Platform/2008-10-21: Difference between revisions
< Platform
Jump to navigation
Jump to search
| (14 intermediate revisions by 5 users not shown) | |||
| Line 4: | Line 4: | ||
* Tuesday - 11:00am Pacific, 2:00pm Eastern, 19:00 UTC | * Tuesday - 11:00am Pacific, 2:00pm Eastern, 19:00 UTC | ||
* Mozilla Building S - <script> conference room | * Mozilla Building S - <script> conference room | ||
* 650-903-0800 or 650-215-1282 | * 650-903-0800 or 650-215-1282 x92 Conf# 8605 (US/INTL) | ||
* 1-800-707-2533 (pin 369) Conf# 8605 (US) | * 1-800-707-2533 (pin 369) Conf# 8605 (US) | ||
* irc.mozilla.org #shiretoko for backchannel | * irc.mozilla.org #shiretoko for backchannel | ||
| Line 21: | Line 21: | ||
* Approach: keep non-critical history data in temporary tables, periodically flushing to permanent tables in the background. | * Approach: keep non-critical history data in temporary tables, periodically flushing to permanent tables in the background. | ||
* Problems | * Problems | ||
** Ts | ** Ts ({{bug|459197}}) | ||
** Tp | *** Initial landing caused a regression of ~3% on linux and ~4% on windows ([https://bugzilla.mozilla.org/show_bug.cgi?id=459197#c2 data]) | ||
*** With fixes, still seeing regressions: ([https://bugzilla.mozilla.org/show_bug.cgi?id=459197#c12 data]) | |||
**** Linux is at 1-2% (30-40ms) | |||
**** Windows XP is at ~2% (~20ms) | |||
**** Windows Vista is at 1.2-2.5% (7-15ms) depending on which boxes you look at | |||
** Tp ({{bug|459235}}) | |||
*** Initial landing caused a regression of about ~3% on OS X; possible regression on windows ([https://bugzilla.mozilla.org/show_bug.cgi?id=459235#c1 data]) | |||
*** With fixes, possibly only small regression on mac, and ~1% win on linux ([https://bugzilla.mozilla.org/show_bug.cgi?id=459235#c10 data]) | |||
* Solutions | |||
** This work is not going to make 3.1; push to next release and make no improvements in this area | |||
** slip beta 2 until we get the performance issues fixed with other performance improvements | |||
** let this work land and accept the fact that we'll take some small hits in some performance benchmarks | |||
=== GFX 1.9.1 Update === | === GFX 1.9.1 Update === | ||
* [http://tinyurl.com/5xfmvp GFX blocking 1.9.1+] | * [http://tinyurl.com/5xfmvp GFX blocking 1.9.1+] | ||
* [http://tinyurl.com/4zv2kn GFX wanted 1.9.1+] | * [http://tinyurl.com/4zv2kn GFX wanted 1.9.1+] | ||
=== Mac OS X 1.9.1 Update === | |||
* working on blockers | |||
* Josh working on tests and various plugin bugs as well | |||
* further filesystem work still in review | |||
=== Layout 1.9.1 Update === | === Layout 1.9.1 Update === | ||
| Line 41: | Line 57: | ||
=== JS 1.9.1 === | === JS 1.9.1 === | ||
*[http://tinyurl.com/5gmg8y JS 1.9.1 Bugs] | *[http://tinyurl.com/5gmg8y JS 1.9.1 Bugs] | ||
m-c blockers hurting daily browsing: | |||
* Google Maps malfunction (https://bugzilla.mozilla.org/show_bug.cgi?id=460024) | |||
Blocking JIT by default: | |||
* multi trees (https://bugzilla.mozilla.org/show_bug.cgi?id=450833) | |||
* slow script timeout (https://bugzilla.mozilla.org/show_bug.cgi?id=450000) | |||
* v8-richards misbehavior (https://bugzilla.mozilla.org/show_bug.cgi?id=456384) | |||
* Running Dijit Tree Test with Content JIT / Ubiquity (https://bugzilla.mozilla.org/show_bug.cgi?id=451843) | |||
* correctness bug with undefined turning to NaN (https://bugzilla.mozilla.org/show_bug.cgi?id=456479) | |||
Not blocking JIT by default: | |||
* jsMSX crash (https://bugzilla.mozilla.org/show_bug.cgi?id=458227) | |||
* SCUMM VM crash (https://bugzilla.mozilla.org/show_bug.cgi?id=455547) | |||
* eval/let combo (https://bugzilla.mozilla.org/show_bug.cgi?id=457065) | |||
* CRASHED signal 10 SIGBUS using regress-271716-n.js (https://bugzilla.mozilla.org/show_bug.cgi?id=460518) | |||
* js1_8/genexps/regress-380237-01.js: InternalError: too much recursion (64bit) (https://bugzilla.mozilla.org/show_bug.cgi?id=457660) | |||
* LIRBufWriter null checking (https://bugzilla.mozilla.org/show_bug.cgi?id=451322) | |||
=== Mobile 1.9.1 Update === | === Mobile 1.9.1 Update === | ||
| Line 54: | Line 91: | ||
=== Roundtable === | === Roundtable === | ||
* feedback on try-talos? | * feedback on try-talos? Please add to {{bug|461012}} | ||
* mothballing 1.8 / 1.9.0 machines? | |||
* any upcoming tool changes after branch for FF3.1? Nope. | |||
Latest revision as of 21:58, 21 October 2008
« previous week | index | next week »
Details
- Tuesday - 11:00am Pacific, 2:00pm Eastern, 19:00 UTC
- Mozilla Building S - <script> conference room
- 650-903-0800 or 650-215-1282 x92 Conf# 8605 (US/INTL)
- 1-800-707-2533 (pin 369) Conf# 8605 (US)
- irc.mozilla.org #shiretoko for backchannel
Notices / Schedule
- Beta 2 Schedule
- en-US string freeze is Thursday, October 30th at 11:59pm PDT
- code freeze is Tuesday, November 4th at 11:59pm PDT
- Please take a look at the feature list and get things missing added, and update status for things you own: Firefox3.1/Features#Gecko_1.9.1
- See the security review schedule page to get your review scheduled ASAP!
Firefox 3.1 Update
Places Fsync-reduction Changes
- Goal: reduce fsyncs by writing to the disk less often, reducing hangs/freezes/sluggishness while browsing (especially hurting mobile, usb and some linux).
- Approach: keep non-critical history data in temporary tables, periodically flushing to permanent tables in the background.
- Problems
- Ts (bug 459197)
- Tp (bug 459235)
- Solutions
- This work is not going to make 3.1; push to next release and make no improvements in this area
- slip beta 2 until we get the performance issues fixed with other performance improvements
- let this work land and accept the fact that we'll take some small hits in some performance benchmarks
GFX 1.9.1 Update
Mac OS X 1.9.1 Update
- working on blockers
- Josh working on tests and various plugin bugs as well
- further filesystem work still in review
Layout 1.9.1 Update
- 1.9.1 Layout Bugs
- Video (roc/cdouble):
- Acid3: See comment 66 in the bug
- SVG CSS, roc:
Content 1.9.1 Update
- Content 1.9.1 Bugs
- Worker Threads:
- Cross-site XHR:
JS 1.9.1
m-c blockers hurting daily browsing:
- Google Maps malfunction (https://bugzilla.mozilla.org/show_bug.cgi?id=460024)
Blocking JIT by default:
- multi trees (https://bugzilla.mozilla.org/show_bug.cgi?id=450833)
- slow script timeout (https://bugzilla.mozilla.org/show_bug.cgi?id=450000)
- v8-richards misbehavior (https://bugzilla.mozilla.org/show_bug.cgi?id=456384)
- Running Dijit Tree Test with Content JIT / Ubiquity (https://bugzilla.mozilla.org/show_bug.cgi?id=451843)
- correctness bug with undefined turning to NaN (https://bugzilla.mozilla.org/show_bug.cgi?id=456479)
Not blocking JIT by default:
- jsMSX crash (https://bugzilla.mozilla.org/show_bug.cgi?id=458227)
- SCUMM VM crash (https://bugzilla.mozilla.org/show_bug.cgi?id=455547)
- eval/let combo (https://bugzilla.mozilla.org/show_bug.cgi?id=457065)
- CRASHED signal 10 SIGBUS using regress-271716-n.js (https://bugzilla.mozilla.org/show_bug.cgi?id=460518)
- js1_8/genexps/regress-380237-01.js: InternalError: too much recursion (64bit) (https://bugzilla.mozilla.org/show_bug.cgi?id=457660)
- LIRBufWriter null checking (https://bugzilla.mozilla.org/show_bug.cgi?id=451322)
Mobile 1.9.1 Update
General 1.9.1 Updates
- offline cache, dcamp:
- localStorage, honzab:
- HTML5 drag-drop:
Security
Security Reviews
Roundtable
- feedback on try-talos? Please add to bug 461012
- mothballing 1.8 / 1.9.0 machines?
- any upcoming tool changes after branch for FF3.1? Nope.