[Webkit-unassigned] [Bug 72155] Replace Qt QThread threading back-end with pthread/Win32 threading back-ends

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=72155


Andrew Wason <rectalogic at rectalogic.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #115875|0                           |1
        is obsolete|                            |
 Attachment #116639|                            |review?, commit-queue?
               Flag|                            |




--- Comment #17 from Andrew Wason <rectalogic at rectalogic.com>  2011-11-25 14:37:07 PST ---
Created an attachment (id=116639)
 --> (https://bugs.webkit.org/attachment.cgi?id=116639&action=review)
initialize webkit_main_thread_invoker() on main thread

Updated the patch to initialize webkit_main_thread_invoker() on the main thread. The recent changes to dispatchFunctionsFromMainThread() from bug 72704 prevent the native threads from being infected by QAdoptedThread - but we need to only create webkit_main_thread_invoker() on the main thread or else the thread it is first created on becomes a QAdoptedThread leading to problems discussed in bug 72538.

So I believe this patch now also fixes bug 72538 - i.e. Qt no longer needs to wait for worker threads to exit when shutting down since the workers are plain native threads.

However, my test application still crashes occasionally - but not necessarily during shutdown. I was able to reproduce the crash in a test app that never shuts down (just periodically reloads a QWebPage with workers) - so I think that is a different bug.

So I think we should proceed with applying this patch, assuming it passes review, and then I'll close bug 72538 and open a new bug for this new worker crash issue.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.