connection (e.g. a seek). This fixes the theoretical case where a server
sends a file first using chunked encoding, and then using non-chunked
encoding.
Originally committed as revision 23665 to svn://svn.ffmpeg.org/ffmpeg/trunk
them from being sent using chunked encoding (I don't think this ever happened,
but either way it would be wrong).
Originally committed as revision 23664 to svn://svn.ffmpeg.org/ffmpeg/trunk
This removes some useless copying of handles, and simplifies error handling.
Patch by Josh Allmann, joshua dot allmann at gmail
Originally committed as revision 23648 to svn://svn.ffmpeg.org/ffmpeg/trunk
Since rtsp_hd isn't assigned to rt->rtsp_hd until after the setup phase,
the initialized URLContext could be leaked on failures.
Originally committed as revision 23643 to svn://svn.ffmpeg.org/ffmpeg/trunk
Earlier, the function only returned when the enough data to fill the
requested buffer was available. This lead to high latency when receiving
low-bandwidth streams.
Originally committed as revision 23642 to svn://svn.ffmpeg.org/ffmpeg/trunk
(extradata_size must be at least 5 bytes) the H.264 MP4 syntax check
in rtpenc.c
Originally committed as revision 23638 to svn://svn.ffmpeg.org/ffmpeg/trunk
When symbol versioning is enabled, moving symbols from one library to
another breaks binary compatibility. This adds wrappers with the old
version tag for the av_*packet functions recently moved to lavc.
Originally committed as revision 23611 to svn://svn.ffmpeg.org/ffmpeg/trunk
Since the parsing of Vorbis/Theora fmtp headers is handled by the
parse_sdp_a_line function pointer now, the buffer in sdp_parse_fmtp
doesn't need to be this large any longer.
Patch by Josh Allmann, joshua dot allmann at gmail
Originally committed as revision 23599 to svn://svn.ffmpeg.org/ffmpeg/trunk
Also use the correct value for unknown size (instead of a truncated
and sign extended value).
Originally committed as revision 23591 to svn://svn.ffmpeg.org/ffmpeg/trunk
This interrupt parsing and goes one ebml level up when encountering
a cluster start inside an unknown sized element which is not supposed to
contain a cluster.
fix issue1991
Originally committed as revision 23589 to svn://svn.ffmpeg.org/ffmpeg/trunk
Now that we save the currently parsed ebml element ID after an interruption,
we don't need to special case the parsing of the first cluster.
Originally committed as revision 23588 to svn://svn.ffmpeg.org/ffmpeg/trunk
This allows to interrupt parsing after reading an ID, and then properly
recover parsing.
Originally committed as revision 23587 to svn://svn.ffmpeg.org/ffmpeg/trunk
different magic in the header (mainly targeted to Sony's .oma/.aa3
format).
Patch by Michael Karcher, ffmpeg A mkarcher dialup fu-berlin de
Originally committed as revision 23583 to svn://svn.ffmpeg.org/ffmpeg/trunk
BeOS support has been broken for many years and the "maintainer" of the port
has not reacted to countless requests to get the port fixed.
approved by Mans
Originally committed as revision 23562 to svn://svn.ffmpeg.org/ffmpeg/trunk
This makes url_fsize return correct values for delay opened connections
that have not yet been initialized.
This fixes using the image2 demuxer with http sources.
Originally committed as revision 23546 to svn://svn.ffmpeg.org/ffmpeg/trunk
If http_connect fails, we've already stored the new connection handle in s->hd,
so clear it so http_close won't double-free it.
10l to me for not spotting it during review
Originally committed as revision 23529 to svn://svn.ffmpeg.org/ffmpeg/trunk
Initialize the buffer av_strlcatf writes into.
Add the dash at the end of the Range header.
10l to me for not spotting it earlier
Originally committed as revision 23528 to svn://svn.ffmpeg.org/ffmpeg/trunk
The LFG is only used in one place, so there is no need to keep it in
the context.
Originally committed as revision 23527 to svn://svn.ffmpeg.org/ffmpeg/trunk
