Re: improving cm-super package
Klaus Schneider <klaus.schneider@iup.uni-heidelberg.de> wrote:
> * I contacted the maintainer of the old debian stuff of the cm-super package
> at ctan, and he is willing to update the original package. He suggested to
> make the package suitable for both tetex 2 and tetex 3, i.e. the package
> detect the tetex version at installation time (i.e. by evaluating the output
> of dpkg -l tetex-base) and generate symbolic links for encoding and map files
> in the /usr/share/tetex tree accordingly; my suggestion for this would be to
> put the real files to /usr/share/cm-super or so.
I doubt that it's a good idea to run "dpkg -l" while dpkg is running and
executes maintainer scripts. Even if the output is predictable
currently, the fact is undocumented and subject to change.
There might be other possibilities to find out where to place symlinks.
But I'm not sure that it is worth the effort. Everybody who installs
the new cm-super package from a Debian repository will have teTeX 3.0
available. Therefore the only people who would benefit from this are
the ones who run sarge without backports, take cm-super from CTAN and
build a Debian package from it.
I think that it is rather a bad habit that upstream authors provide
debian directories in their distribution; most of the time they are
either outdated or inapplicable to stable, and they make packaging
harder. (Note also that there are other distributions, like Ubuntu, who
also use the *.deb format, and might have different policy.) Therefore
I wouldn't want to encourage people using upstream's debian directory to
create *deb, anyway.
The current debian directory is actually a good example why it is a bad
idea: The packages installs /usr/bin/t1c2pfb, and it had no reason not
to do so when it was created. Meanwhile, the file is already in Debian,
at the same place, in its own Debian package, and anybody who installs
the deb from CTAN will get a file conflict when or as soon as
pfb2t1c2pfb is also installed. And nobody is to blame...
Sorry for turning you suggestion down that way, but that's how I see
it.
Regards, Frank
--
Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX)
Reply to: