> Could this be handled in the select on a writeset, and when the socket can
> write it writes the header followed by the data. (I assume the data must
> come in one seq write?) This way the event loop can still process up to the
> point of the send?
No, there's no way that the code could be
Thanks for the report. I have included these fixes in the latest version of
the code (2013.12.04).
Ross Finlayson
Live Networks, Inc.
http://www.live555.com/
___
live-devel mailing list
live-devel@lists.live555.com
http://lists.live555.com/mailman/lis
liveMedia\RTSPClient.cpp 916
while (line[paramIndex] != '\0' && (line[paramIndex] == ' ' || line[paramIndex]
== '\t')) ++paramIndex;
if (&line[paramIndex] == '\0') return False;
I think should be:
if (line[paramIndex] == '\0') return False;
-
liveMedia\MediaSessi
Could this be handled in the select on a writeset, and when the socket can
write it writes the header followed by the data. (I assume the data must come
in one seq write?) This way the event loop can still process up to the point
of the send?
Thanks,
Craig
On 12/04/2013 12:52 AM, Ross Finlay
In your previous message, you mentioned the possibility - in the
"RTPInterface::sendDataOverTCP()" function - of the "send()" call succeeding in
delivering partial (but not complete) data. I have had no other reports of
this ever happening, but - on looking at the code - I realized that the cod
(First, please remember to trim your quoted text when replying)
The existing code is correct; see
http://lists.live555.com/pipermail/live-devel/2013-June/017137.html
(when someone asked about this same code back in June.)
Note, in particular:
"Your write operation is blocking beca
