Rene Engelhard rene-at-deian.org |DebianBug| wrote:
severity 494644 important
tag 494644 + confirmed
thanks
Hi,
Brendon Green wrote:
Package: openoffice.org
Version: 1:2.4.1-6
Severity: grave
Justification: causes non-serious data loss
No. Just because a simgle documents (which even is guessed to be malfformed)
crashes is *NOT* a grave bug.
I wasn't completely sure of that either. My justification was that the
user would lose unsaved documents when the crash is triggered, hence
minor data loss.
When opening some externally-generated documents, OOo immediately
crashes. The worst part of this is that OOo will try to recover all
open documents the next time it starts, including the one that crashed
it, resulting in another crash.
Yes, that is worse, but you can select to not recover it..
True, but then the user has no way to selectively recover other
documents open at the time of the crash. I should probably open another
bug about that.
These particular documents were generated by Inkscape. If they are
indeed malformed (which seems likely, as they are much smaller than
expected), then that is obviously a different bug.
Yes..
Ideally, however, the program should not crash no matter how fubared
the input is. I would rather see a message informing me that the file
is unreadable/corrupted than have the app spiral into the crash loop
described above.
Attached is a backtrace of the crash, and a tarball of (tiny) sample
documents to play with.
OK, I guess I am allowd to give those files also to upstream so that
they can fix it?
Do whatever you like with them. The data they contain (if any) has no
confidentiality.
That said, I also can reproduce it on the 3.0 release branch of OOo...
Mmm, OOo 3.0 beta sounds yummy. Pity it hasn't hit Sid yet (AFAIK).
Regards,
Rene
Cheers
Brendon
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
