[Webkit-unassigned] [Bug 183348] REGRESSION(r222772): [GTK][WPE] WebProcess from WebKitGtk+ 2.19.9x SIGSEVs in WebKit::WebProcess::ensureNetworkProcessConnection() at Source/WebKit/WebProcess/WebProcess.cpp:1127

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=183348

Carlos Garcia Campos <cgarcia at igalia.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |magomez at igalia.com

--- Comment #27 from Carlos Garcia Campos <cgarcia at igalia.com> ---
(In reply to Michael Catanzaro from comment #24)
> Comment on attachment 337489 [details]
> Patch
> 
> Wow.
> 
> Just... wow.
> 
> How did you pin this to r222772? I'd kind of favor reverting that instead,
> since I don't like the platform-specific behavior here, but I guess this is
> OK and I like the test. It needs an owner, though.

Reverting is another option, Miguel has confirmed that r222772 didn't fix the issue in all the cases and we are still not cleaning up everything related to GStreamer. I don't want to make big changes about this in 2.20, so I'm going to merge this commit there to release 2.20.1. For trunk, Miguel is checking it, but I'm afraid any solution will require platform ifdefs anyway. My other option was to go back to exit on sync failure, but changing the exit for a main loop quit, and explicitly release whatever we need (cookies, GStreamer, etc) in the main function after the main loop quits.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20180410/ff6b8e1c/attachment-0002.html>