Open Bug 1401764 Opened 7 years ago Updated 2 years ago

Reduce content process leak threshold to 0

Categories

(Testing :: General, enhancement, P3)

Version 3
enhancement

Tracking

(Not tracked)

People

(Reporter: mccr8, Assigned: mccr8)

References

(Depends on 2 open bugs)

Details

(Whiteboard: [MemShrink:P2])

I could have sworn the threshold had already been reduced to 0, but it appears not! Hopefully we can reduce it without too much trouble.
I did a try push with the leak threshold reduced to 8, on Linux and Windows:
  https://treeherder.mozilla.org/#/jobs?repo=try&revision=740a084371d01fd0869c574e60652269dd3037e1&selectedJob=132521783

There's a permaleak in dom/base, at least on Linux and Linux64. This looks the same as bug 1189646.

There's some kind of leak only on Windows, in most mochitests:
TEST-UNEXPECTED-FAIL | leakcheck | tab process: 12 bytes leaked (nsStringBuffer, nsTArray_base)

This was a push without bug 1401412, so it is possible that 12 byte leak is the same thing, just slightly bigger on Windows. But I do see it on recent treeherder pushes, so I'm guessing it is different. Digging into the Windows failures isn't worthwhile until I have a push with the threshold increased to 12 to avoid this.
Whiteboard: [MemShrink]
Depends on: 1404482
Depends on: 1404486
Depends on: 1408523
Depends on: 1408552
Depends on: 1408554
Depends on: 1408558
Depends on: 1409075
Whiteboard: [MemShrink] → [MemShrink:P2]
Depends on: 1411259
Depends on: 1411416
Depends on: 1411644
Depends on: 1412408
Priority: -- → P3
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.