MS RTSP spec states that the UDP port pair used in subsequent SETUP
requests for various streams must be identical (since there will actually
be only 1 stream of muxed asf packets). Following traditional specs and
using different port pairs in the SETUPs for separate streams will result
in all but the first one failing and only one stream being streamed.
So, in appropriate circumstances, retry UDP SETUP using previously used
port pair. Fixes#552650.
When we are dealing with connected sockets shared between a udpsrc and a udpsink
we might receive ICMP connection refused error messages in udpsrc that will
cause it to go into a bursty loop because the poll returns right away without a
message to read.
Instead of looping, read the error message from the error queue in udpsrc.
Fixes#567857.
sort_by_frame_size is declared static and only used inside
an ifdef, so use the same ifdef to define the function. Fixes#572185
Signed-off-by: David Schleef <ds@schleef.org>
Reading integers from random memory addresses will result
in SIGBUS on some architectures if the memory address
is not correctly aligned. This can happen at two
places in avidemux so we should use GST_READ_UINT32_LE
and friends here. Fixes bug #572256.
g_atomic_int_(get|set) only work on ints and the flags are
an enum (which on most architectures is stored as an int).
Also the way the flags were accessed atomically would still
leave a possible race condition and we don't do it in any
other mixer track implementation, let alone at any other
place where an integer could be changed from different
threads. Removing the g_atomic_int_(get|set) will only
introduce a new race condition on architectures where
integers could be half-written while reading them
which shouldn't be the case for any modern architecture
and if we really care about this we need to use
g_atomic_int_(get|set) at many other places too.
Apart from that g_atomic_int_(set|get) will result in
aliasing warnings if their argument is explicitely
casted to an int *. Fixes bug #571153.
rather than PA thread.
pa_threaded_mainloop_lock() (a.o.) and by extension get_property should
not be done from a PA thread, but the latter may occur as a result of a
property change notification. Fixes#571204 (though current situation
not ideal, e.g. post message rather than signal).
stps atoms contain "partial sync" information, which means that it's
a sync point where pts != dts. This is needed to properly handle
MPEG2, H.264, Dirac, etc., in quicktime.
Remove non-embedded mode. Embed mode becomes default and only mode.
embed property is retained for binary compatibility.
Added autorelease pools around all objc functions that might be called
from a non-main thread.
Ensure we create the ringbuffer segment size as a multiple of the
bytes per sample (fixes 6-channel output).
Reset the segoffset when acquiring the ringbuffer, so we don't retain
a bogus offset when caps change.
