Vicken Simonian [EMAIL PROTECTED] added the comment:
sorry for the delay I wanted to do some extra testing on both Linux and Windows
and break down the code changes that made the difference the most. . .I just
tried you latest patch on Windows and it did not work at all for me. (I pulled
todays
Vicken Simonian [EMAIL PROTECTED] added the comment:
Works on Linux as well.
So the changes are:
1) Adding udp_set_url...
if (s-is_multicast !(h-flags URL_WRONLY))
udp_set_url(my_addr, hostname, s-local_port);
2) Moving the MC options (udp_join_multicast_group) to right before bind
Vicken Simonian [EMAIL PROTECTED] added the comment:
hold off on that, moving the multicast join options will not work on windows
like i originally said. . .i was reading that the multicast add membership
should come after a bind() call.
probably needs another/more windows users to verify
Vicken Simonian [EMAIL PROTECTED] added the comment:
Confirmed it! Both ffplay instance work perfectly now.
very well done!
__
FFmpeg issue tracker [EMAIL PROTECTED]
https://roundup.mplayerhq.hu/roundup/ffmpeg/issue481
__
Vicken Simonian [EMAIL PROTECTED] added the comment:
next, will you be checking it into the trunk?
__
FFmpeg issue tracker [EMAIL PROTECTED]
https://roundup.mplayerhq.hu/roundup/ffmpeg/issue481
__
Vicken Simonian [EMAIL PROTECTED] added the comment:
I built from the latest trunk, patched udp.c in libavformat, and re-ran on
windows and this is the output. It does not even play the first stream anymore.
It is much more difficult for me to debug on windows, than in Linux so I am
having
New submission from Vicken Simonian [EMAIL PROTECTED]:
In at least r12790, both the linux and mingw compile is broken now.
The 'make' result message in Linux
make: *** No rule to make target `libavdevice/libavdevice.so', needed by `all'.
Stop.
The 'make' result message in Mingw
make
Vicken Simonian [EMAIL PROTECTED] added the comment:
r12670
windows compile still has the issue with
rtsp.c:1097: error: 'EPROTONOSUPPORT' undeclared (first use in this function)
The replacement with AVERROR_NOTSUPP workaround works, is this going in the
trunk?
Im not seeing problems