Mozilla Stuff

I spend a lot of time working on Mozilla. I've been involved with testing the new layout engine since late summer of 1998 and I moved from mainly testing to working on code (and analyzing leaks and performance problems) gradually in late 1999 and early 2000. I'm currently the Style System module owner (I'm responsible, along with peers, for the code that handles CSS stylesheets and computes style data from CSS stylesheets and HTML attributes) and bugzilla component owner (so I get all the incoming bugs for such things assigned to me by default. I also work a good bit on the Layout code, mainly in the parts dealing with the implementation of block (things like P) and inline (things like A) layout. Beyond that, I'm one of the super-reviewers (for the second level of code review) and one of the drivers (responsible for release management decisions, such as when releases happen and what gets checked in during the time leading up to the release).

Parser and layout modes

These documents are now obsolete, and newer versions are on the site.

Standards compliance

Writing code

I wrote a little code before I had a CVS account, but I've written good bit more code through 2002 and more code from the start of 2003 until the mercurial switch in 2008, and then more changes and more changes in Mercurial. My patch queue is also publicly available. I've also checked in some changes to the website, and more changes from the start of 2003.

Also see my user repositories where I publish code (since we've switched to Mercurial).


Crash reports (2001-2002)

These are manually updated, whenever I think of it (that is, whenever I run the script that updates them). I also manually maintain the list of bug numbers that the script uses.

Crash reports (2009)


Mozilla X-Remote

On my local machine (Red Hat 8.0), I use a .htmlviewrc file to change the behavior of /usr/bin/htmlview.

I set up a hack that allows someone using a mailreader like mutt on a remote Unix mailreader to view HTML email using a local Mozilla by encoding it as a data URL. It uses a perl script to send the data URL to a modified version (with all the features of the original still there, although I'm not sure whether that was a good idea) of the original reference X-Remote client via a pipe rather than a command-line parameter (so it doesn't show the contents of the whole file on the command line, and perhaps for large files as well). (The perl script requires Perl5 and the perl module, which needs to be installed in MIME/ in some directory in perl's include path.)

Memory Leaks



See talks page.

Valid HTML 4.0!

(Back to David Baron)

LDB,, 2000-05-20, 2000-07-11