[MPlayer-dev-eng] [PATCH] Prevent ao_win32 from hanging during uninit(0)

[David Bolen]

Reimar Döffinger <Reimar.Doeffinger at stud.uni-karlsruhe.de> writes:

> On Thu, Mar 05, 2009 at 02:47:59PM -0500, David Bolen wrote:
>> Any downside to just bounding the loop for now?  I do have a
>> self-interest in looking into this further but in the meantime, it's a
>> hang exposure which the patch corrects.  Certainly it's an
>> improvement, even if imperfect, over the existing code.
>
> Well, I am undecided if it is worth the effort, I consider just going
> with the ao_sdl solution (which in some ways is also more correct),
> namely:
>>  if (!immed)
>>    usec_sleep(get_delay() * 1000 * 1000);

Yep, you win, that's cleaner :-)

Ironically, the "calculation" patch was a cleanup of the actual
"sledgehammer" approach of a fixed time limit I had put in to stop the
pain.  I guess I should have looked around further for other examples
such as ao_sdl.  Still trying to get a handle on the code base...

Was your "worth the effort" point in regards to any patch at all or
just my particular one?  I still think preventing the possibility of
an infinite loop in uninit() to be worth it, even if it doesn't solve
all the problems with the module.

-- David