An ecore loop based libpurple client...

["Marco Trevisan (Treviño)"]

Daniel Atallah wrote:
> On Tue, Apr 14, 2009 at 7:24 PM, "Marco Trevisan (Treviño)"
> <mail at 3v1n0.net> wrote:
>> Marco Trevisan (Treviño) wrote:
>>> However during those days I've also fixed any other bug and now the
>>> client (look at the attachment!) seems to work without any segmentation
>>> fault, abort or any other error, however my connection troubles aren't
>>> over yet. :/
>> I performed new tests after upgrading to ecore svn r39932 (which
>> implemented the FD EBADF handling by Andre Dieb that fixes the
>> fd-related aborts) but after some time (it can vary from few seconds to
>> many minutes) that my client is connected to the prpl-msn protocol (I
>> used the HTTP method) I get a crash due to a SIGPIPE exception.
> 
> <snip />
> 
>> Program received signal SIGPIPE, Broken pipe.
>> [Switching to Thread -1214285024 (LWP 30051)]
>> 0xffffe410 in __kernel_vsyscall ()
> 
> Regardless of whatever else is going on here, this isn't actually a
> crash, you should be ignoring SIGPIPE while running in the debugger.
> 
> See http://d.pidgin.im/wiki/GetABacktrace for instructions on how to do this.

Ok, thanks. By The way also when I was running the client without the
debugger, if I recall correctly, I got some crashes caused by similar
errors...

However I'll perform my next tests ignoring the SIGPIPE signal at all.

Thanks.

-- 
Treviño's World - Life and Linux
http://www.3v1n0.net/