325048 - [PostgreSQL] Changing the status of a bug from UNCO to NEW or ASSIGNED sets everconfirmed to 1 but leave the status as UNCONFIRMED

Status: RESOLVED FIXED

(Bugzilla :: Creating/Changing Bugs, defect)

Description

[Frédéric Buclin]

I cannot reproduce on MySQL, but both qa222pg and my local Pg installations are affected.

Step to reproduce:

- Open a bug as UNCONFIRMED;
- Click " Accept bug (confirm bug, and change status to ASSIGNED) "
- Submit
- Look at the bug activity: everconfirmed = 1 as expected, but status still as UNCONFIRMED!

Comment 1

[Frédéric Buclin]

And of course, sanitycheck.cgi complains:

Bugs that are UNCONFIRMED but have everconfirmed set: 2653, 2654, 2655

Comment 2

[Frédéric Buclin]

Interesting: if you come back to the bug and click "Accept bug" again, then it becomes ASSIGNED:

[e-mail] 2006-01-28 17:38:14 CET Ever Confirmed 0 1
[e-mail] 2006-01-28 17:38:24 CET Status UNCONFIRMED ASSIGNED

Note the delay of 10 seconds, which is the time I needed to come back to the page and click on "Accept bug" again.

Comment 3

[Frédéric Buclin]

OK, the bug also appears when changing the bug status from UNCO to NEW.

I suspect the problem comes from ChangeStatus() in process_bug.cgi due to:

    $::query .= "bug_status = CASE WHEN bug_status IN($open_state) THEN " .
                                "(CASE WHEN everconfirmed = 1 THEN " .
                                    SqlQuote($str) . " ELSE " .
                                    " 'UNCONFIRMED' END) ELSE " .
                                "bug_status END";

Probably PostgreSQL is unable to change everconfirmed from 0 to 1 and update the bug status based on this change at the same time, i.e. for PostgreSQL, everconfirmed is still 0 when it does this check.
Probably also that MySQL realizes it is updating everconfirmed to 1 and so has to take this update into account when determining the new bug status.

I have no better conclusion. mkanat?

Comment 4

[Frédéric Buclin]

Attachment: patch, v1

We cannot release 2.22rc1 without this fix IMO. Tested on both MySQL and PostgreSQL.

Comment 5

[Max Kanat-Alexander]

Comment on attachment 211026 [details] [diff] [review]
patch, v1

Yep, works. Tested on landfill.

Comment 6

[Frédéric Buclin]

Checking in process_bug.cgi;
/cvsroot/mozilla/webtools/bugzilla/process_bug.cgi,v  <--  process_bug.cgi
new revision: 1.302; previous revision: 1.301
done