I completed my GRE patch, waiting for a review now. Luckily I think it will still be possible to run current galeon/epiphany code with just configure changes (keep to link to libxpcom at compile time basically).
On the bad side I doubt we will be able to have multi screen support any time soon. Mozilla widget/gfx code is not multihead aware. Most of the code would be easy to fix (apart the mess of having to keep gtk 2.0 compatibility), though there are some services that are global while they should be per-screen or per-display (most notably the look and feel service). Fixing this looks .... hard :/ It's a very frustrating situation: the incompatibilities between mozilla and GNOME platforms (or more in general linux platform, see the situation with translations) are a major block to further integration work. It's so demotivating to have the whole picture easily figured out on the GNOME side and then block on hardly solvable incompatibilities.
I think it's now clear that the two communities need to merge their efforts and build a common roadmap. The sad browser situation is only one of the issues. GNOME needs several mozilla.org functionalities (rendering engine and XUL for example) and mozilla.org platform needs a desktop (as a set of technologies, an user interfaces system, a set of guidelines, a GUI design philosophy) to be grounded on. Web technologies are going to increase (even more) their importance on the desktop and we need to ensure they are properly integrated in our platform ASAP. GNOME has no resources to reimplement XUL or a rendering engine. Mozilla.org has no resources to reinvent the desktop. If you add a components system and the ability to use multiple languages on the top of it well ... that's more or less what we are looking for. A native frontend is a good solution to the browser problem, the only viable on the short time, but it only goes so far.
Brendan "proposal" is interesting and we should carefully consider it. If mozilla.org is finally considering the free desktop as a primary target, and they are willing to promote _real_ integration with the free platform, then I think there is space for a merge. It's a lot of work and not only code to write or design. There are social spaces to share, conventions, habits, tools, roadmaps to merge. Though I think it's a necessary step for linux to succeed on the desktop.
(hoping to not get out of my dream to discover this is just a marketing strategy to push some applications in Linux distributions...)