MacOSX fixes

[Christophe Massiot]

À (At) 23:32 +0100 26/11/01, Florian G. Pflug écrivait (wrote) :

>sorry to answer my own mail..

No pbm

>Looks like p_bit_stream->p_data->p_next would point to nowhere (but it's not
>NULL). When p_bit_stream->p_data becomes p_bit_stream->p_data->p_next (on
>line 134 in input_ext-dec.c), p_bit_stream->p_data is invalid,
>which leads to a crash in the while-loop condition
>(p_bit_stream->p_data->p_payload_start ==
>p_bit_stream->p_data->p_payload_end).
>
>At have no idea why p_data->p_next is bogus - Could someone with more
>background on the decoder structure give me a hand?

This is quite interesting. I have no idea, especially since it works 
on all other platforms. Could you try some stuff :
- find the guilty ES, play with --novideo, --noaudio, or both ;
- with only one ES activated, how many NextDataPacket do we go 
through before it crashes ?

I think I know where to look for the bug...


-- 
Christophe Massiot.