967467 - Jit-test basic\bug698584.js timesout on Windows XP test machines for debug builds

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Dan Minor [:dminor]]

+++ This bug was initially created as a clone of Bug #959155 +++

It turns out the failure on Windows XP debug was a timeout, not an out of memory, and needs to be dealt with separately.

Comment 1

[Dan Minor [:dminor]]

Based on the try run here [1], this will run for at least 20 minutes and the test run will be killed due to 1200 seconds of no output.

[1] https://tbpl.mozilla.org/?tree=Try&rev=6f78e053742c&showall=1

Comment 2

[Dan Minor [:dminor]]

Given more time, they will run to completion in 9776 seconds.

Comment 3

[Dan Minor [:dminor]]

From what I can tell without having a profiler available, it looks like the allocations are just taking a really, really long time in this case.

I think the next step would be to ask releng to install development tools on the loaner test machine so that I and debug and/or profile to see exactly what is going wrong here.

Before I do that, I wanted to make sure that this test is worth spending more time on or if we should just remove it from the test suite or mark it as 'slow'.

Comment 4

[Terrence Cole [:terrence]]

This test takes 0.18s to run locally and uses 696M of memory. The specific thing it is testing is a missing OOM check when JaegerMonkey stubs into our regex machinery. Given that JM is just gone, I can't see this regression test being particularly useful in future.

r=me to remove this test.

Comment 5

[Dan Minor [:dminor]]

Thanks! pushed to https://hg.mozilla.org/integration/mozilla-inbound/rev/7ec2f0199218

Comment 6

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/7ec2f0199218