625058 - Enable crashtest-ipc for Mac machines

Status: RESOLVED FIXED

(Release Engineering :: General, defect, P5)

Description

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

See also bug 623613 and bug 624851.

Tentatively blocking on bug 622229 so that we have the option of running these tests on 10.6 slaves.  The multi-process gfx code *must* USE_WIDGET_LAYERS or else most tests will fail.  If need be, we could limit the tests to 10.5 machines for now to unblock.

We should be fine to run reftest-ipc with or without GL compositing on mac; it should work equally well for both.  Using whatever happens to be default on the machine works for me.

Comment 1

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

I'm seeing firefox-bin/plugin-container just sit apparently idle for long periods of time during startup.  I suspect that message notifications are being dropped in plugin-container.  The reftest harness is very unforgiving of these bugs because it relies on specific sequences of small numbers of messages; there's not periodic spam to bail out p-c.  Unfortunately, I can't run the IPDL unit tests because ipdlunittest-bin is failing to start up in a scary-looking way.  I'll see if there are upstream changes we can pull.

Comment 2

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

Oops, that was supposed to be posted in bug 625060.

Comment 3

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

The reftests are mainly failing for niggly pixel differences due to the duct-tape-and-rubber-band rendering setup.  We're never going to ship out-of-process content with this code, so I don't think it's worth investing time sorting out these failures.

I *do* think it's worth running the crashtests so that we know of big problems caused by platform changes.

Comment 4

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

I think this should block because it's going to be our only automated way to catch broken mac nightlies of fennec, which extension developers are expected to use.

Comment 5

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

In my last try push (263d877928ac),
 - the opt tests timed out in two tests,

TEST-UNEXPECTED-FAIL | file:///Users/cltbld/talos-slave/test/build/reftest/tests/layout/style/crashtests/192408-1.html
TEST-UNEXPECTED-FAIL | file:///Users/cltbld/talos-slave/test/build/reftest/tests/layout/svg/crashtests/612662-1.svg

 - the debug tests were timing out and also failing the same assertion checks as in bug 623613 comment 32

These aren't ready to enable yet.

Comment 6

[Armen [:armenzg]]

Hi cjones according to your comment:
Would you be OK to close this bug until it is ready for us?
Then file another bug or REOPEN when ready.
Sounds good?

Comment 7

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

That's fine.

Comment 8

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

This is an old bug, but some of the analysis may still be valid, so reopening instead of filing anew.

Comment 9

[Aki Sasaki (not active)]

Fixed in bug 624851.