[FFmpeg-cvslog] r16717 - trunk/ffmpeg.c

[Baptiste Coudurier]

Michael Niedermayer wrote:
> On Thu, Jan 22, 2009 at 11:35:59AM -0800, Baptiste Coudurier wrote:
>> Hi Michael,
>>
>> Michael Niedermayer wrote:
>>> On Thu, Jan 22, 2009 at 01:04:01PM +0100, michael wrote:
>>>> Author: michael
>>>> Date: Thu Jan 22 13:04:01 2009
>>>> New Revision: 16717
>>>>
>>>> Log:
>>>> Handle av_read_frame() EAGAIN return.
>>> This seems to break the ffm regression test as ffmdec.c returns EAGAIN at
>>> EOF which causes an infinite loop ...
>>> Iam not sure how to fix this, ideas are welcome!
>> Yes it seems design needs to be changed:
>> Add a new flag FLAG_EOF, and write a ffm packet with it in
>> ffm_write_trailer, then return -1 when this packet is read.
> 
> i can but this still will make truncated ffm files lock the player
> which in some cases like automated processing of files, especially
> user provided ones is not good (uploading ffm to youtube ...)

Hummm, well, true, we have a problem here.
So either we treat ffm in a special manner in ffmpeg.c temporarly, or we 
disable ffm regression tests for now, up to you, it will not break FFserver.

 >> [...]
>>
>> Also it's been some time Im thinking of adding EOS and EOF return values
>> instead of return -1 (an error). What do you think about that ?
> 
> after 1sec thinking (= no thinking) it sounds like a good idea :)

Ok, I'll think about it and try to come back with something.

-- 
Baptiste COUDURIER                              GnuPG Key Id: 0x5C1ABAAA
Key fingerprint                 8D77134D20CC9220201FC5DB0AC9325C5C1ABAAA
checking for life_signs in -lkenny... no
FFmpeg maintainer                                  http://www.ffmpeg.org