Bug#291545: tetex-bin: initex fails to generate some etmf files

[frank@kuesterei.ch (Frank Küster)]

severity 291545 normal
retitle 291545 upgrade of files in texmf.d shows diff twice (plain ucf and update-texmf)
stop

Margarita Manterola <marga@marga.com.ar> schrieb:

> Now, the fact is that it asked me the same question twice (now that the
> mistery is solved, I have the sensation that I remember looking at the diff
> and thinking "isn't this the same diff I saw before?"), therefore causing
> my making the mistake of not accepting it twice.
>
> Is there a way to prevent this?  Some way of "preseeding" update-texmf so
> that it doesn't happen?

Actually, I fear it is not possible. We *could* make things easy and
straightforward: If we would put the generated texmf.cnf file into
/var/lib/texmf, then only the files in texmf.d would be configuration
files, and only there could a question come. By the way, the reason for
all this is to enable other packages to add their own configuration to
texmf.cnf, like e.g. jadetex does.

However, the problem with this is that texmf.cnf is known as "the"
configuration file for teTeX, it is referenced in many README files or
tutorials, and, to make things worse, since it must always be possible
for TeX to find it, users can find it with the same means:

 kpsewhich --format='web2c files' texmf.cnf
/usr/share/texmf/web2c/texmf.cnf

(actually you can use it without the --format stuff). This is why we
have chosen to keept texmf.cnf in /etc/texmf, as a configuration file -
people would have edited anyway. But this makes the double checking
necessary. 

> Now, this bug was not a bug, but a human's (namely, me) fault, yet I feel 
> that something should be done to prevent this from happening again, don't 
> you think?

Yes, it would be good. I'll think it over, and keep the bug open under a
different title, and with lower severity.

Regards, Frank
-- 
Frank Küster
Inst. f. Biochemie der Univ. Zürich
Debian Developer