usefulness of package failure bugs

[Michael Vogt]

Hi,

I would like to ask for opinions (especially by the heavy bug triagers
:) about the apport package failure report feature in
libapt/synaptic/update-manager that is responsible for generating
apport reports if package maintainer scripts fail and generate dpkg
errors. 

We discussed this in the distro meeting (see [1]). Some people
objected that the current reports are not useful for various
reasons. I tried to summarize the discussion in [2]. 

My impression is that the usefulness of the reports heavily depends on
the set of packages. They are e.g. not very useful for e.g. gnome
packages that always run "update-icon-cache", "update-scrollkeeper" in
pretty much every package with very little "custom" maintainer
scripts. For server oriented package the usefulness seems to be much
bigger because there are more maintainer scripts and usually they are
very custom.

I would like to have your opinion about the reports. One of the
solutions that was discussed was to move the reports initially to a
new pseudo package like "package-failure" that would then need initial
triage to move the reports to the right package. Do you think that
this is a sensible course of action? Do you think the QA team would
have resources to help with the triage of this? 

Cheers,
 Michael

P.S. Please CC me as I'm not on the QA list.

[1] https://lists.ubuntu.com/archives/ubuntu-desktop/2008-August/001650.html
[2] https://lists.ubuntu.com/archives/ubuntu-desktop/2008-August/001659.html