Candidate revision checkbox_0.8.6

I have removed the previous branch and replaced it with another branch which disables the apport prompt plugin instead of just disabling xorg bugs. This will prevent checkbox from prompting the user to report a bug when a test fails.

Is the branch supposed to be linked to this bug report?

I've found the branch now and it looking at it seems that checkbox has its own switch for turning off bug filing. This concerns me a bit in that we have three automated bug reporting tools (checkbox, apport, and kerneloops) all with their own separate on / off bit. (One might also count update-manager and its modification of /etc/default/apport for distribution-upgrades.)

I can understand kerneloops having a different switch as kernel freeze is earlier than other items in the release cycle. But going forward I'd rather than an effort was made to consolidate these wherever possible. Subsequently, I'd really prefer that checkbox checked '/etc/default/apport' to see if it was enabled or not.

On Wed, Nov 04, 2009 at 08:53:44PM -0000, Brian Murray wrote:
> I've found the branch now and it looking at it seems that checkbox has
> its own switch for turning off bug filing. This concerns me a bit in
> that we have three automated bug reporting tools (checkbox, apport, and
> kerneloops) all with their own separate on / off bit. (One might also
> count update-manager and its modification of /etc/default/apport for
> distribution-upgrades.)

Typically update-manager hasn't modified it, correct? We've relied in
the past on the default value as shipped in the apport package, I
thought -- it's only with it not being set correctly in karmic that
we've changed it (which was particularly annoying for those of us that
intend to keep it enabled by default).

> I can understand kerneloops having a different switch as kernel freeze
> is earlier than other items in the release cycle. But going forward I'd
> rather than an effort was made to consolidate these wherever possible.
> Subsequently, I'd really prefer that checkbox checked
> '/etc/default/apport' to see if it was enabled or not.

I concur, though do note that the toggle for apport is the automatic
crash catching behavior, not generally the ability to file bugs
[snarky comment about the desirability of such a toggle elided].
Automated tools such as kerneloops and checkbox certaintly fall into
the same hash bucket, though.

Re kerneloops, one of the interesting bits is that (AFAIK)
kerneloops.org coped just fine with our added load of reports, due to
being to identify duplicate submissions capably. Perhaps this is an
area where, for certain classes of reports, apport needs to improve
as well.

But perhaps this is a discussion better had on one of the mailing lists
than in a random bug report.

--
Steve Beattie
<email address hidden>
http://NxNW.org/~steve/

Brian, I love the idea of checking '/etc/default/apport'. This also makes the merge proposal much simpler since the changes are strictly limited to the apport prompt plugin. I have linked the new branch and proposed for merging.

Following suggestions from the review, I have submitted another branch for review which now sources the default file instead of parsing it.

Package uploaded to karmic-proposed. Note that the SRU instructions are in bug 465447 (which is also the bug referred to by the changelog entry).

0.8.6 is in karmic-updates. Marking Fixed released.