Re: [Ekiga-devel-list] Crashes and freezes in the account edition

[Julien Puydt <jpuydt free fr>]

Le 17/09/2012 09:57, Julien Puydt a écrit :
> Le 17/09/2012 09:11, Damien Sandras a écrit :
> > It's a freeze (deadlock), right ?
>
> No, it's a nice crash (segmentation fault). And it's quite difficult to
> reproduce (which is quite typical of race conditions).


I haven't managed to reproduce the crash when the attached patch is applied.

Can others try to reproduce the problem, apply the patch, and try again?

Snark

commit 01deeb9a0c86d3815b0ac386f23e44f34546ec8d
Author: Julien Puydt <jpuydt free fr>
Date:   Wed Sep 19 09:45:13 2012 +0200

    Remove code whose usefulness is unclear, but without which ekiga crashes less

diff --git a/lib/engine/components/opal/opal-account.cpp b/lib/engine/components/opal/opal-account.cpp
index fb71516..098d1c9 100644
--- a/lib/engine/components/opal/opal-account.cpp
+++ b/lib/engine/components/opal/opal-account.cpp
@@ -316,16 +316,6 @@ void Opal::Account::disable ()
   // registered (because a registration failed, for example), then the
   // next action won't change the status.
   endpoint->unsubscribe (*this, presentity);
-  if (presentity) {
-
-    // FIXME : the following actions should probably be done by opal itself,
-    // remembering what ekiga asked...
-    for (std::set<std::string>::iterator iter = watched_uris.begin ();
-         iter != watched_uris.end (); ++iter) {
-      presentity->UnsubscribeFromPresence (PString (*iter));
-      Ekiga::Runtime::run_in_main (boost::bind (&Opal::Account::presence_status_in_main, this, (*iter), "unknown", ""));
-    }
-  }
 }