[MPlayer-dev-eng] vo_pts ao_pts

[David Holm]

Felix Buenemann wrote:

>On Sunday, 18. November 2001 02:35, you wrote:
>
>>Hi,
>>
>>>static int play(void* data,int len,int flags)
>>>{
>>>    ioctl( fd_audio, EM8300_IOCTL_AUDIO_SETPTS, &ao_pts );
>>>    return write(fd_audio,data,len);
>>>}
>>>
>>>write? yea, I'm quite sure this is no standard glibc funtion and it must
>>>be magically resolved at linktime through some lib. Actually it seems to
>>>be C++ lib function that returns an iostream.
>>>
>>man 2 write
>>it's standard kernel syscall.
>>
>oops, I just checked man 3 and grepped through /usr/include/* :(
>actually I didn't think of a syscall at all...
>
>>>I'm quite sure the right thing(tm) is to replace write with standard libc
>>>guntion fread (look at man 3 fread).
>>>
>>heh? fwrite maybe? but Holm did the right thing this time.
>>
>yes of course fwrite.
>
>well actually it's easy to track down by monitoring function returns, I just 
>wonder ao_dxr3 seems to have no debug code for doing such stuff as I already 
>told David that it's problem with returns of timing functions in ao driver 
>long time ago, maybe he overread my mail.
>
Thanks for the tips guys. Since it stayed ==0 all the time I thought 
perhaps it was unimplemented. The reason I have no debug code is that 
I'm from the old school of dos ;). Where you learned to keep cpu cycles 
to an ABSOLUTE minimum, unnecessary tests etc would slow your 
application down, nowadays this is a handicap, I'm doing my best to 
break an old habit. But old habits die hard, as well as "Assumption is 
the mother of all fuckups" (in relation to that I assumed ao_pts==0 => 
ao_pts is unimplemented or I need to pass a flag somewhere)...
Thanks again

//David Holm

>
>>A'rpi / Astral & ESP-team
>>
>