|
Here are some top level highlights of where we are on platform parity:
We achieved the main goal of the milestone which was to remove
the major roadblocks to development work on Linux and Mac.
These two platforms have seen some dramatic improvements in
stability and functionally thanks to a number of great contributions.
88 Platform Parity related bugs were fixed in the last few weeks. For details run this bugzilla query.
The bigger indication of parity might be the fact that apprunner
on Windows, Linux, and Mac is loading a common XUL file and
hyatt, ramiro, and saari have a pretty hot demo running off
of code that is all checked in.
One of the other goals of platform parity was to get
a good assessment of how much additional work is needed
to bring the platforms into alignment for smaller
non-development-blocking things and get a listing of
those tasks reflected in the bug systems. There is still
some work to be done.
This bugzilla query shows some of the PP work that will be continuing.
The next milestone is what we are calling "Dogfood Ready"
and the goal is to get a version of the client put together
that can be used for all or parts of daily work.
Here is a rough Draft Acceptance Criteria For Internal
Mozilla/Netscape Dogfood Communicator.
- Browser:
- Be able to load and view all the major search engines, top 100 sites, and
seamonkey engineering frequently visited sites.
- home.netscape.com
- my.netscape.com including support for basic functionality of Cookies
including support for basic functionality of Javascript
- Read, write, and manage buglist on Bugzilla.
- View and interact with tinderbox and bonsai.
- Pass all the basic gecko tests and be able to view and Navigate some
list of sites along the lines of the top 100, browser buster, or other,
with MBTF of X (where X is about an hour)
- Be able to view the always troublesome http://w3 page and search for
phone numbers on http://w3
- Have some notion of URL history
- Be able to control the operation of the browser with buttons:
- back
- forward
- reload
- home
- bugzilla
- stop
- simple text entry in a URL location window
- copy and paste a URL location?
- Provide status bar to indicate page download and loading progress.
- Make use of cookies provided by all of the sites listed above
- Download files via FTP
- Ability to read webmail (possibly messenger express)?
- Ability to ready and write 5.0 preferences to control the operation of the
browser?
- Ability to set and unset prefs with pref UI elements?
- Ability to migrate prefs, bookmarks, certs, address book,
mail, news and other user profile data created with 4.x
- (watch for new newsgroups and module owners for these areas.)
- Print a page
- Mail:
- download POP mail with leave on server option (IMAP is later)
- handle smtp
- display messages
- display folder lists
- delete, reply, forward messages
- construct mail using simple text editor
- uses new mail db
- works on Mac, Linux, Win
- Uses various XPFE: tree view, tool bars, buttons, sidebar, mail button in
browser.
- We will also be looking hard at adding some I18n improvements that are targeted
at allowing:
- browser busting evaluation on Windows, Linux, and Mac the top 100 JA sites
- mail reading Japanese HTML and plain text messages
- display Japanese mail folder lists
- delete, reply, forward on Japanese messages
- create Japanese mail via copy&paste
Furball has this update for us:
Grendel is coming along rather nicely at the moment. The menus are
working finally. They're built out of XML. It's something like
XUL. I'm building a dialog box builder that reads from XML
description. In the long run, it would be extremely nice to get a
widget builder that would read from XML description and build a
complete window. That's a little further in the future.
Edwin Woudt hacked the preference file manually as the dialog box for
preferences isn't available yet and was able to verify that IMAP
stores does indeed work. A screenshot can be found here
http://www.djo.nl/grendel_with_imap.gif. This is a very good sign.
So here's what we are working on:
- Edwin - verify that POP3 and NNTP is working.
- Jeff - Getting the address book to work with LDAP and such.
- Me - put the UI back together potentially hack Swing so the
message display will suck less.
I think that if we get dialog boxes up and running and mail sending
works (none of that fancy attachments handling yet), then we are
looking at an almost usable mail client. Getting the addressbook
working would be an added bonus.
Jeff and I are hoping that Grendel could be made usable again and have
set ourselves a deadline for the mozilla.anniversary.
Chris Waterson writes in with this RDF update:
This has been a week of many meetings and much gnashing of teeth as far as RDF goes.
We are slowly making progress with the Mail/News folks, getting their data sources designed, etc. There has been a lot of discussion (internally and on n.p.m.rdf)
about sorting and selecting data from datasources.
Hyatt rewrote the tree widget (for the third time!), and now it uses NGLayout frames based on the table code. I am in the process of modifying the
RDF-to-treewidget content model builder to compensate for the slight differences between the old and new content models. He has an awesome demo that you
should bug him about: I think it was posted to n.p.m.xpfe this week. (Note: Click here to see screenshots!) Here's what Dave Hyatt had to say about the tree widget:
What makes this so cool is simply that arbitrary HTML can be
placed inside the tree cells, which means that you can create a very
complex and rich UI for things like bookmarks, history, mail, etc.
Making a tree widget is now as easy as making an HTML4 table, and you
can do all of the things in the tree widget that you can do with tables
(since the tree is derived from and extending tables).
RJC is putting the finishing touches on the file system data source, which went into the tree earlier this week. He'll then be enjoying the sorting code...
Here's Akkana's update on the status of the editor (composer):
It's almost starting to be usable! Selection is much improved -- it
crosses frame boundaries, and shortly it will be possible to select
non-text items as well as text. The backwards-typing bug has been
fixed -- text is now inserted in the right order, and if something is
selected, it is deleted before the text is inserted. A Copy/Paste
framework now exists, hooked to the selection code, and sample
implementations have been written for Windows and Mac (no Unix as yet).
Preliminary specs are up on the web site
(http://www.mozilla.org/editor/) for editor dialogs and for table
editing.
Troy writes in with this update:
On the layout front, Kipp's going to be developing full-time under
Linux, so this week he's been setting machines up and getting his
development environment going.
I'm implementing support for fixed positioning and doing various bug
fixes.
Vidur Apparao has the DOM update this week:
The implementation of the DOM Level 1 Core interfaces is complete and ready for
exhaustive testing. While Level 1 HTML and Level 0 are largely working, I'm next
doing a completion pass on both.
Vidur Apparao has the DOM update this week:
Matthew Zahorik gave me
a BeOS port of NSPR this week and I reviewed
and checked it in.
This NSPR BeOS port has been successfully used
in early ports of mozilla and Apache 2.0 pre-development
builds.
He thanks the following people for their contributions to
the NSPR BeOS port:
- Michael Bayne
- Steve Wingard
- Richard Hess
- Steve Sakoman
- Howard Berkey
- and the Bezilla team and the staff at Be Inc.
Previous Updates
|