CycleCollector/CCStatus-2007-03-16: Difference between revisions
Jump to navigation
Jump to search
m (correct navigation link) |
|||
| (6 intermediate revisions by 4 users not shown) | |||
| Line 1: | Line 1: | ||
<small>[[CycleCollector/CCStatus-2007-03-07|« previous meeting]] | [[CycleCollector/Status_Meetings|index]] | [[CycleCollector/CCStatus-2007-03-23|next meeting »]]</small> | |||
=== Agenda === | === Agenda === | ||
* Update from most recent discovery and patch to reduce | * Update from most recent discovery and patch to reduce leas by peterv. | ||
* Update from graydon on landing 368869. | * Update from graydon on landing {{bug|368869}}. | ||
* Update from jst on CC timing code. | * Update from jst on CC timing code. | ||
* Update from dbaron on debugging tools. | * Update from dbaron on debugging tools. | ||
| Line 9: | Line 11: | ||
=== Leak Numbers === | === Leak Numbers === | ||
peterv: Patch ( | peterv: Patch ({{bug|374096}} ) was landed today. Leaks dropped from 700k to 6k. Catches most of it, but there's some tuning to due. | ||
jst: Has a patch for CC timing. Will create a bug and attach patch for review. No update on the XPConnect stuff yet. | jst: Has a patch for CC timing. Will create a bug and attach patch for review. No update on the XPConnect stuff yet. | ||
| Line 25: | Line 27: | ||
=== Next Steps === | === Next Steps === | ||
* Performance: Boris has filed two bugs (bug 373693 (addref) and bug 373694 (release) ) because of address/release. Need to look at these items. David has been doing some analysis of collection of why he's seeing slow down. | * Performance: Boris has filed two bugs ({{bug|373693}} (addref) and {{bug|373694}} (release) ) because of address/release. Need to look at these items. David has been doing some analysis of collection of why he's seeing slow down. | ||
* Collectively look for performance enhancements. | * Collectively look for performance enhancements. | ||
* The tinderbox doesn't show major leaks, but maybe we should go through all the leaks tests. | * The tinderbox doesn't show major leaks, but maybe we should go through all the leaks tests. | ||
* Peterv will look at both of the bugs above for next week. | * Peterv will look at both of the bugs above for next week. | ||
* jst will land the CC timing code | * jst will land the CC timing code ({{bug|374239}}). | ||
=== Audio of This Meeting === | === Audio of This Meeting === | ||
* [ | * [http://shotoku.sicore.org:8080/file-access/default/mozilla/podcasts/Cycle_Collector_Meeting_2007-03-16_11_30am_PST.mp3 MP3 Format 10.6 MB] | ||
* You can subscribe to these meetings as a podcast using [http://shotoku.sicore.org:8080/feeds/mozilla/podcasts/rss2 this RSS2 link]. (Accountability Disclaimer: I didn't test this podcast to make sure it is compatible with anything other than iTunes. Sorry. :] ) | * You can subscribe to these meetings as a podcast using [http://shotoku.sicore.org:8080/feeds/mozilla/podcasts/rss2 this RSS2 link]. (Accountability Disclaimer: I didn't test this podcast to make sure it is compatible with anything other than iTunes. Sorry. :] ) | ||
Latest revision as of 12:26, 26 March 2007
« previous meeting | index | next meeting »
Agenda
- Update from most recent discovery and patch to reduce leas by peterv.
- Update from graydon on landing bug 368869.
- Update from jst on CC timing code.
- Update from dbaron on debugging tools.
- Discuss next steps.
Leak Numbers
peterv: Patch (bug 374096 ) was landed today. Leaks dropped from 700k to 6k. Catches most of it, but there's some tuning to due.
jst: Has a patch for CC timing. Will create a bug and attach patch for review. No update on the XPConnect stuff yet.
dbaron: Nothing of interest. Going to do some lower level leak tests. Seeing the usual mess of font config junk.
Status of Pending CC Code in Bugzilla
- Add cycle collection to RDF datasources
- Make XPConnect traverse more JS edges
- Support cycle collection of refcounted non-xpcom objects (aka, the XBL patch)
- Make cycle collection suspect all native wrapper roots
- cycle collector shutdown dump based on known should-have-been-freed objects
Next Steps
- Performance: Boris has filed two bugs (bug 373693 (addref) and bug 373694 (release) ) because of address/release. Need to look at these items. David has been doing some analysis of collection of why he's seeing slow down.
- Collectively look for performance enhancements.
- The tinderbox doesn't show major leaks, but maybe we should go through all the leaks tests.
- Peterv will look at both of the bugs above for next week.
- jst will land the CC timing code (bug 374239).
Audio of This Meeting
- MP3 Format 10.6 MB
- You can subscribe to these meetings as a podcast using this RSS2 link. (Accountability Disclaimer: I didn't test this podcast to make sure it is compatible with anything other than iTunes. Sorry. :] )