- lots of bug fixes going in.
- triaging continues.
- landed patch for bug 99165 "freeze nsIInputStream nsIOutputStream"
- fixed bug 63878 "Implement alert for 'document contains no data' (was Going to lexmark.com does nothing, but www.lexmark works [Connection reset by peer]"
- developed patch for bug 128508 "freeze nsIChannel (nsIRequest)"
- developed patch for bug 116365 "[RFE] Cache partial documents; automatically issue byte range requests"
- developed patch for bug 126755 "should try next ip address on connection timeout failure"
- discussed disk cache stream wrappers w/ gordon (bug 81724)
- reviewed a lot of patches
- triaged bugs
- Working on updating xpcom samples to mozilla 1.0 api.
- Working on xpcom glue library. Will meet with Mike and Brendan to sort out some of their issues with the current design.
- Mime Service was being accessed from a non-ui thread. This causes all kinds of problems. Found that content type is not needed while opening most stream io's. 127918.
- Continued work on bug 81724 "disk cache needs stream wrappers" needed to support partial cache entries.
- Working on bug 109939 - plugins registered in appreg each time they are used, using a flatfile to cache plugin info, instead of using libreg registry.
- Triaged bugs.
- worked on bug 97960 "First time, the Netscape feedback app started - subsequently, the entire Mozilla application closes without warning (does not crash, just terminates)". uploaded the observations on the bug. awaiting comments from brenden.
- worked on bug 125471 "Problems with location HTTP headers, when using method POST". Uploaded the observations. The reporter has indicated that a part of the bug has fixed with 0.9.9. Need more input from reporter to progress further.
- fixed the bug 94938 "Animated image prefs don't apply to background images "
- have patch for the bug 83804 " N621, M098 & Trunk crashes [@ HaveDecodedRow]"
- working on bug 42224 "server JPG push fails after several frames and displays broken image icon". Observed that the problem is in the nsMultiMixedConv.cpp in onDataAvailable.
Plans for next week
- land patches for bug 128508, 116365, and 126755
- fix more bugs :-)
- to work on the assigned bugs
- check in fix for 83289.
- Continue fixing *+ bugs
- Fixed a problem, caused by a short-hand-tag representation of a HTML
tag!, that truncated the character preceding a forward slash (bug 125647).
- Worked on JS line number problem. Discovered a problem
with newlines in end-tags in comments, in NOSCRIPT, and fixed them (bug 111576).
- Worked on a problem pertaining to nested FORM tags
that was causing incorrect FORM submission. Deduced a reduced testcase and
found out that Mozilla was, in fact, doing the correct thing. However, I'm
not sure if this should be in mozilla, for backwards compatibility, or should
be evangelized (bug 128521).
- Investigated a hang in mozilla and found out that the
behavior was somehow related to nsAPromiseString (bug 121841). Reassigned bug to jaggernaut.
- Worked on P3P policy viewer (bug 124503).
- Did code reviews for Shanjian Li, Roger Sidje, and
- Triaged nsbeta1 and nsbeta1+ bugs.
- Worked on my iframe loading changes (bug 52334),
mostly done, had QA test my builds and no large problems found. The remaining
problem is a focus related problem that causes accelerator keys and context
menus to not work.
- Prepared a techtalk on the DOM Implementation and Scriptability
of the DOM.
- Did some W3C DOM standards work.
- Did lots of sr's
- Was virtual Clayton for half the week.
- Did more analysis on bug 81446 (Reloading
leaks memory). At this point, leaking is not the main cause of memory growth
across reloads. The dominating causes now are unbounded growth of certain
data structures and memory fragmentation. See bug for details.
- Looked into bug 77989 (doc.write()
without doc.close() waits 1.2s before painting) and marked it a dupe of bug 81980.
- Worked with Michael Spertus of Geodesic to resolve
a license file issue to unblock me from evaluating Great Circle, their memory
- Fixed hashtable memory-use warning and leak (bug 128861)
- Removed insecure 'trusted codebase' feature (bug 131025)
- Added pref change listener for caps policy db (bug 128697)
- Some minor bug fixes in caps (131342, 131340)
- Talked to PSM folks about lock icon bugs
- Patch uploaded for Bug#121481, 94475.
- Marked 87102 as dupe of 25141.
- Spent some time on 15378, 107927 and 69792.
- Working on 114204.
- Code reviews for harishd, Axel Hecht, Jonas Sicking
- Performance tests for Jonas Sicking.
- Checked in the fix for 129608, parsererror should be
in its own namespace.
- Checked in a fix for 110544, comment parsing is slow
etc. This also fixed bug 130654, an IBM page that we had never displayed correctly but do so now. Investigating
why performance did not improve after checkin.
- Made fix for 126473, unitialized variables in DOM class
- Worked on 86583, can't drag XML links. Found the file
where this happens. This drag problem can also happen on HTML AREA tags.
- Focal feedback almost done.
- Bug triage, including nsbeta1 triage. Untriaged number
has grown too big, but nsbeta1 numbers looking better.
- Working on details regarding summer interns.
Plans for next week
- Work on bug 124503.
- Work on nsbeta1+ bugs.
- Triage untrageted bugs.
- Finish up and present the DOM Implementation and Scriptability talk.
- Land my iframe loading changes if I resolve the remaining
- Do more W3C DOM standards work.
- Finish up a few large sr's.
- On vacation from Monday through Thursday of next week. Will work on
Saturday and Sunday, though.
- Come back from vacation and attack Mozilla 1.0 bugs
approved by ADT/EDT/MozDrivers.
- Keep fixing 1.0 bugs
- Publish first draft of reviewers' guide
- Triage untargeted bugs.
- Work on details regarding summer interns.
Drivers approved about 240 checkins this last week.
We're 1 week from the 1.0 freeze date so approvals are slowing down and drivers are moving to require more testing and more information in approval requests. Please be sure to include in all of your requests a brief description of the problem, your solution, the risk it poses and the testing that's been done to assure that it doesn't regress anything.
We're coming in to the home stretch and we need everyones help to keep the tree green and to keep the builds improving every day.