Fixing Firefox and Adapting to It

At Kloudshare, we are developing a platform that, among other things, pushes the web browser further than most applications, and perhaps further than almost any application has yet pushed the browser. When I say push I don't mean in the sense that we expect it to go particularly faster than others do, but that the way that we are exercising it is that we are using things in presumably unexpected combinations.

Up until now, Firefox has been OK, but there had been some visual glitches that we would like to eliminate. This is hard because, as far as we can tell, we are doing everything exactly right in a standards compliant sort of way. The bugs of most concern seem not to have any obvious thread, and most tellingly, they do not appear in other browsers.

The visual glitches are a bit annoying, but can be lived with. But yesterday we hit one that is really weird. When changing the source of an iframe, the browser seems to change its zoom level. Everything gets bigger. When you click to another tab and then come back, everything goes back to normal. This is a real killer for us. Its not just a few pixels intermittently out of place, and so it makes the experience unusable. The problem is not consistent in the sense that on my Mac it seems to happen every time, but on our Linux machine the problem is more intermittent.

The thing I have been struggling with is trying to figure out how to both report this and whether there are ways to work around this. We are not live yet so we can't provide some kind of public link on Bugzilla. We have thought about trying to capture the HTML but many of these bugs require interaction. More troublingly we really want to do more than just provide information that helps make the next version of the browser better, though obviously that would be great. Ideally we'd like to know what is going on internally, and if there is a way to avoid or work around these problems so that we can satisfy the existing installed base without removing important features.

So the real question here is, what is the right strategy to resolve this. Is there a proper procedure for resolving such problems? I suppose we could try to develop a specific use case outside our software that reproduces the bug. But given the highly circumstantial nature, this would be a huge burden on us, given our small size. Another thought is that we could hire someone who is a contributor to see if s/he can reproduce the bug and help us both identify the bug and, if possible, find a work around. The problem with this solution is I am not sure where to find such contributors. I guess I could try hanging out on IRC, but that seems pretty inefficient. Then again perhaps not.

Anyway, this post is more of a call for help and suggestions. Any ideas greatly appreciated.