Highlights
- Mac and Windows modal dialogs are functional. Mac modal
windows even behave modally. Windows modal windows don't
behave modally, though they actually are. (danm, pinkerton)
- Doing gcc/Solaris builds in the background so we can start
purifying/quantifying the build. Already Quantify/Solaris verified
ramiro's image rendering fix, we went from 90%+ cpu usage
to 5% for showinganimated gifs! (mcafee, ramiro)
- Completed initial infrastructure for Clipboard support on
Windows. (rods)
- Wrote Clipboard Viewer to test clipboard support. It
may even be a nice addition to the product. (rods)
Lowlights
- We are having to decommit on getting system colors to work
with CSS. We thought we had something usable, but the code
we submitted was rejected by the module owner. As was pointed
out to us, this is not our area of responsibility anyway.
Accomplishments
- Changes to XP Clipboard class (transferable, dataflavor,
formatconverter) are about 80% done.
- Landed new box layout manager described here last week (some
widgets didn't like getting resized very small). Corresponding
new navigator.xul file should be checked in this afternoon. (evaughan)
- Created XPToolkit
MyNetscape channel. (pinkerton)
- Got command-key binding limping/bolted on to menu handling
on Mac. Still have some issues to resolve and other platform
work to do. (sdagley, saari)
- Completed initial XP interface for file/folder dialogs. (sdagley)
- Modal window API is as final as it will get until clients
start using it and shooting holes in it. (danm)
- Added Editor to the Tools menu, it Just Worked, that
was cool to be able to hack the file like that. (mcafee)
- Updated the Unix page: http://www.mozilla.org/unix
(mcafee)
Decisions
Priorities
- Finish M4 milestone features in time for them to be used by
apps developers.
- Fix M4 milestone showstopper bugs.
- Get clipboard working on GTK. (mcafee, rods)
- Get drag & drop working on Mac & Linux GTK. (mcafee,
rods, pinkerton).
- File/folder Picker (sdagley, rods, mcafee)
- Get key binding working and checked in on all platforms. (sdagley,
saari)
- Check in sample dialog code and update the docs. (danm)
- Get windows off into the GUI thread where they belong. (danm)
- Make Windows modal windows feel more modal. (danm)
- Write box layout documentation. (evaughan)
- Build XP scrollbar widget. (evaughan)
Issues
- The Linux event loop model is much different from Windows'
and the Mac's, bringing into question the viability of our entire
modal window story.
- Most of us could do a lot better at communication and
coordination. Please keep the appropriate people in the loop about
what you will be doing, are doing, and have done. Work with
people rather than just let them find out what you did afterward.
If you're checking stuff in, post a 1-2 liner to xpfe.checkins.
Don't forget this is open source folks, nearly everything should
be happening on mozilla.org.
- Clipboard API /support will not handle large amounts of data
well. We will have to add streamable support for large data chunks.
People
- DanM, SDagley on vacation the week of April 12.
Jan Leger has this week's QA update:
- "Our initial posting of the Mozilla QA Home Page brought
in 7 volunteer testers within the first 48 hours! More help is
pouring in.
- Entries to the TCMS (Testcase Management System) Naming Contest
is now closed. We received 36 entries which will be posted for
voting next week.
- Have you hugged your bugs lately? Please continue to encourage
folks to read the new Bug Writing Guidelines for complete, accurate
bugzilla problem report information. Help our developers fix them
bugs fast! http://www.mozilla.org/quality/bug-writing-guidelines.html
"
Troy writes in with this update:
"Not much status for layout this week. I was on vacation
last week, which is why there was even less status. :-)
Mostly we're all fixing bugs. Kipp and Chris have their remaining
feature work on hold until their bug counts are reasonable."
Giao has this update for us:
"Many exciting things happened this week. Unfortunately,
they didn't make it into the repository but hopefully this problem
would be fixed by Monday. These are the patches I received this
week.
- From Joel York (Jazilla hacker!) - updates to Berkeley store
to implement a few methods that we never got around to finishing.
- From Mauro Botehlo - MBD interfaces. Mauro's shooting to implement
an MDB style system for Grendel. Grendel currently doesn't have
a DB for addressbook entries yet. If my memory serves me, we currently
use a LDIF file.
Here's what the regular Grendel hackers have managed:
- Edwin finished the multiple identity hacks. This means you
can now send from multiple stores and reconfigure the SMTP based
on the identity you have chosen.
- I committed patches to Grendel's XUL-like menu building process
so that you can isolate individual menu items from the constructed
menu. This means we can start looking towards context sensitive
menus.
What's still plaguing us:
- TreeTable suckness is still upon us. I have no clean solution
and I've been short on time to implement any ideas.
- XUL-like dialog building has suckness written all over it for
constructed widgets. This must be fixed and soon.
- Still having problems with message displays.
I'm going to work on getting the preference editor in sync with
Edwin's patches (after all this time!) to support the multiple
identities feature.
This is going to kick major ass."
Phil Peterson has news regarding the Mail/News client:
This week
- The Mork implementation of MDB is now reading and writing
databases
- IMAP test app is running and talking a little bit of protocol
- I18n work in message DB
- News component created, news DB subclass started
- New toolbar icons added into a new messenger-specific resource
directory in dist.
- Mail/news code is included on Linux Tinderbox
- Started message move/copy code
- Planning work for new netlib and multiple identities
- Updated mozilla mail/news web page
Next week
- Mork performance, robustness work
- Continuing work on IMAP
- Continuing work on News
- Continuing work on move/copy messages
- Mail/news code to be included on Mac Tinderbox
Here's Scott Furman's update on the status of XPConnect:
"In case you missed it, as of a week or so ago we have
the full XPConnect system up and working with XPIDL-generated
typelibs. John
Bandauer and Mike McCabe are beginning the process of transitioning
XPConnect from experimental to mainstream
use. XPConnect and its supporting libraries (libxpt, xptinfo,
xptcall) are now part of the client build on Unix and Win32. Work
on the Mac is proceeding, but lags the other platforms. The rough
edges of the XPIDL compiler are being smoothed out and it's being
ported to other platforms.
In case you haven't been following XPConnect in detail, libxpt
contains code for reading and writing typelibs, xptinfo presents
an xpcom interface for accessing typelibs, and xptcall is the
platform-specific assembly code that implements method dispatch
and impersonation of XPCOM interfaces. xptcall was recently extracted
from xpconnect and made more independent for use in a generic
method dispatch 'scriptable' system that Warren
is building for RDF and other things. Work on the Sparc port of
xptcall is mostly done. We are twisting arms :) to get a Mac PPC
port done at Netscape. Assembly-savvy volunteers willing and able
to port xptcall
to other architectures are appreciated!
Part of the team is ECMA-wrangling in Redmond this week. (ECMA
TC39 is the committee responsible for the standardization
of JavaScript.) There are still some serious battles being fought
among the committee members on the shape of the next-generation
JavaScript language. If you've got a spare minute or a thousand,
you can read Waldemar's
vision of the future
evolution of the JavaScript language.
Finally, Mike McCabe
whipped the unix build system for Perlconnect
into shape. (Perlconnect is the technology that bridges JavaScript
and Perl. Mmmmm - Two great tastes in one scripting language !)
"
Henry Sobotka has this update for us on the status of the OS/2
port:
"This week John Fairhurst released the first apprunner
binaries for OS/2 (mozm3ish.zip at Hobbes and Leo); while they
don't quite make it through the smoke tests, the beast is up and
running and impressively functional. We're now planning to sync
our code and check in a new OS/2 branch within the next couple
weeks."
Previous Updates
|