Bug#150511: tetex-bin: the postinst fails to preserve user changes in /etc/texmf/texmf.cnf and others
>>"Atsuhito" == Atsuhito Kohda <kohda@pm.tokushima-u.ac.jp> writes:
>> My point in fact is that we, in Debian, have been going around
>> telling people that it is OK for them to edit things under /etc, and
>> we shall opreserve changes. Policy says so unambigusously.
Atsuhito> Okay but I say once again that it is OK to edit things
Atsuhito> in /etc/texmf/texmf.d/
As long as I am asked about having my hand crafted
/etc/texmf/texmf.conf removed, that would be fine.
>>
>> Not quite good enough. If one were used to running TeX on a
>> non Debian machine, editing texmf.conf comes naturally. Trying to
>> maintian a common tetex environment across various flavours of UNIX
>> is not uncommon (I have a texmf.conf that has grown over the years,
>> with commnets, and toehr details).
>>
>> This violates the principle of least surprise.
Atsuhito> Do you claim that although we Debian could provide much more
Atsuhito> sophisticated mechanism than other UNIX systems, we should
Atsuhito> follow the criteria of other less featured UNIX systems?
When it comes to not blowing away user changes, yes.
Atsuhito> If you maintain a TeX environment across flavours of UNIX
Atsuhito> then texmf.cnf of Debian would be the best candidate to use
Atsuhito> commonly.
Not if it keeps blowing away my comments and other changes.
Atsuhito> Our scheme is reasonable enough because it preserves user's
Atsuhito> changes so long as a user changes texmf.cnf through modifying
Atsuhito> files in /etc/texmf/texmf.d/ but we might/should provide
Atsuhito> more good documentation or announcement on it.
The policy does not say preserve user changes only if the user
jumps through hoops we tell them to. I think
Atsuhito> Further, IMHO, it seems policy says nothing explicitly on the
Atsuhito> preservation of changes when a package removes a conffile
Atsuhito> (texmf.cnf) and installs new conffiles (/etc/texmf/texmf.d/05TeXMF
Atsuhito> etc.).
>>
>> Policy states quite clearly that _all_ configuration files,
>> whteher or not they are conffiles, MUST have user changes preserved.
Atsuhito> This is not an answer to my question. A conffile must
Atsuhito> preserve user changes and a cofiguration file must preserve
Atsuhito> user changes, it's okay. But when a conffile is changed to
Atsuhito> a configuration file (or vice versa) it is almost meaningless
Atsuhito> to claim it must preserve user changes and policy says nothing
Atsuhito> in such case, IMHO.
If user changes are to be preserved before, and user changes
are to be poreserved after the change, what make you think it is OK
to screw the use during the transition?
Atsuhito> I belive this feature of update-texmf is much, much, much
Atsuhito> more important than you can imagine. These packages should
Atsuhito> stop to work if we stop to use update-texmf.
So don't stop. Just ask the user before update-texmf blows
away changes. preserve old version in the backups directory (after
asking).
manoj
--
Oh, I am a C programmer and I'm okay I muck with indices and structs
all day And when it works, I shout hoo-ray Oh, I am a C programmer
and I'm okay
Manoj Srivastava <srivasta@debian.org> <http://www.debian.org/%7Esrivasta/>
1024R/C7261095 print CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E
1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
--
To UNSUBSCRIBE, email to debian-tetex-maint-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: