This is caused by the patches in bug 727960, unapplying them fixes the problem.

I should be more specific: The native front end starts up but Gecko does not.

(In reply to Benoit Girard (:BenWa) from comment #0) > This is caused by the patches in bug 727960, unapplying them fixes the > problem. Did you try unapplying them one by one?

No, both together, do you have a guess and I'll try it?

1419f0146049 is a likely suspect

Is this actually maple-specific? Presuming not, I'll remove the whiteboard tag.

With 1419f0146049 unapplied I receive a corrupted stack crash. I do receive some log from Gecko so it is starting up: E/Gecko (26551): ### Acquire() E/Gecko (26551): Registered Compositor I/Gecko (26551): ... requesting window surface from bridge I/Gecko (26551): got surface 0x3c6e70 I/Gecko (26551): -- creating basic, not accelerated E/Gecko (26551): ### Basic layers drawing E/Gecko (26551): ### Calling EndDrawing()!

From a quick test, it looks like GeckoThread.run is not even called.

This is actually a regression from bug 728240. Reverting it fixes debug intent for me.

I wouldnt recommend reverting the patch. Instead, could you please add a condition there to do debug intent? Also, onResume() is called whenever Fennec comes to foreground. So, this intent should be processed only once in initialize()

This patch adds the debug intent processing to the startup path.

Comment on attachment 603335 [details] [diff] [review] Patch Review of attachment 603335 [details] [diff] [review]: ----------------------------------------------------------------- It looks good to me (and, more importantly, works), but I won't pretend I know this code to review it.

https://hg.mozilla.org/mozilla-central/rev/f1a5fd7d0baf (Please set the assignee when attaching and check when landing, thanks :-))