[tex-live] texmf-dist/doc/xelatex/xltxtra/xltxtra.pdf

[Reinhard Kotucha]

>>>>> "Sanjoy" == Sanjoy Mahajan <sanjoy at mrao.cam.ac.uk> writes:

  >> > === texmf-dist/doc/generic/vntex/vn-min.pdf ====
  >> 
  >> Could you tell me the version number of vntex, please?  Look at
  >> the very first line in the file
  >> texmf-dist/doc/generic/vntex/README.

  > vntex-3.1.4

Hi Sanjoy,
good that you have the latest release, this makes things easier here.

I guess that there is a problem with Bitstream Charter fonts indeed.
All fonts in vn-min.pdf define RD and ND but the Charter fonts put
this definition to userdict.

 userdict /RD {string currentfile exch readstring pop} executeonly put
 userdict /ND {noaccess def} executeonly put

In PostScript a dictionary is an associative list, or, with other
words, a named list which contains key/value pairs.

A font itself is a dictionary and it should contain all the resources
it needs instead of defining some procedures in other
dictionaries.  In my opinion, even if it is allowed, it's wrong and
dangerous. 

Also, not every Type1 rendering engine is based on a PostScript
interpreter and I expect more problems with other programs.

I suppose that ghostscript people provided a workaround in newer
versions.  If something is not found in the dictionary which is on top
of the dictionary stack it steps down the dictionary stack until it
finds the procedure.

I replaced the lines mentioned above by 

 /RD {string currentfile exch readstring pop} executeonly def
 /ND {noaccess def} executeonly def

and compiled vn-min.pdf again.  The error/warning messages do not
appear any more even with gs-8.50.

All the other fonts define RD and ND like this.

You can download vn-min.pdf which includes the modified fonts from

    http://ms25.ath.cx/vn-min.pdf

You can try the modified fonts yourself:

    http://ms25.ath.cx/charter.zip

is supposed to be extracted in texmf-dist/fonts/type1/bitstrea.  There
are three files for each font.

   *.pfb are the modified fonts.
   *.pfb.gz are the original fonts.
   *.src are the disassembled modified fonts.


Karl, how shall I proceed?  I'm sure that it's better to provide the
modified fonts and it's no problem for me to upload them to the
repository but there might be some conflicts with .tpm files if I just
replace the files.  It would be nice if some people try 

    http://ms25.ath.cx/vn-min.pdf

on screen and with printers.

It's a pity that Tom Kacvinsky is so silent now.  Barbara Beeton told
me that he he's still alive but probably quite busy, but there is
someone else now is responsible for PostScript and fonts at AMS.  I
don't remember his name and do not know his email address but a
comment from him is highly desired. 


Sanjoy,
I do not regard ghostscript as a tool which is able to validate PDF
files.  Their aim is to make everything work, whether it's broken or
not.  Similar with HTML.  You can't use a web browser to determine
whether your HTML file is valid because all developers of web browsers
think that it is a good idea to make their products error tolerant.

Regards,
  Reinhard

-- 
----------------------------------------------------------------------------
Reinhard Kotucha			              Phone: +49-511-4592165
Marschnerstr. 25
D-30167 Hannover	                      mailto:reinhard.kotucha at web.de
----------------------------------------------------------------------------
Microsoft isn't the answer. Microsoft is the question, and the answer is NO.
----------------------------------------------------------------------------