Re: [DRAFT]: The Debian TeX Policy
Florent Rougon <f.rougon@free.fr> schrieb:
> Hi,
>
> In addition to what Richard already suggested...
>
> Frank Küster <frank@debian.org> wrote:
>
>> The latest copy of this document can be found in the
>> `Debian-TeX-policy' files in the `tetex-bin' package.
>
> Not very important, but... why not tetex-base ? It feels more natural to
> me to have it there (arch-indep, can be installed without tetex-bin...).
You're right. The only reason why it is in tetex-bin is a "historical"
one: I started editing tetex-bin's README.Debian, which contains
information about the update-* programs, in order to develop a
policy, and put it into a parallel README.Developers.
,----
| README.Developers for the teTeX packages
| =========================================
|
| This document contains information that is relevant for developers of
| TeX-related packages in Debian. It is intended that it will once
| become some kind of "TeX policy"
`----
And when I decided to start a new policy document, I did it there. Yes,
we should move it to tetex-base.
> should be written "Metafont"...
fixed.
>> architecture-indpendent package.
> ^
> e
fixed
>> 3. `/usr/share/texmf-site/', referenced as SITETEXMF
>
> Ah, there at last, you'll have something to hand out to the people who
> request $latex_package to be updated to the latest CTAN version. ;-)
Or packages that really depend on the functionality of a specific
version, if they use LaTeX as a backend to output pdf or the like.
>
>> TEXMFMAIN, can place them into SITETEXMF. The package must make sure
>> that the newer version is backward-compatible, that is: It must not
> ^
> i
No capital after a colon?
>> determinded by `/var/lib/texmf/web2c/fmtutil.cnf'. All three files
> ^
fixed
>> Maintainer scripts should call the configuration update programs
>> without any options (excepct `-v') to allow for internal changes, e.g.
> ^
>
>> of the directories where the generated files are placed.
>
> Reminds me of something...
Any reason why you did not yet change it? (I'm writing this offline, so
I cannot follow your hint in postinst and read bug #213310).
>> `Depends:' on the package providing the executable, and _only_ do
>> that. Any additional checks, e.g. for the existence of format files,
> ^
>
>> is unnecessary and harmful, causing internal changes (e.g. of format
> ^
>
> debiandoc2txt thinks we have end-of-sentence periods here. I suppose
> there is no way to tell him that they do *not* mark the end of a
> sentence...
frank@alhambra:~$ zgrep sentence /usr/share/doc/debiandoc-sgml-doc/debiandoc-sgml.en.txt.gz
frank@alhambra:~$ zgrep period /usr/share/doc/debiandoc-sgml-doc/debiandoc-sgml.en.txt.gz
frank@alhambra:~$
Could be worth a bug report against debiandoc-sgml
>
>> Packages should be able to delay running of mktexlsr, updmap and
>> perhaps even "fmtutil -all" until all TeX-related packages that want
>
> According to the manpage, it is "--all".
Fixed.
> OK, it seems to be good starting point. I suppose more is to come to
> this policy (things such as the best order when running update-updmap,
> mktexlsr and updmap) but things will probably change in a substantial
> way when teTeX 3 enters sid, so I think it is a bit too early to write
> that.
Yes, of course. But maybe not too early to think about it. As for the
order:
I think mktexlsr often has to be run twice, especially when updmap is
involved (and we wanted to kick out update-updmap, anyway). But I don't
know exactly in which cases, that is, in which cases it will search only
the ls-R files, or also the disk. If you delete latex.efmt, mktexlsr,
and regenerate it, it will be found by kpsewhich even before mktexlsr is
run again.
One more thing that should perhaps be added is a convention "what is a
configuration file". Are map files for a particular font configuration
files?
Thanks for your comments,
Frank
--
Frank Küster, Biozentrum der Univ. Basel
Abt. Biophysikalische Chemie
Reply to: