November 2018 - FFmpeg-trac - lists.ffmpeg.org

Re: [FFmpeg-trac] #3356(undetermined:open): feature request: Segment HLS streams on SCTE 35 markers
by FFmpeg 06 Nov '18

06 Nov '18

#3356: feature request: Segment HLS streams on SCTE 35 markers -------------------------------------+------------------------------------- Reporter: adamscybot | Owner: Type: enhancement | Status: open Priority: wish | Component: Version: git-master | undetermined Keywords: hls bounty | Resolution: Blocking: | Blocked By: Analyzed by developer: 0 | Reproduced by developer: 0 -------------------------------------+------------------------------------- Comment (by vmm2017): Replying to [comment:29 TheTroll]: > Hi, > > any progress on this subject ? > > Thanks Any progress on this ? Thanks -- Ticket URL: <https://trac.ffmpeg.org/ticket/3356#comment:30> FFmpeg <https://ffmpeg.org> FFmpeg issue tracker

1 0

#7531(trac:new): registration is kind of broken
by FFmpeg 05 Nov '18

05 Nov '18

#7531: registration is kind of broken -------------------------------------+------------------------------------- Reporter: | Type: defect BlackLotus89 | Priority: normal Status: new | Version: Component: trac | unspecified Keywords: | Blocked By: Blocking: | Reproduced by developer: 0 Analyzed by developer: 0 | -------------------------------------+------------------------------------- Summary of the bug: I tried registering at the trac a few times and always got a 76.68% probability that I was spam. After solving the google captcha (which I had to manually allow since I'm using umatrix) I was rejected. So were my following attempts. How to reproduce: Try to register with a password like 2epKyHUKaaFa9zEa (auto-generated) use a gmail address like yourgmailaddress+ffmpegtrac(a)gmail.com and enter your name as optional information. After I added a 89 to the username, changed my password to a 8 character password and didn't provide any additional information it worked. I'm not sure which of that measures worked, but it was a frustrating experience. Creating an account shouldn't be so hard. -- Ticket URL: <https://trac.ffmpeg.org/ticket/7531> FFmpeg <https://ffmpeg.org> FFmpeg issue tracker

1 2

#7481(undetermined:new): ffmpeg/ffprobe do not work with an FTP server
by FFmpeg 05 Nov '18

05 Nov '18

#7481: ffmpeg/ffprobe do not work with an FTP server -------------------------------------+------------------------------------- Reporter: boris- | Type: defect petrov | Priority: important Status: new | Version: Component: | unspecified undetermined | Blocked By: Keywords: | Reproduced by developer: 0 Blocking: | Analyzed by developer: 0 | -------------------------------------+------------------------------------- Running ffmpeg/ffprobe on this FTP server causes them both to block for a minute or two and then output an "Input/output error": {{{ % ffmpeg -i ftp://ffmpeg:ffmpeg2018@ftp.pbteu.com/ExportHD.mpg }}} Happens both on ffmpeg version 3.4.2 and on 4.0.2. The FTP server is FileZilla running on a Windows 7. -- Ticket URL: <https://trac.ffmpeg.org/ticket/7481> FFmpeg <https://ffmpeg.org> FFmpeg issue tracker

1 10

#5615(avformat:new): avformat_open_input on dead UDP source link hangs after interrupt_callback signaled exit
by FFmpeg 04 Nov '18

04 Nov '18

#5615: avformat_open_input on dead UDP source link hangs after interrupt_callback signaled exit -------------------------------------+------------------------------------- Reporter: Marlon | Type: defect Status: new | Priority: normal Component: avformat | Version: git- Keywords: udp open | master callback hang | Blocked By: Blocking: | Reproduced by developer: 0 Analyzed by developer: 0 | -------------------------------------+------------------------------------- Udp source is not active. After returning 1 from interrupt_callback (exit), avformat_open_input does not return. Using Win32 dll from Zeranoe on W10 64bit, Microsoft compiler 2015-05-22 Working https://ffmpeg.zeranoe.com/builds/win32/shared/ffmpeg-20160522-git- 566be4f-win32-shared.7z 2015-05-25 Not working https://ffmpeg.zeranoe.com/builds/win32/shared/ffmpeg-20160525-git- 9591ca7-win32-shared.7z How to reproduce: GetTickCount used can be any time function returning miliseconds. {{{ unsigned long TimePrev=0; //callback static int cb_open(void *ctx) { //timeout after 3 seconds if(GetTickCount() - TimePrev > 3000) { printf("---interrupt_cb: Exit\n"); return 1; } return 0; } void main() { avformat_network_init(); AVFormatContext *pFmtCtx = avformat_alloc_context(); pFmtCtx->interrupt_callback.callback = cb_open; pFmtCtx->interrupt_callback.opaque = 0; const char *url="udp://127.0.0.1:8000"; printf("avformat_open_input %s\n", url); TimePrev=GetTickCount(); int nRet=avformat_open_input(&pFmtCtx, url, NULL, NULL); //here never called }}} -- Ticket URL: <https://trac.ffmpeg.org/ticket/5615> FFmpeg <https://ffmpeg.org> FFmpeg issue tracker

1 14

#7521(build system:new): FFmpeg failed to build tests/api/api-h264-slice-test.c with error C2065 with MSVC
by FFmpeg 03 Nov '18

03 Nov '18

#7521: FFmpeg failed to build tests/api/api-h264-slice-test.c with error C2065 with MSVC -------------------------------------+------------------------------------- Reporter: KarenHuang | Type: defect Status: new | Priority: normal Component: build | Version: git- system | master Keywords: msvc | Blocked By: Blocking: | Reproduced by developer: 0 Analyzed by developer: 0 | -------------------------------------+------------------------------------- Summary of the bug: FFmpeg failed to build tests/api/api-h264-slice-test.c(183): error C2065: 'ssize_t': undeclared identifier when build with VS2017 on Windows. This issue can be reproduced from master revision fb7925b. I think 'ssize_t' type maybe undeclared, if we change 'ssize_t' to 'size_t' type and build pass. Could you please help take a look at this? Thank you! Steps to reproduce: 1. git clone https://github.com/FFmpeg/FFmpeg.git D:\FFmpeg\src 2. Open a VS 2017 x86 prompt and browse to D:\FFmpeg 3. set FFMPEG_TARGET=i386 4. set FFMPEG_SRC_PATH=D:\FFmpeg/src 5. set FFMPEG_FATE_PATH=../fate-suite/ 6. set PATH=D:\FFmpeg\tools;%PATH% 7. set MSYS2_PATH_TYPE=inherit 8. .\tools\msys32\usr\bin\sh.exe --login -i -c "/D/ffmpeg/tools/build.sh" 9. .\tools\msys32\usr\bin\sh.exe --login -i -c "/D/ffmpeg/tools/test.sh" Actual result: tests/api/api-h264-slice-test.c(183): error C2065: 'ssize_t': undeclared identifier tests/api/api-h264-slice-test.c(183): error C2146: syntax error: missing ';' before identifier 'ret' tests/api/api-h264-slice-test.c(183): error C2065: 'ret': undeclared identifier tests/api/api-h264-slice-test.c(184): error C2065: 'ret': undeclared identifier tests/api/api-h264-slice-test.c(187): error C2065: 'ret': undeclared identifier tests/api/api-h264-slice-test.c(190): error C2065: 'ret': undeclared identifier tests/api/api-h264-slice-test.c(191): error C2065: 'ret': undeclared identifier tests/api/api-h264-slice-test.c(195): error C2065: 'ret': undeclared identifier make: *** [ffbuild/common.mak:60: tests/api/api-h264-slice-test.o] Error 2 make: *** Waiting for unfinished jobs.... -- Ticket URL: <https://trac.ffmpeg.org/ticket/7521> FFmpeg <https://ffmpeg.org> FFmpeg issue tracker

1 3

#7433(avcodec:new): libaomenc.c: undeclared symbol AOM_ERROR_RESILIENT_PARTITIONS
by FFmpeg 03 Nov '18

03 Nov '18

#7433: libaomenc.c: undeclared symbol AOM_ERROR_RESILIENT_PARTITIONS ---------------------------------+-------------------------------------- Reporter: LigH | Type: defect Status: new | Priority: normal Component: avcodec | Version: git-master Keywords: | Blocked By: Blocking: | Reproduced by developer: 0 Analyzed by developer: 0 | ---------------------------------+-------------------------------------- Building ffmpeg with libaom fails in libavcodec/libaomenc.c:744:109: error: 'AOM_ERROR_RESILIENT_PARTITIONS' undeclared here (not in a function); did you mean 'AOM_ERROR_RESILIENT_DEFAULT'? {{{ " is still done over the partition boundary.", 0, AV_OPT_TYPE_CONST, {.i64 = AOM_ERROR_RESILIENT_PARTITIONS}, 0, 0, VE, "er"}, ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ AOM_ERROR_RESILIENT_DEFAULT }}} Attached archive of logs and other diagnostic files was generated by media-autobuild_suite (running in MSYS2/MinGW) -- Ticket URL: <https://trac.ffmpeg.org/ticket/7433> FFmpeg <https://ffmpeg.org> FFmpeg issue tracker

1 8

#7530(avcodec:new): cuviddec/nvdec/nvenc: driver checking errors
by FFmpeg 02 Nov '18

02 Nov '18

#7530: cuviddec/nvdec/nvenc: driver checking errors -------------------------------------+------------------------------------- Reporter: msiders | Type: defect Status: new | Priority: normal Component: avcodec | Version: git- Keywords: cuviddec, | master nvdec, nvenc | Blocked By: Blocking: | Reproduced by developer: 0 Analyzed by developer: 0 | -------------------------------------+------------------------------------- Summary of the bug: When executing the FFmpeg tool with an ancient NVidia driver, the output ins't clear about the incompatibility of the driver. This is the regular output when this error appears: {{{ [AVHWDeviceContext @ 0x1f2d000] Loaded lib: libcuda.so.1 [AVHWDeviceContext @ 0x1f2d000] Loaded sym: cuInit [AVHWDeviceContext @ 0x1f2d000] Loaded sym: cuDeviceGetCount [AVHWDeviceContext @ 0x1f2d000] Loaded sym: cuDeviceGet [AVHWDeviceContext @ 0x1f2d000] Loaded sym: cuDeviceGetName [AVHWDeviceContext @ 0x1f2d000] Cannot load cuDeviceGetUuid [AVHWDeviceContext @ 0x1f2d000] Could not dynamically load CUDA [h264_cuvid @ 0x1eec3c0] Error creating a CUDA device }}} Take note that the driver is installed and the card is working. However, when using a ''recent'' version of FFmpeg, in this case the error appears... as the driver version is too old. So, ''no errors at all'' when using a previous version of FFmpeg. Here are the three points where the driver version is checked: https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/cuviddec.c#L727 https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/nvdec.c#L97 https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/nvenc.c#L122 First of all, the function to check the driver version should be centralized (perhaps in "hwcontext_cuda.c"). And second, the function needs to log every time the minimum version required and the current version. This will help the user to identificate the problem. I hope you can fix this soon. Regards. -- Ticket URL: <https://trac.ffmpeg.org/ticket/7530> FFmpeg <https://ffmpeg.org> FFmpeg issue tracker

1 11

#7529(build system:new): cuda filters not compiled
by FFmpeg 02 Nov '18

02 Nov '18

#7529: cuda filters not compiled -------------------------------------+------------------------------------- Reporter: msiders | Type: defect Status: new | Priority: normal Component: build | Version: git- system | master Keywords: | Blocked By: Blocking: | Reproduced by developer: 0 Analyzed by developer: 0 | -------------------------------------+------------------------------------- Summary of the bug: When compiling FFmpeg with CUDA (./configure --enable-cuda) the filters based on this library isn't automatically enabled. See this example of a build of git-master when searching for CUDA filters: {{{ root@ffmpeg:~/FFmpeg# ./ffmpeg -filters | grep cuda ffmpeg version N-92330-gd6d407d Copyright (c) 2000-2018 the FFmpeg developers built with gcc 4.7 (Debian 4.7.2-5) configuration: --enable-static --disable-shared --disable-doc --disable- dxva2 --enable-small --enable-pthreads --enable-stripping --enable-bsfs --disable-devices --disable-hwaccels --disable-ffprobe --enable-gpl --enable-libx264 --enable-cuda --enable-nonfree --enable-cuvid --enable- nvenc --extra-cflags=-I/usr/local/cuda/include/ --extra- ldflags=-L/usr/local/cuda/lib64/ libavutil 56. 23.100 / 56. 23.100 libavcodec 58. 36.100 / 58. 36.100 libavformat 58. 21.100 / 58. 21.100 libavdevice 58. 6.100 / 58. 6.100 libavfilter 7. 41.100 / 7. 41.100 libswscale 5. 4.100 / 5. 4.100 libswresample 3. 4.100 / 3. 4.100 libpostproc 55. 4.100 / 55. 4.100 ... hwupload_cuda V->V (null) }}} As you can see the "--enable-cuda" is enabled in the build. But filters like "scale_cuda" and "thumbnail_cuda" are missing. To compile them is ''required'' to enable each one in the configure call. However, it has sense to enable them '''automatically''' when CUDA_SDK is set in the configuration. Please, fix this! -- Ticket URL: <https://trac.ffmpeg.org/ticket/7529> FFmpeg <https://ffmpeg.org> FFmpeg issue tracker

1 2

#7446(avcodec:new): x264 init race/corruption
by FFmpeg 01 Nov '18

01 Nov '18

#7446: x264 init race/corruption -------------------------------------+------------------------------------- Reporter: lukealonso | Type: defect Status: new | Priority: normal Component: avcodec | Version: Keywords: | unspecified Blocking: | Blocked By: Analyzed by developer: 0 | Reproduced by developer: 0 -------------------------------------+------------------------------------- Summary of the bug: I have a program that creates many concurrent instances of x264 via AVCodecContext. Occasionally, I'll get the following output: {{{ x264 [error]: invalid tune '[RANDOM JUNK]' }}} and avcodec_open2 will fail. I've attached a stripped down program that reproduces the issue, at least on Ubuntu 16.04. This appears to happen with all x264 versions, I tried the latest stable commit, and one from several years ago. This does *NOT* happen with older FFmpeg versions. I git bisected and found that the problem started happening on commit: {{{ [abaa12263e08ae7b2099f00fa4597ae6a86d7a9c] avcodec: Don't lock during open if the codec has threadsafe init] }}} .... which is not surprising, but not terribly helpful in finding what exactly is happening in X264_init that causes the issue. Hopefully this is just me being dumb, any pointers would be appreciated. How to reproduce: Build enc_test.c: {{{ gcc -o enc_test enc_test.c -lavcodec -lavformat -lavutil -lpthread }}} Run 1-10 times, with num_threads between 4-32, whatever is appropriate for your machine. Happens more often with higher thread counts. It should happen early on, so ctrl-c and try again if it hasn't happened with in a few seconds. {{{ ./enc_test [num_threads] }}} -- Ticket URL: <https://trac.ffmpeg.org/ticket/7446> FFmpeg <https://ffmpeg.org> FFmpeg issue tracker

1 5

#7505(avdevice:new): Coding Error in decklink_dec.cpp
by FFmpeg 01 Nov '18

01 Nov '18

#7505: Coding Error in decklink_dec.cpp ----------------------------------+-------------------------------------- Reporter: dbigler | Type: defect Status: new | Priority: important Component: avdevice | Version: git-master Keywords: decklink | Blocked By: Blocking: | Reproduced by developer: 0 Analyzed by developer: 0 | ----------------------------------+-------------------------------------- Summary of the bug: Just located a coding error in decklink_dec.cpp lines 1170 and 1171 should be swapped: st->codecpar->codec_tag = avcodec_pix_fmt_to_codec_tag((enum AVPixelFormat)st->codecpar->format); st->codecpar->format = AV_PIX_FMT_0RGB; as: st->codecpar->format = AV_PIX_FMT_0RGB; st->codecpar->codec_tag = avcodec_pix_fmt_to_codec_tag((enum AVPixelFormat)st->codecpar->format); lines 1176 and 1177 should also be swapped: st->codecpar->codec_tag = avcodec_pix_fmt_to_codec_tag((enum AVPixelFormat)st->codecpar->format); st->codecpar->format = AV_PIX_FMT_BGR0; as: st->codecpar->format = AV_PIX_FMT_BGR0; st->codecpar->codec_tag = avcodec_pix_fmt_to_codec_tag((enum AVPixelFormat)st->codecpar->format); Otherwise the value set in format will have no effect in the call to avcodec_pix_fmt_to_codec_tag(). -- Ticket URL: <https://trac.ffmpeg.org/ticket/7505> FFmpeg <https://ffmpeg.org> FFmpeg issue tracker

1 2