* qatar/master:
http: Add the url_shutdown function for https, too
http: Simplify code by removing a local variable
http: Clear the old URLContext pointer when closed
tcp: Try enabling SO_REUSEADDR when listening
tcp: Check the return values from bind and accept
avisynth: Make sure the filename passed to avisynth is in the right code page
Conflicts:
libavformat/http.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Without this, the compiler in theory could attempt to keep
things in float registers across the emms which would cause
bad things to happen.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This fixes one (disabled) fate test.
Thanks-to: ubitux to pointing at the odd filter order and stereo/mono messup
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This fixes issues with opening http urls that have authentication
or redirects, introduced in commit e999b641.
Signed-off-by: Martin Storsjö <martin@martin.st>
ra144 uses floats so bitexactness cannot be guranteed
This should fix a long standing issue with icc
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
avisynth is a non-unicode application and cannot accept UTF-8
characters. Therefore, the input filename should be converted to
the correct code page that it expects.
Signed-off-by: Martin Storsjö <martin@martin.st>
* qatar/master:
fate: Fix fate-ac3-fixed-encode for pre-ssse3 x86 machines
http: Pass the proper return code of net IO operations
http: Add 'post_data', a new option which sets custom HTTP post data
lavfi: amix: check active input count before calling request_samples
vp8: move block coeff arithcoder on stack.
mp3/ac3 probe: search for PES headers to prevent probing MPEG-PS as MP3.
Conflicts:
libavformat/ac3dec.c
libavformat/mp3dec.c
tests/fate/ac3.mak
Merged-by: Michael Niedermayer <michaelni@gmx.at>
This fixes out of global array reads.
The alternative solutions of checking the index or modifying the VLC tables
to prevent the index going outside are each about 1-2 cpu cyclces slower
per coded 4x4 block.
The alternative of padding the global tables directly is more ugly and
moving them to the context should benefit cache locality.
Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
The old implementation, upon receiving a frame on the main
input, would request an overlay frame if necessary. This could
generate an infinite recursion, off-by-one frame mismatch and
other harmful effects, especially it the split filter is present
upgraph.
The new implementation uses the linear approach: it uses two
buffer queues for frames received out of turn and forwards
request_frame calls to the input where a frame is necessary.
Maybe someone can add a check in the second gif encoder (rgb24), because I'm not sure where it should be added.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
