Re: Configuration of teTeX broken (send once more due to list problems)
Andreas Tille writes:
> On 12 May 1998, Olaf Weber wrote:
> Only putting the real file to /etc/texmf could require some changes
> (but I don't know why because I can't think that texconfig or other
> programs check, whether /usr/lib/texmf/web2c/texmf.cnf is a real
> file or only a symlink to /etc/texmf/texmf.cnf).
The manner in which the file is updated would be important here.
AFAICT texconfig uses 'ed' for updates, so this would probably be OK
in the presence of symlinks. The files themselves are (typically)
only touched by root, during updates when /usr should be writable
anyway, so having /usr readonly should not pose a problem.
>> You might make texmf/web2c itself a symlink to /etc/texmf, but note
>> that this directory might contain largish files which IMHO are not
>> conffiles. The letter of the law might insist on those being moved to
>> a /var/texmf/web2c directory.
> Sure this would be a bad solution. Why we don't go the way of the old
> teTeX-0.4 packages???
Which was? (Incidentally, the files I'm thinking of are the .fmt
files and their relatives.)
[...]
> May be that teTeX 0.9 is more strict in following the TDS and enforces
> the user to stay at this. Now it works for me.
Unlike 0.4, 0.9 supports multiple texmf trees, but the underlying
assumption is that each texmf is TDS-compliant, or at least a close
match.
As for why things that worked under 0.4 failed under 0.9, you did not
provide enough info about what you did in 0.4, and how you attempted
to do it for 0.9 for me to be able to tell how you were bitten. My
best guess at present is that your local tree wasn't suffiently close
to TDS to be usable.
--
Olaf Weber
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: