Bug#350992: tetex-bin: Still installs files in TEXMFMAIN
Ralf Stubner <ralf.stubner@physik.uni-erlangen.de> wrote:
> Frank Küster wrote:
>> Package: tetex-bin
>> Version: 3.0-12sarge13
>> Severity: normal
>>
>> actually, lots of them.
>
> Is that really a problem? I had the impression that the files in
> TEXMFMAIN are related to the actual binaries (sort of like upstream has
> it). I don't think that it is possible to achieve interoperability with
> TeX Live for binary packages. And I am not sure if it where good if
> other packages would shadow these files. I only had a casual look on the
> files in TEXMFMAIN, though.
You may be right, but we wrote in our policy:
,----
| The role of the trees TEXMFMAIN and TEXMFDIST in Debian differ from
| upstream's original usage. Upstream uses TEXMFMAIN for the files that
| have to match the binary executables and TEXMFDIST for other TeX input
| files that are replaced when a new texmf tarball appears; this
| distinction is not necessary on a system with a decent package
| managment system. Instead, the basic TeX packages install their files
| into their TEXMFDIST directories [1] , while TEXMFMAIN is used by TeX
| add-on packages for their files and allows them to shadow older
| versions provided by the basic TeX packages.
`----
This implies, IMHO, that the basic TeX packages should not install their
files into TEXMFMAIN. If we think that these files should not be
possibly shadowed, we should change the wording of the policy.
But I'm not sure that shadowing is in fact a problem for all of them.
Some of them are also kind of configuration files, for example
$ grep "some prefer" /usr/share/texmf/web2c/mktex.opt
: ${ps_to_pk=gsftopk} # some prefer ps2pk
and it might make sense to change this when a particular package is
installed.
Regards, Frank
--
Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX)
Reply to: