[tex-live] dvipdfm(1) regression with TL 2009

[Jukka Salmi]

Jonathan Kew --> tex-live (2009-08-05 09:32:24 +0100):
> On 5 Aug 2009, at 09:27, Jukka Salmi wrote:
> 
> >Hello,
> >
> >I just built TL binaries from svn trunk and installed a TL 2009
> >pretest
> >version on a NetBSD/i386 system.  With a LaTeX document which includes
> >several .eps graphics I'm seeing a regression with dvipdfm(1)
> >(which is
> >actually dvipdfmx now IIUC): with TL 2008's (non-dvipdfmx) dvipdfm(1)
> >the resulting .pdf looks fine, but with TL 2009's dvipdfm(1) some
> >included graphics seem to be replaced by graphics which are shown
> >elsewhere in the same document.  The .dvi and .ps (generated by
> >dvips(1)) are fine, so it's definitely dvipdfm(1) which causes the
> >problem...
> >
> >(Unfortunately I can't publish the document in question, but I
> >could try
> >to reproduce the problem with a small example document if this might
> >help...)
> 
> Yes, I think that would be helpful.

Ok, I've put an example [1]online.  Running latex(1) on the .tex file
gives the expected results, but then running dvipdfm(1) does [2]not.
Note that instead of g27.eps, g01.eps is shown in the resulting .pdf;
somehow displaying the .eps files seems to wrap after 26 files...

BTW, I can only reproduce this with dvipdfmx-20090708 (TL 2009), but not
with dvipdfmx-20080607 (TL 2008).  The 2008 version works fine, as does
dvipdfm 0.13.2d (TL 2008).


Regards, Jukka

[1] http://salmi.ch/~jukka/TL/dvipdfmx/
[2] http://salmi.ch/~jukka/TL/dvipdfmx/dvipdfm-test.pdf

-- 
This email fills a much-needed gap in the archives.