[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#170382: acknowledged by developer (Bug#170382: fixed in tetex-base 1.0.2+20021025-4)



On [07/01/03 10:52], Atsuhito Kohda wrote:
> From: Christian Kurz <shorty@debian.org>
> Subject: Re: Bug#170382: acknowledged by developer (Bug#170382: fixed in tetex-base 1.0.2+20021025-4)
> Date: Mon, 6 Jan 2003 10:01:43 +0100

> > > But I cann't believe this makes tetex stable and reasonable 
> > > package unless we have a resonable, good, global design.

> > Pardon? Why should changing the package structure have any effect on the
> > stability of tetex itself? 

> Did you read "unless we have a reasonable, good, global design."?

Yes, but if I'm not mistaken, Debian ships with the three big
tetex-packages at least since the release 1.3 (Hamm). I remember seeing
them in 2.0 (potato) and now in 3.0 (woody). So I'm wondering how long
it takes to find a reasonable, good, global design. Maybe you only
started working on this design very late, then I'm going to apologize.

> If we splitted tetex to 20 or 30 small packages, then
> the situation would be much complicated and the dependency
> could be messed up.

Well, that's certainly an issue, but I think you (the tetex
maintainers) should be able to find a solution to this problem. So far
it seems Debian was always able to address problems when they occured.
So I hope that this is going to be the case here too. 

> > > then you already installed extra 38 Mb package, so you could 

> > koma-script. Some time ago I found that one can even download
> > koma-script as a seperate package and install it locally. After I

> # Hmm, I'm not sure but it might be realistic solution to 
> # request packages, e.g. koma-script or perl-tk, etc
> # which can work and coexist only with tetex-bin and tetex-base.

Interesting idea, it might also solve at least some of my problems with
the upgrade to the latest tetex packages. Maybe it would solve them
completely.

> I have a feeling you think teTeX as a union of mutually
> disjoint components but it was wrong.  
> As I said before teTeX consisted of TeX itself and many other
> related components and Thomas rearranged them in a consistent
> way with fairy good design.  If we split it again then there 
> is almost no reason to use teTeX.

Well, I think of tetex having a core and other helpful components, like
koma-script. So considering that lots of those components are seperate
packages that can even be downloaded from CTAN, it should be possible to
have seperate packages for them. I agree that splitting the core would
be difficult and need lots of consideration.

> For example, texdoctk is splitted in the teTeX so we have 
> /usr/bin/texdoctk in tetex-bin and its supporting stuffs 
> in tetex-extra.

> kohda@nsx:~/public_html/debian/experiment$ dpkg -L tetex-bin | grep texdoctk
> /usr/share/man/man1/texdoctk.1.gz
> /usr/bin/texdoctk
> kohda@nsx:~/public_html/debian/experiment$ dpkg -L tetex-extra | grep texdoctk
> /etc/texdoctk
> /etc/texdoctk/texdoc-100.dat
> /etc/texdoctk/texdoc-102.dat
> /etc/texdoctk/texdoc-103.dat
> /etc/texdoctk/texdocrc
> /usr/share/doc/texmf/texdoctk
> /usr/share/doc/texmf/texdoctk/README.gz
> /usr/share/texmf/texdoctk

> so we need texdoctk-bin and texdoctk-base if we split
> tetex?  But when texdoctk was its own package, they 
> were all in one.

Well, I don't hink so. Considering that only the files above are
belonging to texdoctk, I think they can be safely put in one package, if
they aren't very big. I fail to see why you want to create two packages,
-bin and -base. 

> So if it is really needed to split TeX (and related) 
> packages to small packages I guess it is much better not
> to use teTeX from the beginning but it would be another hard
> task (and then perhaps we need another Thomas in Debian)

Well, maybe this question should be discussed in a more public manner to
see what other developers think about the issue. I've no idea of the
majority of them is happy with the current three packages or not. But if
the majority is preferring to keep them, I'm going to shut up and look
for other solutions to my problem.

Christian
-- 
           Debian Developer (http://www.debian.org)
1024D/B7CEC7E8 44BD 1F9E A997 3BE2 A44F  96A4 1C98 EEF3 B7CE C7E8

Attachment: pgpbgTe_cJnUT.pgp
Description: PGP signature


Reply to: