MailNews:

• RTM++ Bugs Fixed last week (15) Bugscape (2)
• non-mail Bugs! Fixed last week (4) Bugscape (0) (bugs fixed by other checkins)
• not-rtm++ Bugs Fixed last week (14) Bugscape (0) (bugs fixed by other checkins, CCK, some trunk work)
• Total Bugs closed last week (104) Bugscape (6)
• new untriaged RTM (0) Bugscape (0)
• open RTM Need Info (5) Bugscape (0)
• open RTM+ (0) Bugscape (0)
  
• open RTM++ (0) Bugscape (0)
• Held our first 6.5 performance analysis meeting, will continue bi-weekly for a while
• Fixed a number of rtm candidate bugs that won't go in the release.
  
• Seth files 23 new bugs this week! 39 for the group. Thanks for using the product and filing bugs!
  

DATE	Open RTM++	RTM Need Info/RTM+	UnTriaged RTM	Fixed RTM++	Fixed non-RTM++	Verified for RTM
10/2	0	23	13	0	0	8
10/9	5	17	6	27	9 WFMs	13
10/16	1	8	6	27	17 CCK, WFMs, minor trunk work	20 33 bugs filed
10/23	1	18	1	9	6 CCK, WFMs, minor trunk work	21 20 bugs filed
10/30	0	19	0	9	23 CCK, WFMs, minor trunk work	28
11/6	0	5	0	17	14 WFMs, minor trunk work	8 Helped QA w/candidate testing for Editor

Profile Manager:

• One outstanding bug w/JA profile names
  

CCK:

• No current activity, some new minor bugs on horizon

Summary

• Lots of planning underway for life after 6.0

Highlights

• darin
  • Investigated bug 57679 "failure to read remaining data from PR_POLL_HUP'd channel"
  • Accepted patch from harbaugh@ncifcrf.gov with minor revision.
  • Investigated bug 58655 "nsSyncHelper uses a native event queue"--Testing patch suggested by dougt.
  • Investigated bug 32359 "Wrong referrer given to images after redirect" --Verified that it is no longer a bug; our behavior matches NS 4.75.
  • Confirmed crasher bug 57589 "Clicking on URL from within GAIM".
  • SyncHTTP: planning the API, evaluating how much of a performance gain it would mean.
  • Developing test cases. Added a test program to the mozilla/netwerk/test on the trunk
  • Asynchronous DNS on Unix: discussion with Gordon and netlib newsgroup. Coded a thread-pool based reference implementation for testing.
  • Investigated bug 58099 "what's related shows up blank"-- Looks like a server problem; can duplicate the problem using telnet.
  • Finally received PO for windows machine!
• neeti
  • Currrently investigating top crasher bug 55616, 52818.
  • Created a document for post 6.0 cache work.
• pnunn
  • Closed bugs: 56608, 56425, 58318, 57748, 57773,
  • Helped Judson locate some test urls for cgi generated images. (bug 55914)
  • Wrote up drafts of ImgLib 6.x plans. Posted on status.mcom.com
  • Studied new crasher bugs: 52912, 58071-- neither turned out to be show stoppers.
• gordon
  • Landed fix for Bugscape bug 3003 "Unable to activate a new/existing member on Mac only" on RTM branch
  • Wan-Teh is on vacation, so I was unable to get approval to land on other NSPR branches this week.
  • Closed Bugzilla bug 52534 "On the Mac, opening cache file 4 times" as INVALID -- it was a simple misunderstanding concerning how the cache works.
  • Began planning for future (ImgLib 2.0 meeting, Async DNS discussions w/Darin, investigating problems with NSPR & Security)
• gagan
  • misc planning work.

Lowlights

• neeti
  • Hard to get traction on crasher bugs 55616, 52818 as they are difficult to reproduce.

Plans

• pnunn
  • Continue writing up ImgLib 6.x Plans
• neeti
  • Work on bugs 52818, 55616
  • Synchronous HTTP development, work on time schedule.
• gordon
  • Start to look forward (Async DNS, Cache renovation, NSPR cleanup, etc.)
• darin
  • Run SyncHTTP evaluation tests.
  • Work on API changes with Neeti

Issues/Escalations

• None

Highlights

• Came up with a report for jhstaff that describes memory consumption over time. Also breaks down code/DLLs by size. The report shows that both heap and code reduction are a significant factor in achieving our footprint goals.

Engineer Status

• jband
  • no status
• kandrot
  • no status
• dprice
  • Helped AOL get their CDT working properly (still not quite there)
  • worked on getting SCC's string factoring tests to compile on win32
  • modified the zip cache test, made it more flexable for automated testing
  • wrote diffing-test.pl which runs a test program and compares its output to expected output
  • various meetings
• waterson
  • Presented first cut at memory status to jhstaff. Thanks to kandrot, vidur, warren for helping me get data together.
  • Gathered some profiles for mail/news performance meeting.
  • Spent most of the week trying to determine where the memory's going. I've been banging on the tools, partly to get them more usable, partly just to collect data for roll-up and to determine what to attack next.
  • Improved GC's RTTI heuristics on Linux. Makes us crash less when looking for leaks.
  • Wrote a "leak soup lite" perl script that lowers the bar to entry for using beard's GC stuff to find leaks. Need to get this checked in.
  • Modified brendan's trace-malloc stuff; combined with "leak soup" ideas to create a tool that dumps a comprehensive object map. In other words, it shows you everything that's in memory, all threaded up. Hope to refine this and use it to find the "big fish".
  • Helped create confusion with respect to Bug 59102
  • vidur
    • Pariticipated in task-list reviews for architecture, embedding, layout, rendering and XML/DOM groups.
    • Participated in presenting memory status and near-term strategy to Jim Hamerly's staff.
    • Reviewed bug fixes for harishd, jst, heikki, nisheeth.
    • Wrote initial wording for some requirements XML Protocol Working Group. Pariticipated in weekly conference call and WG list discussion.
    • Minor additional fixes to parser "no-copy" code.
    • Started writing new web applications document.
  • waldemar
    • Prepared for four-day ECMA meeting next week (which we're hosting).
    • Made many changes in the ECMAScript proposed specification.
  • rpotts
    • Bug #54725 - Java does not shut down correctly and hangs the browser. I spent *alot* of time isolating this problem and trying to come up witha solution. It turns out that until Javasoft can ship a new versionof its Java Plugin (with a patch for this bug) java applets will *not*run correctly when mozilla is embedded :-( I have a "low risk" band-aidpatch that allows the current Java Plugin to shutdown when inside of SeaMonkey. I'm trying to get this patch checked into the N6 branch.
    • Bug #54004 - Printing documents with framesets causes random crashes. I tracked down the cause of this problem and talked with dcone about it... If no progress is made, I'll submit a patch...
    • Spent time investigating how mozilla can be built without C++ exceptions when the containing application builds with exceptions enabled - and still get a version of operator new that returns NULL...
    • Lots of code reviews...
  • scc
    • wrote code, analyzed/filed bugs, or discussed solutions (primary engineering)
      • filed bug #58625: `copy' shouldn't include elements that are { display: none; }
      • bug #43311: investigating AIX compiler behavior
    • reviewed code: bug #'s 52653, 54203, 56998, 58491
      • bug #52653: reviewed shaver's regxpcom fix
      • bug #54203: reviewed hyatt's command dispatch/XUL refactoring
      • bug #56998: reviewed Conrad's FSSpec patch
      • bug #58491: reviewed mkaply's plugin interface fix
    • general assistance and/or mentoring
      • more investigation into embedding problems, in particular, new
      • provided some assistance to dprice with respect to the strings test suite
      • helped Hixie with Javascript and sidebars
      • helped Aaron Leventhol with Javascript and XBL handlers
      • helped hyatt with nsCOMPtrs
      • helped jag with C++ idiom, and later (a little) with IDL
      • helped Eazel's Mike Fleming (or tried to) with DOM questions
    • wrote documentation
      • started Getting CVS Write Access to Mozilla
      • updated Hacking Mozilla and the CVS Contributor Form to mention the new rules
    • growth
    • attended meetings and/or general process
      • mozilla.org staff meeting 1 hour
    • miscellaneous
      • now the week entries in my journal are actually separate pages that get assembled into this page; that means that you can get to them individually, and I've made them sensitive to queries to select the initial view, e.g., you could get to a summary view of this week with http://ScottCollins.net/journal/status/2000.10.30.shtml?summary
      • bug #55450: talked with leaf, cls, and blizzard about ykoehler. He does have CVS access; he should be able to commit his own patch.
  • rayw
    • Participated in 4 W3C teleconferences last week, for: XML Protocols, UI-tech task force, XML Protocols, Embedded DOM / Aggregate DOM
    • Spending time resolving travel issues
    • Dealing with Netscape's W3C votes membership, participation by Netscape employees, and correcting old issues, as well as preparing for meetings related to AC issues.
    • Gearing up for DOM WG meeting is coming up in two weeks, and the W3C AC meeting in 4 weeks.
    • Held meetings with Web Apps group.
    • Spent a few hours writing a prototype Java app to monitor, track, and display anything that comes down the pipe.
  • thesteve
    • no status

Summary

• The XPToolkit team resolved 43 bugs in the last week, fixing 29.
• Zarro Boogs on our RTM radar, and zarro nominated. Nothing left to fix for N6!
  
• Will now be able to start fleshing out our post-Seamonkey plans.

Highlights

• Mike Pinkerton (pinkertoni)
  • 56156minimize and restore a scrolled window, always scrolled to t
• Chris Saari (saari)
  • 33844Ability to scroll is lost when using context menu "Back"
  • 43399main window loses focus when selecting items from boomarks o
  • 35741shift key event is not recognized
• Daniel Matejka (danm)
  • 53953crash if i close the "view source" window
  • Fixed a minor regression in nspr logging.
• Eric Vaughan (evaughan)
  • 56310Collapsing server causes all tree rows to disappear
  • 55899ftp and file picker columns get out of synch when scrolling
• Stuart Parmenter (pavlov)
  • 56010Can't copy some international characters into the clipboard.
• David Hyatt (hyatt)
  • Landed 57139 on the branch. This patch fixed numerous tree widget painting and reflow issues.
  • landed two fixes for the chrome registry that prevented installation of "safe" skins and correct installation of add-on packages (like Chatzilla)
  • Added two sections to the XBL spec.
  • Fixed 54203. Removed the command dispatcher dependency from layout and made it so that builtin key bindings work with viewer and with embedding.
  • 54150 Style hint for moz-binding should be framechange, not reflow
  • 56979 Tree doesn't repaint after multi-delete
  • 53218 Theme park can't be used to install skins with the current c
  • 57733 Add-on packages don't work with Netscape 6
  • 55909 crash dragging items in trees, e..g. "manage bookmarks" windo
  • 57139 Subject list scrolls undesirably while reading messages
  • 52392 nsXBLService is leaked
  • 55332 Folders dissapear from the Mail pane
  • 48036 Columns in thread pane are improperly aligned with their con
  • 57483 Scrollbar position hosed by insertion/removal above the visi
  • 34400 [skins/xbl] behavior property doesn't work with <window> tag
  • 49096 category tree flickers when you open prefs dialog
  • 44154 Chrome Registry should process installed-chrome.txt on start
  • 47241 removeChildren fails when window scrolled
  • 54526 assertions from layout when I start up with "News cnn.com" a
  • 58797 Preview Image for Skins in Prefs Doesn't Work!
  • 51227 Scrolling in preferences window duplicates last item
  • 32486 Directory structure for installation should be nicer
• Dan Rosen (dr)
• Brian Ryner (bryner)
  • 58437 - patch from jrgm to force min-height of 19px on tree rows
  • 57178 - make sure the filepicker's directory menulist doesn't expand
  • 50597 - mousewheel nonfunctional if set to full-page mode (win32)
  • 56065 - extra *.* filter appended in filepicker

Priorities

• Keep triaging, testing, fixing for RTM, if needed.
• Start fleshing out post-Seamonkey development plans.
  

Summary

• This status report was largely prepared by Heikki Toivonen. Thanksa lot, Heikki!
• We have no rtm nominated, rtm need info, or rtm++ bugs.
• We will continue with post Netscape 6 planning work.
• The XML/DOM team resolved ~24 bugs last week (bugs not assigned to us do not show up).
• We triaged the bugs filed on the Layout component.
• We helped QA the limbo 1 rtm candidate build in the DOM Level 1, Client side Javascript areas.
• We have ordered more memory for our team's NT machines to be able to effectively run Purify & Quantify. We have also placed an order for a Macintosh computer for the team.
  

People

• Nisheeth will attend the XML Query face to face meeting in Menlo Park, California from November 14th to 16th
• Johnny and Tom will attend the DOM face to face meeting in RedwoodShores, California on November 16th and 17th
• Tom will be in Mountain View from November 14th to November 21st.
  
• Nisheeth will be on vacation from November 20th to December 17th.

Accomplishments of last week

• Harish Dhurvasula (harishd)
  • Fixed rtm++ bug 27006.
  • Did XML testing for limbo1 & limbo2.
  • Triaged my bug list.
    
• Tom Pixley (joki)
  • Worked more on event fixes for the trunk and looking through old bugs.
  • Worked on blocker bug for tree opening.
  • Also still setting up linux machine for local builds.
• Heikki Toivonen (heikki)
  • Triaged Layout bugs.
    
  • Was effectively transformed into a QA person: run Client-side JavaScript tests and verified a bunch of bugs.
  • Resolved 55975 (worksforme).
  • Attended a 2 hour meeting on Tuesday to discuss the task lists rolled up by Clayton's teams.
  • Arranged for new/additional hardware orders for the team.
• Johnny Stenback (jst)
  • Fixed bug 25821. Checked in a a bandaid for the real problem with javascript:links that have a target attribute, my checkin fixes most of the immediate problem while the targetting code is beingrewritten. This fix wasn't taken into the rtm branch.
  • Fixed rtm++ limbo bug 58256. Removed the strong PresShell reference from nsComputedDOMStyle so thatnsComputedDOMStyle's that are left in the JS GC don't interfere with the lifetime of thepresshell. This fix is only checked in to fix a crasher untill the PresShell/ViewManager ownership model is cleaned up.
  • Fixed rtm++ limbo bug 57495. Made element.getElementsByTagName() case insensitive if element is an HTML element for compatibility with the DOM spec.
  • On top of all that I fixed a number of bugs that never got rtm++'edso next week I'll hopefully be able to land some of those changes on the trunk.
    
• Nisheeth Ranjan (nisheeth)
  • Attended a 2 hour meeting on Tuesday to discuss the task lists rolled up by Clayton's teams.
  • Attended the 2 hour XML Query working group teleconference on Wednesday.
  • Attended Vidur's Web Applications meeting on Tuesday.
  • Reviewed patches to rtm bugs for Johnny, Harish.
  • Conducted bug triage and management for checkins into the rtm branch.
    

Plans for this week

• Harish Dhurvasula (harishd)
  • General bug fixes.
  • Bug triage.
• Heikki Toivonen (heikki)
  • Start reorganizing currently Futured bugs.
• Johnny Stenbäck (jst)
  • I'll continue with my bugtriaging and I'll start looking into someof the bloat reduction I've planned on doing for some time now...

Lowlights

• None.

Escalations

• None.

Although we're still under the gun for some N6 and embedding work, I wanted to share with you our next short term steps.

Once N6 ships, our main short term goal will be to integrate PSM into the Mozilla builds system.

This change will have the following benefits:

1. It will dramatically reduce the amount of time the OEMs spend to build a version of N6 for their platform, and the number of hours they take from our team.
2. PSM will be built against NSS 3.1, the completely open source version of NSS. No BSAFE or other 3rd party code will be required.
3. PSM will be able to live in the same directory as the Mozilla or Netscape client, which means it can reuse other core components (rather than bundling an identical copy, which we do now). This change should help reduce PSM footprint, and therefore download size.
4. Since PSM will use NSS 3.1, it will be possible to take advantage of the new loadable root technology. The benefit is that we won't have to respin NSS to get new roots.
5. Mozilla users will no longer have to wait for us to produce a Mozilla XPI file (which we don't always have time to produce and QA). PSM will always be present.

In addition to this integration, we're going to fix a number of outstanding Macintosh issues.

We're also working on a longer term plan for PSM, but there's nothing final at this point.

We have a newsgroup! mozilla-accessibility@mozilla.org.

We're still at the stage of making the widgets accesible. Here's a list of the widgets I've worked on so far, and the status of making them talk, with blocker bugs listed next to them.

Button

working. However, I need additive XBL to work so that they don't lose their graphical stylngs (bug 51261) Checkbox: working

Textfield

blocked by bug 58850 (selectionStart undefined)

Menus and menulist

can't get keystrokes (bug 58681 and 58847)

Radio

working

Tree

basics working! These are also used for tables in XUL, such as in email box panes. However, we don't have a way to move horizontally one cell at a time right now, because you have to select an entire line. In addition, there is no way to read the unselectable rows yet.

Tab bars

can't get keystrokes because their is no way to focus on a tab widget (bug 58987).

Sliders, progressbars, scrollbars

With these, we will just want to speak % done/thru, when the user asks. Not done.

Common dialogs

I have the Find dialog working, which proves the feasibility.

Browser widgets

need to be able to attach XBL to browser widgets (bug 57209)

Browser content

need to be able to getSelection in browser, and use caret visible (bug 49508, 48575)

Not touched yet

stacks, decks, tooltips, bulletinboards, titledbox, grid, slider, progressbar.

We have been receiving fixes from non-IBM contributors. Thanks to Dainis Jonitis and Dmitry Kubov.

4.X plugins are working on OS/2 - we are still waiting for final approval to check the changes into the tree. They involved modifying the current ns4xPlugin interface to use pass C rather than C++ functions to the plugin.

We currently have a team doing testing of the OS/2 browser and they have been very effective in finding bugs. We expect to make a great deal of progress in the next couple of weeks.

We fixed a major bug in printing, so printing should be mostly there.

We currently have the code merged to about a week ago. We are still waiting for approval to start checking it in.

Previous Updates