maintained by Tim Rowley <firstname.lastname@example.org>
Last Updated Wednesday October 24, 2001
This status update page is updated every
week. To get updates and news throughout the
week you are invited to check out mozillaZine
a site devoted to Mozilla advocacy.
- Mozilla 1.0 manifesto
posted; 1.0 tracking metabug created.
Please do not change the dependencies for this bug. Please do not
begin a discussion in this bug, but in a newsgroup or by mail. If you
add yourself to the CC list, please do not add a comment.
- 0.9.5 Milestone
Speakers lined up for Nov. 9 Developer Day
- New document on Mozilla Modules and Module Ownership now available
- Michael Ang to speak at the Silicon Valley Linux Users Group on December
5. Topic is Building Applications with Mozilla and XPCOM
I have checked in the code for [Bug 76070] we need namespace support from the DOM. This code gives the
transformiix DOM namespace IDs and atoms for localnames. This will give
us some great tools for both performance and correctness and was one of
our major architecture blockers. Thanks to Jonas Sicking for starting
this and both him and peterv for great input after I took over this bug.
Kudos for sr go to jst.
- fixed bug 103979 "Crash if autoconfig proxies prefs enabled by the default"
- fixed bug 102743 "crash pressing back/forward on oreilly.com"
- nearing completion of the new standard url implementation
-- probably 2 more days worth of work. matches behavior of current standard
url implementation on all of our URL tests.
- tested a linux build with mach-o patch applied -- seems
to work fine.
- triaged some bugs.
- Fixed 103991 "remove IsAsciiString from nsDNSService and nsHttpChannel".
- Fixed 103167 "netwerk MANIFEST files out of date, remove
obsolete cache source files"
- Fixed 101857 "support cache entries with no data"
- Fixed 92908 "prefs inconsistency causes cache session
creation to fail"
- Fix 101651 "nsStdURL::SetSpec can trash memory"
- Landed API freeze to nsIServiceManager, nsIObserver, and nsIObserverService.
- attended a meeting on jud's proposal for 'delivering
software in CPD'
- Finally got review for bug 96461- ComponentManager cleanup.
Will check it in shortly.
- Triaged a lot of bugs.
- bug 83163. fixed FMW in gtk timer code.
- bug 104999. checked in the first few parts... working
on shrinking the footprint by doing a better job of storing images in the
- bug 104999. Helps startup a little by making winmm.dll
not be loaded at startup but instead when it is first needed.
- bug 92453. Removed <img> lowsrc support.
- Laptop died. Ticket been filed-- hoping for a replacement
- nsIObserver landing was not as graceful as I would have liked. Change
touch many subsystems, some of which could not be verified. I should have
cut a branch and produced experimental builds.
Plans for next week
- triage more bugs
- land standard url patch
- fix more bugs for 0.9.6
- Start looking at the nsIComponent freeze dependencies.
- Continue on bug 104999.
- Continue fixing bugs.
- jst went mad and fixed 11 bugs! If you see him sleepwalking
on the campus, please guide him to the B21 bridge area and gently lay him
down on the sofa to recover from the ordeal ;)
- We have too many bugs on our 0.9.6 list, needs triaging.
jst is doomed with his untriaged list (over 100) and would need some help.
- heikki will be on vacation 10/26.
- Fixed crasher bug 104774. Added null check in nsStyleLinkElement
before calling CallQueryInterface() on mStyleSheet, which will be null if
there is no stylesheet.
- Fixed bug 70613. element.innerHTML came up empty if element
was not part of a document hierarchy.
- Fixed bug 79133. Image onerror events cause window.onerror
handlers to be called since the events bubble up the tree.
- Fixed bug 103188. Made window.prompter [noscript] since
it's not accessible from web scripts anyway.
- Fixed bug 76049. Made access past the end of a 'DOM array'
return undefined and not null.
- Fixed bug 104031. Bad string code in nsHTMLDocument::WriteCommon().
- Fixed bug 102583. Fixing bad old string usage in nsXMLContentSink.
- Fixed part of bug 104297. Made calling window.close()
on an already closed window not throw exceptions and not assert.
- Fixed bug 101492. Made self.close() do nothing when called
from within a frame in a frameset (or from an iframe) for backwards compatibility.
- Fixed most of bug 62536. Adding support for the properties
scrollTop, scrollLeft, scrollHeight, scrollWidth, clientHeight, clientWidth
and the method scrollIntoView() to elements for compatibility with IE. Patch
- Fixed bug 105572. Don't bother GC'ing on context destruction
when we're tearing down a frame in a frameset (or an iframe) since the tear-down
of the frameset will do a GC.
- Reviewed (almost done) John Keiser's form submission rewrite.
- Landed fixes for a bunch of 0.9.6 bugs.
- Worked on parser performance issues.
- Discussed with Dave Barrowman about the plans for landing
compact policy changes.
- Code reviews.
- Checked in some cleanup of Transformiix code.
- Looked at my topembed bug (102209 - When copying text
from composer to AOL, it takes out spacing randomly).
- Worked on the XSLT output rewrite.
- Did reviews for sicking, Pike, jst.
- I will be on vacation 10/26.
- Attended AOLU course on Thursday and Friday.
- Worked on 41983, lots of testing and some fine tuning.
- Code reviews for harishd.
- Bug triage.
Plans for next week
- Fix some more of my 096 bugs.
- Bug triage.
- Triage... Triage... Triage... bugs.
- Land compact policy changes.
- Work on parser performance issue.
- If time permits (!!) work on 0.9.6 bugs.
- Work on my topembed bug (102209 - When copying text from composer to
AOL, it takes out spacing randomly) and the XSLT output rewrite.
- Land 41938.
- Work on FIXptr.
- Good progress on Message Compose performance and Folder Loading
made this week.
- A number of top crashers fixed this week.
- Resolved as fixed
26 bugzilla bugs.
Performance and Footprint
- Seth started working on
104989. This would improve the performance of opening a compose
window on the next window after a send. We'll hide the most recently
used compose window and then reuse it on the next compose operation. Results
look promising so far. If this works out we'll look into the costs
of making it behave this way for the first compose window.
- Cavin and David worked on
folder loading performance. They have an improvement for UpdateFolder()
which looks to have some promising results. The overall improvement
appears to range from 4%-8% based on 1000 and 4000 message folders, where
the larger folders see the larger improvement. The improvement to UpdateFolder()
itself ranges from 17%-30% based on the size of the folder. The improvement
involved a reduction in the number of times we call memset and memcpy.
- Made it so that when removing an account from the Account Settings
dialog, the current panel doesn't stay on the deleted account.
- Worked on performance changes for UpdateFolder() which will speed
up folder loading.
- Completed Quick Search. The review is under way.
- Made the pop3 message downloading limit size pref work again.
- Fixed a bug where there was no error when creating a duplicate local
folder through the filter ui.
- Created an initial design doc for the
manual execution of filters.
- Fixed an imap top crasher.
- Created XBL widgets for the message headers.
- The new XBL widgets were used to do View All Headers.
- Started work on the Labels feature. He worked on getting the
new column to show up and started working on the preference panel.
- Fixed a crash that occurred while emptying the trash folder.
- Fixed a crash when you tried to rename a local folder while it's
- Fixed a MIME crash.
- Fixed a problem which caused a bunch of assertions when searching
newsgroups if you hadn't opened the newsgroup.
- Worked with Cavin on the folder loading optimizations mentioned above.
- He started rewriting the way we store attachment information in the
message compose window. This will let us store more information about
the content-type and the attachment name. We have a number of bugs
where we don't show a good name for the attachment or where garbage shows
up in the attachment pane in international builds. This should allow
us to fix these.
- Prefill Filters is going through review.
- Started working on Block Addresses.
- Investigated using the new Wizard for the Account Wizard. Converting
to this will solve a lot of the accessibility problems we have with the current
- Started working on the compose window performance improvement mentioned
- Bulletproofed a topcrash in outliner.
- Created a default height for the thread pane, thereby preventing
an extra paint on startup.
- Started working on landing Address Book Search.