Re: Potential blockers for 2.16

["Elijah Newren" <newren gmail com>]

On 9/2/06, Brent Smith <gnome nextreality net> wrote:
> So I was thinking about how I could identify some of the major issues
> that might be blockers for this release tonight, and the following is
> what I came up with:
>
> mysql> select b.bug_id, b.short_desc, count(1)
>           from bugs b, duplicates d
>          where b.creation_ts  > DATE_SUB(CURDATE(), INTERVAL 30 DAY)
>            and b.bug_status != 'RESOLVED'
>            and b.bug_id = d.dupe_of
>       group by b.bug_id
>         having count(1) >= 5;
>
> +--------+------------------------------------------------+----------+
> | bug_id | short_desc                                     | count(1) |
> +--------+------------------------------------------------+----------+
> | 350758 | Nautilus crashes after browsing some folders   |       14 |
> | 350975 | crash on Computer                              |        7 |
> | 351651 | Crash creating the Python module               |        7 |
> | 351713 | crash on Computer                              |        6 |
> | 352686 | crash in file chooser (gtk_file_system_handle_ |       13 |
> | 352736 | crash in Bug Report Tool: properties in the ico|        5 |
> | 353348 | crash in Bug Report Tool:                      |        6 |
> +--------+------------------------------------------------+----------+
> 7 rows in set (0.96 sec)
>
> For the non-SQL savvy, this query finds all bugs in the last
> month that are not resolved, and have 5 or more duplicates filed against
> them.
>
> Out of these, it appears that only 353348 has a (untested) patch to fix
> the issue.
>
> 350758 and 352686 seem like the most critical blockers here, as they
> have the most duplicates.
> I think we should block the release on these two bugs.
>
> Please discuss (let me know if I should send this to another m-l for
> discussion - I don't know if release-team@ is the appropriate place for
> discussion to take place)

Awesome, thanks for looking at this!

Bug 350758 is actually irrelevant to Gnome 2.16; the stack trace shows
it crashing in liboobs which, as previously announced, will not be
part of the 2.16 release.  (we are using gnome-system-tools 2.14.x
instead).  Bug 351376 is irrelevant as well.

All duplicates of bug 352736 were filed by the same person as the
reporter of bug 352736; in fact, all descriptions and stack traces are
identical, so this bug represents someone just resubmitting like
crazy.

Fer has a patch in bug 353348, which he says he asked for r-t approval
for in the bug (which I think I remember seeing the last couple days;
I'll double check...).  So, hopefully that one should be taken care
of.

Of the bugs in your list, 4 are needinfo.  I don't like holding up a
release for a needinfo bug, because it could be weeks or even months
before the additional information requested is received.  However, a
quick call to action on the bugsquad or gnome-love to try to get
people to reproduce and provide the extra needed information might be
useful.

I agree with you about bug 352686, as does Kjartan as mentioned in a
separate email.  We should probably make an announcement / call for
help.  The evo bug brought up by Nickolay should probably be added to
the list too given the sheer number of duplicates.  So the libgnomeui
bug you point out and Nickolay's evo bug sound like good blocker
candidates to me.

How does everyone else feel about blocking the release on these two
bugs?  If we do so, we should probably send out an email soon, stating
that we're doing so, reminding that hard code freeze lasts until the
release, and calling for help in fixing those bugs.

Thanks Brent,
Elijah