is "Foo". If I remove the "graphicx" the error goes away. Doing "texdoc graphics" reveals nothing to illuminate this or why it would suddenly be different from TL2015. -jack From martin at oneiros.de Mon Jun 6 23:44:33 2016 From: martin at oneiros.de (=?UTF-8?Q?Martin_Schr=C3=B6der?=) Date: Mon, 6 Jun 2016 23:44:33 +0200 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> Message-ID: 2016-06-06 23:24 GMT+02:00 : > If I remove the "graphicx" the error goes away. Doing "texdoc graphics" > reveals nothing to illuminate this or why it would suddenly be different > from TL2015. Section 2 explains the driver option. IIRC it was always needed. Best Martin From news3 at nililand.de Mon Jun 6 23:57:40 2016 From: news3 at nililand.de (Ulrike Fischer) Date: Mon, 6 Jun 2016 23:57:40 +0200 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> Message-ID: <99srjjzqqonp.dlg@nililand.de> Am Mon, 06 Jun 2016 15:24:31 -0600 schrieb tljack at forallx.net: > I just got TL2016 installed. There is a LaTeX document that with TL2015 > typesets (pdfLaTeX) fine with no errors or warnings, but when I typset > it with pdfLaTeX from 2016 I see: > > /tl/texmf-dist/tex/latex/graphics/graphics.sty:94: Package graphics > Error: > No driver specified. I don't get such an error. > > I don't know why it's saying this or why it's asking for a driver. I > don't have a funny document or a funny preamble. All it does is this: Check the log file if you have some local graphics.cfg. -- Ulrike Fischer http://www.troubleshooting-tex.de/ From kakuto at fuk.kindai.ac.jp Tue Jun 7 00:05:22 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Tue, 7 Jun 2016 07:05:22 +0900 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> Message-ID: > /tl/texmf-dist/tex/latex/graphics/graphics.sty:94: Package graphics > Error: > No driver specified. Not confirmed here. % pdflatex \documentclass[a4paper]{article} \usepackage{enumitem,geometry,array,graphicx,amsmath,amssymb} \begin{document} Foo. \end{document} works ok. Do you have texmf-dist/tex/latex/graphics-cfg/graphics.cfg ? Best, Akira From w.m.l at gmx.net Tue Jun 7 01:24:53 2016 From: w.m.l at gmx.net (Robert) Date: Tue, 7 Jun 2016 01:24:53 +0200 Subject: [tex-live] pdftex segfault with \pdflastmatch In-Reply-To: References: Message-ID: On 05.06.16 23:09, David Carlisle wrote: > \pdfmatch{a}{b} > \pdflastmatch0 > \bye On Mac OS X (x86_64-darwin), there is no crash, but the second number in the output (eg. the 8 in "0 8-?") randomly changes on every run (seems to be between 0 and 15). I haven't tested with the patch. Best, -- Robert From werner at suse.de Tue Jun 7 13:49:39 2016 From: werner at suse.de (Dr. Werner Fink) Date: Tue, 7 Jun 2016 13:49:39 +0200 Subject: [tex-live] TeXLive 2016 source tar ball with object and library files? Message-ID: <20160607114939.2e6qzkxeex5l4ite@boole.suse.de> Hi, just tried to archivemount texlive-20160523-source.tar.xz and recoginzed that the tar ball is much larger then those of 2015 (252126K to 44395K) due a lot of binary, object, and library files. Why this? Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From daleif at math.au.dk Tue Jun 7 16:43:24 2016 From: daleif at math.au.dk (Lars Madsen) Date: Tue, 7 Jun 2016 14:43:24 +0000 Subject: [tex-live] Download did not succeed Message-ID: <92d9c7b2ffca40828638f0419112a153@Exch08.uni.au.dk> I tried installing TL2016 today I keep getting TLUtils::check_file: removing /tmp/bd1BBkrSDi/uEpFRVo_Yx/pdfpagediff.doc.tar.xz, sizes differ: TLUtils::check_file: TL=56330, arg=740820 TLPDB::_install_package: downloading did not succeed Installation failed. Rerunning the installer will try to restart the installation. Or you can restart by running the installer with: install-tl --profile installation.profile [EXTRA-ARGS] install-tl: Writing log in current directory: /home/daleif/tmp/install-tl-20160523/install-tl.log at random places in the package list. I just rerun it, and when I come back to it some time later, it has given a similar error at a different package. Any ideas? Using Automated TeX Live installation using profile: installation.profile Trying to verify cryptographic signatures! Loading http://ftp.snt.utwente.nl/pub/software/tex/systems/texlive/tlnet/tlpkg/texlive.tlpdb Installing TeX Live 2016 from: http://ftp.snt.utwente.nl/pub/software/tex/systems/texlive/tlnet (verified) Platform: x86_64-linux => 'GNU/Linux on x86_64' Distribution: net (downloading) Using URL: http://ftp.snt.utwente.nl/pub/software/tex/systems/texlive/tlnet Directory for temporary files: /tmp/bd1BBkrSDi I'll see if the Danish mirror is working and try to use that instead /Lars Madsen Institut for Matematik / Department of Mathematics Aarhus Universitet / Aarhus University Mere info: http://au.dk/daleif at math / More information: http://au.dk/en/daleif at math From daleif at math.au.dk Tue Jun 7 17:08:03 2016 From: daleif at math.au.dk (Lars Madsen) Date: Tue, 7 Jun 2016 15:08:03 +0000 Subject: [tex-live] Download did not succeed In-Reply-To: <92d9c7b2ffca40828638f0419112a153@Exch08.uni.au.dk> References: <92d9c7b2ffca40828638f0419112a153@Exch08.uni.au.dk> Message-ID: <1465312091062.15766@math.au.dk> seems to have been a bad mirror Restarted everything against a Norwegian mirror, no issues at all, download done in ~20min sorry for the noise /Lars Madsen Institut for Matematik / Department of Mathematics Aarhus Universitet / Aarhus University Mere info: http://au.dk/daleif at math / More information: http://au.dk/en/daleif at math ________________________________________ From: tex-live on behalf of Lars Madsen Sent: 07 June 2016 16:43 To: tex-live at tug.org Subject: [tex-live] Download did not succeed I tried installing TL2016 today I keep getting TLUtils::check_file: removing /tmp/bd1BBkrSDi/uEpFRVo_Yx/pdfpagediff.doc.tar.xz, sizes differ: TLUtils::check_file: TL=56330, arg=740820 TLPDB::_install_package: downloading did not succeed Installation failed. Rerunning the installer will try to restart the installation. Or you can restart by running the installer with: install-tl --profile installation.profile [EXTRA-ARGS] install-tl: Writing log in current directory: /home/daleif/tmp/install-tl-20160523/install-tl.log at random places in the package list. I just rerun it, and when I come back to it some time later, it has given a similar error at a different package. Any ideas? Using Automated TeX Live installation using profile: installation.profile Trying to verify cryptographic signatures! Loading http://ftp.snt.utwente.nl/pub/software/tex/systems/texlive/tlnet/tlpkg/texlive.tlpdb Installing TeX Live 2016 from: http://ftp.snt.utwente.nl/pub/software/tex/systems/texlive/tlnet (verified) Platform: x86_64-linux => 'GNU/Linux on x86_64' Distribution: net (downloading) Using URL: http://ftp.snt.utwente.nl/pub/software/tex/systems/texlive/tlnet Directory for temporary files: /tmp/bd1BBkrSDi I'll see if the Danish mirror is working and try to use that instead /Lars Madsen Institut for Matematik / Department of Mathematics Aarhus Universitet / Aarhus University Mere info: http://au.dk/daleif at math / More information: http://au.dk/en/daleif at math From tljack at forallx.net Tue Jun 7 17:34:28 2016 From: tljack at forallx.net (tljack at forallx.net) Date: Tue, 07 Jun 2016 09:34:28 -0600 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> Message-ID: <03610f21a8464e9ec4df0d9dfd752908@forallx.net> On 2016-06-06 16:05, Akira Kakuto wrote: >> /tl/texmf-dist/tex/latex/graphics/graphics.sty:94: Package graphics >> Error: >> No driver specified. > > Not confirmed here. > % pdflatex > \documentclass[a4paper]{article} > \usepackage{enumitem,geometry,array,graphicx,amsmath,amssymb} > \begin{document} > Foo. > \end{document} > works ok. > Do you have > texmf-dist/tex/latex/graphics-cfg/graphics.cfg ? I assume that path should end with /graphics/graphics.cfg. Anyway, no I don't have that file either in TL2016, or in TL2015. TL2016 is the only one with the error. -jack From d.p.carlisle at gmail.com Tue Jun 7 17:42:13 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Tue, 7 Jun 2016 16:42:13 +0100 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: <03610f21a8464e9ec4df0d9dfd752908@forallx.net> References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <03610f21a8464e9ec4df0d9dfd752908@forallx.net> Message-ID: On 7 June 2016 at 16:34, wrote: > On 2016-06-06 16:05, Akira Kakuto wrote: > >> /tl/texmf-dist/tex/latex/graphics/graphics.sty:94: Package graphics Error: >>> No driver specified. >>> >> >> Not confirmed here. >> % pdflatex >> \documentclass[a4paper]{article} >> \usepackage{enumitem,geometry,array,graphicx,amsmath,amssymb} >> \begin{document} >> Foo. >> \end{document} >> works ok. >> Do you have >> texmf-dist/tex/latex/graphics-cfg/graphics.cfg ? >> > > I assume that path should end with /graphics/graphics.cfg. > > Anyway, no I don't have that file either in TL2016, or in TL2015. TL2016 > is the only one with the error. > > -jack > If you do not have graphics.cfg then you need to supply an explicit driver option to the graphics package but one has been part of texlive since before it was called texlive. in tl2015 (and possibly tl2016) what do you get returned for kpsewhich graphics.cfg If you have not got that file you can get it from https://www.ctan.org/tex-archive/macros/latex/contrib/graphics-cfg but how did you install texlive? David -------------- next part -------------- An HTML attachment was scrubbed... URL: From martin at oneiros.de Tue Jun 7 17:47:55 2016 From: martin at oneiros.de (=?UTF-8?Q?Martin_Schr=C3=B6der?=) Date: Tue, 7 Jun 2016 17:47:55 +0200 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: <03610f21a8464e9ec4df0d9dfd752908@forallx.net> References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <03610f21a8464e9ec4df0d9dfd752908@forallx.net> Message-ID: 2016-06-07 17:34 GMT+02:00 : > On 2016-06-06 16:05, Akira Kakuto wrote: >> Do you have >> texmf-dist/tex/latex/graphics-cfg/graphics.cfg ? > > > I assume that path should end with /graphics/graphics.cfg. Do you have the directory texmf-dist/tex/latex/graphics-cfg/ and what does it contain? Anway: Your installation looks broken. It's probably easier if you reinstall. Best Martin From tljack at forallx.net Tue Jun 7 17:53:55 2016 From: tljack at forallx.net (tljack at forallx.net) Date: Tue, 07 Jun 2016 09:53:55 -0600 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <03610f21a8464e9ec4df0d9dfd752908@forallx.net> Message-ID: <78ac2bbe0636794a5369dc1aa820c338@forallx.net> On 2016-06-07 09:42, David Carlisle wrote: > On 7 June 2016 at 16:34, wrote: > >> On 2016-06-06 16:05, Akira Kakuto wrote: >> /tl/texmf-dist/tex/latex/graphics/graphics.sty:94: Package graphics >> Error: >> No driver specified. >> >> Not confirmed here. >> % pdflatex >> documentclass[a4paper]{article} >> usepackage{enumitem,geometry,array,graphicx,amsmath,amssymb} >> begin{document} >> Foo. >> end{document} >> works ok. >> Do you have >> texmf-dist/tex/latex/graphics-cfg/graphics.cfg ? > > I assume that path should end with /graphics/graphics.cfg. > > Anyway, no I don't have that file either in TL2016, or in TL2015. > TL2016 is the only one with the error. > > -jack > > If you do not have graphics.cfg then you need to supply an explicit > driver option to the graphics package but one has been part of texlive > since before it was called texlive. > > in tl2015 (and possibly tl2016) what do you get returned for > > kpsewhich graphics.cfg Ah, now we're getting somewhere. That command returns nothing at all in TL2016, but in TL2015 it returns: /tl/texmf-dist/tex/latex/latexconfig/graphics.cfg I don't know why it wouldn't be there this year. > If you have not got that file you can get it from > > https://www.ctan.org/tex-archive/macros/latex/contrib/graphics-cfg [1] > > but how did you install texlive? For the last few years, I've done it this way: I rsync the entire distro to my machine, then I run the tlnet/install-tl script and install the "minimal" scheme. I then make sure that everything is on my $PATH (by doing e.g., `which tex`). Then I tell tlmgr to use my local clone as the repo, and then I install some more packages. Prior to that I did not rsync the entire distro, but just downloaded the small tarball that holds "install-tl", and ran that. -jack From tljack at forallx.net Tue Jun 7 17:59:30 2016 From: tljack at forallx.net (tljack at forallx.net) Date: Tue, 07 Jun 2016 09:59:30 -0600 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <03610f21a8464e9ec4df0d9dfd752908@forallx.net> Message-ID: <99cf6d18c536b2688c06a7e0b6710fc6@forallx.net> On 2016-06-07 09:47, Martin Schr?der wrote: > 2016-06-07 17:34 GMT+02:00 : >> On 2016-06-06 16:05, Akira Kakuto wrote: >>> Do you have >>> texmf-dist/tex/latex/graphics-cfg/graphics.cfg ? >> >> >> I assume that path should end with /graphics/graphics.cfg. > > Do you have the directory texmf-dist/tex/latex/graphics-cfg/ > and what does it contain? No, I have no directories ending in "graphics-cfg". I have "graphics", but it contains only .sty and .def files. I don't see how my installation could be broken unless some big change happened since last year. Even if I reinstall TL2016 using the ".profile" from TL2015 (but changing the year), I still see this. Did the missing folder get moved into a different package from where it previously was kept? Which package should contain it? -jack From d.p.carlisle at gmail.com Tue Jun 7 18:01:32 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Tue, 7 Jun 2016 17:01:32 +0100 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: <78ac2bbe0636794a5369dc1aa820c338@forallx.net> References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <03610f21a8464e9ec4df0d9dfd752908@forallx.net> <78ac2bbe0636794a5369dc1aa820c338@forallx.net> Message-ID: On 7 June 2016 at 16:53, wrote: > On 2016-06-07 09:42, David Carlisle wrote: > >> On 7 June 2016 at 16:34, wrote: >> >> On 2016-06-06 16:05, Akira Kakuto wrote: >>> /tl/texmf-dist/tex/latex/graphics/graphics.sty:94: Package graphics >>> Error: >>> No driver specified. >>> >>> Not confirmed here. >>> % pdflatex >>> documentclass[a4paper]{article} >>> usepackage{enumitem,geometry,array,graphicx,amsmath,amssymb} >>> begin{document} >>> Foo. >>> end{document} >>> works ok. >>> Do you have >>> texmf-dist/tex/latex/graphics-cfg/graphics.cfg ? >>> >> >> I assume that path should end with /graphics/graphics.cfg. >> >> Anyway, no I don't have that file either in TL2016, or in TL2015. >> TL2016 is the only one with the error. >> >> -jack >> >> If you do not have graphics.cfg then you need to supply an explicit >> driver option to the graphics package but one has been part of texlive >> since before it was called texlive. >> >> in tl2015 (and possibly tl2016) what do you get returned for >> >> kpsewhich graphics.cfg >> > > Ah, now we're getting somewhere. That command returns nothing at all in > TL2016, but in TL2015 it returns: > /tl/texmf-dist/tex/latex/latexconfig/graphics.cfg > > I don't know why it wouldn't be there this year. > there's been some rearrangement of the locations, in particular that file is now on ctan (and pulled from there to texlive) rather than being maintained directly in texlive. so in 2016 it is 2016/texmf-dist/tex/latex/graphics-cfg/graphics.cfg > > If you have not got that file you can get it from >> >> https://www.ctan.org/tex-archive/macros/latex/contrib/graphics-cfg [1] >> >> but how did you install texlive? >> > > For the last few years, I've done it this way: I rsync the entire distro > to my machine, then I run the tlnet/install-tl script and install the > "minimal" scheme. I leave it to the texlive maintainers to say what should or shouldn't be in minimal but if you install minimal then it's natural that you have to install some additional things so that real-world documents run, so it's not necessarily a bug that graphics.cfg isn't in minimal. (The graphics package hopefully clearly documents that it is optional and what you need to do to specify a driver if you don't have a cfg file) David -------------- next part -------------- An HTML attachment was scrubbed... URL: From tljack at forallx.net Tue Jun 7 18:06:00 2016 From: tljack at forallx.net (tljack at forallx.net) Date: Tue, 07 Jun 2016 10:06:00 -0600 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <03610f21a8464e9ec4df0d9dfd752908@forallx.net> <78ac2bbe0636794a5369dc1aa820c338@forallx.net> Message-ID: <6fc429fff496f9bc80abd054380f1fbb@forallx.net> On 2016-06-07 10:01, David Carlisle wrote: > On 7 June 2016 at 16:53, wrote: > >> On 2016-06-07 09:42, David Carlisle wrote: >> On 7 June 2016 at 16:34, wrote: >> >> On 2016-06-06 16:05, Akira Kakuto wrote: >> /tl/texmf-dist/tex/latex/graphics/graphics.sty:94: Package graphics >> Error: >> No driver specified. >> >> Not confirmed here. >> % pdflatex >> documentclass[a4paper]{article} >> usepackage{enumitem,geometry,array,graphicx,amsmath,amssymb} >> begin{document} >> Foo. >> end{document} >> works ok. >> Do you have >> texmf-dist/tex/latex/graphics-cfg/graphics.cfg ? >> >> I assume that path should end with /graphics/graphics.cfg. >> >> Anyway, no I don't have that file either in TL2016, or in TL2015. >> TL2016 is the only one with the error. >> >> -jack >> >> If you do not have graphics.cfg then you need to supply an explicit >> driver option to the graphics package but one has been part of >> texlive >> since before it was called texlive. >> >> in tl2015 (and possibly tl2016) what do you get returned for >> >> kpsewhich graphics.cfg > > Ah, now we're getting somewhere. That command returns nothing at all > in TL2016, but in TL2015 it returns: > /tl/texmf-dist/tex/latex/latexconfig/graphics.cfg > > I don't know why it wouldn't be there this year. > > there's been some rearrangement of the locations, in particular that > file is now on ctan (and pulled from there to texlive) rather than > being maintained directly in texlive. so in 2016 it is > 2016/texmf-dist/tex/latex/graphics-cfg/graphics.cfg > >>> If you have not got that file you can get it from >>> >>> https://www.ctan.org/tex-archive/macros/latex/contrib/graphics-cfg >>> [1] [1] >>> >>> but how did you install texlive? >> >> For the last few years, I've done it this way: I rsync the entire >> distro to my machine, then I run the tlnet/install-tl script and >> install the "minimal" scheme. > > I leave it to the texlive maintainers to say what should or shouldn't > be in minimal > > but if you install minimal then it's natural that you have to install > some additional > > things so that real-world documents run, so it's not necessarily a bug > that graphics.cfg isn't in minimal. (The graphics package hopefully > clearly documents that it is optional and what you need to do to > specify a driver if you don't have a cfg file) I don't just install the minimal scheme. I must've forgot to mention that in my email. After the minimal scheme, I install several hundred additional packages. -jack From tljack at forallx.net Tue Jun 7 18:08:27 2016 From: tljack at forallx.net (tljack at forallx.net) Date: Tue, 07 Jun 2016 10:08:27 -0600 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <03610f21a8464e9ec4df0d9dfd752908@forallx.net> <78ac2bbe0636794a5369dc1aa820c338@forallx.net> Message-ID: On 2016-06-07 10:01, David Carlisle wrote: > On 7 June 2016 at 16:53, wrote: > >> On 2016-06-07 09:42, David Carlisle wrote: >> On 7 June 2016 at 16:34, wrote: >> >> On 2016-06-06 16:05, Akira Kakuto wrote: >> /tl/texmf-dist/tex/latex/graphics/graphics.sty:94: Package graphics >> Error: >> No driver specified. >> >> Not confirmed here. >> % pdflatex >> documentclass[a4paper]{article} >> usepackage{enumitem,geometry,array,graphicx,amsmath,amssymb} >> begin{document} >> Foo. >> end{document} >> works ok. >> Do you have >> texmf-dist/tex/latex/graphics-cfg/graphics.cfg ? >> >> I assume that path should end with /graphics/graphics.cfg. >> >> Anyway, no I don't have that file either in TL2016, or in TL2015. >> TL2016 is the only one with the error. >> >> -jack >> >> If you do not have graphics.cfg then you need to supply an explicit >> driver option to the graphics package but one has been part of >> texlive >> since before it was called texlive. >> >> in tl2015 (and possibly tl2016) what do you get returned for >> >> kpsewhich graphics.cfg > > Ah, now we're getting somewhere. That command returns nothing at all > in TL2016, but in TL2015 it returns: > /tl/texmf-dist/tex/latex/latexconfig/graphics.cfg > > I don't know why it wouldn't be there this year. > > there's been some rearrangement of the locations, in particular that > file is now on ctan (and pulled from there to texlive) rather than > being maintained directly in texlive. so in 2016 it is > 2016/texmf-dist/tex/latex/graphics-cfg/graphics.cfg My question, then, is which package should I tell tlmgr to install, so that I get that file? -jack From tljack at forallx.net Tue Jun 7 18:13:58 2016 From: tljack at forallx.net (tljack at forallx.net) Date: Tue, 07 Jun 2016 10:13:58 -0600 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <03610f21a8464e9ec4df0d9dfd752908@forallx.net> <78ac2bbe0636794a5369dc1aa820c338@forallx.net> Message-ID: <5da059b2b1823c1a1d7cdc543408a1cd@forallx.net> On 2016-06-07 10:01, David Carlisle wrote: > On 7 June 2016 at 16:53, wrote: > >> On 2016-06-07 09:42, David Carlisle wrote: >> On 7 June 2016 at 16:34, wrote: >> >> On 2016-06-06 16:05, Akira Kakuto wrote: >> /tl/texmf-dist/tex/latex/graphics/graphics.sty:94: Package graphics >> Error: >> No driver specified. >> >> Not confirmed here. >> % pdflatex >> documentclass[a4paper]{article} >> usepackage{enumitem,geometry,array,graphicx,amsmath,amssymb} >> begin{document} >> Foo. >> end{document} >> works ok. >> Do you have >> texmf-dist/tex/latex/graphics-cfg/graphics.cfg ? >> >> I assume that path should end with /graphics/graphics.cfg. >> >> Anyway, no I don't have that file either in TL2016, or in TL2015. >> TL2016 is the only one with the error. >> >> -jack >> >> If you do not have graphics.cfg then you need to supply an explicit >> driver option to the graphics package but one has been part of >> texlive >> since before it was called texlive. >> >> in tl2015 (and possibly tl2016) what do you get returned for >> >> kpsewhich graphics.cfg > > Ah, now we're getting somewhere. That command returns nothing at all > in TL2016, but in TL2015 it returns: > /tl/texmf-dist/tex/latex/latexconfig/graphics.cfg > > I don't know why it wouldn't be there this year. > > there's been some rearrangement of the locations, in particular that > file is now on ctan (and pulled from there to texlive) rather than > being maintained directly in texlive. so in 2016 it is > 2016/texmf-dist/tex/latex/graphics-cfg/graphics.cfg I asked what package this is in, because if I do: "tlmgr info graphics-cfg", I get this: Unknown directive ...containerchecksum c59200574a316416a23695c258edf3a32531fbda43ccdc09360ee105c3f07f9fb77df17c4ba4c2ea4f3a5ea6667e064b51e3d8c2fe6c984ba3e71b4e32716955... , please fix it! at /opt/tl/tlpkg/TeXLive/TLPOBJ.pm line 210, line 5579. -jack From preining at logic.at Tue Jun 7 18:17:22 2016 From: preining at logic.at (Norbert Preining) Date: Wed, 8 Jun 2016 01:17:22 +0900 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: <6fc429fff496f9bc80abd054380f1fbb@forallx.net> References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <03610f21a8464e9ec4df0d9dfd752908@forallx.net> <78ac2bbe0636794a5369dc1aa820c338@forallx.net> <6fc429fff496f9bc80abd054380f1fbb@forallx.net> Message-ID: Well, without telling which packages you installed nobody can help you. In particular, of you only install packages one but one you might have forgotten the necessary packages that got singled out into a separate package as explained. It is up to you to install the necessary packages if you don't use the collections we provide. Simply installing collection-latex would fix your problems I guess. Norbert >that in my email. After the minimal scheme, I install several hundred >additional packages. -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From preining at logic.at Tue Jun 7 18:17:58 2016 From: preining at logic.at (Norbert Preining) Date: Wed, 8 Jun 2016 01:17:58 +0900 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <03610f21a8464e9ec4df0d9dfd752908@forallx.net> <78ac2bbe0636794a5369dc1aa820c338@forallx.net> Message-ID: Use tlmgr search ? -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From d.p.carlisle at gmail.com Tue Jun 7 18:20:57 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Tue, 7 Jun 2016 17:20:57 +0100 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <03610f21a8464e9ec4df0d9dfd752908@forallx.net> <78ac2bbe0636794a5369dc1aa820c338@forallx.net> Message-ID: On 7 June 2016 at 17:08, wrote: > > My question, then, is which package should I tell tlmgr to install, so > that I get that file? > > -jack > > seems to be graphics-cfg: $ tlmgr info graphics-cfg package: graphics-cfg category: Package shortdesc: Sample configuration files for LaTeX color and graphics longdesc: This bundle includes color.cfg and graphics.cfg files that set default "driver" options for the color and graphics packages. It contains support for defaulting the new LuaTeX option which was added to graphics and color in the 2016-02-01 release. The LuaTeX option is only used for LuaTeX versions from 0.87, older versions use the pdfTeX option as before. installed: Yes revision: 40269 sizes: doc: 5k, run: 9k relocatable: No cat-date: 2016-03-01 21:50:00 +0100 cat-license: pd cat-topics: graphics colour cat-related: color graphics collection: collection-latex David -------------- next part -------------- An HTML attachment was scrubbed... URL: From komascript at gmx.info Tue Jun 7 19:12:18 2016 From: komascript at gmx.info (Markus Kohm) Date: Tue, 7 Jun 2016 19:12:18 +0200 Subject: [tex-live] TL2016 brings "no driver" error where previously there was none In-Reply-To: <5da059b2b1823c1a1d7cdc543408a1cd@forallx.net> References: <647200a3759eed02dd99c9be7e75b8a1@forallx.net> <5da059b2b1823c1a1d7cdc543408a1cd@forallx.net> Message-ID: <4075266.4px6TCTdWA@x61s.fritz.box> Am Dienstag, 7. Juni 2016, 10:13:58 schrieb tljack at forallx.net: > Unknown directive ...containerchecksum Your tlmgr is outdated. Maybe you are using tlmgr from TL2015. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 181 bytes Desc: This is a digitally signed message part. URL: From joseph.wright at morningstar2.co.uk Tue Jun 7 23:51:59 2016 From: joseph.wright at morningstar2.co.uk (Joseph Wright) Date: Tue, 7 Jun 2016 22:51:59 +0100 Subject: [tex-live] (u)p(La)TeX dependency tree Message-ID: <2feaf360-d84d-8a56-1fc5-69d03d3ee1d1@morningstar2.co.uk> Hello all, Updating some automated tests from TL'15 to TL'16, I find that there are some oddities with (u)pTeX/(u)pLaTeX. In particular, with TL'15 a minimised system with tlmgr install babel babel-english latex latex-bin \ latex-fonts latexconfig xetex tlmgr install --no-depends ptex uptex would allow basic testing with (u)pTeX including with (u)pLaTeX. To achieve the same in TL'16 I need tlmgr install babel babel-english latex latex-bin \ latex-fonts latexconfig xetex tlmgr install --no-depends ptex uptex ptex-base \ uptex-base ptex-fonts uptex-fonts platex uplatex What's odd in particular is: - You need the two 'fonts' packages to get the formats to be listed in fmtutil.cnf - The (u)platex packages don't depend on the (u)ptex ones (I use the --no-depends here as the aim is to have a truly minimal system which has just enough to work: it's automatically reinstalled for every run, so there is a need to avoid any unused material.) Joseph From kakuto at fuk.kindai.ac.jp Wed Jun 8 03:06:03 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Wed, 8 Jun 2016 10:06:03 +0900 Subject: [tex-live] xindy.pl In-Reply-To: <2feaf360-d84d-8a56-1fc5-69d03d3ee1d1@morningstar2.co.uk> References: <2feaf360-d84d-8a56-1fc5-69d03d3ee1d1@morningstar2.co.uk> Message-ID: <62D8C6CFFE7A434EBA2AEBC67BF934BA@CJ3001517A> Hello all, In r41316, we have updated xindy.pl. In the new xindy.pl, kpsewhich is used to search for xindy modules. The performance is lost. However user modules specified by the -M option can be under /xindy/modules/. Best, Akira From thomasgmeyer at online.de Wed Jun 8 17:21:49 2016 From: thomasgmeyer at online.de (Thomas Meyer) Date: Wed, 8 Jun 2016 17:21:49 +0200 Subject: [tex-live] LuaLaTex MacTex 2016 doesn't work In-Reply-To: <2382684.l2dEdiHFUk@x61s.fritz.box> References: <023373fb-d377-334e-f5b9-8b8ab60ca2ef@online.de> <2382684.l2dEdiHFUk@x61s.fritz.box> Message-ID: <75d6839e-6a0f-755a-7472-75028a2adfdc@online.de> Hi, thank you all for your answers. Now I can work with. Greetings Thomas Am 06.06.16 um 20:09 schrieb Markus Kohm: > Am Montag, 6. Juni 2016, 19:55:08 schrieb Thomas Meyer: >> What's going wrong? > > This is not a TeX Live bug, it's a known KOMA-Script bug: > > ? > > Markus > --- > Infos zu LaTeX: > Infos zu KOMA-Script: > Aktuelles KOMA-Script: > KOMA-Script-Anleitung: > -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 841 bytes Desc: OpenPGP digital signature URL: From vincent.torri at gmail.com Wed Jun 8 22:20:20 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Wed, 8 Jun 2016 22:20:20 +0200 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf Message-ID: Hello I am writing a library using the kpathsea library. kpathsea reports, on Windows, this message: warning: kpathsea: configuration file texmf.cnf not found in these directories: etc... Here are the steps that described what and how I install and use TeX Live: 1) I installed TeX Live 2016 (using install-tl-windows.exe) in the directory E:\Documents\programmes_x86\texlive, so I have texmf.cnf in E:\Documents\programmes_x86\texlive\2016 2) just after the installation of TeX Live, I have added the binary directory to PATH (that is : E:\Documents\programmes_x86\texlive\2016\bin\win32) 3) As there is only the kpathsea DLL (no import library, no header file for developping with kpathsea), I have loaded the DLL with LoadLibrary and retrieve the symbols with GetProcAdress(). The code can be find here : https://github.com/vtorri/libdvi/blob/master/src/lib/dvi_kpathsea.c (lines 281 and after, and the macros are defined lines 230 and after) 4) I test my code with a DVI file using the font files cmr12.tfm and cmsy10.tfm, with the code of the above file, function dvi_kpathsea_path_name_get(), line 357. 5) I install my program in /opt/efl_i686/bin, that is : E:/Documents/MSYS2/opt/efl_i686/bin 6) When I run my program, with MSYS2, the kpathsea message is : warning: kpathsea: configuration file texmf.cnf not found in these directories: E:/Documents/MSYS2/opt/efl_i686/bin;E:/Documents/MSYS2/opt/efl_i686/bin/share/texmf-local/web2c;E:/Documents/MSYS2/opt/efl_i686/bin/share/texmf-dist/web2c;E:/Documents/MSYS2/opt/efl_i686/bin/share/texmf/web2c;E:/Documents/MSYS2/opt/efl_i686/bin/texmf-local/web2c;E:/Documents/MSYS2/opt/efl_i686/bin/texmf-dist/web2c;E:/Documents/MSYS2/opt/efl_i686/bin/texmf/web2c;E:/Documents/MSYS2/opt/efl_i686;E:/Documents/MSYS2/opt/efl_i686/share/texmf-local/web2c;E:/Documents/MSYS2/opt/efl_i686/share/texmf-dist/web2c;E:/Documents/MSYS2/opt/efl_i686/share/texmf/web2c;E:/Documents/MSYS2/opt/efl_i686/texmf-local/web2c;E:/Documents/MSYS2/opt/efl_i686/texmf-dist/web2c;E:/Documents/MSYS2/opt/efl_i686/texmf/web2c;E:/Documents/MSYS2/texmf-local/web2c;E:/Documents/MSYS2/opt;E:/Documents/MSYS2/opt/share/texmf-local/web2c;E:/Documents/MSYS2/opt/share/texmf-dist/web2c;E:/Documents/MSYS2/opt/share/texmf/web2c;E:/Documents/MSYS2/opt/texmf-local/web2c;E:/Documents/MSYS2/opt/texmf-dist/web2c;E:/Documents/MSYS2/opt/texmf/web2c. So, it seems to search texmf.cnf only in subdirectories of E:/Documents/MSYS2/opt/. Maybe it is because I load kpathsea DLL with LoadLibrary, and kpathsea searches texmf.cnf from where my program is installed. Maybe it is the code of dvi_kpathsea_path_name_get() : I call kpathsea_set_program_name() and kpathsea_init_prog() with the name of my program ("dvi.exe", installed in E:/Documents/MSYS2/opt/efl_i686/bin) I tried to set TEXMF and TEXMFCNF in the shell to some values, without success. But maybe i've wrongly set those environment variables. Does someone have ideas to solve this issue ? thank you Vincent Torri From kakuto at fuk.kindai.ac.jp Thu Jun 9 01:42:52 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Thu, 9 Jun 2016 08:42:52 +0900 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf In-Reply-To: References: Message-ID: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> Dear Vincent, > I install my program in /opt/efl_i686/bin, that is : > E:/Documents/MSYS2/opt/efl_i686/bin You have to install your program into E:/Documents/programmes_x86/texlive/2016/bin/win32 in order to search for texmf.cnf's correctly. Best, Akira From vincent.torri at gmail.com Thu Jun 9 04:02:25 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Thu, 9 Jun 2016 04:02:25 +0200 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf In-Reply-To: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> References: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> Message-ID: Hello On Thu, Jun 9, 2016 at 1:42 AM, Akira Kakuto wrote: > Dear Vincent, > >> I install my program in /opt/efl_i686/bin, that is : >> E:/Documents/MSYS2/opt/efl_i686/bin > > > You have to install your program into > E:/Documents/programmes_x86/texlive/2016/bin/win32 > > in order to search for texmf.cnf's correctly. There is no other option, like setting an environment variable (TEXMF* environment variable) ? Otherwise, that's very annoying as it force me to install all the program using my library in that directory. There is certainly another option, as kpsewhich can find the directory where cmr12.tfm is located. Vincent Torri From kakuto at fuk.kindai.ac.jp Thu Jun 9 04:32:12 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Thu, 9 Jun 2016 11:32:12 +0900 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf In-Reply-To: References: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> Message-ID: Dear Vincent, > There is no other option, like setting an environment variable (TEXMF* > environment variable) ? Otherwise, that's very annoying as it force me > to install all the program using my library in that directory. > so I have texmf.cnf in > E:/Documents/programmes_x86/texlive/2016 Define the env variable TEXMFCNF as TEXMFCNF=E:/Documents/programmes_x86/texlive/2016; The trailing semicolon is important. If the semicolon does not exist, texmf.cnf is searched only in E:/Documents/programmes_x86/texlive/2016. Best, Akira From vincent.torri at gmail.com Thu Jun 9 04:33:25 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Thu, 9 Jun 2016 04:33:25 +0200 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf In-Reply-To: References: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> Message-ID: Hello 1) Is setting TEXMFCNF in some way can help ? 2) If i call kpathsea_set_program_name(kpse, "kpsewhich.exe", NULL); that is not wth "dvi.exe" but with "kpsewhich.exe", it works. Is this reasonable to call kpathsea_set_program_name() like that ? thank you Vincent Torri On Thu, Jun 9, 2016 at 4:02 AM, Vincent Torri wrote: > Hello > > On Thu, Jun 9, 2016 at 1:42 AM, Akira Kakuto wrote: >> Dear Vincent, >> >>> I install my program in /opt/efl_i686/bin, that is : >>> E:/Documents/MSYS2/opt/efl_i686/bin >> >> >> You have to install your program into >> E:/Documents/programmes_x86/texlive/2016/bin/win32 >> >> in order to search for texmf.cnf's correctly. > > There is no other option, like setting an environment variable (TEXMF* > environment variable) ? Otherwise, that's very annoying as it force me > to install all the program using my library in that directory. > > There is certainly another option, as kpsewhich can find the directory > where cmr12.tfm is located. > > Vincent Torri From vincent.torri at gmail.com Thu Jun 9 04:38:49 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Thu, 9 Jun 2016 04:38:49 +0200 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf In-Reply-To: References: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> Message-ID: On Thu, Jun 9, 2016 at 4:32 AM, Akira Kakuto wrote: > Dear Vincent, > >> There is no other option, like setting an environment variable (TEXMF* >> environment variable) ? Otherwise, that's very annoying as it force me >> to install all the program using my library in that directory. > > >> so I have texmf.cnf in >> E:/Documents/programmes_x86/texlive/2016 > > > Define the env variable TEXMFCNF as > TEXMFCNF=E:/Documents/programmes_x86/texlive/2016; > > The trailing semicolon is important. If the semicolon does not exist, > texmf.cnf is searched only in E:/Documents/programmes_x86/texlive/2016. 1) I've just tried this. My program seems to hang. 2) I tried E:/Documents/programmes_x86/texlive/2016: (with : at the end) and I get warning: kpathsea: configuration file texmf.cnf not found in these directories: E;E:/Documents/MSYS2/Documents/programmes_x86/texlive/2016. Vincent Torri From bernhard.kleine at gmx.net Thu Jun 9 09:22:06 2016 From: bernhard.kleine at gmx.net (Bernhard Kleine) Date: Thu, 9 Jun 2016 09:22:06 +0200 Subject: [tex-live] sanitize-umlaut in CTAN but not in texlive Message-ID: <4762883f-0a8e-3927-2e20-add5626827f4@gmx.net> Is there an issue with licences? Regards Bernhard -- -- spitzhalde9 79853 lenzkirch bernhard.kleine at gmx.net www.b-kleine.com, www.urseetal.net - thunderbird mit enigmail GPG schl?ssel: D5257409 fingerprint: 08 B7 F8 70 22 7A FC C1 15 49 CA A6 C7 6F A0 2E D5 25 74 09 From news3 at nililand.de Thu Jun 9 14:08:12 2016 From: news3 at nililand.de (Ulrike Fischer) Date: Thu, 9 Jun 2016 14:08:12 +0200 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf References: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> Message-ID: Am Thu, 9 Jun 2016 04:38:49 +0200 schrieb Vincent Torri: >> The trailing semicolon is important. If the semicolon does not exist, >> texmf.cnf is searched only in E:/Documents/programmes_x86/texlive/2016. > > 1) I've just tried this. My program seems to hang. I'm using TEXMFCNF=D:\xxx\yyy\; on windows 7 and 10 to get texlive to use a special texmf.cnf and it works fine. -- Ulrike Fischer http://www.troubleshooting-tex.de/ From zdenek.wagner at gmail.com Thu Jun 9 14:19:39 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Thu, 9 Jun 2016 14:19:39 +0200 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf In-Reply-To: References: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> Message-ID: 2016-06-09 4:38 GMT+02:00 Vincent Torri : > On Thu, Jun 9, 2016 at 4:32 AM, Akira Kakuto > wrote: > > Dear Vincent, > > > >> There is no other option, like setting an environment variable (TEXMF* > >> environment variable) ? Otherwise, that's very annoying as it force me > >> to install all the program using my library in that directory. > > > > > >> so I have texmf.cnf in > >> E:/Documents/programmes_x86/texlive/2016 > > > > > > Define the env variable TEXMFCNF as > > TEXMFCNF=E:/Documents/programmes_x86/texlive/2016; > > > > The trailing semicolon is important. If the semicolon does not exist, > > texmf.cnf is searched only in E:/Documents/programmes_x86/texlive/2016. > > 1) I've just tried this. My program seems to hang. > > 2) I tried E:/Documents/programmes_x86/texlive/2016: (with : at the > end) and I get > Unix-like forward slashes instead of windows-like backslashes are probably OK but what matters is the unix-like colon at the end, it certailny does not work. You must use a semicolon in Windows because a colon is used to separate a drive letter. > > warning: kpathsea: configuration file texmf.cnf not found in these > directories: E;E:/Documents/MSYS2/Documents/programmes_x86/texlive/2016. > > Vincent Torri > Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz -------------- next part -------------- An HTML attachment was scrubbed... URL: From preining at logic.at Thu Jun 9 15:50:24 2016 From: preining at logic.at (Norbert Preining) Date: Thu, 9 Jun 2016 22:50:24 +0900 Subject: [tex-live] TeXLive 2016 source tar ball with object and library files? In-Reply-To: <20160607114939.2e6qzkxeex5l4ite@boole.suse.de> References: <20160607114939.2e6qzkxeex5l4ite@boole.suse.de> Message-ID: <20160609135024.GQ16030@logic.at> On Tue, 07 Jun 2016, Dr. Werner Fink wrote: > just tried to archivemount texlive-20160523-source.tar.xz and recoginzed > that the tar ball is much larger then those of 2015 (252126K to 44395K) > due a lot of binary, object, and library files. No idea what has happened here ... Karl? Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From news3 at nililand.de Thu Jun 9 15:53:45 2016 From: news3 at nililand.de (Ulrike Fischer) Date: Thu, 9 Jun 2016 15:53:45 +0200 Subject: [tex-live] sanitize-umlaut in CTAN but not in texlive References: <4762883f-0a8e-3927-2e20-add5626827f4@gmx.net> Message-ID: Am Thu, 9 Jun 2016 09:22:06 +0200 schrieb Bernhard Kleine: > Is there an issue with licences? No. Probably texlive simply didn't yet catch up all new packages. After the freeze it can take some time until everything is there. -- Ulrike Fischer http://www.troubleshooting-tex.de/ From preining at logic.at Thu Jun 9 16:01:37 2016 From: preining at logic.at (Norbert Preining) Date: Thu, 9 Jun 2016 23:01:37 +0900 Subject: [tex-live] (u)p(La)TeX dependency tree In-Reply-To: <2feaf360-d84d-8a56-1fc5-69d03d3ee1d1@morningstar2.co.uk> References: <2feaf360-d84d-8a56-1fc5-69d03d3ee1d1@morningstar2.co.uk> Message-ID: <20160609140137.GR16030@logic.at> Hi Joseph, > - You need the two 'fonts' packages to get the formats to be listed > in fmtutil.cnf > - The (u)platex packages don't depend on the (u)ptex ones We don't track package - package dependencies completely seriously. The uptex rework was very recent, and thus probably needs some adjustments. If you have suggetions (which depends to be added to which package) simple mail them to Karl and me. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From werner at suse.de Thu Jun 9 16:01:51 2016 From: werner at suse.de (Dr. Werner Fink) Date: Thu, 9 Jun 2016 16:01:51 +0200 Subject: [tex-live] TeXLive 2016 source tar ball with object and library files? In-Reply-To: <20160609135024.GQ16030@logic.at> References: <20160607114939.2e6qzkxeex5l4ite@boole.suse.de> <20160609135024.GQ16030@logic.at> Message-ID: <20160609140151.bennnt643maxsqun@boole.suse.de> On Thu, Jun 09, 2016 at 10:50:24PM +0900, Norbert Preining wrote: > On Tue, 07 Jun 2016, Dr. Werner Fink wrote: > > just tried to archivemount texlive-20160523-source.tar.xz and recoginzed > > that the tar ball is much larger then those of 2015 (252126K to 44395K) > > due a lot of binary, object, and library files. > > No idea what has happened here ... Meanwhile I've cleaned out the source tar ball, no only binary, object, and library files but also config.status and the dependency files as well ;) du -ck texlive-20160523-source.tar.xz 46656 texlive-20160523-source.tar.xz 46656 total Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From henrimenke at gmail.com Thu Jun 9 16:05:53 2016 From: henrimenke at gmail.com (Henri Menke) Date: Thu, 9 Jun 2016 16:05:53 +0200 Subject: [tex-live] Debugging symbols for binaries Message-ID: <575977C1.5020205@gmail.com> Dear list, recently I found that pdftex segfaults [1]. When trying to find the issue, I ran pdftex in gdb to obtain a full backtrace. Unfortunately, without debugging symbols these backtraces are somewhat worthless, as they only show memory addresses. I then tried to compile pdftex from source, but soon noticed that a lot of handwork has to be done to make it run (installing header files, building formats, etc.). Therefore, I ask humbly, whether it would be possible to include debugging symbols for binaries in TeXlive (perhaps in an extra tree like documentation and sources). It is not necessary to have extra binaries with symbols, but symbols can be extracted to an extra file after compilation [2], so there is little to no redundancy. Below I have listed some pros and cons: [+] Experienced users have the ability to debug crashes [+] Developers might receive higher quality bug reports [+] Full stack traces reveal why it ?doesn't work on my machine? [-] Little to no use for most of the users [-] Additional disk space (approx. 5 times the size of the exe) [-] Platform dependent Thank you very much for your consideration. Kind regards, Henri [1] https://www.tug.org/pipermail/tex-live/2016-June/038664.html [2] https://stackoverflow.com/questions/866721 From karl at freefriends.org Thu Jun 9 20:07:45 2016 From: karl at freefriends.org (Karl Berry) Date: Thu, 9 Jun 2016 18:07:45 +0000 Subject: [tex-live] Debugging symbols for binaries In-Reply-To: <575977C1.5020205@gmail.com> Message-ID: <201606091807.u59I7jUf003435@freefriends.org> Hi Henri - thanks for the suggestion and the links. I am sympathetic to the idea, as I have been stymied myself plenty of times on crashes with various distro-supplied binaries which (of course) have no symbols and are not feasible to recompile. (Or useful, as most likely the crash is not easily reproducible anyway.) I have the vague impression that Debian started supplying binaries with symbols in a sibling repo at one point, but I don't know. Anyway, back to TL, our standard build process does create the binaries with symbols; they're only stripped as part of going into the final bin/ directory. So it should be straightforward to make tarballs of the debuggable binaries available on the web. (I don't think it's worth it to try to include them in the distro or the repo.) I'll think about it and discuss it with the builders. Thanks again. P.S. As for the problem at hand, Akira came up with a patch (below). -k --- utils.c.orig Fri Feb 19 11:45:32 2016 +++ utils.c Mon Jun 06 07:06:49 2016 @@ -870,7 +870,8 @@ int size, len = 0; /* to avoid warning about uninitialized use of len */ boolean found = i < sub_match_count - && match_string != NULL && pmatch[i].rm_so >= 0 && i >= 0; + && match_string != NULL && pmatch[i].rm_so >= 0 && i >= 0 + && pmatch[i].rm_eo >= pmatch[i].rm_so; if (found) { len = pmatch[i].rm_eo - pmatch[i].rm_so; From karl at freefriends.org Thu Jun 9 23:47:32 2016 From: karl at freefriends.org (Karl Berry) Date: Thu, 9 Jun 2016 21:47:32 +0000 Subject: [tex-live] TeXLive 2016 source tar ball with object and library files? In-Reply-To: <20160607114939.2e6qzkxeex5l4ite@boole.suse.de> Message-ID: <201606092147.u59LlWmB028458@freefriends.org> due a lot of binary, object, and library files. Why this? I obviously made a mistake somewhere in my scripts. I'll do the same cutting you did and update the file when I have a chance. Sorry. -k From reinhard.kotucha at web.de Fri Jun 10 00:57:11 2016 From: reinhard.kotucha at web.de (Reinhard Kotucha) Date: Fri, 10 Jun 2016 00:57:11 +0200 Subject: [tex-live] Debugging symbols for binaries In-Reply-To: <201606091807.u59I7jUf003435@freefriends.org> References: <575977C1.5020205@gmail.com> <201606091807.u59I7jUf003435@freefriends.org> Message-ID: <22361.62535.96968.324708@zaphod.ms25.net> On 2016-06-09 at 18:07:45 +0000, Karl Berry wrote: > Hi Henri - thanks for the suggestion and the links. I am sympathetic to > the idea, as I have been stymied myself plenty of times on crashes with > various distro-supplied binaries which (of course) have no symbols and > are not feasible to recompile. (Or useful, as most likely the crash is > not easily reproducible anyway.) I have the vague impression that > Debian started supplying binaries with symbols in a sibling repo at one > point, but I don't know. > > Anyway, back to TL, our standard build process does create the binaries > with symbols; they're only stripped as part of going into the final bin/ > directory. So it should be straightforward to make tarballs of the > debuggable binaries available on the web. (I don't think it's worth it > to try to include them in the distro or the repo.) > > I'll think about it and discuss it with the builders. Thanks again. Hi Karl, is it feasible/reasonable to add an option --nostrip to the trunk/Build/source/Build shell script and adapt the makefiles accordingly? Since all we need is to avoid that the program strip(1) is run, I suppose that a platform-independent solution is possible. Regards, Reinhard -- ------------------------------------------------------------------ Reinhard Kotucha Phone: +49-511-3373112 Marschnerstr. 25 D-30167 Hannover mailto:reinhard.kotucha at web.de ------------------------------------------------------------------ From kakuto at fuk.kindai.ac.jp Fri Jun 10 06:48:56 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Fri, 10 Jun 2016 13:48:56 +0900 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf In-Reply-To: References: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> Message-ID: >> TEXMFCNF=E:/Documents/programmes_x86/texlive/2016; >> >> The trailing semicolon is important. If the semicolon does not exist, >> texmf.cnf is searched only in E:/Documents/programmes_x86/texlive/2016. > > 1) I've just tried this. My program seems to hang. > > 2) I tried E:/Documents/programmes_x86/texlive/2016: (with : at the > end) and I get Probably you are using a non-standard shell in which the semicolon is a special character. But the kpathsea for w32 requires DOS path separator, semicolon. So maybe the following works: export TEXMFCNF='E:/Documents/programmes_x86/texlive/2016;' Best, Akira From vincent.torri at gmail.com Fri Jun 10 06:57:49 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Fri, 10 Jun 2016 06:57:49 +0200 Subject: [tex-live] kpathsea: adding import library on Windows (and if possible header files too) Message-ID: Hello Would it be possible to add the import library for kpathsea on Windows, when creating the DLL ? if gcc is used, then it is easy. The function are already declared with the dllexport storage class attribute, so when creating the DLL, you can do something like that in the make rule: $(SHAREDLIB): $(OBJS) $(CC) -shared -Wl,--out-implib,$(IMPLIB) $(LDFLAGS) -o $@ $(OBJS) $(STRIP) $@ usually, the name of the import library, stored in $(IMPLIB), is libkpathsea.dll.a with gcc and is installed in a lib/ subdirectory. Here, TeX Live is a bit different from what is commonly used in other projects : the binaries are not in a bin/ subdir but in bin/win32/. If you follow the same scheme, the import library would be installed in lib/win32/ Changing this to win32/bin and win32/lib would be a bit more what is commonly used. Anyway, in whatever directory you install the files, it would be nice to have such import library on Windows thank you Vincent Torri From vincent.torri at gmail.com Fri Jun 10 07:09:37 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Fri, 10 Jun 2016 07:09:37 +0200 Subject: [tex-live] 64 bits version of TeX Live on Windows Message-ID: Hello I've downloaded the installation binary from https://www.tug.org/texlive/acquire-netinstall.html, and it installs a 32 bits version of TeX Live I would like to know if there is a 64 bits Windows version of TeX Live. thank you Vincent Torri From vincent.torri at gmail.com Fri Jun 10 07:14:13 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Fri, 10 Jun 2016 07:14:13 +0200 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf In-Reply-To: References: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> Message-ID: Hello On Fri, Jun 10, 2016 at 6:48 AM, Akira Kakuto wrote: >>> TEXMFCNF=E:/Documents/programmes_x86/texlive/2016; >>> >>> The trailing semicolon is important. If the semicolon does not exist, >>> texmf.cnf is searched only in E:/Documents/programmes_x86/texlive/2016. >> >> >> 1) I've just tried this. My program seems to hang. >> >> 2) I tried E:/Documents/programmes_x86/texlive/2016: (with : at the >> end) and I get > > > Probably you are using a non-standard shell in which > the semicolon is a special character. > But the kpathsea for w32 requires DOS path separator, semicolon. > So maybe the following works: > export TEXMFCNF='E:/Documents/programmes_x86/texlive/2016;' I use MSYS2. It indeed suppresses the warning, but with that code: char *name = "cmsy10.tfm"; kpathsea kpse; char *n; kpse = kpathsea_new(); kpathsea_set_program_name(kpse, "dvi.exe", NULL); kpathsea_init_prog (kpse, "dvi.exe", 300, NULL, NULL); n = kpathsea_find_file(kpse, name, kpse_tfm_format, 1); kpathsea_finish(kpse); n is NULL. If I replace "dvi.exe" with "kpsewhich.exe", n is filled with the correct value. Any idea why ? thank you Vincent Torri From kakuto at fuk.kindai.ac.jp Fri Jun 10 07:27:52 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Fri, 10 Jun 2016 14:27:52 +0900 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: References: Message-ID: <6172301186B14EC8B7F17C04AE3A0C79@CJ3001517A> > I would like to know if there is a 64 bits Windows version of TeX Live. See CTAN/systems/win32/w32tex/TLW64/ Please read 00README.TLW64 carefully. Best, Akira From dbitouze at wanadoo.fr Fri Jun 10 08:04:10 2016 From: dbitouze at wanadoo.fr (=?utf-8?Q?Denis_Bitouz=C3=A9?=) Date: Fri, 10 Jun 2016 08:04:10 +0200 Subject: [tex-live] Error with TL 2015 (involving `floatrow') In-Reply-To: ("Denis \=\?utf-8\?Q\?Bitouz\=C3\=A9\?\= \=\?utf-8\?Q\?\=22's\?\= message of "Fri, 19 Jun 2015 22:49:58 +0200") References: Message-ID: Hi, Le 19/06/15 ? 22h49, Denis Bitouz? a ?crit : > the following M (non) WE example: > > \documentclass{article} > \usepackage{kpfonts} > \usepackage{booktabs} > \usepackage{siunitx} > \usepackage{pgfplots} > \usepackage{floatrow} > \begin{document} > Foo > \end{document} > > used to compile like a charm with TL 2014 but fails to compile with an > up to date TL 2015 with error: > > ????? > ? (/usr/local/texlive/2015/texmf-dist/tex/latex/floatrow/floatrow.sty > ? (/usr/local/texlive/2015/texmf-dist/tex/latex/caption/caption3.sty) > ? ! No room for a new \dimen . > ? \ch at ck ...\else \errmessage {No room for a new #3} > ? \fi > ? l.485 ...finable\flrow at foot{\newinsert\flrow at foot} > ? > ? ? > ????? I noticed that this bug, due to (quoting David Carlisle): "Line 485 of floatrow.sty should be \@ifdefinable\flrow at foot{\newbox\flrow at foot} not \@ifdefinable\flrow at foot{\newinsert\flrow at foot}" has been fixed in TL 2016 and even in an updated TL 2015 (great!), though `floatrow.sty' hasn't been updated (still 2008/08/02 v0.3b). Do you know what happened? Thanks in anticipation. -- Denis From d.p.carlisle at gmail.com Fri Jun 10 08:41:04 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Fri, 10 Jun 2016 07:41:04 +0100 Subject: [tex-live] Error with TL 2015 (involving `floatrow') In-Reply-To: References: Message-ID: On 10 June 2016 at 07:04, Denis Bitouz? wrote: > Hi, > > Le 19/06/15 ? 22h49, Denis Bitouz? a ?crit : > > > the following M (non) WE example: > > > > \documentclass{article} > > \usepackage{kpfonts} > > \usepackage{booktabs} > > \usepackage{siunitx} > > \usepackage{pgfplots} > > \usepackage{floatrow} > > \begin{document} > > Foo > > \end{document} > > > > used to compile like a charm with TL 2014 but fails to compile with an > > up to date TL 2015 with error: > > > > ????? > > ? (/usr/local/texlive/2015/texmf-dist/tex/latex/floatrow/floatrow.sty > > ? (/usr/local/texlive/2015/texmf-dist/tex/latex/caption/caption3.sty) > > ? ! No room for a new \dimen . > > ? \ch at ck ...\else \errmessage {No room for a new #3} > > ? \fi > > ? l.485 ...finable\flrow at foot{\newinsert\flrow at foot} > > ? > > ? ? > > ????? > > I noticed that this bug, due to (quoting David Carlisle): > > "Line 485 of floatrow.sty should be > > \@ifdefinable\flrow at foot{\newbox\flrow at foot} > > not > > \@ifdefinable\flrow at foot{\newinsert\flrow at foot}" > > has been fixed in TL 2016 and even in an updated TL 2015 (great!), > though `floatrow.sty' hasn't been updated (still 2008/08/02 v0.3b). > > Do you know what happened? > > Thanks in anticipation. > -- > Denis > > I did actually get an updated version of floatrow at the time but it never got to ctan (there were some communication/language issues that complicated things) but meanwhile we made latex aware that it's running in etex and has rather more registers available (although not more inserts actually) and so you are far less likely to get the no room for a new dimen error in latex releases after 2015/10/01. (None of this is related to texlive really: it's all macro layer changes and applies to any tex distribution). David -------------- next part -------------- An HTML attachment was scrubbed... URL: From angelo.graziosi at alice.it Fri Jun 10 15:15:18 2016 From: angelo.graziosi at alice.it (Angelo Graziosi) Date: Fri, 10 Jun 2016 15:15:18 +0200 Subject: [tex-live] 64 bits version of TeX Live on Windows Message-ID: Akira Kakuto wrote: > See CTAN/systems/win32/w32tex/TLW64/ > Please read 00README.TLW64 carefully. I wonder if we can install TL2016 from MSYS2 shell using the Perl script install-tl with the "-custom-bin" option. BTW, I tried to build TL2016 many months ago with MSYS2/MINGW64 [1] and there were some build issues.. I tried to call attention [2] but none took care of it. I think that TL W64 binaries can be created adding MSYS2/MINGW64 as new TL platform but it need the help of TL gurus.. (as we did when Cygwin was added...) This could/should be one of the goals of the next 2017 release.. Ciao, Angelo. --- [1] MSYS2 is a fork from Cygwin which aims to build apps natively for W64 (https://sourceforge.net/p/msys2/wiki/Home). [2] http://tug.org/mailman/htdig/tlbuild/2015q3/003317.html https://sourceforge.net/p/msys2/tickets/197 From mojca.miklavec.lists at gmail.com Fri Jun 10 15:43:13 2016 From: mojca.miklavec.lists at gmail.com (Mojca Miklavec) Date: Fri, 10 Jun 2016 15:43:13 +0200 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: References: Message-ID: On 10 June 2016 at 15:15, Angelo Graziosi wrote: > Akira Kakuto wrote: >> >> See CTAN/systems/win32/w32tex/TLW64/ >> Please read 00README.TLW64 carefully. > > I wonder if we can install TL2016 from MSYS2 shell using the Perl script > install-tl with the "-custom-bin" option. It would probably be nice to support installing TL under windows from a shell script (either from MSYS2 or from the new shell that Windows 10 has introduced). This alone doesn't require official addition of new binaries, but maybe some patches would be needed. However adding *both* binaries (one compiled with Visual Studio and one compiled with MinGW for the *same* architecture; please note that I'm not talking about 32 vs. 64-bit here) makes no sense in my opinion. (The cygwin binaries actually also have very little added value if you ask me; cygwin users could just as well install the other windows binaries from the shell script.) Supporting 64-bit windows binaries is another story though. Akira already compiles and provides 64-bit binaries built with Visual Studio, so there is no need to compile yet another set (you can of course do that at home and complain when something breaks). The problem is just that they are not officially included. I would also be in favour of adding them, but Karl is afraid of the added complexity with not that much "added value" (properly supporting 64-bit Windows also means providing 64-bit perl, ghostscript etc., doubling the space required for all those binaries and properly deciding whether the Windows in 32-bit or 64-bit, ...). > BTW, I tried to build TL2016 many months ago with MSYS2/MINGW64 [1] and > there were some build issues.. I tried to call attention [2] but none took > care of it. > https://sourceforge.net/p/msys2/tickets/197 It's not that TeX Live calls clang. It's ICU that calls clang. I would suggest you to try to compile the official ICU under the same window. I would say that this is a bug that ICU developers have to fix (or perhaps a bug in clang ? not impossible either; Akira is probably not using clang) and you probably need to complain to them. Just for fun you could try to uninstall clang[++] and make a new clean build attempt. Maybe you'll then succeed. I didn't know that mingw64-w64 was actually shipping clang[++] by default. I looked at it a while ago and only found gcc plus some unofficial builds or instructions that needed quite a bit of patience to make it work. > I think that TL W64 binaries can be created adding MSYS2/MINGW64 as new TL > platform but it need the help of TL gurus.. (as we did when Cygwin was > added...) There are several different unrelated goals: - making 64-bit mingw builds work out of the box (you should really try to pursue that further) - supporting installation of existing or "third-party" binaries from the msys shell - adding a new set of binaries and support to TL Karl was not particularly in favour of achieving the last goal, but fixing the build can be taken care of separately. Mojca From vincent.torri at gmail.com Fri Jun 10 15:48:13 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Fri, 10 Jun 2016 15:48:13 +0200 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: References: Message-ID: On Fri, Jun 10, 2016 at 3:15 PM, Angelo Graziosi wrote: > Akira Kakuto wrote: >> >> See CTAN/systems/win32/w32tex/TLW64/ >> Please read 00README.TLW64 carefully. > > > I wonder if we can install TL2016 from MSYS2 shell using the Perl script > install-tl with the "-custom-bin" option. > > BTW, I tried to build TL2016 many months ago with MSYS2/MINGW64 [1] and > there were some build issues.. I tried to call attention [2] but none took > care of it. > > I think that TL W64 binaries can be created adding MSYS2/MINGW64 as new TL > platform but it need the help of TL gurus.. (as we did when Cygwin was > added...) > > This could/should be one of the goals of the next 2017 release.. > > Ciao, > Angelo. > > --- > [1] MSYS2 is a fork from Cygwin which aims to build apps natively for W64 > (https://sourceforge.net/p/msys2/wiki/Home). > > [2] http://tug.org/mailman/htdig/tlbuild/2015q3/003317.html > https://sourceforge.net/p/msys2/tickets/197 Maybe this could help : http://stackoverflow.com/a/13792251/688348 Vincent Torri From kakuto at fuk.kindai.ac.jp Fri Jun 10 16:02:36 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Fri, 10 Jun 2016 23:02:36 +0900 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: References: Message-ID: Dear Mojca, > I would also be in favour of adding them, but Karl is afraid of the > added complexity with not that much "added value" (properly supporting > 64-bit Windows also means providing 64-bit perl, ghostscript etc., > doubling the space required for all those binaries and properly > deciding whether the Windows in 32-bit or 64-bit, ...). I prefer not to include 64bit binaries for windows in TeX Live, because it introduces unnecessary complexities, and WOW64 is great. I assume that the present TLW64 is enough for interested 64bit users. Best, Akira From d.p.carlisle at gmail.com Fri Jun 10 16:15:39 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Fri, 10 Jun 2016 15:15:39 +0100 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: References: Message-ID: On 10 June 2016 at 14:43, Mojca Miklavec wrote: > > The cygwin binaries actually also have very little added > value if you ask me; cygwin users could just as well install the other > windows binaries from the shell script.) > Just to comment on that aspect, personally I find the cygwin builds invaluable. Calling native windows binaries from cygwin is of course possible but since they have a different view of the filesystem paths, you always have to be aware of the file system mapping and so path searching, shell escape etc would never work in a natural way from a cygwin shell if using a native windows tex build, would they? I do this all the time with calling a native windows Java application from cygwin, but it's good not to have the same issues with tex. David -------------- next part -------------- An HTML attachment was scrubbed... URL: From angelo.graziosi at alice.it Fri Jun 10 16:26:15 2016 From: angelo.graziosi at alice.it (Angelo Graziosi) Date: Fri, 10 Jun 2016 16:26:15 +0200 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: References: Message-ID: <835e5d64-dbb1-7149-4c52-10133c517211@alice.it> Il 10/06/2016 15:43, Mojca Miklavec ha scritto: > > However adding *both* binaries (one compiled with Visual Studio and > one compiled with MinGW for the *same* architecture; please note that > I'm not talking about 32 vs. 64-bit here) makes no sense in my > opinion. (The cygwin binaries actually also have very little added > value if you ask me; cygwin users could just as well install the other > windows binaries from the shell script.) I would remove VS binaries in favor of Cygwin and MSYS2-MINGW64 ones. Notice that Cygwin already gives a method to install and use TL in the same manner one uses TL on GNU/Linux and OSX (I don't use MacTex): same scripts to setup PATHs, same build documents commands and so on... With MSYS2-MINGW64 would be the same, almost.. > > I would also be in favour of adding them, but Karl is afraid of the > added complexity with not that much "added value" (properly supporting > 64-bit Windows also means providing 64-bit perl, ghostscript etc., > doubling the space required for all those binaries and properly > deciding whether the Windows in 32-bit or 64-bit, ...). why then having double binaries for other platforms (i386/x86_64 for linux etc..) BTW, having the same app running on different systems helps in finding subtle bugs.. > > It's not that TeX Live calls clang. It's ICU that calls clang. > > I would suggest you to try to compile the official ICU under the same > window. I would say that this is a bug that ICU developers have to fix > (or perhaps a bug in clang ? not impossible either; Akira is probably > not using clang) and you probably need to complain to them. Just for > fun you could try to uninstall clang[++] and make a new clean build > attempt. Maybe you'll then succeed. No, the fix was to build TL using a command like this (I did this many months ago and don't remember exactly): CC=gcc ./Build... ## force to use GCC that helped in building ICU but then the build stopped in another step of the build.. In any case, my was only a suggestion. If none of the TL guys take care of this, let's stop to discuss it, then. Angelo From kakuto at fuk.kindai.ac.jp Fri Jun 10 16:26:01 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Fri, 10 Jun 2016 23:26:01 +0900 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf In-Reply-To: References: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> Message-ID: <3601A79C6A3143F7BE85A8E6948167D2@CJ3001517A> > Any idea why ? Because TEXMF in TFMFONTS = .;{$TEXMF/fonts,$VARTEXFONTS}/tfm// is defined by using $SELFAUTOLOC, the directory where the exec file exists. As you find in this exapmle, you have to install your program into bin/win32 in order to fully use the kpathsea. Best, Akira From zdenek.wagner at gmail.com Fri Jun 10 16:49:06 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Fri, 10 Jun 2016 16:49:06 +0200 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: <835e5d64-dbb1-7149-4c52-10133c517211@alice.it> References: <835e5d64-dbb1-7149-4c52-10133c517211@alice.it> Message-ID: Is it possible to use cygwin binaries without cygwin? I know a lot of Windows users who do not need cygwin at all and do not need it. It would be very bad to inform them that TL will cease to work. You will send them to MiKTeX as the only Windows friendly distributin in such a case. Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz 2016-06-10 16:26 GMT+02:00 Angelo Graziosi : > > > Il 10/06/2016 15:43, Mojca Miklavec ha scritto: > >> >> However adding *both* binaries (one compiled with Visual Studio and >> one compiled with MinGW for the *same* architecture; please note that >> I'm not talking about 32 vs. 64-bit here) makes no sense in my >> opinion. (The cygwin binaries actually also have very little added >> value if you ask me; cygwin users could just as well install the other >> windows binaries from the shell script.) >> > > I would remove VS binaries in favor of Cygwin and MSYS2-MINGW64 ones. > Notice that Cygwin already gives a method to install and use TL in the same > manner one uses TL on GNU/Linux and OSX (I don't use MacTex): same scripts > to setup PATHs, same build documents commands and so on... With > MSYS2-MINGW64 would be the same, almost.. > > >> I would also be in favour of adding them, but Karl is afraid of the >> added complexity with not that much "added value" (properly supporting >> 64-bit Windows also means providing 64-bit perl, ghostscript etc., >> doubling the space required for all those binaries and properly >> deciding whether the Windows in 32-bit or 64-bit, ...). >> > > why then having double binaries for other platforms (i386/x86_64 for linux > etc..) > > BTW, having the same app running on different systems helps in finding > subtle bugs.. > > > >> It's not that TeX Live calls clang. It's ICU that calls clang. >> >> I would suggest you to try to compile the official ICU under the same >> window. I would say that this is a bug that ICU developers have to fix >> (or perhaps a bug in clang ? not impossible either; Akira is probably >> not using clang) and you probably need to complain to them. Just for >> fun you could try to uninstall clang[++] and make a new clean build >> attempt. Maybe you'll then succeed. >> > > No, the fix was to build TL using a command like this (I did this many > months ago and don't remember exactly): > > CC=gcc ./Build... ## force to use GCC > > that helped in building ICU but then the build stopped in another step of > the build.. > > In any case, my was only a suggestion. If none of the TL guys take care of > this, let's stop to discuss it, then. > > > > Angelo > -------------- next part -------------- An HTML attachment was scrubbed... URL: From enrico.gregorio at univr.it Fri Jun 10 17:07:13 2016 From: enrico.gregorio at univr.it (Enrico Gregorio) Date: Fri, 10 Jun 2016 17:07:13 +0200 Subject: [tex-live] Xdvi not working Message-ID: Situation: TeX Live 2016 (MacTeX) on Mac OS X (10.11.5); XQuartz 2.7.9 Running tex ?story\bye' xdvi story results in a segmentation fault Warning: locale not supported by Xlib, locale set to C xdvi-xaw 22.87.03 (Xaw toolkit): ../../../texk/xdvik/events.c:4450: Shouldn't happen: Segmentation fault - trying to clean up and aborting ... Abort trap: 6 The issue affects all other Xdvi binaries I have on my machine (TL 2012, 2013, 2014 and 2015). On the TeX.SX chat, Joseph Wright told it worked for him, until he upgraded from XQuartz 2.7.8 to 2.7.9. The release notes of XQuartz 2.7.9 tell that some fixes have been done on libXaw, but I could not gather much about them: ? libXaw ? Remove incorrect export of vendorShellClassRec and vendorShellWidgetClass Maybe some expert can shed light on the matter Ciao Enrico From angelo.graziosi at alice.it Fri Jun 10 17:11:51 2016 From: angelo.graziosi at alice.it (Angelo Graziosi) Date: Fri, 10 Jun 2016 17:11:51 +0200 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: References: Message-ID: <3aab8ac6-06f8-0196-9909-471cccd8d088@alice.it> Il 10/06/2016 15:15, Angelo Graziosi ha scritto: > I wonder if we can install TL2016 from MSYS2 shell using the Perl script > install-tl with the "-custom-bin" option. I tried this: $ /e/tlrepo/install-tl -custom-bin bin/win64 being /e/tlrepo a copy (rsync) of a CTAN/tlnet repository of TL2016. It fails in post-install step with [...] running fmtutil-sys --no-error-if-no-engine=luajittex,mfluajit --all ...done /e/tlrepo/install-tl: fmtutil-sys --no-error-if-no-engine=luajittex,mfluajit --all failed: running package-specific postactions /tlpkg/tlpostcode/dvipdfmx.pl": No such file or directory /e/tlrepo/install-tl: calling post action script tlpkg/tlpostcode/dvipdfmx.pl did not succeed in /home/angelo, status 2 at /e/tlrepo/tlpkg/TeXLive/TLUtils.pm line 1575. /tlpkg/tlpostcode/ptex2pdf-tlpost.pl": No such file or directory /e/tlrepo/install-tl: calling post action script tlpkg/tlpostcode/ptex2pdf-tlpost.pl did not succeed in /home/angelo, status 2 at /e/tlrepo/tlpkg/TeXLive/TLUtils.pm line 1575. /tlpkg/tlpostcode/xetex.pl": No such file or directory /e/tlrepo/install-tl: calling post action script tlpkg/tlpostcode/xetex.pl did not succeed in /home/angelo, status 2 at /e/tlrepo/tlpkg/TeXLive/TLUtils.pm line 1575. finished with package-specific postactions See /opt/texlive/index.html for links to documentation. The TeX Live web site contains updates and corrections: http://tug.org/texlive. TeX Live is a joint project of the TeX user groups around the world; please consider supporting it by joining the group best for you. The list of user groups is on the web at http://tug.org/usergroups.html. Add /opt/texlive/texmf-dist/doc/info to INFOPATH. Add /opt/texlive/texmf-dist/doc/man to MANPATH (if not dynamically found). Most importantly, add /opt/texlive/bin/custom to your PATH for current and future sessions. [...] I am not sure if the failure depend on the installation method or somewhere else.. Angelo From angelo.graziosi at alice.it Fri Jun 10 17:26:40 2016 From: angelo.graziosi at alice.it (Angelo Graziosi) Date: Fri, 10 Jun 2016 17:26:40 +0200 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: References: <835e5d64-dbb1-7149-4c52-10133c517211@alice.it> Message-ID: Il 10/06/2016 16:49, Zdenek Wagner ha scritto: > Is it possible to use cygwin binaries without cygwin? I know a lot of > Windows users who do not need cygwin at all and do not need it. It would be > very bad to inform them that TL will cease to work. You will send them to > MiKTeX as the only Windows friendly distributin in such a case. Is it possible to use Windows binaries without Windows? Don't they need the VS DLL? Also the win32 binaries contain the msvcr*dll. In any case, my proposal to remove VS binaries was quite rhetoric.. Angelo From zdenek.wagner at gmail.com Fri Jun 10 17:46:32 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Fri, 10 Jun 2016 17:46:32 +0200 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: References: <835e5d64-dbb1-7149-4c52-10133c517211@alice.it> Message-ID: 2016-06-10 17:26 GMT+02:00 Angelo Graziosi : > > > Il 10/06/2016 16:49, Zdenek Wagner ha scritto: > >> Is it possible to use cygwin binaries without cygwin? I know a lot of >> Windows users who do not need cygwin at all and do not need it. It would >> be >> very bad to inform them that TL will cease to work. You will send them to >> MiKTeX as the only Windows friendly distributin in such a case. >> > > Is it possible to use Windows binaries without Windows? Don't they need > the VS DLL? Also the win32 binaries contain the msvcr*dll. > > In any case, my proposal to remove VS binaries was quite rhetoric.. > If you can include just a few cygwin DLL's that will be installed as a part of TL, it will be fine. > > > Angelo > Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz -------------- next part -------------- An HTML attachment was scrubbed... URL: From vincent.torri at gmail.com Fri Jun 10 19:05:47 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Fri, 10 Jun 2016 19:05:47 +0200 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf In-Reply-To: <3601A79C6A3143F7BE85A8E6948167D2@CJ3001517A> References: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> <3601A79C6A3143F7BE85A8E6948167D2@CJ3001517A> Message-ID: Is it a problem to use "kpsewhich.exe" when I call kpathsea_set_program_name() ? thank you Vincent Torri On Fri, Jun 10, 2016 at 4:26 PM, Akira Kakuto wrote: >> Any idea why ? > > > Because TEXMF in > TFMFONTS = .;{$TEXMF/fonts,$VARTEXFONTS}/tfm// > is defined by using $SELFAUTOLOC, the directory where > the exec file exists. > As you find in this exapmle, you have to install your > program into bin/win32 in order to fully use the kpathsea. > > Best, > Akira > From faro49 at gmail.com Fri Jun 10 20:40:26 2016 From: faro49 at gmail.com (Roberto Fanciulli) Date: Fri, 10 Jun 2016 20:40:26 +0200 Subject: [tex-live] TexLive manager does not load any repository Message-ID: I use WinEdt 10.1 Build: 20,160,526 (v. 10.1) - 32-bit on Texlive, under Windows 10. After several days of inactivity I wanted to update the Tex packages. In WinEDT I started Texlive manager 2015, but I can not load any repository. Texlive manager can not load any database, in the window the repository name remains always fixed on the same server even if I try to change it. No database server is loaded and no package of the remote server appears. The debug log file contains: *Loading local TeX Live database (C:/texlive/2014/tlpkg/texlive.tlpdb)This may take some time, please be patient ...... caricamento completo.D:TLPDB.pm: trying to initialize from ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb D:trying to download ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb.xz to C:\Users\User\AppData\Local\Temp\WGcc0d54MbD:persistent connection set up, trying to get ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb.xz (for C:\Users\User\AppData\Local\Temp\WGcc0d54Mb)D:downloading file via persistent connection succeededD:un-xzing C:\Users\User\AppData\Local\Temp\WGcc0d54Mb to C:\Users\User\AppData\Local\Temp\N1ElYvHivbD:found the uncompressed xz fileD:TLPDB.pm: trying to initialize from ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb D:trying to download ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb.xz to C:\Users\User\AppData\Local\Temp\R_sWKng9bfD:persistent connection set up, trying to get ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb.xz (for C:\Users\User\AppData\Local\Temp\R_sWKng9bf)D:downloading file via persistent connection succeededD:un-xzing C:\Users\User\AppData\Local\Temp\R_sWKng9bf to C:\Users\User\AppData\Local\Temp\Xr0UI3jgQrD:found the uncompressed xz file* What can I do to restore normal operation of the TeXlive manager? Thanks! -------------- next part -------------- An HTML attachment was scrubbed... URL: From ziegenhagen at gmail.com Fri Jun 10 23:59:17 2016 From: ziegenhagen at gmail.com (Uwe Ziegenhagen) Date: Fri, 10 Jun 2016 23:59:17 +0200 Subject: [tex-live] TexLive manager does not load any repository In-Reply-To: References: Message-ID: I recommend a new installation of TeX Live 2016 as there are no more updates for 2014 and 2015. 2016 is brandnew. Uwe 2016-06-10 20:40 GMT+02:00 Roberto Fanciulli : > I use WinEdt 10.1 Build: 20,160,526 (v. 10.1) - 32-bit on Texlive, under > Windows 10. > After several days of inactivity I wanted to update the Tex packages. > In WinEDT I started Texlive manager 2015, but I can not load any > repository. > Texlive manager can not load any database, in the window the repository > name remains always fixed on the same server even if I try to change it. No > database server is loaded and no package of the remote server appears. > The debug log file contains: > > > > > > > > > > > > > > > > > *Loading local TeX Live database > (C:/texlive/2014/tlpkg/texlive.tlpdb)This may take some time, please be > patient ...... caricamento completo.D:TLPDB.pm: trying to initialize from > ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb > D:trying > to download > ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb.xz > > to C:\Users\User\AppData\Local\Temp\WGcc0d54MbD:persistent connection set > up, trying to get > ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb.xz > > (for C:\Users\User\AppData\Local\Temp\WGcc0d54Mb)D:downloading file via > persistent connection succeededD:un-xzing > C:\Users\User\AppData\Local\Temp\WGcc0d54Mb to > C:\Users\User\AppData\Local\Temp\N1ElYvHivbD:found the uncompressed xz > fileD:TLPDB.pm: trying to initialize from > ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb > D:trying > to download > ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb.xz > > to C:\Users\User\AppData\Local\Temp\R_sWKng9bfD:persistent connection set > up, trying to get > ftp://ftp.rrzn.uni-hannover.de/pub/mirror/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb.xz > > (for C:\Users\User\AppData\Local\Temp\R_sWKng9bf)D:downloading file via > persistent connection succeededD:un-xzing > C:\Users\User\AppData\Local\Temp\R_sWKng9bf to > C:\Users\User\AppData\Local\Temp\Xr0UI3jgQrD:found the uncompressed xz file* > > What can I do to restore normal operation of the TeXlive manager? > Thanks! > -- Dr. Uwe Ziegenhagen -------------- next part -------------- An HTML attachment was scrubbed... URL: From karl at freefriends.org Sat Jun 11 00:11:57 2016 From: karl at freefriends.org (Karl Berry) Date: Fri, 10 Jun 2016 22:11:57 +0000 Subject: [tex-live] Xdvi not working In-Reply-To: Message-ID: <201606102211.u5AMBvtk026341@freefriends.org> tex ????story\bye' xdvi story Works fine for me on i386-linux. Situation: TeX Live 2016 (MacTeX) on Mac OS X (10.11.5); XQuartz 2.7.9 I'll tell Dick. -k From richardmoncriefkoch at icloud.com Sat Jun 11 05:03:50 2016 From: richardmoncriefkoch at icloud.com (Richard Koch) Date: Fri, 10 Jun 2016 20:03:50 -0700 Subject: [tex-live] Xdvi not working Message-ID: <34FE27CE-9859-4C00-BCE1-37286685EA17@icloud.com> Folks, As already noted on this list, xdvi works if you have XQuartz 2.7.8, but fails if you updated to XQuartz 2.7.9, which was released on May 5, 2016. Just now, I installed XQuartz 2.7.9 on El Capitan and recompiled. Then I installed the new xdvi and xdvi-xaw. After that, xdvi ran fine on El Capitan. But there is a complication. I actually compile 64 bit binaries on Snow Leopard, so the binaries will run on many different systems. That machine had XQuartz 2.3.6. It turns out that XQuartz 2.7.9 can be installed on Snow Leopard (it is the oldest version of OS X supported by XQuartz). I?ve updated and am compiling xdvi now. With luck, this version will work on all the current 64 bit architectures if XQuartz is updated to the current version. (With even greater luck, perhaps the binary will work without updating XQuartz). A lot more testing is required, however, because Ghostscript in MacTeX was compiled both with and without X11 support, and the X11 support used XQuartz 2.3.6. We wouldn?t want to fix xdvi for the very few people who use it, but break Ghostscript. Next week is WWDC, which I?m going to watch from home using appleTV. This note explains that we are looking at the xdvi problem, but doesn?t promise a fix very soon. Dick Koch From Herbert.Voss at FU-Berlin.DE Sat Jun 11 11:23:35 2016 From: Herbert.Voss at FU-Berlin.DE (Herbert Voss) Date: Sat, 11 Jun 2016 11:23:35 +0200 Subject: [tex-live] update failure Message-ID: <575BD897.5040904@FU-Berlin.DE> Is it a problem with that mirror? It works if I choose for example dante.de Herbert voss at shania:~/Dropbox/DTK/.Versand> tlmgr --self --all update /usr/local/texlive/2016/tlpkg/installer/xz/xzdec.x86_64-linux: (stdin): File format not recognized /usr/local/texlive/current/bin/x86_64-linux/tlmgr: signature verification error of /tmp/H6cVJlGfHy/Y8fxEmWIBt from http://get-software.net/systems/texlive/tlnet#main/tlpkg/texlive.tlpdb: cryptographic signature verification of /tmp/H6cVJlGfHy/x2tAb_xZWh against http://get-software.net/systems/texlive/tlnet#main/tlpkg/texlive.tlpdb.sha512.asc failed. Output was gpg: no valid OpenPGP data found. gpg: the signature could not be verified. Please remember that the signature file (.sig or .asc) should be the first file given on the command line. Please report to texlive at tug.org From preining at logic.at Sat Jun 11 11:25:01 2016 From: preining at logic.at (Norbert Preining) Date: Sat, 11 Jun 2016 18:25:01 +0900 Subject: [tex-live] TexLive manager does not load any repository In-Reply-To: References: Message-ID: <20160611092501.GF25579@logic.at> > In WinEDT I started Texlive manager 2015, but I can not load any repository. TL 2015? Your log file says 2014? Both of them are not updated anymore and the repositories don't work unless you change to a copy of the historic repos of old TL releases. > What can I do to restore normal operation of the TeXlive manager? Update to TL2016. There are anyway no package updates for 2014 or 2015. So no need for tlmgr ;-) Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From preining at logic.at Sat Jun 11 11:35:07 2016 From: preining at logic.at (Norbert Preining) Date: Sat, 11 Jun 2016 18:35:07 +0900 Subject: [tex-live] update failure In-Reply-To: <575BD897.5040904@FU-Berlin.DE> References: <575BD897.5040904@FU-Berlin.DE> Message-ID: <20160611093507.GH25579@logic.at> > /usr/local/texlive/2016/tlpkg/installer/xz/xzdec.x86_64-linux: > (stdin): File format not recognized Seems to be an update in progress - the texlive.tlpdb.xz was not fully updated and thus xzdec complained - and then gpg complained ;-) Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From Rolf.Niepraschk at gmx.de Sat Jun 11 11:37:02 2016 From: Rolf.Niepraschk at gmx.de (Rolf Niepraschk) Date: Sat, 11 Jun 2016 11:37:02 +0200 Subject: [tex-live] update failure In-Reply-To: <575BD897.5040904@FU-Berlin.DE> References: <575BD897.5040904@FU-Berlin.DE> Message-ID: <16593d96-c46e-f7ee-be08-01193ca88ae1@gmx.de> Am 11.06.2016 um 11:23 schrieb Herbert Voss: > Is it a problem with that mirror? > It works if I choose for example dante.de > No Problem here with (automatically selected) ftp://ftp.tu-chemnitz.de/pub/tex/systems/texlive/tlnet ...Rolf From Herbert.Voss at FU-Berlin.DE Sat Jun 11 11:41:12 2016 From: Herbert.Voss at FU-Berlin.DE (Herbert Voss) Date: Sat, 11 Jun 2016 11:41:12 +0200 Subject: [tex-live] update failure In-Reply-To: <20160611093507.GH25579@logic.at> References: <575BD897.5040904@FU-Berlin.DE> <20160611093507.GH25579@logic.at> Message-ID: <575BDCB8.30402@FU-Berlin.DE> Am 11.06.2016 um 11:35 schrieb Norbert Preining: >> /usr/local/texlive/2016/tlpkg/installer/xz/xzdec.x86_64-linux: >> (stdin): File format not recognized > > Seems to be an update in progress - the texlive.tlpdb.xz was not fully > updated and thus xzdec complained - and then gpg complained ;-) The files look okay to me: -rwxr-xr-x 1 voss users 74832 17. Apr 20:39 xzdec.x86_64-linux -rwxr-xr-x 1 voss users 198656 17. Apr 20:39 xz.x86_64-linux I had the problem always with that mirror and not with another one. Herbert From preining at logic.at Sat Jun 11 11:48:37 2016 From: preining at logic.at (Norbert Preining) Date: Sat, 11 Jun 2016 18:48:37 +0900 Subject: [tex-live] update failure In-Reply-To: <575BDCB8.30402@FU-Berlin.DE> References: <575BD897.5040904@FU-Berlin.DE> <20160611093507.GH25579@logic.at> <575BDCB8.30402@FU-Berlin.DE> Message-ID: <20160611094837.GA10246@logic.at> On Sat, 11 Jun 2016, Herbert Voss wrote: > -rwxr-xr-x 1 voss users 74832 17. Apr 20:39 xzdec.x86_64-linux > -rwxr-xr-x 1 voss users 198656 17. Apr 20:39 xz.x86_64-linux Not those, the texlive.tlpdb.xz which is updated on every sync. xzdec tries to decompress it and finds rubbish, thus it errors out. > I had the problem always with that mirror and not with > another one. Interesting. Let me check .. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From preining at logic.at Sat Jun 11 11:50:38 2016 From: preining at logic.at (Norbert Preining) Date: Sat, 11 Jun 2016 18:50:38 +0900 Subject: [tex-live] update failure In-Reply-To: <575BD897.5040904@FU-Berlin.DE> References: <575BD897.5040904@FU-Berlin.DE> Message-ID: <20160611095038.GB10246@logic.at> On Sat, 11 Jun 2016, Herbert Voss wrote: > verification error of /tmp/H6cVJlGfHy/Y8fxEmWIBt from http://get-software.net/systems/texlive/tlnet#main/tlpkg/texlive.tlpdb: Aemm something is strange here ... why do you have the #main/tlpkg/texlive.tlpdb attached to the URL? Did you edit the texlive.tlpdb by hand? Can you show me the output of tlmgr option tlmgr repository and/or the head of your texlive.tlpdb (yours). Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From Herbert.Voss at FU-Berlin.DE Sat Jun 11 11:54:38 2016 From: Herbert.Voss at FU-Berlin.DE (Herbert Voss) Date: Sat, 11 Jun 2016 11:54:38 +0200 Subject: [tex-live] update failure In-Reply-To: <20160611095038.GB10246@logic.at> References: <575BD897.5040904@FU-Berlin.DE> <20160611095038.GB10246@logic.at> Message-ID: <575BDFDE.4080003@FU-Berlin.DE> Am 11.06.2016 um 11:50 schrieb Norbert Preining: > On Sat, 11 Jun 2016, Herbert Voss wrote: >> verification error of /tmp/H6cVJlGfHy/Y8fxEmWIBt from http://get-software.net/systems/texlive/tlnet#main/tlpkg/texlive.tlpdb: > > Aemm something is strange here ... why do you have the > #main/tlpkg/texlive.tlpdb > attached to the URL? Ah, I got it ... A problem with adding another repository. And I suppose that I did something wrong with using "tlmgr add ...". But cannot remember what I did wrong. However, I deleted the wrong line and now its ok. Sorry for the noise ... Herbert From bernhard.kleine at gmx.net Sat Jun 11 12:35:59 2016 From: bernhard.kleine at gmx.net (Bernhard Kleine) Date: Sat, 11 Jun 2016 12:35:59 +0200 Subject: [tex-live] sanitize-umlaut in CTAN but not in texlive In-Reply-To: References: <4762883f-0a8e-3927-2e20-add5626827f4@gmx.net> Message-ID: Am 09.06.2016 um 15:53 schrieb Ulrike Fischer: > Am Thu, 9 Jun 2016 09:22:06 +0200 schrieb Bernhard Kleine: > >> Is there an issue with licences? > No. Probably texlive simply didn't yet catch up all new packages. > After the freeze it can take some time until everything is there. It seems to be forgotten. I have updated every day and packages where updates arrived much later are updated, and new packages are installed as well, however, sanitize-umlaut is not picked up. Maybe that could change? -- spitzhalde9 79853 lenzkirch bernhard.kleine at gmx.net www.b-kleine.com, www.urseetal.net - thunderbird mit enigmail GPG schl?ssel: D5257409 fingerprint: 08 B7 F8 70 22 7A FC C1 15 49 CA A6 C7 6F A0 2E D5 25 74 09 -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: OpenPGP digital signature URL: From preining at logic.at Sat Jun 11 16:07:42 2016 From: preining at logic.at (Norbert Preining) Date: Sat, 11 Jun 2016 23:07:42 +0900 Subject: [tex-live] sanitize-umlaut in CTAN but not in texlive In-Reply-To: References: <4762883f-0a8e-3927-2e20-add5626827f4@gmx.net> Message-ID: <2CCC88D9-5239-4F66-A9FE-71C9945A9876@logic.at> Please calm down and give us (Karl) some time. On June 11, 2016 7:35:59 PM GMT+09:00, Bernhard Kleine wrote: >Am 09.06.2016 um 15:53 schrieb Ulrike Fischer: >> Am Thu, 9 Jun 2016 09:22:06 +0200 schrieb Bernhard Kleine: >> >>> Is there an issue with licences? >> No. Probably texlive simply didn't yet catch up all new packages. >> After the freeze it can take some time until everything is there. >It seems to be forgotten. I have updated every day and packages where >updates arrived much later are updated, and new packages are installed >as well, however, sanitize-umlaut is not picked up. Maybe that could >change? > > >-- >spitzhalde9 >79853 lenzkirch >bernhard.kleine at gmx.net >www.b-kleine.com, www.urseetal.net >- >thunderbird mit enigmail >GPG schl?ssel: D5257409 >fingerprint: >08 B7 F8 70 22 7A FC C1 15 49 CA A6 C7 6F A0 2E D5 25 74 09 -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From preining at logic.at Sat Jun 11 16:48:51 2016 From: preining at logic.at (Norbert Preining) Date: Sat, 11 Jun 2016 23:48:51 +0900 Subject: [tex-live] TexLive manager does not load any repository In-Reply-To: References: <20160611092501.GF25579@logic.at> Message-ID: <20160611144851.GI10246@logic.at> Hi Please keep the list in Cc > Thanks for your help. Texlive 2016 properly updates the packages. Good to hear. > However, because I want to keep both installed MiKTeX and Texlive, I would Fine with us. > like to ask you how to eliminate the following warning that is shown in the > Warning: conflicting [pdf] tex program found on the > system path in C: \ Program Files \ MiKTeX 2.9 \ miktex \ bin \; No. There is a reason this warning is there, and it is to warn people that having multiple (epsecially old, not here though!) tex programs hanging around will give a bad outcome. This warning has a reason, is correct, and thus will not be removed. BTW, how do you switch between MikTeX and TeX Live? Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From ml at mariochiari.net Sat Jun 11 16:52:07 2016 From: ml at mariochiari.net (mario chiari) Date: Sat, 11 Jun 2016 16:52:07 +0200 Subject: [tex-live] Unknown directive ...containerchecksum In-Reply-To: References: <87twhxaoqv.fsf@skimble.plus.com> Message-ID: <1465656727.3811.1.camel@mariochiari.net> hi I am suffering the same issue. Which script should I get, which server should I get it from? thanks cheers mario On Mon, 2016-05-16 at 18:51 +0100, David Carlisle wrote: > On 16 May 2016 at 18:36, Sharon Kimble wrote: > > > > I had a problem with tlmgr, so I used ' tlmgr update --self' which > > failed so I then tried ' ./update-tlmgr-latest.sh' which failed with > > this error report - > > Did you get??update-tlmgr-latest.sh from the tlpretest mirror, not the > one linked from the top level texlive page (which is 2015 version) > I've seen that error after makingtthat mistake myself:-) > > David > > > > > > > --8<---------------cut here---------------start------------->8--- > > tlmgr update --self > > Unknown directive ...containerchecksum > > c59200574a316416a23695c258edf3a32531fbda43ccdc09360ee105c3f07f9fb77df17c4ba4 > > c2ea4f3a5ea6667e064b51e3d8c2fe6c984ba3e71b4e32716955... , please fix it! at > > /usr/local/texlive/2016/tlpkg/TeXLive/TLPOBJ.pm line 210, line 36. > > --8<---------------cut here---------------end--------------->8--- > > > > Now tlmgr is failing to work. Is it possible to copy over tlmgr from > > texlive 2015 to get things working properly again? If not, how do I get > > texlive2016-pretest working again please? > > > > Thanks > > Sharon. > > -- > > A taste of linux = http://www.sharons.org.uk > > TGmeds = http://www.tgmeds.org.uk > > Debian 8.4, fluxbox 1.3.7, emacs 25.0.93 > > From preining at logic.at Sat Jun 11 17:09:36 2016 From: preining at logic.at (Norbert Preining) Date: Sun, 12 Jun 2016 00:09:36 +0900 Subject: [tex-live] Unknown directive ...containerchecksum In-Reply-To: <1465656727.3811.1.camel@mariochiari.net> References: <87twhxaoqv.fsf@skimble.plus.com> <1465656727.3811.1.camel@mariochiari.net> Message-ID: <20160611150936.GJ10246@logic.at> On Sat, 11 Jun 2016, mario chiari wrote: > I am suffering the same issue. Since you sent this email from a host called fedora-pc I assume you are on Linux (but this is a guess ..) What is the output of tlmgr version If it says ... TeX Live (http://tug.org/texlive) version 2016 then ... > Which script should I get, which server should I get it from? get http://mirror.ctan.org/systems/texlive/tlnet/update-tlmgr-latest.sh then run bash update-tlmgr-latest.sh --info That should give you lots of output, but most importatntly: update-tlmgr-r41280.sh somewhere. That is the correct one (or any rNNNNNN that is bigger than 40000) After that, make sure you have tlmgr in the path which tlmgr should return the path to your tlmgr. And finally, run bash update-tlmgr-latest.sh (Can we post this somewhere that I don't have to retype it???) Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From ml at mariochiari.net Sat Jun 11 17:15:28 2016 From: ml at mariochiari.net (mario chiari) Date: Sat, 11 Jun 2016 17:15:28 +0200 Subject: [tex-live] http://ctan.mirror.garr.it/mirrors/CTAN/systems/texlive/tlnet/update-tlmgr-latest.sh Message-ID: <1465658128.3811.2.camel@mariochiari.net> Hi is?http://ctan.mirror.garr.it/mirrors/CTAN/systems/texlive/tlnet/update-tlmgr-la test.sh?broken (for texlive2015). thanks m. From preining at logic.at Sat Jun 11 17:25:50 2016 From: preining at logic.at (Norbert Preining) Date: Sun, 12 Jun 2016 00:25:50 +0900 Subject: [tex-live] http://ctan.mirror.garr.it/mirrors/CTAN/systems/texlive/tlnet/update-tlmgr-latest.sh In-Reply-To: <1465658128.3811.2.camel@mariochiari.net> References: <1465658128.3811.2.camel@mariochiari.net> Message-ID: <20160611152550.GK10246@logic.at> On Sat, 11 Jun 2016, mario chiari wrote: > is?http://ctan.mirror.garr.it/mirrors/CTAN/systems/texlive/tlnet/update-tlmgr-la > test.sh?broken (for texlive2015). No. It is the correct for TeX Live 2016. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From beebe at math.utah.edu Sat Jun 11 17:28:46 2016 From: beebe at math.utah.edu (Nelson H. F. Beebe) Date: Sat, 11 Jun 2016 09:28:46 -0600 Subject: [tex-live] http://ctan.mirror.garr.it/mirrors/CTAN/systems/texlive/tlnet/update-tlmgr-latest.sh In-Reply-To: <1465658128.3811.2.camel@mariochiari.net> Message-ID: Mario Chiari asks on Sat, 11 Jun 2016 17:15:28 +0200: >> is >> http://ctan.mirror.garr.it/mirrors/CTAN/systems/texlive/tlnet/update-tlmgr-latest.sh >> broken (for texlive2015). The mirror status reports site at http://dante.ctan.org/mirmon/ shows that garr.it is up-to-date. I fetched the above *.sh file, and compared it with our mirror in Utah at http://ctan.math.utah.edu/tex-archive/systems/texlive/tlnet/update-tlmgr-latest.sh They are identical. The timestamp on the file is May 22 16:34 (MDT = UTC/GMT - 6). ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - University of Utah FAX: +1 801 581 4148 - - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - ------------------------------------------------------------------------------- From ml at mariochiari.net Sat Jun 11 18:15:52 2016 From: ml at mariochiari.net (mario chiari) Date: Sat, 11 Jun 2016 18:15:52 +0200 Subject: [tex-live] 2016 or 2015? Message-ID: <1465661752.3811.6.camel@mariochiari.net> Hi I am trying top install texlive 2016 (and I wish to keep 2015).? But if I launch install-tl-20160610/install-tl, it seems it tries to install into the 2015 folder. Why so? do I need to fix something? thanks m.? [root at fedora-pc install-tl-20160610]# install-tl --gui=text Loading /usr/local/texlive/2015/tlpkg/texlive.tlpdb Installing TeX Live 2015 from: /usr/local/texlive/2015 ? ?<<<<<<< Platform: x86_64-linux => 'GNU/Linux on x86_64' Distribution: live (uncompressed) Directory for temporary files: /tmp From beebe at math.utah.edu Sat Jun 11 18:51:10 2016 From: beebe at math.utah.edu (Nelson H. F. Beebe) Date: Sat, 11 Jun 2016 10:51:10 -0600 Subject: [tex-live] 2016 or 2015? In-Reply-To: <1465661752.3811.6.camel@mariochiari.net> Message-ID: Mario Chiari asks today about installing TeX Live 2016 while preserving the 2015 version. That is certainly possible; I have TeX Live versions back to 2003 on some of our servers here. >> # install-tl --gui=text >> Loading /usr/local/texlive/2015/tlpkg/texlive.tlpdb It looks like your install-tl script is coming from the 2015 release. You can empty the environment with "env -i" and then set PATH appropriately: # env -i PATH=/path/to/texlive/2016:/bin:/usr/bin install-tl --gui=text ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - University of Utah FAX: +1 801 581 4148 - - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - ------------------------------------------------------------------------------- From P.Taylor at Rhul.Ac.Uk Sat Jun 11 19:14:24 2016 From: P.Taylor at Rhul.Ac.Uk (Philip Taylor) Date: Sat, 11 Jun 2016 18:14:24 +0100 Subject: [tex-live] TexLive manager does not load any repository In-Reply-To: <20160611144851.GI10246@logic.at> References: <20160611092501.GF25579@logic.at> <20160611144851.GI10246@logic.at> Message-ID: <4e2d0f01-fc71-a23f-4bea-a6fb2a46184c@Rhul.Ac.Uk> Norbert Preining wrote: > in C: \ Program Files \ MiKTeX 2.9 \ miktex \ bin \; > No. There is a reason this warning is there, and it is to warn people > that having multiple (epsecially old, not here though!) tex programs > hanging around will give a bad outcome. "Will" ? Surely "may". -- Philip Taylor -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: Signature.jpg Type: image/jpeg Size: 12104 bytes Desc: not available URL: From richardmoncriefkoch at icloud.com Sat Jun 11 21:08:14 2016 From: richardmoncriefkoch at icloud.com (Richard Koch) Date: Sat, 11 Jun 2016 12:08:14 -0700 Subject: [tex-live] Xdvi not working Message-ID: <7CA44916-B8C2-461E-9C74-991C2BC6E4C1@icloud.com> Folks, Recall the problem. On 64 bit Macintoshes, xdvi in TeX Live 2016 works with XQuartz 2.7.8, but immediately crashes with XQuartz 2.7.9, which was released on May 5, 2016. The 64 bit version of xdvi in MacTeX was compiled on a machine running Snow Leopard, with XQuartz 2.3.6. ????? Tests: The file xdvi is just a script which ultimately calls xdvi-xaw. I recompiled TeX Live on El Capitan on a machine with XQuartz 2.7.9. Then I moved the file xdvi-xaw from this compile into the binary directory of TeX Live. After that xdvi worked. (But the binary didn?t work on older machines, as expected.) Next I installed XQuartz 2.7.9 on my Snow Leopard machine and recompiled TeX Live. Then I moved the file xdvi-xaw from this compile into the binary directory of TeX Live on that Snow Leopard machine running XQuartz 2.7.9. After that, xdvi worked. I have another machine running Snow Leopard. The system was installed from scratch during the TeX Live testing period, so it has X11 version 2.3.6. I moved the xdvi-xaw file created above to the binary directory of TeX Live on the other Snow Leopard machine. After that, xdvi continued to work. Next I repeated this expeiment on an El Capitan machine still running XQuartz 2.7.8. After that, xdvi continued to work. So far, so good. If we upgrade XQuartz on the Snow Leopard machine used to compile 64-bit binaries, then xdvi works on all systems supporting 68 bits, and WITHOUT UPDATING XQUARTZ. This is exactly what we need. So there is one more test. I moved the xdvi-xaw file created above to an El Capitan machine running XQuartz 2.7.9. After that, xdvi failed in the standard way. ???? Conclusion: XQuartz and xdvi break the underlying assumption behind TeX Live precompiled binaries: ?we can compile binaries on an older machine and get code which runs on any version of OS X" I?ll report these experiments to the XQuartz people. If you want to use xdvi, don?t update to XQuartz 2.7.9. Dick Koch From faro49 at gmail.com Sat Jun 11 22:05:02 2016 From: faro49 at gmail.com (Roberto Fanciulli) Date: Sat, 11 Jun 2016 22:05:02 +0200 Subject: [tex-live] TexLive manager does not load any repository In-Reply-To: <4e2d0f01-fc71-a23f-4bea-a6fb2a46184c@Rhul.Ac.Uk> References: <20160611092501.GF25579@logic.at> <20160611144851.GI10246@logic.at> <4e2d0f01-fc71-a23f-4bea-a6fb2a46184c@Rhul.Ac.Uk> Message-ID: Thanks for this new help too. > BTW, how do you switch between MikTeX and TeX Live? I use WindEdt 10.1 as tex editor and by the "Options->Configurations" in the main menus it is possible to switch between MikTeX and TeXLive Roberto 2016-06-11 19:14 GMT+02:00 Philip Taylor : > > > Norbert Preining wrote: > > in C: \ Program Files \ MiKTeX 2.9 \ miktex \ bin \; > > No. There is a reason this warning is there, and it is to warn people > that having multiple (epsecially old, not here though!) tex programs > hanging around will give a bad outcome. > > > "Will" ? Surely "may". > -- > > Philip Taylor > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: Signature.jpg Type: image/jpeg Size: 12104 bytes Desc: not available URL: From karl at freefriends.org Sun Jun 12 00:24:51 2016 From: karl at freefriends.org (Karl Berry) Date: Sat, 11 Jun 2016 22:24:51 +0000 Subject: [tex-live] Xdvi not working In-Reply-To: <7CA44916-B8C2-461E-9C74-991C2BC6E4C1@icloud.com> Message-ID: <201606112224.u5BMOpbY007784@freefriends.org> Conclusion: XQuartz and xdvi break the underlying assumption I really doubt it is xdvi, per se, that is breaking anything. xdvi hasn't notably changed. And there is no problem on any other platform. If you want to use xdvi, don????t update to XQuartz 2.7.9. How about if you send me an xdvi-for-xquartz-2.7.9 binary (just the one xdvi-xaw executable file) and I'll make it available from the TL bugs page? Or you could make it available from the mactex page. Then at least when people have the problem -- and we can be sure it will be reported ad infinitum -- we have an easy answer. We could ship both binaries and do tweaking in the shell script to invoke the right one, but I don't want to! --karl From beebe at math.utah.edu Sun Jun 12 01:35:48 2016 From: beebe at math.utah.edu (Nelson H. F. Beebe) Date: Sat, 11 Jun 2016 17:35:48 -0600 Subject: [tex-live] TeX Live 2016 binaries status, and TeX Live 2016 reaches some Linux distributions Message-ID: I've made several reports on this list about building TeX Live 2016 binary distributions for numerous platforms. The current state of affairs is documented at http://www.math.utah.edu/pub/texlive/ ftp://ftp.math.utah.edu/pub/texlive/ in the top-level README.html file. That page reports today a total of 88 systems, among them CentOS 7 on IBM POWER8 little-endian, and GNU Hurd on x86. As I was running system updates on various machines today, I noticed that TeX Live 2016 has finally reached some Linux distributions: Kali Linux Debian 8-testing Debian 8-unstable All are on x86-64, but presumably there are, or will soon be, distributions for other supported CPUs. On the Debian systems, I had to do "apt-get dist-upgrade" to get the latest TeX Live. I have several flavors of Ubuntu, which uses Debian-style packaging, but the latest version, 16.04 LTS (Long-Term Support -- to 2021), is still supplying TeX Live 2015. CentOS 7 (and thus, Red Hat 7 and Scientific Linux 7), the latest version in that popular distribution family, still has only TeX Live 2013. For another popular platform with commercial support, openSUSE 13, and 42.1 supply only TeX Live 2013. However, the latest release, openSUSE TumbleWeed (rolling updates) has TeX Live 2015. I conclude that, although a few of us developers are at the latest TeX Live release, most of the (La)TeX users on the large installed based of Linux systems are years behind. ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - University of Utah FAX: +1 801 581 4148 - - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - ------------------------------------------------------------------------------- From richardmoncriefkoch at icloud.com Sun Jun 12 03:30:33 2016 From: richardmoncriefkoch at icloud.com (Richard Koch) Date: Sat, 11 Jun 2016 18:30:33 -0700 Subject: [tex-live] Xdvi not working Message-ID: <3745E153-97E4-4F8C-A4F6-DD1769BA2E82@icloud.com> Karl wrote ?How about if you send me an xdvi-for-xquartz-2.7.9 binary and I?ll make it available from the TL bugs page? Or you could make it available from the mactex page.? If only life were that simple. Our 64 bit binaries run on seven different versions of OS X: Snow Leopard Lion Mountain Lion Mavericks Yosemite El Capitan The system Apple will unveil on Monday Every one of these could have several different versions of XQuartz. The current binaries work UNLESS XQuartz 2.7.9 is installed. It can be installed on each of the seven systems. I happen to have two computers which can run Snow Leopard and two computers which can run El Capitan. So I installed XQuartz 2.7.9 on one Snow Leopard computer and one El Capitan computer, and I recompiled xdvi on both. The Snow Leopard version doesn?t fix the problem on the El Capitan machine. The El Capitan version doesn?t even run on Snow Leopard. Thus potentially we might need seven different copies of xdvi-xaw for the seven systems when they are upgraded to XQuartz 2.7.9. I have computers running all seven systems, but I don?t want to install XQuartz 2.7.9 on the five remaining systems, for fear of making compiling and testing much harder in the future. Thus we?d need five people willing to sacrifice five systems for testing, and potentially compile five copies of the TeX Live binaries. Better to wait and see if my bug report to XQuartz yields any fruit. Dick From vincent.torri at gmail.com Sun Jun 12 07:09:34 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Sun, 12 Jun 2016 07:09:34 +0200 Subject: [tex-live] TeX Live 2016 binaries status, and TeX Live 2016 reaches some Linux distributions In-Reply-To: References: Message-ID: Hello On Sun, Jun 12, 2016 at 1:35 AM, Nelson H. F. Beebe wrote: > > I've made several reports on this list about building TeX Live 2016 > binary distributions for numerous platforms. The current state of > affairs is documented at > > http://www.math.utah.edu/pub/texlive/ > ftp://ftp.math.utah.edu/pub/texlive/ openindiana is mentioned only for i386 CPU and not x86_64. Is it normal ? Vincent Torri > in the top-level README.html file. That page reports today a total of > 88 systems, among them CentOS 7 on IBM POWER8 little-endian, and GNU > Hurd on x86. > > As I was running system updates on various machines today, I noticed > that TeX Live 2016 has finally reached some Linux distributions: > > Kali Linux > Debian 8-testing > Debian 8-unstable > > All are on x86-64, but presumably there are, or will soon be, > distributions for other supported CPUs. On the Debian systems, I had > to do "apt-get dist-upgrade" to get the latest TeX Live. > > I have several flavors of Ubuntu, which uses Debian-style packaging, > but the latest version, 16.04 LTS (Long-Term Support -- to 2021), is > still supplying TeX Live 2015. > > CentOS 7 (and thus, Red Hat 7 and Scientific Linux 7), the latest > version in that popular distribution family, still has only TeX Live > 2013. > > For another popular platform with commercial support, openSUSE 13, and > 42.1 supply only TeX Live 2013. However, the latest release, openSUSE > TumbleWeed (rolling updates) has TeX Live 2015. > > I conclude that, although a few of us developers are at the latest TeX > Live release, most of the (La)TeX users on the large installed based > of Linux systems are years behind. > > ------------------------------------------------------------------------------- > - Nelson H. F. Beebe Tel: +1 801 581 5254 - > - University of Utah FAX: +1 801 581 4148 - > - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - > - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - > - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - > ------------------------------------------------------------------------------- From vincent.torri at gmail.com Sun Jun 12 08:27:47 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Sun, 12 Jun 2016 08:27:47 +0200 Subject: [tex-live] Windows (MSYS2) : kpathsea does not find texmf.cnf In-Reply-To: References: <90EFD62E9E6148D59255BE42586502CD@CJ3001517A> <3601A79C6A3143F7BE85A8E6948167D2@CJ3001517A> Message-ID: Hello On Fri, Jun 10, 2016 at 7:05 PM, Vincent Torri wrote: > Is it a problem to use "kpsewhich.exe" when I call kpathsea_set_program_name() ? actually, from the source code, argv0 is only used to set some environment variables like SELFAUTOLOC from its path, so kpsewhich is sufficient Vincent Torri > thank you > > Vincent Torri > > > On Fri, Jun 10, 2016 at 4:26 PM, Akira Kakuto wrote: >>> Any idea why ? >> >> >> Because TEXMF in >> TFMFONTS = .;{$TEXMF/fonts,$VARTEXFONTS}/tfm// >> is defined by using $SELFAUTOLOC, the directory where >> the exec file exists. >> As you find in this exapmle, you have to install your >> program into bin/win32 in order to fully use the kpathsea. >> >> Best, >> Akira >> From zdenek.wagner at gmail.com Sun Jun 12 13:48:17 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Sun, 12 Jun 2016 13:48:17 +0200 Subject: [tex-live] TexLive manager does not load any repository In-Reply-To: <4e2d0f01-fc71-a23f-4bea-a6fb2a46184c@Rhul.Ac.Uk> References: <20160611092501.GF25579@logic.at> <20160611144851.GI10246@logic.at> <4e2d0f01-fc71-a23f-4bea-a6fb2a46184c@Rhul.Ac.Uk> Message-ID: 2016-06-11 19:14 GMT+02:00 Philip Taylor : > > > Norbert Preining wrote: > > in C: \ Program Files \ MiKTeX 2.9 \ miktex \ bin \; > > No. There is a reason this warning is there, and it is to warn people > that having multiple (epsecially old, not here though!) tex programs > hanging around will give a bad outcome. > > > "Will" ? Surely "may". > Yes, it is a common problem in Linux where TeX is installed with the OS, a user installs TeX Live, both are in PATH and quite often (because TeX Live is usually more complete but further in PATH), some binaries are take from one TeX distro and some binaries from the other. Each binary reads its own settings and use its own packages that have different versions. The result is a strange error message or even a crash. Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz > -- > > Philip Taylor > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: Signature.jpg Type: image/jpeg Size: 12104 bytes Desc: not available URL: From enrico.gregorio at gmail.com Sun Jun 12 15:50:00 2016 From: enrico.gregorio at gmail.com (Enrico Gregorio) Date: Sun, 12 Jun 2016 15:50:00 +0200 Subject: [tex-live] Xdvi not working In-Reply-To: <3745E153-97E4-4F8C-A4F6-DD1769BA2E82@icloud.com> References: <3745E153-97E4-4F8C-A4F6-DD1769BA2E82@icloud.com> Message-ID: Too bad indeed! Thanks for the effort. I don't think there are many people around who use Xdvi on Mac OS X; it just happened that I wanted to check an answer also with it and, to my dismay, I discovered the problem. Let's hope the bug report about XQuartz succeeds. Otherwise, short instructions on how to compile Xdvi (possibly without building all the binaries) or a binary to download might be an acceptable workaround. Ciao Enrico 2016-06-12 3:30 GMT+02:00 Richard Koch : > Karl wrote ?How about if you send me an xdvi-for-xquartz-2.7.9 binary > and I?ll make it available from the TL bugs page? Or you could make it > available from the mactex page.? > > If only life were that simple. Our 64 bit binaries run on seven different > versions of OS X: > > Snow Leopard > Lion > Mountain Lion > Mavericks > Yosemite > El Capitan > The system Apple will unveil on Monday > > Every one of these could have several different versions of > XQuartz. The current binaries work UNLESS XQuartz 2.7.9 > is installed. It can be installed on each of the seven systems. > > I happen to have two computers which can run Snow Leopard > and two computers which can run El Capitan. So I installed XQuartz 2.7.9 > on one Snow Leopard computer and one El Capitan computer, > and I recompiled xdvi on both. The Snow Leopard version > doesn?t fix the problem on the El Capitan machine. > The El Capitan version doesn?t even run on Snow Leopard. > > Thus potentially we might need seven different copies of xdvi-xaw > for the seven systems when they are upgraded to XQuartz 2.7.9. > > I have computers running all seven systems, but I don?t want to install > XQuartz 2.7.9 on the five remaining systems, for fear of making compiling > and testing much harder in the future. Thus we?d need five > people willing to sacrifice five systems for testing, and potentially > compile five copies of the TeX Live binaries. > > Better to wait and see if my bug report to XQuartz yields any fruit. > > Dick > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jfbu at free.fr Sun Jun 12 16:01:12 2016 From: jfbu at free.fr (jfbu) Date: Sun, 12 Jun 2016 16:01:12 +0200 Subject: [tex-live] Xdvi not working In-Reply-To: References: <3745E153-97E4-4F8C-A4F6-DD1769BA2E82@icloud.com> Message-ID: Le 12 juin 2016 ? 15:50, Enrico Gregorio a ?crit : > Too bad indeed! Thanks for the effort. > > I don't think there are many people around who use Xdvi > on Mac OS X; indeed, but for info I do use xdvi sometimes, hence many thanks for this whole thread which has led me to refuse the 2.7.9 install... (on mavericks) Best Jean-Fran?ois > it just happened that I wanted to check > an answer also with it and, to my dismay, I discovered > the problem. > > Let's hope the bug report about XQuartz succeeds. > Otherwise, short instructions on how to compile Xdvi > (possibly without building all the binaries) or a binary > to download might be an acceptable workaround. > > Ciao > Enrico > > > 2016-06-12 3:30 GMT+02:00 Richard Koch : > Karl wrote ?How about if you send me an xdvi-for-xquartz-2.7.9 binary > and I?ll make it available from the TL bugs page? Or you could make it > available from the mactex page.? > > If only life were that simple. Our 64 bit binaries run on seven different > versions of OS X: > > Snow Leopard > Lion > Mountain Lion > Mavericks > Yosemite > El Capitan > The system Apple will unveil on Monday > > Every one of these could have several different versions of > XQuartz. The current binaries work UNLESS XQuartz 2.7.9 > is installed. It can be installed on each of the seven systems. > > I happen to have two computers which can run Snow Leopard > and two computers which can run El Capitan. So I installed XQuartz 2.7.9 > on one Snow Leopard computer and one El Capitan computer, > and I recompiled xdvi on both. The Snow Leopard version > doesn?t fix the problem on the El Capitan machine. > The El Capitan version doesn?t even run on Snow Leopard. > > Thus potentially we might need seven different copies of xdvi-xaw > for the seven systems when they are upgraded to XQuartz 2.7.9. > > I have computers running all seven systems, but I don?t want to install > XQuartz 2.7.9 on the five remaining systems, for fear of making compiling > and testing much harder in the future. Thus we?d need five > people willing to sacrifice five systems for testing, and potentially > compile five copies of the TeX Live binaries. > > Better to wait and see if my bug report to XQuartz yields any fruit. > > Dick > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From vivrii at gmail.com Sun Jun 12 16:13:35 2016 From: vivrii at gmail.com (Victor Ivrii) Date: Sun, 12 Jun 2016 10:13:35 -0400 Subject: [tex-live] Xdvi not working In-Reply-To: References: <3745E153-97E4-4F8C-A4F6-DD1769BA2E82@icloud.com> Message-ID: xdvi works for me on XQuartz 2.7.9 and MacOSX 10.7.5 -------------- next part -------------- An HTML attachment was scrubbed... URL: From richardmoncriefkoch at icloud.com Sun Jun 12 17:24:05 2016 From: richardmoncriefkoch at icloud.com (Richard Koch) Date: Sun, 12 Jun 2016 08:24:05 -0700 Subject: [tex-live] Xdvi not working Message-ID: <346F6FBA-8D1F-4F13-BC9C-2EC3C75F3BC5@icloud.com> Folks, My bug report to XQuartz led to important information from them. Our problem may be solvable. Please have patience. I should know later today. Dick Koch From ml at mariochiari.net Sun Jun 12 18:18:02 2016 From: ml at mariochiari.net (mario chiari) Date: Sun, 12 Jun 2016 18:18:02 +0200 Subject: [tex-live] calling mktexnam, bug or misconf? Message-ID: <1465748282.3535.4.camel@mariochiari.net> Hi trying texlive2016, i get the following error msg: ----- LaTeX Font Warning: Font shape `T2A/wncyr/m/n' undefined (Font)??????????????using `T2A/cmr/m/n' instead on input line 15. kpathsea: Running mktextfm larm1200 /usr/share/texlive/texmf-dist/web2c/mktexnam: Could not map source abbreviation??for larm1200. /usr/share/texlive/texmf-dist/web2c/mktexnam: Need to update ? ---- clearly, it should call /usr/share/texlive/***2016***/texmf-dist/web2c/mktexnam instead. (i fixed my compilation by running > mktextfm larm1200; ?directly) Is it a bug or a misconfiguration on my side? Comments welcome thanks cheers mario From ml at mariochiari.net Sun Jun 12 18:23:35 2016 From: ml at mariochiari.net (mario chiari) Date: Sun, 12 Jun 2016 18:23:35 +0200 Subject: [tex-live] 2016 or 2015? In-Reply-To: <1465661752.3811.6.camel@mariochiari.net> References: <1465661752.3811.6.camel@mariochiari.net> Message-ID: <1465748615.3535.9.camel@mariochiari.net> Hi trying ? #??perl install-tl -gui it works very fine. It still bother me that #install-tl does not work for me thanks? cheers mario On Sat, 2016-06-11 at 18:15 +0200, mario chiari wrote: > Hi > > I am trying top install texlive 2016 (and I wish to keep 2015).? > But if I launch install-tl-20160610/install-tl, it seems it tries to install > into the 2015 folder. > Why so? do I need to fix something? > thanks > m.? > > [root at fedora-pc install-tl-20160610]# install-tl --gui=text > Loading /usr/local/texlive/2015/tlpkg/texlive.tlpdb > Installing TeX Live 2015 from: /usr/local/texlive/2015 ? ?<<<<<<< > Platform: x86_64-linux => 'GNU/Linux on x86_64' > Distribution: live (uncompressed) > Directory for temporary files: /tmp From komascript at gmx.info Sun Jun 12 18:36:52 2016 From: komascript at gmx.info (Markus Kohm) Date: Sun, 12 Jun 2016 18:36:52 +0200 Subject: [tex-live] 2016 or 2015? In-Reply-To: <1465748615.3535.9.camel@mariochiari.net> References: <1465661752.3811.6.camel@mariochiari.net> <1465748615.3535.9.camel@mariochiari.net> Message-ID: <1651126.kOWM2cz86e@x61s.fritz.box> Am Sonntag, 12. Juni 2016, 18:23:35 schrieb mario chiari: > [root at fedora-pc install-tl-20160610]# install-tl --gui=text Are you sure, that this should call the correct install-tl? What does which install-tl say? Current directory should not be part of PATH, i.e., for root. Generally you have to add a (relative) path to the call of install-tl in the current directory. Try: ./install-tl --gui=text instead of install-tl --gui=text -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 181 bytes Desc: This is a digitally signed message part. URL: From martin at oneiros.de Sun Jun 12 18:38:40 2016 From: martin at oneiros.de (=?UTF-8?Q?Martin_Schr=C3=B6der?=) Date: Sun, 12 Jun 2016 18:38:40 +0200 Subject: [tex-live] TeX Live 2016 binaries status, and TeX Live 2016 reaches some Linux distributions In-Reply-To: References: Message-ID: 2016-06-12 1:35 GMT+02:00 Nelson H. F. Beebe : > I have several flavors of Ubuntu, which uses Debian-style packaging, > but the latest version, 16.04 LTS (Long-Term Support -- to 2021), is > still supplying TeX Live 2015. Nelson, sorry - but I have to ask: Do you know what "Long-Term Support" means? I doubt it - otherwise you wouldn't imply that 16.04 should update to TL2016. > CentOS 7 (and thus, Red Hat 7 and Scientific Linux 7), the latest > version in that popular distribution family, still has only TeX Live > 2013. Small wonder: RHEL7 was released on 2014-06-10. OTOH Fedora currently has TL2015. Best Martin From richardmoncriefkoch at icloud.com Sun Jun 12 18:40:55 2016 From: richardmoncriefkoch at icloud.com (Richard Koch) Date: Sun, 12 Jun 2016 09:40:55 -0700 Subject: [tex-live] Xdvi not working Message-ID: <2A8E6FF5-5CA6-40F3-923C-D7391D0D804A@icloud.com> Folks, Using information from the XQuartz bug system, I?ve compiled xdvi on Snow Leopard for XQuartz 2.7.9. This version will be supplied by tlmgr and TeX Live Utility in a few days as an update. The new version runs on Snow Leopard and El Capitan, and I assume on all intermediate systems, but users MUST update to XQuartz 2.7.9. The update is available for Snow Leopard and above, and thus on all systems supported by x86_64-darwin. The 64 bit binaries for the Mac in TeX Live are compiled on Snow Leopard, for compatibility with older systems. In most cases, TeX Live contains its own copies of libraries, rather than relying on system libraries. But X11 is an exception. Until now, I have used XQuartz 2.3.6 on Snow Leopard for maximum compatibility. This failed with the introduction of XQuartz 2.7.9, so from now on I will update to the latest version of XQuartz before making TeX Live. Users who need X11 should get in the habit of updating it regularly. Dick Koch From zdenek.wagner at gmail.com Sun Jun 12 18:48:07 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Sun, 12 Jun 2016 18:48:07 +0200 Subject: [tex-live] calling mktexnam, bug or misconf? In-Reply-To: <1465748282.3535.4.camel@mariochiari.net> References: <1465748282.3535.4.camel@mariochiari.net> Message-ID: The TeX from your Linux distro is before TeX from TL in your PATH. Change the order and it will work. It is one of the most frequents errors in PATH settings. Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz 2016-06-12 18:18 GMT+02:00 mario chiari : > Hi > > trying texlive2016, i get the following error msg: > > ----- > LaTeX Font Warning: Font shape `T2A/wncyr/m/n' undefined > (Font) using `T2A/cmr/m/n' instead on input line 15. > > kpathsea: Running mktextfm larm1200 > /usr/share/texlive/texmf-dist/web2c/mktexnam: Could not map source > abbreviation for larm1200. > /usr/share/texlive/texmf-dist/web2c/mktexnam: Need to update ? > ---- > > clearly, it should call > /usr/share/texlive/***2016***/texmf-dist/web2c/mktexnam > instead. (i fixed my compilation by running > mktextfm larm1200; directly) > > Is it a bug or a misconfiguration on my side? > Comments welcome > > thanks > cheers > mario > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From zdenek.wagner at gmail.com Sun Jun 12 18:52:20 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Sun, 12 Jun 2016 18:52:20 +0200 Subject: [tex-live] TeX Live 2016 binaries status, and TeX Live 2016 reaches some Linux distributions In-Reply-To: References: Message-ID: 2016-06-12 18:38 GMT+02:00 Martin Schr?der : > 2016-06-12 1:35 GMT+02:00 Nelson H. F. Beebe : > > I have several flavors of Ubuntu, which uses Debian-style packaging, > > but the latest version, 16.04 LTS (Long-Term Support -- to 2021), is > > still supplying TeX Live 2015. > > Nelson, sorry - but I have to ask: Do you know what "Long-Term Support" > means? > I doubt it - otherwise you wouldn't imply that 16.04 should update to > TL2016. > > > CentOS 7 (and thus, Red Hat 7 and Scientific Linux 7), the latest > > version in that popular distribution family, still has only TeX Live > > 2013. > > Small wonder: RHEL7 was released on 2014-06-10. > It is explicitely written in the CentOS documentation. If the major version of CentOS is not changed, software is never updated to a new major version even if it contains known bugs (as was the gace of ghostsvript in CentOS 4.x). The newer TeX Live will appear in CentOS 8.0. > > OTOH Fedora currently has TL2015. > > Best > Martin > Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz -------------- next part -------------- An HTML attachment was scrubbed... URL: From ml at mariochiari.net Sun Jun 12 19:04:48 2016 From: ml at mariochiari.net (mario chiari) Date: Sun, 12 Jun 2016 19:04:48 +0200 Subject: [tex-live] 2016 or 2015? In-Reply-To: <1651126.kOWM2cz86e@x61s.fritz.box> References: <1465661752.3811.6.camel@mariochiari.net> <1465748615.3535.9.camel@mariochiari.net> <1651126.kOWM2cz86e@x61s.fritz.box> Message-ID: <1465751088.3535.10.camel@mariochiari.net> On Sun, 2016-06-12 at 18:36 +0200, Markus Kohm wrote: > Am Sonntag, 12. Juni 2016, 18:23:35 schrieb mario chiari: > > > > [root at fedora-pc install-tl-20160610]# install-tl --gui=text > Are you sure, that this should call the correct install-tl? What does > > which install-tl > > say? > > Current directory should not be part of PATH, i.e., for root. Generally you? > have to add a (relative) path to the call of install-tl in the current? > directory. > > Try:? > > ./install-tl --gui=text yes, it works, thanks (by the way, why?) cheers mario > > instead of > > install-tl --gui=text From karl at freefriends.org Sun Jun 12 20:19:59 2016 From: karl at freefriends.org (Karl Berry) Date: Sun, 12 Jun 2016 18:19:59 +0000 Subject: [tex-live] pdftex segfault with \pdflastmatch In-Reply-To: <248EACE4D9714086BE38F6AEA223AE07@CJ3001517A> Message-ID: <201606121819.u5CIJxpj013815@freefriends.org> dc> This plain tex document demonstrates \pdflastmatch failing if dc> used after a \pdfmatch that returned no matches. Thanks for the report, of course, though it's sad. ak> The crash disappears if I apply an attached utils.c.diff. ak> But I don't know that the patch is right or not. I was still seeing garbage for the position reported by \pdflastmatch unless I added an explicit check (and variable) for whether the preceding match succeeded. (Because otherwise the various arrays are accessed without necessarily being set.) Patch (minus the documentation, etc.) below; I installed it in pdftex and TL. Akira (or anyone), let me know you see problems with it. Thanks, Karl My test file (with the example from the pdftex manual): %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % This file public domain. Bug report of \pdflastmatch crash: % http://tug.org/pipermail/tex-live/2016-June/038664.html % \catcode`\{=1 \catcode`\}=2 \catcode`^=7 \newlinechar=`^^J % \message{^^Jpdflastmatch0 before any match: \pdflastmatch0} \message{^^J}% % \message{^^Jmatch a to b: \pdfmatch{a}{b}} \message{^^J}% \message{pdflastmatch0: \pdflastmatch0} \message{^^J}% % \message{^^Jmatch example: \pdfmatch subcount 3 {ab(cd)*ef(gh)(ij)}{abefghij}} \message{^^J}% \message{pdflastmatch0 (0->abefghij): \pdflastmatch0 ^^J} \message{pdflastmatch1 (-1->): \pdflastmatch1 ^^J} \message{pdflastmatch2 (4->gh): \pdflastmatch2 ^^J} \message{pdflastmatch3 (-1->): \pdflastmatch3} \message{^^J}% \end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% And the patch: --- pdftexdir/utils.c (revision 757) +++ pdftexdir/utils.c (working copy) @@ -817,12 +817,16 @@ pdf_printf("/ModDate (%s)\n", start_time_str); } + #define DEFAULT_SUB_MATCH_COUNT 10 static int sub_match_count = DEFAULT_SUB_MATCH_COUNT; static regmatch_t *pmatch = NULL; static char *match_string = NULL; +static int last_match_succeeded = 0; -void matchstrings(strnumber s, strnumber t, int subcount, boolean icase) +/* Implements \pdfmatch */ +void +matchstrings(strnumber s, strnumber t, int subcount, boolean icase) { regex_t preg; int cflags = REG_EXTENDED; @@ -857,20 +861,31 @@ pmatch = xtalloc(sub_match_count, regmatch_t); } ret = regexec(&preg, str, sub_match_count, pmatch, eflags); + xfree(match_string); - match_string = xstrdup(str); - strpool[poolptr++] = ((ret == 0) ? '1' : '0'); + match_string = xstrdup(str); /* save searched-in string, used below */ + last_match_succeeded = ret == 0; /* save whether match succeeded */ + strpool[poolptr++] = ((ret == 0) ? '1' : '0'); /* in string pool too */ } regfree(&preg); } -void getmatch(int i) +/* Implements \pdflastmatch */ + +void +getmatch(int i) { - int size, len = 0; /* to avoid warning about uninitialized use of len */ + int size; + int len = 0; /* avoid spurious uninitialized warning */ - boolean found = i < sub_match_count - && match_string != NULL && pmatch[i].rm_so >= 0 && i >= 0; + boolean found + = i >= 0 /* should always be so due to pdftex.web */ + && i < sub_match_count /* if >subcount, not found by definition */ + && match_string != NULL /* first call, and just in case */ + && last_match_succeeded /* if no match, not found */ + && pmatch[i].rm_so >= 0 /* if no starting position, not found */ + && pmatch[i].rm_eo >= pmatch[i].rm_so; /* just in case */ if (found) { len = pmatch[i].rm_eo - pmatch[i].rm_so; From komascript at gmx.info Sun Jun 12 20:40:30 2016 From: komascript at gmx.info (Markus Kohm) Date: Sun, 12 Jun 2016 20:40:30 +0200 Subject: [tex-live] 2016 or 2015? In-Reply-To: <1465751088.3535.10.camel@mariochiari.net> References: <1465661752.3811.6.camel@mariochiari.net> <1651126.kOWM2cz86e@x61s.fritz.box> <1465751088.3535.10.camel@mariochiari.net> Message-ID: <6664896.Goedeo5Ftj@x61s.fritz.box> Am Sonntag, 12. Juni 2016, 19:04:48 schrieb mario chiari: > by the way, why? Because it does not start the intall_tl from PATH but install_tl located at the current working directory. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 181 bytes Desc: This is a digitally signed message part. URL: From simon.b.may at gmail.com Sun Jun 12 21:45:18 2016 From: simon.b.may at gmail.com (Simon May) Date: Sun, 12 Jun 2016 21:45:18 +0200 Subject: [tex-live] Tlmgr closing takes fairly long In-Reply-To: References: <2f7dbcca-290b-9db0-9be7-4baca67e3802@gmx.net> <84eg8j3gl6.fsf@siart.de> <22348.45767.11629.528210@zaphod.ms25.net> Message-ID: <575DBBCE.90306@gmail.com> Apologies for digging this out again, but I would like to comment on this in the interest of educating the public. It is certainly not true that Java (OpenJDK) is not freely redistributable. The OpenJDK is licensed under the GPL and has been the reference implementation of Java since version 7 (see [1]). The ?Oracle JDK? and OpenJDK are virtually identical in any recent version. I do not understand what you mean by ?the GNU Java for Linux? and that it ?lacks GUI?. I am not aware of any Java implementations which omit the graphical user interface toolkits (AWT and Swing) as these are part of the Java standard library. There are no ?standard classes that are not distributed with JRE but JDK is required?. Java has a standard library, required to be implemented by all conforming implementations (like the C standard library for the C language) and which is distributed in what is called the JRE, or Java Runtime Environment. The JDK (Java Development Kit) is for development, as the name suggests. Any program relying specifically on the presence of a JDK is doing something seriously wrong. Yes, historically, there have been some differences and difficulties between open Java implementations and the Sun/Oracle proprietary distribution, but as explained above, with OpenJDK as the reference implementation, these have been gone for years. Without any detailed information, I cannot say why the application you dealt with ran on your Linux machine, but not on another Windows machine. From experience, I would rather suggest that the issue is that program installation (especially runtimes like Java, which are used by many independent applications) under Windows is a complete mess due to the lack of package management. Often, Windows users have many different versions of the JRE are installed at the same time, causing all sorts of trouble. [1]: https://en.wikipedia.org/wiki/OpenJDK Best wishes, Simon May On 30.05.2016 23:56, Zdenek Wagner wrote: > > I don't know anything about Java, but I've been told that current > implementations are quite fast. > > > Yes, it is fast but the GNU Java for Linux lacks GUI, Java for other > platforms is not freely redistributable, users will have to download and > install it themselves. It need not be trivial. A few month ago I found > that a Java app working fine on my Linux did not work on a Windows > computer and the error message was not useful. The reason was that the > app required standard classes that are not distributed with JRE but JDK > is required. It can easily be figured out by a Java expert but not by > normal users, they will just report that it does not work. > > > Zden?k Wagner > http://ttsm.icpf.cas.cz/team/wagner.shtml > http://icebearsoft.euweb.cz > -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: OpenPGP digital signature URL: From cereda.paulo at gmail.com Mon Jun 13 00:26:54 2016 From: cereda.paulo at gmail.com (Paulo Roberto Massa Cereda) Date: Sun, 12 Jun 2016 19:26:54 -0300 Subject: [tex-live] Tlmgr closing takes fairly long In-Reply-To: <575DBBCE.90306@gmail.com> References: <2f7dbcca-290b-9db0-9be7-4baca67e3802@gmx.net> <84eg8j3gl6.fsf@siart.de> <22348.45767.11629.528210@zaphod.ms25.net> <575DBBCE.90306@gmail.com> Message-ID: 'ello! :) [Simon May] > There are no ?standard classes that are not distributed with JRE but > JDK is required?. Java has a standard library, required to be > implemented by all conforming implementations (like the C standard > library for the C language) and which is distributed in what is > called the JRE, or Java Runtime Environment. The JDK (Java > Development Kit) is for development, as the name suggests. Any > program relying specifically on the presence of a JDK is doing > something seriously wrong. It is worth mentioning that some Linux distros deliberately separate Java components in order to avoid unnecessary bindings. In Fedora, there's a headless version: [paulo at cambridge ~] $ rpm -qa | grep headless java-1.8.0-openjdk-headless-1.8.0.91-6.b14.fc23.x86_64 This package has no graphical components. As the name indicates, it is used in environments where there is no need for a GUI (there are no X11 bindings). This is my Java setup: [paulo at cambridge ~] $ rpm -qa | grep openjdk java-1.8.0-openjdk-1.8.0.91-6.b14.fc23.x86_64 java-1.8.0-openjdk-headless-1.8.0.91-6.b14.fc23.x86_64 java-1.8.0-openjdk-devel-1.8.0.91-6.b14.fc23.x86_64 The first two ensure normal execution of Java programs, including the ones with GUI features. The last one regards development and it is not needed for the average user. About execution, as time goes by, each version gets enhanced (or bloated, pick the term you like) with new features. In general, you can run older code in newer JVMs, but the opposite is not always true. Also, bytecode compatibility does not always guarantee proper execution in older JVMs, but that's another chapter. :) From karl at freefriends.org Mon Jun 13 00:51:29 2016 From: karl at freefriends.org (Karl Berry) Date: Sun, 12 Jun 2016 22:51:29 +0000 Subject: [tex-live] Unknown directive ...containerchecksum In-Reply-To: <20160611150936.GJ10246@logic.at> Message-ID: <201606122251.u5CMpTTQ030799@freefriends.org> Subject: Re: [tex-live] Unknown directive ...containerchecksum np> ... get http://mirror.ctan.org/systems/texlive/tlnet/update-tlmgr-latest.sh (Can we post this somewhere that I don't have to retype it???) I added a note on http://tug.org/texlive/bugs.html and explicitly mentioned the "Unknown directive" error on http://tug.org/texlive/tlmgr.html (#disaster). -k From enrico.gregorio at gmail.com Mon Jun 13 09:52:33 2016 From: enrico.gregorio at gmail.com (Enrico Gregorio) Date: Mon, 13 Jun 2016 09:52:33 +0200 Subject: [tex-live] magyar.ldf hacked? Message-ID: Please look at http://tug.org/svn/texlive/trunk/Master/texmf-dist/tex/generic/babel-hungarian/magyar.ldf?sortby=date&r1=41421&r2=41420&pathrev=41421 Ciao Enrico -------------- next part -------------- An HTML attachment was scrubbed... URL: From preining at logic.at Mon Jun 13 10:32:09 2016 From: preining at logic.at (Norbert Preining) Date: Mon, 13 Jun 2016 17:32:09 +0900 Subject: [tex-live] magyar.ldf hacked? In-Reply-To: References: Message-ID: <20160613083209.GZ10246@logic.at> On Mon, 13 Jun 2016, Enrico Gregorio wrote: > http://tug.org/svn/texlive/trunk/Master/texmf-dist/tex/generic/babel-hungarian/magyar.ldf?sortby=date&r1=41421&r2=41420&pathrev=41421 Thanks. Bad. CTAN is also bad. I removed the file and will inform CTAN and the authors. THANKS!!! Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From preining at logic.at Mon Jun 13 10:43:56 2016 From: preining at logic.at (Norbert Preining) Date: Mon, 13 Jun 2016 17:43:56 +0900 Subject: [tex-live] magyar.ldf hacked? In-Reply-To: <20160613083209.GZ10246@logic.at> References: <20160613083209.GZ10246@logic.at> Message-ID: <20160613084356.GC10246@logic.at> > I removed the file and will inform CTAN and the authors. Reinstalled the v1.5c verion, informed CTAN and author. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From P.Taylor at Rhul.Ac.Uk Mon Jun 13 10:45:50 2016 From: P.Taylor at Rhul.Ac.Uk (Philip Taylor) Date: Mon, 13 Jun 2016 09:45:50 +0100 Subject: [tex-live] magyar.ldf hacked? In-Reply-To: <20160613083209.GZ10246@logic.at> References: <20160613083209.GZ10246@logic.at> Message-ID: <314a9388-e6fc-96b4-2d30-6ffed949b382@Rhul.Ac.Uk> "disgustedobjective.com" -> "holdhelptime.com"; WHOIS for "holdhelptime.com" : > Domain Name: holdhelptime.com > Registry Domain ID: D400707386 > Registrar WHOIS Server: Whois.domainerschoice.com > Updated date: 2016-04-15T15:47:37Z > Creation date: 2016-04-15T15:00:17Z > Registrar Registration Expiration date: > Registrar: Nanjing Imperiosus Technology Co. Ltd > Registrar IANA ID: 953 > Registrar Abuse Contact Email: abuse at domainerschoice.com > Registrar Abuse Contact Phone: +86.2584752360 > Registrar Abuse Website: http://www.domainerschoice.com/report_abuse > Domain Status: ok > Registry Registrant ID: > Registrant Name: Private Registration > Registrant Organization: WhoisGuardService.com > Registrant Street: Tian Hong Shan Zhuang, BLd. 7, Office 104 > Registrant City: Nanjing > Registrant State/Province : Jiangsu > Registrant Postal Code: 210049 > Registrant Country: CHINA > Registrant Phone: 86.2584752362 > Registrant Phone Ext: > Registrant Fax: 86.2584752360 > Registrant Fax Ext: > Registrant Email: holdhelptime.com at domainersidshield.com > Registry Admin ID: > Admin Name: Private Registration > Admin Organization: WhoisGuardService.com > Admin Street: Tian Hong Shan Zhuang, BLd. 7, Office 104 > Admin City: Nanjing > Admin State/Province : Jiangsu > Admin Postal Code : 210049 > Admin Country: CHINA > Admin Phone: 86.2584752362 > Admin Phone Ext: > Admin Fax: 86.2584752360 > Admin Fax Ext: > Admin Email: holdhelptime.com at domainersidshield.com > Registry Tech ID: > Tech Name: Private Registration > Tech Organization: WhoisGuardService.com > Tech Street: Tian Hong Shan Zhuang, BLd. 7, Office 104 > Tech City: Nanjing > Tech State/Province: Jiangsu > Tech Postal Code: 210049 > Tech Country: CHINA > Tech Phone: 86.2584752362 > Tech Phone Ext: > Tech Fax: 86.2584752360 > Tech Fax Ext: > Tech Email: holdhelptime.com at domainersidshield.com > Name Server: NS1.PQL.NET > Name Server: NS2.PQL.NET > DNSSEC: UnSigned > URL of the ICANN WHOIS Data Problem Reporting System: > http://wdprs.internic.net/ > Last update of WHOIS database: 2016-04-15T15:47:37Z Just across the bay from you, Norbert -- perhaps you could call in and have a word with them :-) ** Phil. From beebe at math.utah.edu Mon Jun 13 14:43:59 2016 From: beebe at math.utah.edu (Nelson H. F. Beebe) Date: Mon, 13 Jun 2016 06:43:59 -0600 Subject: [tex-live] TeX Live 2016 binaries status, and TeX Live 2016 reaches some Linux distributions In-Reply-To: Message-ID: Vincent Torri asks on Sun, 12 Jun 2016 07:09:34 +0200: >> openindiana is mentioned only for i386 CPU and not x86_64. Is it normal ? Yes, it is normal. The SPARC architecture that has been the core CPU of Solaris since the introduction of SPARC in 1987 as Sun Microsystems' version of the Berkeley RISC design, replacing Sun's previous use of the Motorola 68000 family processors, has been a mixed 32-bit and 64-bit system since SPARC V8 in 1990. It is possible to compile software on Solaris (and its OpenSolaris relative, OpenIndiana) in either mode, but the compiler default for both SPARC and Intel x86/x86-64 CPUs, has always been 32-bit mode. Indeed, a check of the executables in /bin (== /usr/bin) on both CPU platforms shows that all are 32-bit files. Thus, we have % cc hello.c && file a.out a.out: ELF 32-bit MSB executable, SPARC32PLUS, V8+ Required, version 1 (SYSV), dynamically linked (uses shared libs), not stripped % cc hello.c && file a.out a.out: ELF 32-bit LSB executable, Intel 80386, version 1, dynamically linked (uses shared libs), not stripped, uses FPU % cc -m64 hello.c && file a.out a.out: ELF 64-bit MSB executable, SPARC V9, total store ordering, version 1 (SYSV), dynamically linked (uses shared libs), not stripped % cc -m64 hello.c && file a.out a.out: ELF 64-bit LSB executable, x86-64, version 1, dynamically linked (uses shared libs), not stripped, uses FPU FXSR SSE For typesetting applications, consider that a typical book is about 1MB of text, I find it hard to imagine that an address space of more than 4096MB would be needed for the majority of users, especially because TeX processes only a page at a time. The memory hogs are instead the document viewers that try to hold everything in memory for rapid navigation. Nevertheless, if there is a feeling that a 64-bit build of TeX Live 2016 for OpenIndiana would be of interest, I can certainly try it. The Open Community Software Project (OpenCSW) https://en.wikipedia.org/wiki/OpenCSW that provides an archive of about 3850 open-source and free software packages built for Solaris and OpenIndiana currently has TeX Live, but only the 2012 version. Thus, as I reported on Saturday on this list, major software channels for easy installation of binary distributions are often several years behind. By offering TeX binaries for as many platforms as possible, we can bring the end-user community up to the latest available TeXware. ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - University of Utah FAX: +1 801 581 4148 - - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - ------------------------------------------------------------------------------- From mojca.miklavec.lists at gmail.com Mon Jun 13 14:53:23 2016 From: mojca.miklavec.lists at gmail.com (Mojca Miklavec) Date: Mon, 13 Jun 2016 14:53:23 +0200 Subject: [tex-live] TeX Live 2016 binaries status, and TeX Live 2016 reaches some Linux distributions In-Reply-To: References: Message-ID: On 13 June 2016 at 14:43, Nelson H. F. Beebe wrote: > > Nevertheless, if there is a feeling that a 64-bit build of TeX Live > 2016 for OpenIndiana would be of interest, I can certainly try it. Official binaries do include 64-bit binaries for Solaris. I assume they work just fine on OpenIndiana. > The Open Community Software Project (OpenCSW) > > https://en.wikipedia.org/wiki/OpenCSW > > that provides an archive of about 3850 open-source and free software > packages built for Solaris and OpenIndiana currently has TeX Live, but > only the 2012 version. Indeed. This should be upgraded at some point and I'm sure they would warmly welcome any contribution, but good luck upgrading it ... there are 5380 lines of code to package TL 2012 ... https://buildfarm.opencsw.org/source/xref/opencsw/csw/mgar/pkg/texlive/trunk/Makefile https://www.opencsw.org/packages/texlive/ Mojca From arthur.reutenauer at normalesup.org Mon Jun 13 14:58:28 2016 From: arthur.reutenauer at normalesup.org (Arthur Reutenauer) Date: Mon, 13 Jun 2016 13:58:28 +0100 Subject: [tex-live] TeX Live 2016 binaries status, and TeX Live 2016 reaches some Linux distributions In-Reply-To: References: Message-ID: <20160613125828.GA3906224@phare.normalesup.org> > The Open Community Software Project (OpenCSW) > > https://en.wikipedia.org/wiki/OpenCSW > > that provides an archive of about 3850 open-source and free software > packages built for Solaris and OpenIndiana currently has TeX Live, but > only the 2012 version. As I understand, an OpenCSW developer actually tried compiling more recent binaries but did run into serious technical problems. Mojca has more details on that. Best, Arthur From beebe at math.utah.edu Mon Jun 13 15:09:36 2016 From: beebe at math.utah.edu (Nelson H. F. Beebe) Date: Mon, 13 Jun 2016 07:09:36 -0600 Subject: [tex-live] TeX Live 2016 binaries status, and TeX Live 2016 reaches some Linux distributions In-Reply-To: Message-ID: Martin Schr{\"o}der asks on Sun, 12 Jun 2016 18:38:40 +0200: >> but I have to ask: Do you know what "Long-Term Support" means? >> I doubt it - otherwise you wouldn't imply that 16.04 should update to TL2016. >> ... >> > CentOS 7 (and thus, Red Hat 7 and Scientific Linux 7), the latest >> > version in that popular distribution family, still has only TeX Live >> > 2013. >> >> Small wonder: RHEL7 was released on 2014-06-10. >> >> OTOH Fedora currently has TL2015. >> ... The shelf-life of most Linux (and BSD) distributions is fairly short, especially in comparison to Solaris where I can still run binaries build in 1990 (and IBM mainframe O/Ses can sometimes run 60-year-old binaries). For Ubuntu, releases appear twice a year, and according to https://en.wikipedia.org/wiki/Ubuntu_%28operating_system%29#History_and_development_process every fourth release is labeled long-term support (LTS), and since version 12.04 LTS (on 26 April 2012), support is guaranteed for five years. Fedora's major releases seem to appear every few months: according to https://fedoraproject.org/wiki/Releases major version 1 appeared in 2003, and version 24 in 2016. Because many Linux distributions drop previously-released shared libraries, user-produced executables often fail to run after major version upgrades. That is a HUGE problem at many sites, commercial, government, and academic, and is a strong reason for the commercial success of companies like Red Hat and SUSE who offer customers support for more than a decade on each release. See https://access.redhat.com/support/policy/updates/errata/ https://www.suse.com/support/policy As an indication of the size of the problem, on the CentOS 7 desktop on which I'm writing this letter, I have 23,667 executables in my PATH, and 60% of those are not from the CentOS channel. It has never seemed to me that the release of an O/S major version should mean that all additional software should be frozen at the levels available sometime before the date of version n.0 of the kernel of that O/S. ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - University of Utah FAX: +1 801 581 4148 - - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - ------------------------------------------------------------------------------- From beebe at math.utah.edu Mon Jun 13 17:07:55 2016 From: beebe at math.utah.edu (Nelson H. F. Beebe) Date: Mon, 13 Jun 2016 09:07:55 -0600 Subject: [tex-live] 64-bit binaries for OpenIndiana 11: success Message-ID: I've just successfully built 64-bit binaries for TeX Live 2016 on OpenIndiana 11. The first attempt with -m64 added to CFLAGS and CXXFLAGS failed: gmake[6]: Leaving directory `/local/build/c89/texlive-20160513/source/Work/libs/luajit/native' native/buildvm -m bcdef -o lj_bcdef.h lib_base.c lib_math.c lib_bit.c lib_string.c lib_table.c lib_io.c lib_os.c lib_package.c lib_debug.c lib_ji\ t.c lib_ffi.c native/buildvm -m ffdef -o lj_ffdef.h lib_base.c lib_math.c lib_bit.c lib_string.c lib_table.c lib_io.c lib_os.c lib_package.c lib_debug.c lib_ji\ t.c lib_ffi.c Error: pointer size mismatch in cross-build. Try: make HOST_CC="gcc -m32" CROSS=... I deleted the source/Work directory and tried again: % env TL_CONFIGURE_FLAGS='--disable-luajittex --disable-mfluajit' \ ./build-texlive-2016.sh -m64 texlive-20160513 That succeeded, with 402 files in the bin directory. They are now available at the usual location http://www.math.utah.edu/pub/texlive/ I noticed that clisp is available in the CSW distribution, so I've just installed it, and am now redoing 32-bit and 64-bit builds on OpenIndiana 11 in the hope of getting xindy into the binary distributions. More later.... ---------------------------------------- P.S. Normally, we try to minimize the dependence of TeX Live executables on packages and libraries that are not provided by the main vendor distribution channel. For these new OpenIndiana builds, I needed ghostscript, freeglut, and mesa libraries from the OpenCSW channel, and their scores of dependents, all installed this morning. Here are the library dependendents from that channel: % /usr/uumath/src/texlive/show-texlive-libs.sh . | grep /opt 1 /opt/csw/lib/64/libGL.so.1 1 /opt/csw/lib/64/libGLU.so.1 1 /opt/csw/lib/64/libOSMesa.so.7 1 /opt/csw/lib/64/libgcc_s.so.1 1 /opt/csw/lib/64/libglut.so.3 1 /opt/csw/lib/64/libncurses.so.5 1 /opt/csw/lib/64/libreadline.so.5 1 /opt/csw/lib/64/libsigsegv.so.0 1 /opt/csw/lib/64/libz.so.1 Here are the executables that use those OpenCSW libraries: libGL asy libGLU asy libOSMesa asy libglut asy libncurses asy libreadline asy libsigsegv asy libz asy asy xelatex xelollipop xetex libgcc_s amstex asy bibtexu cfftot1 cslatex csplain \ dvilualatex dviluatex eplain etex \ jadetex latex lollipop lualatex \ lualollipop luatex mex mllatex mltex \ mmafm mmpfb otfinfo otftotfm \ pdfcslatex pdfcsplain pdfetex \ pdfjadetex pdflatex pdfmex pdftex \ pdftosrc pdfxmltex t1dotlessj t1lint \ t1rawafm t1reencode t1testpage \ teckit_compile texlua texluac texsis \ ttftotype42 upmendex utf8mex xelatex \ xelollipop xetex xmltex Regrettably, the -lgcc_s dependence is present in far too many executables, so users of these binaries will need to install several OpenCSW packages. ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - University of Utah FAX: +1 801 581 4148 - - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - ------------------------------------------------------------------------------- From beebe at math.utah.edu Mon Jun 13 22:45:14 2016 From: beebe at math.utah.edu (Nelson H. F. Beebe) Date: Mon, 13 Jun 2016 14:45:14 -0600 Subject: [tex-live] TeX Live 2016 binaries for OpenIndiana 11 Message-ID: Thanks to a question earlier today on this list, I'm pleased to announce the availability of both 32-bit and 64-bit binaries for OpenIndiana 11 (aka OpenSolaris) on x86 and x86-64 systems, with the necessary support of libraries from the OpenCSW distribution channel. The binaries now include asympotote and xindy, thanks to the clisp package from OpenCSW. See http://www.math.utah.edu/pub/texlive/ and two links to openindiana files. All is not entirely rosy, however. I've yet to succeed in building TeX Live 2016 on Hipster, the rolling-update version of OpenSolaris. Fortunately, the two OpenIndiana builds appear to work correctly on Hipster, and I'm using that approach for Hipster until native builds succeed. ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - University of Utah FAX: +1 801 581 4148 - - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - ------------------------------------------------------------------------------- From mojca.miklavec.lists at gmail.com Mon Jun 13 23:18:32 2016 From: mojca.miklavec.lists at gmail.com (Mojca Miklavec) Date: Mon, 13 Jun 2016 23:18:32 +0200 Subject: [tex-live] 64-bit binaries for OpenIndiana 11: success In-Reply-To: References: Message-ID: On 13 June 2016 at 17:07, Nelson H. F. Beebe wrote: > I've just successfully built 64-bit binaries for TeX Live 2016 on > OpenIndiana 11. The first attempt with -m64 added to CFLAGS and > CXXFLAGS failed: > > gmake[6]: Leaving directory `/local/build/c89/texlive-20160513/source/Work/libs/luajit/native' > > native/buildvm -m bcdef -o lj_bcdef.h lib_base.c lib_math.c lib_bit.c lib_string.c lib_table.c lib_io.c > lib_os.c lib_package.c lib_debug.c lib_ji\ > t.c lib_ffi.c > > native/buildvm -m ffdef -o lj_ffdef.h lib_base.c lib_math.c lib_bit.c lib_string.c lib_table.c lib_io.c > lib_os.c lib_package.c lib_debug.c lib_ji\ > t.c lib_ffi.c > > Error: pointer size mismatch in cross-build. > > Try: make HOST_CC="gcc -m32" CROSS=... > > I deleted the source/Work directory and tried again: > > % env TL_CONFIGURE_FLAGS='--disable-luajittex --disable-mfluajit' \ > ./build-texlive-2016.sh -m64 texlive-20160513 You can check my notes at http://tug.org/svn/texlive/trunk/Build/source/doc/README.solaris?view=markup If you want to build LuaJIT, you need to use export CC="/path/to/gcc -m64" export CXX="/path/to/g++ -m64" rather than setting those variables in CFLAGS/CXXFLAGS. I would say that's due to a bug in LuaJIT's configuration, but at the moment this is the only way to get the two additional binaries. > Here are the executables that use those OpenCSW libraries: > > libGL asy > libGLU asy > libOSMesa asy > libglut asy > libncurses asy > libreadline asy > libsigsegv asy You cannot avoid dependency on OpenGL anyway unless you want a "featureless" Asymptote. > libz asy asy xelatex xelollipop xetex In my case libz came from the OS: > ldd xetex libfontconfig.so.1 => /usr/lib/libfontconfig.so.1 libsocket.so.1 => /lib/libsocket.so.1 libm.so.2 => /lib/libm.so.2 libc.so.1 => /lib/libc.so.1 libfreetype.so.6 => /usr/sfw/lib/libfreetype.so.6 libz.so.1 => /usr/lib/libz.so.1 libexpat.so.0 => /usr/sfw/lib/libexpat.so.0 libnsl.so.1 => /lib/libnsl.so.1 libmp.so.2 => /lib/libmp.so.2 libmd.so.1 => /lib/libmd.so.1 libscf.so.1 => /lib/libscf.so.1 libdoor.so.1 => /lib/libdoor.so.1 libuutil.so.1 => /lib/libuutil.so.1 libgen.so.1 => /lib/libgen.so.1 /platform/SUNW,SPARC-Enterprise-T5220/lib/libc_psr.so.1 /platform/SUNW,SPARC-Enterprise-T5220/lib/libmd_psr.so.1 > libgcc_s amstex asy bibtexu cfftot1 cslatex csplain \ > dvilualatex dviluatex eplain etex \ > jadetex latex lollipop lualatex \ > lualollipop luatex mex mllatex mltex \ > mmafm mmpfb otfinfo otftotfm \ > pdfcslatex pdfcsplain pdfetex \ > pdfjadetex pdflatex pdfmex pdftex \ > pdftosrc pdfxmltex t1dotlessj t1lint \ > t1rawafm t1reencode t1testpage \ > teckit_compile texlua texluac texsis \ > ttftotype42 upmendex utf8mex xelatex \ > xelollipop xetex xmltex > > Regrettably, the -lgcc_s dependence is present in far too many > executables, so users of these binaries will need to install several > OpenCSW packages. I didn't get any dependency on libgcc_s at all. I assume that must have something to do with the way you used or installed the compiler. Mojca From werner at suse.de Tue Jun 14 18:35:50 2016 From: werner at suse.de (Dr. Werner Fink) Date: Tue, 14 Jun 2016 18:35:50 +0200 Subject: [tex-live] TL2016: Stack smashed in upmendex fprint_uchar() function Message-ID: <20160614163550.xoedjzx7wp27oq2b@boole.suse.de> Hi, during test I've seen a catched stack smash done by glibc fortify. It seems that in texk/upmendex/fwrite.c the function fprint_uchar() will be overflowed by the UChar inital array in indwrite() UChar datama[256],initial[INITIALLENGTH],initial_prev[INITIALLENGTH]; [...] fprint_uchar(fp,initial,lethead_flag,-1); as the macro INITIALLENGTH had shown 18 members during debugging whereas in the original fprint_uchar() only 5 members are handled static void fprint_uchar(FILE *fp, const UChar *a, const int mode, const int len) { int k; char str[15], *ret; UChar istr[5]; int olen, wclen; UErrorCode perr; if (len<0) { for (k=0; a[k] || k<4; k++) istr[k]=a[k]; wclen=k; } else { wclen = is_surrogate_pair(a) ? 2 : 1; istr[0]=a[0]; if (wclen==2) istr[1]=a[1]; } istr[wclen]=L'\0'; ^^^^^^^^^^^^^^^^^^ Stack smash maybe the macro INITIALLENGTH has other values for other builds as this crash only happens on a view build environments here around ;) Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -------------- next part -------------- A non-text attachment was scrubbed... Name: x Type: text/x-patch Size: 652 bytes Desc: Patch URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From norbert at preining.info Wed Jun 15 08:08:16 2016 From: norbert at preining.info (Norbert Preining) Date: Wed, 15 Jun 2016 15:08:16 +0900 Subject: [tex-live] TL2016: Stack smashed in upmendex fprint_uchar() function In-Reply-To: <20160614163550.xoedjzx7wp27oq2b@boole.suse.de> References: <20160614163550.xoedjzx7wp27oq2b@boole.suse.de> Message-ID: <20160615060816.GI10246@logic.at> Hi Werner, > during test I've seen a catched stack smash done by glibc fortify. > It seems that in texk/upmendex/fwrite.c the function fprint_uchar() > will be overflowed by the UChar inital array in indwrite() Hmm, that looks interesting. I have seen several test failures due to segfaults on Debian packages, too, and I will give this patch a try. Thanks! Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From werner at suse.de Wed Jun 15 10:08:10 2016 From: werner at suse.de (Dr. Werner Fink) Date: Wed, 15 Jun 2016 10:08:10 +0200 Subject: [tex-live] How to configure and build luajit for PPC on PPC64? Message-ID: <20160615080810.3hggqouiiua7ndti@boole.suse.de> Hi, as subject says I'd like to compile the luajit tree (libs and binaries) on PPC64 but for PPC (should be doable with -m32) but AFAICS the autoconfigure does not support to compile a part with a different bit environment? The reason is that ppc(32) is out of the focus but luajit does not (yet?) support ppc64. Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From werner at suse.de Wed Jun 15 10:13:58 2016 From: werner at suse.de (Dr. Werner Fink) Date: Wed, 15 Jun 2016 10:13:58 +0200 Subject: [tex-live] TL2016: Stack smashed in upmendex fprint_uchar() function In-Reply-To: <20160615060816.GI10246@logic.at> References: <20160614163550.xoedjzx7wp27oq2b@boole.suse.de> <20160615060816.GI10246@logic.at> Message-ID: <20160615081358.cut6em4gwgaspywb@boole.suse.de> On Wed, Jun 15, 2016 at 03:08:16PM +0900, Norbert Preining wrote: > Hi Werner, > > > during test I've seen a catched stack smash done by glibc fortify. > > It seems that in texk/upmendex/fwrite.c the function fprint_uchar() > > will be overflowed by the UChar inital array in indwrite() > > Hmm, that looks interesting. I have seen several test failures > due to segfaults on Debian packages, too, and I will give this > patch a try. Thanks! Even if the patch avoids the crash here I've no idea if the solution is fully correct, that is that fprint_uchar() does really its job for all 18 UChar members as the final char array is a bit small I guess. But IMHO this is a job for upstream of upmendex ;) Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From norbert at preining.info Wed Jun 15 10:24:30 2016 From: norbert at preining.info (Norbert Preining) Date: Wed, 15 Jun 2016 17:24:30 +0900 Subject: [tex-live] How to configure and build luajit for PPC on PPC64? In-Reply-To: <20160615080810.3hggqouiiua7ndti@boole.suse.de> References: <20160615080810.3hggqouiiua7ndti@boole.suse.de> Message-ID: <20160615082430.GO10246@logic.at> Hi Werner, > as subject says I'd like to compile the luajit tree (libs and binaries) > on PPC64 but for PPC (should be doable with -m32) but AFAICS the Hmm, AFAIR the luajit part is not supported on ppc, anyof them. On Debian we only build for some archs: LUAJIT_GOOD_ARCHS := amd64 armel armhf hurd-i386 i386 kfreebsd-amd64 kfreebsd-i386 powerpc FOr the others --disable-luajittex --disable-mfluajit All the best Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From werner at suse.de Wed Jun 15 10:59:46 2016 From: werner at suse.de (Dr. Werner Fink) Date: Wed, 15 Jun 2016 10:59:46 +0200 Subject: [tex-live] How to configure and build luajit for PPC on PPC64? In-Reply-To: <20160615082430.GO10246@logic.at> References: <20160615080810.3hggqouiiua7ndti@boole.suse.de> <20160615082430.GO10246@logic.at> Message-ID: <20160615085946.dfaldpstkllkmh7b@boole.suse.de> On Wed, Jun 15, 2016 at 05:24:30PM +0900, Norbert Preining wrote: > Hi Werner, > > > as subject says I'd like to compile the luajit tree (libs and binaries) > > on PPC64 but for PPC (should be doable with -m32) but AFAICS the > > Hmm, AFAIR the luajit part is not supported on ppc, anyof them. You meant ppc64 and ppc64le (whereas I prefere big endian as natural endianess), as here powerpc is given ;) > On Debian we only build for some archs: > LUAJIT_GOOD_ARCHS := amd64 armel armhf hurd-i386 i386 kfreebsd-amd64 kfreebsd-i386 powerpc > > FOr the others > --disable-luajittex --disable-mfluajit > > All the best Similar here, from spec file: # # LuaJIT -- does not build nor support all architectures # Current status is available at https://github.com/LuaJIT/LuaJIT # Compare with libs/luajit/LuaJIT-/src/lj_arch.h # %ifarch %ix86 x86_64 ppc arm mips %global with_LuaJIT 1 %endif %bcond_with LuaJIT [...] %if %{with LuaJIT} --enable-luajittex \ --enable-mfluajit \ %else --disable-luajittex \ --disable-mfluajit \ %endif even arm64 (aarch64) seems to be doable. Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From mojca.miklavec.lists at gmail.com Wed Jun 15 12:11:24 2016 From: mojca.miklavec.lists at gmail.com (Mojca Miklavec) Date: Wed, 15 Jun 2016 12:11:24 +0200 Subject: [tex-live] How to configure and build luajit for PPC on PPC64? In-Reply-To: <20160615082430.GO10246@logic.at> References: <20160615080810.3hggqouiiua7ndti@boole.suse.de> <20160615082430.GO10246@logic.at> Message-ID: > On 15 June 2016 at 10:08, Dr. Werner Fink wrote: >> >> as subject says I'd like to compile the luajit tree (libs and binaries) >> on PPC64 but for PPC (should be doable with -m32) but AFAICS the >> autoconfigure does not support to compile a part with a different bit >> environment? The problem is that you are not compiling "just that tiny bit". I'm not sure how you do the compilation on SuSE(?), but compiling LuaJIT means that you probably want to end up with a luajittex binary which has to link (either statically or dynamically) with other libraries like libpng. So you also need a lot of 32-bit libraries, not just luajit and I don't see any easy way to do that in a single step (unless you were compiling "fat" libraries and binaries containing multiple architectures like it used to be popular on Mac OS X). On 15 June 2016 at 10:24, Norbert Preining wrote: > > Hmm, AFAIR the luajit part is not supported on ppc, anyof them. That's not entirely true. It is supposed to work on PPC on Linux. It even compiles on Mac and some basic functionality works (I can even make a ConTeXt format file and typeset a basic document on Mac OS X 10.5/PPC), but there are examples that dump core and the typeset document is somewhat junk (some weird scaling of fonts). But on the other hand luajittex doesn't really bring any extra functionality, it is just supposed to be slightly optimized in some aspects. (One could in principle make a symlink from luajittex to luatex and users wouldn't even notice.) If I was packaging TeX Live on a distribution that doesn't have native support for luajit, I would not bother about ugly hacks to compile luajittex. Mojca From herbs at wideopenwest.com Wed Jun 15 16:04:11 2016 From: herbs at wideopenwest.com (Herbert Schulz) Date: Wed, 15 Jun 2016 09:04:11 -0500 Subject: [tex-live] today's pdftex update Message-ID: <0AF15335-D112-42A8-A1A5-2DA903DEA0C1@wideopenwest.com> Howdy, With today's update to pdftex I no longer can get automatic conversion of file.eps->file-eps-converted-to.pdf without using --shell-escape. I get the error message !! Error: Option forbidden in restricted mode: --device system returned with code 65280 Good Luck, Herb Schulz (herbs at wideopenwest dot com) From d.p.carlisle at gmail.com Wed Jun 15 16:07:43 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Wed, 15 Jun 2016 15:07:43 +0100 Subject: [tex-live] today's pdftex update In-Reply-To: <0AF15335-D112-42A8-A1A5-2DA903DEA0C1@wideopenwest.com> References: <0AF15335-D112-42A8-A1A5-2DA903DEA0C1@wideopenwest.com> Message-ID: On 15 June 2016 at 15:04, Herbert Schulz wrote: > Howdy, > > With today's update to pdftex I no longer can get automatic conversion of > file.eps->file-eps-converted-to.pdf without using --shell-escape. I get the > error message > > !! Error: Option forbidden in restricted mode: --device > system returned with code 65280 > > Good Luck, > > Herb Schulz > (herbs at wideopenwest dot com) > > > The code was re-arranged, but it's supposed to work, let me check..... -------------- next part -------------- An HTML attachment was scrubbed... URL: From ttk at t-lab.opal.ne.jp Wed Jun 15 16:45:02 2016 From: ttk at t-lab.opal.ne.jp (TANAKA Takuji) Date: Wed, 15 Jun 2016 23:45:02 +0900 Subject: [tex-live] TL2016: Stack smashed in upmendex fprint_uchar() function In-Reply-To: <20160615081358.cut6em4gwgaspywb@boole.suse.de> References: <20160614163550.xoedjzx7wp27oq2b@boole.suse.de> <20160615060816.GI10246@logic.at> <20160615081358.cut6em4gwgaspywb@boole.suse.de> Message-ID: Hi Werner, Thank you for your report. I (the author of upmendex) will fix the issue later. Takuji On 2016/06/15 17:13, Dr. Werner Fink wrote: > On Wed, Jun 15, 2016 at 03:08:16PM +0900, Norbert Preining wrote: >> Hi Werner, >> >>> during test I've seen a catched stack smash done by glibc fortify. >>> It seems that in texk/upmendex/fwrite.c the function fprint_uchar() >>> will be overflowed by the UChar inital array in indwrite() >> >> Hmm, that looks interesting. I have seen several test failures >> due to segfaults on Debian packages, too, and I will give this >> patch a try. Thanks! > > Even if the patch avoids the crash here I've no idea if the solution > is fully correct, that is that fprint_uchar() does really its job for > all 18 UChar members as the final char array is a bit small I guess. > > But IMHO this is a job for upstream of upmendex ;) > > Werner > From d.p.carlisle at gmail.com Wed Jun 15 16:58:42 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Wed, 15 Jun 2016 15:58:42 +0100 Subject: [tex-live] today's pdftex update In-Reply-To: References: <0AF15335-D112-42A8-A1A5-2DA903DEA0C1@wideopenwest.com> Message-ID: On 15 June 2016 at 15:07, David Carlisle wrote: > > > On 15 June 2016 at 15:04, Herbert Schulz wrote: > >> Howdy, >> >> With today's update to pdftex I no longer can get automatic conversion of >> file.eps->file-eps-converted-to.pdf without using --shell-escape. I get the >> error message >> >> !! Error: Option forbidden in restricted mode: --device >> system returned with code 65280 >> >> Good Luck, >> >> Herb Schulz >> (herbs at wideopenwest dot com) >> >> >> > The code was re-arranged, but it's supposed to work, let me check..... > Oh, I spotted a couple of issues with the graphics package re-arrangement so I may have to update but I think the underlying error isn't my fault this time (for once:-) I get on a test file more or less as you reported: (./rotex.exepstopdf ($Id: epstopdf.pl 41287 2016-05-28 18:09:45Z karl $) 2.24 !!! Error: Option forbidden in restricted mode: --device Karl, is this a change in epstopdf? David -------------- next part -------------- An HTML attachment was scrubbed... URL: From norbert at preining.info Wed Jun 15 17:11:24 2016 From: norbert at preining.info (Norbert Preining) Date: Thu, 16 Jun 2016 00:11:24 +0900 Subject: [tex-live] today's pdftex update In-Reply-To: References: <0AF15335-D112-42A8-A1A5-2DA903DEA0C1@wideopenwest.com> Message-ID: <20160615151124.GS10246@logic.at> Hi Herb, hi all. > >> !! Error: Option forbidden in restricted mode: --device Yes, support for the device options was removed in the restricted mode. It should not be necessary at all. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From d.p.carlisle at gmail.com Wed Jun 15 17:22:38 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Wed, 15 Jun 2016 16:22:38 +0100 Subject: [tex-live] today's pdftex update In-Reply-To: <20160615151124.GS10246@logic.at> References: <0AF15335-D112-42A8-A1A5-2DA903DEA0C1@wideopenwest.com> <20160615151124.GS10246@logic.at> Message-ID: On 15 June 2016 at 16:11, Norbert Preining wrote: > Hi Herb, hi all. > >> >> !! Error: Option forbidden in restricted mode: --device > > Yes, support for the device options was removed in the restricted > mode. It should not be necessary at all. > > Norbert > yes but... the macros don't appear to be using that option the log shows runsystem(repstopdf --outfile=example-image-eps-converted-to.pdf example-image. epstopdf ($Id: epstopdf.pl 41287 2016-05-28 18:09:45Z karl $) 2.24 !!! Error: Option forbidden in restricted mode: --device system returned with code 65280 eps)...executed safely (allowed). so if I read that right, the option isn't used, and repstopdf is allowed to run but then complains about the option? David From news3 at nililand.de Wed Jun 15 22:57:40 2016 From: news3 at nililand.de (Ulrike Fischer) Date: Wed, 15 Jun 2016 22:57:40 +0200 Subject: [tex-live] today's pdftex update References: <0AF15335-D112-42A8-A1A5-2DA903DEA0C1@wideopenwest.com> <20160615151124.GS10246@logic.at> Message-ID: <1pdgo38no5ld7.dlg@nililand.de> Am Wed, 15 Jun 2016 16:22:38 +0100 schrieb David Carlisle: > On 15 June 2016 at 16:11, Norbert Preining wrote: >> Hi Herb, hi all. >> >>> >> !! Error: Option forbidden in restricted mode: --device >> >> Yes, support for the device options was removed in the restricted >> mode. It should not be necessary at all. > > yes but... the macros don't appear to be using that option > > the log shows > > runsystem(repstopdf --outfile=example-image-eps-converted-to.pdf example-image. > epstopdf ($Id: epstopdf.pl 41287 2016-05-28 18:09:45Z karl $) 2.24 > !!! Error: Option forbidden in restricted mode: --device > system returned with code 65280 > eps)...executed safely (allowed). > > so if I read that right, the option isn't used, and repstopdf is > allowed to run but then complains about the option? Hm. epstopdf.pl contains $::opt_device = $default_device; and later if ($::opt_device) { if ($restricted) { error "Option forbidden in restricted mode: --device"; } else { debug "Switching from $default_device to $::opt_device"; } } This looks wrong. -- Ulrike Fischer http://www.troubleshooting-tex.de/ From ml at mariochiari.net Thu Jun 16 00:01:09 2016 From: ml at mariochiari.net (mario chiari) Date: Thu, 16 Jun 2016 00:01:09 +0200 Subject: [tex-live] 2016 or 2015? In-Reply-To: <6664896.Goedeo5Ftj@x61s.fritz.box> References: <1465661752.3811.6.camel@mariochiari.net> <1651126.kOWM2cz86e@x61s.fritz.box> <1465751088.3535.10.camel@mariochiari.net> <6664896.Goedeo5Ftj@x61s.fritz.box> Message-ID: <1466028069.14972.4.camel@mariochiari.net> On Sun, 2016-06-12 at 20:40 +0200, Markus Kohm wrote: > Am Sonntag, 12. Juni 2016, 19:04:48 schrieb mario chiari: > > > > by the way, why? > Because it does not start the intall_tl from PATH but install_tl located at? > the current working directory. yes, of course thanks mario From norbert at preining.info Thu Jun 16 02:34:35 2016 From: norbert at preining.info (Norbert Preining) Date: Thu, 16 Jun 2016 09:34:35 +0900 Subject: [tex-live] today's pdftex update In-Reply-To: <1pdgo38no5ld7.dlg@nililand.de> References: <0AF15335-D112-42A8-A1A5-2DA903DEA0C1@wideopenwest.com> <20160615151124.GS10246@logic.at> <1pdgo38no5ld7.dlg@nililand.de> Message-ID: <20160616003435.GV10246@logic.at> > $::opt_device = $default_device; Yes, I will fix if for today. Thanks! Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From werner at suse.de Thu Jun 16 10:44:25 2016 From: werner at suse.de (Dr. Werner Fink) Date: Thu, 16 Jun 2016 10:44:25 +0200 Subject: [tex-live] hyphen -- nearly all languages added by default Message-ID: <20160616084425.adu5xvcea6zygzkw@boole.suse.de> Hi, just get a report that language.dat nearly all hyphen are added by default (package hyphen-base). What is the reason for this? Shouldn't this be done by AddHyphen execute of the specific hyphen package? Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From mojca.miklavec.lists at gmail.com Thu Jun 16 12:06:12 2016 From: mojca.miklavec.lists at gmail.com (Mojca Miklavec) Date: Thu, 16 Jun 2016 12:06:12 +0200 Subject: [tex-live] hyphen -- nearly all languages added by default In-Reply-To: <20160616084425.adu5xvcea6zygzkw@boole.suse.de> References: <20160616084425.adu5xvcea6zygzkw@boole.suse.de> Message-ID: On 16 June 2016 at 10:44, Dr. Werner Fink wrote: > Hi, > > just get a report that language.dat nearly all hyphen are added by > default (package hyphen-base). This must be a misinformation. The hyphen-base package only adds the Knuth's English patterns. It is true though that the "default" full installation of TeX Live installs all hyphenation patterns. > Shouldn't this be done by AddHyphen execute of the specific hyphen > package? This is already the case. It used to be the case until TL 2016 that hyph-utf8 shipped all the patterns and hyphen- packages only enabled those patterns in language.dat(.def). But since recently even the patterns themselves are shipped by individual packages. See http://tug.org/svn/texlive/trunk/Master/tlpkg/tlpsrc/hyphen-german.tlpsrc?revision=40340&view=markup for example. Mojca From arthur.reutenauer at normalesup.org Thu Jun 16 12:26:38 2016 From: arthur.reutenauer at normalesup.org (Arthur Reutenauer) Date: Thu, 16 Jun 2016 11:26:38 +0100 Subject: [tex-live] hyphen -- nearly all languages added by default In-Reply-To: References: <20160616084425.adu5xvcea6zygzkw@boole.suse.de> Message-ID: <20160616102638.GI330775@phare.normalesup.org> > This must be a misinformation. The hyphen-base package only adds the > Knuth's English patterns. > > It is true though that the "default" full installation of TeX Live > installs all hyphenation patterns. I can confirm that. We need more information to diagnose the problem; it may be that the user who reported this just installed the default scheme of the net installer, and hence got all the patterns. Best, Arthur From norbert at preining.info Thu Jun 16 13:11:44 2016 From: norbert at preining.info (Norbert Preining) Date: Thu, 16 Jun 2016 20:11:44 +0900 Subject: [tex-live] hyphen -- nearly all languages added by default In-Reply-To: <20160616084425.adu5xvcea6zygzkw@boole.suse.de> References: <20160616084425.adu5xvcea6zygzkw@boole.suse.de> Message-ID: <20160616111144.GZ10246@logic.at> Hi Werner, > just get a report that language.dat nearly all hyphen are added by > default (package hyphen-base). What is the reason for this? With the very unclear information it is hard to diagnose. Do you speak about upstream TeX LIve (TUG) or the SuSE packaging? I *can* imagine that you ship in the TL package the language.dat that is included in TL - this is the one that containss *ALL* patterns. FOr an actual installation the AddHyphen directives need to be evaluated. In TeX Live from TUG the installer/tlmgr doe that and generates a fitting language.dat. In Debian every package ships a snippet for language.dat (.def,dat.lua) adn those are combined at installation time. So bottomline: without further information it is hard to give a good explanation. All the best Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From norbert at preining.info Thu Jun 16 13:12:23 2016 From: norbert at preining.info (Norbert Preining) Date: Thu, 16 Jun 2016 20:12:23 +0900 Subject: [tex-live] How to configure and build luajit for PPC on PPC64? In-Reply-To: <20160615085946.dfaldpstkllkmh7b@boole.suse.de> References: <20160615080810.3hggqouiiua7ndti@boole.suse.de> <20160615082430.GO10246@logic.at> <20160615085946.dfaldpstkllkmh7b@boole.suse.de> Message-ID: <20160616111223.GA10246@logic.at> > You meant ppc64 and ppc64le (whereas I prefere big endian as natural endianess), > as here powerpc is given ;) Of course, sorry ... Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From werner at suse.de Thu Jun 16 14:48:07 2016 From: werner at suse.de (Dr. Werner Fink) Date: Thu, 16 Jun 2016 14:48:07 +0200 Subject: [tex-live] hyphen -- nearly all languages added by default In-Reply-To: <20160616111144.GZ10246@logic.at> References: <20160616084425.adu5xvcea6zygzkw@boole.suse.de> <20160616111144.GZ10246@logic.at> Message-ID: <20160616124807.h5sgnqrenpqwnvso@boole.suse.de> On Thu, Jun 16, 2016 at 08:11:44PM +0900, Norbert Preining wrote: > Hi Werner, > > > just get a report that language.dat nearly all hyphen are added by > > default (package hyphen-base). What is the reason for this? > > With the very unclear information it is hard to diagnose. Do > you speak about upstream TeX LIve (TUG) or the SuSE packaging? > > I *can* imagine that you ship in the TL package the language.dat > that is included in TL - this is the one that containss *ALL* > patterns. > > FOr an actual installation the AddHyphen directives need to be > evaluated. In TeX Live from TUG the installer/tlmgr doe that > and generates a fitting language.dat. > > In Debian every package ships a snippet for language.dat (.def,dat.lua) > adn those are combined at installation time. > > So bottomline: without further information it is hard to > give a good explanation. Doing a tar Oxf archive/hyphen-base.tar.xz tex/generic/config/language.dat shows that *all* hyphens had already been added. The same for language.def and language.dat.lua. And clearly this file has been packaged for openSUSE as I was not aware of this major change. Indeed thos caused a bug that all hyphens are required for generating the formats now. Also the perl update of those files are a nightmare for rpm based distributions. Using tlmgr is not an option ... I've looked at TeXLive/TL*.pm and with this I see, after some small changes, that without a local copy of the already installed hyphens the last installed hyphen does overwrite all others :( Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From norbert at preining.info Thu Jun 16 15:07:45 2016 From: norbert at preining.info (Norbert Preining) Date: Thu, 16 Jun 2016 22:07:45 +0900 Subject: [tex-live] hyphen -- nearly all languages added by default In-Reply-To: <20160616124807.h5sgnqrenpqwnvso@boole.suse.de> References: <20160616084425.adu5xvcea6zygzkw@boole.suse.de> <20160616111144.GZ10246@logic.at> <20160616124807.h5sgnqrenpqwnvso@boole.suse.de> Message-ID: Dear Werner Only quick from the mobile On June 16, 2016 9:48:07 PM GMT+09:00, "Dr. Werner Fink" wrote: > tar Oxf archive/hyphen-base.tar.xz tex/generic/config/language.dat > >shows that *all* hyphens had already been added. The same for Yes, but it has been like this since long. A distribution (including TeX Live itself) that allows for partial install may not use this file, by either generating a fitting copy and optionally not even including this file. How did you do this till 2015? You still had to activate patterns on demand at package installation time, right? (Same with font maps and format definitions!) If you don't like the Perl modules, at package (rpm) creation date you can parse the tlpdb trivially in shell (enough examples are floating around) to pull out the AddHyphen lines used in the packages to be built, and adjust the post-install routine to activate the hyphens (and deactivate on uninstall) in the system language.* If you can use Perl, just load the TL modules, in particular TLPDB to parse, and TLUtils to generate the language.* files. All the best Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From werner at suse.de Thu Jun 16 15:27:25 2016 From: werner at suse.de (Dr. Werner Fink) Date: Thu, 16 Jun 2016 15:27:25 +0200 Subject: [tex-live] hyphen -- nearly all languages added by default In-Reply-To: <20160616124807.h5sgnqrenpqwnvso@boole.suse.de> References: <20160616084425.adu5xvcea6zygzkw@boole.suse.de> <20160616111144.GZ10246@logic.at> <20160616124807.h5sgnqrenpqwnvso@boole.suse.de> Message-ID: <20160616132725.jonwa4cslactzpl3@boole.suse.de> On Thu, Jun 16, 2016 at 02:48:07PM +0200, Werner Fink wrote: > On Thu, Jun 16, 2016 at 08:11:44PM +0900, Norbert Preining wrote: > > Hi Werner, > > > > > just get a report that language.dat nearly all hyphen are added by > > > default (package hyphen-base). What is the reason for this? > > > > With the very unclear information it is hard to diagnose. Do > > you speak about upstream TeX LIve (TUG) or the SuSE packaging? > > > > I *can* imagine that you ship in the TL package the language.dat > > that is included in TL - this is the one that containss *ALL* > > patterns. > > > > FOr an actual installation the AddHyphen directives need to be > > evaluated. In TeX Live from TUG the installer/tlmgr doe that > > and generates a fitting language.dat. > > > > In Debian every package ships a snippet for language.dat (.def,dat.lua) > > adn those are combined at installation time. > > > > So bottomline: without further information it is hard to > > give a good explanation. > > Doing a > > tar Oxf archive/hyphen-base.tar.xz tex/generic/config/language.dat > > shows that *all* hyphens had already been added. The same for > language.def and language.dat.lua. And clearly this file has been > packaged for openSUSE as I was not aware of this major change. > Indeed thos caused a bug that all hyphens are required for generating > the formats now. > > Also the perl update of those files are a nightmare for rpm based > distributions. Using tlmgr is not an option ... I've looked at > TeXLive/TL*.pm and with this I see, after some small changes, that > without a local copy of the already installed hyphens the last > installed hyphen does overwrite all others :( Indeed using TeXLive::TLPDB creates via our $hyphen = "$texdir/tex/generic/config/language.dat"; our $hlocal = "/var/lib/texmf/config/language.dat"; TeXLive::TLUtils::create_language_dat($tlpdb, $hyphen, $hlocal); the same content as already in archive/hyphen-base.tar.xz ... beside the first two lines (Generated by / $Id:). Hacking TeXLive/TLPOBJ.pm to have a method root and skipping the usermode in TeXLive/TLUtils.pm does only write out one entry. Do I miss something here? Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From werner at suse.de Thu Jun 16 15:32:24 2016 From: werner at suse.de (Dr. Werner Fink) Date: Thu, 16 Jun 2016 15:32:24 +0200 Subject: [tex-live] hyphen -- nearly all languages added by default In-Reply-To: References: <20160616084425.adu5xvcea6zygzkw@boole.suse.de> <20160616111144.GZ10246@logic.at> <20160616124807.h5sgnqrenpqwnvso@boole.suse.de> Message-ID: <20160616133224.3rxgkxscydjltyog@boole.suse.de> On Thu, Jun 16, 2016 at 10:07:45PM +0900, Norbert Preining wrote: > Dear Werner > > Only quick from the mobile > > > On June 16, 2016 9:48:07 PM GMT+09:00, "Dr. Werner Fink" wrote: > > tar Oxf archive/hyphen-base.tar.xz tex/generic/config/language.dat > > > >shows that *all* hyphens had already been added. The same for > > Yes, but it has been like this since long. A distribution (including TeX Live itself) that allows for partial install may not use this file, by either generating a fitting copy and optionally not even including this file. > > How did you do this till 2015? You still had to activate patterns on demand at package installation time, right? (Same with font maps and format definitions!) For 2015 the spec file post scriptlets had worked well. And indeed I generate those scriptlets from the texlive.tlpdb.xz during generating the spec files (in perl). > > If you don't like the Perl modules, at package (rpm) creation date you can parse the tlpdb trivially in shell (enough examples are floating around) to pull out the AddHyphen lines used in the packages to be built, and adjust the post-install routine to activate the hyphens (and deactivate on uninstall) in the system language.* > > If you can use Perl, just load the TL modules, in particular TLPDB to parse, and TLUtils to generate the language.* files. That is what I've tried but the TLPDB does include all AddHyphen lines found in texlive.tlpdb but do not skip the not installed packages. Also it would be a great win if TLPDB could read xz-zipped texlive.tlpdb ;) Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From preining at logic.at Thu Jun 16 15:59:07 2016 From: preining at logic.at (Norbert Preining) Date: Thu, 16 Jun 2016 22:59:07 +0900 Subject: [tex-live] hyphen -- nearly all languages added by default In-Reply-To: <20160616133224.3rxgkxscydjltyog@boole.suse.de> <20160616132725.jonwa4cslactzpl3@boole.suse.de> Message-ID: <20160616135907.GA21093@logic.at> Hi Werner, On Thu, 16 Jun 2016, Dr. Werner Fink wrote: > Indeed using TeXLive::TLPDB creates via > > our $hyphen = "$texdir/tex/generic/config/language.dat"; > our $hlocal = "/var/lib/texmf/config/language.dat"; > TeXLive::TLUtils::create_language_dat($tlpdb, $hyphen, $hlocal); > > the same content as already in archive/hyphen-base.tar.xz ... But from *which* tlpdb did you generate the language.dat? My guess is that you are misundersstanding TLPDBs. They describe the status of the TeX Live installation at *that*point*. That means, the TLPDB from tlnet (or DVD, or svn) contains *all* packages - because all are installed, and thus also all AddHyphen lines. > That is what I've tried but the TLPDB does include all AddHyphen lines > found in texlive.tlpdb but do not skip the not installed packages. This is incorrect - or you are using the wrong TLPDB. I just started install-tl and selected scheme-minimal, and then grepped fro AddHyphen lines [~/tlfoo/2016/tlpkg] grep AddHyphen texlive.tlpdb norbert at wienerschnitzel: 1 [~/tlfoo/2016/tlpkg] Nothing. THe reason is that the installer *creates* a new TLPDB, and adds only those packages that are actually installed. On the Debian side we don't do this, because we don't allow for tlmgr to do anything but user mode, and thus we *know* at package build time what are the execute actions. Again my guess: You are doing this from the *full* tlpdb from which you are selecting package to be installed. My suggestion: You need to create a *new* tlpdb, and for each package $pkg installed into the package you include in the rpm, you need to collect the execute lines ($tlpobj->executes plus some utility functions in TLUtils). Another approach would be that: * each rom ships the .tlpobj files for teh TL packages included in it * on installation the tlpobj are *added* to the user computer tlpdb * language.dat (etc) is generated in the postinst form the user computer tlpdb But that sounds complicated. Don't you have something like "many packages contribute to one configuration file via snippets" ? How does SuSE do that with emacs packages, python, perl, etc? > Also it would be a great win if TLPDB could read xz-zipped texlive.tlpdb ;) It does already, but in a slightly tricky way ;-) SOrry for that my $tlpdb = TeXLive::TLPDB->new; my $nr_pkgs_loaded = $tlpdb->from_file("file://path/to/root/") *should* try to load /path/to/root/tlpkg/texlive.tlpdb.xz and only then the non-xz variant. But this is a hardly travelled code path ... so I don't put my hand into the fire for it ;-) But I will think about to extend this to reading .xz files from TeXLive::TLPDB->new(root => "/path/to/root") Hope the above explanations help - although I am still stymed what has changed so deeply to 2016 that your scripts broke. ALl the best Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From samuel.c.p.neves at gmail.com Fri Jun 17 01:29:34 2016 From: samuel.c.p.neves at gmail.com (Samuel Neves) Date: Fri, 17 Jun 2016 00:29:34 +0100 Subject: [tex-live] xcolor broken with pdflatex Message-ID: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> Minimal example: \documentclass{standalone} \usepackage{xcolor} \begin{document} \end{document} This fails with pdflatex with an undefined control sequence error. Including graphicx before xcolor fixes this. This appears related to this change: https://github.com/ho-tex/pdftex-def/commit/ea30c0fbc033ce69f7dd43ea266270b9f09dd7d3 From d.p.carlisle at gmail.com Fri Jun 17 01:43:20 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Fri, 17 Jun 2016 00:43:20 +0100 Subject: [tex-live] xcolor broken with pdflatex In-Reply-To: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> References: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> Message-ID: On 17 June 2016 at 00:29, Samuel Neves wrote: > Minimal example: > > \documentclass{standalone} > > \usepackage{xcolor} > > \begin{document} > > \end{document} > > > This fails with pdflatex with an undefined control sequence error. Including graphicx before xcolor fixes this. This appears related to this change: https://github.com/ho-tex/pdftex-def/commit/ea30c0fbc033ce69f7dd43ea266270b9f09dd7d3 sigh, thanks for the report I'll fix (could also be fixed in xcolor, but I guess I should do it. If you don't want to load graphics, \makeatletter \newif\ifGin at setpagesize \makeatother would also avoid the error. David From zarniwhoop at ntlworld.com Fri Jun 17 02:04:19 2016 From: zarniwhoop at ntlworld.com (Ken Moffat) Date: Fri, 17 Jun 2016 01:04:19 +0100 Subject: [tex-live] Asymptote 3D and gcc-6.1 ? Message-ID: <20160617000419.GA3140@milliways.localdomain> Verging on O/T, but I'm not sure where to ask (I suspect the problem is in freeglut) and I'm making no headway trying to debug this. The linux x86_64 2016 binary works fine on my Beyond Linuxfromscratch 7.9 system, where gcc is 5.3.1. But on a development system using gcc-6.1 it fails to create 3D drawings and I can't get close to identifying the exact problem (the 2D drawings seem fine). The binary uses ncurses5, readline5, and freeglut from the host, but since the 2D drawing is fine I am guessing that freeglut is the problem. I updated my test scripts to include a 3D drawing: http://www.linuxfromscratch.org/~ken/tex-testfiles/latex-test-20160523.tar.xz (run ./configure - it won't install anything - and then make cylinder) but that juist craps out with: (./cylinder-1.tex LaTeX Warning: File `cylinder-1+0_0.pdf' not found on input line 14. ! Package pdftex.def Error: File `cylinder-1+0_0.pdf' not found. See the pdftex.def package documentation for explanation. Type H for immediate help. ... l.14 ...-1+0_0.pdf}}}{\ASYprefix\jobname-1+0.prc}} and sure enough, all I have is cylinder-1.0.{pdf,prc} as well as cylinder-1.{asy,pre,tex} : I tried diffing the stdout from runs on gcc-5.3.1 and gcc-6.1 but it was too different for me to make any sense of it. Then I tried building asymptote from source on gcc-6.1, but of course nothing changed (I did add extra deps such as fftw and Imaging). I've tried kicking freeglut's cmake until I managed to get it to build with no optimization at all (other gcc-6.1 bugs have sometimes been worked around by disabling optimization), and then adding all the -fno-this-and-that flags I've had to add to other packages, but no change. Can anybody offer any pointers, please ? ?en -- War is Peace Freedom is Slavery Ignorance is Strength From d.p.carlisle at gmail.com Fri Jun 17 03:26:10 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Fri, 17 Jun 2016 02:26:10 +0100 Subject: [tex-live] xcolor broken with pdflatex In-Reply-To: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> References: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> Message-ID: I just updated latex-graphics, luatex-def and pdftex-def on ctan (for dvips.def, luatex.def and pdftex.def) to guard against the files being loaded without the new pagesize option being set up. Sorry about that , David From vk489 at uowmail.edu.au Thu Jun 16 03:31:17 2016 From: vk489 at uowmail.edu.au (Vafa Khalighi) Date: Thu, 16 Jun 2016 01:31:17 +0000 Subject: [tex-live] Lollipop is broken in TeXlive 2016 Message-ID: Hi Karl lollipop is broken in TeXLive 2016. Create a simple lollipop document \Start This is a test. \Stop and process it with `lollipop test.tex' and this is the error you get %%%%%%%%%% This is pdfTeX, Version 3.14159265-2.6-1.40.17 (TeX Live 2016) (preloaded format=lollipop) restricted \write18 enabled. kpathsea: Running mktexfmt lollipop.fmt mktexfmt: mktexfmt is using the following fmtutil.cnf files (in precedence order): mktexfmt: /usr/local/texlive/2016/texmf-dist/web2c/fmtutil.cnf mktexfmt: mktexfmt is using the following fmtutil.cnf file for writing changes: mktexfmt: /Users/vafa/Library/texlive/2016/texmf-config/web2c/fmtutil.cnf mktexfmt [INFO]: writing formats under /Users/vafa/Library/texlive/2016/texmf-var/web2c mktexfmt [INFO]: --- remaking lollipop with tex mktexfmt: running `tex -ini -jobname=lollipop -progname=lollipop lollipop.ini' ... This is TeXk, Version 3.14159265 (TeX Live 2016) (INITEX) restricted \write18 enabled. (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop.ini (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop.tex (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-plain.tex (/usr/local/texlive/2016/texmf-dist/tex/generic/hyphen/hyphen.tex)) (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-tools.tex) (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-define.tex) (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-fonts.tex) (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-fontdefs.tex) (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-text.tex) (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-document.tex) (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-heading.tex) (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-output.tex) (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-float.tex) (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-lists.tex) ) ) Beginning to dump on file lollipop.fmt (preloaded format=lollipop 2016.6.16) 3422 strings of total length 48580 33354 memory locations dumped; current usage is 158&33023 2325 multiletter control sequences \font\nullfont=nullfont \font\exx=cmex10 \font\ttx=cmtt10 \font\ComputerModernromanx=cmr10 \font\ttix=cmtt9 \font\ttxi=cmtt10 at 10.95pt \font\ttxii=cmtt10 at 12.0pt \font\ttxiv=cmtt10 at 14.4pt \font\ttxvii=cmtt10 at 17.28pt \font\ttxx=cmtt10 at 20.73999pt \font\ComputerModernboldx=cmbx10 \font\ComputerModernslantx=cmsl10 \font\ComputerModernitalicx=cmti10 2706 words of font info for 12 preloaded fonts 19 hyphenation exceptions Hyphenation trie of length 6075 has 181 ops out of 35111 181 for language 0 No pages of output. Transcript written on lollipop.log. mktexfmt [INFO]: /Users/vafa/Library/texlive/2016/texmf-var/web2c/tex/lollipop.fmt installed. mktexfmt [INFO]: Disabled formats: 6 mktexfmt [INFO]: Successfully rebuilt formats: 1 mktexfmt [INFO]: Not selected formats: 43 mktexfmt [INFO]: Total formats: 50 mktexfmt [INFO]: exiting with status 0 ---! /Users/vafa/Library/texlive/2016/texmf-var/web2c/tex/lollipop.fmt was written by tex (Fatal format file error; I'm stymied) %%%%%%%%%%%%%%%%%%%%%% This is because, lollipop format should be based on Knuth tex and not pdftex. Indeed, when I look at TeXLive repository, I can confirm that only netbsd platforms have lollipop based on Knuth TeX (link tex) and all other platforms link pdftex. I believe I sent you an email and explained this clearly to you when you asked me to delete xelollipop. lollipop is the only tex format that I use myself and this breaks all my documents. Could you please fix this? Thanks -------------- next part -------------- An HTML attachment was scrubbed... URL: From hendry at webconverger.com Thu Jun 16 17:22:58 2016 From: hendry at webconverger.com (Kai Hendry) Date: Thu, 16 Jun 2016 16:22:58 +0100 Subject: [tex-live] Letter template Message-ID: Hi there, I created a Docker image with texlive to help generate a letter from a Web form in the very occasional times I need to. However I recently got a complaint from my wife that the template the font is too small and the borders are too wide and the signature is odd for being offset to the right. This is the current template I am using and it must be 10 years old. Can anyone offer offer pointers to a better (more modern) template? https://github.com/kaihendry/letterly/blob/master/www/pdf/letter.template \documentclass[british,a4paper]{letter} \usepackage{fontspec} \defaultfontfeatures{Mapping=tex-text} \usepackage{xunicode} \usepackage{babel} \begin{document} Many thanks! From leonid.sinev at yandex.ru Thu Jun 16 21:14:56 2016 From: leonid.sinev at yandex.ru (Leonid Sinev) Date: Thu, 16 Jun 2016 22:14:56 +0300 Subject: [tex-live] texcount on Windows fail because of Perl file missing from TeX Live distribution Message-ID: <306071466104496@web16h.yandex.ru> Greetings, This problem already discussed at https://www.tug.org/pipermail/tex-live/2015-August/037293.html but nothing changed in TeXLive 2016. Nothing changed after http://tex.stackexchange.com/a/145766 advice of Win32::Console::ANSI from CPAN I have copied file from this package to C:\texlive\2016\tlpkg\tlperl\site\lib\Win32\Console\ANSI.pm and got this: Can't locate loadable object for module Win32::Console::ANSI in @INC (@INC conta ins: C:/texlive/2016/tlpkg/tlperl/lib c:/texlive/2016/tlpkg/tlperl/site/lib c:/t exlive/2016/tlpkg/tlperl/lib .) at c:\texlive\2016\texmf-dist\scripts\texcount\t excount.pl line 11. Compilation failed in require at c:\texlive\2016\texmf-dist\scripts\texcount\tex count.pl line 11. BEGIN failed--compilation aborted at c:\texlive\2016\texmf-dist\scripts\texcount \texcount.pl line 14. C:\texlive\2016\bin\win32\runscript.tlu:679: command failed with exit code 2: perl.exe c:\texlive\2016\texmf-dist\scripts\texcount\texcount.pl Can there be a patch or update for TeXLive for windows that can make windows users use Perl scripts like texcount without need of additional perl installation or messing with additional packaging for tlperl? --? Best regards, Leonid From royalts at gmail.com Thu Jun 16 21:20:32 2016 From: royalts at gmail.com (Tobias Schmidt) Date: Thu, 16 Jun 2016 19:20:32 +0000 Subject: [tex-live] =?utf-8?q?MacTeX_=E2=80=9CThe_installation_failed?= =?utf-8?b?4oCd?= Message-ID: Hi all, I'm trying to install the current MacTeX distribution (mactex-20160603.pkg) but am getting an "The installation failed" error. I downloaded via Safari and checked the md5 checksum and so am sure that it's not just a corrupt download. Here's the complete install.log: Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: LSExceptions [0x7fa471e18aa0] loaded Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: @(#)PROGRAM:Install PROJECT:Install-1000 Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: @(#)PROGRAM:Installer PROJECT:Installer-853 Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Hardware: MacBookPro11,5 @ 2.80 GHz (x 8), 16384 MB RAM Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Running OS Build: Mac OS X 10.11.3 (15D21) Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: __CF_USER_TEXT_ENCODING=0x1F5:0x0:0x0 Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: TMPDIR=/var/folders/wl/klf5ry4n2vg4lfmtqk5tv5sr0000gn/T/ Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: HOME=/Users/tobiasschmidt Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: SHELL=/bin/bash Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: SSH_AUTH_SOCK=/private/tmp/com.apple.launchd.gWKsCfbdVh/Listeners Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: Apple_PubSub_Socket_Render=/private/tmp/com.apple.launchd.eTsiSMBp81/Render Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: PATH=/usr/bin:/bin:/usr/sbin:/sbin Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: LOGNAME=tobiasschmidt Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: XPC_SERVICE_NAME=com.apple.xpc.launchd.oneshot.0x10000062.Installer Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: COMMAND_MODE=unix2003 Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: USER=tobiasschmidt Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Env: XPC_FLAGS=0x0 Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: MacTeX-2016 Installation Log Jun 12 18:22:41 Tobiass-MacBook-Pro Installer[68769]: Opened from: /Users/tobiasschmidt/Downloads/mactex-20160603.pkg Jun 12 18:22:42 Tobiass-MacBook-Pro Installer[68769]: Product archive /Users/tobiasschmidt/Downloads/mactex-20160603.pkg trustLevel=202 Jun 12 18:22:46 Tobiass-MacBook-Pro Installer[68769]: InstallerStatusNotifications plugin loaded Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: Administrator authorization granted. Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: ================================================================================ Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: User picked Standard Install Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: Choices selected for installation: Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: Upgrade: "MacTeX-2016" Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: Upgrade: "TeXLive-2016" at path "//" Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: mactex-20160603.pkg#root.pkg : org.tug.mactex.texlive2016 : 1.0 Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: Upgrade: "GUI-Applications" at path "//" Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: mactex-20160603.pkg#applications.pkg : org.tug.mactex.gui2016 : 1.0 Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: Upgrade: "Ghostscript-9.19" at path "//" Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: mactex-20160603.pkg#local.pkg : org.tug.mactex.ghostscript9.19 : 1.0 Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: ================================================================================ Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: It took 0.00 seconds to summarize the package selections. Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: -[IFDInstallController(Private) _buildInstallPlanReturningError:]: location = file://localhost Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: -[IFDInstallController(Private) _buildInstallPlanReturningError:]: file://localhost/Users/tobiasschmidt/Downloads/mactex-20160603.pkg#root.pkg Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: -[IFDInstallController(Private) _buildInstallPlanReturningError:]: file://localhost/Users/tobiasschmidt/Downloads/mactex-20160603.pkg#applications.pkg Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: -[IFDInstallController(Private) _buildInstallPlanReturningError:]: file://localhost/Users/tobiasschmidt/Downloads/mactex-20160603.pkg#local.pkg Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: Set authorization level to root for session Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: Will use PK session Jun 12 18:22:51 Tobiass-MacBook-Pro Installer[68769]: Using authorization level of root for IFPKInstallElement Jun 12 18:22:52 Tobiass-MacBook-Pro Installer[68769]: LSExceptions [0x7fa471e18aa0] unloaded Jun 12 18:22:53 Tobiass-MacBook-Pro Installer[68769]: Starting installation: Jun 12 18:22:53 Tobiass-MacBook-Pro Installer[68769]: Configuring volume "Macintosh HD" Jun 12 18:22:53 Tobiass-MacBook-Pro Installer[68769]: Preparing disk for local booted install. Jun 12 18:22:53 Tobiass-MacBook-Pro Installer[68769]: Free space on "Macintosh HD": 26,96 GB (26955264000 bytes). Jun 12 18:22:53 Tobiass-MacBook-Pro Installer[68769]: Create temporary directory "/var/folders/wl/klf5ry4n2vg4lfmtqk5tv5sr0000gn/T//Install.68769gACYZl" Jun 12 18:22:53 Tobiass-MacBook-Pro Installer[68769]: IFPKInstallElement (3 packages) Jun 12 18:22:53 Tobiass-MacBook-Pro installd[341]: PackageKit: Adding client PKInstallDaemonClient pid=68769, uid=501 (/System/Library/CoreServices/Installer.app/Contents/MacOS/Installer) Jun 12 18:22:53 Tobiass-MacBook-Pro Installer[68769]: PackageKit: Enqueuing install with framework-specified quality of service (utility) Jun 12 18:22:53 Tobiass-MacBook-Pro installd[341]: PackageKit: ----- Begin install ----- Jun 12 18:22:53 Tobiass-MacBook-Pro installd[341]: PackageKit: request=PKInstallRequest <3 packages, destination=/> Jun 12 18:22:53 Tobiass-MacBook-Pro installd[341]: PackageKit: packages=( "PKLeopardPackage ", "PKLeopardPackage ", "PKLeopardPackage " ) Jun 12 18:22:59 Tobiass-MacBook-Pro installd[341]: PackageKit: Extracting file://localhost/Users/tobiasschmidt/Downloads/mactex-20160603.pkg#root.pkg (destination=/var/folders/zz/zyxvpxvq6csfxvn_n0000000000000/C/PKInstallSandboxManager/8BD855B4-CAA1-4A63-8137-72E3A47D8F49.activeSandbox/Root, uid=0) Jun 12 18:23:40 Tobiass-MacBook-Pro installd[341]: PackageKit: Extracting file://localhost/Users/tobiasschmidt/Downloads/mactex-20160603.pkg#applications.pkg (destination=/var/folders/zz/zyxvpxvq6csfxvn_n0000000000000/C/PKInstallSandboxManager/8BD855B4-CAA1-4A63-8137-72E3A47D8F49.activeSandbox/Root/Applications, uid=0) Jun 12 18:23:42 Tobiass-MacBook-Pro installd[341]: PackageKit: Extracting file://localhost/Users/tobiasschmidt/Downloads/mactex-20160603.pkg#local.pkg (destination=/var/folders/zz/zyxvpxvq6csfxvn_n0000000000000/C/PKInstallSandboxManager/8BD855B4-CAA1-4A63-8137-72E3A47D8F49.activeSandbox/Root/usr/local, uid=0) Jun 12 18:23:43 Tobiass-MacBook-Pro installd[341]: PackageKit: prevent user idle system sleep Jun 12 18:23:43 Tobiass-MacBook-Pro installd[341]: PackageKit: suspending backupd Jun 12 18:23:43 Tobiass-MacBook-Pro install_monitor[68901]: Temporarily excluding: /Applications, /Library, /System, /bin, /private, /sbin, /usr Jun 12 18:23:43 Tobiass-MacBook-Pro install_monitor[68901]: Re-included: /Applications, /Library, /System, /bin, /private, /sbin, /usr Jun 12 18:23:43 Tobiass-MacBook-Pro installd[341]: PackageKit: releasing backupd Jun 12 18:23:43 Tobiass-MacBook-Pro installd[341]: PackageKit: allow user idle system sleep Jun 12 18:23:43 Tobiass-MacBook-Pro installd[341]: PackageKit: Install Failed: Error Domain=NSPOSIXErrorDomain Code=2 "No such file or directory" UserInfo={NSFilePath=/private/tmp/PKInstallSandbox.XXXXXX} { NSFilePath = "/private/tmp/PKInstallSandbox.XXXXXX"; } Jun 12 18:23:43 Tobiass-MacBook-Pro installd[341]: PackageKit: Running idle tasks Jun 12 18:23:43 Tobiass-MacBook-Pro Installer[68769]: install:didFailWithError:Error Domain=NSPOSIXErrorDomain Code=2 "No such file or directory" UserInfo={NSFilePath=/private/tmp/PKInstallSandbox.XXXXXX} Jun 12 18:23:43 Tobiass-MacBook-Pro installd[341]: PackageKit: Removing client PKInstallDaemonClient pid=68769, uid=501 (/System/Library/CoreServices/Installer.app/Contents/MacOS/Installer) Jun 12 18:23:44 Tobiass-MacBook-Pro Installer[68769]: Install failed: The Installer encountered an error that caused the installation to fail. Contact the software manufacturer for assistance. Jun 12 18:23:44 Tobiass-MacBook-Pro Installer[68769]: IFDInstallController 74036930 state = 8 Jun 12 18:23:44 Tobiass-MacBook-Pro Installer[68769]: Displaying 'Install Failed' UI. Jun 12 18:23:44 Tobiass-MacBook-Pro Installer[68769]: 'Install Failed' UI displayed message:'The Installer encountered an error that caused the installation to fail. Contact the software manufacturer for assistance.'. Jun 12 18:23:53 Tobiass-MacBook-Pro installd[341]: PackageKit: Done with sandbox removals Jun 12 18:23:53 Tobiass-MacBook-Pro installd[341]: PackageKit: Sandbox /var/folders/zz/zyxvpxvq6csfxvn_n0000000000000/C/PKInstallSandboxManager/563D4C1A-7052-4D64-9A27-42AFEBA305D2.sandbox: relevance 72 Any idea where I might be going wrong? Best, Tobias -------------- next part -------------- An HTML attachment was scrubbed... URL: From kakuto at fuk.kindai.ac.jp Fri Jun 17 05:18:28 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Fri, 17 Jun 2016 12:18:28 +0900 Subject: [tex-live] Lollipop is broken in TeXlive 2016 In-Reply-To: References: Message-ID: <553EA6C97769458DA05A3D0650CEB477@CJ3001517A> Dear Vafa, > lollipop is broken in TeXLive 2016. Create a simple lollipop document > \Start > > This is a test. > > \Stop Here lollipop test.tex works fine with pdftex: This is pdfTeX, Version 3.14159265-2.6-1.40.17 (TeX Live 2016/W32TeX) (preloaded format=lollipop) restricted \write18 enabled. entering extended mode (./test.tex >>>> This is Lollipop TeX (v1.07, 21 May 2016) <<<< (./test.aux) [1.1] (./test.aux) ) Output written on test.dvi (1 page, 216 bytes). Transcript written on test.log. Generated test.dvi looks fine. I think that pdftex is not incompatible with the Knuth TeX. But Karl may change to use tex instead of pdftex if you want. Best, Akira From languasc at gmail.com Fri Jun 17 06:26:36 2016 From: languasc at gmail.com (Alessandro Languasco) Date: Fri, 17 Jun 2016 06:26:36 +0200 Subject: [tex-live] Lollipop is broken in TeXlive 2016 In-Reply-To: References: Message-ID: <88B12276-2625-47D2-B7CD-8196826F52E4@gmail.com> I see the same problem Vafa Khalighi wrote both on my mac with mactex-2016 and my linux box with texlive 2016. Bests, Alessandro > On 16 Jun 2016, at 03:31, Vafa Khalighi wrote: > > Hi Karl > > lollipop is broken in TeXLive 2016. Create a simple lollipop document > > \Start > This is a test. > \Stop > > and process it with `lollipop test.tex' and this is the error you get > > %%%%%%%%%% > This is pdfTeX, Version 3.14159265-2.6-1.40.17 (TeX Live 2016) (preloaded format=lollipop) > restricted \write18 enabled. > > kpathsea: Running mktexfmt lollipop.fmt > mktexfmt: mktexfmt is using the following fmtutil.cnf files (in precedence order): > mktexfmt: /usr/local/texlive/2016/texmf-dist/web2c/fmtutil.cnf > mktexfmt: mktexfmt is using the following fmtutil.cnf file for writing changes: > mktexfmt: /Users/vafa/Library/texlive/2016/texmf-config/web2c/fmtutil.cnf > mktexfmt [INFO]: writing formats under /Users/vafa/Library/texlive/2016/texmf-var/web2c > mktexfmt [INFO]: --- remaking lollipop with tex > mktexfmt: running `tex -ini -jobname=lollipop -progname=lollipop lollipop.ini' ... > This is TeXk, Version 3.14159265 (TeX Live 2016) (INITEX) > restricted \write18 enabled. > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop.ini > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop.tex > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-plain.tex > (/usr/local/texlive/2016/texmf-dist/tex/generic/hyphen/hyphen.tex)) > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-tools.tex) > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-define.tex) > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-fonts.tex) > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-fontdefs.tex) > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-text.tex) > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-document.tex) > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-heading.tex) > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-output.tex) > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-float.tex) > (/usr/local/texlive/2016/texmf-dist/tex/lollipop/lollipop-lists.tex) > ) ) > Beginning to dump on file lollipop.fmt > (preloaded format=lollipop 2016.6.16) > 3422 strings of total length 48580 > 33354 memory locations dumped; current usage is 158&33023 > 2325 multiletter control sequences > \font\nullfont=nullfont > \font\exx=cmex10 > \font\ttx=cmtt10 > \font\ComputerModernromanx=cmr10 > \font\ttix=cmtt9 > \font\ttxi=cmtt10 at 10.95pt > \font\ttxii=cmtt10 at 12.0pt > \font\ttxiv=cmtt10 at 14.4pt > \font\ttxvii=cmtt10 at 17.28pt > \font\ttxx=cmtt10 at 20.73999pt > \font\ComputerModernboldx=cmbx10 > \font\ComputerModernslantx=cmsl10 > \font\ComputerModernitalicx=cmti10 > 2706 words of font info for 12 preloaded fonts > 19 hyphenation exceptions > Hyphenation trie of length 6075 has 181 ops out of 35111 > 181 for language 0 > No pages of output. > Transcript written on lollipop.log. > mktexfmt [INFO]: /Users/vafa/Library/texlive/2016/texmf-var/web2c/tex/lollipop.fmt installed. > mktexfmt [INFO]: Disabled formats: 6 > mktexfmt [INFO]: Successfully rebuilt formats: 1 > mktexfmt [INFO]: Not selected formats: 43 > mktexfmt [INFO]: Total formats: 50 > mktexfmt [INFO]: exiting with status 0 > ---! /Users/vafa/Library/texlive/2016/texmf-var/web2c/tex/lollipop.fmt was written by tex > (Fatal format file error; I'm stymied) > > %%%%%%%%%%%%%%%%%%%%%% > > > This is because, lollipop format should be based on Knuth tex and not pdftex. Indeed, when I look at TeXLive repository, I can confirm that only netbsd platforms have lollipop based on Knuth TeX (link tex) and all other platforms link pdftex. I believe I sent you an email and explained this clearly to you when you asked me to delete xelollipop. > > lollipop is the only tex format that I use myself and this breaks all my documents. > > Could you please fix this? > > Thanks -------------- next part -------------- An HTML attachment was scrubbed... URL: From kakuto at fuk.kindai.ac.jp Fri Jun 17 07:12:17 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Fri, 17 Jun 2016 14:12:17 +0900 Subject: [tex-live] Lollipop is broken in TeXlive 2016 In-Reply-To: <88B12276-2625-47D2-B7CD-8196826F52E4@gmail.com> References: <88B12276-2625-47D2-B7CD-8196826F52E4@gmail.com> Message-ID: <955ABE217C2342FBAE84549D737DDD3F@CJ3001517A> Dear Vafa, Alessandro, > I see the same problem Vafa Khalighi wrote both on my mac with mactex-2016 > and my linux box with texlive 2016. I have found the reason. Karl already had changed fmtutil.cnf to use tex instead of pdftex in fmtutil.cnf: # from lollipop: lollipop tex - lollipop.ini However he forgot to change symlink for lollipop, which is still linked to pdftex. Please wait a day. Thanks, Akira From zdenek.wagner at gmail.com Fri Jun 17 10:09:18 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Fri, 17 Jun 2016 10:09:18 +0200 Subject: [tex-live] Letter template In-Reply-To: References: Message-ID: Hi, this question has nothing to do with TeX Live maintenance so this should be posted somewhere else. Anyway, I am afraid that a good general answer does not exist. Although various internationalization projects exist so that you can prepare a general template containing macros as \from, \to, \cc etc., the universe is not that grey. First, there are different conventions in different countries one of them being the position of the signature. Generally it can be said that for A4 or letter size paper the font size should be 11pt or 12pt, margins should be around 1in (the old Czech standard required 25mm which is almost the same dimension). Setting of \parindent and \parskip is also a matter of convention. What is the most important is "corporate identity". I was asked to prepare letter templates for several companies. The layout was designed by a graphuical design studio. The people from the studio do not know LaTeX and the companies requested LaTeX so I have implemented it for them. In short, if you wish a professional design, ask a graphical designer. If you think that you can do it yourself. look at letters that you receive from others and try to do it in a similar way. Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz 2016-06-16 17:22 GMT+02:00 Kai Hendry : > Hi there, > > I created a Docker image with texlive to help generate a letter from a > Web form in the very occasional times I need to. > > However I recently got a complaint from my wife that the template the > font is too small and the borders are too wide and the signature is > odd for being offset to the right. > > This is the current template I am using and it must be 10 years old. > Can anyone offer offer pointers to a better (more modern) template? > https://github.com/kaihendry/letterly/blob/master/www/pdf/letter.template > > \documentclass[british,a4paper]{letter} > \usepackage{fontspec} > \defaultfontfeatures{Mapping=tex-text} > \usepackage{xunicode} > \usepackage{babel} > \begin{document} > > > Many thanks! > -------------- next part -------------- An HTML attachment was scrubbed... URL: From zdenek.wagner at gmail.com Fri Jun 17 10:13:37 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Fri, 17 Jun 2016 10:13:37 +0200 Subject: [tex-live] texcount on Windows fail because of Perl file missing from TeX Live distribution In-Reply-To: <306071466104496@web16h.yandex.ru> References: <306071466104496@web16h.yandex.ru> Message-ID: This is strange, cannot it be a problem of permissions? The error message lists the @INC settings and it is clear that you have it in the right directory. Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz 2016-06-16 21:14 GMT+02:00 Leonid Sinev : > Greetings, > > This problem already discussed at > https://www.tug.org/pipermail/tex-live/2015-August/037293.html > but nothing changed in TeXLive 2016. > > Nothing changed after > http://tex.stackexchange.com/a/145766 advice > of Win32::Console::ANSI from CPAN > I have copied file from this package to > C:\texlive\2016\tlpkg\tlperl\site\lib\Win32\Console\ANSI.pm > and got this: > > Can't locate loadable object for module Win32::Console::ANSI in @INC (@INC > conta ins: C:/texlive/2016/tlpkg/tlperl/lib > c:/texlive/2016/tlpkg/tlperl/site/lib c:/t exlive/2016/tlpkg/tlperl/lib .) > at c:\texlive\2016\texmf-dist\scripts\texcount\t excount.pl line 11. > Compilation failed in require at > c:\texlive\2016\texmf-dist\scripts\texcount\tex count.pl line 11. > BEGIN failed--compilation aborted at > c:\texlive\2016\texmf-dist\scripts\texcount \texcount.pl line 14. > C:\texlive\2016\bin\win32\runscript.tlu:679: command failed with exit code > 2: perl.exe c:\texlive\2016\texmf-dist\scripts\texcount\texcount.pl > > Can there be a patch or update for TeXLive for windows that can make > windows users use Perl scripts like texcount without need of additional > perl installation or messing with additional packaging for tlperl? > > -- > Best regards, > Leonid > -------------- next part -------------- An HTML attachment was scrubbed... URL: From preining at logic.at Fri Jun 17 12:16:09 2016 From: preining at logic.at (Norbert Preining) Date: Fri, 17 Jun 2016 19:16:09 +0900 Subject: [tex-live] =?utf-8?q?MacTeX_=E2=80=9CThe_installation_failed?= =?utf-8?b?4oCd?= In-Reply-To: References: Message-ID: <20160617101609.GD21093@logic.at> > I'm trying to install the current MacTeX distribution (mactex-20160603.pkg) Can you please contact the mactex mailing list - it is the most appropriate one. None of the devs here are working on the MacTeX personally (AFAIR). Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From preining at logic.at Fri Jun 17 12:23:45 2016 From: preining at logic.at (Norbert Preining) Date: Fri, 17 Jun 2016 19:23:45 +0900 Subject: [tex-live] xcolor broken with pdflatex In-Reply-To: References: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> Message-ID: <20160617102345.GE21093@logic.at> Thanks David, > I just updated latex-graphics, luatex-def and pdftex-def on ctan (for updated for tomorrow. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From preining at logic.at Fri Jun 17 13:43:07 2016 From: preining at logic.at (Norbert Preining) Date: Fri, 17 Jun 2016 20:43:07 +0900 Subject: [tex-live] Asymptote 3D and gcc-6.1 ? In-Reply-To: <20160617000419.GA3140@milliways.localdomain> References: <20160617000419.GA3140@milliways.localdomain> Message-ID: <20160617114307.GA16040@logic.at> On Fri, 17 Jun 2016, Ken Moffat wrote: > development system using gcc-6.1 it fails to create 3D drawings and Do you mean that when you recompile with gcc-6 you get problems? Or are you using the shipped binaries on a system with gcc-6? (libstdc++ and some other libs would change I guess). asy is not the only problem, on Debian I see other failures with gcc-6, and some of them are for sure compiler bugs. WHether this is the case for asy/3d, I don't know. So all in all - yes - I expect lots of problems with gcc6, at least in the beginning. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From josemaria.alcaide at ehu.eus Fri Jun 17 15:51:14 2016 From: josemaria.alcaide at ehu.eus (=?utf-8?Q?Jos=C3=A9_Mar=C3=ADa_Alcaide?=) Date: Fri, 17 Jun 2016 15:51:14 +0200 Subject: [tex-live] pdftex.def 0.06f breaks xcolor package Message-ID: Hello. I have TeX Live 2016 installed (MacTeX 2016). After updating TL using tlmgr, I found that loading the xcolor package causes this error with pdftex: /usr/local/texlive/2016/texmf-dist/tex/latex/pdftex-def/pdftex.def:444: Undefined control sequence. l.444 \ifGin at setpagesize And, indeed, pdftex.def was among the files updated (from v0.06d to v0.06f). Restoring pdftex v0.06d makes that error to go away. This problem also affects other packages which use xcolor (such as the minted package). -- Jos? Mar?a Alcaide Prof. titular de Lenguajes y sistemas inform?ticos Universidad del Pa?s Vasco Dpto. de Electricidad y Electr?nica Facultad de Ciencia y Tecnolog?a 48940 Leioa (Spain) From josemaria.alcaide at ehu.eus Fri Jun 17 15:59:37 2016 From: josemaria.alcaide at ehu.eus (=?utf-8?Q?Jos=C3=A9_Mar=C3=ADa_Alcaide?=) Date: Fri, 17 Jun 2016 15:59:37 +0200 Subject: [tex-live] pdftex.def 0.06f breaks xcolor package In-Reply-To: References: Message-ID: > El 17 jun 2016, a las 15:51, Jos? Mar?a Alcaide escribi?: > > Hello. > > I have TeX Live 2016 installed (MacTeX 2016). After updating TL using tlmgr, I found that loading the xcolor package causes this error with pdftex: > > /usr/local/texlive/2016/texmf-dist/tex/latex/pdftex-def/pdftex.def:444: Undefined control sequence. > l.444 \ifGin at setpagesize > > And, indeed, pdftex.def was among the files updated (from v0.06d to v0.06f). Restoring pdftex v0.06d makes that error to go away. > > This problem also affects other packages which use xcolor (such as the minted package). > Ouch. I have just found another thread from a few hours ago reporting this same problem. Sorry for the duplicate! -- Jos? Mar?a Alcaide Prof. titular de Lenguajes y sistemas inform?ticos Universidad del Pa?s Vasco Dpto. de Electricidad y Electr?nica Facultad de Ciencia y Tecnolog?a 48940 Leioa (Spain) From preining at logic.at Fri Jun 17 16:41:35 2016 From: preining at logic.at (Norbert Preining) Date: Fri, 17 Jun 2016 23:41:35 +0900 Subject: [tex-live] pdftex.def 0.06f breaks xcolor package In-Reply-To: References: Message-ID: <20160617144135.GA14671@logic.at> > /usr/local/texlive/2016/texmf-dist/tex/latex/pdftex-def/pdftex.def:444: Undefined control sequence. > l.444 \ifGin at setpagesize Should be fixed with tomorrow's update. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From herbs at wideopenwest.com Fri Jun 17 17:04:04 2016 From: herbs at wideopenwest.com (Herbert Schulz) Date: Fri, 17 Jun 2016 10:04:04 -0500 Subject: [tex-live] pdftex.def 0.06f breaks xcolor package In-Reply-To: <20160617144135.GA14671@logic.at> References: <20160617144135.GA14671@logic.at> Message-ID: <8C707D24-8950-40EE-997A-AD7A83CEE915@wideopenwest.com> > On Jun 17, 2016, at 9:41 AM, Norbert Preining wrote: > >> /usr/local/texlive/2016/texmf-dist/tex/latex/pdftex-def/pdftex.def:444: Undefined control sequence. >> l.444 \ifGin at setpagesize > > Should be fixed with tomorrow's update. > > Norbert Howdy, In the mean time, assuming you're using the graphicx package also, load graohicx before xcolor. Good Luck, Herb Schulz (herbs at wideopenwest dot com) From zarniwhoop at ntlworld.com Fri Jun 17 17:20:14 2016 From: zarniwhoop at ntlworld.com (Ken Moffat) Date: Fri, 17 Jun 2016 16:20:14 +0100 Subject: [tex-live] Asymptote 3D and gcc-6.1 ? In-Reply-To: <20160617114307.GA16040@logic.at> References: <20160617000419.GA3140@milliways.localdomain> <20160617114307.GA16040@logic.at> Message-ID: <20160617152014.GA1725@milliways.localdomain> On Fri, Jun 17, 2016 at 08:43:07PM +0900, Norbert Preining wrote: > On Fri, 17 Jun 2016, Ken Moffat wrote: > > development system using gcc-6.1 it fails to create 3D drawings and > > Do you mean that when you recompile with gcc-6 you get problems? > Or are you using the shipped binaries on a system with gcc-6? > (libstdc++ and some other libs would change I guess). > This is for the shipped binary - I have not got around to building everything from source. Hmm, yes, a different libstdc++. > asy is not the only problem, on Debian I see other failures with gcc-6, > and some of them are for sure compiler bugs. WHether this is the case > for asy/3d, I don't know. > For the moment, this is the binary - the same one that works on a gcc-5.3.1 system. > So all in all - yes - I expect lots of problems with gcc6, at least > in the beginning. > > Norbert > Yes, I/we (BLFS) have had fun in several of packages. I've only got back to this now that I've got firefox-47 stable. Thanks for those comments. ?en -- War is Peace Freedom is Slavery Ignorance is Strength From leonid.sinev at yandex.ru Fri Jun 17 18:19:46 2016 From: leonid.sinev at yandex.ru (Leonid Sinev) Date: Fri, 17 Jun 2016 19:19:46 +0300 Subject: [tex-live] texcount on Windows fail because of Perl file missing from TeX Live distribution In-Reply-To: References: <306071466104496@web16h.yandex.ru> Message-ID: <462441466180386@web3m.yandex.ru> An HTML attachment was scrubbed... URL: From jfbu at free.fr Fri Jun 17 19:05:51 2016 From: jfbu at free.fr (jfbu) Date: Fri, 17 Jun 2016 19:05:51 +0200 Subject: [tex-live] xcolor broken with pdflatex In-Reply-To: <20160617102345.GE21093@logic.at> References: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> <20160617102345.GE21093@logic.at> Message-ID: <42490C09-B19D-4E6D-A76B-8D02015B7F7D@free.fr> Hi, should I worry about these tlmgr error messages below ? (I don't as the first compilation I tried which had failed before now worked) Best, Jean-Fran?ois $ tlmgr restore graphics Available backups for graphics: 41015 (2016-06-17 12:13) $ tlmgr restore graphics 41015 Do you really want to restore graphics to revision 41015 (y/N): y Restoring graphics, 41015 from /usr/local/texlive/2016/tlpkg/backups/graphics.r41015.tar.xz Can't locate object method "_install_package" via package "TeXLive::TLPDB" at /Library/TeX/texbin/tlmgr line 1768, line 1. $ tlmgr restore luatex-def Available backups for luatex-def: 41398 (2016-06-17 12:13) $ tlmgr restore luatex-def 41398 Do you really want to restore luatex-def to revision 41398 (y/N): y Restoring luatex-def, 41398 from /usr/local/texlive/2016/tlpkg/backups/luatex-def.r41398.tar.xz Can't locate object method "_install_package" via package "TeXLive::TLPDB" at /Library/TeX/texbin/tlmgr line 1768, line 1. $ tlmgr restore pdftex-def Available backups for pdftex-def: 22653 (2016-06-17 12:13) $ tlmgr restore pdftex-def 22653 Do you really want to restore pdftex-def to revision 22653 (y/N): y Restoring pdftex-def, 22653 from /usr/local/texlive/2016/tlpkg/backups/pdftex-def.r22653.tar.xz Can't locate object method "_install_package" via package "TeXLive::TLPDB" at /Library/TeX/texbin/tlmgr line 1768, line 1. [19:03:28] jfb at mba-de-jfb:~ ~$ Le 17 juin 2016 ? 12:23, Norbert Preining a ?crit : > Thanks David, > >> I just updated latex-graphics, luatex-def and pdftex-def on ctan (for > > updated for tomorrow. > > Norbert > > -- > PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info > GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From jfbu at free.fr Fri Jun 17 20:05:25 2016 From: jfbu at free.fr (jfbu) Date: Fri, 17 Jun 2016 20:05:25 +0200 Subject: [tex-live] xcolor broken with pdflatex In-Reply-To: <42490C09-B19D-4E6D-A76B-8D02015B7F7D@free.fr> References: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> <20160617102345.GE21093@logic.at> <42490C09-B19D-4E6D-A76B-8D02015B7F7D@free.fr> Message-ID: Le 17 juin 2016 ? 19:05, jfbu a ?crit : > Hi, > > should I worry about these tlmgr error messages below ? apparently I should because there simply isn't anymore a pdftex.def on my installation... (sorry for not complete report but I am doing other things with latex compilation as a side thing once in a while and I currently can not investigated too much the latex problems, reverted to TL2015 for my immediate needs) after a compilation failed and kpsewhich pdftex.def reported nothing quick try: $ ls /usr/local/texlive/2016/texmf-dist/tex/latex/graphics ls: /usr/local/texlive/2016/texmf-dist/tex/latex/graphics: No such file or directory $ ls /usr/local/texlive/2015/texmf-dist/tex/latex/graphics color.sty dvipsone.def graphics.sty pctex32.def tcidvi.def dvipdf.def dviwin.def graphicx.sty pctexhp.def trig.sty dvips.def emtex.def keyval.sty pctexps.def truetex.def dvipsnam.def epsfig.sty lscape.sty pctexwin.def but in backups I see (extract) -rw-r--r-- 1 xxxxxx yyyyyy 13592 Jun 17 12:13 pdftex-def.r22653.tar.xz -rw-r--r-- 1 xxxxxx yyyyyy 175036 Jun 17 12:13 luatex85.r41395.tar.xz -rw-r--r-- 1 xxxxxx yyyyyy 8980 Jun 17 12:13 luatex-def.r41398.tar.xz -rw-r--r-- 1 xxxxxx yyyyyy 12886592 Jun 17 12:13 latex.r40218.tar.xz -rw-r--r-- 1 xxxxxx yyyyyy 2400 Jun 17 12:13 graphics-cfg.r40269.tar.xz -rw-r--r-- 1 xxxxxx yyyyyy 1685216 Jun 17 12:13 graphics.r41015.tar.xz -rw-r--r-- 1 xxxxxx yyyyyy 36972 Jun 17 12:13 epstopdf.r41388.tar.xz Jean-Fran?ois > (I don't as the first compilation I tried which had failed before now worked) > > Best, > > Jean-Fran?ois > > $ tlmgr restore graphics > Available backups for graphics: 41015 (2016-06-17 12:13) > > $ tlmgr restore graphics 41015 > Do you really want to restore graphics to revision 41015 (y/N): y > Restoring graphics, 41015 from /usr/local/texlive/2016/tlpkg/backups/graphics.r41015.tar.xz > Can't locate object method "_install_package" via package "TeXLive::TLPDB" at /Library/TeX/texbin/tlmgr line 1768, line 1. > > $ tlmgr restore luatex-def > Available backups for luatex-def: 41398 (2016-06-17 12:13) > > $ tlmgr restore luatex-def 41398 > Do you really want to restore luatex-def to revision 41398 (y/N): y > Restoring luatex-def, 41398 from /usr/local/texlive/2016/tlpkg/backups/luatex-def.r41398.tar.xz > Can't locate object method "_install_package" via package "TeXLive::TLPDB" at /Library/TeX/texbin/tlmgr line 1768, line 1. > > $ tlmgr restore pdftex-def > Available backups for pdftex-def: 22653 (2016-06-17 12:13) > > $ tlmgr restore pdftex-def 22653 > Do you really want to restore pdftex-def to revision 22653 (y/N): y > Restoring pdftex-def, 22653 from /usr/local/texlive/2016/tlpkg/backups/pdftex-def.r22653.tar.xz > Can't locate object method "_install_package" via package "TeXLive::TLPDB" at /Library/TeX/texbin/tlmgr line 1768, line 1. > > [19:03:28] jfb at mba-de-jfb:~ > ~$ > > > > > Le 17 juin 2016 ? 12:23, Norbert Preining a ?crit : > >> Thanks David, >> >>> I just updated latex-graphics, luatex-def and pdftex-def on ctan (for >> >> updated for tomorrow. >> >> Norbert >> >> -- >> PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info >> GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 > From karl at freefriends.org Sat Jun 18 00:43:35 2016 From: karl at freefriends.org (Karl Berry) Date: Fri, 17 Jun 2016 22:43:35 +0000 Subject: [tex-live] texcount on Windows fail because of Perl file missing from TeX Live distribution In-Reply-To: <462441466180386@web3m.yandex.ru> Message-ID: <201606172243.u5HMhZsi025686@freefriends.org> It has never been a goal for the w32 Perl distributed with TeX Live to have every Perl module possibly needed by every package. Rather, its purpose is to allow Windows users to run the TL infrastructure programs written in Perl without installing a Perl themselves. Thus, if you want to be able to run texcount, you need to install another Perl with the extra packages. Or install the extra packages somewhere and arrange for our Perl to see it, if that's feasible on Windows. Maybe someone else here can help more specifically ... -k From preining at logic.at Sat Jun 18 02:48:54 2016 From: preining at logic.at (Norbert Preining) Date: Sat, 18 Jun 2016 09:48:54 +0900 Subject: [tex-live] xcolor broken with pdflatex In-Reply-To: <42490C09-B19D-4E6D-A76B-8D02015B7F7D@free.fr> References: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> <20160617102345.GE21093@logic.at> <42490C09-B19D-4E6D-A76B-8D02015B7F7D@free.fr> Message-ID: <20160618004854.GC14671@logic.at> On Fri, 17 Jun 2016, jfbu wrote: > Can't locate object method "_install_package" via package "TeXLive::TLPDB" at /Library/TeX/texbin/tlmgr line 1768, line 1. Bad, fixed for today, need to use tlcritical for now. Thanks for finding this Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From jfbu at free.fr Sat Jun 18 07:36:33 2016 From: jfbu at free.fr (jfbu) Date: Sat, 18 Jun 2016 07:36:33 +0200 Subject: [tex-live] xcolor broken with pdflatex In-Reply-To: <20160618004854.GC14671@logic.at> References: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> <20160617102345.GE21093@logic.at> <42490C09-B19D-4E6D-A76B-8D02015B7F7D@free.fr> <20160618004854.GC14671@logic.at> Message-ID: <5498EC78-CE7C-4C5B-99A1-89EF39E500DC@free.fr> Le 18 juin 2016 ? 02:48, Norbert Preining a ?crit : > On Fri, 17 Jun 2016, jfbu wrote: >> Can't locate object method "_install_package" via package "TeXLive::TLPDB" at /Library/TeX/texbin/tlmgr line 1768, line 1. > > > Bad, fixed for today, need to use tlcritical for now. > > Thanks for finding this > just to confirm that after updating tlmgr itself from tlcritical the restoring of graphics, luatex-def, and pdftex-def from backuped versions then worked, Thanks ! the new versions do not seem to have made it yet to mirrors, I get $ tlmgr update --list tlmgr: package repository http://ftp.oleane.net/pub/CTAN/systems/texlive/tlnet (verified) tlmgr: saving backups to /usr/local/texlive/2016/tlpkg/backups keep: texlive.infra : local: 41476, source: 41280 update: graphics [1924k]: local: 41015, source: 41447 update: luatex-def [10k]: local: 41398, source: 41450 update: pdftex-def [15k]: local: 22653, source: 41449 and I think graphics-41447 is the broken one. Jean-Fran?ois From ziegenhagen at gmail.com Sat Jun 18 07:37:59 2016 From: ziegenhagen at gmail.com (Uwe Ziegenhagen) Date: Sat, 18 Jun 2016 07:37:59 +0200 Subject: [tex-live] General test suite for TeX-Live Message-ID: Hi everyone, I think it might be a good idea to create a general test suite for TeX Live that we could use during the preparations of each release. Interactions between different components are likely to grow so a set of tests could help us determine if the release is fine or not. What do you think? Uwe PS: If you agree that it would be a good thing I suggest to create a Subversion repository which we could fill either on per-tester but maybe better on per-topic basis. -- Dr. Uwe Ziegenhagen -------------- next part -------------- An HTML attachment was scrubbed... URL: From preining at logic.at Sat Jun 18 07:38:41 2016 From: preining at logic.at (Norbert Preining) Date: Sat, 18 Jun 2016 14:38:41 +0900 Subject: [tex-live] xcolor broken with pdflatex In-Reply-To: <5498EC78-CE7C-4C5B-99A1-89EF39E500DC@free.fr> References: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> <20160617102345.GE21093@logic.at> <42490C09-B19D-4E6D-A76B-8D02015B7F7D@free.fr> <20160618004854.GC14671@logic.at> <5498EC78-CE7C-4C5B-99A1-89EF39E500DC@free.fr> Message-ID: <2C1A255A-9732-4430-8534-A0D93D141172@logic.at> The current versions, including the graphics one, should be the fixed one. Norbert On June 18, 2016 2:36:33 PM GMT+09:00, jfbu wrote: > >Le 18 juin 2016 ? 02:48, Norbert Preining a ?crit : > >> On Fri, 17 Jun 2016, jfbu wrote: >>> Can't locate object method "_install_package" via package >"TeXLive::TLPDB" at /Library/TeX/texbin/tlmgr line 1768, line >1. >> >> >> Bad, fixed for today, need to use tlcritical for now. >> >> Thanks for finding this >> > >just to confirm that after updating tlmgr itself from tlcritical >the restoring of graphics, luatex-def, and pdftex-def from backuped >versions then worked, Thanks ! > >the new versions do not seem to have made it yet to mirrors, I get > >$ tlmgr update --list >tlmgr: package repository >http://ftp.oleane.net/pub/CTAN/systems/texlive/tlnet (verified) >tlmgr: saving backups to /usr/local/texlive/2016/tlpkg/backups >keep: texlive.infra : local: 41476, source: 41280 >update: graphics [1924k]: local: 41015, source: 41447 >update: luatex-def [10k]: local: 41398, source: 41450 >update: pdftex-def [15k]: local: 22653, source: 41449 > >and I think graphics-41447 is the broken one. > >Jean-Fran?ois -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From jfbu at free.fr Sat Jun 18 07:43:57 2016 From: jfbu at free.fr (jfbu) Date: Sat, 18 Jun 2016 07:43:57 +0200 Subject: [tex-live] xcolor broken with pdflatex In-Reply-To: <2C1A255A-9732-4430-8534-A0D93D141172@logic.at> References: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> <20160617102345.GE21093@logic.at> <42490C09-B19D-4E6D-A76B-8D02015B7F7D@free.fr> <20160618004854.GC14671@logic.at> <5498EC78-CE7C-4C5B-99A1-89EF39E500DC@free.fr> <2C1A255A-9732-4430-8534-A0D93D141172@logic.at> Message-ID: Le 18 juin 2016 ? 07:38, Norbert Preining a ?crit : > The current versions, including the graphics one, should be the fixed one. however if I reinstall the current versions, I get a failed compilation (/usr/local/texlive/2016/texmf-dist/tex/latex/xcolor/xcolor.sty (/usr/local/texlive/2016/texmf-dist/tex/latex/graphics-cfg/color.cfg) (/usr/local/texlive/2016/texmf-dist/tex/latex/pdftex-def/pdftex.def (/usr/local/texlive/2016/texmf-dist/tex/generic/oberdiek/infwarerr.sty) (/usr/local/texlive/2016/texmf-dist/tex/generic/oberdiek/ltxcmds.sty) ! Undefined control sequence. l.444 \ifGin at setpagesize ? on a project of mine, which seems to be the earlier reported problem (this project uses a style file which loads xcolor first then graphicx) the update looked like: $ tlmgr update --all tlmgr: package repository http://ftp.oleane.net/pub/CTAN/systems/texlive/tlnet (verified) tlmgr: saving backups to /usr/local/texlive/2016/tlpkg/backups [1/4, ??:??/??:??] update: graphics [1924k] (41015 -> 41447) ... done [2/4, 00:04/00:04] update: graphics-cfg [2k] (40269 -> 41448) ... done [3/4, 00:05/00:05] update: luatex-def [10k] (41398 -> 41450) ... done [4/4, 00:06/00:06] update: pdftex-def [15k] (22653 -> 41449) ... done tlmgr: package log updated: /usr/local/texlive/2016/texmf-var/web2c/tlmgr.log running mktexlsr ... (in between I had restore graphics-cfg, this is why it gets updated here again) best, JF PS: I have set autobackup to 2 in the meantime ;-) > > Norbert > > On June 18, 2016 2:36:33 PM GMT+09:00, jfbu wrote: >> >> Le 18 juin 2016 ? 02:48, Norbert Preining a ?crit : >> >>> On Fri, 17 Jun 2016, jfbu wrote: >>>> Can't locate object method "_install_package" via package >> "TeXLive::TLPDB" at /Library/TeX/texbin/tlmgr line 1768, line >> 1. >>> >>> >>> Bad, fixed for today, need to use tlcritical for now. >>> >>> Thanks for finding this >>> >> >> just to confirm that after updating tlmgr itself from tlcritical >> the restoring of graphics, luatex-def, and pdftex-def from backuped >> versions then worked, Thanks ! >> >> the new versions do not seem to have made it yet to mirrors, I get >> >> $ tlmgr update --list >> tlmgr: package repository >> http://ftp.oleane.net/pub/CTAN/systems/texlive/tlnet (verified) >> tlmgr: saving backups to /usr/local/texlive/2016/tlpkg/backups >> keep: texlive.infra : local: 41476, source: 41280 >> update: graphics [1924k]: local: 41015, source: 41447 >> update: luatex-def [10k]: local: 41398, source: 41450 >> update: pdftex-def [15k]: local: 22653, source: 41449 >> >> and I think graphics-41447 is the broken one. >> >> Jean-Fran?ois > > -- > PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info > GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From preining at logic.at Sat Jun 18 07:52:16 2016 From: preining at logic.at (Norbert Preining) Date: Sat, 18 Jun 2016 14:52:16 +0900 Subject: [tex-live] xcolor broken with pdflatex In-Reply-To: References: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> <20160617102345.GE21093@logic.at> <42490C09-B19D-4E6D-A76B-8D02015B7F7D@free.fr> <20160618004854.GC14671@logic.at> <5498EC78-CE7C-4C5B-99A1-89EF39E500DC@free.fr> <2C1A255A-9732-4430-8534-A0D93D141172@logic.at> Message-ID: <20160618055216.GG14671@logic.at> On Sat, 18 Jun 2016, jfbu wrote: > however if I reinstall the current versions, I get a failed compilation YUour server is not uptodate. > [1/4, ??:??/??:??] update: graphics [1924k] (41015 -> 41447) ... done > [2/4, 00:04/00:04] update: graphics-cfg [2k] (40269 -> 41448) ... done > [3/4, 00:05/00:05] update: luatex-def [10k] (41398 -> 41450) ... done > [4/4, 00:06/00:06] update: pdftex-def [15k] (22653 -> 41449) ... done Here I get: [ 4/10, 00:04/00:12] update: graphics [2000k] (41447 -> 41468 at main) ... done [ 8/10, 00:09/00:13] update: luatex-def [10k] (41450 -> 41466 at main) ... done [ 9/10, 00:10/00:15] update: pdftex-def [15k] (41449 -> 41467 at main) ... done Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From jfbu at free.fr Sat Jun 18 07:53:46 2016 From: jfbu at free.fr (jfbu) Date: Sat, 18 Jun 2016 07:53:46 +0200 Subject: [tex-live] xcolor broken with pdflatex In-Reply-To: References: <5834b23c-92f5-200d-b91a-73c449bda09a@gmail.com> <20160617102345.GE21093@logic.at> <42490C09-B19D-4E6D-A76B-8D02015B7F7D@free.fr> <20160618004854.GC14671@logic.at> <5498EC78-CE7C-4C5B-99A1-89EF39E500DC@free.fr> <2C1A255A-9732-4430-8534-A0D93D141172@logic.at> Message-ID: <0D40250C-C72C-4EDA-A6AD-507AEB10F3AD@free.fr> Le 18 juin 2016 ? 07:43, jfbu a ?crit : >> The current versions, including the graphics one, should be the fixed one. > I just saw announcement on c.t.tex from CTAN, but it appears that from where I am the mirror which was chosen by the multiplexer does not yet serve these updates, that's all similarly the announced luaotfload 2.7-fix-3 is not yet updated via my use of tlmgr best, Jean-Fran?ois From preining at logic.at Sat Jun 18 07:59:18 2016 From: preining at logic.at (Norbert Preining) Date: Sat, 18 Jun 2016 14:59:18 +0900 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: References: Message-ID: <20160618055918.GH14671@logic.at> Hi Uwe, I think this is a nice idea, but needs good thinking. I don't want just a bunch of examples - all must be automatized and best included in our daily checks. I have pushed the L3 Team to include support for reproducible builds (SOURCE_DATE_EPOCH) into the l3build suite, and I am now working/thinking about setting up some environment so that package authors can upload tests, or we write them. These test would be run against l3build tester automatically during the nightly rebuild. For now I am thinking aobut hte following: * loading test: - for every latex package (.sty file) if nothing is extra specified for testts we try a simple document like: \documentclass{article} \usepackage{....} \begin{document}Hellow Wolrd\end{document} and make sure that hte generated pdf does not change under S_D_E settings. - one can override the documentclass, engine, etc etc on a per sty file - one can submit more involved tests * other tests coming (combinations, user supplied) Well, that is all in very early stages and if I would have more time I would have written some code already - maybe during TUG 2016. Are you coming? Maybe intereted people can make a brain storming session during TUG. Also, I would prefer a git repository on github/lab or similar. On Sat, 18 Jun 2016, Uwe Ziegenhagen wrote: > I think it might be a good idea to create a general test suite for TeX Live All the best Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From hendry at webconverger.com Sat Jun 18 08:17:45 2016 From: hendry at webconverger.com (Kai Hendry) Date: Sat, 18 Jun 2016 07:17:45 +0100 Subject: [tex-live] Letter template In-Reply-To: References: Message-ID: Thank you for your reply Zdenek, On 17 June 2016 at 09:09, Zdenek Wagner wrote: > this question has nothing to do with TeX Live maintenance so this should be > posted somewhere else. Sorry, please let me know a better forum. Perhaps http://tex.stackexchange.com/ ? > Anyway, I am afraid that a good general answer does not exist. Although > various internationalization projects exist so that you can prepare a > general template containing macros as \from, \to, \cc etc., the universe is > not that grey. First, there are different conventions in different countries > one of them being the position of the signature. Generally it can be said > that for A4 or letter size paper the font size should be 11pt or 12pt, > margins should be around 1in (the old Czech standard required 25mm which is I think that's really the problem I have. I don't even know how to change the font size or if I even need this lines: \usepackage{fontspec} \defaultfontfeatures{Mapping=tex-text} \usepackage{xunicode} \usepackage{babel} > almost the same dimension). Setting of \parindent and \parskip is also a > matter of convention. What is the most important is "corporate identity". I > was asked to prepare letter templates for several companies. The layout was > designed by a graphuical design studio. The people from the studio do not > know LaTeX and the companies requested LaTeX so I have implemented it for > them. In short, if you wish a professional design, ask a graphical designer. > If you think that you can do it yourself. look at letters that you receive > from others and try to do it in a similar way. I don't have the skills to hack tex templates. I would be just shooting in the dark. I was hoping to find a canonical minimal letter template and go from there but I was struggling. I did find a couple of more templates on a Github search: https://github.com/khemanta/Rescv/blob/master/CL.tex <-- bit complex https://github.com/albcorp/alb-latex/blob/master/make/serial/example-letter.tex <-- better So I've since minimised my template: https://github.com/kaihendry/letterly/commit/45dd0a0de6590fb2466205f86ffa3cab7db5a780 I find marrying tex with a Web form quite a challenge. Perhaps someone has already done the work? For example certain characters like # need escaping. Is there a list? There are certain commands I need to disable like braces https://github.com/kaihendry/letterly/blob/master/www/pdf/index.php#L3 I am running in a container which should prevent an attacker "owning" my system, but it seems very difficult to secure latex from malicious input. Has anyone converted markdown the latex? I simply want to be able to write lists and embolden text and have that reflected in the output without the jarring latex syntax. How does one handle emojis or unicode? Since in my tests it doesn't work at all. And last but not least, why is texlive-bin texlive-core (both seems typically needed) about 1GB !? Can't help but feel tex is headed into obsolescence. Kind regards, From ziegenhagen at gmail.com Sat Jun 18 08:33:28 2016 From: ziegenhagen at gmail.com (Uwe Ziegenhagen) Date: Sat, 18 Jun 2016 08:33:28 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <20160618055918.GH14671@logic.at> References: <20160618055918.GH14671@logic.at> Message-ID: I can't come to TUG, too many things keeping me busy here right now. Automated testing is a good idea. I can start an Etherpad this weekend to collect the requirements. Git is fine for me of course. Uwe 2016-06-18 7:59 GMT+02:00 Norbert Preining : > Hi Uwe, > > I think this is a nice idea, but needs good thinking. I don't want > just a bunch of examples - all must be automatized and best > included in our daily checks. > > I have pushed the L3 Team to include support for reproducible builds > (SOURCE_DATE_EPOCH) into the l3build suite, and I am now working/thinking > about setting up some environment so that package authors > can upload tests, or we write them. These test would be run against > l3build tester automatically during the nightly rebuild. > > For now I am thinking aobut hte following: > * loading test: > - for every latex package (.sty file) if nothing is extra specified > for testts we try a simple document like: > \documentclass{article} > \usepackage{....} > \begin{document}Hellow Wolrd\end{document} > and make sure that hte generated pdf does not change under S_D_E > settings. > - one can override the documentclass, engine, etc etc on a per sty > file > - one can submit more involved tests > * other tests coming (combinations, user supplied) > > Well, that is all in very early stages and if I would have more > time I would have written some code already - maybe during TUG 2016. > Are you coming? > > Maybe intereted people can make a brain storming session during TUG. > > Also, I would prefer a git repository on github/lab or similar. > > On Sat, 18 Jun 2016, Uwe Ziegenhagen wrote: > > I think it might be a good idea to create a general test suite for TeX > Live > > All the best > > Norbert > > -- > PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info > GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 > -- Dr. Uwe Ziegenhagen -------------- next part -------------- An HTML attachment was scrubbed... URL: From komascript at gmx.info Sat Jun 18 08:42:23 2016 From: komascript at gmx.info (Markus Kohm) Date: Sat, 18 Jun 2016 08:42:23 +0200 Subject: [tex-live] Letter template In-Reply-To: References: Message-ID: <3737341.2bnjbHTBLL@x61s.fritz.box> Am Samstag, 18. Juni 2016, 07:17:45 schrieb Kai Hendry: > I don't even know how to > change the font size or if I even need this lines You should at least read an introduction to LaTeX, e.g. or even . After this you should be prepared to search and ask on . -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 181 bytes Desc: This is a digitally signed message part. URL: From jfbu at free.fr Sat Jun 18 09:31:04 2016 From: jfbu at free.fr (jfbu) Date: Sat, 18 Jun 2016 09:31:04 +0200 Subject: [tex-live] 3.4.4 System font configuration for XeTEX and LuaTEX In-Reply-To: References: <04B5F63F-0F54-44EC-9C48-3D477FB4A899@free.fr> <201605132249.u4DMnYWS030466@freefriends.org> Message-ID: Hi all, with luaotfload 2.7-fix-3 now on CTAN, the symlinks are automatically filtered out from the database Hence doing on mac os x ln -s /Library/TeX/Root/texmf-dist/fonts/opentype ~/Library/Fonts/TeXOpenType ln -s /Library/TeX/Root/texmf-dist/fonts/truetype ~/Library/Fonts/TeXTrueType to make all opentype/truetype fonts inside the texlive/mactex tree visible to XeTeX hasn't anymore an impact on increased font database for LuaTeX Jean-Fran?ois Le 16 mai 2016 ? 01:24, Zdenek Wagner a ?crit : > Hi, > > this is good to know. A friend of mine just copied all the font files which I do not consider good. Imagine that one year later you will have to remove TL fonts and install the new versions. Copying all files can then turn into a nightmare. This is why I sggested to try to symlink the directory tree. It is good news that it works. > > I have not only the TL fonts but also a huge collection of commercial fonts. Yes, when I want to see the font list in any Linux application, it takes some time but I do not need it frequently. I prefer having the fonts available. > > > Zden?k Wagner > http://ttsm.icpf.cas.cz/team/wagner.shtml > http://icebearsoft.euweb.cz > > 2016-05-15 16:38 GMT+02:00 jfbu : > Hi > > > Le 14 mai 2016 ? 09:37, Zdenek Wagner a > ?crit : > > >> 2016-05-14 0:49 GMT+02:00 Karl Berry : > >> because in recent years I don't think that was the case. > >> > >> I suppose you're right. I'll have to ask about what can be > >> sensibly said instead. Thanks. > >> > >> Else one must use filenames, or fontspec with ExternalLocation > >> > >> Well, that's preferable anyway :). --karl > > As I remember, someone has already mentioned that the font > > instructions for Mac OSX are wrong. > > I think that was me, a few weeks ago ;-) and perhaps also > last year ;-) > > Then the formulation changed but the new formulation says that > things are automatically set-up for XeTeX on mac os x, which is > wrong. > > > > I do not have Mac so I cannot try whether it is really necessary > > to copy the fonts, maybe the symlink would be sufficient. And I do > > not know whether the font files must be symlinked, maybe just the > > parent directory may be symlinked. I do not know whether Mac can > > look for fonts into subdirectories. > > > > 1. I had not tried earlier the symlink approach because the Mac > comes with FontBook app which has a concept of "Font Library" > which one can add to it without having to copy or symlink the > fonts. Also I feared that perhaps I would have to symlink > individually files or many repertories. > > The "Font Library" method initially seemed to work but I soon > realized it was a nightmare, because FontBook.app is seriously > broken (at least on Mavericks, but Herb indicates he also has had > bad experiences). > > Thus I dropped it some months ago and entered resignation mode. > > (Michael Sharpe has some very nice extras to his font packages so > that even when accessing a font by filename with xelatex one does > not have to list explicitely the bold, italic, etc.. variants, but > doing this for all fonts obviously requires extra work) > > 2. But Herb Schulz gave me very good advice to simply symlink the > texmf-dist/fonts/opentype (and I did also texmf-dist/fonts/truetype) > from $HOME/Library/Fonts. The miracle is : > > a. FontBook.app is still usable after that. There is only a > slight delay in opening it, about two seconds during which the font > list is prepared. > > b. all fonts in sub-repertories are found nicely > > c. there are a few duplicates between .otf and .ttf (FreeSerif, > FreeSans, FreeMono in particular) but the FontBook GUI allows to > choose which fonts to validate and this works with no hiccup. > (which is surprising regarding FontBook.app) > > d. there is slight disadvantage that some fonts with names > starting with "drm" occupies many (dozens) lines in FontBook > display, but that's cosmetic. > > After my bad experience with the concept of "Font Library" I was > very much surprised that simply putting symlinks worked out so > well. Thanks Herb ! > > And I could confirm that after that, all the TeXLive distributed > opentype/truetype fonts become generally available to Mac > Applications such as TextEdit, OpenOffice, etc... and to XeTeX > naturally. The delay I mentioned above applies only to FontBook.app, > applications such as OpenOffice have their font popup instantly > available with TeX Gyre Termes etc... and naturally XeTeX works > with no delay either. > > 3. There is perhaps an impact on LuaTeX+luaotfload, because the > TeXLive distributed fonts will be indexed twice. I do not know if > this slows down LuaTeX. I opened a ticked on github site of > luaotfload about this asking for info. > > https://github.com/lualatex/luaotfload/issues/359 > > 4. The specific way to symlink by Herb takes advantage of the > TeXDist structure which has been part of MacTeX install for some > years, in /Library/TeX. This way the symlinks are dynamical and > obey on the fly change of TeXLive distribution year. > > ln -s /Library/TeX/Root/texmf-dist/fonts/opentype > ~/Library/Fonts/TeXOpenType > > I did also > > ln -s /Library/TeX/Root/texmf-dist/fonts/truetype > ~/Library/Fonts/TeXTrueType > > but had to resolve a few duplicates then as mentioned above. > > (the point is that /Library/TeX/Root is already itself a symlink, > and when changing distribution year via the former Panel or via > TLU, its target is automatically updated) > > 5. To make my story complete: took me a while to go to the root > of the following problem. I was surprised to see that log files > with \XeTeXtracingfonts1 indicated the opentype fonts were picked > always from texlive/2015, even if working with TL2014 or TL2016 > pretest. > > Turns out that my earlier attempt at adding TeXLive OpenType fonts > to general mac os availability via a "Font Library" added to FontBook.app > had resulted in some non-removed cache in /private/var/folders. > > I looked > in vain everywhere in my $HOME and in obvious system locations, > for such a file, as it was obviously the only explanation to the > observed behavior and was finally saved by the tip at > > http://osxdaily.com/2015/01/08/clear-font-caches-databases-mac-os-x/ > > which is to do $ sudo atsutil databases -remove > > it is important to reboot immediately after that. > > This last item is only of relevance to people who messed up with > FontBook.app before, I guess. > > To sum up: For smooth experience with XeTeX on Mac OS X, some > people (me following Herb) have found that > > ln -s /Library/TeX/Root/texmf-dist/fonts/opentype ~/Library/Fonts/TeXOpenType > > and > > ln -s /Library/TeX/Root/texmf-dist/fonts/truetype ~/Library/Fonts/TeXTrueType > > is a well-working approach. In some rare cases, it might be needed to do > (before or after it does not matter) > > sudo atsutil databases -remove > > and > to reboot. That's only to clean-up a pre-existing dysfunctional font situation. > > Hope it helps. > > Jean-Fran?ois > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jfbu at free.fr Sat Jun 18 10:44:40 2016 From: jfbu at free.fr (jfbu) Date: Sat, 18 Jun 2016 10:44:40 +0200 Subject: [tex-live] xdvi on mac os Message-ID: Hi, on mac os x 10.9.5, XQuartz 2.7.8 (xorg-server 1.16.4) I have a problem with xdvi. Unfortunately I can't recall last time I used it. There is no backup available from tlmgr retore (besides, when the pretest period ended, I removed all backups). $ tlmgr restore xdvi.x86_64-darwin Available backups for xdvi.x86_64-darwin: (nothing) Now, from an xterm launched in XQuartz $ which xdvi /Library/TeX/texbin/xdvi $ xdvi & [1] 55250 $ dyld: Library not loaded: /opt/X11/lib/libXt.7.dylib Referenced from: /Library/TeX/texbin/xdvi-xaw Reason: image not found [1]+ Trace/BPT trap: 5 xdvi After reverting to TL2015, works fine. (without even re-launching xterm). State of memory makes it impossible for me to say if I tested during pretest or since. I believe Yes for pretest-period, but not sure. The XQuartz as stated above is the 2.7.8, not the 2.7.9. Best Jean-Fran?ois From siepo at cybercomm.nl Sat Jun 18 12:38:11 2016 From: siepo at cybercomm.nl (Siep Kroonenberg) Date: Sat, 18 Jun 2016 12:38:11 +0200 Subject: [tex-live] texcount on Windows fail because of Perl file missing from TeX Live distribution In-Reply-To: <306071466104496@web16h.yandex.ru> References: <306071466104496@web16h.yandex.ru> Message-ID: <20160618103811.GA2920@bitmuis.thuis.internal> On Thu, Jun 16, 2016 at 10:14:56PM +0300, Leonid Sinev wrote: > Greetings, > > This problem already discussed at > https://www.tug.org/pipermail/tex-live/2015-August/037293.html > but nothing changed in TeXLive 2016. > > Nothing changed after > http://tex.stackexchange.com/a/145766 advice > of Win32::Console::ANSI from CPAN > I have copied file from this package to > C:\texlive\2016\tlpkg\tlperl\site\lib\Win32\Console\ANSI.pm > and got this: > > Can't locate loadable object for module Win32::Console::ANSI in @INC (@INC conta ins: C:/texlive/2016/tlpkg/tlperl/lib c:/texlive/2016/tlpkg/tlperl/site/lib c:/t exlive/2016/tlpkg/tlperl/lib .) at c:\texlive\2016\texmf-dist\scripts\texcount\t excount.pl line 11. > Compilation failed in require at c:\texlive\2016\texmf-dist\scripts\texcount\tex count.pl line 11. > BEGIN failed--compilation aborted at c:\texlive\2016\texmf-dist\scripts\texcount \texcount.pl line 14. > C:\texlive\2016\bin\win32\runscript.tlu:679: command failed with exit code 2: perl.exe c:\texlive\2016\texmf-dist\scripts\texcount\texcount.pl > > Can there be a patch or update for TeXLive for windows that can make windows users use Perl scripts like texcount without need of additional perl installation or messing with additional packaging for tlperl? > > --? > Best regards, > Leonid 1. This module includes a dll which would need to be added too. 2. MetaCPAN https://metacpan.org/pod/Win32::Console::ANSI lists a dependency Win32::Pipe which is not in tlperl either As Karl noted, the included perl for windows does not aim to support all included perl scripts. -- Siep Kroonenberg From mojca.miklavec.lists at gmail.com Sat Jun 18 13:10:41 2016 From: mojca.miklavec.lists at gmail.com (Mojca Miklavec) Date: Sat, 18 Jun 2016 13:10:41 +0200 Subject: [tex-live] xdvi on mac os In-Reply-To: References: Message-ID: On 18 June 2016 at 10:44, jfbu wrote: > Hi, > > on mac os x 10.9.5, XQuartz 2.7.8 (xorg-server 1.16.4) > > I have a problem with xdvi. You need to upgrade XQuartz to 2.7.9. There have been recent updates of binaries (after the TL release) and there is no way to support both 2.7.8 and 2.7.9 at the same time with a single binary, so Dick decided to support the latest version only. Mojca From jfbu at free.fr Sat Jun 18 13:18:43 2016 From: jfbu at free.fr (jfbu) Date: Sat, 18 Jun 2016 13:18:43 +0200 Subject: [tex-live] xdvi on mac os In-Reply-To: References: Message-ID: <3D1282CD-7DD4-43D9-9CA9-0D33123C38E9@free.fr> Le 18 juin 2016 ? 13:10, Mojca Miklavec a ?crit : > On 18 June 2016 at 10:44, jfbu wrote: >> Hi, >> >> on mac os x 10.9.5, XQuartz 2.7.8 (xorg-server 1.16.4) >> >> I have a problem with xdvi. > > You need to upgrade XQuartz to 2.7.9. There have been recent updates > of binaries (after the TL release) and there is no way to support both > 2.7.8 and 2.7.9 at the same time with a single binary, so Dick decided > to support the latest version only. > Thanks! I thought I had followed that thread, but obviously I did not quite get it. But will XQuartz 2.7.9 be backwards compatible with xdvi binaries of bygone years ? It is very important to me to be able to switch back to one of the earlier TeXLive on my computer, JF From chuck.newman at hpe.com Sat Jun 18 04:34:46 2016 From: chuck.newman at hpe.com (Newman, Chuck) Date: Sat, 18 Jun 2016 02:34:46 +0000 Subject: [tex-live] Error installing install-tl-unx.tar.gz Message-ID: # ./install-tl cannot contact mirror.ctan.org, returning a backbone server! Loading http://www.ctan.org/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb ./tlpkg/installer/xz/xzdec.x86_64-linux: (stdin): File format not recognized ./install-tl: signature verification error of /tmp/0etZ_Fw4Ec/paQlkCVfEv from http://www.ctan.org/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb: cryptographic signature verification of /tmp/0etZ_Fw4Ec/GiM4SAMCXi against http://www.ctan.org/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb.sha512.asc failed. Output was gpg: no valid OpenPGP data found. gpg: the signature could not be verified. Please remember that the signature file (.sig or .asc) should be the first file given on the command line. Please report to texlive at tug.org From preining at logic.at Sat Jun 18 16:08:46 2016 From: preining at logic.at (Norbert Preining) Date: Sat, 18 Jun 2016 23:08:46 +0900 Subject: [tex-live] Error installing install-tl-unx.tar.gz In-Reply-To: References: Message-ID: <20160618140846.GK14671@logic.at> Hi Chuck, > Loading http://www.ctan.org/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb > ./tlpkg/installer/xz/xzdec.x86_64-linux: (stdin): File format not recognized Is this reproducible? I just tried myself with the very same mirror and didn't have any problems. It seems that the download of the tlpdb.xz file failed. Can you please retry and see if it works? Do you have any proxies or filters on the way? > ./install-tl: signature verification error of /tmp/0etZ_Fw4Ec/paQlkCVfEv from http://www.ctan.org/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb: cryptographic signature verification of All the rest is just a continuation of the problems above. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From herries.press at earthlink.net Sat Jun 18 20:03:11 2016 From: herries.press at earthlink.net (Peter Wilson) Date: Sat, 18 Jun 2016 19:03:11 +0100 Subject: [tex-live] Letter template In-Reply-To: <3737341.2bnjbHTBLL@x61s.fritz.box> References: <3737341.2bnjbHTBLL@x61s.fritz.box> Message-ID: <57658CDF.3060609@earthlink.net> On 18/06/16 07:42, Markus Kohm wrote: > Am Samstag, 18. Juni 2016, 07:17:45 schrieb Kai Hendry: >> I don't even know how to >> change the font size or if I even need this lines > You should at least read an introduction to LaTeX, e.g. > or even . > > After this you should be prepared to search and ask on > . Don't forget the standard LaTeX letter class. From chuck.newman at hpe.com Sat Jun 18 20:46:52 2016 From: chuck.newman at hpe.com (Newman, Chuck) Date: Sat, 18 Jun 2016 18:46:52 +0000 Subject: [tex-live] Error installing install-tl-unx.tar.gz In-Reply-To: <20160618140846.GK14671@logic.at> References: <20160618140846.GK14671@logic.at> Message-ID: Yes, I still get the same error, although the /tmp filename is different. Below is what I got just now. Yes, there is a proxy server that I have to go through. # ./install-tl cannot contact mirror.ctan.org, returning a backbone server! Loading http://www.tex.ac.uk/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb ./tlpkg/installer/xz/xzdec.x86_64-linux: (stdin): File format not recognized ./install-tl: signature verification error of /tmp/Y6yDS8DtAw/Qeut_ibPT4 from http://www.tex.ac.uk/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb: cryptographic signature verification of /tmp/Y6yDS8DtAw/9PCwZUJXXc against http://www.tex.ac.uk/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb.sha512.asc failed. Output was gpg: no valid OpenPGP data found. gpg: the signature could not be verified. Please remember that the signature file (.sig or .asc) should be the first file given on the command line. Please report to texlive at tug.org -----Original Message----- From: Norbert Preining [mailto:preining at logic.at] Sent: Saturday, 18 June, 2016 10:09 To: Newman, Chuck Cc: texlive at tug.org Subject: Re: [tex-live] Error installing install-tl-unx.tar.gz Hi Chuck, > Loading http://www.ctan.org/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb > ./tlpkg/installer/xz/xzdec.x86_64-linux: (stdin): File format not recognized Is this reproducible? I just tried myself with the very same mirror and didn't have any problems. It seems that the download of the tlpdb.xz file failed. Can you please retry and see if it works? Do you have any proxies or filters on the way? > ./install-tl: signature verification error of /tmp/0etZ_Fw4Ec/paQlkCVfEv from http://www.ctan.org/tex-archive/systems/texlive/tlnet/tlpkg/texlive.tlpdb: cryptographic signature verification of All the rest is just a continuation of the problems above. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From esbati at gmx.de Sat Jun 18 23:47:20 2016 From: esbati at gmx.de (Arash Esbati) Date: Sat, 18 Jun 2016 23:47:20 +0200 Subject: [tex-live] egreg.txt in LaTeX base Message-ID: Hi all, with the latest update for LaTeX base, there comes a file named `egreg.txt': $ tlmgr info --list latex | grep egreg texmf-dist/doc/latex/base/egreg.txt The file looks like a snippet from a chat on tex.sx. Is this on purpose? Best, Arash From d.p.carlisle at gmail.com Sun Jun 19 00:24:30 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Sat, 18 Jun 2016 23:24:30 +0100 Subject: [tex-live] egreg.txt in LaTeX base In-Reply-To: References: Message-ID: oops:-) will remove for next time, thanks David On 18 June 2016 at 22:47, Arash Esbati wrote: > Hi all, > > with the latest update for LaTeX base, there comes a file named > `egreg.txt': > > $ tlmgr info --list latex | grep egreg > texmf-dist/doc/latex/base/egreg.txt > > The file looks like a snippet from a chat on tex.sx. Is this on > purpose? > > Best, Arash > -- http://dpcarlisle.blogspot.com/ From koch at uoregon.edu Sun Jun 19 01:28:30 2016 From: koch at uoregon.edu (Richard Koch) Date: Sat, 18 Jun 2016 16:28:30 -0700 Subject: [tex-live] xdvi on mac os Message-ID: <68C7A3F4-C18E-4E22-94CF-D8738D012645@uoregon.edu> JF asked "but will XQuartz 2.7.9 be backward compatible with xdvi binaries of bygone years?". I'm in San Francisco for the opera: Carmen last night, Don Carlos tonight, Jenufu tomorrow. The answer is that you cannot have everything (and thus the answer is NO). TeX Live protects itself against library changes by including its own copy of most important libraries. Thus if you run TeX Live 2014, you are running binaries compiled with an older version of libpng, or popler, or icu, but you are also running older versions of these libraries. One exception is common libraries like the standard C library, but these are usually upward compatible. Mac 64 bit binaries, for example, were compiled on Snow Leopard. X11 is a special case because it is a fundamental library on Linux machines, so TeX Live doesn't contain it. You can install the latest XQuartz on Snow Leopard and above, so all our machines can run the latest binaries. But do older binaries of xdvi run on these machines after they have been updated to the latest XQuartz. Luckily, my traveling portable has the 2012, 2013, and 2014 binaries, so I can test, and the answer is NO. Usually users need xdvi for "isolated checks", and it is easy enough to switch to TeXLive 2016 temporarily and make that check. Since this email is essentially a piece of bad news, let me end with some good news. I am running MacTeX-2016 and TeX Live 2016, and doing these checks, in Sierra. Dick Koch From reinhard.kotucha at web.de Sun Jun 19 04:12:44 2016 From: reinhard.kotucha at web.de (Reinhard Kotucha) Date: Sun, 19 Jun 2016 04:12:44 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <20160618055918.GH14671@logic.at> References: <20160618055918.GH14671@logic.at> Message-ID: <22373.65436.46521.354964@zaphod.ms25.net> On 2016-06-18 at 14:59:18 +0900, Norbert Preining wrote: > Hi Uwe, > > I think this is a nice idea, but needs good thinking. I don't want > just a bunch of examples - all must be automatized and best > included in our daily checks. Hi Uwe and Norbert, maybe it's worthwhile to know how Thomas Esser tested all the packages before they were included in teTeX. Thomas said that his tests were never supposed to detect all problems but a package should at least be able to compile its own documentation. Thus he created a Makefile for each and every package he supported. Even this was an enormous amount of work. teTeX was a small subset of TeX Live, hence it was feasible. But TeX Live was always much bigger and grew rapidly since then. Compiling package documentation is a good starting point. If a package comes with a .dtx file, the process can probably be automatized, but I fear that most packages require manual intervention. An other point comes to my mind. Occasionally people report that some dependencies in texlive.tlpdb are incorrect. Of course, it's difficult to get everything correct. But if there is a halfway comprehensive test suite we can probably determine dependencies from the log files. Another approach is to run TeX with the --recorder option on these files and evaluate the .fls files afterwards. I don't think that the dependencies in texlive.tlpdb can be determined automatically but I'm convinced that a halfway comprehensive test suite can provide useful information. Uwe's idea is, as far as I understand, to create files which load many packages at once. This reduces the mumber of required tests significantly. The caveat is that some packages are overlooked because Uwe doesn't use them. After all, I like Uwe's idea. In TeX parlance: \let\uwe\proceed Regards, Reinhard -- ------------------------------------------------------------------ Reinhard Kotucha Phone: +49-511-3373112 Marschnerstr. 25 D-30167 Hannover mailto:reinhard.kotucha at web.de ------------------------------------------------------------------ From preining at logic.at Sun Jun 19 07:31:11 2016 From: preining at logic.at (Norbert Preining) Date: Sun, 19 Jun 2016 14:31:11 +0900 Subject: [tex-live] Error installing install-tl-unx.tar.gz In-Reply-To: References: <20160618140846.GK14671@logic.at> Message-ID: <20160619053111.GP14671@logic.at> Hi CHuck, again, I am seriously believe that something in your network setup is broken. It seems that the proxy needs some login, or whatever. > cannot contact mirror.ctan.org, returning a backbone server! First of all, this is *strange*. It means that you cannot connect. So my guess is the following: Your proxy allows only a certain set of "User Agents" - that means that a User Agent of LWP or so is blocked. That would explain If you have lwp-download installed, can you try lwp-download http://mirror.ctan.org and send the output of it, probably saved into "index" named file. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From jfbu at free.fr Sun Jun 19 08:32:00 2016 From: jfbu at free.fr (jfbu) Date: Sun, 19 Jun 2016 08:32:00 +0200 Subject: [tex-live] xdvi on mac os In-Reply-To: <68C7A3F4-C18E-4E22-94CF-D8738D012645@uoregon.edu> References: <68C7A3F4-C18E-4E22-94CF-D8738D012645@uoregon.edu> Message-ID: <81DDA609-5D3D-4809-89A0-19E022884B0E@free.fr> Hi Dick, > JF asked "but will XQuartz 2.7.9 be backward compatible with > xdvi binaries of bygone years?". > > I'm in San Francisco for the opera: Carmen last night, Don Carlos > tonight, Jenufu tomorrow. that's quite a treat ! reminds me of the technique my dentist employed last time to get me relaxed ;-) > > The answer is that you cannot have everything (and thus the > answer is NO). > [...] thanks for taking time to explain. It was implicit or even explicit from earlier messages, but I did not then read closely enough. I always have the possibility to archive my current XQuartz 2.7.8, and ressuscitating it in my /opt for one-shot occasions could be as easy as executing a script (although I am not sure about the /Applications part). And, it is true that I don't expect too much trouble if my xdvi use is tied to TL2016. Jean-Fran?ois From zdenek.wagner at gmail.com Sun Jun 19 09:30:51 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Sun, 19 Jun 2016 09:30:51 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <22373.65436.46521.354964@zaphod.ms25.net> References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> Message-ID: Hi all, the idea of automated tested was here more than a year ago and I explained that I consider it impossible. Nobody knows which set of packages will be used in a document created by each user. It is only possible to create subsets of test files. It happened to my zwpagelayout package, it did not work for one user at all. After the bug report I found conflicting packages, therefore I added code to cope with it. I still cannot guarantee that it will work with every package (it does not work with xcolor and the support of RTL is very limited). Without loading such packages you cannot find the problem and it took a few years before somebody did it. In this case the error appeared when the package was loaded, but for instance if you have path.sty together with any package loading url.sty (for instance hyperref), the packages wil load without errors but it will fail when typesetting the URL. If you load a lot of packages into a single document, it is more probable that new clashes will be detected but then it might be difficult to find the reason. Moreover, it sometimes depends on the order of packages. Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz 2016-06-19 4:12 GMT+02:00 Reinhard Kotucha : > On 2016-06-18 at 14:59:18 +0900, Norbert Preining wrote: > > > Hi Uwe, > > > > I think this is a nice idea, but needs good thinking. I don't want > > just a bunch of examples - all must be automatized and best > > included in our daily checks. > > Hi Uwe and Norbert, > maybe it's worthwhile to know how Thomas Esser tested all the packages > before they were included in teTeX. > > Thomas said that his tests were never supposed to detect all problems > but a package should at least be able to compile its own > documentation. Thus he created a Makefile for each and every package > he supported. > > Even this was an enormous amount of work. teTeX was a small subset of > TeX Live, hence it was feasible. But TeX Live was always much bigger > and grew rapidly since then. > > Compiling package documentation is a good starting point. If a > package comes with a .dtx file, the process can probably be > automatized, but I fear that most packages require manual > intervention. > > An other point comes to my mind. Occasionally people report that some > dependencies in texlive.tlpdb are incorrect. Of course, it's > difficult to get everything correct. But if there is a halfway > comprehensive test suite we can probably determine dependencies from > the log files. Another approach is to run TeX with the --recorder > option on these files and evaluate the .fls files afterwards. > > I don't think that the dependencies in texlive.tlpdb can be determined > automatically but I'm convinced that a halfway comprehensive test > suite can provide useful information. > > Uwe's idea is, as far as I understand, to create files which load many > packages at once. This reduces the mumber of required tests > significantly. The caveat is that some packages are overlooked > because Uwe doesn't use them. > > After all, I like Uwe's idea. In TeX parlance: > > \let\uwe\proceed > > Regards, > Reinhard > > -- > ------------------------------------------------------------------ > Reinhard Kotucha Phone: +49-511-3373112 > Marschnerstr. 25 > D-30167 Hannover mailto:reinhard.kotucha at web.de > ------------------------------------------------------------------ > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jfbu at free.fr Sun Jun 19 09:38:37 2016 From: jfbu at free.fr (jfbu) Date: Sun, 19 Jun 2016 09:38:37 +0200 Subject: [tex-live] xdvi on mac os In-Reply-To: <81DDA609-5D3D-4809-89A0-19E022884B0E@free.fr> References: <68C7A3F4-C18E-4E22-94CF-D8738D012645@uoregon.edu> <81DDA609-5D3D-4809-89A0-19E022884B0E@free.fr> Message-ID: <9E2799C3-A3AA-48C5-865B-ECEBFF13F4E6@free.fr> >> >> JF asked "but will XQuartz 2.7.9 be backward compatible with >> xdvi binaries of bygone years?". > >> >> The answer is that you cannot have everything (and thus the >> answer is NO). >> [...] > perhaps I should simply use the currently distributed xdvi, xdvi-xaw as compiled by Dick to work with XQuartz 2.7.9 and copy them or symlink them from 2016 to the earlier TeXLive ? wouldn't that work ? thus my earlier texlive/YYYY would be slightly altered but with a functional xdvi ? Best, Jean-Fran?ois From mojca.miklavec.lists at gmail.com Sun Jun 19 09:42:38 2016 From: mojca.miklavec.lists at gmail.com (Mojca Miklavec) Date: Sun, 19 Jun 2016 09:42:38 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> Message-ID: On 19 June 2016 at 09:30, Zdenek Wagner wrote: > Hi all, > > the idea of automated tested was here more than a year ago and I explained > that I consider it impossible. It is certainly impossible to create a complete test suite that would catch any problem one could run into. But that doesn't mean (and shouldn't be an excuse) that a group of volunteers couldn't come up with a "small" test suite to cover at least the most basic functionality and to prevent cases with packages not loading at all. I would imagine that such a test suite should be able to raise an alarm during the latest magyar.ldf update (even though I still argue that it wouldn't prevent a savy TeX user from tricking CTAN and TL into accepting a formally valid tex file with super weird functionality). Mojca From jfbu at free.fr Sun Jun 19 09:54:23 2016 From: jfbu at free.fr (jfbu) Date: Sun, 19 Jun 2016 09:54:23 +0200 Subject: [tex-live] xdvi on mac os In-Reply-To: <9E2799C3-A3AA-48C5-865B-ECEBFF13F4E6@free.fr> References: <68C7A3F4-C18E-4E22-94CF-D8738D012645@uoregon.edu> <81DDA609-5D3D-4809-89A0-19E022884B0E@free.fr> <9E2799C3-A3AA-48C5-865B-ECEBFF13F4E6@free.fr> Message-ID: <826480F0-B2CF-4E5D-99E1-988506DF8A0A@free.fr> Le 19 juin 2016 ? 09:38, jfbu a ?crit : > > >>> >>> JF asked "but will XQuartz 2.7.9 be backward compatible with >>> xdvi binaries of bygone years?". >> >>> >>> The answer is that you cannot have everything (and thus the >>> answer is NO). >>> [...] >> > > perhaps I should simply use the currently distributed xdvi, xdvi-xaw > as compiled by Dick to work with XQuartz 2.7.9 and copy them or > symlink them from 2016 to the earlier TeXLive ? wouldn't that work ? > Hello world example works (tried with TL2012's xdvi-xaw replaced by a symlink to TL2016's xdvi-xaw) I don't know anything about the innards of xdvi and which libraries it uses and whether symlinking it in this way to something away from PATH is ok. (I recall that some years ago it started using some t1lib which avoided pk fonts, that's all) > thus my earlier texlive/YYYY would be slightly altered but with > a functional xdvi ? JF From preining at logic.at Sun Jun 19 10:17:16 2016 From: preining at logic.at (Norbert Preining) Date: Sun, 19 Jun 2016 17:17:16 +0900 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> Message-ID: <20160619081716.GR14671@logic.at> > It is certainly impossible to create a complete test suite that would > catch any problem one could run into. > > But that doesn't mean (and shouldn't be an excuse) that a group of > volunteers couldn't come up with a "small" test suite to cover at For me the "one big test file" was anyway never of interest. Of course we cannot check all code path in a sty file in an automated way - this must be left for the author. What I want is * automated testing of simple load functionality * provide infrastructure (via l3build and maybe some extra infos) for package authros to provide more complete tests that can be part of an automated testing environment for TL > I would imagine that such a test suite should be able to raise an > alarm during the latest magyar.ldf update (even though I still argue Yes, typical case which a simple load test would exhibit. Same with many other case. Eg the recent shellesc which should have been loadable by plain tex, but wasn't. These things are testable in an automated way. We only need to start providing a easy to use infra and start writing a few tests. ... *only* ... Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From ziegenhagen at gmail.com Sun Jun 19 10:39:20 2016 From: ziegenhagen at gmail.com (Uwe Ziegenhagen) Date: Sun, 19 Jun 2016 10:39:20 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <20160619081716.GR14671@logic.at> References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> <20160619081716.GR14671@logic.at> Message-ID: <5DD3F71D-C62B-48D8-BAA5-1BA5256F8CAA@gmail.com> I have a few ideas, let me get back to You later today Von meinem iPhone gesendet Am 19.06.2016 um 10:17 schrieb Norbert Preining : >> It is certainly impossible to create a complete test suite that would >> catch any problem one could run into. >> >> But that doesn't mean (and shouldn't be an excuse) that a group of >> volunteers couldn't come up with a "small" test suite to cover at > > For me the "one big test file" was anyway never of interest. > > Of course we cannot check all code path in a sty file in an automated > way - this must be left for the author. > > What I want is > * automated testing of simple load functionality > * provide infrastructure (via l3build and maybe some extra infos) > for package authros to provide more complete tests that can > be part of an automated testing environment for TL > >> I would imagine that such a test suite should be able to raise an >> alarm during the latest magyar.ldf update (even though I still argue > > Yes, typical case which a simple load test would exhibit. > > Same with many other case. Eg the recent shellesc which should have > been loadable by plain tex, but wasn't. These things are > testable in an automated way. We only need to start providing > a easy to use infra and start writing a few tests. > > ... *only* ... > > Norbert > > -- > PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info > GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From zdenek.wagner at gmail.com Sun Jun 19 11:05:31 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Sun, 19 Jun 2016 11:05:31 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <20160619081716.GR14671@logic.at> References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> <20160619081716.GR14671@logic.at> Message-ID: 016-06-19 10:17 GMT+02:00 Norbert Preining : > > > It is certainly impossible to create a complete test suite that would > > catch any problem one could run into. > > > > But that doesn't mean (and shouldn't be an excuse) that a group of > > volunteers couldn't come up with a "small" test suite to cover at > > For me the "one big test file" was anyway never of interest. > > Of course we cannot check all code path in a sty file in an automated > way - this must be left for the author. > > What I want is > * automated testing of simple load functionality > * provide infrastructure (via l3build and maybe some extra infos) > for package authros to provide more complete tests that can > be part of an automated testing environment for TL I would be very happy to have such infrastructure. I have some tiny tests used during development of my packages but they are private, I do not include them in the version for CTAN. If you decide to offer suchan infreastructure, I can publish my tests that could be automated (some test files require visual check). > > > > I would imagine that such a test suite should be able to raise an > > alarm during the latest magyar.ldf update (even though I still argue > > Yes, typical case which a simple load test would exhibit. > > Same with many other case. Eg the recent shellesc which should have > been loadable by plain tex, but wasn't. These things are > testable in an automated way. We only need to start providing > a easy to use infra and start writing a few tests. > > ... *only* ... > > Norbert Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz 2 > > > -- > PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info > GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From zdenek.wagner at gmail.com Sun Jun 19 11:06:35 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Sun, 19 Jun 2016 11:06:35 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <5DD3F71D-C62B-48D8-BAA5-1BA5256F8CAA@gmail.com> References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> <20160619081716.GR14671@logic.at> <5DD3F71D-C62B-48D8-BAA5-1BA5256F8CAA@gmail.com> Message-ID: 2016-06-19 10:39 GMT+02:00 Uwe Ziegenhagen : > I have a few ideas, let me get back to You later today > I will also be interested. Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz > Von meinem iPhone gesendet > > Am 19.06.2016 um 10:17 schrieb Norbert Preining : > >>> It is certainly impossible to create a complete test suite that would >>> catch any problem one could run into. >>> >>> But that doesn't mean (and shouldn't be an excuse) that a group of >>> volunteers couldn't come up with a "small" test suite to cover at >> >> For me the "one big test file" was anyway never of interest. >> >> Of course we cannot check all code path in a sty file in an automated >> way - this must be left for the author. >> >> What I want is >> * automated testing of simple load functionality >> * provide infrastructure (via l3build and maybe some extra infos) >> for package authros to provide more complete tests that can >> be part of an automated testing environment for TL >> >>> I would imagine that such a test suite should be able to raise an >>> alarm during the latest magyar.ldf update (even though I still argue >> >> Yes, typical case which a simple load test would exhibit. >> >> Same with many other case. Eg the recent shellesc which should have >> been loadable by plain tex, but wasn't. These things are >> testable in an automated way. We only need to start providing >> a easy to use infra and start writing a few tests. >> >> ... *only* ... >> >> Norbert >> >> -- >> PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info >> GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From reinhard.kotucha at web.de Sun Jun 19 11:13:05 2016 From: reinhard.kotucha at web.de (Reinhard Kotucha) Date: Sun, 19 Jun 2016 11:13:05 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> Message-ID: <22374.25121.269329.71735@zaphod.ms25.net> On 2016-06-19 at 09:30:51 +0200, Zdenek Wagner wrote: > Hi all, > > the idea of automated tested was here more than a year ago and I > explained that I consider it impossible. Nobody knows which set of > packages will be used in a document created by each user. It is > only possible to create subsets of test files. Dear Zden?k, I fully agree with you. But I don't think that a test is useless only because it doesn't catch all possible errors. What I propose is to \let\uwe\proceed Zden?k, I'm certainly much more pessimistic than you, but I think that we should allow Uwe to prove that our assumptions are wrong. The only alternative I can imagine is to do nothing. Regards, Reinhard -- ------------------------------------------------------------------ Reinhard Kotucha Phone: +49-511-3373112 Marschnerstr. 25 D-30167 Hannover mailto:reinhard.kotucha at web.de ------------------------------------------------------------------ From ziegenhagen at gmail.com Sun Jun 19 21:00:25 2016 From: ziegenhagen at gmail.com (Uwe Ziegenhagen) Date: Sun, 19 Jun 2016 21:00:25 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <22374.25121.269329.71735@zaphod.ms25.net> References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> <22374.25121.269329.71735@zaphod.ms25.net> Message-ID: I created a Etherpad instance to collect some thoughts: I think that even visual checks might not be necessary if we calculate statistical summaries for a given output. https://piratenpad.de/p/TeXLiveTesting I commute each day approx. 2 hours, I'll start some design approach tomorrow- Uwe 2016-06-19 11:13 GMT+02:00 Reinhard Kotucha : > On 2016-06-19 at 09:30:51 +0200, Zdenek Wagner wrote: > > > Hi all, > > > > the idea of automated tested was here more than a year ago and I > > explained that I consider it impossible. Nobody knows which set of > > packages will be used in a document created by each user. It is > > only possible to create subsets of test files. > > Dear Zden?k, > I fully agree with you. But I don't think that a test is useless > only because it doesn't catch all possible errors. > > What I propose is to > > \let\uwe\proceed > > Zden?k, I'm certainly much more pessimistic than you, but I think that > we should allow Uwe to prove that our assumptions are wrong. > > The only alternative I can imagine is to do nothing. > > Regards, > Reinhard > > -- > ------------------------------------------------------------------ > Reinhard Kotucha Phone: +49-511-3373112 > Marschnerstr. 25 > D-30167 Hannover mailto:reinhard.kotucha at web.de > ------------------------------------------------------------------ > > -- Dr. Uwe Ziegenhagen -------------- next part -------------- An HTML attachment was scrubbed... URL: From zdenek.wagner at gmail.com Sun Jun 19 21:25:12 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Sun, 19 Jun 2016 21:25:12 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> <22374.25121.269329.71735@zaphod.ms25.net> Message-ID: Wouldn't it be better to use lua instead of python? Python is automatically installed on Linux but not on Windows, on the other hand lua is a part of TeX Live, hence it always exists. Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz 2016-06-19 21:00 GMT+02:00 Uwe Ziegenhagen : > I created a Etherpad instance to collect some thoughts: I think that even > visual checks might not be necessary if we calculate statistical summaries > for a given output. > > https://piratenpad.de/p/TeXLiveTesting > > I commute each day approx. 2 hours, I'll start some design approach > tomorrow- > > Uwe > > 2016-06-19 11:13 GMT+02:00 Reinhard Kotucha : >> >> On 2016-06-19 at 09:30:51 +0200, Zdenek Wagner wrote: >> >> > Hi all, >> > >> > the idea of automated tested was here more than a year ago and I >> > explained that I consider it impossible. Nobody knows which set of >> > packages will be used in a document created by each user. It is >> > only possible to create subsets of test files. >> >> Dear Zden?k, >> I fully agree with you. But I don't think that a test is useless >> only because it doesn't catch all possible errors. >> >> What I propose is to >> >> \let\uwe\proceed >> >> Zden?k, I'm certainly much more pessimistic than you, but I think that >> we should allow Uwe to prove that our assumptions are wrong. >> >> The only alternative I can imagine is to do nothing. >> >> Regards, >> Reinhard >> >> -- >> ------------------------------------------------------------------ >> Reinhard Kotucha Phone: +49-511-3373112 >> Marschnerstr. 25 >> D-30167 Hannover mailto:reinhard.kotucha at web.de >> ------------------------------------------------------------------ >> > > > > -- > Dr. Uwe Ziegenhagen > From ziegenhagen at gmail.com Sun Jun 19 21:30:12 2016 From: ziegenhagen at gmail.com (Uwe Ziegenhagen) Date: Sun, 19 Jun 2016 21:30:12 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> <22374.25121.269329.71735@zaphod.ms25.net> Message-ID: Maybe but my Lua Skills are non-existing, my Python skills improving on a daily basis. Let's see if it works at all, then we can find a way. Uwe 2016-06-19 21:25 GMT+02:00 Zdenek Wagner : > Wouldn't it be better to use lua instead of python? Python is > automatically installed on Linux but not on Windows, on the other hand > lua is a part of TeX Live, hence it always exists. > Zden?k Wagner > http://ttsm.icpf.cas.cz/team/wagner.shtml > http://icebearsoft.euweb.cz > > > 2016-06-19 21:00 GMT+02:00 Uwe Ziegenhagen : > > I created a Etherpad instance to collect some thoughts: I think that even > > visual checks might not be necessary if we calculate statistical > summaries > > for a given output. > > > > https://piratenpad.de/p/TeXLiveTesting > > > > I commute each day approx. 2 hours, I'll start some design approach > > tomorrow- > > > > Uwe > > > > 2016-06-19 11:13 GMT+02:00 Reinhard Kotucha : > >> > >> On 2016-06-19 at 09:30:51 +0200, Zdenek Wagner wrote: > >> > >> > Hi all, > >> > > >> > the idea of automated tested was here more than a year ago and I > >> > explained that I consider it impossible. Nobody knows which set of > >> > packages will be used in a document created by each user. It is > >> > only possible to create subsets of test files. > >> > >> Dear Zden?k, > >> I fully agree with you. But I don't think that a test is useless > >> only because it doesn't catch all possible errors. > >> > >> What I propose is to > >> > >> \let\uwe\proceed > >> > >> Zden?k, I'm certainly much more pessimistic than you, but I think that > >> we should allow Uwe to prove that our assumptions are wrong. > >> > >> The only alternative I can imagine is to do nothing. > >> > >> Regards, > >> Reinhard > >> > >> -- > >> ------------------------------------------------------------------ > >> Reinhard Kotucha Phone: +49-511-3373112 > >> Marschnerstr. 25 > >> D-30167 Hannover mailto:reinhard.kotucha at web.de > >> ------------------------------------------------------------------ > >> > > > > > > > > -- > > Dr. Uwe Ziegenhagen > > > -- Dr. Uwe Ziegenhagen -------------- next part -------------- An HTML attachment was scrubbed... URL: From d.p.carlisle at gmail.com Sun Jun 19 21:34:53 2016 From: d.p.carlisle at gmail.com (David Carlisle) Date: Sun, 19 Jun 2016 20:34:53 +0100 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> <22374.25121.269329.71735@zaphod.ms25.net> Message-ID: On 19 June 2016 at 20:30, Uwe Ziegenhagen wrote: > Maybe but my Lua Skills are non-existing, my Python skills improving on a > daily basis. > > Let's see if it works at all, then we can find a way. > > Uwe > the existing (lua) l3build framework could be used as a model of a cross platform tex test suite. Using texlua as the scripting engine has proved pretty successful in making testing and building packages work in all texlive supported environments. David From karl at freefriends.org Sun Jun 19 23:33:41 2016 From: karl at freefriends.org (Karl Berry) Date: Sun, 19 Jun 2016 21:33:41 +0000 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: Message-ID: <201606192133.u5JLXf7r019275@freefriends.org> All I have to add is that my previous attempt at automated testing (years and years ago) is in ./Build/tests. The directory names will give the idea: ./Build/tests/ Makefile dvi-latex0-small2e/ largefile/ README dvi-latex1-sample2e/ pdf-context0-hello/ asytestlibs.asy dvi-latex2-pdfprim/ pdf-context4-select/ checkdvi.pl* dvi-latex5-tugboat/ pdf-latex0-small2e/ checkpdf.pl* dvi-tex0-story/ pdf-tex0-story/ common.mak dvi-tex5-tugboat/ tryconcat/ My approach for PDF (DVI is easy) was to make images and compare them. Turned out to be impractical, not surprisingly. The advent of l3build and "reproducible builds" for PDF should make the idea much more viable. -k P.S. > https://piratenpad.de/p/TeXLiveTesting I don't understand why you put ideas in some read-only temporary url. Here is the text you wrote, for reading/archival on the mailing list like everything else in the thread. Thoughts regarding automated TeX LIVE Testing Requirements * must run on at least Mac OS, Windows and Linux, more platforms are appreciated * shall not require manual checks * shall be self contained * shall test a majority of user scenarios (letters, reports, books, etc.) * many small tests may be helpful, bigger test scenarios must not be omitted (I, for example, ran a 600 pages dissertation as part of my testing) * cannot test each and every scenario users may come across Design Ideas * use python (as I know it best) * make use of its unit test facilities * run a testcase by calling TeX engine n-times on specific file * check if the process ended properly * proper process ending does not mean that file is correct * Gather statistics on file: Size, number of pages, ... * Convert generated file to image and run imagemagick to collect statistical data. * Compare statistical data to known result. If theshold > level ==> testcase failed. From karl at freefriends.org Mon Jun 20 00:08:40 2016 From: karl at freefriends.org (Karl Berry) Date: Sun, 19 Jun 2016 22:08:40 +0000 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <22373.65436.46521.354964@zaphod.ms25.net> Message-ID: <201606192208.u5JM8edJ026549@freefriends.org> Another approach is to run TeX with the --recorder option on these files and evaluate the .fls files afterwards. For the record, some months ago I wrote a script ./Master/tlpkg/bin/check-fmttriggers to do this -- get package dependencies from .fls files -- for format builds. It's how the fmttriggers= settings in tlpdb are defined. I run it regularly since changes in the dependencies, while not frequent, certainly happen from time to time. If long lists of dependencies were being created for lots of packages, updating them would have to be automated also, one way or another. Certainly doable, no particular problem, just something to remember and not start down the road of doing it manually. Anyway, the hard part is setting up meaningful and re-runnable tests. Good luck. -k From zdenek.wagner at gmail.com Mon Jun 20 00:20:06 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Mon, 20 Jun 2016 00:20:06 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <201606192208.u5JM8edJ026549@freefriends.org> References: <22373.65436.46521.354964@zaphod.ms25.net> <201606192208.u5JM8edJ026549@freefriends.org> Message-ID: And I have just found that my xevlna.sty silently does not work ? because the number of character classes has changed... I will have to fix it... Such things can hardly be verified automatically... Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz 2016-06-20 0:08 GMT+02:00 Karl Berry : > Another approach is to run TeX with the --recorder > option on these files and evaluate the .fls files afterwards. > > For the record, some months ago I wrote a script > ./Master/tlpkg/bin/check-fmttriggers to do this -- get package > dependencies from .fls files -- for format builds. It's how the > fmttriggers= settings in tlpdb are defined. I run it regularly since > changes in the dependencies, while not frequent, certainly happen from > time to time. > > If long lists of dependencies were being created for lots of packages, > updating them would have to be automated also, one way or another. > Certainly doable, no particular problem, just something to remember and > not start down the road of doing it manually. > > Anyway, the hard part is setting up meaningful and re-runnable tests. > Good luck. -k From zarniwhoop at ntlworld.com Mon Jun 20 00:45:54 2016 From: zarniwhoop at ntlworld.com (Ken Moffat) Date: Sun, 19 Jun 2016 23:45:54 +0100 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <201606192133.u5JLXf7r019275@freefriends.org> References: <201606192133.u5JLXf7r019275@freefriends.org> Message-ID: <20160619224554.GB10411@milliways.localdomain> On Sun, Jun 19, 2016 at 09:33:41PM +0000, Karl Berry wrote: (replying to Uwe, I think) > All I have to add is that my previous attempt at automated testing > (years and years ago) is in ./Build/tests. The directory names will > give the idea: > > ./Build/tests/ > Makefile dvi-latex0-small2e/ largefile/ > README dvi-latex1-sample2e/ pdf-context0-hello/ > asytestlibs.asy dvi-latex2-pdfprim/ pdf-context4-select/ > checkdvi.pl* dvi-latex5-tugboat/ pdf-latex0-small2e/ > checkpdf.pl* dvi-tex0-story/ pdf-tex0-story/ > common.mak dvi-tex5-tugboat/ tryconcat/ > > My approach for PDF (DVI is easy) was to make images and compare them. > Turned out to be impractical, not surprisingly. Which has an impact on the points you've copied in below (thanks for doing that - it always helps if links can be read at some point in the future). > > The advent of l3build and "reproducible builds" for PDF should make the > idea much more viable. -k > > P.S. > > https://piratenpad.de/p/TeXLiveTesting > > I don't understand why you put ideas in some read-only temporary url. > Here is the text you wrote, for reading/archival on the mailing list > like everything else in the thread. > > Thoughts regarding automated TeX LIVE Testing > > Requirements > > * must run on at least Mac OS, Windows and Linux, more platforms are appreciated > * shall not require manual checks Nice if you can achieve it. > * shall be self contained I don't think I understand what you mean - e.g. you mention using ImageMagik which is a separate program, and probably there will be other external programs. > * shall test a majority of user scenarios (letters, reports, books, etc.) In my limited experience (and to be honest, I'm dubious about posting on this because my latex skills are so limited) the fun comes from various things within the texsphere - letters, reports, etc should be tested but they may be tangential to likely problems. I have my own tests for Beyond Linuxfromscratch which I mentioned the other day, but they are fairly minimal and do require manual review (I adapted my xindy tests to make sure the index did get created - but only after some changes in how _I_ built it caused failures). And they are definitely not useful in general although they should work on any 'nix-ish system with bash and a PDF viewer. > * many small tests may be helpful, bigger test scenarios must not be omitted (I, for example, ran a 600 pages dissertation as part of my testing) I know that CPU power continues to increase, and tests can be run in parallel if you can keep the details separate, but for repeated testing smaller is probably better. > * cannot test each and every scenario users may come across > Indeed, that is why people have to test for themselves. > > Design Ideas > > * use python (as I know it best) > * make use of its unit test facilities For a proof of concept, whatever you are familiar with. > * run a testcase by calling TeX engine n-times on specific file This I do not understand. I understand using the engine the required number of times - and interspersing that with calls to other parts of TeX such as xindy or asymptote - but I'm not at all clear what you are proposing. > * check if the process ended properly I don't know about your preferred python (I try to avoid scripting languages where whitespace is important ;) but when my own tests run from a Makefile hang in TeX I have to key 'x' to stop. And sometimes TeX reports normal status but part of the test did not work. > * proper process ending does not mean that file is correct > * Gather statistics on file: Size, number of pages, ... > * Convert generated file to image and run imagemagick to collect statistical data. This goes back to what Karl wrote above - what guarantee is there that the version of ImageMagick you use today will give the same results as the next version you happen to use ? Also, I have no experience in using ImageMagick to split a multi-page PDF into individual page images, but I suspect there might be a large overhead and perhaps other tools such as qpdf may be better for that. > * Compare statistical data to known result. If theshold > level ==> testcase failed. I'm not a statistician, but I suspect that the degree of acceptable variation will probably differ for each test. Interesting ideas, and don't let me discourage you. ?en -- War is Peace Freedom is Slavery Ignorance is Strength From werner at suse.de Mon Jun 20 09:13:26 2016 From: werner at suse.de (Dr. Werner Fink) Date: Mon, 20 Jun 2016 09:13:26 +0200 Subject: [tex-live] TL2016: latexpand perl script is fragile as it depends on perl version Message-ID: <20160620071326.k65vtndvfzp2auvy@boole.suse.de> Hi, ... lines like my ($_, $prefix, $file) = @_; in sub process_line or if (!$keep_comments && $line =~ /^[ \t]*\\end{document}/) { in sub process_file do not work well with all perl versions out there, compare with https://bugzilla.opensuse.org/show_bug.cgi?id=983823 Werner -- "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From zdenek.wagner at gmail.com Mon Jun 20 11:08:05 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Mon, 20 Jun 2016 11:08:05 +0200 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <20160619224554.GB10411@milliways.localdomain> References: <201606192133.u5JLXf7r019275@freefriends.org> <20160619224554.GB10411@milliways.localdomain> Message-ID: 2016-06-20 0:45 GMT+02:00 Ken Moffat : > On Sun, Jun 19, 2016 at 09:33:41PM +0000, Karl Berry wrote: > > (replying to Uwe, I think) > >> All I have to add is that my previous attempt at automated testing >> (years and years ago) is in ./Build/tests. The directory names will >> give the idea: >> >> ./Build/tests/ >> Makefile dvi-latex0-small2e/ largefile/ >> README dvi-latex1-sample2e/ pdf-context0-hello/ >> asytestlibs.asy dvi-latex2-pdfprim/ pdf-context4-select/ >> checkdvi.pl* dvi-latex5-tugboat/ pdf-latex0-small2e/ >> checkpdf.pl* dvi-tex0-story/ pdf-tex0-story/ >> common.mak dvi-tex5-tugboat/ tryconcat/ >> >> My approach for PDF (DVI is easy) was to make images and compare them. >> Turned out to be impractical, not surprisingly. > > Which has an impact on the points you've copied in below (thanks for > doing that - it always helps if links can be read at some point in > the future). >> >> The advent of l3build and "reproducible builds" for PDF should make the >> idea much more viable. -k >> >> P.S. >> > https://piratenpad.de/p/TeXLiveTesting >> >> I don't understand why you put ideas in some read-only temporary url. >> Here is the text you wrote, for reading/archival on the mailing list >> like everything else in the thread. >> >> Thoughts regarding automated TeX LIVE Testing >> >> Requirements >> >> * must run on at least Mac OS, Windows and Linux, more platforms are appreciated >> * shall not require manual checks > > Nice if you can achieve it. > >> * shall be self contained > > > I don't think I understand what you mean - e.g. you mention using > ImageMagik which is a separate program, and probably there will be > other external programs. > I hope I understand this idea. It replaces the need of visual check. Suppose you have an output from TeX Live "year-1" and the output from the latest versions. Both have equal number of pages. You can now convert the pages to images and use ImageMagick to subtract them. Ideally the resul should be white pages. You can them make a test to distinguish different pages from rounding errors. >> * shall test a majority of user scenarios (letters, reports, books, etc.) > > In my limited experience (and to be honest, I'm dubious about > posting on this because my latex skills are so limited) the fun > comes from various things within the texsphere - letters, reports, > etc should be tested but they may be tangential to likely problems. > > I have my own tests for Beyond Linuxfromscratch which I mentioned > the other day, but they are fairly minimal and do require manual > review (I adapted my xindy tests to make sure the index did get > created - but only after some changes in how _I_ built it caused > failures). And they are definitely not useful in general although > they should work on any 'nix-ish system with bash and a PDF viewer. > There is a philosophy around the trip and trap tests made by D. E. Knuth. He tests a lot of things that can be verified automatically, does not do such tests that require visual check because users will spot it during normal work. >> * many small tests may be helpful, bigger test scenarios must not be omitted (I, for example, ran a 600 pages dissertation as part of my testing) > > I know that CPU power continues to increase, and tests can be run in > parallel if you can keep the details separate, but for repeated > testing smaller is probably better. > Minimal tests are much better not becaus they run faster but because they clearly say what went wrong. >> * cannot test each and every scenario users may come across >> > Indeed, that is why people have to test for themselves. > >> >> Design Ideas >> >> * use python (as I know it best) >> * make use of its unit test facilities > > For a proof of concept, whatever you are familiar with. > >> * run a testcase by calling TeX engine n-times on specific file > > This I do not understand. I understand using the engine the > required number of times - and interspersing that with calls to > other parts of TeX such as xindy or asymptote - but I'm not at all > clear what you are proposing. > >> * check if the process ended properly > > I don't know about your preferred python (I try to avoid scripting > languages where whitespace is important ;) but when my own tests run > from a Makefile hang in TeX I have to key 'x' to stop. And > sometimes TeX reports normal status but part of the test did not > work. > I have the same feeling, I do not like a language where a white space is an important element. I do not need python for my work (I just have it installed) but as you could see in another thread, even perl scripts distributed with TL do not work in Windows. I would not be surprized if a Windows version of python were incomplete. The existence of lua in TL motivates me to learn lua, not python. The status code is not the best way to use in such tests. It will be necessary to analyze the log file. The test files can be designed in such a way that the analysis of the log should not be difficult. >> * proper process ending does not mean that file is correct >> * Gather statistics on file: Size, number of pages, ... >> * Convert generated file to image and run imagemagick to collect statistical data. > > This goes back to what Karl wrote above - what guarantee is there > that the version of ImageMagick you use today will give the same > results as the next version you happen to use ? > There are basic operations that must always work. First the number of pages must match and the page size must match, if not, the output is not equivalent. Unless there is a bug in ImageMagick, subtraction must work and calculation of non-white pixels must work. It is not necessary to use sophisticated filters that may (and probably will) depend on the version of ImageMagick. Anyway, this test need not be the most important. > Also, I have no experience in using ImageMagick to split a > multi-page PDF into individual page images, but I suspect there > might be a large overhead and perhaps other tools such as qpdf may > be better for that. > >> * Compare statistical data to known result. If theshold > level ==> testcase failed. > > I'm not a statistician, but I suspect that the degree of acceptable > variation will probably differ for each test. > Not for the text but for the font. If you use the same resolution, the roundoff errors in both documents will be the same but if the font is updated and the glyphs are modified, you will get a different output even if the metrics of the font remained unchanged. It might be useful to subtract the images and before counting the non-white pixels apply a filter based on algorithms of mathematical gnostics. It will remove noise but retain patterns that signal real differences. > Interesting ideas, and don't let me discourage you. > > ?en Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz From beebe at math.utah.edu Mon Jun 20 14:07:09 2016 From: beebe at math.utah.edu (Nelson H. F. Beebe) Date: Mon, 20 Jun 2016 06:07:09 -0600 Subject: [tex-live] mirroring beebe's binaries In-Reply-To: Message-ID: >> This no longer works. >> rsync://ftp.math.utah.edu/texlive/bin is no longer there. That is because of a small change that I made yesterday after prompting from Karl Berry about an inconsistency in naming. % rsync rsync://ftp.math.utah.edu/ CTAN all of ftp://ctan.tug.org/ [huge: > 31GB] bib TeX User Group bibliography archive [> 5GB] bibnet BibNet Project bibliography archive [about 0.3GB] historic archive of past TeX distributions and files [huge: > 150GB] texlive all of ftp://ftp.math.utah.edu/pub/ctan/tex-archive/systems/texlive [> 6GB] texlive-pretest all of ftp://ftp.math.utah.edu/pub/texlive-pretest/ [> 4GB] TUHS The Unix Heritage Society (TUHS) mirror [> 3GB] Thus, you can pull the pretest tree with the binaries over like this: rsync -a rsync://ftp.math.utah.edu/texlive-pretest texlive-pretest The pretest tree is not intended to be a part of TeX Live, just a place where I can record experiences from building the system on many machines, and deposit binaries for them. ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - University of Utah FAX: +1 801 581 4148 - - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - ------------------------------------------------------------------------------- From chuck.newman at hpe.com Mon Jun 20 17:00:58 2016 From: chuck.newman at hpe.com (Newman, Chuck) Date: Mon, 20 Jun 2016 15:00:58 +0000 Subject: [tex-live] Error installing install-tl-unx.tar.gz In-Reply-To: <20160619053111.GP14671@logic.at> References: <20160618140846.GK14671@logic.at> <20160619053111.GP14671@logic.at> Message-ID: Norbert -- The server is on a hidden network (10. address) behind a server inside our corporate environment. I suppose there are several ways the process could break down. I am able to install .rpms from the web using yum, so not everything is broken. I've decided instead to install from DVD, so am downloading the .iso image from my target machine: firefox 'http://mirror.unl.edu/ctan/systems/texlive/Images/texlive.iso' I expect that should work well. If not, I'll install on a different server and move /usr/local/texlive over to my target server. Thank you for your quick replies. -- Chuck Newman -----Original Message----- From: Norbert Preining [mailto:preining at logic.at] Sent: Sunday, 19 June, 2016 01:31 To: Newman, Chuck Cc: texlive at tug.org Subject: Re: [tex-live] Error installing install-tl-unx.tar.gz Hi CHuck, again, I am seriously believe that something in your network setup is broken. It seems that the proxy needs some login, or whatever. > cannot contact mirror.ctan.org, returning a backbone server! First of all, this is *strange*. It means that you cannot connect. So my guess is the following: Your proxy allows only a certain set of "User Agents" - that means that a User Agent of LWP or so is blocked. That would explain If you have lwp-download installed, can you try lwp-download http://mirror.ctan.org and send the output of it, probably saved into "index" named file. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From joseph.wright at morningstar2.co.uk Mon Jun 20 17:51:57 2016 From: joseph.wright at morningstar2.co.uk (Joseph Wright) Date: Mon, 20 Jun 2016 16:51:57 +0100 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> <22374.25121.269329.71735@zaphod.ms25.net> Message-ID: <96054516-5df1-6a2d-526c-dde892deebc0@morningstar2.co.uk> On 19/06/2016 20:34, David Carlisle wrote: > On 19 June 2016 at 20:30, Uwe Ziegenhagen wrote: >> Maybe but my Lua Skills are non-existing, my Python skills improving on a >> daily basis. >> >> Let's see if it works at all, then we can find a way. >> >> Uwe >> > > the existing (lua) l3build framework could be used as a model of a > cross platform tex test suite. > > Using texlua as the scripting engine has proved pretty successful in > making testing and building > packages work in all texlive supported environments. > > David I'd strongly suggest taking a look over l3build before anyone expends effort in scripting something new. Whilst it's about more than just testing, the core idea goes back a long way (see http://tug.org/TUGboat/tb35-3/tb111mitt-l3build.pdf). As detailed in http://mirrors.ctan.org/macros/latex/contrib/l3build/l3build.pdf, it does log-based testing with Lua doing the 'grind' and uses system utilities for the diff step (if required). (The team and so l3build needs to work cross-platform, so this works on *nix and Windows.) As Norbert has noted, we've recently added PDF-based comparisons as this will be potentially-useful for cases where log-based methods are not so useful. I'd also point to the automated testing we've set up: https://travis-ci.org/latex3/latex3. (Travis-CI is one of many choices, this is just a convenient way of running the tests on a VM we don't have to maintain.) In amongst our own tests there are a small number for third-party code (https://github.com/latex3/latex3/tree/master/contrib). I talked about this at the UK-TUG meeting last year: http://river-valley.zeeba.tv/why-does-it-always-fail-on-me-automating-testing-using-travis-ci/. Joseph From karl at freefriends.org Tue Jun 21 01:23:18 2016 From: karl at freefriends.org (Karl Berry) Date: Mon, 20 Jun 2016 23:23:18 +0000 Subject: [tex-live] TL2016: latexpand perl script is fragile as it depends on perl version In-Reply-To: <20160620071326.k65vtndvfzp2auvy@boole.suse.de> Message-ID: <201606202323.u5KNNIMb019003@freefriends.org> Hi Werner, Subject: [tex-live] TL2016: latexpand perl script is fragile as it depends I wish Perl had not made that backward-incompatible change, but anyway, latexpand is not maintained by TL, we just import the package from CTAN. You can find the maintainer email at the top of the script. --best, karl. From karl at freefriends.org Tue Jun 21 01:35:42 2016 From: karl at freefriends.org (Karl Berry) Date: Mon, 20 Jun 2016 23:35:42 +0000 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <96054516-5df1-6a2d-526c-dde892deebc0@morningstar2.co.uk> Message-ID: <201606202335.u5KNZgRM021161@freefriends.org> > * Convert generated file to image and run imagemagick to collect > * statistical data. FWIW,I have done some image comparisons over the years using imagemagick (there are some excellent web pages from anthony thyssen about it), and I am extremely skeptical that statistical image comparison will work. jw> I'd strongly suggest taking a look over l3build FWIW, I fully agree. l3build is the way! km> using ImageMagick to split a multi-page PDF into individual page images Imagemagick just delegates to gs for PDF operations, as far as I know. Personally, I'd just use gs directly. It could also be done with pdf(la)tex itself, e.g., with the pdfjam scripts for convenience. If it's really needed to split PDF output in the first place, which is certainly premature to be discussing. km> when my own tests run from a Makefile hang in TeX I have to key 'x' Aside, but FWIW, alternatives: tex '\nonstopmode\input ...' or -interaction=nonstopmode, or References: <20160618055918.GH14671@logic.at> <22373.65436.46521.354964@zaphod.ms25.net> <22374.25121.269329.71735@zaphod.ms25.net> Message-ID: <22376.34422.711429.520771@zaphod.ms25.net> On 2016-06-19 at 21:30:12 +0200, Uwe Ziegenhagen wrote: > Maybe but my Lua Skills are non-existing, my Python skills > improving on a daily basis. Vice versa here. However, I've bought a book about Python some time ago and have the impression that the main concepts are the same or at least very similar. Thus, you benefit a lot from your knowledge about Python. Since Lua is a small language, there is much less you have to learn. As other people mentioned already, TeXLua has two important advantages: it's the most portable programming language in the TeX world, and the LaTeX team is using TeXLua already for regression tests. The latter is an important point because the infrastructure exists already. Uwe, I know that it is an extra burden to learn two programming languages at the same time. But Lua is much less complex than Python. The definite source is the book "Programming in Lua", see: https://www.lua.org/docs.html This book is nicely written, easy to understand, and though it's not very verbose at all, it contains everything you need. The current version of the Reference Manual is available online (see link above). If you just want to know how things are done in Lua which you already can do in Python, the Lua Tutorial is quite helpful: http://lua-users.org/wiki/TutorialDirectory You definitely don't have to study all tutorials. Some of them explain things you never need. > Let's see if it works at all, then we can find a way. I'm convinced that if you provide a Python script and someone converts it to Lua, you will write your next script in Lua anyway because you recognize that it's quite easy. ;) Please consult the tutorial in order to compare Python with Lua. Your knowledge about Python is helpful because the concepts are very similar. And if you have a question, don't hesitate to ask. In my opinion TeXLua is by far the best approach. Regards, Reinhard -- ------------------------------------------------------------------ Reinhard Kotucha Phone: +49-511-3373112 Marschnerstr. 25 D-30167 Hannover mailto:reinhard.kotucha at web.de ------------------------------------------------------------------ From Gunner.Gewiss at posteo.de Mon Jun 20 23:42:35 2016 From: Gunner.Gewiss at posteo.de (=?UTF-8?Q?Gunner_Gewi=c3=9f?=) Date: Mon, 20 Jun 2016 23:42:35 +0200 Subject: [tex-live] ERROR: cannot build FMT files since update to TeXLive 2016 (Linux: OpenSuSE) Message-ID: <730f2b8f-1fd4-2844-0c11-38457c9d2d4f@posteo.de> Dear all, my system is OpenSuSE Leap 42.1, with the following repositories as its base: . Since my update to TeXLive 2016 I cannot build needed FMT files any more. The log says "! I can't find file `pdftexconfig.tex'." and "! I can't find file `pdftexconfig.tex'." as well as "! I can't find file `loadhyph-af.tex'." This did not happen before. Any ideas, any help? Very appreciated. TIA (many thanks). Thank you and kind regards, Gunner From rpspringuel at gmail.com Tue Jun 21 03:42:40 2016 From: rpspringuel at gmail.com (R. Padraic Springuel) Date: Mon, 20 Jun 2016 21:42:40 -0400 Subject: [tex-live] General test suite for TeX-Live In-Reply-To: <201606202335.u5KNZgRM021161@freefriends.org> References: <201606202335.u5KNZgRM021161@freefriends.org> Message-ID: <5328D804-69C7-42FC-826B-8F2DF7DF8AF7@gmail.com> As an example of using imagemagik to compare pdf outputs, people might want to take a look at the gregorio-test repository on GitHub (https://github.com/gregorio-project/gregorio-test). While our infrastructure is bash based, we use imagemagik to check the output from using our package against a previously produced known good model. Furthermore, by specifying the paper size and a font which is provided locally in the test repository, we're able to do this across both Mac and Linux systems running in both the US and Europe. ???????????????????????? Br. Samuel, OSB (R. Padraic Springuel) PAX ? ??????? On Jun 20, 2016, at 7:35 PM, Karl Berry wrote: >> * Convert generated file to image and run imagemagick to collect >> * statistical data. > > FWIW,I have done some image comparisons over the years using imagemagick > (there are some excellent web pages from anthony thyssen about it), and > I am extremely skeptical that statistical image comparison will work. > > jw> I'd strongly suggest taking a look over l3build > > FWIW, I fully agree. l3build is the way! > > km> using ImageMagick to split a multi-page PDF into individual page images > > Imagemagick just delegates to gs for PDF operations, as far as I know. > Personally, I'd just use gs directly. It could also be done with > pdf(la)tex itself, e.g., with the pdfjam scripts for convenience. > If it's really needed to split PDF output in the first place, which is > certainly premature to be discussing. > > km> when my own tests run from a Makefile hang in TeX I have to key 'x' > > Aside, but FWIW, alternatives: > tex '\nonstopmode\input ...' > or -interaction=nonstopmode, > or > (I run TeX from a Makefile 99% of the time, and I never want interaction.) > > k -------------- next part -------------- An HTML attachment was scrubbed... URL: From preining at logic.at Tue Jun 21 07:40:47 2016 From: preining at logic.at (Norbert Preining) Date: Tue, 21 Jun 2016 14:40:47 +0900 Subject: [tex-live] ERROR: cannot build FMT files since update to TeXLive 2016 (Linux: OpenSuSE) In-Reply-To: <730f2b8f-1fd4-2844-0c11-38457c9d2d4f@posteo.de> References: <730f2b8f-1fd4-2844-0c11-38457c9d2d4f@posteo.de> Message-ID: Hi Gunner This looks like a SuSE packaged version of TeX Live - we cannot help here a lot as this mailing list is for upstream TeX Live. I guess it is best for you to send a bug report via the usual SuSE channels. It simply looks like some files are missing in the packaging. All the best Norbert On June 21, 2016 6:42:35 AM GMT+09:00, "Gunner Gewi?" wrote: >Dear all, > >my system is OpenSuSE Leap 42.1, with the following repositories as its >base: >. > >Since my update to TeXLive 2016 I cannot build needed FMT files any >more. > >The log > >says >"! I can't find file `pdftexconfig.tex'." >and >"! I can't find file `pdftexconfig.tex'." >as well as >"! I can't find file `loadhyph-af.tex'." > >This did not happen before. > >Any ideas, any help? Very appreciated. TIA (many thanks). > >Thank you and kind regards, > >Gunner -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From Gunner.Gewiss at posteo.de Tue Jun 21 10:08:38 2016 From: Gunner.Gewiss at posteo.de (=?UTF-8?Q?Gunner_Gewi=c3=9f?=) Date: Tue, 21 Jun 2016 10:08:38 +0200 Subject: [tex-live] ERROR: cannot build FMT files since update to TeXLive 2016 (Linux: OpenSuSE) In-Reply-To: References: <730f2b8f-1fd4-2844-0c11-38457c9d2d4f@posteo.de> Message-ID: SOLVED! :-) I declared the TeXLive installation on my OpenSuSE as broken and completely re-installed it. Now, it works. Sorry for any inconvenience. Gunner Am 21.06.2016 um 07:40 schrieb Norbert Preining: > I guess it is best for you to send a bug report via the usual SuSE channels. It simply looks like some files are missing in the packaging. > On June 21, 2016 6:42:35 AM GMT+09:00, "Gunner Gewi?" wrote: >> Since my update to TeXLive 2016 I cannot build needed FMT files any >> more. From norbert at preining.info Tue Jun 21 10:15:03 2016 From: norbert at preining.info (Norbert Preining) Date: Tue, 21 Jun 2016 17:15:03 +0900 Subject: [tex-live] TeX Live Testing - call for participants Message-ID: <20160621081503.GG14671@logic.at> Dear all, to get something started on a more substantial level, I have created an organization at github called TeX-Live https://github.com/TeX-Live and within that a first reposiotry tltesting https://github.com/TeX-Live/tltesting by now there is only a minimal README, but I have collected my and Uwe's ideas plus other things that went over the line here. I invite everyone interested to apply for membership in this organization and we can work on tltesting together. Since the ideas page is a wiki pafge, everyone can add new ideas and design proposals for now. In the next days I will upload simple examples for l3build testing of loading and maybe pdf testing. Hope that helps Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From norbert at preining.info Tue Jun 21 10:23:29 2016 From: norbert at preining.info (Norbert Preining) Date: Tue, 21 Jun 2016 17:23:29 +0900 Subject: [tex-live] TeX Live Testing - call for participants In-Reply-To: <20160621081503.GG14671@logic.at> References: <20160621081503.GG14671@logic.at> Message-ID: <20160621082329.GI14671@logic.at> Hi all, forgot the Wiki URL: On Tue, 21 Jun 2016, Norbert Preining wrote: > Dear all, > > to get something started on a more substantial level, I have created > an organization at github called > TeX-Live > https://github.com/TeX-Live > and within that a first reposiotry > tltesting > https://github.com/TeX-Live/tltesting > by now there is only a minimal README, but I have collected my > and Uwe's ideas plus other things that went over the line here. ... into a Wiki page: https://github.com/TeX-Live/tltesting/wiki > I invite everyone interested to apply for membership in this > organization and we can work on tltesting together. > > Since the ideas page is a wiki pafge, everyone can add new ideas > and design proposals for now. > > In the next days I will upload simple examples for l3build testing > of loading and maybe pdf testing. Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From zdenek.wagner at gmail.com Tue Jun 21 11:07:04 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Tue, 21 Jun 2016 11:07:04 +0200 Subject: [tex-live] TeX Live Testing - call for participants In-Reply-To: <20160621082329.GI14671@logic.at> References: <20160621081503.GG14671@logic.at> <20160621082329.GI14671@logic.at> Message-ID: 2016-06-21 10:23 GMT+02:00 Norbert Preining : > Hi all, forgot the Wiki URL: > Hi all, I will certainly be interested. It will even help me to learn git (I have to learn it anyway). > On Tue, 21 Jun 2016, Norbert Preining wrote: >> Dear all, >> >> to get something started on a more substantial level, I have created >> an organization at github called >> TeX-Live >> https://github.com/TeX-Live >> and within that a first reposiotry >> tltesting >> https://github.com/TeX-Live/tltesting >> by now there is only a minimal README, but I have collected my >> and Uwe's ideas plus other things that went over the line here. > > ... into a Wiki page: > https://github.com/TeX-Live/tltesting/wiki > >> I invite everyone interested to apply for membership in this >> organization and we can work on tltesting together. >> >> Since the ideas page is a wiki pafge, everyone can add new ideas >> and design proposals for now. >> >> In the next days I will upload simple examples for l3build testing >> of loading and maybe pdf testing. > > Norbert > > -- > PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info > GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz From beebe at math.utah.edu Tue Jun 21 23:46:47 2016 From: beebe at math.utah.edu (Nelson H. F. Beebe) Date: Tue, 21 Jun 2016 15:46:47 -0600 Subject: [tex-live] rsync to Utah TeX Live 2016 binaries, redux Message-ID: I have made what I hope will be a final change in the rsync configuration for the Utah TeX Live 2016 binaries: % rsync rsync://ftp.math.utah.edu/ ... texlive-utah all of ftp://ftp.math.utah.edu/pub/texlive-utah/ [> 4GB] ... The path component texlive-utah is a symlink to the local directory texlive, so the tree can also be accessed via FTP and Web clients as ftp://ftp.math.utah.edu/pub/texlive/ http://www.math.utah.edu/pub/texlive/ http://www.math.utah.edu/pub/texlive-utah/ In retrospect, it was a mistake on my part to have named the local pretest work directory texlive, but that has now been recorded by at least the Google search engine, so it cannot be changed. As far as I know, our University does not pay any of the search engine companies for more frequent global indexing, and it seems that new additions to our HTTP trees may wait for several weeks before they are discovered by search engines. ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - University of Utah FAX: +1 801 581 4148 - - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - ------------------------------------------------------------------------------- From ttk at t-lab.opal.ne.jp Wed Jun 22 16:33:08 2016 From: ttk at t-lab.opal.ne.jp (TANAKA Takuji) Date: Wed, 22 Jun 2016 23:33:08 +0900 Subject: [tex-live] TL2016: Stack smashed in upmendex fprint_uchar() function In-Reply-To: <20160615081358.cut6em4gwgaspywb@boole.suse.de> References: <20160614163550.xoedjzx7wp27oq2b@boole.suse.de> <20160615060816.GI10246@logic.at> <20160615081358.cut6em4gwgaspywb@boole.suse.de> Message-ID: Hi Werner, I think r.41497 & r.41498 have fixed the issue. Takuji On 2016/06/15 17:13, Dr. Werner Fink wrote: >> >>> during test I've seen a catched stack smash done by glibc fortify. >>> It seems that in texk/upmendex/fwrite.c the function fprint_uchar() >>> will be overflowed by the UChar inital array in indwrite() >> From norbert at preining.info Wed Jun 22 17:10:41 2016 From: norbert at preining.info (Norbert Preining) Date: Thu, 23 Jun 2016 00:10:41 +0900 Subject: [tex-live] TL2016: Stack smashed in upmendex fprint_uchar() function In-Reply-To: References: <20160614163550.xoedjzx7wp27oq2b@boole.suse.de> <20160615060816.GI10246@logic.at> <20160615081358.cut6em4gwgaspywb@boole.suse.de> Message-ID: <20160622151041.GP840@logic.at> > I think r.41497 & r.41498 have fixed the issue. I can confirm that the Debian packages where I included these two patches are now building and all test succeeding on all architectures. Thanks a lot Norbert -- PREINING Norbert + TeX Live & Debian Developer + http://www.preining.info GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From siepo at cybercomm.nl Thu Jun 23 11:41:00 2016 From: siepo at cybercomm.nl (Siep Kroonenberg) Date: Thu, 23 Jun 2016 11:41:00 +0200 Subject: [tex-live] texcount on Windows fail because of Perl file missing from TeX Live distribution In-Reply-To: <20160618103811.GA2920@bitmuis.thuis.internal> References: <306071466104496@web16h.yandex.ru> <20160618103811.GA2920@bitmuis.thuis.internal> Message-ID: <20160623094100.GA3303@bitmuis.nl> On Sat, Jun 18, 2016 at 12:38:11PM +0200, Siep Kroonenberg wrote: > On Thu, Jun 16, 2016 at 10:14:56PM +0300, Leonid Sinev wrote: > > Greetings, > 1. This module includes a dll which would need to be added too. > > 2. MetaCPAN https://metacpan.org/pod/Win32::Console::ANSI lists a > dependency Win32::Pipe which is not in tlperl either > > As Karl noted, the included perl for windows does not aim to > support all included perl scripts. Maybe some user of texcount should contact the author of the script about making the ANSI modules optional. -- Siep Kroonenberg From zarniwhoop at ntlworld.com Fri Jun 24 09:32:12 2016 From: zarniwhoop at ntlworld.com (Ken Moffat) Date: Fri, 24 Jun 2016 08:32:12 +0100 Subject: [tex-live] Bad perms in 2016 texmf ? Message-ID: <20160624073212.GA28723@milliways.localdomain> Maybe this is me (I'm not at my best any more), but after my first attempt to install from source for 2016 crapped out when I tried to use context (unreadable directories/files for a user when mtxrun was invoked) I compared it to the binary I had been using : several directories in the texmf tarball seem to not be usable by users, and a number of files are similarly not readable. I'm currently rebuilding (for that, installing in /opt is nice) and will be trying to change all directories in texmf-dist to 775 and all files to 644. Meanwhile I have noted the following by untarring the texmf tarball : texmf-dist/doc/latex/acroterm 750 texmf-dist/doc/fonts/source/lh/*/* 600 texmf-dist/doc/latex/acroterm/* 640 some texmf-dist/doc/latex/*/README are 700 gmiflink 600 gmverse NB in my binary install gmverse/README was 664! texmf-dist/makeindex/gmdoc/gmglo.ist 700 texmf-dist/source/fonts/lh/*.mfj 600 texmf-dist/source/fonts/lh/tex/*.t* 600 texmf-dist/source/fonts/lh/tex/wrk/*/* 600 texmf-dist/source/latex/acroterm/* 640 texmf-dist/source/latex/lh/* - my notes on this directory look flakey (like me at the moment), but I think some of the files are 600 and some 640. texmf-dist/tex/latex/acroterm 750 texmf-dist/tex/latex/lh 700 texmf-dist/tex/latex/acroterm/* { acroterm.sty ) 640 texmf-dist/tex/latex/lh/* 600 for files and 700 for directories texmf-dist/tex/plain/lh 700 texmf-dist/tex/plain/lh/* (*.tex) 600 I *hope* this is me again doing something wrong, but for the moment I think texmf this year has bad perms. ?en -- Yes, you CAN fool most of the people some of the time. From karl at freefriends.org Sat Jun 25 00:37:02 2016 From: karl at freefriends.org (Karl Berry) Date: Fri, 24 Jun 2016 22:37:02 +0000 Subject: [tex-live] Bad perms in 2016 texmf ? In-Reply-To: <20160624073212.GA28723@milliways.localdomain> Message-ID: <201606242237.u5OMb2Jp003615@freefriends.org> I *hope* this is me again doing something wrong, but for the moment I think texmf this year has bad perms. You're speaking of texlive-20160523-texmf.tar.xz, I presume. From tar tvf, looks like you are right. Sigh. Will check the others and fix all when I have a chance. --thanks, karl. From zarniwhoop at ntlworld.com Sat Jun 25 01:34:11 2016 From: zarniwhoop at ntlworld.com (Ken Moffat) Date: Sat, 25 Jun 2016 00:34:11 +0100 Subject: [tex-live] Bad perms in 2016 texmf ? In-Reply-To: <201606242237.u5OMb2Jp003615@freefriends.org> References: <20160624073212.GA28723@milliways.localdomain> <201606242237.u5OMb2Jp003615@freefriends.org> Message-ID: <20160624233411.GA28044@milliways.localdomain> On Fri, Jun 24, 2016 at 10:37:02PM +0000, Karl Berry wrote: > I *hope* this is me again doing something wrong, but for the moment > I think texmf this year has bad perms. > > You're speaking of texlive-20160523-texmf.tar.xz, I presume. From tar > tvf, looks like you are right. Sigh. Will check the others and fix all > when I have a chance. --thanks, karl. I've had a bad day with this (forcing *all* texmf files to 644 was a bad idea because I had overlooked the scripts). For the moment I'm rebuilding again, using the following commands post-install : find /opt/texlive/${TEXYEAR}/texmf-dist -type d -perm 2750 -exec \ chmod -v 755 {} \; find /opt/texlive/${TEXYEAR}/texmf-dist -type f -perm 744 -exec \ chmod -v 644 {} \; find /opt/texlive/${TEXYEAR}/texmf-dist -type f -perm 600 -o -perm \ 640 -exec chmod -v 644 {} \; The first catches a few directories such as doc/latex/acroterm where context (mtxrun) initially blew up when I ran it as a user. I guess that only a root install (and perhaps a root untar) preserves the sgid bit. The second gets over 2000 files where the perms look dodgy but I doubt they cause any real harm. The last one is where I keep making mistakes / typos : it is for files like texmf-dist/source/fonts/lh/*.mfj texmf-dist/source/fonts/lh/tex/*.t* texmf-dist/source/fonts/lh/tex/wrk/*/* (all 600) texmf-dist/source/latex/acroterm/* and texmf-dist/source/latex/lh/* (mostly 640) I think I will still be failing to address texmf-dist/makeindex/gmdoc/gmglo.ist which got installed as 700 instead of 644. ?en -- Yes, you CAN fool most of the people some of the time. From reinhard.kotucha at web.de Sat Jun 25 03:02:18 2016 From: reinhard.kotucha at web.de (Reinhard Kotucha) Date: Sat, 25 Jun 2016 03:02:18 +0200 Subject: [tex-live] Bad perms in 2016 texmf ? In-Reply-To: <20160624233411.GA28044@milliways.localdomain> References: <20160624073212.GA28723@milliways.localdomain> <201606242237.u5OMb2Jp003615@freefriends.org> <20160624233411.GA28044@milliways.localdomain> Message-ID: <22381.55322.168113.967958@zaphod.ms25.net> On 2016-06-25 at 00:34:11 +0100, Ken Moffat wrote: > I've had a bad day... If your internet connection isn't extraordinary slow, it's better to throw everything away and install TeX Live again. Fixing file permissions manually is time-consuming and you can never be sure that everything you've done is correct. Regards, Reinhard -- ------------------------------------------------------------------ Reinhard Kotucha Phone: +49-511-3373112 Marschnerstr. 25 D-30167 Hannover mailto:reinhard.kotucha at web.de ------------------------------------------------------------------ From zarniwhoop at ntlworld.com Sat Jun 25 04:16:14 2016 From: zarniwhoop at ntlworld.com (Ken Moffat) Date: Sat, 25 Jun 2016 03:16:14 +0100 Subject: [tex-live] Bad perms in 2016 texmf ? In-Reply-To: <22381.55322.168113.967958@zaphod.ms25.net> References: <20160624073212.GA28723@milliways.localdomain> <201606242237.u5OMb2Jp003615@freefriends.org> <20160624233411.GA28044@milliways.localdomain> <22381.55322.168113.967958@zaphod.ms25.net> Message-ID: <20160625021614.GA31374@milliways.localdomain> On Sat, Jun 25, 2016 at 03:02:18AM +0200, Reinhard Kotucha wrote: > On 2016-06-25 at 00:34:11 +0100, Ken Moffat wrote: > > > I've had a bad day... > > If your internet connection isn't extraordinary slow, it's better to > throw everything away and install TeX Live again. > > Fixing file permissions manually is time-consuming and you can never > be sure that everything you've done is correct. > > Regards, > Reinhard I'm installing from source (I moved the initial binary install out of the way after I had modified both my asymptote tests to work as expected on a system where all the system libs were built with gcc6). For a binary install, yes, reinstall (hopefully, with better downloads after an update). But some of us prefer building from source. And we get all the pleasure of discovering new breakages when system packages change (not just gcc or libstdc++, I think it was grep which caused problems in xindy in the last year or so). Fortunately I have the moved binary files to compare (and they are fine). At the moment I'm reinstalling again (I've lost count of the times I've done this, but at least my i7 is quite quick) : find -perm 600 -o -perm 640 -o -perm 700 -exec chmod ... is only chmodding the files matched by the last test. Retrying with xargs. ?en -- Yes, you CAN fool most of the people some of the time. From texlive at schoepfer.info Sat Jun 25 09:20:00 2016 From: texlive at schoepfer.info (texlive at schoepfer.info) Date: Sat, 25 Jun 2016 09:20:00 +0200 Subject: [tex-live] Bad perms in 2016 texmf ? In-Reply-To: <20160625021614.GA31374@milliways.localdomain> References: <20160624073212.GA28723@milliways.localdomain> <201606242237.u5OMb2Jp003615@freefriends.org> <20160624233411.GA28044@milliways.localdomain> <22381.55322.168113.967958@zaphod.ms25.net> <20160625021614.GA31374@milliways.localdomain> Message-ID: <6b2c8289a3a64ec7787ad1d06983dc59@schoepfer.info> > At the moment I'm reinstalling again (I've lost count of the times > I've done this, but at least my i7 is quite quick) : > > find -perm 600 -o -perm 640 -o -perm 700 -exec chmod ... > > is only chmodding the files matched by the last test. Retrying with > xargs. Maybe chmod -R u+w,go-w,a+rX-st PATH does the trick for you. Regards, Johannes From zarniwhoop at ntlworld.com Sat Jun 25 19:10:00 2016 From: zarniwhoop at ntlworld.com (Ken Moffat) Date: Sat, 25 Jun 2016 18:10:00 +0100 Subject: [tex-live] Bad perms in 2016 texmf ? In-Reply-To: <6b2c8289a3a64ec7787ad1d06983dc59@schoepfer.info> References: <20160624073212.GA28723@milliways.localdomain> <201606242237.u5OMb2Jp003615@freefriends.org> <20160624233411.GA28044@milliways.localdomain> <22381.55322.168113.967958@zaphod.ms25.net> <20160625021614.GA31374@milliways.localdomain> <6b2c8289a3a64ec7787ad1d06983dc59@schoepfer.info> Message-ID: <20160625171000.GC18696@milliways.localdomain> On Sat, Jun 25, 2016 at 09:20:00AM +0200, texlive at schoepfer.info wrote: > > At the moment I'm reinstalling again (I've lost count of the times > > I've done this, but at least my i7 is quite quick) : > > > > find -perm 600 -o -perm 640 -o -perm 700 -exec chmod ... > > > > is only chmodding the files matched by the last test. Retrying with > > xargs. > > Maybe > chmod -R u+w,go-w,a+rX-st PATH > does the trick for you. > I'm told I needed parentheses because there is an implicit AND before the -exec, with a higher priority than OR. But yes, recursive chmod like that might be simpler. ?en -- Yes, you CAN fool most of the people some of the time. From vincent.torri at gmail.com Sun Jun 26 22:01:50 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Sun, 26 Jun 2016 22:01:50 +0200 Subject: [tex-live] memory leak with kpathsea_find_file() ? Message-ID: hello my little library uses kpathsea_find_file(). The actual code to retrieve the full path of a TFM file is the following: ------------- kpathsea kpse; char *n; kpse = kpathsea_new(); kpathsea_set_program_name(kpse, "kpsewhich", NULL); kpathsea_init_prog (kpse, "LIBDVI", 300, NULL, NULL); n = kpathsea_find_file(kpse, name, kpse_tfm_format, 1); kpathsea_finish(kpse); ------------- Drmemory (or valgrind) reports the following unfreed memory : # 1 ll!xmalloc +0x16 # 2 ll!hash_create +0x11 # 3 ll!concatn +0x8a0 # 4 ll!kpathsea_cnf_get +0x2e # 5 ll!kpathsea_var_value +0xa5 # 6 ll!kpse_set_suffixes +0x54 # 7 ll!kpathsea_init_format_return_varlist+0x1f0 # 8 ll!kpathsea_find_file_generic +0x70 # 9 ll!kpathsea_find_file +0x1b #10 dvi_kpathsea_path_name_get [src/lib/dvi_kpathsea.c:354] #11 dvi_font_define [src/lib/dvi_font.c:332] trying to free 'n' with free() leads to errors in drmemory (or valgrind). Is this unfreed memory normal ? or my code is not fully correct ? thank you Vincent Torri -------------- next part -------------- An HTML attachment was scrubbed... URL: From karl at freefriends.org Sun Jun 26 23:48:23 2016 From: karl at freefriends.org (Karl Berry) Date: Sun, 26 Jun 2016 15:48:23 -0600 Subject: [tex-live] memory leak with kpathsea_find_file() ? In-Reply-To: (message from Vincent Torri on Sun, 26 Jun 2016 22:01:50 +0200) Message-ID: <86oa6naafs.fsf@frenzy.freefriends.org> n = kpathsea_find_file(kpse, name, kpse_tfm_format, 1); ... trying to free 'n' with free() leads to errors I don't understand that. As far as I can recall, kpathsea_find_file should always return newly-malloced memory, aside perhaps from pathological cases, certainly not yours. Is this unfreed memory normal ? Yes, it is normal. I never tried to make kpse 100% memory efficient. A few lost bytes did not matter in 1986, and they matter even less now. or my code is not fully correct ? Your code looks ok to me. -k From karl at freefriends.org Mon Jun 27 00:31:14 2016 From: karl at freefriends.org (Karl Berry) Date: Sun, 26 Jun 2016 22:31:14 +0000 Subject: [tex-live] TeXLive 2016 source tar ball with object and library files? In-Reply-To: <20160607114939.2e6qzkxeex5l4ite@boole.suse.de> Message-ID: <201606262231.u5QMVEQB018440@freefriends.org> Hi Werner, texlive-20160523-source.tar.xz and recoginzed that the tar ball is much larger then those of 2015 (252126K to 44395K) due a lot of binary, object, and library files. Sorry for the delay. I have finally rebuilt texlive-20160523-source.tar.xz without the useless bloat, now at ftp://tug.org/historic/systems/texlive/2016/texlive-20160523b-source.tar.xz (and .sha512 and .sha512.asc). The content should be the same, only the excess files removed. This happened because I accidentally/stupidly made one last-minute rebuild out of my working directory instead of the clean reference tree. (Which is also the cause of some, though not all, of the permission problems in the texmf tarball noted by Ken M. Sigh.) Let me know if problems, of course ... -k From ul at openlilylib.org Mon Jun 27 01:06:44 2016 From: ul at openlilylib.org (Urs Liska) Date: Mon, 27 Jun 2016 01:06:44 +0200 Subject: [tex-live] Directory structure for package including files Message-ID: <2810541d-3a91-ddde-6418-4f143d4fa44f@openlilylib.org> Hi all, I'm mentoring a GSoC student who will as part of his project produce a LaTeX package and upload it to CTAN. (As some background that is not directly relevant to the question: This package will take sets of LaTeX commands generated from annotations in LilyPond music scores and typeset a critical report from them.) The package will provide extensive configuration options for the visual appearance of the output, mainly implemented through \setkeys. As the definition of the *default* appearance will use a significant number of code lines the student suggested to move that to a separate file, e.g. `default-stylesheet.inp` (which I find good as it's also a good example file for end-users). The question is: Should we take any precautions regarding the directory structure when it comes to uploading the package to CTAN? Of course in the context of a CTAN package it's probably the most straightforward to just put it in the same folder as the .sty file. But is that also a good approach when we expect the package to be incorporated in distributions like tex-live? Best Urs From karl at freefriends.org Mon Jun 27 03:01:07 2016 From: karl at freefriends.org (Karl Berry) Date: Mon, 27 Jun 2016 01:01:07 +0000 Subject: [tex-live] Bad perms in 2016 texmf ? In-Reply-To: <20160624233411.GA28044@milliways.localdomain> Message-ID: <201606270101.u5R117x7031179@freefriends.org> Hi Ken - you probably have no reason to spend more time on it, but FYI, I've recreated the texmf tarball as texlive-20160523b-texmf.tar.xz in ftp://tug.org/historic/systems/texlive/2016/ (also .sha512 and .asc). Hopefully the permissions are more reasonable. I found some Japanese property list (.pl) files that were spuriously executable due to the conflict with .pl for Perl. I fixed those. Aside from that, under doc/ and source/, there may well be some scripts which are not executable (especially if they don't have an extension), but I've never worried about that. (People can always invoke them "manually".) I'm more interested in fixing files that are executable and shouldn't be, if you find any of those. (gmglo.ist is fixed.) Thanks/sorry, Karl From vincent.torri at gmail.com Mon Jun 27 07:30:10 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Mon, 27 Jun 2016 07:30:10 +0200 Subject: [tex-live] memory leak with kpathsea_find_file() ? In-Reply-To: <86oa6naafs.fsf@frenzy.freefriends.org> References: <86oa6naafs.fsf@frenzy.freefriends.org> Message-ID: On Sun, Jun 26, 2016 at 11:48 PM, Karl Berry wrote: > > n = kpathsea_find_file(kpse, name, kpse_tfm_format, 1); > ... > trying to free 'n' with free() leads to errors > > I don't understand that. As far as I can recall, kpathsea_find_file > should always return newly-malloced memory, aside perhaps from > pathological cases, certainly not yours. if I free the result given by kpathsea_find_file(), after kpathsea_finish(), i get : Error #1: INVALID HEAP ARGUMENT to free 0x03bec028 # 0 replace_free [d:\drmemory_package\common\alloc_replace.c:2706] # 1 libdvi-0.dll!dvi_font_define [src/lib/dvi_font.c:342] # 2 libdvi-0.dll!dvi_interpreter_fnt_def_parse [src/lib/dvi_interpreter.c:505] # 3 libdvi-0.dll!dvi_interpreter_postamble_parse [src/lib/dvi_interpreter.c:369] # 4 libdvi-0.dll!dvi_document_new [src/lib/dvi_document.c:153] # 5 main [src/bin/dvi.c:16] Note: @0:00:27.083 in thread 2704 Note: refers to -1 byte(s) before next malloc Note: next higher malloc: 0x03bec028-0x03bec07f Note: refers to -87 byte(s) beyond last valid byte in prior malloc Note: prev lower malloc: 0x03bec028-0x03bec07f The code is equivalent to this: kpathsea kpse; char *n; kpse = kpathsea_new(); kpathsea_set_program_name(kpse, "kpsewhich", NULL); kpathsea_init_prog (kpse, "LIBDVI", 300, NULL, NULL); n = kpathsea_find_file(kpse, name, kpse_tfm_format, 1); kpathsea_finish(kpse); /* i do something */ free(n); Vincent Torri > > Is this unfreed memory normal ? > > Yes, it is normal. I never tried to make kpse 100% memory efficient. > A few lost bytes did not matter in 1986, and they matter even less now. > > or my code is not fully correct ? > > Your code looks ok to me. -k From vincent.torri at gmail.com Mon Jun 27 07:36:35 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Mon, 27 Jun 2016 07:36:35 +0200 Subject: [tex-live] question about kpathsea_new() and kpathsea_finish(), and thread safety Message-ID: Hello I'm using these 2 functions in my library. Is it reasonnable to call kpathsea_new() when my library starts and kpathsea_finish() when my library finishes, and in between i can use all the kpathsea_* functions I want ? Another question: is kpathsea thread safe and reentrant ? I want to find function names and glyphs of several fonts in different threads, one thread per font. thank you Vincent Torri From kakuto at fuk.kindai.ac.jp Mon Jun 27 08:48:35 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Mon, 27 Jun 2016 15:48:35 +0900 Subject: [tex-live] question about kpathsea_new() and kpathsea_finish(), and thread safety In-Reply-To: References: Message-ID: > Another question: is kpathsea thread safe and reentrant ? On Unix platforms, I believe that kpathsea is reentrant, thanks to Taco. On windows, unfortunately, there are, probably, two global variables, which Peter and I had left for simplicity. Best, Akira From zdenek.wagner at gmail.com Mon Jun 27 10:47:50 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Mon, 27 Jun 2016 10:47:50 +0200 Subject: [tex-live] Directory structure for package including files In-Reply-To: <2810541d-3a91-ddde-6418-4f143d4fa44f@openlilylib.org> References: <2810541d-3a91-ddde-6418-4f143d4fa44f@openlilylib.org> Message-ID: Hi Urs, first of all you should read the CTAN guidelines. CTAN itself is not primarily used for installation, the contents must be easily browsable. The maintainers require flat structure, everything in the single directory. If the package is too complex, it is possible to separate the files to a directory containing the macros, a directory containing documentation etc. It is explained on CTAN. In addition, you can upload a TDS zip file. In TDS (used in nowadays TeX distros) the macros for your package will reside in $TEXMF/tex/latex/your_package_name, hence it is not necessary to split it to other directories unless you need different files for different engines but having the same name. Anyway, even without the TDS zip file the TeX Live maintainers can cope with it. As each package must contain a README file, this is the best place where to describe the structure of the package so that the distro maintainers know how to install it. Hope this helps. Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz 2016-06-27 1:06 GMT+02:00 Urs Liska
    : > Hi all, > > I'm mentoring a GSoC student who will as part of his project produce a > LaTeX package and upload it to CTAN. (As some background that is not > directly relevant to the question: This package will take sets of LaTeX > commands generated from annotations in LilyPond music scores and typeset > a critical report from them.) > > The package will provide extensive configuration options for the visual > appearance of the output, mainly implemented through \setkeys. As the > definition of the *default* appearance will use a significant number of > code lines the student suggested to move that to a separate file, e.g. > `default-stylesheet.inp` (which I find good as it's also a good example > file for end-users). > > The question is: Should we take any precautions regarding the directory > structure when it comes to uploading the package to CTAN? Of course in > the context of a CTAN package it's probably the most straightforward to > just put it in the same folder as the .sty file. But is that also a good > approach when we expect the package to be incorporated in distributions > like tex-live? > > Best > Urs > From ul at openlilylib.org Mon Jun 27 10:54:17 2016 From: ul at openlilylib.org (Urs Liska) Date: Mon, 27 Jun 2016 10:54:17 +0200 Subject: [tex-live] Directory structure for package including files In-Reply-To: References: <2810541d-3a91-ddde-6418-4f143d4fa44f@openlilylib.org> Message-ID: <0ce86c19-1d9a-5904-5986-daedefe778d8@openlilylib.org> Hi Zdenek, thank you for the explanation. When I published my last package things were more complicated (providing additional Python scripts, font files etc.) so I probably didn't think of it being that simple and "flat" in simpler cases. But then I was strongly discouraged to provide a TDS file. Have recommendations changed in the last few years? Urs Am 27.06.2016 um 10:47 schrieb Zdenek Wagner: > Hi Urs, > > first of all you should read the CTAN guidelines. CTAN itself is not > primarily used for installation, the contents must be easily > browsable. The maintainers require flat structure, everything in the > single directory. If the package is too complex, it is possible to > separate the files to a directory containing the macros, a directory > containing documentation etc. It is explained on CTAN. In addition, > you can upload a TDS zip file. In TDS (used in nowadays TeX distros) > the macros for your package will reside in > $TEXMF/tex/latex/your_package_name, hence it is not necessary to split > it to other directories unless you need different files for different > engines but having the same name. Anyway, even without the TDS zip > file the TeX Live maintainers can cope with it. As each package must > contain a README file, this is the best place where to describe the > structure of the package so that the distro maintainers know how to > install it. > > Hope this helps. > Zden?k Wagner > http://ttsm.icpf.cas.cz/team/wagner.shtml > http://icebearsoft.euweb.cz > > > 2016-06-27 1:06 GMT+02:00 Urs Liska
      : >> Hi all, >> >> I'm mentoring a GSoC student who will as part of his project produce a >> LaTeX package and upload it to CTAN. (As some background that is not >> directly relevant to the question: This package will take sets of LaTeX >> commands generated from annotations in LilyPond music scores and typeset >> a critical report from them.) >> >> The package will provide extensive configuration options for the visual >> appearance of the output, mainly implemented through \setkeys. As the >> definition of the *default* appearance will use a significant number of >> code lines the student suggested to move that to a separate file, e.g. >> `default-stylesheet.inp` (which I find good as it's also a good example >> file for end-users). >> >> The question is: Should we take any precautions regarding the directory >> structure when it comes to uploading the package to CTAN? Of course in >> the context of a CTAN package it's probably the most straightforward to >> just put it in the same folder as the .sty file. But is that also a good >> approach when we expect the package to be incorporated in distributions >> like tex-live? >> >> Best >> Urs >> From zdenek.wagner at gmail.com Mon Jun 27 11:08:31 2016 From: zdenek.wagner at gmail.com (Zdenek Wagner) Date: Mon, 27 Jun 2016 11:08:31 +0200 Subject: [tex-live] Directory structure for package including files In-Reply-To: <0ce86c19-1d9a-5904-5986-daedefe778d8@openlilylib.org> References: <2810541d-3a91-ddde-6418-4f143d4fa44f@openlilylib.org> <0ce86c19-1d9a-5904-5986-daedefe778d8@openlilylib.org> Message-ID: Hi Urs, the current instructions for CTAN upload are here: http://www.ctan.org/upload If you upload a package in TDS only, the CTAN team would not like it. TeX Live maintainers do not need the TDS zip file, they have a script to put the files to correct directories. Zden?k Wagner http://ttsm.icpf.cas.cz/team/wagner.shtml http://icebearsoft.euweb.cz 2016-06-27 10:54 GMT+02:00 Urs Liska
        : > Hi Zdenek, > > thank you for the explanation. > When I published my last package things were more complicated (providing > additional Python scripts, font files etc.) so I probably didn't think > of it being that simple and "flat" in simpler cases. > > But then I was strongly discouraged to provide a TDS file. Have > recommendations changed in the last few years? > > Urs > > > Am 27.06.2016 um 10:47 schrieb Zdenek Wagner: >> Hi Urs, >> >> first of all you should read the CTAN guidelines. CTAN itself is not >> primarily used for installation, the contents must be easily >> browsable. The maintainers require flat structure, everything in the >> single directory. If the package is too complex, it is possible to >> separate the files to a directory containing the macros, a directory >> containing documentation etc. It is explained on CTAN. In addition, >> you can upload a TDS zip file. In TDS (used in nowadays TeX distros) >> the macros for your package will reside in >> $TEXMF/tex/latex/your_package_name, hence it is not necessary to split >> it to other directories unless you need different files for different >> engines but having the same name. Anyway, even without the TDS zip >> file the TeX Live maintainers can cope with it. As each package must >> contain a README file, this is the best place where to describe the >> structure of the package so that the distro maintainers know how to >> install it. >> >> Hope this helps. >> Zden?k Wagner >> http://ttsm.icpf.cas.cz/team/wagner.shtml >> http://icebearsoft.euweb.cz >> >> >> 2016-06-27 1:06 GMT+02:00 Urs Liska
          : >>> Hi all, >>> >>> I'm mentoring a GSoC student who will as part of his project produce a >>> LaTeX package and upload it to CTAN. (As some background that is not >>> directly relevant to the question: This package will take sets of LaTeX >>> commands generated from annotations in LilyPond music scores and typeset >>> a critical report from them.) >>> >>> The package will provide extensive configuration options for the visual >>> appearance of the output, mainly implemented through \setkeys. As the >>> definition of the *default* appearance will use a significant number of >>> code lines the student suggested to move that to a separate file, e.g. >>> `default-stylesheet.inp` (which I find good as it's also a good example >>> file for end-users). >>> >>> The question is: Should we take any precautions regarding the directory >>> structure when it comes to uploading the package to CTAN? Of course in >>> the context of a CTAN package it's probably the most straightforward to >>> just put it in the same folder as the .sty file. But is that also a good >>> approach when we expect the package to be incorporated in distributions >>> like tex-live? >>> >>> Best >>> Urs >>> > From karl at freefriends.org Mon Jun 27 23:37:55 2016 From: karl at freefriends.org (Karl Berry) Date: Mon, 27 Jun 2016 21:37:55 +0000 Subject: [tex-live] Directory structure for package including files In-Reply-To: <0ce86c19-1d9a-5904-5986-daedefe778d8@openlilylib.org> Message-ID: <201606272137.u5RLbthO013528@freefriends.org> But then I was strongly discouraged to provide a TDS file. I continue to strongly discourage people from making a .tds.zip file unless their package is very very complicated. Have recommendations changed in the last few years? Nope. In short: for a small package like this, it is best to put all files in one directory for CTAN, omit files derived from the .dtx (if it uses .dtx), and no .tds.zip. Good luck with the project! -k From reinhard.kotucha at web.de Tue Jun 28 00:32:42 2016 From: reinhard.kotucha at web.de (Reinhard Kotucha) Date: Tue, 28 Jun 2016 00:32:42 +0200 Subject: [tex-live] Directory structure for package including files In-Reply-To: <0ce86c19-1d9a-5904-5986-daedefe778d8@openlilylib.org> References: <2810541d-3a91-ddde-6418-4f143d4fa44f@openlilylib.org> <0ce86c19-1d9a-5904-5986-daedefe778d8@openlilylib.org> Message-ID: <22385.43402.899210.927664@zaphod.ms25.net> On 2016-06-27 at 10:54:17 +0200, Urs Liska wrote: > Hi Zdenek, > > thank you for the explanation. > When I published my last package things were more complicated (providing > additional Python scripts, font files etc.) so I probably didn't think > of it being that simple and "flat" in simpler cases. > > But then I was strongly discouraged to provide a TDS file. Have > recommendations changed in the last few years? Not within the last few years. More than a decade ago TDS files were preferred. But Karl Berry said that if a TDS file isn't 100% correct, it's not helpful at all and even causes much more extra work. However, IMO it's very helpful if there is one level of subdirectories on CTAN, for instance doc/, tex/, source/, ... I suppose that most CTAN users are interested in documentation only. It's also quite helpful to have distinct directories for macros and documentation because this allows the author to make clear whether a particular file is only part of the documentation (has to be found by texdoc) or must be accessible at run-time (has to be found by TeX). VnTeX, for instance, comes with a lot of test files which can't even be compiled easily on all platforms. I prefer to have these files in source/ instead of doc/. With a completely flat hierarchy (only one directory for everything) all my intentions certainly get lost. A flat hierarchy (only one level) is quite helpful IMO in order to preserve the author's intents. This is also described at http://www.ctan.org/upload but I think that instead of distinguishing between small and very large packages, it makes more sense to assume that your package is very large if particular files in your package have to be installed at certain places. If there are no such ambiguities you can assume that your package is small. Regards, Reinhard -- ------------------------------------------------------------------ Reinhard Kotucha Phone: +49-511-3373112 Marschnerstr. 25 D-30167 Hannover mailto:reinhard.kotucha at web.de ------------------------------------------------------------------ From karl at freefriends.org Tue Jun 28 00:50:53 2016 From: karl at freefriends.org (Karl Berry) Date: Mon, 27 Jun 2016 22:50:53 +0000 Subject: [tex-live] memory leak with kpathsea_find_file() ? In-Reply-To: Message-ID: <201606272250.u5RMorAS021379@freefriends.org> kpathsea_new() when my library starts and kpathsea_finish() when my library finishes, and in between i can use all the kpathsea_* functions I want ? Sounds reasonable to me. Error #1: INVALID HEAP ARGUMENT to free 0x03bec028 Do you check if n is NULL? Does it make a difference if the free is before the kpathsea_finish call? (kpathsea_finish should be nearly a no-op, but ...) Are you using a distro's kpathsea or the original from TL? I have no idea what distros do to kpathsea, but I'm pretty sure they don't leave it alone. Aside from those trivialities, I don't have drmemory and can't reproduce it with valgrind if I set name = "cmr10.tfm". So I don't know what to say. All I can suggest is running it under the debugger to see what's going on. Good luck, Karl From kakuto at fuk.kindai.ac.jp Tue Jun 28 06:46:49 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Tue, 28 Jun 2016 13:46:49 +0900 Subject: [tex-live] memory leak with kpathsea_find_file() ? In-Reply-To: <201606272250.u5RMorAS021379@freefriends.org> References: <201606272250.u5RMorAS021379@freefriends.org> Message-ID: > n = kpathsea_find_file(kpse, name, kpse_tfm_format, 1); > ... > trying to free 'n' with free() leads to errors On windows, if I link by using the import library, I cannot reproduce the error. However if I use the LoadLibrary() method, I can reproduce the error. I don't understand why. Presumably my DLL lacks something necessary, though it has been useful if used with the import library. Best, Akira From vincent.torri at gmail.com Tue Jun 28 07:30:58 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Tue, 28 Jun 2016 07:30:58 +0200 Subject: [tex-live] memory leak with kpathsea_find_file() ? In-Reply-To: References: <201606272250.u5RMorAS021379@freefriends.org> Message-ID: On Tue, Jun 28, 2016 at 6:46 AM, Akira Kakuto wrote: >> n = kpathsea_find_file(kpse, name, kpse_tfm_format, 1); >> ... >> trying to free 'n' with free() leads to errors > > > On windows, if I link by using the import library, where is the import library in the texlive 2016 tree ?? > I cannot reproduce the error. > However if I use the LoadLibrary() method, that's what i use as i can't find the import library > I can reproduce > the error. I don't understand why. Presumably my DLL lacks > something necessary, though it has been useful if used with > the import library. you can seee all the exported symbols in the DLL with dependency walker in casee it helps you Vincent Torri From vincent.torri at gmail.com Tue Jun 28 08:05:59 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Tue, 28 Jun 2016 08:05:59 +0200 Subject: [tex-live] memory leak with kpathsea_find_file() ? In-Reply-To: References: <201606272250.u5RMorAS021379@freefriends.org> Message-ID: On Tue, Jun 28, 2016 at 6:46 AM, Akira Kakuto wrote: >> n = kpathsea_find_file(kpse, name, kpse_tfm_format, 1); >> ... >> trying to free 'n' with free() leads to errors > > > On windows, if I link by using the import library, > I cannot reproduce the error. > However if I use the LoadLibrary() method, I can reproduce another problem: as I use LoadLibrary and as texlive provides only a 32bits DLL, i have an error if my library is compiled in 64bits mode: error 193 (ERROR_BAD_EXE_FORMAT): "%1 is not a valid Win32 application." Vincent Torri > the error. I don't understand why. Presumably my DLL lacks > something necessary, though it has been useful if used with > the import library. > > Best, > Akira > From vincent.torri at gmail.com Tue Jun 28 08:09:04 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Tue, 28 Jun 2016 08:09:04 +0200 Subject: [tex-live] kpathsea PDF document has wrong table of contents Message-ID: Hello kpathsea.pdf (in texlive\2016\texmf-dist\doc\kpathsea) seems to have a wrong table of contents (or the table of contents is correct but the content is wrong) Vincent Torri From kakuto at fuk.kindai.ac.jp Tue Jun 28 08:18:49 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Tue, 28 Jun 2016 15:18:49 +0900 Subject: [tex-live] memory leak with kpathsea_find_file() ? In-Reply-To: References: <201606272250.u5RMorAS021379@freefriends.org> Message-ID: <50623A2575FC4FD383E44CD988155DCC@CJ3001517A> Dear Vincent, > where is the import library ... I already sent you off-list kpathsea.tar.gz, which contains 32bit and 64bit DLLs and import libs. Please check spam mails also, due to the attachment. Best, Akira From vincent.torri at gmail.com Tue Jun 28 12:30:02 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Tue, 28 Jun 2016 12:30:02 +0200 Subject: [tex-live] memory leak with kpathsea_find_file() ? In-Reply-To: <50623A2575FC4FD383E44CD988155DCC@CJ3001517A> References: <201606272250.u5RMorAS021379@freefriends.org> <50623A2575FC4FD383E44CD988155DCC@CJ3001517A> Message-ID: On Tue, Jun 28, 2016 at 8:18 AM, Akira Kakuto wrote: > Dear Vincent, > >> where is the import library ... > > > I already sent you off-list kpathsea.tar.gz, which contains > 32bit and 64bit DLLs and import libs. sorry, but I have no private mail from you. Only a mail with a link to get the 64 bits dll and exe (CTAN/systems/win32/w32tex/TLW64/) > Please check spam mails also, due to the attachment. nor in spam mails. Can you resend it please ? thank you Vincent Torri From P.Taylor at Rhul.Ac.Uk Tue Jun 28 20:00:22 2016 From: P.Taylor at Rhul.Ac.Uk (Philip Taylor) Date: Tue, 28 Jun 2016 19:00:22 +0100 Subject: [tex-live] TeX Live 2016 release available In-Reply-To: <201606062109.u56L96k4020573@freefriends.org> References: <201606062109.u56L96k4020573@freefriends.org> Message-ID: <6281de42-3cfa-b591-492b-1fcecc71b799@Rhul.Ac.Uk> On 22/07/2011 at 00:07, Karl Berry wrote: > Date: Thu, 21 Jul 2011 23:07:45 +0000 > From: Karl Berry > To: , > For the sake of anyone who doesn't follow ctan-ann ... > > TeX Live 2016 has been released. Could I just check the date of this announcement ? According to the message headers, TeX Live 2016 was announced as released on Thu, 21 Jul 2011 at 23:07:45 +0000; however, the first "Received: " header does not appear until Mon, 6 Jun 2016 at 21:09:06 GMT. Where was the message lurking for five years, and how was Karl able so accurately to predict the future ?! -- Philip Taylor -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: Signature.jpg Type: image/jpeg Size: 12104 bytes Desc: not available URL: From jacobgyntelberg at me.com Tue Jun 28 20:50:57 2016 From: jacobgyntelberg at me.com (Jacob Gyntelberg) Date: Tue, 28 Jun 2016 20:50:57 +0200 Subject: [tex-live] TeX Live 2016 release available In-Reply-To: <6281de42-3cfa-b591-492b-1fcecc71b799@Rhul.Ac.Uk> References: <201606062109.u56L96k4020573@freefriends.org> <6281de42-3cfa-b591-492b-1fcecc71b799@Rhul.Ac.Uk> Message-ID: <6D4428BB-DC3F-43AD-9BC9-18DB0FE0EA82@me.com> Karl rocks :-) Sent from my iPhone On 28 Jun 2016, at 20:00, Philip Taylor wrote: On 22/07/2011 at 00:07, Karl Berry wrote: > Date: Thu, 21 Jul 2011 23:07:45 +0000 > From: Karl Berry > To: , > For the sake of anyone who doesn't follow ctan-ann ... > > TeX Live 2016 has been released. Could I just check the date of this announcement ? According to the message headers, TeX Live 2016 was announced as released on Thu, 21 Jul 2011 at 23:07:45 +0000; however, the first "Received: " header does not appear until Mon, 6 Jun 2016 at 21:09:06 GMT. Where was the message lurking for five years, and how was Karl able so accurately to predict the future ?! -- Philip Taylor -------------- next part -------------- An HTML attachment was scrubbed... URL: From kakuto at fuk.kindai.ac.jp Tue Jun 28 23:15:47 2016 From: kakuto at fuk.kindai.ac.jp (Akira Kakuto) Date: Wed, 29 Jun 2016 06:15:47 +0900 Subject: [tex-live] memory leak with kpathsea_find_file() ? In-Reply-To: References: <201606272250.u5RMorAS021379@freefriends.org> <50623A2575FC4FD383E44CD988155DCC@CJ3001517A> Message-ID: <922532B9924C4F99B9A45EAC6EBEE641@CJ3001517A> Dear Vincent, > sorry, but I have no private mail from you. You can obtain it at http://members2.jcom.home.ne.jp/wt1357ak/kpathsea2016.zip Best, Akira From karl at freefriends.org Tue Jun 28 23:36:42 2016 From: karl at freefriends.org (Karl Berry) Date: Tue, 28 Jun 2016 21:36:42 +0000 Subject: [tex-live] kpathsea PDF document has wrong table of contents In-Reply-To: Message-ID: <201606282136.u5SLagkb004616@freefriends.org> Subject: [tex-live] kpathsea PDF document has wrong table of contents So it does. Strange. I'll regenerate it and check the others that I did similarly. Thanks for the report. -k From karl at freefriends.org Tue Jun 28 23:46:27 2016 From: karl at freefriends.org (Karl Berry) Date: Tue, 28 Jun 2016 21:46:27 +0000 Subject: [tex-live] Directory structure for package including files In-Reply-To: <22385.43402.899210.927664@zaphod.ms25.net> Message-ID: <201606282146.u5SLkRNL005487@freefriends.org> all my intentions certainly get lost. In the case when an author has a straightforward-enough package but wants certain files in certain places, it is fine to either include an installation note in the README, or with the CTAN upload (they pass it on to me), or, if I install it wrong, report the problem and I will fix it. Not an issue. Much better than making a .tds.zip just to "work around" my bug. Notwithstanding, I think .tds.zip is fine and appropriate for vntex. -k From reinhard.kotucha at web.de Wed Jun 29 01:54:01 2016 From: reinhard.kotucha at web.de (Reinhard Kotucha) Date: Wed, 29 Jun 2016 01:54:01 +0200 Subject: [tex-live] Directory structure for package including files In-Reply-To: <201606282146.u5SLkRNL005487@freefriends.org> References: <22385.43402.899210.927664@zaphod.ms25.net> <201606282146.u5SLkRNL005487@freefriends.org> Message-ID: <22387.3609.234358.545676@zaphod.ms25.net> On 2016-06-28 at 21:46:27 +0000, Karl Berry wrote: > all my intentions certainly get lost. > > In the case when an author has a straightforward-enough package but > wants certain files in certain places, it is fine to either include > an installation note in the README, or with the CTAN upload (they > pass it on to me), or, if I install it wrong, report the problem > and I will fix it. Not an issue. Much better than making a > .tds.zip just to "work around" my bug. Sure. I actually didn't have tds.zip files in mind. What I suggest is a flat hierarchy on CTAN like packagename/{doc,source,tex,tfm,...} The CTAN team assumes that users want to browse the files. But web browsers display only files with particular filename extensions and offer to download everything else. So it's useful to have browsable and unrecognized files in different directories. In this case you get the information where an author wants to have his files installed for free. Admittedly, I never used the CTAN web interface for accessing anything but documentation. Browsing anything else is nearly impossible. I suppose that even if an author provides everything in a single directory, the CTAN team sorts out the documentation, at least. > Notwithstanding, I think .tds.zip is fine and appropriate for > vntex. Good to know. It saves us a lot of time because our svn repository has been TDS compliant from the beginnig. Regards, Reinhard -- ------------------------------------------------------------------ Reinhard Kotucha Phone: +49-511-3373112 Marschnerstr. 25 D-30167 Hannover mailto:reinhard.kotucha at web.de ------------------------------------------------------------------ From vincent.torri at gmail.com Wed Jun 29 17:59:26 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Wed, 29 Jun 2016 17:59:26 +0200 Subject: [tex-live] memory leak with kpathsea_find_file() ? In-Reply-To: References: <201606272250.u5RMorAS021379@freefriends.org> Message-ID: On Tue, Jun 28, 2016 at 6:46 AM, Akira Kakuto wrote: >> n = kpathsea_find_file(kpse, name, kpse_tfm_format, 1); >> ... >> trying to free 'n' with free() leads to errors > > > On windows, if I link by using the import library, > I cannot reproduce the error. > However if I use the LoadLibrary() method, I can reproduce > the error. I don't understand why. Presumably my DLL lacks > something necessary, though it has been useful if used with > the import library. maybe some initialization not done. Maybe look at DllMain() to initialize stuff when the DLL is loaded https://msdn.microsoft.com/en-us/library/windows/desktop/ms682583(v=vs.85).aspx https://msdn.microsoft.com/en-us/library/windows/desktop/ms682596(v=vs.85).aspx regards Vincent Torri From zarniwhoop at ntlworld.com Wed Jun 29 22:59:14 2016 From: zarniwhoop at ntlworld.com (Ken Moffat) Date: Wed, 29 Jun 2016 21:59:14 +0100 Subject: [tex-live] Bad perms in 2016 texmf ? In-Reply-To: <201606270101.u5R117x7031179@freefriends.org> References: <20160624233411.GA28044@milliways.localdomain> <201606270101.u5R117x7031179@freefriends.org> Message-ID: <20160629205914.GA675@milliways.localdomain> On Mon, Jun 27, 2016 at 01:01:07AM +0000, Karl Berry wrote: (replying to the list now that I've had time to use it) > Hi Ken - you probably have no reason to spend more time on it, but FYI, > I've recreated the texmf tarball as texlive-20160523b-texmf.tar.xz in > ftp://tug.org/historic/systems/texlive/2016/ (also .sha512 and .asc). > Hopefully the permissions are more reasonable. > > I found some Japanese property list (.pl) files that were spuriously > executable due to the conflict with .pl for Perl. I fixed those. > > Aside from that, under doc/ and source/, there may well be some scripts > which are not executable (especially if they don't have an extension), > but I've never worried about that. (People can always invoke them > "manually".) I'm more interested in fixing files that are executable > and shouldn't be, if you find any of those. (gmglo.ist is fixed.) > > Thanks/sorry, > Karl Works fine for me. In the previous version, it was only the 'unreadable by others' files and directories that made me look. Those are all gone, and all my limited tests now pass. Thanks again. ?en -- Democracy is the theory that the common people know what they want, and deserve to get it good and hard -- H.L. Mencken From vincent.torri at gmail.com Thu Jun 30 12:08:25 2016 From: vincent.torri at gmail.com (Vincent Torri) Date: Thu, 30 Jun 2016 12:08:25 +0200 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: <835e5d64-dbb1-7149-4c52-10133c517211@alice.it> References: <835e5d64-dbb1-7149-4c52-10133c517211@alice.it> Message-ID: Hello On Fri, Jun 10, 2016 at 4:26 PM, Angelo Graziosi wrote: > > > I would remove VS binaries in favor of Cygwin and MSYS2-MINGW64 ones. Notice > that Cygwin already gives a method to install and use TL in the same manner > one uses TL on GNU/Linux and OSX (I don't use MacTex): same scripts to setup > PATHs, same build documents commands and so on... With MSYS2-MINGW64 would > be the same, almost.. As the Windows maintainer of several projects, I can confirm that with MSYS2 + mingw-w64, the compilation is the same (and even, cross compiling on linux is even faster as MSYS2 is quite slow). it's a matter of using i686-w64-mingw32-gcc or x86_64-w64-mingw32-gcc as compiler, that's all. I think that you use the autotools, in that case, configure should just be called like that (in MSYS2 or on linux for cross compilation): ./configure --host=i686-w64-mingw32 etc.... or ./configure --host=x86_64-w64-mingw32 etc.... On the coding side, what is important is that on Windows 64 bits, 'long int' is a 32 bits type, unlike on UNIX 64 bits. >> >> I would also be in favour of adding them, but Karl is afraid of the >> added complexity with not that much "added value" (properly supporting >> 64-bit Windows also means providing 64-bit perl, ghostscript etc., >> doubling the space required for all those binaries and properly >> deciding whether the Windows in 32-bit or 64-bit, ...). > > > why then having double binaries for other platforms (i386/x86_64 for linux > etc..) > > BTW, having the same app running on different systems helps in finding > subtle bugs.. indeed. Vincent Torri From angelo.graziosi at alice.it Thu Jun 30 15:37:11 2016 From: angelo.graziosi at alice.it (Angelo Graziosi) Date: Thu, 30 Jun 2016 15:37:11 +0200 Subject: [tex-live] 64 bits version of TeX Live on Windows In-Reply-To: References: <835e5d64-dbb1-7149-4c52-10133c517211@alice.it> Message-ID: <9127527f-093f-0aa3-29dd-d545c9c5a9d1@alice.it> Ciao Vincent, Il 30/06/2016 12:08, Vincent Torri ha scritto: > > As the Windows maintainer of several projects, I can confirm that with > MSYS2 + mingw-w64, the compilation is the same (and even, cross > compiling on linux is even faster as MSYS2 is quite slow). > > it's a matter of using i686-w64-mingw32-gcc or x86_64-w64-mingw32-gcc > as compiler, that's all. is not clear if you have successfully build TL (say from master) using this: > > I think that you use the autotools, in that case, configure should > just be called like that (in MSYS2 or on linux for cross compilation): > > ./configure --host=i686-w64-mingw32 etc.... > > or > > ./configure --host=x86_64-w64-mingw32 etc.... I tried "CXX=g++ ./Build" (or something like that) but the build stopped after a while (not on ICU) (now I don't remember...) If you have built TL, may you give a recipe both for MSYS2-MINGW64 build and the cross build on linux? > On the coding side, what is important is that on Windows 64 bits, > 'long int' is a 32 bits type, unlike on UNIX 64 bits. Yes, this is an old story.. TIA, Angelo.