[Ffmpeg-devel-irc] ffmpeg-devel.log.20190314

Fri Mar 15 03:05:03 EET 2019

[01:06:35 CET] <cone-144> ffmpeg 03Alex Mogurenko 07release/4.1:7e204f72607c: avcodec/prores_ks: Fix luma quantization if q >= MAX_STORED_Q
[01:06:35 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:1a82246caed8: avcodec/sbrdsp_fixed.c: remove input value limit for sbr_sum_square_c()
[01:06:35 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:d25f388584a3: avformat/mov: Do not use reference stream in mov_read_sidx() if there is no reference stream
[01:06:35 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:8066cb3556e7: avcodec/mpeg4videodec: Clear interlaced_dct for studio profile
[01:06:35 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:4d1fcd734e87: avformat/matroskadec: Do not leak queued packets on sync errors
[01:06:36 CET] <cone-144> ffmpeg 03Kevin Backhouse via RT 07release/4.1:cc5361ed18ab: avcodec/htmlsubtitles: Fixes denial of service due to use of sscanf in inner loop for tag scaning
[01:06:36 CET] <cone-144> ffmpeg 03Kevin Backhouse via RT 07release/4.1:f7f3937494f6: avcodec/htmlsubtitles: Fixes denial of service due to use of sscanf in inner loop for handling braces
[01:06:37 CET] <cone-144> ffmpeg 03Wenxiang Qian 07release/4.1:11375cd10147: avformat/ftp: Fix Out-of-Bounds Access and Information Leak in ftp.c:393
[01:06:37 CET] <cone-144> ffmpeg 03Wenxiang Qian 07release/4.1:ec22b46a4d34: avformat/http: Fix Out-of-Bounds access in process_line()
[01:06:38 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:339f40f61818: avformat/webmdashenc: Check id in adaption_sets
[01:06:39 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:9495228df028: avcodec/h264_direct: Fix overflow in POC comparission
[01:06:40 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:4e624c89fddd: avcodec/jvdec: Check available input space before decode8x8()
[01:06:41 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:ff491b154442: avcodec/zmbv: obtain frame later
[01:06:42 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:92335fc02b2c: avcodec/mlpdec: Insuffient typo
[01:06:43 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:b8dd1d2d4bde: avcodec/error_resilience: Use a symmetric check for skipping MV estimation
[01:06:44 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:662b6351c80a: avcodec/jpeg2000dwt: Fix integer overflow in dwt_decode97_int()
[01:06:45 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:09683e1f4e21: avcodec/bethsoftvideo: Check block_type
[01:06:46 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:29619a8ac275: avcodec/gdv: Check for truncated tags in decompress_5()
[01:06:47 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:6c0124d3921d: avcodec/aic: Check remaining bits in aic_decode_coeffs()
[01:06:48 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:c90836cc3df3: avcodec/qpeg: Limit copy in qpeg_decode_intra() to the available bytes
[01:06:49 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:67d030787e0b: avcodec/scpr: Fix use of uninitialized variable
[01:06:50 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:ad12d9df1ed9: avcodec/dxv: Correct integer overflow in get_opcodes()
[01:06:51 CET] <cone-144> ffmpeg 03Michael Niedermayer 07release/4.1:1e50a327c649: avcodec/mpeg4videodec: Check idx in mpeg4_decode_studio_block()
[01:43:03 CET] <vel0city> do you guys have any GSoC candidates?
[10:10:34 CET] <JEEB> i think I am at peak masochist when I start writing a TLV parser >_>
[10:33:10 CET] <durandal_1707> i'm trying harder and harder and harder but can not RE this shit
[10:36:40 CET] <j-b> JEEB: in rust?
[10:40:13 CET] <JEEB> j-b: I... didn't think of that since I wanted the nice binary reading helpers of lavu :D
[10:40:19 CET] <JEEB> but that might be interestdusting
[10:40:30 CET] <j-b> JEEB: use Rust and Nom
[10:41:11 CET] <JEEB> yes, Nom I did know of
[10:43:12 CET] <BradleyS> michaelni: these can probably be backported, handbrake needs at least the first one
[10:43:13 CET] <BradleyS> https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/402bf262375dfecd0e90d7acc67c238abe952fc3
[10:43:13 CET] <BradleyS> https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/d9b2668766e3e924d4ebb3c6531b449874e13666
[10:43:13 CET] <BradleyS> https://git.ffmpeg.org/gitweb/ffmpeg.git/commitdiff/db332832a17c6da68067869da815dfcef174dca3
[10:44:34 CET] <nevcairiel> why are you not using pkg-config? :)
[10:47:53 CET] <j-b> especiallys since libvpx provides a vpx.pc now
[10:48:51 CET] <BradleyS> and pain
[10:49:05 CET] Action: BradleyS runs
[10:49:21 CET] <j-b> fighting against .pc in 2019 is beyond idiotic
[10:49:27 CET] <j-b> pkg-config has won, deal with it.
[10:49:31 CET] <JEEB> pkgconf doesn't require gobject
[10:49:38 CET] <JEEB> it also can be built with MSVC IIRC
[10:49:41 CET] <j-b> pkgconflite either
[10:49:43 CET] <j-b> etc..
[10:49:43 CET] <BradleyS> i meant libvpx :P
[10:49:43 CET] <JEEB> https://github.com/pkgconf/pkgconf/
[10:49:54 CET] <BradleyS> pkg-config is fine
[10:50:03 CET] <JEEB> different distros have moved to pkgconf as the pkg-config provider
[10:51:04 CET] Action: BradleyS wishes he had a cluster to throw at vp9/av1
[10:51:13 CET] <BradleyS> JEEB: which?
[10:51:32 CET] <j-b> VP9 is dead :0
[10:51:37 CET] <j-b> also, fork libvpx
[10:52:38 CET] Action: BradleyS sticks a fork in j-b's hat and runs faster
[10:52:43 CET] <j-b> :D
[10:55:48 CET] <JEEB> BradleyS: fedora quite a while ago, and I think I saw some other distro recently switch over during a package upgrade
[10:55:53 CET] <JEEB> not sure if it was ubuntu or what
[10:56:07 CET] <BradleyS> ah
[11:04:01 CET] <JEEB> ah yes, cygwin was actually the one I saw switch most recently
[12:25:04 CET] <cone-198> ffmpeg 03Matthew Fearnley 07master:032137060183: libavcodec/zmbvenc: Add support for RGB formats
[12:26:35 CET] <thardin> cutting edge high-color zmbv
[12:56:13 CET] <JEEB> durandal_1707: were you interested in poking at a seemingly incorrectly decoding APE file?
[12:57:10 CET] <durandal_1707> JEEB: what version of APE?
[12:57:39 CET] <durandal_1707> and how it decodes incorrecly?
[12:57:58 CET] <JEEB> I haven't listened to it
[12:58:05 CET] <JEEB> I just got a sample from an audio player dev
[12:59:09 CET] <JEEB> having him test with latest master if possible first
[12:59:16 CET] <JEEB> https://www.foobar2000.org/temp/Viande.ape is the sample
[13:01:14 CET] <JEEB> was just tested with d227ed5d59 which is not too old of a master revision and it seems to still be borked there :)
[13:02:55 CET] <faLUCE> I think that the line with "if ((ret = mjpeg_decode_app(s)) < 0) av_log(avctx, AV_LOG_ERROR, "unable to decode APP fields: %s\n", av_err2str(ret));"  in mjpegdec.h  is overkill in many cases. Currently lot of usb cameras have multiple APP fields with uncorrect length, so they raise continously this warning, which is ennoying... Since Ffmpeg 3.3.3 there was not this warning... would it be better to remove it?
[13:03:09 CET] <faLUCE> (mjpegdec.c  not, h)
[13:03:40 CET] <faLUCE> at lease IMHO should be removed for APP = AVI1
[13:03:46 CET] <faLUCE> (at least)
[13:06:00 CET] <faLUCE> in addition, MJPEG specs are very obscure. I only found an old doc, with google. So, there's no need to be so much strict with checkings...
[16:26:12 CET] <cone-129> ffmpeg 03Michael Niedermayer 07master:38381400fca4: avformat/gdv: Check fps
[16:26:12 CET] <cone-129> ffmpeg 03Michael Niedermayer 07master:9a9f0e239c1c: avcodec/cdgraphics: Use ff_set_dimensions()
[16:26:12 CET] <cone-129> ffmpeg 03Michael Niedermayer 07master:a8c5ae451184: avcodec/dvbsubdec: Check object position
[16:28:22 CET] <JEEB> seems like that stuff pops up every now and then due to android forcing PIC
[16:37:06 CET] <J_Darnley> Does that mean android still only does 32bit?
[16:38:59 CET] <nevcairiel>  there is also x64 android
[16:42:48 CET] <J_Darnley> Dude should upgrade his cpu to one from this decade then
[16:59:41 CET] <Gramner> for some reason people really like using 32-bit x86 instead of 64-bit with android
[17:00:50 CET] <gnafu> My recently purchased, current generation Fire TV Stick 4K appears to run 32-bit Android.  As far as I can tell, it has a 64-bit processor; I just can't install 64-bit apps.
[17:31:02 CET] <JEEB> yea they usually have 64bit kernel even
[17:31:21 CET] <JEEB> except those that intel.missed the driver builds with
[17:31:37 CET] <JEEB> those were thus limited to 32bit for the age of those devices
[17:51:45 CET] <cone-129> ffmpeg 03Guo, Yejun 07release/4.1:9bf40978c657: configure: add missing pthreads extralibs dependency for libvpx-vp9
[17:51:46 CET] <cone-129> ffmpeg 03Guo, Yejun 07release/4.1:83bfd4f3b5b6: configure: use vpx_codec_vp8_dx/cx for libvpx-vp8 checking
[17:51:47 CET] <cone-129> ffmpeg 03Michael Niedermayer 07release/4.1:5d208aac5253: avformat/gdv: Check fps
[17:51:48 CET] <cone-129> ffmpeg 03Michael Niedermayer 07release/4.1:04ce4cc0728e: avcodec/cdgraphics: Use ff_set_dimensions()
[17:51:49 CET] <cone-129> ffmpeg 03Michael Niedermayer 07release/4.1:8cee4190f368: avcodec/dvbsubdec: Check object position
[17:51:50 CET] <cone-129> ffmpeg 03Michael Niedermayer 07release/4.1:77d244e7a937: Update for 4.1.2
[18:47:35 CET] <BtbN> tfw some "I get paid by Google and have an important 12h meeting with them tomorrow and need you to acknowledge the 1200 zero day exploits I found!!!" guy turns up right before your shift ends.
[18:53:03 CET] <JEEB> ouch
[19:44:29 CET] <cone-129> ffmpeg 03Diego Biurrun 07master:a70eac7a9b19: tests: Convert image2pipe tests to non-legacy test scripts
[19:44:30 CET] <cone-129> ffmpeg 03Diego Biurrun 07master:eb8a81159944: tests: Convert audio-only lavf tests to non-legacy test scripts
[19:44:31 CET] <cone-129> ffmpeg 03Diego Biurrun 07master:a957e9379d11: tests: Convert lavf image tests to non-legacy test scripts
[19:44:32 CET] <cone-129> ffmpeg 03Diego Biurrun 07master:896fe15dbb7b: tests: Convert lavf pixfmt conversion tests to non-legacy test scripts
[19:44:33 CET] <cone-129> ffmpeg 03Diego Biurrun 07master:618d02c1fa9e: tests: Convert lavf container tests to non-legacy test scripts
[19:44:34 CET] <cone-129> ffmpeg 03Diego Biurrun 07master:f8df5e2f31a5: tests: Add a convenience function for video-only lavf tests
[19:44:35 CET] <cone-129> ffmpeg 03James Almer 07master:d0c43e32427e: Merge commit 'a70eac7a9b193e8434b5bed90bd72aa4cb688363'
[19:44:36 CET] <cone-129> ffmpeg 03James Almer 07master:29929fc8e244: Merge commit 'eb8a8115994434b548523cf0bca6a4a74784e79c'
[19:44:37 CET] <cone-129> ffmpeg 03James Almer 07master:c2631dfd0a0a: Merge commit 'a957e9379d11f2982d615f92c30580a57ea8bb40'
[19:44:38 CET] <cone-129> ffmpeg 03James Almer 07master:ace96d2e693d: Merge commit '896fe15dbb7b78de495c4a7dd75e7faec66778da'
[19:44:39 CET] <cone-129> ffmpeg 03James Almer 07master:fda424b300c1: Merge commit '618d02c1fa9e74d490cace64a7d15762656b521c'
[19:44:40 CET] <cone-129> ffmpeg 03James Almer 07master:a200bf8094bf: tests: Convert lavf container tests with samples to non-legacy test scripts
[19:44:41 CET] <cone-129> ffmpeg 03James Almer 07master:ac4b5d862220: Merge commit 'f8df5e2f31a5ba7b30a0e1caaaf5a03c753b3f9b'
[19:46:28 CET] <jamrial> those took a while. the build system has deviated quite a bit by now
[19:53:56 CET] <cone-129> ffmpeg 03Luca Barbato 07master:899ee03088d5: Unbreak travis on macos
[19:53:57 CET] <cone-129> ffmpeg 03James Almer 07master:ecb5546e97c4: Merge commit '899ee03088d55152a48830df0899887f055da1de'
[20:12:10 CET] <cone-129> ffmpeg 03Magnus Röös 07master:0801853e6406: libavcodec: vp8 neon optimizations for aarch64
[20:12:11 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:26d7af4c381e: aarch64: vp8: Fix assembling with clang
[20:12:12 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:2eeac79936e8: aarch64: vp8: Fix assembling with armasm64
[20:12:13 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:85bfaa4949f4: aarch64: vp8: Use the proper aarch64 form for conditional branches
[20:12:14 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:ad32f7b1264d: aarch64: vp8: Remove superfluous includes
[20:12:15 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:b4b27dce95a6: aarch64: vp8: Move the vp8dsp makefile entries to the right places
[20:12:16 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:f1011ea28a40: aarch64: vp8: Reorder the function pointer inits to match the arm original
[20:12:17 CET] <cone-129> ffmpeg 03James Almer 07master:a2ae381b5a6f: Merge commit '0801853e640624537db386727b36fa97aa6258e7'
[20:12:18 CET] <cone-129> ffmpeg 03James Almer 07master:94346ab593f3: Merge commit '2eeac79936e83c4495cbe5905064ab797e9b45ff'
[20:12:19 CET] <cone-129> ffmpeg 03James Almer 07master:398000abcf98: Merge commit '85bfaa4949f4afcde19061def3e8a18988964858'
[20:12:20 CET] <cone-129> ffmpeg 03James Almer 07master:b8eb0827f053: Merge commit 'b4b27dce95a6d40bfcd78043d3abec7d80dae143'
[20:12:21 CET] <cone-129> ffmpeg 03James Almer 07master:4f9a8d3fe2f9: Merge commit 'f1011ea28a4048ddec97794ca3e9901474fe055f'
[20:16:14 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:c513fcd7d235: aarch64: vp8: Fix a typo in a comment
[20:16:14 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:52c9b0a6c0d0: aarch64: vp8: Port vp8_luma_dc_wht and vp8_idct_dc_add4uv from arm version
[20:16:15 CET] <cone-129> ffmpeg 03James Almer 07master:39278ff0de5b: Merge commit 'c513fcd7d235aa4cef45a6c3125bd4dcc03bf276'
[20:16:17 CET] <cone-129> ffmpeg 03James Almer 07master:79025da3f2e7: Merge commit '52c9b0a6c0d02cff6caebcf6989e565e05b55200'
[20:17:15 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:cc7ba00c35fa: aarch64: vp8: Port missing epel8 functions from arm version
[20:17:16 CET] <cone-129> ffmpeg 03James Almer 07master:c6892f59eb0e: Merge commit 'cc7ba00c35faf0478f1f56215e926f70ccb31282'
[20:18:26 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:58d154922707: aarch64: vp8: Port epel4 functions from arm version
[20:18:27 CET] <cone-129> ffmpeg 03James Almer 07master:2ac399d7faa5: Merge commit '58d154922707bfeb873cb3a7476e0f94b17463dd'
[20:19:03 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:e39a9212ab37: aarch64: vp8: Port bilin functions from arm version
[20:19:04 CET] <cone-129> ffmpeg 03James Almer 07master:34a0a9746b2f: Merge commit 'e39a9212ab37a55b346801c77487d8a47b6f9fe2'
[20:21:33 CET] <dominiques> Hello please, can you HELP me... i really need solve using ffmpeg cutting exactly 00:04:00 second audio fragments ... i need cut ideally aac or wav, for example, i have 60 second audio file, and i need to cut them to exactly 15*4sec audio fragments i am using for example ffmpeg -i $DIR_TMP/"${file%.mp4}.wav" -segment_time 4 -f segment $DIR_TMP/"${file%.mp4}%03d.wav"
[20:22:00 CET] <dominiques> please PM to me 
[20:22:33 CET] <J_Darnley> With uncompressed you should be able to cut to *exactly* 15 seconds
[20:22:58 CET] <J_Darnley> with aac youre gonna have to hope and pray to the frequency transform gods
[20:23:34 CET] <J_Darnley> oh sorry, 4 seconds
[20:24:11 CET] <J_Darnley> Also this isnt the tech support channel
[20:24:12 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:cef914e08310: arm: vp8: Optimize put_epel16_h6v6 with vp8_epel8_v6_y2
[20:24:13 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:37394ef01b04: aarch64: vp8: Optimize put_epel16_h6v6 with vp8_epel8_v6_y2
[20:24:14 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:49f9c4272c40: aarch64: vp8: Skip saturating in shrn in ff_vp8_idct_add_neon
[20:24:15 CET] <cone-129> ffmpeg 03Martin Storsjö 07master:7e42d5f0ab2a: aarch64: vp8: Optimize vp8_idct_add_neon for aarch64
[20:24:16 CET] <cone-129> ffmpeg 03James Almer 07master:d6b62ce1aced: Merge commit 'cef914e08310166112ac09567e66452a7679bfc8'
[20:24:17 CET] <cone-129> ffmpeg 03James Almer 07master:fbd607dd560a: Merge commit '37394ef01b040605f8e1c98e73aa12b1c0bcba07'
[20:24:18 CET] <cone-129> ffmpeg 03James Almer 07master:409e684e79b6: Merge commit '49f9c4272c4029b57ff300d908ba03c6332fc9c4'
[20:24:19 CET] <cone-129> ffmpeg 03James Almer 07master:5c363d3e595a: Merge commit '7e42d5f0ab2aeac811fd01e122627c9198b13f01'
[20:32:39 CET] <cone-129> ffmpeg 03Janne Grunau 07master:186bd30aa3b6: h264/arm64: implement missing 4:2:2 chroma loop filter neon functions
[20:32:40 CET] <cone-129> ffmpeg 03Janne Grunau 07master:f8abf7d4dfa0: checkasm/h264: test 4:2:2 chroma loop filter functions
[20:32:41 CET] <cone-129> ffmpeg 03James Almer 07master:92219ef4ac01: Merge commit '186bd30aa3b6c2b29b4dbf18278700b572068b1e'
[20:32:42 CET] <cone-129> ffmpeg 03James Almer 07master:0dda0f3bdb7e: Merge commit 'f8abf7d4dfa0504f7f65e4f1fd9d22e01cb371cc'
[21:17:12 CET] <dominiques> I CAN PAY FOR YOUR HELP ! REALLY NEED SOLVE THIS .. PLEASE CHECK https://codeshare.io/aVnze3
[21:17:51 CET] <cone-129> ffmpeg 03James Almer 07master:85649b6523a6: fate/lavf-image: fix passed arguments for some high bit depth tests
[21:36:57 CET] <J_Darnley> another damn pastebin that requires javascript
[23:11:51 CET] <uau> qtrle now has code like this:
[23:11:54 CET] <uau>                   AV_WN16A(rgb + pixel_ptr, rg);
[23:11:54 CET] <uau>                     rgb[pixel_ptr + 2] = b;
[23:11:54 CET] <uau>                     pixel_ptr += 3;
[23:12:23 CET] <uau> that combination of AV_WN16A and ptr+=3 looks very suspicious
[23:12:44 CET] Action: JEEB double-blinks
[23:16:27 CET] <jamrial> uau: i mentioned a qtrle fate failure introduced by that to the author of that code
[23:19:51 CET] <atomnuker> uau: its rgb so 24 bits, we don't have writes by 24 bits so it looks okay
[23:24:23 CET] <uau> atomnuker: what do you mean?
[23:25:47 CET] <uau> to clarify, that code looks like it'd violate alignment, and i don't see how your comment would say it doesn't
[23:29:00 CET] <atomnuker> alignment?
[23:29:49 CET] <uau> yes, as in the 'A' at the end of 'AV_WN16A'
[23:31:51 CET] <atomnuker> right, forgot about that
[00:00:00 CET] --- Fri Mar 15 2019