Re: [Gnash-dev] firefox segfault again

[strk]

FYI: the crash is fixed by commenting out
the 1-second timer attached to the glib (?) main loop.

I guess it is because the plugin instance dies and thus
all its memory is reclaimed and not assigned at time of
invocation. It is unsafe to register a callback to
an object data in the object destructor...

What's the problem that timer is there to solve ?

--strk;

On Wed, Mar 31, 2010 at 09:05:28AM -0400, dolphinling wrote:
> On 03/31/2010 08:07 AM, strk wrote:
> >
> >I can reproduce 100% by going here:
> >http://www.youtube.com/watch?gl=US&feature=related&v=haZeXUAQRtc
> >
> >And then, while it plays [1], clicking on the YouTube logo on top-left.
> >
> >Firefox version: 1.9.0.18
> >Backtrace: http://strk.keybit.net/tmp/BT
> >
> >[1] I'm advertising Flash 8.0 r0 in order for that to play
> 
> If you can get a mozilla.org build of firefox instead of a distro-provided 
> one, firefox's built-in crash handler (breakpad) provides a lot more 
> information than a simple backtrace. Here's a sample report of the adobe 
> player crashing: 
> http://crash-stats.mozilla.com/report/index/12202c88-8fff-4301-9379-6cdac2100331
> 
> It may be worth looking at the alpha builds of firefox: they run plugins in 
> a separate process, so when the plugin crashes the browser doesn't. Plugin 
> crashes are reported somehow, but I'm not sure if the information in them 
> is better or worse or the same as in older versions of firefox.
> 
> -- 
> dolphinling
> <http://dolphinling.net/>
> 
> 
> _______________________________________________
> Gnash-dev mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/gnash-dev

-- 

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html