Re: Event loop hangup when using TimeoutSource and IdleSource

[Jim Hodapp <james hodapp gmail com>]

Thanks Chris,

    I'll be trying to track this down. I believe it's a problem with  
my Emsac instance reference possibly being invalid from time to time  
as I had a run where the exact same thing happened where it segfaulted  
when trying to set the text of a label and the *this* pointer address  
was not NULL this time. I suspect it was invalid though. I'll keep you  
updated if I have more questions.

Thanks for the prompt help today!

Jim


On Sep 24, 2009, at 6:06 PM, Chris Vine wrote:

> On Thu, 24 Sep 2009 17:46:40 -0400
> Jim Hodapp <james hodapp gmail com> wrote:
> > Also, you mentioned the boost signals2 library. How do you use this
> > library instead of sigc with glibmm objects like Glib::IdleSource or
> > Glib::TimeoutSource. They seem to want sigc slots specifically.
>
> You can't, although it is relatively trivial to make your own wrapper
> for g_idle_add_full() and g_timeout_add_full(), both of which are
> thread safe, using boost::function and/or boost::signal2 if you so  
> want.
>
> But I think this is a red herring, since you do not (you say) allow
> more than one thread to get into any of your application functions.  
> You
> have a much simpler null pointer issue by the looks of it.
>
> Chris