Mozilla

Thanks to:

• Alec Flett for everything he's done for the Mail team! Congratulations and good luck on the Browser team!!

Issues:

Vacations & Sabbaticals!

MailNews:

• RTM++ Bugs Fixed last week (23) Bugscape (4)
• non-mail ! ; Bugs Fixed last week (4) Bugscape (0)
• not-rtm++ Bugs Fixed last week (8) Bugscape (1)
• Total Bugs closed last week (68) Bugscape (6)
• new untriaged RTM (5) Bugscape (1)
• open RTM Need Info (13) Bugscape (4)
• open RTM+ (0) Bugscape (0)
  
• op! en RTM++ (3) Bugscape (2)
• New clean-all script for QA to remove all profile info for migration testing
• Helper apps now have standalone progress window!
• Fixed some topcrash bugs for RTM

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

Profile Manager:

• Ja migration fix being tested on Mac after some last tweaks

CCK:

• Bug investigation, will need a week for CCK & Shrimp bugs
  

Summary

• The XPToolkit team resolved 48 bugs in the last week, fixing 25 of these, including 14 '++'. For details, see our resolved bug list.
• We still have 18 bugs that we think should be fixed for rtm.

Highlights

• Mike Pinkerton (pinkertoni)
  • 53121 Mouse pointer shows busy icon when filepicker open; and duri
  • 47413 Dragging InetShortcuts created by IE into Nav doesn't work
  • fixed a nasty COM refcounting bug that would cause other apps to crash on quit after dragging through mozilla (bug 54803, rtm+)
  • got FizzillaCFM optimized builds running (ftp://www.mozilla.org/contrib/fizzillacfm/10_4_00_opt.sit)
  • reviewed clipboard portions of the noXIF branch for jst
• Chris Saari (saari)
  • 55080 nsMenu::AttributeChanged fails to return a value
  • 53771 can't open Composer window with Command-Shift-N keybinding
• Daniel Matejka (danm)
  • 53827 MTBF - Crash in nsXULWindow::Destroy - dereferencing a null
  • 53116 crash viewing saved cookies
  • 55275 Unable to get window.open().document, get NS_ERROR_FAILURE
  • 53953 crash if i close the "view source" window
  • 49615 Crash when clicking on [x] close in Profile dialogs
  • 54235 window objects (locationbar, statusbar, etc) visible propert
  • 54747 <tabcontrol><tabbox> crashes in nsXBLPrototypeBinding::Locat
  • 55349 Start with command-line parameters, get blank window
• Eric Vaughan (evaughan)
• Stuart Parmenter (pavlov)
  • 52275 Page causes Mozilla to crash (nsImageGTK::DrawComposited) - Patch from Tim Rowley <tor AT cs.cmu.edu>
  • Reviewing patches
• David Hyatt (hyatt)
  • 44437 (almost) landed on branch
  • 53417 Create prototype XBL bindings to reduce bloat
  • 54874 Double click on a bookmark folder does not open it
  • 41372 control-click should toggle selection in thread pane
  • 54524 FindFrameWithContent doesn't deal with XBL <children> insert
  • 53871 Double clicking on parent tree item won't toggle visibility
  • 53632 Text in div is duplicated on mouseover unless wrapped in <ul
  • 47763 always leak bindingattached/detached nsXBLEventHandlers
• Dan Rosen (dr)
• Brian Ryner (bryner)
  • 53697 xp filepicker not opening because content/global/filepicker.
  • 54798 Save All Attachments appends "(Not Downloaded)" to files
  • 54789 Crash saving all attachments - GetFolder mode handled incorr
  • 54306 can only select two items in a tree w/ shift-click
  • 53723 gray text input boxes are ugly.

Lowlights

Priorities

• pinkerton:
  • 54803 Mozilla can cause other win32 apps to crash after URL D&D
  • 52565 Crashed when spinning the cursor
  • 52225 Long <menulist> does not autoscroll in MacClassic or Modern
  • applying several patches to see if other bugs are fixed for other teams
• saari:
  • Probably start verifying bugs, keep using the product
• hyatt:
  • 44093 crash scrolling folder pane [@ nsXULTreeOuterGroupFrame::Fin
  • 44437 Current web page disappears when switching themes
  • 54097 If '/' in path, chrome files generated incorrectly, app will
  • 54784 Chrome registry should flush stringbundle and XUL caching af
  • 54150 Style hint for moz-binding should be framechange, not reflow
  • 55067 :hover attached to an element does not work while mousing ov
  • 5693 [ESM/CSS] problems with :hover
  • 43248 XBL in skins is executing scripts!
• evaughan:
  • 46877 Scroll position in page not being remembered in session hist
  • 39655 Switch folder after resize msg pane hides header envelope un
  • 54049 folder pane, thread pane and Buddy list often fails to refre
• pavlov:
  • continue helping other people with their bugs
  • working to try and get RTM out on time
• danm:
  • 52548 Sidebar Links opening in Composer window.
  • 50705 root_points_to_gcArenaPool assertion / crashes due to leaked
• dr:
  • 53322 Possible to put toolbars into unrecoverable state
• bryner:
  • 51833 Expanding tree folders sometimes causes lower-down folders t

Dogfood

Only have one: 43248 XBL in skins is executing scripts!

RTM radar

We still have 18 bugs that we think should be fixed for rtm.

Risk

• Still finding lots of ugly bugs in dogfood
• Performance regressions (e.g. today's branch bits on Win98) make app unusable too often.

Issues

Escalation

People

Highlights

• Held Mozilla Architecture Offsite on Thursday. Came up with a large list of "big ticket" items to be addressed during the next release cycle that are not necessarily end-user feature related. Notes forthcoming (to porkjockeys@mozilla.org).
  

Engineering Status

• waterson
  • Fix crasher with form controls after printing (http://bugzilla.mozilla.org/show_bug.cgi?id=53969)
  • Tracked down and fixed several problems with my thread pane minimization patch. (http://bugzilla.mozilla.org/show_bug.cgi?id=53627)
  • More mail scrolling analysis. See http://www.maubi.net/mozilla/journals/2000-10.html#04
  • Generate patch to use string iterators for RDF hotspot. (http://bugzilla.mozilla.org/show_bug.cgi?id=54715)
  • Profile new browser window opening. See http://www.maubi.net/mozilla/journals/2000-10.html#05
  • Planning meeting, code reviews, shuffled some layout bugs.
• vidur
  • Attended Architecture offsite
  • Attended Clayton's staff offsite
  • Helped land NOXIF branch onto trunk. Code on the branch eliminates the use of XIF during clipboard copying, saving from the editor and HTML to plaintext translation for mail, giving us performance and bloat improvements. Fixed several bugs in HTML content serialization.
  • Published footprint tracking numbers for NT 4.0 no porkjockeys.
  • Additional work on parser performance improvements. I hope to land on the trunk later this week or early next week.
    
  • Did some internal web applications writing.
    
• rayw
  • no status
• waldemar
  • Got caught up after coming back from the ECMA TC39 meeting in England.
  • Participated in architecture off-site.
  • Worked on JS2 spec.
• rpotts
  • no status
  • attended architecture offsite on Thursday
• jband
  • fixed 55139 - JS Component Loader JSObject parenting wackiness.
  • fixed 55151 - regression that was blocking deferred JS component registration.
  • helped dprice with jar cache flusher.
  • helped brendan with various JS bugs.
  • fixed 55200 - xpcom/proxy pointer copy error.
  • participated in c&t offsite.
  • helped chase various crashers.
• warren
  • code reviews
  • attended architecture offsite on Thursday
  • attended Clayton's staff offsite on Friday
• dprice
  • no status
  • working on brown orifice bug for 4.x
• scc
  • string work
    • working on a new string class for Vidur's parser work: nsSlidingSubstringList
    • bug #54714: got the appropriate reviews and checked in the patch to make life good for waterson
    • bug #54601: checked in my string allocation fix on the trunk and the branch (thanks to everyone who helped me test this patch)
  • helped other people; reviewed, tested, and/or checked-in their patches
    • bug #54013: reviewed blizzard's chrome patch
    • bug #54345: reviewed Adam Locke's build patch
    • bug #54740: helped jag and reviewed his |.get()| patch
    • bug #55157: reviewed chris seawood's rhapsody build patch
    • bug #48155: reviewed dveditz' chrome path relativizing patch
    • bug #54715: reviewed waterson's string manipulation patch
    • bug #53417: helped hyatt land the Mac project part of his XBL changes
    • bug #53999: reviewed sfraser's build patches
    • bugscape bug #2666: reviewed pchen's search fix
    • bug #54938: reviewed pchen's theme URL fix
    • bug #4830: approved granrose's Mac release build switch-throwing (turning off traceback tables)
    • bug #46645: reviewed blake's patch to help the sidebar obey the pref
    • bug #53403: reviewed morse's drag-and-drop patch
    • helped Pink with some string work under Carbon
    • bug #55587: helped jst fix a bug in FindChar
    • helped Rahda with C++/XPCOM type navigation
    • helped ftang with C++ tricks and standard practices; and discussed string allocation policy
  • wrote documentation
    • started a more concise document about what's down the road for the string classes
    • starting to talk about substrings how to use them and how they might change
    • answered (here, and here), at least partially, XPCOM questions on the newsgroup, and got a good response. More answers are needed here, though.
  • attended meetings
    • mozilla.org staff meeting (1 hour)
    • footprint meeting (I only stayed for one topic: 0.5 hours)
    • all-hands meeting (0.5 hours)
    • impromptu meeting on the phrase "Netscape Gecko" and other labels with ekrock and Todd Pringle (about 20 minutes)
    • architecture off-site (7.5 hours)
  • miscellaneous
    • sheriffed Friday
• wade
  • no status
• kandrot
  • Continued, and put aside the JavaScript performance debate.
  • Meet with rayw to get info on xpcom.
  • Started back into looking at memory leaks.
  • Out sick a bit.
• bruce AT puremagic.com
  • no status

Highlights

• Some significant composer bug fixes checked in on branch and trunk
• Lots of work on NOXIF branch (work to eliminate the XIF step from document output; should help performance and bugs in copy/paste, outputting to files etc).
• the team resolved 32 bugs last week
  

Lowlights

• 32 rtm need info bugs open
• 56 open non-RTM bugs in M19
• Bugs need triage to get to 3 RTM bugs per engineer, heading for zarro boogs.
  

Issues

• Server outages (e.g. mail servers) still cost time

RTM status

• Currently have 32 [rtm need info] bugs

Goals for this coming week

• Akkana
  • More NOXIF work.
  • Any other RTM bugs that come up.
  • Help whoever needs help.
  • More dogfood eating.
• Anthony
  • RTM bugs
  • Dogfooding
  • Helping out others where necessary
• Joe
  • sleep
  • speed up copy
  • correct any noxif problems that appear
  • finally work on whitespace issues, at long last
• Kathy
  • Triage my bug list and supply the necessary info
  • Fix approved bugs
  • Eat dogfood
• Kin
  • Try to plow through my rtm+ bugs.
  • More code reviews.
  • Eat dogfood.
• Mike
  • Working on RTM bugs, checking out new bugs
• Simon
  • Work on remaining RTM bugs
  • Be virtual beppe this week
  • Eat dogfood, file bugs

Last week's progress

• Akkana
  • Worked on the noxif branch, testing and making small changes.
  • Lots of dogfood eating, bug filing, bug commenting, and bug verifying.
  • Lots of time talking to jag, Pike and others (mostly on IRC) regarding key binding problems, e.g. bug 54694 (currently marked minus, but it's a frequently-encountered crash on platforms affected by it, mostly solaris).
  • Spent time diagnosingbug 54283, only to find eventually that it was a dup of one of Harish's. Confirmed Harish's fix, which is now checked in.
    
  • Sheriffed on Thursday.
• Anthony
  • Ate dogfood, client looks much better, and is more stable, Composer of course is running smooth.
  • Tiger Team: Looked at talkback reports, trying to find bugs to file so that MTBF will improive if they are fixed.
  • Reviewed and triaged my bugs.
  • Worked on getting bugs fixed and reviewed and begged to get them into RTM, such as:
  • Fixed and Checked in 54253, Drag and Drop causes data loss.
  • Fixed and Checked in 54222, the 'noshade' on horizontal rule not being set. Kin gave me this bug, and helped me work on it. This bug also should fix several other problems dealing with attributes that do not require values to be set.
  • Resolved 35414 as WORKSFORME, 3rd party apps crash with SeaMonkey.
  • More profiling, and quantifying of SeaMonkey. Editor still looks really really good.
  • Pulled Sheriff duty Thursday night.
• Charley
  • On sabbatical
    
• Joe
  • helped land NOXIF, lots of work on this branch.
• Kathy
  • triaged bug list
  • checked in approved fixes
  • reviewed/tested patches submitted (rejected some; still testing others)
  • reviewed for rtm++ fixes
  • ate dogfood; filed bugs
• Kin
  • Traded emails with international group about various spellchecker issues.
  • Did several code reviews for akk, brade, sfraser, anthonyd.
  • Checked in fixes for 50570, 53326, 54731 on the branch and trunk.
  • Checked in cmanske's fixes for 54426 and 53733 on the Netscape_20000922_BRANCH.
  • Helped build/debug/test Win32 problems on the NOXIF_20000911_BRANCH.
  • Had several discussions with jfrancis about Copy/Paste issues he is working on for the no XIF branch. I must say Copy/Paste is looking really good on theNOXIF branch!!
  • anthonyd graciously took some bugs off my plate. Spent some time talking through some of the issues involved.
  • Talked Hurricane into starting a project that would allow us to use ASpell so that the mozilla community can use an Open Source Spellchecker. This should be pretty easy once he figures out ASpell's requirements. Some of the glue code I wrote for the INSO spellchecker should probably be moved into some sortof SpellChecker controller that we can share.
    
  • Ate dogfood off the trunk and branch, and filed bugs.
• Mike
  • Working on RTM bugs 55325, 49041, 49772, 48575, 51772
  • Need to look at bugs, 48064, 48868
    
• Simon
  • Found and fixed a nasty bug in the Mac unicode font mapping code (55140)
  • Fixed bug 53742(titlebar for open location dialog)
  • Fixed bug 55080, (missing return value in Mac menu code)
  • Worked on several outstanding rtm bugs:
    • 52808-- fix several editor teardown issues
    • 20743-- low-memory crashes still a problem
  • Attended whole-day offsite on future planning for the C&T group.
    

Personnel

• Charley Manske on sabbatical.

Summary

• The biggest crasher for HTTP is now resolved. Thanks to darin, dougt and danm's work.
• We have about 10 need info(work in progress bugs) and 25 nominations!
  

Highlights

• darin
  • Sorting thru untargeted Necko bugs
  • Learning HTTP and more on the Necko way of doing things.
• gordon
  • Fix for 49990 "Networking Fails after sleep/wakeup cycle" got reviewed, modified, and super-reviewed
• neeti
  • Have fixes ready for bugs 54072, 55405, 42606. Waiting for a review and super review to check them in.
  • Resolved bugs 55145, 54304
• pnunn
  • Bugs Closed: #54996, #31918, #54876, #55354, #45902:, #23882, #52807, bugscape#2750
  • Bug Work: helped AOL folk to get art component working with gecko.
• gagan
  • Triaging bugs for the whole team.
  • Prepared a presentation for Clayton's offsite.
    

Lowlights

• ruslan
  • no status.
• gordon
  • Fix for 49990 "Networking Fails after sleep/wakeup cycle" waiting to get double-plused.
  • Turn around for getting reviews and approvals seem way too slow.

Issues/Escalations

• None.

On Thursday, Oct 5th we had a Mozilla Architecture Offsite meeting to discuss issues surrounding the continued evolution of the mozilla platform architecture, particularly focused on the needs of the relatively short-term embedding effort. Here is my summary of the offsite compiled from notes taken by Chris Waterson and Clayton Lewis.

In attendance: me, Clayton Lewis, Chris Waterson, Rick Potts, John Bandhauer, Simon Fraser, Patrick Beard, Waldemar Horwat, Vidur Apparao, Brendan Eich (after lunch), Dave Hyatt (briefly), Jud Valeski, Scott Collins, Ray Whitmer.

Opening remarks were by Clayton, and he offered the following goals and agenda:

• goals (for next 6 months):
  • focus on embedding in other products
  • memory target = 32 meg device
  • performance
• agenda:
  • how are we going to tackle memory problems
  • how are we going to develop a layered architecture for various product requirements
  • what do we need for SDKs?

A good part of the day was then spent discussing various projects we've had in the back of our minds for some time now, and how they fit into the specified goals. Projects were then categorized into short-term (ST) and long-term (LT) according to how important they were with respect to the specified embedding and footprint goals.

Projects

• reducing number of DLLs
  • 3.5Mb overhead on linux. why is the loader costing us so much?
  • cost of downloading bits to users
  • build system issue: big lump vs. lotsa dlls
  • where to apply component model; remove XPCOM interfaces from places where not appropriate (e.g., GFX)
  • getting rid of interfaces, getting rid of XPCOM interfaces, keep
  • XPCOM interface but "friends" don't need to use those interfaces.
  • ST: more analysis, build process changes on all platforms
  • LT: analyze eliminating component model boundaries, identify which components are replacable.
• examine multithreaded concerns
  • ST: fix proxy code, fix atom table
• making frames extensible and changing the way that frame ownership works
  • plug ins for mathML, svg, etc.
• make GFX not XPCOM
  • ST: profile GFX on all platforms
  • LT: pavlov's GFX2
• remove static linkage of DOM JS, replace with XPConnect
  • ST: 3 months ETA
• emulating native widgets
  • accessibility
  • test harness (e.g. QA Partner)
  • ST:
    • compile new frame classes that talk to native widgets, or
    • build out XUL and XBL to emulate
• repartition code for working set
  • ST: get XUL, chrome, and rdf out of distribution
  • tool-based approach
• attack allocations
  • using new string implementations
  • ST: construction site analysis (wallet, layout)
  • savings on common strings
  • string compaction (e.g., UTF-8 encode everywhere?)
• imagelib redesign
  • image depth; space wasted?
  • image cache analysis
  • tor's imagelib 1.5
  • ST: tor will land some good things; e.g., platform code for scaling
• style system
  • style context sharing
  • evaluate bloaty structures
  • ST: both
• network cache
  • ST: memory & disk cache needs stronger ownership
• object cache analysis
  • ST: bloatblame, dynamic (page-to-page) analysis
• DLL unloading, plugin unloading
  • actually unload modules that claim they can be unloaded
  • unload unused services
  • pick a simple test case to unload, like ftp
  • ST: turn on unloading when a module returns CanUnload().
• ender "lite" analysis
  • ST: analyze allocation during blur, focus, typing
• bi-direction text landing
  • ST: let it happen; accept patches turned off
• new view manager
  • remove xpcom interfaces?
  • ST: let it happen
• security audit
  • ST: starting to happen
  • LT: privacy
• consistent gamma correction
  • CSS work; e.g., PNG and named colors get same color correction
  • inhibit use of PNGs in skins
  • ST: tor is working on this
  • LT: CSS work
• in-process PSM
  • psm does its own html UI -- inflexible
  • ST: yes (?)
  • LT:
    • eliminate PSM glue, protocols, threads
    • cert management UI
• separate wallet from cookie
  • ST: ok
• change content sink interfaces so that we can hook expat up directly to sinks
  • LT
• layout memory analysis
  • ST
• kill all leaks
  • ST
• memory visualization tools
  • ST
• get serious about unit testing
  • ST:
    • "make test"?
    • investigate QA partner integration (somebody having success @iplanet)
    • getting at DOM-level stuff? DOM-based automation? DOM recorder?
• finish API cleanup
  • ST
• make seamonkey be a "standard" embedding client
  • LT

SDK Issues

• sdk covers:
  • embedding APIs
  • plugin API
  • necko (? - relatively mature)
  • xpcom
• contents:
  • XPIDL, header files, docs, examples, binaries, cf plugin SDK
  • Release and debug builds, can these binaries be interchangeable?
• issues:
  • IID should be forever stable, any changes need a new IID
  • need discipline/testing to ensure this
  • need methodology for deprecating APIs over time

Teams needed

• Unit testing team
  • consists of development and QA engineers
• Memory analysis team
  • find memory problems and hand-off work to be fixed
  • develop tools
  • develop way to measure progress
• Tools team
  • need someone to help us improve build system, tinderbox, bugzilla, etc.
  • perl hacker

Summary

• We currently have 14 rtm nominated bugs, 4 rtm++ bugs, 3 rtm+ bugs awaiting PDT approval. Triaging, with Eric Krock's help, is continuing.
• The XML/DOM team resolved 23 bugs last week.

People

• Tom Pixley is back from vacation and busy with his rtm bugs.

Accomplishments of last week

• Harish Dhurvasula (harishd)
  • Checked in fix for bug 50965 on the branch.
  • Checked in fixes for bugs 53935,54448,53489,53695 on the trunk [ has been PDT approved for rtm ]
  • Provided a patch for rtm+ bug 54571 ( a=hyatt ), working on a patch for another rtm+ bug 54845.
  • Worked on rtm nominated bug 55462 ( very close to a fix )
  • Triaged my bug list.
• Heikki Toivonen (heikki)
  • Triaged joki's untargeted bugs.
  • Fixed 51211, image load event caused problems in chrome, for example the search sidebar.
  • Resolved some bugs: 54622 (duplicate), 53460 (duplicate), 52620 (worksforme), 51268 (worksforme), 51059 (worksforme).
  • Reviewed patches from toml@us.ibm.com to the XMLExtras component (bugs 55508 and 55599).
  • RTM nominated:
    • 55377 [HAVE FIX] FMM nsRange::CreateContextualFragment
    • 30936 can't tab from last to first page element
    RTM need info:
    • 53763 Crash closing event dialog and source changed
    • 54233 Crash on quit via keyboard
    • 54323 [HAVE FIX, CAN'T REPRO] IPR nsGlobalWindow::WinHasOption
• Johnny Stenback (jst)
  • Checked in a fix for rtm++ bug 53317 that caused a crash in nsFrameImageLoader::NotifyFrames() due to references to pointers that are invalidated during a callback. Patch done by bienvenu@netscape.com.
  • Fixed rtm++ bug 55587. nsAReadableString::FindChar() was returning the index from the offset to the character if found in stead of the index from the beginning of the string.
  • Worked on the NOXIF changes and finally landed the changes on friday night/saturday morning.
• Nisheeth Ranjan (nisheeth)
  • Ongoing bug triage for the XML/DOM team.
  • Participated in Clayton's daylong offsite for managers on Friday.

Plans for this week

• Harish Dhurvasula (harishd)
  • Checkin fixes for bugs 53935,54448,53489,53695 on the trunk
  • Fix bugs 54845, 55462, and 55036
  • Verify bugs that aren't verified yet, if time permits
  • Triage my bug list.
• Heikki Toivonen (heikki)
  • Work on RTM bugs.
• Johnny Stenback (jst)
  • Setting out some small fires caused by the NOXIF landing
  • Working on my rtm bugs.
• Nisheeth Ranjan (nisheeth)
  • Work on rtm bugs.

Lowlights

• None.

Escalations

• None.

We originally planned to deliver the NSS 3.1 release candidate last Friday (Oct. 6). We missed that date. The current estimate of the slippage is three days to a week. I believe we fixed the last P1 bug last night, so if we don't run into any build machine problems today, we should be able to finish the QA cycles and deliver a release candidate by tonight. After that, we will let it bake for a week before declaring it final.

We added several new tests in this release to increase the test coverage and we will continue to do so in future releases to improve the quality and performance of NSS.

It was another very successful week for the Warpzilla project.

Highlights

• Major timer fixes went in that improve performance considerably
• New font changes improve layout and italics
• Fixed trap related to Modern skin and Verdana font

Nightly builds are available from ftp://ftp.software.ibm.com/ps/products/warpzilla and the current one looks really good.

We are continuing to concentrate on getting our build in sync with the other platforms.

We should have an M18 build available about the same time as all the other platforms.

This week was also quite active. We got NSPR to work with BONE while we're still having problem with the mozilla. The menu are still broken but we got the cursor working too. We're investigating some issues about thread safe and strange image draw occuring. The Gfx has been re-written and also fixed japanese text rendering if you use the proper font it should work! Also tested that 8 bit PNG with alpha is working! And I'm going to be testing my newly CVS account very soon! Things are really exciting as once the input get fixed we will start getting a functioning browser. Ho! Another issue we're looking for fixing is the fact that chrome doesn't work, it seems like the profile doesn't even create a "chrome" folder.

Previous Updates