https://bugs.documentfoundation.org/show_bug.cgi?id=95128

Michael Stahl <mst...@redhat.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |caol...@redhat.com,
                   |                            |t...@libreoffice.org
           Assignee|libreoffice-b...@lists.free |caol...@redhat.com
                   |desktop.org                 |

--- Comment #10 from Michael Stahl <mst...@redhat.com> ---
looks like this should have made it stop crashing:

commit 27ccddf5034ec9e80a5adf93c630247bb5807676
Author:     Caolán McNamara <caol...@redhat.com>
AuthorDate: Mon Oct 19 14:16:50 2015 +0100

    afl-eventtesting: deref of null GetSdrPageView


(In reply to Michael Meeks from comment #6)
> Interesting trace here; nothing to do with GL that I can see; looks like a
> race condition. Michael - I'm curious as to why we're processing any events
> (except) ones incoming around eg. DC & hWnd handling during the
> SolarMutexReleaser re-acquire. By processing random other events - we seem
> to have burned a lot of stack to no good purpose (thoughts?).

interestingly we don't process events recursively as such, rather there is
a mechanism to install a listener with Application::AddPostYieldListener()
and there is evidently one in
 sdlo.dll!sd::SlideshowImpl::PostYieldListener(LinkParamNone * __formal) Line
1798    
that calls Reschedule() - so right after we return from SalInstance::DoYield()
we yield again from ImplYield()!  (there's a Yo Dawg meme going on here)

i, for one, find this is highly un-intuitive.

git log says that Thorsten touched this thing so he clearly knows
what it is doing, see

commit 044c534b2deb985052e7977bba3290c5d2c3ad77
Author:     Thorsten Behrens <tbehr...@novell.com>
AuthorDate: Tue Sep 6 16:25:49 2011 +0200

    Fix hang in slideshow.

    This fixes fdo#32861, by processing all remaining events after the
    post-yield callback comes around (that's supposed to be the time
    when the app can spend time on event processing).

-- 
You are receiving this mail because:
You are the assignee for the bug.
_______________________________________________
Libreoffice-bugs mailing list
Libreoffice-bugs@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs

Reply via email to