September 2026 - ffmpeg-cvslog

[ffmpeg] branch master updated (a7b0af0071 -> 5a54fcf75e)
by ffmpeg-git＠ffbox0-bg.ffmpeg.org 30 Sep '26

30 Sep '26

This is an automated email from the git hooks/post-receive script. Git pushed a change to branch master in repository ffmpeg. from a7b0af0071 avformat/movenc: reject MPEG-H 3D Audio streams with no extradata new 47855fce76 arm, aarch64: Don't use the ccnt / pmccntr_el0 registers for READ_TIME new 5a54fcf75e arm: Remove the arm/timer.h header The 2 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "add" were already present in the repository and have only been added to this reference. Summary of changes: libavutil/aarch64/timer.h | 7 ------- libavutil/arm/timer.h | 46 ---------------------------------------------- libavutil/timer.h | 2 -- 3 files changed, 55 deletions(-) delete mode 100644 libavutil/arm/timer.h -- To stop receiving notification emails like this one, please contact ffmpeg-git(a)ffmpeg.org.

1 0

[ffmpeg] branch master updated: avformat/movenc: reject MPEG-H 3D Audio streams with no extradata
by James Almer 30 Sep '26

30 Sep '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new a7b0af0071 avformat/movenc: reject MPEG-H 3D Audio streams with no extradata a7b0af0071 is described below commit a7b0af0071c0d95f657719372a84bea7c6fd84f3 Author: James Almer <jamrial(a)gmail.com> AuthorDate: Wed Sep 30 13:47:06 2026 -0300 Commit: James Almer <jamrial(a)gmail.com> CommitDate: Wed Sep 30 13:47:06 2026 -0300 avformat/movenc: reject MPEG-H 3D Audio streams with no extradata It's not clear if mpegh3daConfigLength is allowed to be 0, so reject it outright. Signed-off-by: James Almer <jamrial(a)gmail.com> --- libavformat/movenc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/libavformat/movenc.c b/libavformat/movenc.c index 6c2a009ac3..6492f4ffb3 100644 --- a/libavformat/movenc.c +++ b/libavformat/movenc.c @@ -975,7 +975,8 @@ static int mov_write_mhac_tag(AVFormatContext *s, AVIOContext *pb, MOVTrack *tra int64_t pos = avio_tell(pb); int layout = 0; - if (track->extradata_size[track->last_stsd_index] > UINT16_MAX) { + if (!track->extradata_size[track->last_stsd_index] || + track->extradata_size[track->last_stsd_index] > UINT16_MAX) { av_log(s, AV_LOG_ERROR, "Invalid extradata size %d for MPEGH-H stream.\n", track->extradata_size[track->last_stsd_index]); -- To stop receiving notification emails like this one, please contact ffmpeg-git(a)ffmpeg.org.

1 0

[ffmpeg] branch master updated: avcodec/libmpeghdec: don't use extradata for streams with in-band config configuration
by James Almer 30 Sep '26

30 Sep '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new e4fa18c220 avcodec/libmpeghdec: don't use extradata for streams with in-band config configuration e4fa18c220 is described below commit e4fa18c2206bb626ab5ad76c8fb055cf8fa26ba1 Author: James Almer <jamrial(a)gmail.com> AuthorDate: Wed Sep 30 13:39:50 2026 -0300 Commit: James Almer <jamrial(a)gmail.com> CommitDate: Wed Sep 30 13:40:35 2026 -0300 avcodec/libmpeghdec: don't use extradata for streams with in-band config configuration The decoder seems to accept it but then will not output any frame at all. Signed-off-by: James Almer <jamrial(a)gmail.com> --- libavcodec/libmpeghdec.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/libavcodec/libmpeghdec.c b/libavcodec/libmpeghdec.c index 49e7a5b1ec..29627b8f52 100644 --- a/libavcodec/libmpeghdec.c +++ b/libavcodec/libmpeghdec.c @@ -114,7 +114,8 @@ static av_cold int mpegh3dadec_init(AVCodecContext *avctx) return AVERROR_EXTERNAL; } - if (avctx->extradata_size) { + if (avctx->codec_tag != MKTAG('m', 'h', 'm', '1') && + avctx->codec_tag != MKTAG('m', 'h', 'm', '2') && avctx->extradata_size) { if (mpeghdecoder_setMhaConfig(s->decoder, avctx->extradata, avctx->extradata_size)) { av_log(avctx, AV_LOG_ERROR, "Unable to set MHA configuration\n"); -- To stop receiving notification emails like this one, please contact ffmpeg-git(a)ffmpeg.org.

1 0

[ffmpeg] branch master updated (3ed0dd076e -> a70587cc9f)
by ffmpeg-git＠ffbox0-bg.ffmpeg.org 30 Sep '26

30 Sep '26

This is an automated email from the git hooks/post-receive script. Git pushed a change to branch master in repository ffmpeg. from 3ed0dd076e avutil/tests/softfloat: only run the timers with -t new 11de0fc8fd avformat/tls_schannel: support sending client certificates new a70587cc9f avformat/tls_schannel: don't automatically send client cert from users store The 2 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "add" were already present in the repository and have only been added to this reference. Summary of changes: libavformat/tls_schannel.c | 38 ++++++++++++++++++++++++-------------- 1 file changed, 24 insertions(+), 14 deletions(-) -- To stop receiving notification emails like this one, please contact ffmpeg-git(a)ffmpeg.org.

2 2

[ffmpeg] branch master updated (6ae491a26c -> 3ed0dd076e)
by ffmpeg-git＠ffbox0-bg.ffmpeg.org 30 Sep '26

30 Sep '26

This is an automated email from the git hooks/post-receive script. Git pushed a change to branch master in repository ffmpeg. from 6ae491a26c avformat/jpegmpfenc: serialize GContainer XMP APP1 if needed new 8b94094253 avutil/rational: fix the sign handling in av_q2intfloat() new 0bd69b4ba5 avutil/tests/rational: fix the av_q2intfloat() checks new a17a624069 swresample: skip the rematrix stage when there are no samples new 3ed0dd076e avutil/tests/softfloat: only run the timers with -t The 4 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "add" were already present in the repository and have only been added to this reference. Summary of changes: libavutil/rational.c | 34 ++++++++++++++--------------- libavutil/tests/rational.c | 37 ++++++++++++++++++++++++------- libavutil/tests/softfloat.c | 53 ++++++++++++++++++++++++++++++--------------- libswresample/swresample.c | 4 ++-- 4 files changed, 84 insertions(+), 44 deletions(-) -- To stop receiving notification emails like this one, please contact ffmpeg-git(a)ffmpeg.org.

2 4

[ffmpeg] branch master updated (e0c94b2d1c -> 6ae491a26c)
by ffmpeg-git＠ffbox0-bg.ffmpeg.org 30 Sep '26

30 Sep '26

This is an automated email from the git hooks/post-receive script. Git pushed a change to branch master in repository ffmpeg. from e0c94b2d1c avcodec/nvenc: avoid invalid surface destruction after init error new b917a0fe1e avutil/gain_map: add AVGainMapParams and side data type new caa5824750 avformat: add AV_STREAM_GROUP_PARAMS_GAIN_MAP new f14e7eaf7c avfilter: add AV_FRAME_DATA_GAIN_MAP_PARAMS support new 931db4d377 avfilter/vf_gainmap: add filter to compute a gain map new 782d60e3a5 avcodec/gain_map: add ff_gain_map_params_(to/from)_iso21496 new 1b4ca126d8 avcodec/mjpegenc: add support for writing ISO 21496-1 gain map metadata new d3d376d9e6 avformat/jpegmpfenc: add JPEG-MPF muxer new 0de89d621f tests/fate/image: add jpeg-mpf-gainmap test new 7db3357abf avcodec/gain_map: add support for serializing XMP metadata new ebbc9b6600 avcodec/mjpegenc: add -xmp_gain_map option new 6ae491a26c avformat/jpegmpfenc: serialize GContainer XMP APP1 if needed The 11 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "add" were already present in the repository and have only been added to this reference. Summary of changes: doc/APIchanges | 10 + doc/ffmpeg.texi | 13 + doc/filters.texi | 75 ++++ doc/muxers.texi | 21 + fftools/ffmpeg_mux_init.c | 3 + libavcodec/Makefile | 6 +- libavcodec/gain_map.c | 241 +++++++++++ libavcodec/gain_map.h | 89 ++++ libavcodec/mjpegenc.c | 3 +- libavcodec/mjpegenc.h | 3 + libavcodec/mjpegenc_common.c | 83 +++- libavcodec/mjpegenc_common.h | 2 + libavcodec/mpegvideo_enc.c | 3 + libavfilter/Makefile | 1 + libavfilter/allfilters.c | 1 + libavfilter/f_sidedata.c | 1 + libavfilter/vf_gainmap.c | 774 +++++++++++++++++++++++++++++++++++ libavfilter/vf_showinfo.c | 30 ++ libavformat/Makefile | 1 + libavformat/allformats.c | 1 + libavformat/avformat.c | 2 + libavformat/avformat.h | 6 +- libavformat/jpegmpfenc.c | 635 ++++++++++++++++++++++++++++ libavformat/options.c | 3 + libavformat/version.h | 2 +- libavutil/Makefile | 2 + libavutil/frame.h | 7 + libavutil/gain_map.c | 110 +++++ libavutil/gain_map.h | 115 ++++++ libavutil/side_data.c | 1 + libavutil/version.h | 2 +- tests/fate/filter-video.mak | 8 + tests/fate/image.mak | 14 +- tests/filtergraphs/jpeg_mpf_gainmap | 4 + tests/ref/fate/filter-gainmap | 6 + tests/ref/fate/filter-gainmap-fixed | 6 + tests/ref/fate/filter-gainmap-luma | 6 + tests/ref/fate/filter-gainmap-maxrgb | 6 + tests/ref/fate/jpeg-mpf-gainmap | 1 + 39 files changed, 2286 insertions(+), 11 deletions(-) create mode 100644 libavcodec/gain_map.c create mode 100644 libavcodec/gain_map.h create mode 100644 libavfilter/vf_gainmap.c create mode 100644 libavformat/jpegmpfenc.c create mode 100644 libavutil/gain_map.c create mode 100644 libavutil/gain_map.h create mode 100644 tests/filtergraphs/jpeg_mpf_gainmap create mode 100644 tests/ref/fate/filter-gainmap create mode 100644 tests/ref/fate/filter-gainmap-fixed create mode 100644 tests/ref/fate/filter-gainmap-luma create mode 100644 tests/ref/fate/filter-gainmap-maxrgb create mode 100644 tests/ref/fate/jpeg-mpf-gainmap -- To stop receiving notification emails like this one, please contact ffmpeg-git(a)ffmpeg.org.

2 11

[ffmpeg] branch master updated: swscale/rational64: use 64-bit math whenever possible
by Niklas Haas 30 Sep '26

30 Sep '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new 453386217a swscale/rational64: use 64-bit math whenever possible 453386217a is described below commit 453386217a3140f9cfc3ec923390104d5e814151 Author: Niklas Haas <git(a)haasn.dev> AuthorDate: Tue Sep 29 18:53:27 2026 +0200 Commit: Niklas Haas <ffmpeg(a)haasn.dev> CommitDate: Wed Sep 30 11:51:24 2026 +0000 swscale/rational64: use 64-bit math whenever possible The runtime of av_q64() ops are dominated by reduce128()'s 128-bit divisions. We can vastly improve upon this in the common case of values that already fit into 64 bits, by falling back to a 64-bit reduction fast path. On 64-bit: time make fate-sws-ops-list # 3.57s -> 2.38s On 32-bit: time make fate-sws-ops-list # 49.13s -> 6.79s Signed-off-by: Niklas Haas <git(a)haasn.dev> --- libswscale/rational64.c | 102 ++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 98 insertions(+), 4 deletions(-) diff --git a/libswscale/rational64.c b/libswscale/rational64.c index 6577a66613..d1658e2792 100644 --- a/libswscale/rational64.c +++ b/libswscale/rational64.c @@ -28,9 +28,36 @@ #include <limits.h> +#include "libavutil/attributes.h" +#include "libavutil/common.h" #include "libavutil/int128.h" +#include "libavutil/intmath.h" #include "rational64.h" +static av_always_inline uint64_t high64(av_int128 a) +{ + return av_from128u(av_shr128(a, 64)); +} + +/* unsigned version of av_gcd(); see libavutil/mathematics.c */ +static av_always_inline uint64_t gcd_u64(uint64_t a, uint64_t b) +{ + if (!a) + return b; + if (!b) + return a; + + const int k = ff_ctzll(a | b); + a >>= ff_ctzll(a); + do { + b >>= ff_ctzll(b); + if (a > b) + FFSWAP(uint64_t, a, b); + b -= a; + } while (b); + return a << k; +} + static av_int128 gcd128(av_int128 a, av_int128 b) { while (av_test128(b)) { @@ -52,6 +79,27 @@ static AVRational64 reduce64(av_int128 num, av_int128 den) if (den_sign) den = av_sub128(zero, den); + if (!high64(num) && !high64(den)) { + /* Fast path: both values already fit into 64 bits */ + uint64_t n = av_from128u(num), d = av_from128u(den); + const uint64_t gcd = gcd_u64(n, d); + if (gcd) { + n /= gcd; + d /= gcd; + } + + if (!((n | d) >> 63)) { + AVRational64 res = { n, d }; + if (num_sign ^ den_sign) + res.num = -res.num; + return res; + } + + /* Fall back to 128-bit arithmetic for values > INT64_MAX */ + num = av_to128u(n); + den = av_to128u(d); + } + const av_int128 gcd = gcd128(num, den); if (av_test128(gcd)) { num = av_div128(num, gcd); @@ -105,11 +153,47 @@ done:; return res; } +static av_always_inline int fits31(int64_t x) +{ + return x >= -INT32_MAX && x <= INT32_MAX; +} + +static av_always_inline int fits31_q(AVRational64 b, AVRational64 c) +{ + return fits31(b.num) && fits31(b.den) && fits31(c.num) && fits31(c.den); +} + +/* Simplified version of reduce64() for small values */ +static AVRational64 reduce64_fast(int64_t num, int64_t den) +{ + const int sign = (num < 0) ^ (den < 0); + uint64_t n = num < 0 ? -(uint64_t) num : num; + uint64_t d = den < 0 ? -(uint64_t) den : den; + const uint64_t gcd = gcd_u64(n, d); + if (gcd) { + n /= gcd; + d /= gcd; + } + + AVRational64 res = { n, d }; + if (sign) + res.num = -res.num; + return res; +} + + int ff_cmp_q64(AVRational64 a, AVRational64 b) { - const av_int128 p = av_mul128(av_to128i(a.num), av_to128i(b.den)); - const av_int128 q = av_mul128(av_to128i(b.num), av_to128i(a.den)); - const int test = av_cmp128(p, q); + int test; + if (fits31_q(a, b)) { + const int64_t p = a.num * b.den; + const int64_t q = b.num * a.den; + test = (p > q) - (p < q); + } else { + const av_int128 p = av_mul128(av_to128i(a.num), av_to128i(b.den)); + const av_int128 q = av_mul128(av_to128i(b.num), av_to128i(a.den)); + test = av_cmp128(p, q); + } if (test) return (a.den < 0) ^ (b.den < 0) ? -test : test; @@ -123,6 +207,9 @@ int ff_cmp_q64(AVRational64 a, AVRational64 b) AVRational64 ff_mul_q64(AVRational64 b, AVRational64 c) { + if (fits31_q(b, c)) + return reduce64_fast(b.num * c.num, b.den * c.den); + return reduce64(av_mul128(av_to128i(b.num), av_to128i(c.num)), av_mul128(av_to128i(b.den), av_to128i(c.den))); } @@ -132,7 +219,11 @@ AVRational64 ff_div_q64(AVRational64 b, AVRational64 c) return ff_mul_q64(b, ff_inv_q64(c)); } -AVRational64 ff_add_q64(AVRational64 b, AVRational64 c) { +AVRational64 ff_add_q64(AVRational64 b, AVRational64 c) +{ + if (fits31_q(b, c)) + return reduce64_fast(b.num * c.den + c.num * b.den, b.den * c.den); + return reduce64(av_add128(av_mul128(av_to128i(b.num), av_to128i(c.den)), av_mul128(av_to128i(c.num), av_to128i(b.den))), av_mul128(av_to128i(b.den), av_to128i(c.den))); @@ -140,6 +231,9 @@ AVRational64 ff_add_q64(AVRational64 b, AVRational64 c) { AVRational64 ff_sub_q64(AVRational64 b, AVRational64 c) { + if (fits31_q(b, c)) + return reduce64_fast(b.num * c.den - c.num * b.den, b.den * c.den); + return reduce64(av_sub128(av_mul128(av_to128i(b.num), av_to128i(c.den)), av_mul128(av_to128i(c.num), av_to128i(b.den))), av_mul128(av_to128i(b.den), av_to128i(c.den))); -- To stop receiving notification emails like this one, please contact ffmpeg-git(a)ffmpeg.org.

1 0

[ffmpeg] branch master updated: avformat/libcurl: properly lock CURLSH
by Niklas Haas 30 Sep '26

30 Sep '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new a9f159ad05 avformat/libcurl: properly lock CURLSH a9f159ad05 is described below commit a9f159ad0513e1651af7c71fb9eb4d5385e5855f Author: Niklas Haas <git(a)haasn.dev> AuthorDate: Tue Sep 29 20:46:32 2026 +0200 Commit: Niklas Haas <ffmpeg(a)haasn.dev> CommitDate: Wed Sep 30 09:36:03 2026 +0000 avformat/libcurl: properly lock CURLSH This was previously rejected from my PR because "the CURLSH is only active once the CURL * is added to the CURLM *" (paraphrased), however this is not true - as we also curl_easy_setopt(CURLOPT_COOKIELIST) during setup_curl(), called from the main thread, which references (and writes to) the CURLSH, relying only on the lockfunc to synchronize access. Fixes: 6c338a659551c8f395954698f8e0c8607c3e3f4c Fixes: https://github.com/nxtedition/FFmpeg/issues/77 Signed-off-by: Niklas Haas <git(a)haasn.dev> Sponsored-by: nxtedition AB --- libavformat/libcurl.c | 19 ++++++++++++++++++- 1 file changed, 18 insertions(+), 1 deletion(-) diff --git a/libavformat/libcurl.c b/libavformat/libcurl.c index 3e53eadb3c..983b207722 100644 --- a/libavformat/libcurl.c +++ b/libavformat/libcurl.c @@ -91,7 +91,7 @@ typedef struct CurlLoop { CURLM *multi; CURLSH *share; /* shared cookies/HSTS */ - pthread_mutex_t mutex; /* guards the command queue, exit and cmd->done */ + pthread_mutex_t mutex; /* guards the command queue, exit, share and cmd->done */ pthread_cond_t cond; /* signaled when a sync command completes */ CurlCmd *cmd_head, *cmd_tail; int exit; @@ -920,6 +920,20 @@ static int curl_dispatch(CurlLoop *loop, enum cmd_kind kind, CurlContext *c, return 0; } +static void share_lock_callback(CURL *handle, curl_lock_data data, + curl_lock_access access, void *userdata) +{ + CurlLoop *loop = userdata; + pthread_mutex_lock(&loop->mutex); +} + +static void share_unlock_callback(CURL *handle, curl_lock_data data, + void *userdata) +{ + CurlLoop *loop = userdata; + pthread_mutex_unlock(&loop->mutex); +} + static CurlLoop *curl_loop_create(AVFormatContext *avfc) { CurlLoop *loop = av_mallocz(sizeof(*loop)); @@ -945,6 +959,9 @@ static CurlLoop *curl_loop_create(AVFormatContext *avfc) loop->share = curl_share_init(); if (!loop->share) goto fail3; + curl_share_setopt(loop->share, CURLSHOPT_USERDATA, loop); + curl_share_setopt(loop->share, CURLSHOPT_LOCKFUNC, share_lock_callback); + curl_share_setopt(loop->share, CURLSHOPT_UNLOCKFUNC, share_unlock_callback); curl_share_setopt(loop->share, CURLSHOPT_SHARE, CURL_LOCK_DATA_COOKIE); curl_share_setopt(loop->share, CURLSHOPT_SHARE, CURL_LOCK_DATA_HSTS); -- To stop receiving notification emails like this one, please contact ffmpeg-git(a)ffmpeg.org.

1 0

[ffmpeg] branch master updated (18ee27e67b -> ff2059afd9)
by ffmpeg-git＠ffbox0-bg.ffmpeg.org 30 Sep '26

30 Sep '26

This is an automated email from the git hooks/post-receive script. Git pushed a change to branch master in repository ffmpeg. from 18ee27e67b tests/fate/libswresample: run the swresample self test new f88f1041a6 avformat/hls: advance the init section read pointer new 9d49ed18da tests/fate/hlsenc: add a test for a large fMP4 init section new 3c667d0947 avformat/dashdec: copy the init section from the current read offset new ff2059afd9 tests/fate/dashenc: add a test for a large init segment The 4 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "add" were already present in the repository and have only been added to this reference. Summary of changes: libavformat/dashdec.c | 2 +- libavformat/hls.c | 2 +- tests/fate/dashenc.mak | 19 +++++++++++++++++++ tests/fate/hlsenc.mak | 13 +++++++++++++ tests/ref/fate/dash-big-init | 33 +++++++++++++++++++++++++++++++++ tests/ref/fate/hls-fmp4-big-init | 33 +++++++++++++++++++++++++++++++++ 6 files changed, 100 insertions(+), 2 deletions(-) create mode 100644 tests/ref/fate/dash-big-init create mode 100644 tests/ref/fate/hls-fmp4-big-init -- To stop receiving notification emails like this one, please contact ffmpeg-git(a)ffmpeg.org.

3 4

[ffmpeg] branch master updated (09f879b074 -> 18ee27e67b)
by ffmpeg-git＠ffbox0-bg.ffmpeg.org 29 Sep '26

29 Sep '26

This is an automated email from the git hooks/post-receive script. Git pushed a change to branch master in repository ffmpeg. from 09f879b074 swresample: keep integer samples if channels are only copied new b5763f40f3 avutil/tests/rational: return failure on any reported mismatch new d79f7a3530 swscale/tests/rational64: return failure on any reported mismatch new 01385bff33 avutil/tests/log: initialize the console state before forcing use_color new 25976f1908 avutil/tests/pca: fix unsigned wraparound in the generated samples new 8a2d5b727a avutil/tests/pca: free the context new 696d754bba avutil/tests/error: print only the FFmpeg-defined codes new ba8604f8be avutil/tests/utf8: decode built-in sequences when no file is given new 84f2449940 avutil/tests/lls: make the printed values platform independent new e688792d2b tests/fate/libavutil: run the test programs that were only built new 4e9e195f3f tests/fate/libavcodec: run the snow DWT test new 09367fd57f tests/fate/libavcodec: add reconstructed frame tests for snow new 7058bcf78a avfilter/tests/drawutils: print a stable word for unsupported formats new 07565e4666 tests/fate/filter-video: run the nlmeans integral test new 9305f9f2ff swscale/tests/colorspace: free the buffers new 61a9315a94 tests/fate/libswscale: run the rgb2rgb bounds and rational64 tests new 5191917736 swresample/tests/swresample: free the contexts new 739132b015 swresample/tests/swresample: fail when a context cannot be initialized new 755b6bdfd5 swresample/tests/swresample: align the sample buffers new 18ee27e67b tests/fate/libswresample: run the swresample self test The 19 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "add" were already present in the repository and have only been added to this reference. Summary of changes: libavfilter/tests/drawutils.c | 2 +- libavutil/tests/error.c | 42 +++-- libavutil/tests/lls.c | 14 +- libavutil/tests/log.c | 1 + libavutil/tests/pca.c | 5 +- libavutil/tests/rational.c | 14 +- libavutil/tests/utf8.c | 46 ++++-- libswresample/tests/swresample.c | 31 +++- libswscale/tests/colorspace.c | 7 +- libswscale/tests/rational64.c | 13 +- tests/fate/filter-video.mak | 11 +- tests/fate/libavcodec.mak | 12 ++ tests/fate/libavutil.mak | 30 ++++ tests/fate/libswresample.mak | 9 ++ tests/fate/libswscale.mak | 10 ++ tests/ref/fate/error | 29 ++++ tests/ref/fate/filter-drawutils | 342 +++++++++++++++++++-------------------- tests/ref/fate/lls | 300 ++++++++++++++++++++++++++++++++++ tests/ref/fate/pca | 17 ++ tests/ref/fate/snow-dwt | 75 +++++++++ tests/ref/fate/softfloat | 12 ++ tests/ref/fate/utf8 | 31 ++++ 22 files changed, 834 insertions(+), 219 deletions(-) create mode 100644 tests/ref/fate/error create mode 100644 tests/ref/fate/lls create mode 100644 tests/ref/fate/pca create mode 100644 tests/ref/fate/snow-dwt create mode 100644 tests/ref/fate/softfloat create mode 100644 tests/ref/fate/utf8 -- To stop receiving notification emails like this one, please contact ffmpeg-git(a)ffmpeg.org.

2 19