On Wed, 3 Jul 2013 23:01:47 -0300 Jose Souza <zehortig...@profusion.mobi> said:
> Now it increase reference of libdbus_connection not eldbus_connection, some
> callback are removing the last reference of eldbus_connection.
hmm that may be true. i didnt check valgrind. :)
> On Wed, Jul 3, 2013 at 10:57 PM, Carsten Haitzler <ras...@rasterman.com>wrote:
>
> > On Wed, 3 Jul 2013 22:33:18 -0300 Jose Souza <zehortig...@profusion.mobi>
> > said:
> >
> > > On Wed, Jul 3, 2013 at 8:42 PM, Carsten Haitzler - Enlightenment Git <
> > > no-re...@enlightenment.org> wrote:
> > >
> > > > raster pushed a commit to branch master.
> > > >
> > > > commit bdd57af1136dd6c81244c15b0456ec06d10fa59e
> > > > Author: Carsten Haitzler (Rasterman) <ras...@rasterman.com>
> > > > Date: Thu Jul 4 08:41:45 2013 +0900
> > > >
> > > > eldbus - only unref client conns that are not null after dispatch
> > > > ---
> > > > src/lib/eldbus/eldbus_core.c | 3 ++-
> > > > src/lib/ethumb_client/ethumb_client.c | 2 +-
> > > > 2 files changed, 3 insertions(+), 2 deletions(-)
> > > >
> > > > diff --git a/src/lib/eldbus/eldbus_core.c
> > b/src/lib/eldbus/eldbus_core.c
> > > > index 040fe4c..a92d1e2 100644
> > > > --- a/src/lib/eldbus/eldbus_core.c
> > > > +++ b/src/lib/eldbus/eldbus_core.c
> > > > @@ -772,7 +772,8 @@ eldbus_idler(void *data)
> > > > dbus_connection_ref(conn->dbus_conn);
> > > > DBG("Connection@%p: Dispatching", conn);
> > > > dbus_connection_dispatch(conn->dbus_conn);
> > > > - dbus_connection_unref(conn->dbus_conn);
> > > > + // add the if check as dispact may set conn->dbus_conn to NULL in
> > a cb
> > > > + if (conn->dbus_conn) dbus_connection_unref(conn->dbus_conn);
> > > >
> > > Huuum I think should be better increase eldbus_connection reference
> > before
> > > call dbus_connection_dispatch().
> > > conn->dbus_conn only will be null if some cb remove the last reference of
> > > eldbus_connection, so this will ran invalid memory.
> > >
> > > I will fix this tomorrow at work.
> >
> > it already does do the _ref() before dispatch. the problem is that
> > dbus_conn is
> > NULL... some callback dispatch() calls sets it to NULL. actually i have a
> > better solution to this... copy it locally.
> >
> > > > return ECORE_CALLBACK_RENEW;
> > > > }
> > > >
> > > > diff --git a/src/lib/ethumb_client/ethumb_client.c
> > > > b/src/lib/ethumb_client/ethumb_client.c
> > > > index cbd3c96..3d62fa1 100644
> > > > --- a/src/lib/ethumb_client/ethumb_client.c
> > > > +++ b/src/lib/ethumb_client/ethumb_client.c
> > > > @@ -238,7 +238,7 @@ end_connection:
> > > > obj = eldbus_proxy_object_get(client->proxy);
> > > > eldbus_proxy_unref(client->proxy);
> > > > eldbus_object_unref(obj);
> > > > - eldbus_connection_unref(client->conn);
> > > > + if (client->conn) eldbus_connection_unref(client->conn);
> > > >
> > > > if (client->connect.free_data)
> > > > client->connect.free_data(client->connect.data);
> > > >
> > > > --
> > > >
> > > >
> > > >
> > ------------------------------------------------------------------------------
> > > > This SF.net email is sponsored by Windows:
> > > >
> > > > Build for Windows Store.
> > > >
> > > > http://p.sf.net/sfu/windows-dev2dev
> > > >
> > >
> > ------------------------------------------------------------------------------
> > > This SF.net email is sponsored by Windows:
> > >
> > > Build for Windows Store.
> > >
> > > http://p.sf.net/sfu/windows-dev2dev
> > > _______________________________________________
> > > enlightenment-devel mailing list
> > > enlightenment-devel@lists.sourceforge.net
> > > https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
> > >
> >
> >
> > --
> > ------------- Codito, ergo sum - "I code, therefore I am" --------------
> > The Rasterman (Carsten Haitzler) ras...@rasterman.com
> >
> >
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
>
> Build for Windows Store.
>
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> enlightenment-devel mailing list
> enlightenment-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
>
--
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler) ras...@rasterman.com
------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:
Build for Windows Store.
http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
enlightenment-devel mailing list
enlightenment-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
