[MPlayer-dev-eng] [PATCH] wrong behaviour with slave command 'stop'

[Reimar Doeffinger]

On Mon, Aug 04, 2008 at 03:38:23PM +0200, Benjamin Zores wrote:
> Mathieu SCHROETER wrote:
> > Hello,
> > 
> > I've written some weeks ago a patch to add a slave command
> > 'stop' especially for "-slave -idle" modes.
> > 
> > (cvslog r26909)
> > http://lists.mplayerhq.hu/pipermail/mplayer-cvslog/2008-May/034513.html
> > 
> > But it seems to have a problem.
> > 
> > To reproduce (easly reproducible with dvd://):
> > 
> >  $mplayer -idle -slave -quiet
> >   <snip>
> >  loadfile dvd://
> >  stop
> >  loadfile dvd://
> > 
> > And here, an error because dvd_angle = 0 instead of 1.
> > 
> > 
> > And:
> > 
> >  $mplayer -idle -slave -quiet
> >   <snip>
> >  loadfile dvd://
> >  loadfile dvd://
> > 
> > Here there is no error on dvd_angle.
> > 
> > 
> > After a lot of debugging with gdb, I've found my mistake in command.c.
> > 
> > When we use `loadfile` to stop playback, loadfile calls the
> > functions "play_tree_iter(_up)_step()" and these functions calls
> > "m_config_pop()". If I've understood correctly, m_config_pop() restore
> > the previous configuration. And at least, restore "dvd_angle = 1"
> > in my example.
> > 
> > In this case, I've written this patch to fix this behaviour with
> > the slave command 'stop'.
> > 
> > I've looked in "case MP_CMD_LOADFILE:" to adapt for 'stop'.
> > I hope that someone can review this patch, it works fine for me
> > but I'm not sure to use the right way. Thanks..
> 
> I'm gonna apply this patch within the week unless someone
> (random pick ... Ivan ?) objects to.

Have you checked it is still necessary and correct? I think there was an issue due to
there being a m_config_pop too much somewhere and this was fixed already.
Unfortunately I do not have internet at home so I can not properly check right now :-(