[FFmpeg-devel] [PATCH] ffserver misc bug fixes

[Howard Chu]

Luca Abeni wrote:
> Hi Martin,
>
> On 05/17/2010 01:59 PM, Martin Storsj? wrote:
> [...]
>>> +    } else
>>> +        av_strlcatf(buff, size, "c=IN IP4 0.0.0.0\r\n");
>>>    }
>>>
>>
>> This will make the SDP generator emit more c= lines than what is intended.
>> If you want to make sure the ffserver generated SDP contains a c= line,
>> it's better to populate the AVFormatContext->filename field for
>> non-multicast cases, too. The attached patch does this - Howard already
>> tested this and it seemed to work fine.
>
> Yes, thanks; your patch is much better.
> Baptiste, are you ok with Martin's patch?

Yes, thanks again Martin.

While I've still got my nose stuck in this code, I also noticed this silly bug 
- Unix time stamps are not RFC 822 compliant, and ctime() returns a result in 
local time, not in GMT, so these stamps were completely wrong.

ffserver still uses 100% CPU busylooping on input when receiving an ffm feed. 
I hope to find a way to improve this soon, there's no reason it should be such 
a CPU hog.

-- 
   -- Howard Chu
   CTO, Symas Corp.           http://www.symas.com
   Director, Highland Sun     http://highlandsun.com/hyc/
   Chief Architect, OpenLDAP  http://www.openldap.org/project/
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: dif4.txt
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20100517/ad452789/attachment.txt>