[PATCH] fix publishing on different servers
Hi all I'm writing application, that able to publish live rtmp stream. I tested it on wowza/fms/crtmpd. First one works fine with current implementation, but fms works only from 3-5 try. And crmpd connection not working at all. This due servers responding to "releaseStream" and "FCPublish" and sending unneeded "_result" answers, but librtmp using first "_result" as answer to "createStream". This little patch fixes this bug, and now publishing works fine on all servers, that i try.
Sergiy wrote:
Hi all
I'm writing application, that able to publish live rtmp stream.
I tested it on wowza/fms/crtmpd. First one works fine with current implementation, but fms works only from 3-5 try. And crmpd connection not working at all.
This due servers responding to "releaseStream" and "FCPublish" and sending unneeded "_result" answers, but librtmp using first "_result" as answer to "createStream".
This little patch fixes this bug, and now publishing works fine on all servers, that i try.
Thanks for the info. It seems the current code needed to behave the way it does because wowza just doesn't handle those requests properly. Your patch worries me because it will cause other results to be ignored if a player has other times where multiple valid requests are outstanding. Will look into this some more... -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/
Sergiy wrote:
Hi all
I'm writing application, that able to publish live rtmp stream.
I tested it on wowza/fms/crtmpd. First one works fine with current implementation, but fms works only from 3-5 try. And crmpd connection not working at all.
This due servers responding to "releaseStream" and "FCPublish" and sending unneeded "_result" answers, but librtmp using first "_result" as answer to "createStream".
This little patch fixes this bug, and now publishing works fine on all servers, that i try.
Please try SVN -r486 instead and tell us your results, thanks.
Howard Chu:
Please try SVN -r486 instead and tell us your results, thanks.
Now publishing works correctly on all 3 types of servers.
participants (2)
-
Howard Chu -
Sergiy