Re: [E-devel] ecore_con valgrind errors

2011-02-19 Thread Mike Blumenkrantz
On Sun, 20 Feb 2011 13:03:34 +0900 Carsten Haitzler (The Rasterman) wrote: > On Sat, 19 Feb 2011 12:25:05 -0500 Mike Blumenkrantz said: > > > Should be r56254 > > looking at the code... the bt makes no sense. when that struct is freed... > it's removed from the infos list by ecore_con_server_i

Re: [E-devel] ecore_con valgrind errors

2011-02-19 Thread The Rasterman
On Sat, 19 Feb 2011 12:25:05 -0500 Mike Blumenkrantz said: > Should be r56254 looking at the code... the bt makes no sense. when that struct is freed... it's removed from the infos list by ecore_con_server_infos_del. unless the same struct has been added TWICE... and i dont see that as it'd only

Re: [E-devel] ecore_con valgrind errors

2011-02-19 Thread Mike Blumenkrantz
On Sat, 19 Feb 2011 23:28:50 +0900 Carsten Haitzler (The Rasterman) wrote: > On Sat, 12 Feb 2011 00:32:30 -0500 Mike Blumenkrantz said: > > which commit? > > > Hey Raster, > > > > The following errors (when connecting to a server) are new > > since you added that code which fixed the epoll bu

Re: [E-devel] ecore_con valgrind errors

2011-02-19 Thread The Rasterman
On Sat, 12 Feb 2011 00:32:30 -0500 Mike Blumenkrantz said: which commit? > Hey Raster, > > The following errors (when connecting to a server) are new > since you added that code which fixed the epoll bug pre-1.0. It seems that > the lookup data is attempting to clear its own pointer after the c