CycleCollector/CCStatus-2007-05-04

« previous meeting | index | next meeting »

Agenda

• status of patches in progress
• performance requirements
• next steps on performance and on memory leaks

Status of Pending CC Code in Bugzilla

Last week's next steps: Wait for patches (first 4 of 5) to land, then test more:

• Add cycle collection to RDF datasources: leak regressions remaining?
• Support cycle collection of refcounted non-xpcom objects (aka, the XBL patch)
• Make cycle collection suspect all native wrapper roots: backed out.
• Switch nsXPConnect::Traverse to use tracing
• Don't try to collect documents, windows and elements of actively viewed documents

More Next Steps from last week:

bug 378514 was around a 30% performance improvement in the duration of pauses using one sample workload (and that was with an accidental extra call to canonicalize() that was taken out before landing).

Do we know how much bug 378595 helped and how much bug 378987 and doing less XPCCallContext construction (maybe in bug 377884) will help?

Next steps: wait for bug 378987 to land, and for peterv to fix XPCCallContext construction performance problem in nsXPConnect::Traverse (either in bug 368774 or a new bug to be filed). Then measure and profile collection pauses again.

Performance requirements

For discussion: How important is the overall performance degradation (e.g., AddRef/Release overhead)?

Next steps on leaks

• Jonas will write a test case to generate cycles via document.

Next steps on performance (pauses)

Next steps on performance (general degradation)