On Mo, 12 Nov 2007, Felipe Figueiredo wrote:
> I acknowledge it *is* a different bug, but I must insist that if the symlink
> is not created automatically on package installation, this is related. I
Not in even remotly related. The only thing that COULD be the error
(sorry, I cannot check, sitting on a Windows Computer) is that the
fmtutil-sys call breaks BEFORE the update-alternatives code is reached,
so it will be never called. But in this case this is still a different
bug (and 99% probablility that it has been already reported).
> > It is not THAT simple, because we have other xdvi alternatives and we
> > cannot replace a MANUALLY set link (status manual) in any case with
> > --auto, this is not how Debian (Ubuntu?) works, one has to respect the
> > decisions of the system administrator.
>
> Sure. But that's assuming the link is sane, which was not the case for the
> bug
> reporter, myself and others. I'm also not sure this IF was even tested, by
> the reason mentioned above.
Sorry, but obviously you don't understand these package issues: Can you
explain me in simple words what "sane" means?
If the link points to a file on a NFS mounted share, that is NOT mounted
at package upgrade time, I cannot even check on the existence of the
file. But still this is perfectly sane!!!!
So, now it's your turn, explain we what is sane? We are already checking
for those settings we KNOW that are outdated/from tetex, and remove
them. But anything else is OUT OF OUR BUSINESS. That is sysadm business.
I am not saying that we cannot improve the code in the postinst script,
but a simple
update-alternatives --auto ..
is simply NOT the solution. You are invited to contribute more
intelligent code, please contact [EMAIL PROTECTED]
> In my particular case, since I only have one xdvi alternative, setting it
> with --auto solves the problem. I inputted this to help others who might find
> this report looking for help. Sorry if it's inapropriate. Could you please
> suggest a better approach?
No, it is a good approach. First, I assume that the update-alternatives
code was not reached due to a failure PRIOR in the postinst script.
Second, if finally a configure was successfull, and the alternative was
pointing to some strange place, we as maintainer cannot do anything.
> > And the output of
> > LANG=C update-alternatives --display xdvi.bin
> > LANG=C update-alternatives --list xdvi.bin
> > would be nice. The output BEFORE fixing this by hand.
>
> Sorry, I didn't check. I was in a hurry, and I only cared at making it work
> then.
A pity, because we cannot know now the source of the problem.
> I might have another recently upgraded system to test it on, and if it has
> the
> bug, I'll give the details you asked for.
That would be great, thanks a lot.
Best wishes
Norbert
-------------------------------------------------------------------------------
Dr. Norbert Preining <[EMAIL PROTECTED]> Vienna University of Technology
Debian Developer <[EMAIL PROTECTED]> Debian TeX Group
gpg DSA: 0x09C5B094 fp: 14DF 2E6C 0307 BE6D AD76 A9C0 D2BF 4AA3 09C5 B094
-------------------------------------------------------------------------------
ULLOCK (n.)
The correct name for either of the deaf Scandinavian tourists who are
standing two abreast in front of you on the escalator.
--- Douglas Adams, The Meaning of Liff
--
[gutsy] xdvi.bin symlink not installed
https://bugs.launchpad.net/bugs/156628
You received this bug notification because you are a member of Ubuntu
Bugs, which is the bug contact for Ubuntu.
--
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
