653310 - ipc:content-shutdown notification maybe should always set a dumpID property

Status: RESOLVED FIXED

(Core :: IPC, defect)

Description

[Cameron McCormack (:heycam)]

The test dom/ipc/tests/test_process_error.xul fails in my Ubuntu VM, due to breakpad not being able to attach to the crashing content process (which is apparently a known issue).  Since there's no minidump, the property bag that is sent as the subject of the ipc:content-shutdown notification does not have a "dumpID" property.  The test script throws when trying to get it, which causes the test timeout.

Plugin crashes OTOH always set a "pluginDumpID" property for plugin-crash notifications.  If there was no minidump available, it uses an empty string value.  modules/plugin/test/mochitest/test_crash_notify.xul (for example) will still fail, but it won't cause the test to timeout.

It seems like these two should be consistent.  Should we set dumpID = "" in this case?

Comment 1

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Sounds good.  Should be a one-liner; want to write the patch? ;)  See http://mxr.mozilla.org/mozilla-central/source/dom/ipc/ContentParent.cpp#287 .

Comment 2

[Cameron McCormack (:heycam)]

Sure.  And the test should still fail if dumpID == "", right?

Comment 3

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Yep.

Comment 4

[Cameron McCormack (:heycam)]

Attachment: Always put a dumpID property on ipc:content-shutdown notification property bags.

Comment 5

[(not currently active) Ted Mielczarek]

(In reply to comment #2)
> Sure.  And the test should still fail if dumpID == "", right?

I was just going to echo this sentiment. It's fine to have the data be consistent, but if we're failing to write a dump we are still failing the test.

Comment 6

[Cameron McCormack (:heycam)]

Landed: http://hg.mozilla.org/mozilla-central/rev/9f8e09d29a0d