[MPlayer-cvslog] r27106 - trunk/playtree.c

[Benjamin Zores]

Reimar Döffinger a écrit :
> On Sat, Jun 21, 2008 at 09:45:44AM +0200, Reimar Döffinger wrote:
>> On Sat, Jun 21, 2008 at 12:41:14AM +0200, Benjamin Zores wrote:
>>> Aurelien Jacobs a écrit :
>>>> ben wrote:
>>>>
>>>>> Author: ben
>>>>> Date: Fri Jun 20 21:51:43 2008
>>>>> New Revision: 27106
>>>>>
>>>>> Log:
>>>>>
>>>>> Only "pop" subtree params if they had previously been "pushed",
>>>>> and afterwards reset the "pushed" value to 0 again.
>>>>> Similarly only set the PLAY_TREE_RND_PLAYED flag
>>>>> if the entry had been pushed before.
>>>> This patch patch was already submitted and rejected long time ago:
>>>> http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/2003-August/020356.html
>>>>
>>>> Somewhat later, it was revived, with the same conclusion (unacceptable):
>>>> http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/2007-September/053873.html
>>>>
>>>> What makes you so confident that this patch is now OK (to the point that
>>>> you don't even dare to discuss it before commit) ?
>>> Using it on geexbox (but you're aware of that !!) for years without any 
>>> trouble ?
>> I don't know, but for me code is supposed to be correct, not just
>> working, and I have very much the feeling that this at most hides but
>> not fixes a bug - though obviously I can be wrong.
> 
> Hm, since play_tree_iter_step already does this, why is that code needed
> at all??
> Also, I don't really understand why the
> pt->flags |= PLAY_TREE_RND_PLAYED;
> is in all those ifs...

Ok, my bad, I'll remove this one.

Ben