The current Gecko script in Orca:
* Is largely a bunch of hacks to work around Gecko or Firefox bugs.
* Originated in the Firefox 3.3ish days.
* Contains much of the support for Firefox when instead it should
  just be a toolkit script. (A Firefox script should support browser
  specific functionality.)
* Is non-performant.
* Is pretty large.
* Contains some code which should be pulled out and made available to
  all modules within Orca. (label guess, for example)
* Is likely outdated due to fixes made in Gecko or Firefox itself.

Given all of the above issues, and the fact that we're early in the 3.1 release cycle, and the fact that we have a stable GNOME 3 branch and an unstable, alternative branch (orca-xdesktop), now seems like an excellent time to fully analyze the current situation and fix it, staring with just the master branch. 

Depending on what is discovered, it might even be worth scrapping the current Gecko script and more or less starting over from scratch. Not only would this result in a vastly cleaned up Gecko script, but also give us the opportunity to identify issues/bugs in Gecko which really need to be fixed on their end rather than hacked around on ours.

The Gecko script is dead. Long live the Gecko script.