1023719 - Report notable classes in the JS memory reporter

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect, P2)

Description

[Nicholas Nethercote [inactive]]

Part 6 of bug 972712 was backed out for causing intermittent ASAN failures on Mochitest-o. I suspected at the time that the patch was just exposing a latent bug. Recent runs on try server seem to indicate that the old failure isn't happening any more. So I will try landing this again.

Comment 1

[Nicholas Nethercote [inactive]]

Attachment: Report notable classes in the JS memory reporter

Here's the rebased patch.

Comment 2

[Nicholas Nethercote [inactive]]

Comment on attachment 8438186 [details] [diff] [review]
Report notable classes in the JS memory reporter

Review of attachment 8438186 [details] [diff] [review]:
-----------------------------------------------------------------

Carrying over r=till from bug 972712.

Comment 3

[Nicholas Nethercote [inactive]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/277bacd9f2f8

Comment 4

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/277bacd9f2f8

Comment 5

[Nicholas Nethercote [inactive]]

Backed out for causing bug 1024996. The intermittent failures relating to base shapes are still there, sigh. I still think there's an underlying problem with base shapes that this patch is exposing. I'll probably end up changing the patch so that it does the per-class reporting for objects, but not for shapes.

Comment 6

[Nicholas Nethercote [inactive]]

And the backout failed because I was out of date, and now the tree is closed...

Comment 7

[Nicholas Nethercote [inactive]]

... and it's open again:

https://hg.mozilla.org/integration/mozilla-inbound/rev/43735d730427

Comment 8

[Florian Bender]

(In reply to Nicholas Nethercote [:njn] from comment #7)
> ... and it's open again:
> 
> https://hg.mozilla.org/integration/mozilla-inbound/rev/43735d730427

Hmm this didn't seem to land on m-c, or someone forgot to update the bug. What happened?

Comment 9

[Andrew McCreight (out of office until 8/21) [:mccr8]]

You can find the commit for a merge to central by just changing the middle part of the commit URL and seeing if it is there.

merge of backout: https://hg.mozilla.org/mozilla-central/rev/43735d730427

Comment 10

[Nicholas Nethercote [inactive]]

I just did a try run where I triggered M-o 73 times, and ASAN was happy with every run:
https://tbpl.mozilla.org/?tree=Try&rev=73055b184d9f

I'm hopeful that whatever latent bug this patch exposed has been fixed. I'll try landing it again.

Comment 11

[Nicholas Nethercote [inactive]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/d67d68ac4119

Fingers crossed it sticks!

Comment 12

[Nicholas Nethercote [inactive]]

(In reply to Nicholas Nethercote [:njn] from comment #11)
> https://hg.mozilla.org/integration/mozilla-inbound/rev/d67d68ac4119
> 
> Fingers crossed it sticks!

It stuck. Hooray.