Use keys on pifr.queue() batch response [#544026]

Re: http://d6.drupal.org/node/367072

I submitted a broken and a good patch. The failed patch got assigned to the wrong test result from the master. If you click the "View details" link for make_node_optional_01.patch, it will show the test patch page on the master for extra_spaces.patch. Likewise, clicking extra_spaces.patch's "View details" shows the test page for make_node_optional_01.patch.

When make_node_optional_01.patch failed, it was reported as a failure to extra_spaces.patch. The issue status did not get updated because something in the code still recognized that the result belonged to the older patch which shouldn't trigger a status change on failure (only the newest patch should, e.g. the issue status did get updated on http://d6.drupal.org/node/367068 for a failed patch as it should have been).

Comment	File	Size	Author
#4	544026-key-batch.patch	3.21 KB	boombatower

Comments

Comment #1

andypost

he/him

Russian

commented 9 August 2009 at 21:40

Suppoe it depends on sequence of patches. Maybe before changing status need some check for a newer patches in comments.

Comment #2

boombatower commented 13 August 2009 at 01:32

Assigned:

Unassigned

» boombatower

Took a glance at code and I really don't see how this could happen. I'm gonna need some help reproducing.

Summary: both files point to test 21.

Comment #3

boombatower commented 13 August 2009 at 17:43

Seems to be the order that test_ids are returned...no idea how this never popped up before. I am debating between defining a sort algorithm for the XML-RPC API or some sort of arbitrary key system.