Mozilla 2/StatusMeetings/2007-12-19: Difference between revisions
< Mozilla 2 | StatusMeetings
No edit summary |
No edit summary |
||
| (One intermediate revision by one other user not shown) | |||
| Line 1: | Line 1: | ||
<small>[[Mozilla_2/StatusMeetings/2007-12-12|« previous week]] | [[Mozilla_2/StatusMeetings|index]] | [[Mozilla_2/StatusMeetings/2008-01- | <small>[[Mozilla_2/StatusMeetings/2007-12-12|« previous week]] | [[Mozilla_2/StatusMeetings|index]] | [[Mozilla_2/StatusMeetings/2008-01-09|next week »]]</small> | ||
'''Mozilla 2 Meeting Details''' | '''Mozilla 2 Meeting Details''' | ||
| Line 17: | Line 17: | ||
== Updates == | == Updates == | ||
* XPCOMGC status | |||
** Working through 'make check' in necko. Deadlocks mostly solved through magic in nsAutoLock/nsAutoMonitor | |||
** Necko is being problematic: in particular, the current necko behavior of close-stream-on-last-release is not possible with finalizers. This means that's we'll need to retrofit necko (and perhaps necko clients) with explicit-close of channels. Right now I'm just commenting out assertions and working around problems, but this might not scale. | |||
** But this means that getting performance numbers from a running browser may be quite difficult. Should I instead attempt to create some artificial large-object-graph to test GC performance? | |||
** Going to work on tools to make verification easier: simple dehydra scripts to annotate classes that should be GC or stack objects, and a tool to inform valgrind of live/free MMgc memory | |||
* Discussion of Moz2/next Firefox release | * Discussion of Moz2/next Firefox release | ||
Latest revision as of 07:45, 9 January 2008
« previous week | index | next week »
Mozilla 2 Meeting Details
- Wednesdays - 11:00am Pacific, 2:00pm Eastern, 19:00 UTC
- Mozilla Building S - <script> conference room
- 650-903-0800 or 650-215-1282 x91 Conf# 217 (US/INTL)
- 1-800-707-2533 (pin 369) Conf# 217 (US)
- irc.mozilla.org #mmgc for backchannel
Open Action Items
- ActionMonkey status: JavaScript:ActionMonkey
- bsmedberg is generating builds for comparative perftests... this got lost, will have it today
Updates
- XPCOMGC status
- Working through 'make check' in necko. Deadlocks mostly solved through magic in nsAutoLock/nsAutoMonitor
- Necko is being problematic: in particular, the current necko behavior of close-stream-on-last-release is not possible with finalizers. This means that's we'll need to retrofit necko (and perhaps necko clients) with explicit-close of channels. Right now I'm just commenting out assertions and working around problems, but this might not scale.
- But this means that getting performance numbers from a running browser may be quite difficult. Should I instead attempt to create some artificial large-object-graph to test GC performance?
- Going to work on tools to make verification easier: simple dehydra scripts to annotate classes that should be GC or stack objects, and a tool to inform valgrind of live/free MMgc memory
- Discussion of Moz2/next Firefox release