Thanks to:

• Seth for helping Bhuvan w/44764
• Bhuvan for finding a bug with our commercial overlays that was caused by some chrome registry changes.
• Entire mail5 team for tremendous beta3 effort - almost 1000 nsbeta3 bugs triaged, over 300 nsbeta3+ found & fixed!
  

Issues:

• Lots of RTM candidates flooding in already!
  

Vacations & Sabbaticals!

MailNews:

• nsBeta3+ Bugs Fixed last week (2) Bugscape (4)
• non-mail Bugs Fixed last week (3) Bugscape (0)
• not-nsBeta3+ Bugs Fixed last week (0) Bugscape (0)
• Total Bugs closed last week (57) Bugscape (8)
• new untriaged RTM (9) Bugscape (4)
• open RTM+ (20) Bugscape (3)
• Scottip helped fix cache crasher
• Lots of rtm+ bugs w/fix in hand
• Varada helped verify 8 bugs
  

Date	Open nsBeta3+ P1/2	Open nsBeta3+ P3/4/5	New, UnTriaged nsBeta3 Nominees	Fixed nsBeta3+	Fixed non-nsBeta3+
7/31	0	0	630	10	46 (and 8 b2+)
8/7	0	135	57	33	47 (and 9 b2+)
8/14	0	146	43	56	16
8/21	69	62	70	43	6
8/28	54	66	53	37 (26 were P1/2)	13
9/5	28	13	46	44 (29 were P1/2)	7
9/11	23	14	41	24 (16 were P1/2)	4
9/18	21	2	44	28 (20 were P1/2)	4
9/25	17	2	11	23 (all P1/2)	5
10/2	n/a	n/a	n/a	6 (all P1/2)	0
TOTAL			995	304	93 b2+, 55 other

DATE	Open RTM+	Open RTM	Fixed RTM++	Fixed non-RTM++	Verified for RTM
10/2	23	13	0	0	8

Profile Manager:

• Solved Ja problem just a tiny bit too late for beta3, will go into rtm.

CCK:

• Fixed CCK for latest new jar packaging scheme
  

Highlights

• PR3 work is done.
• Big update on W3D DOM Working Group from Ray, below.
  

Lowlights

• This document was edited entirely with Netscape 6 PR3. I did the best I could formatting it, but numerous things are wrong with styles and bulleted lists.

Engineering Status

• waterson
  • Continued to tinker with broken inline margins, border, and padding (3490): I almost have a fix, but will probably need to hack the text transformer to compute whether or not there's any non-whitespace content left in its buffer.
  • Spent some time looking in to two problems (52726 and 53901) that dbaron discovered with how we're setting up the mDocument member in XUL elements. This ended up being a rock with many unpleasant crawlies beneath it: we have a pretty serious problem with the way that we're cleaning up script objects: HTML and XML elements run the risk of losing properties on objects; XUL runs the risk of leaking megabytes! Oy vey. Details in the bugs.
  • Tracked down the Win98 crash on exit, which was introduced with the memory flusher thread.
  • Spent a lot of time profiling mail scrolling:
    • Checked in the fixed template builder stuff on the tip so that I could finish up the attribute-ectomy on the mail thread-pane. This yields a 10% speedup when scrolling "new rows" (but isn't yet checked in, pending review from scottip et. al. in the mail team).
    • Added a check to detect that we're at a leaf when building content. Fix checked in to tip; yields nominal (2%) speedup for scrolling.
    • For more deails, see analysis here.
• vidur
  • Worked with Jud on Monday to try to get a Linux build together which combines a large number of generic modules into a single one. The idea was to quantify on-disk and in-memory savings by reducing the per-shared library overhead. We used Chris Seawood's Linux build modifications plus additional ones of our own, but got stymied by a linker problem. Will continue this week.
  • Fixed Bugscape bug 2533 - document.onload was firing earlier than the last layout pass.
  • Completed first phase of parser string changes. The scanner buffer is now using a nsSlidingString and text tokens are using nsSlidingSubstrings. This eliminates 2 copies of text in the parser. Will do work this week to quantify performance and bloat savings, fix known bugs and consider additional changes to eliminate more copying.
  • Ran Footprint/Performance meeting. Notes posted to porkjockeys.
  • Wrote first draft of process size tracking script (see n.p.m.porkjockeys). This will allow us measure and track process size for common tasks in a consistent manner and determine when regressions and improvements come in.
  • Held internal web applications meeting.
• rayw
  • Spent three days at W3C DOM Working Group and most of a day traveling back due to bad travel arrangements.
  • Some issues that lots of time was spent on include:
    • I made no progress transitioning the XPCOM responsibilities to kandrot, although I need to do that before I can work with the Web Applications team.
    • Load / Save / SAX / related issues.
    • Also, how I/O errors are handled in DOM, since DOM exceptions are designed for programming errors, not system errors.
    • A new CORBA DOM binding by OMG for use in CORBA APIs which pass DOM nodes (I have been asked to reopen contacts and conversations with OMG from prior employer's heavy use of XML in CORBA -- Vidur, this is another protocol / platform where XML exchanges may be used).
    • Incompatibilities between Netscape all versions and W3C DOM HTML spec -- list is being developed.
    • Specific issues in the new core functionality such as the lifetime and lifecycle of a new ID that is being introduced to allow users to create a mapping between DOM nodes and extra user data.
    • How we should handle different variations of read-only DOM and validating DOMs, which raise exceptions in such a way to make them a subset of DOMs which support everything without raising exceptions.
    • Many hours discussion of the new standard for TV content which is likely to become the law WRT hypertext markup content within television delivery mechanisms. They have taken great pains to document browser DOM level 0 behaviors where, and extend HTML level 1 in a few cases where it appeared that important things were missing.
  • Lots of old issues including:
    • null versus empty string and how the spec is written and how bindings adapt, how the documentation clarifies when null may be used, how ECMAScript, CORBA, etc. work where nulls are required.
  • My own responsibilities include:
    • CORBA follow-through
    • More updates identified for the Views and Formatting document.
    • Lots of Pseudo-code examples and sorting out existing use cases, and decisions regarding computed content, view-based events, etc.
    • Create a note to be submitted to W3C on how Mozilla is loading XHTML into the HTML DOM (not the same as the XHTML DOM proposed by the HTML Working Group) and explicit decisions that have been made that might be followed by other browser vendors for compatibility, and how users will have to carefully write their Javascript so that it will work in browsers with and without XHTML support.
  • Looking at some issues in CORE...
    • I will be working with the new embedding editorial team trying to iron out issues of what happens when documents containing multiple namespaces such as XHTML, SVG, MathML, etc. are combined in one document tree, possibly recursively (e.g. XHTML contains SVG contains XHTML), how are CSS properties cascaded, events propogated, DOM implementations combined compatibly, parsers unified, between namespaces.
    • UI Task Force (was Style Coordination Group) irons out lots of issues, especially the collision of different namespaces with respect to naming, behaviors, etc.
    • I could supply lots more details of issues that were discussed.
  • After the DOM meeting, I spent time making edits on the Views and Formatting document, and triaging the endless flow of bugs that come to the default recipient of XPCOM (which means if they can't figure it out, XPCOM must be involved).
  • I made no progress transitioning the XPCOM responsibilities to kandrot, although I need to do that before I can work with the Web Applications team.
    
• waldemar
  • in England all of last week
• rpotts
  • Checked in a fix for bug #41398 (trunk only). The Cache Manager is now a memory pressure observer and will discard the memory cache when notified.
  • Checked in a fix to the memory cache (trunk only) where storage was being double counted when freed...
  • Have fix for bug #54470 - Clearing the cache causes disk cache corruption.. Waiting for approval to check into the trunk. It would be nice to add this fix to the branch before RTM.
  • Have fix for bug #54239 - The PSM component is created on the wrong thread. Waiting for approval to check into the trunk.
• jband
  • Run Purify a lot and filed lots of bugs.
  • Helped diagnose and fix bugs: 51954 53123 53094 53353 53268
  • Did some work on the jar cache - wrote a stress test and proposed a patch for a threading bug.
  • Improved and checked in contributed fixes to bugs: 54264 54275 54310 54352 54473 54462
• warren
  • Cleaning up numerous details after the jar packaging landing.
    
• dprice
  • Spent several hours going through the talkback stuff. Most of the crashers have bugs assigned to them, and are either fixed or close to it. Tried finding test cases / bugs for some of the other ones.
  • Removing the bin/res directory.
  • Working on crasher in zip cache.
• scc
  • string work
    • bug #52721: more API changes to nsSlidingSubstring for Vidur's parser work
    • bug #54601: finally fixed the string allocation policy
    • bug #54714: trying to help waterson with some optimizations by providing this functionality before the rest of the support is in
  • helped other people; reviewed, tested, and/or checked-in their patches
    • bug #32180: helped Chris Seawood analyze C++ problems on Irix
    • bug #53682: came up with a solution (tested by lordpixel) to the namespace problem brought out by CW6
    • bug #50782: reviewed jlnance's patch, and discussed Irix and Mac fixes
    • helped waterson with gcc 2.7.2.3 temporary insanity
    • bug #54740: helped jag diagnose leaks
    • bug #54564: reviewed dougt's autoconf patches
    • bug #17641: reviewed BenB's work, and, given the correlation to current behavior, marked the bug fixed at his request
  • filed other bugs
    • bug #54564: let's get L"..." working under gcc
  • wrote or posted documentation
    • checked warren's Jar Packaging 101 into the mozilla doc tree (thanks to Eric Hodel, who cleaned it up for posting)
    • answered JustinVP's school project interview questions (should I post them here?)
  • attended meetings
    • macdev meeting (1 hour)
    • mozilla staff meeting (2 hours)
• wade
  • no status
    
• kandrot
  • no status
• bruce AT puremagic.com
  • no status

Highlights

• the team worked hard at getting pr3 ready for distribution
• we hit off zarro during the week
• the team resolved 55 bugs last week
  

Lowlights

• good news/bad news Charley has started his sabbatical, returns week of 11/13
  
• 7 bugs have been reopened
  

Issues

• no outstanding issues this week

RTM status

• Currently have 55 rtm+ bugs
  

  Engineer	evaluating/ not +	rtm+/ debugging	rtm+ fix in hand	rtm+ awaiting++
  Akkana	54694	7	1	--
  Anthony	--	2	1	--
  Kathy	--	8	4	--
  Joe	52379,54452,54569	7	--	--
  Kin	54825	6	5	--
  Mike	--	3	--	--
  Simon	54023	10	--	--

  linkto the rtm+ query
  

Goals for this coming week

• Akkana
  • Plug away at rtm bugs (or nsbeta3 last-minute issues if any come up).
  • More dogfood eating (yum!)
• Anthony
  • continue to triage open bug list
  • work on any nsbeta3++ stoppers
  • work on rtm+ bugs
  • eat dogfood and file bugs
• Joe
  • work on rtm+ bugs
  • merging the branch to tip in preperation for landing on tip; begin redoing editor whitespace handling.
    
• Kathy
  • Triage my bug list and supply the necessary info
  • Fix approved bugs
  • Eat dogfood
• Kin
  • Help with any last minute nsbeta3+ bugs, if it hasn't gone out the door already.
  • Start on rtm++ bugs.
  • Checkin code for cmanske.
• Mike
  • work on rtm+ bugs
  • help Akkana, Kin and Simon if they need my help
  • help xpfe folks if they need help
  • do more profiling and look for leaks and other performance problems
• Simon
  • Any remaining nsbeta3++ work that needs going
  • dogfood+ bugs
  • rtm+ bugs

Good stuff from last

• Akkana
  • Monday off.
  • Triaged all my nsbeta3- bugs, marking rtm or not as appropriate.
  • 54016: reviewed Daniel Bratell's fix and checked it into the noxif branch.
  • Found and fixed bug 54449, problem with caret warping after delete-line, and added comment to the related bug 54452 (ubercaret after deleting part of quoted text) suggesting what the problem might be.
  • Lots of dogfood eating, filed lots of bugs and commented on lots of other bugs.
  • Several more rounds of triaging my rtm+ bugs, adding and re-adding comments re-affirming that yes, this is still important.
  • Lots of time spent going back and forth with Sun people regarding meta key weirdness in bugs 43433 and 54694, and with jag and others trying to track down the problems in the + keybinding (text zoom).
  • Spun wheels trying to get a solaris build (gave up, worms just isn't up to the task any more) and a noxif build (haven't yet found a date-pull that's compatible with the branched directories; they need to be updated to the tip and/or branch).
• Anthony
  • ate dogfood.
  • made patches for 2 bugs 49663 and 54253.
  • quantified editor and layout again.
  • triaged my bugs.
  • triaged editor team bugs without a milestone.
  • built and tested the branch builds.
  • Mike and I also triaged several crashers on mozilla.
• Charley
  • Fixed bugs 53733 (nested tables), and 53377 (strip quotes from title for candidate filename) -- these are checked into trunk.
  • Fixed rtm bug candidates that Kathy and Kin will get approved and checkin: 54296 (make colspan/rowspan work in table dialog), 54315, 54316, 54317 (supply state feedback with checks in font face, size and text style submenus under Format), 54222 (noshade attribute not set on HR element), 54426 (infinite loop in Join Table Cells). 53733 (Make nested tables easier to do).
  • Fixed Vidur's new content serializer code to use same rules for "pretty printing" as the to-be-replaced implemention in the HTMLContentSink code.
  • Lots of bug filing and triage.
  • Worked with evaughan to isolate layout problems when using buttons in a XUL grid (bug 51697), but now he claims he can't show the problem with his build! This problem is important and should be followed through -- it causes great difficulty with localization if buttons don't resize when text is longer ( "the German problem.") But Eric promises to follow through and try to find out what's going on.
• Joe
  • spent 2 weeks redoing copy/paste. it's demoable now. will land on branch tonight.
• Kathy
  • skimming/deleting thousands of e-mail messages
  • pulling/building branch/trunk
  • eating dogfood
  • filing bugs
  • querying bugzilla for status on various bugs, etc.
  • taking handoff from Charley
  • investigated/fixed image map editor bug
• Kin
  • id code reviews for brade, cmanske and sfraser.
  • Spent some time trying to help out dcone with bug 24406 ("Password, Textfields and TextAreas prints blank") which was an nsbeta3+ which got minused recently. It's looking like the problem is caused by a missed reflow.
  • Triaged and did some preliminary debugging for editor bugs. Also went through my entire open bug list, updating with comments and looking for things that need to be fixed before rtm.
  • Did some poking around as to why people were reporting that the spellchecker stopped working (Bug 54057). Turns out that some changes were made to the spellchecker code on the commercial trunk, without my knowledge, to get rid of nsIFileLocator usage that prevented the spellchecker from finding it's dictionaries.
  • Fixed 53326 ("Can't select paragraph headers"), will checkin when tree opens for rtm+ fixes.
  • Ate dogfood and filed bugs.
• Mike
  • fixed remaining nsbeta3+ bugs
  • helped xpfe team
  • reviewed open bug list and triaged rtm+ bugs
• Simon
  • Checked in fix for nsbeta3++ bug 52150-- make the caret show in new composer and mail compose windows. Despite fixing editor to show the caret when focussed, we still needed a focus fix elsewhere in the code.
  • Checked in a fix for nsbeta3++ bug 54079,and bugscape bug 2331-- there were a whole bunch of problems using items on the Help menu in the commercial build, so this fix touched several files.
  • Spent some time debugging Flash plugin problems on Mac, participated in discussion of how to solve Mac plugin issues.
  • Reviewed several patches for others, spotted some nasty leaks in one of them.
  • Some minor build script changes, most post-jar tweaking.
  • Attended a small devices meeting on Monday; cute toys, but no immediate relevance to gecko/editor.

Personal

• Kathy will be out for several weeks starting 11/15
  

Summary

• Kept our plate clean of nsbeta3++ bugs for nsbeta3 candidate.
  
• The XPToolkit team resolved 39 bugs in the last week, fixing 11 of these, including 6 '+'. For details, see our lists of resolved bugs, 37 in Bugzilla an! d 2 in Bugscape.
• We have 49 bugs nominated for rtm, 47 in Bugzilla and 2 in Bugscape , of which we have marked 14 rtm+, and have 32 left to triage.
• Need to triage remaining rtm bugs.
  

Highlights

• Mike Pinkerton (pinkertoni)
  • lots of bugs filed, running branch as dogfood
  • Fizzilla CFM debug builds are now current, but optimized builds error at startup
• Chris Saari (saari)
  • Fixed what nsbeta3+ bugs were left on my plate, tested fixes for joki.
  • Using pr3 which is becoming a much more enjoyable experience.
• Daniel Matejka (danm)
  • Tested beta3 candidates.
    
• Eric Vaughan (evaughan)
  • Refactored tree layout to fix many tree problems as well as beta3+ bugs hopefully will make RTM.
  • Made tree paint optimally (no flicker)
    
• Stuart Parmenter (pavlov)
  • tryed to get SSL proxies going for 31174.. not having much luck.
  • reviewing patches, trying to help verify and test bugs
• David Hyatt (hyatt)
  • current web page and session history kept when switching themes (44437). hasn't landed yet though.
  • prototype XBL bindings created to reduce bloat and provide a dramatic speedup. fixed on trunk. waiting word for branch.
  • fixed some fallout from the big typing speedup landing from last week.
  • helped eric with some tree code shuffling to fix many + bug problems. looking sweet.
• Dan Rosen (dr)
  • ??? - no progress report.
    
• Brian Ryner (bryner)
  • helped review patches
  • used dogfood, reported bugs from it
  • just verified all the resolved/unverified bugs I've reported

Lowlights

Priorities

• pinkerton:
  • 52225 Long <menulist> does not autoscroll in MacClassic or Modern
• saari:
  • 49232 Should only be one tab from addressing field to subject line
• hyatt:
  • 44093 crash scrolling folder pane [@ nsXULTreeOuterGroupFrame::Fin
  • 53322 Possible to put toolbars into unrecoverable state
  • 53417 Create prototype XBL bindings to reduce bloat
  • 44437
  • Current web page disappears when switching themes
  • 54049 folder pane (and thread pane) often fails to refresh
• evaughan:
  • 46877 Scroll position in page not being remembered in session hist
  • 39655 Switch folder after resize msg pane hides header envelope un
• pavlov:
  • 52275 Page causes Mozilla to crash (nsImageGTK::DrawComposited)
• danm:
  • Assist dougt in his efforts to thwap nested event queues and not go woozy from the resulting earthquakes.
  • 53827 MTBF - Crash in nsXULWindow::Destroy - dereferencing a null
  • 52548 Sidebar Links opening in Composer window.
  • 53953 crash if i close the "view source" window
• dr:
• bryner:
  • 53723 gray text input boxes are ugly.

Dogfood

• None.

Beta Stoppers

• None.

Risk

Issues

• Most CPD employees are still not using N6 for dogfood.

Escalation

People

Summary

• Darin Fisher joins Necko/Imglib group!
• Exceptional work done by the whole of networking/imagelib team in finishing up on PR3.
• Certain crashers in HTTP are still a problem.

Dogfood and Risks

• none

Highlights

• neeti
  • Fixed bug 52818
  • Currently working on 54072, 54630
• pnunn
  • Bugs Closed:
    • #54588, #54460, #53169, #53338, #52996, #52992 (all dupes)
    • #54700, #54463, #52390**, #53434
    • (reassigned): #41500, #52494, #54465,
  • Bug Work:
    • #41268: animation cache
    • #54463: img problems in attachments.
• gagan
  • Triaging bugs for the whole team.
  • Updated status/hire page to support editing.
  • Investigating the HTTP crasher bugs.

Lowlights

• gordon
  • no status.
• ruslan
  • no status.

Issues/Escalations

• None.

People

• Darin Fisher joins us at the new member of the Networking team. He is likely to be working on HTTP. He can be reached at darin@netscape.com

Summary

• Beta 3 is done! Onwards to RTM!
• We currently have 25 rtm nominated bugs. We will triage this list with Eric Krock on Monday, October 2nd.
• The XML/DOM team resolved 27 bugs last week.

People

• Tom Pixley is away on vacation from Monday, October 2nd through Friday, October 6th.

Accomplishments of last week

• Harish Dhurvasula (harishd)
  • Checked in fixes for bug 50965 on the trunk.
  • Fixed rtm+ bug 53965.
  • Triaged Clayton's and my bug lists.
  • Ran quantify in vain!! It kept giving up on me. Will download the patch from Rational and will try once again this week.
  • Fixed rtm+ bug 53935 and rtm nominated bug 54453.
  • Worked on a rtm nominated linux crasher bug 53829 ( haven't got a clue yet ).
• Heikki Toivonen (heikki)
  • I am pretty much finished with bug 51211 (rtm+), need ++ to check my stuff in and then pass this on to some GUI person for the other issues in this bug. This is about load events not firing as they should + some performance issues + other stuff (yeah this should really be several bugs but the other bugs were duped to this).
  • Triaged 6 bugs from Clayton's list
  • Verified a bunch of nsbeta3 nominated and resolved bugs
  • Resolved bug 54707 (duplicate)
  • Reported a couple of UMRs found with Purify (hyatt made a record in fixing one of these, he had actually checked in the fix 10 minutes before I reported the bug [;)] . ESP or what?
  • Fighting with Quantify. It was happily using all my memory (256MR RAM + 300MB disc) until I finally gave up and set default instrumentation to function level. I was able to set one DLL to line level instr. No useful data yet.
  • Trying to figure out the inner workings of the XMLExtras component.
• Johnny Stenback (jst)
  • Was away to DOM face to face meeting in Boston from Monday through Wednesday.
  • Triaged untargeted bug list from 27 to 5.
  • Caught up with email backlog.
• Nisheeth Ranjan (nisheeth)
  • Fixed nsbeta3++ bug 48382 (http://bugzilla.mozilla.org/show_bug.cgi?id=48382)
  • Checked in fix for nsbeta3++ bug 53219 (http://bugzilla.mozilla.org/show_bug.cgi?id=53219) on joki's behalf.
  • Turned off the "nglayout.events.showHierarchicalHover" pref by default on joki's behalf.
  • Triaged untargeted bugs from 17 to 4.
  • Triaged Clayton's bug list from last week.
  • Attended XML Query teleconference on Wednesday.

Plans for this week

• Harish Dhurvasula (harishd)
  • Checkin fixes for bugs 53935, 53965.
  • Fix bug 53829.
  • Verify bugs that aren't verified yet, if time permits.
  • Triage my bug list.
• Heikki Toivonen (heikki)
  • Check in my fix to 51211 and pass it then to someone in charge of navigator.js or something...
  • Fix a bug or two in XMLExtras.
• Johnny Stenback (jst)
  • Triage remaining untargeted bugs.
  • Start working on rtm++ bugs.
• Nisheeth Ranjan (nisheeth)
  • Triage rtm nominated bugs.
  • Work on XML Query use case for querying RDF. Couldn't do this last week because I got pulled into doing nsbeta3++ stuff.

Lowlights

• None.

Escalations

• None.

Since Jean-Baptiste M. Queru from Be Inc. provided us with a solution that worked for loading that much shared components (on Be, called add-ons) the team has not stopped progressing. We got the component loading fix, the file handling fixed and now the timer and gfx components are working too. We're missing the network and input, they are implemented but broken. On our page, http://www.mozilla.org/ports/beos/, we've put up a new design and we keep updating the status. Plus, there's the screenshot of how Mozilla look without any skin applied, on BeOS!

It was a very successful week for the Warpzilla project.

Highlights

• Font switching is now working
• We are able to view DBCS web pages
• New keyboard support is in - should improve NLS
• Clipboard support is in
• Found unported code in plevent.c - fixed random traps

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

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

We have merged the Bidi code up to the tip and are trying to get it built.

Done

• replaced nsILDAPMessage::{first,next}Attribute methods with a threadsafe, getter_Copies, JS-friendly nsILDAPMessage::GetAttributes that returns an array of strings.
• fixed nsLDAPMessage::GetDn and nsLDAPMessage::GetValues to properly implement XPCOM getter_Copies() semantics
• completed migration to contractids and jars
• improved performance of nsLDAPChannel significantly. Performance on large searches is not yet usable, but this seems to be due to a lack of prioritization in the event system (bug 50104).
• implemented a macro to free XPIDL pointer arrays (nsLDAPInternal.h). currently named NSLDAP_FREE_XPIDL_ARRAY. perhaps it should be renamed NS_FREE_XPIDL_POINTER_ARRAY and moved somewhere into mozilla/xpcom.
• Heavily triaged TODO.txt in preparation for some sort of 0.x release of the LDAP URL handler.

Priorities

• Re-work error-handling in nsLDAPChannel
• Finish implementing nsIProgressEventSink support
• Implement nsLDAPConnection thread-shutdown

Home page