Open Bug 1046271 Opened 10 years ago Updated 1 year ago

Fix up interface for measuring CC pauses

Categories

(Core :: Cycle Collector, defect)

defect

Tracking

()

People

(Reporter: mccr8, Unassigned)

References

(Blocks 1 open bug)

Details

Avi had some suggestions for improving what landed in bug 1019101.  We should also align with what ends up landing for GC.
Blocks: 1017055
Thanks.

The differences between the CC patch which landed already at bug 1019101 and the GC patch at bug 1019611 comment 13 are:

1. The CC patch exposes the functions at Components.utils, while the GC patch uses global chrome functions. The CC approach is already better according to Terrence at bug 1019611 comment 9.

2. The GC patch uses microseconds resolution while the CC patch uses ms. As long as the accuracy is reasonable for the expected values and our usage of them (telemetry buckets which will most probably not be sub-ms), I think ms is OK here.

3. The GC patch returns the last max value from its Clear function. While it's more complete than the existing CC API, I don't believe it would have a meaningful impact on our measurements, if at all.

Overall, I think the CC API is good enough and personally I wouldn't have asked you to put more time into it right now, but your call if you want to prettyfy it and make it match the GC API.
Severity: normal → S3
Component: XPCOM → Cycle Collector
You need to log in before you can comment on or make changes to this bug.