Summary
- Bug fixing, triaging continues
- Documentation on Necko updated at mozilla.org
- http://status/reports/ demo'ed at all-hands
Highlights
- darin
- fixed bug 47617 "Connection to https needs to tell user to install PSM if w/o"
- fixed bug 113949 "Trunk crash [@ nsFileTransport::Process]"
- fixed bug 114660 "data: URLs get sent as http-referer"
- fixed bug 115129 "stream loader optimizations"
- fixed bug 114286 "No way to determine if a http request was a 304 or not."
- filed bugs for a number of small startup/page-load optimizations in necko, roughly 3% worth of startup improvement.
- presented a general overview of necko on monday.
- attended IDN overview and demo meeting on friday.
- worked on necko documentation.
- triaged bugs.
- neeti
- Fixed bug 114439 - nsFileTransportService should cache the mime service.
- Triaged bugs.
- Measured page load times by removing all the monitors using proxies for nsComponentManager. We get very erratic results. Results range from 7% to 0% improvement.
- gordon
- Squishing bugs:
- WONFIX bug 89903 "Slow Page Load with Cache on Remote Drive"
- FIXED bug 84531 "DNS cache eviction count not always decremented"
- WORKSFORME bug 108388 "Caching of DNS entries has stopped"
- WORKSFORME bug 92069 "MacBinary is recognized as Untyped binary data"
- create patch for bug 110405 "Assertions in cache code"
- duped and triaged more bugs.
- worked on bug 105344 "Memory cache pref should be a percentage of physical RAM"
- worked on bug 81724 "disk cache needs stream wrappers"
- dougt
- Changed file protocol to optionally generate html content for embedding customers. 115135
- Investigated cost of PRLock locking against a simple Spin LOCK in component manager - No detectible difference.
- (Again) asked for feedback on the nsIComponentManager interface changes.
- Investigated reducing stat calls when file code need to produce a url string. 106815.
- Reviewed a bunch of patches including Pavlov's timer code 115473.
- pavlov
- Bug 78611. Landed new XP timer code.
- Bug 108161 has a fix for the multiple reloading of images on reload/shift reload. Probably won't land it until I get back from vacation.
- gagan
- attending several meetings as virtual Clayton. Still alive.
- submitted patches for bug 89701
- added support for automatic bug links in status reports (converting bug numbers to bugzilla/bugscape links)
- finally demo'ed status reports at the all hands.
- helped DP with his linux problems :-)
- attended exec reviews for AIM, Navigator and JS2(dry run)
- attended Focal Pt. Training
- pushed updated documentation for Necko onto mozilla.org site.
Lowlights
- gordon
- The tree never seems open during the day...
Plans for next week
- darin
- fix more bugs for 0.9.8 before taking off for the holidays.
- triage bugs for 0.9.8 milestone.
- gordon
- pavlov
- More bug triage.
- Starting vacation on December 19. I'll be back on January 2.
Issues
Summary
- Lots of bug triage across the board; untriaged number is now about half
of what it was last week (=85 now). Numbers for 0.9.8 too high though :(
- We have some important fixes almost ready to land:
FIXptr
, P3P compact
policy
and XML parsing
speedup
. The new SOAP implementation is also getting tested on the trunk, and a
huge XSLT patch that Peter has been working on for months is finally in super-review
phase.
- Good performance/footprint wins on several fronts, see
bug 114320
, bug 113657
, bug 92618
and bug 114539
.
- jst will go on vacation 12/18, heikki 12/21; both back
1/2/2002. harishd will be on vacation 12/22-12/26. Nisheeth will be back
second week of Jannuary.
Accomplishments
- harishd
- Fix in hand for
bug 114320
[ Removed use of malloc from nsExpatTokenizer::LoadStream ]
- Fix in hand for
bug 113657
[ Shrunk token size - Thanks to dann@godzilla.ics.uci.edu & Randell
Jesup for their help]
- Fix in hand for
bug 92618
[ Identified code that shouldn't belong in optimized build and moved 'em
under DEBUG flag ]
- Fix in hand for
bug 115262
[ nested FORM regression ]
- Worked on
bug 105137
[ Clean up - moved a few of nsIContentSink APIs into nsIHTMLContenSink ]
- Worked on
bug 98687
[ Deduced a reduced testcase ]
- Did code reivew for heikki.
- Triaged bugs.
- jst
- Fixed bug
114539
. Got rid of mInner in all our 'data' nodes and made those classes inherit
nsGenericDOMDataNode in stead. Also removed mContentID from nsTextNode and
pushed mRangeList and mListenerManager into hashtables since they are used
so rarely. This results in a size reduction of nsTextNode from 44 bytes to
28 bytes (~36%).
- Fixed
bug 114581
. Made table.deleteRow() throw a INDEX_SIZE_ERR exception if the index is
out of range, as per the DOM HTML spec.
- Fixed
bug 112217
. DOM interface cleanup.
- Fixed
bug 112120
. Cleaned up the GetPrimaryFrameFor() code in nsGenericHTMLElement and making
it obvious from the method name that this method is form control specific.
- Did some random code cleanup here n' there.
- Worked on making iframes load their documents even
if the iframe is not shown.
- I'll be traveling to Finland on Tuesday (and back on
the 29th), should be back in action once I've settled down in Finland.
- mstoltz
- Rewriting security policy storage for better performance.
- Wrote patch to prevent unwanted window.close(), still
needs a confirmation dialog.
- heikki
- Finished review of the large XML parsing perf patch in
bug 105137
.
- Made fix for
bug 74800
, implemented FIXptr with scripting.
- Made fix for
bug 92127
, bookmarking image map links gets nothing for name, top100 bug.
- peterv and I switched our team's Mac to OS X, and investigated
remote builds. Seems like there is still some more work to be done in the
build system.
Plans for next week
- harishd
- Work on P3P compact policy.
- Work on performance issues.
- Work on 0.9.8 bugs.
- Triage bugs.
- jst
- Continue working on bloat issues.
- mstoltz
- More work on new policy db.
- Work on reviewers' guide.
- heikki
- Land the fixes in my tree.
- Still some bug triage to be done.
Here is a link to the Navigator team's progress report for last
week.
http://mozilla.org/xpapps/progress/2001-12-14.html
Tree
- Mozilla 0.9.7 froze and branched in much better shape than nightly builds
the week before.
- The tracking bug for fixes needed for 0.9.7 is
14455
.
Infrastructure
- MozillaZine
looking for donations to help it move to an ISP that can provide adequate
bandwidth
- working on getting
Distributed Checksum Clearinghouse
(DCC) installed on gila in order to remove spam from the mailing lists;
in the meantime, myk now killing spam in the rdf, unix webtools and xpfe
newsgroups
- still waiting for review of request tracker patch for inclusion in
Bugzilla
Events
Previous Updates
|