Re: Question for A. Towns - NM
Henning Makholm <henning@makholm.net> wrote:
> Frank Küster <frank@debian.org> wrote on -vote:
>
>> P.S. should anybody be in search of a package to devote their work to,
>> and have some bandwidth and disk space available, and be interested in
>> TeX, please join debian-tetex-maint!
>
> OK.
>
> I have bandwidth, disk space, and a fairly good command of TeX arcana
> and BibTeX stylefiles. I maintain the bibclean package (a validator
> and prettyprinter for BibTeX database files). Regrettably not much
> familiarity with what happens to fonts outside the TeX program itself
> (though I did once read the Metafontbook).
That sounds great, thank you very much, and welcome in the team!
> So what needs work?
Mainly teTeX 3, see below
> Triaging new bugs,
That's always welcome, of course, but currently there aren't so many
(knock on wood), and Helge Kreutzmann also started to do some "first
user contacts" recently - Helge, by the way, do you plan to be active on
the list on a constant basis?
> fixing specific old bugs,
Hilmar Preusse is doing a good job on keeping track of old bugs. There
are some which we delayed to until sarge is released, and/or to teTeX
3. Other could still deserve that somebody looked at them, both old
outstanding upstream bugs, and packaging stuff like more libpaper
support.
And then there's always #218105 - we should specifiy directly in the
copyright file which license (and, ideally, copyright) applies to the
individual parts of teTeX. I've started that with
http://cvs.debian.org/tetex-base/debian/copyright?cvsroot=tetex
http://cvs.debian.org/tetex-base/debian/Copyright.Files?cvsroot=tetex
http://cvs.debian.org/tetex-base/debian/tetex-legalese?cvsroot=tetex
However, this is a really big piece of work with no particular reason
that would make it fun doing it, and I fear that it will never be done
if we want to do it in one big effort. Therefore I suggested a "policy"
that everytime someone looks at a package, e.g. because there's a bug
report about it and he tries to find a patch, he should at the same time
check its licensing/copyright situation, and provide a patch to
copyright and Copyright.Files, or just check it in.
As for bugs, almost a year ago I have tried to group them, but I didn't
keep the result up-to-date. It might still be interesting to look at
it, it's at
http://people.debian.org/~frank/tetex-base_bugs_sorted.html
http://people.debian.org/~frank/tetex-bin_bugs_sorted.html
> packaging or testing teTeX 3?
That's the biggest issue, and there are two aspects of it:
a) any chance of getting teTeX-3.0 into sarge?
b) address all the issues that have been delayed previously, like a
clever splitting scheme, writing the Policy text, etc.
As for a), I have written up a document on that, and will post it soon.
The short summary is:
For the use of tetex as a typesetter of generated code (from docbook,
debiandoc, texinfo, etc.) (and of unmaintained old code), a tetex
upgrade is undesirable, but for users that directly write code
themselves it would be highly desirable. I still hope that sarge will
be released soon; but if it's going to be released only in the second
half of this year, I think we should try to get it in. Because of the
many packages build-depending on tetex, we must be _very_ careful . For
the details, see the text I'll send separately.
As for b):
- We have always desired to come up with a better splitting scheme, with
the goals of having smaller packages, better integration of the
non-TeX-specific parts (mainly Type1 fonts) into Debian, more choices
(e.g. xdvi with different toolkits), possibility to replace parts of
the packages (e.g. if somebody packages pdfxtex, the development
branch of pdftex), etc.
- I have many ideas how the packaging of teTeX-3.0, as currently in
experimental and in the CVS, could be improved. Some are in the TODO
files, many are not, because I was the only one that actually did
coding on it, the others only tested and commented. I would also
appreciate criticism on what I've done so far, especially from
somebody who is not very familiar with the way 2.0.2 was packaged.
Although I did a couple of things different, it could be that I still
keep some old cruft that doesn't really make sense, or have not taken
over things I didn't understand, but would have in fact been good.
Some further notes: For 2.0.2, the orig.tar.gz file isn't the original
one. The original one unpacks in the current directory what later ends
up as subdirectories of /usr/share/texmf, and it has been repacked to
unpack into tetex-base-2.0.2c.orig. For version 3.0, I stopped the
repackaging, because meanwhile I learned that such an orig.tar.gz file
is no problem for dpkg (see http://bugs.debian.org/278524 which I
submitted against developers-reference).
The nasty side effect is that our CVS for tetex-base has a double
structure: Subdirectories (and files therein) exist in their 2.0.2
version beneath http://cvs.debian.org/tetex-base/texmf/?cvsroot=tetex,
and in their 3.0 versions directly in
http://cvs.debian.org/tetex-base/. Since we don't have any
Debian-specific changes in these files (everything is in patches), I
didn't mind the move; but it has the side effect that currently, as long
as HEAD still is 2.0.2 and 3.0 is in the experimental branch, you get
after a checkout either the empty 3.0 directories and the 2.0.2 files,
or the 3.0 files and the empty texmf directory from 2.0.2, and have to
do a cvs update -P to get rid of them. Which on the other hand had the
good effect that it made it easier to find CVS bug #290531; I have not
yet found time to make a reproducible, small example for it.
Regards, Frank
--
Frank Küster
Inst. f. Biochemie der Univ. Zürich
Debian Developer
Reply to: