[tex-live] hyperref/puenc.def broken after upgrade

[Pander]

 2012-08-14 17:44, Robin Fairbairns wrote:
> Pander <pander at users.sourceforge.net> wrote:
> 
>> On 2012-08-14 16:48, Robin Fairbairns wrote:
>>> Pander <pander at users.sourceforge.net> wrote:
>>>
>>>> About testing hyperref, the following isn't even working with xelatex:
>>>>
>>>> \documentclass{article}
>>>> \usepackage{hyperref}
>>>> \begin{document}
>>>> \end{document}
>>>>
>>>> Perhaps this could be part of some automated testing in TeX Live
>>>> whenever anything related to hyperref changes.
>>>
>>> wouldn't help; i've just (a) updated tl, (b) run your test with xelatex
>>> (and pdflatex and plain latex) without any problem.
>>
>> Here not :(
>>
>> $ date
>> Tue Aug 14 17:10:46 CEST 2012
>>
>> # tlmgr update --all
>> tlmgr: package repository
>> http://ftp.snt.utwente.nl/pub/software/tex/systems/texlive/tlnet
>> tlmgr: saving backups to /usr/local/texlive/2012/tlpkg/backups
>> tlmgr: no updates available
>>
>> $ cat hyperref.tex
>> \documentclass{article}
>> \usepackage{hyperref}
>> \begin{document}
>> asdf
>> \end{document}
>>
>> $ xelatex hyperref.tex
>> This is XeTeX, Version 3.1415926-2.4-0.9998 (TeX Live 2012)
>>  restricted \write18 enabled.
>> entering extended mode
>> (./hyperref.tex
>> [...]
>> ! LaTeX Error: Missing \begin{document}.
>>
>> See the LaTeX manual or LaTeX Companion for explanation.
>> Type  H <return>  for immediate help.
>>  ...
>>
>> l.1591 \DeclareTextCommand[\textBeam}{P
>>                                        U}{\9043\223}%* U+2393
> 
> sounds to me as if i'm using a more up-to-date tl than you are.  (note i
> take my updates over nfs from the ctan server in the next corridor.)
> 
> looking at your offending package, i see
> 
> $ ls -l `kpsewhich puenc.def`
> -rw-r--r--. 1 rf10 rf10 122261 2012-08-13 23:37 /local/texlive/2012/texmf-dist/tex/latex/hyperref/puenc.def
> 
> i.e., the package was installed yesterday, and would have appeared for
> the first time in this morning's mirror.  are you *sure* you have an up
> to date texlive?


# tlmgr update --self
tlmgr: package repository
http://ftp.snt.utwente.nl/pub/software/tex/systems/texlive/tlnet
tlmgr: saving backups to /usr/local/texlive/2012/tlpkg/backups
tlmgr: no updates for tlmgr present.
tlmgr: no updates available

# tlmgr update --all
tlmgr: package repository
http://ftp.snt.utwente.nl/pub/software/tex/systems/texlive/tlnet
tlmgr: saving backups to /usr/local/texlive/2012/tlpkg/backups
tlmgr: no updates available

What more can one do? Except changing mirrors...

$ ls -l `kpsewhich puenc.def`
-rw-r--r-- 1 root root 118162 Aug 12 23:26
/usr/local/texlive/2012/texmf-dist/tex/latex/hyperref/puenc.def

> 
>> What example do you use?
> 
> the null document you previously posted.  i've now added some text in
> the document body, but it plainly has no relevance
> 
>> If hyperref is a high risk upgrade, some simple testing would be in
>> place. People should be able to expect some quality when using TeX Live.
>> I value TeX Live distribution a lot so some extra tests would be very
>> welcome to keep on guaranteeing that.
>>
>> Just start out with a simple test such as the one above and each time a
>> problem arises with new packages, just add that particular test. In this
>> way, updating is less risky as it apparently is now.
>>
>>> if we're going to run tests for every package that's updated, there's
>>> going to need to be an awful lot more work done on the tl release
>>> infrastructure, including writing a test suite per package.
>>
>> All packages should include their own tests.
> 
> in an ideal world.  very few do (you wouldn't want many of them -- for
> example, the latex tests are of the same order of size as latex itself).
> 
> in the real world, with software being developed by people in their
> spare time, your "should" is one of those inaccessible pies, floating in
> the sky, that no-one can get at.
> 
>> The tests I refer to are
>> for integration testing checking if all packages can coexist peacefully.
> 
> but *everyone* knows that there are packages that can't coexist --
> several (ones doing the same job in different ways) you wouldn't even
> _want_ to coexist.
> 
>>> sounds excessive, to me.
> 
> i agree with myself.

LOL

> you've not even mentioned when you last updated, nor which repository
> you were updating from.  (have you tried updating since you first
> observed the effect, even?)

Yes. I updated Monday and all was OK and today, then it broke down.

> 
> robin 
>