[Ffmpeg-devel-irc] ffmpeg-devel.log.20151119

burek burek021 at gmail.com
Fri Nov 20 02:05:02 CET 2015


[00:00:23 CET] <kierank> I think it's a false positive though
[00:32:12 CET] <atomnuker> tested my harddrive:
[00:32:16 CET] <atomnuker> "Completed: read failure       70%     15220         576"
[00:32:47 CET] <atomnuker> common sense demands that there would be at least a few reallocated sectors
[00:32:53 CET] <BBB> kierank: if its entirely bugfree, that would be very cool
[00:33:43 CET] <rcombs> nevcairiel: put automatic BSF support through FATE; passes everything
[00:34:06 CET] <rcombs> nevcairiel: resent patch series with the name tweak you suggested and addition of the deinit API
[00:41:36 CET] <wm4> kierank: lol "file world" vs. "standard compliant world"
[00:44:49 CET] <cone-223> ffmpeg 03Michael Niedermayer 07master:50ce510ac4e3: avcodec/pthread_slice: remove dummy_ret hack
[01:16:30 CET] <nevcairiel> rcombs: sounds good, will try to give it a final review on the ML tomorrow
[01:16:39 CET] <rcombs> thanks
[01:18:00 CET] <nevcairiel> having a quick look, is write trailer the correc point to call deinit? Don't we have a "close" function that might seem more logical, somehow?
[01:19:55 CET] <nevcairiel> avformat_free_context, thats it
[01:20:03 CET] <nevcairiel> probably make no difference whatsoever
[01:21:33 CET] <nevcairiel> anway sleepy time
[01:57:17 CET] <atomnuker> my hard drive was a walking zombie, I've ran badblocks for 5 minutes and it already has around 300 bad sectors
[01:58:07 CET] <atomnuker> but somehow it still boots and home appears to mount fine
[02:13:21 CET] <Compn> good time to backup
[02:13:44 CET] <atomnuker> alot of things seem to be gone, and I had no backups, though I had most stuff of real value in text files
[02:14:01 CET] <atomnuker> well that'll teach me to keep sha512s of valuables
[02:19:00 CET] <atomnuker> thinking of installing arch this time instead of debian
[02:19:18 CET] <atomnuker> I hate how everything was in *dev packages
[02:21:58 CET] <atomnuker> or maybe fedora rawhide
[03:40:16 CET] <kierank> wm4: well if you generate timestamps which are in the past its a problem because mpegts isnt like TCP formats where there's a ton of buffering
[03:48:18 CET] <cone-486> ffmpeg 03Michael Niedermayer 07n2.7.3:HEAD: avcodec/pthread_slice: remove dummy_ret hack
[09:06:16 CET] <wm4> atomnuker: what brand is that drive? (not that it matters)
[09:16:24 CET] <wm4> ubitux: so I dropped mpv's libass-based ass loader, and just within a week there was already a report of some ass files not working (they didn't have the headers in a form lavf wanted - extra newline before the script header)
[10:02:04 CET] <atomnuker> wm4: a western digital blue
[10:02:38 CET] <atomnuker> extremely abused, load cycle count of something like 10 000 000
[10:02:50 CET] <atomnuker> g sense of ~10 000
[10:03:03 CET] <atomnuker> and yet only 1 pending sector and 0 reallocated
[10:04:55 CET] <wm4> the smart data is probably just lying
[10:07:36 CET] <atomnuker> I did do regular long tests, but still nothing
[10:07:56 CET] <atomnuker> then again, it's a 2.5 inch model, they're not meant to survive
[10:35:01 CET] <ubitux> wm4: sample?
[10:35:28 CET] <ubitux> so, nv to rgb in neon are done, but we don't have the same results as the c version
[10:35:33 CET] <ubitux> but... it looks better
[10:36:29 CET] <fritsch> hehe
[10:36:47 CET] <ubitux> http://b.pkh.me/nv12-rgba-ref.png http://b.pkh.me/nv12-rgba-neon.png
[10:36:55 CET] <fritsch> some itu colors missing?
[10:37:00 CET] <fritsch> bt601 / bt709?
[10:37:03 CET] <fritsch> while converting?
[10:37:16 CET] <ubitux> no it looks like some issue in the chroma averaging or something
[10:37:21 CET] <nevcairiel> the  ref seems to have extra ringing
[10:37:25 CET] <ubitux> yeah
[10:37:37 CET] <wm4> ubitux: no sample, but https://github.com/mpv-player/mpv/issues/2506
[10:38:41 CET] <fritsch> mmh, i would not say this edges (if you look at the right squares) are ringing
[10:38:56 CET] <wm4> vsfilter is in general _very_ tolerant about such errors
[10:39:50 CET] <fritsch> but yeah the ref version seems to have "smoothed" borders
[10:40:01 CET] <fritsch> while the neon one has very rough edges
[10:40:11 CET] <fritsch> probably ringing compensation by butterworth on the ref?
[10:40:15 CET] <wm4> (in fact, I don't think vsfilter reads/needs the script header line at all)
[11:28:03 CET] <cone-326> ffmpeg 03wm4 07release/2.6:29fde41906a0: rawdec: fix mjpeg probing
[11:28:04 CET] <cone-326> ffmpeg 03wm4 07release/2.6:da7d36ef8037: rawdec: fix mjpeg probing buffer size check
[11:28:05 CET] <cone-326> ffmpeg 03Emanuel Czirai 07release/2.6:ec328dc46182: libavcodec/aacdec_template: Use init_get_bits8() in aac_decode_frame()
[11:28:06 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:dd19374ed71a: avcodec/dcaenc: clear bitstream end
[11:28:07 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:6c32c9e1eff3: avcodec/svq1enc: Check dimensions
[11:28:08 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:8e597b57d0b1: avcodec/flashsvenc: Correct max dimension in error message
[11:28:09 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:dea4fd94467b: avformat/mux: Update sidedata in ff_write_chained()
[11:28:10 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:5062059c531e: ffmpeg: check avpicture_fill() return value
[11:28:11 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:74ce026e684f: ffmpeg: Check for RAWVIDEO and do not relay only on AVFMT_RAWPICTURE
[11:28:12 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:9c3d81481798: ffmpeg: Check av_parser_change() for failure
[11:28:13 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:462c1dacaef7: ffmpeg: Use correct codec_id for av_parser_change() check
[11:28:14 CET] <cone-326> ffmpeg 03Arthur Grant 07release/2.6:d405a10743e5: avformat/hevc: Fix parsing errors
[11:28:15 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:10870af7e46b: avformat/hevc: Check num_long_term_ref_pics_sps to avoid potentially long loops
[11:28:16 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:c3ee01b9da83: avcodec/libopusenc: Fix infinite loop on flushing after 0 input
[11:28:17 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:2d9f536cd95f: avcodec/mpeg12dec: Set dimensions in mpeg1_decode_sequence() only in absence of errors
[11:28:18 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:ba1715f52a96: avcodec/truemotion1: Check for even width
[11:28:19 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:c23902b6717b: avformat/dump: Fix integer overflow in aspect ratio calculation
[11:28:20 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:3311efe2c229: avutil/common: Document FFABS() corner case
[11:28:21 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:ca89c0afdb14: avutil/common: Add FFNABS()
[11:28:22 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:fb84d92079af: avformat/mov: Fix integer overflow in FFABS
[11:28:23 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:3e9ccd001901: swresample/swresample: Fix integer overflow in seed calculation
[11:28:24 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:719f17541f7e: avcodec/vp3: Check init_get_bits8() for failure
[11:28:25 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:75d186ab6ea3: avcodec/tta: Check init_get_bits8() for failure
[11:28:26 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:a2e0602157ad: avcodec/svq1dec: Check init_get_bits8() for failure
[11:28:27 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:9a90f61d998a: avcodec/g2meet: Fix potential overflow in tile dimensions check
[11:28:28 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:cb1bb8ca459d: avcodec/g2meet: Also clear tile dimensions on header_fail
[11:28:29 CET] <cone-326> ffmpeg 03Ganesh Ajjanagadde 07release/2.6:a2d38168fdda: avfilter/af_asyncts: use llabs for int64_t
[11:28:30 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:4ef257f702bd: avcodec/mjpegdec: Fix decoding RGBA RCT LJPEG
[11:28:31 CET] <cone-326> ffmpeg 03Simon Thelen 07release/2.6:b096f160957b: lavf/webvttenc: Require webvtt file to contain exactly one WebVTT stream.
[11:28:32 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:e77ad4b06954: avcodec/rangecoder: Check e
[11:28:33 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:17bbabc141a1: avcodec/ffv1dec: Explicitly check read_quant_table() return value
[11:28:34 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:53f11a47a350: avcodec/ffv1dec: Fix off by 1 error in quant_table_count check
[11:28:35 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:873938b83fe3: avcodec/x86/sbrdsp: Fix using uninitialized upper 32bit of noise
[11:28:36 CET] <cone-326> ffmpeg 03Andrey Utkin 07release/2.6:289e74d72dbd: avformat/httpauth: Add space after commas in HTTP/RTSP auth header
[11:28:37 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:2da580a6bbdc: avcodec/ffv1dec: Clear slice coordinates if they are invalid or slice header decoding fails for other reasons
[11:28:38 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:f8951f4c7af8: avcodec/ffv1dec: update progress in case of broken pointer chains
[11:28:39 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:a456a5285d15: avcodec/ffv1: Initialize vlc_state on allocation
[11:28:40 CET] <cone-326> ffmpeg 03Kieran Kunhya 07release/2.6:d52d5553b8dc: opusdec: Don't run vector_fmul_scalar on zero length arrays
[11:28:41 CET] <cone-326> ffmpeg 03Tobias Rapp 07release/2.6:e0569172d644: avutil/file_open: avoid file handle inheritance on Windows
[11:28:42 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:4fec2df719b3: avcodec/mjpegdec: Check index in ljpeg_decode_yuv_scan() before using it
[11:28:43 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:dd6274d904ac: avcodec/mjpegdec: Reinitialize IDCT on BPP changes
[11:28:44 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:139a96b6ebf3: avcodec/ffv1dec: Check for 0 quant tables
[11:28:45 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:cbafd8e1a7eb: libavutil/channel_layout: Check strtol*() for failure
[11:28:46 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:371970e45faa: avcodec/mpeg12dec: Do not call show_bits() with invalid bits
[11:28:47 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:0b3b993bae22: avformat/xmv: factor return check out of if/else
[11:28:48 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:36b65f1c50f4: avformat/xmv: Discard remainder of packet on error
[11:28:49 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:23c7dbc16401: avcodec/dirac_parser: Fix undefined memcpy() use
[11:28:50 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:a420cb8e45f5: avcodec/microdvddec: Check for string end in 'P' case
[11:28:51 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:6a82ec8844a1: avcodec/hevc_ps: Check chroma_format_idc
[11:28:52 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:9f6e755272e7: avcodec/jpeg2000dec: Clip all tile coordinates
[11:28:53 CET] <cone-326> ffmpeg 03Andreas Cadhalpun 07release/2.6:bbeae2c690f1: jvdec: avoid unsigned overflow in comparison
[11:28:54 CET] <cone-326> ffmpeg 03Simon Thelen 07release/2.6:c3467bbc5250: ffmpeg: Don't try and write sdp info if none of the outputs had an rtp format.
[11:28:55 CET] <cone-326> ffmpeg 03Simon Thelen 07release/2.6:d9bff0843f67: doc/ffmpeg: Clarify that the sdp_file option requires an rtp output.
[11:28:56 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:53869b687c4e: avcodec/jpeg2000dec: Check for duplicate SIZ marker
[11:28:57 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:896c8e38915c: avcodec/utils: Better check for channels in av_get_audio_frame_duration()
[11:28:58 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:c00d40095bf2: avcodec/ivi: Check image dimensions
[11:28:59 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:1c73f9d84c8c: avcodec/flashsv: Check size before updating it
[11:29:00 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:7ab1f5b0d4b6: avcodec/dpx: Move need_align to act per line
[11:29:01 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:cb060668f129: avcodec/error_resilience: avoid accessing previous or next frames tables beyond height
[11:29:02 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:3f4a4f174a41: avcodec/dxtory: Fix input size check in dxtory_decode_v1_420()
[11:29:03 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:9bccaf7cf276: avcodec/dxtory: Fix input size check in dxtory_decode_v1_410()
[11:29:04 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:2d23cd42e35c: avcodec/takdec: Skip last p2 sample (which is unused)
[11:29:05 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:cf2d26eacf62: avcodec/smacker: Check that the data size is a multiple of a sample vector
[11:29:06 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:f28838903173: avcodec/wmaprodec: Check for overread in decode_packet()
[11:29:07 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:8930b2d532c9: avcodec/jpeg2000: Use av_image_check_size() in ff_jpeg2000_init_component()
[11:29:08 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:f86984653747: avcodec/jpeg2000: Check comp coords to be within the supported size
[11:29:09 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:912dfe167fbb: avcodec/jpeg2000dec: Check SIZ dimensions to be within the supported range
[11:29:10 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:b7d98b2fe7f6: avcodec/jpeg2000dec: Fix potential integer overflow with tile dimensions
[11:29:11 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:7c3fc6bf0641: avformat/utils: Do not init parser if probing is unfinished
[11:29:12 CET] <cone-326> ffmpeg 03Michael Niedermayer 07release/2.6:89612acc81d1: avformat/matroskadec: Check subtitle stream before dereferencing
[11:29:13 CET] <cone-326> ffmpeg 03Martin Storsjö 07release/2.6:949f5dbcfbb8: rtmpcrypt: Do the xtea decryption in little endian mode
[11:35:11 CET] <TimNich> wow thats  lot of patches in one hit!
[11:35:29 CET] <ubitux> so for 1080p frames, the convert function (arm/neon) takes about 30ms, but it seems to spend 2/3 of the time in the write (vst) call
[11:35:42 CET] <ubitux> TimNich: release/2.6, it's backports
[11:35:57 CET] <ubitux> so, any idea how we can speed up the writing?
[11:35:58 CET] <TimNich> Yeah, I just noticed that :(
[11:38:18 CET] <ubitux> the write is: vst2.32 {q6, q7}, [\dst,:128]!
[13:16:24 CET] <cone-326> ffmpeg 03Michael Niedermayer 07master:bfd0e02dd64e: avcodec/aacsbr: Split pre dequantization noise factors table
[13:16:25 CET] <cone-326> ffmpeg 03Michael Niedermayer 07master:d877b88f5188: avcodec/aacsbr_template: Check values read in read_sbr_noise()
[13:55:47 CET] <cone-326> ffmpeg 03Michael Niedermayer 07master:58fba4b02f65: avcodec/aacsbr: Replace a noise_facs_q check by an av_assert0
[14:21:03 CET] <ubitux> we don't have a stream select filter, right?
[14:52:26 CET] <Daemon404> nevcairiel, how did you ever manage to merge "cleanly"
[14:52:39 CET] <Daemon404> because if someone pushes stuff in the meantime while you merge
[14:52:46 CET] <Daemon404> you can't "rebasE" that merge
[14:52:50 CET] <nevcairiel> Well
[14:52:52 CET] <nevcairiel> Kinda
[14:52:54 CET] <Daemon404> and you end up with a horrible "merge ffmpeg/master" commit
[14:53:14 CET] <nevcairiel> You can rebase merges with rebase -p and rerere
[14:53:41 CET] <Daemon404> ive never heard of rerere
[14:53:44 CET] <Daemon404> or -p for rebase
[14:54:10 CET] <nevcairiel> P ia preserve merges, and rerere records conflict resolutions
[14:54:42 CET] <nevcairiel> So when no new conflicts show up, its pretty easy
[14:54:45 CET] <Daemon404> not sure i follow the latter
[14:55:13 CET] <Daemon404> so git fetch, git rebase -p, then...?
[14:55:25 CET] <Daemon404> (cause afaict you cannot pass params to pull --rebase)
[14:55:27 CET] <nevcairiel> rebase -p doesn't really "preserve" the merge commit, but rather re-applies it
[14:55:35 CET] <nevcairiel> So the old conflicts show up again
[14:55:47 CET] <nevcairiel> Which is where rerere comes in
[14:55:57 CET] <Daemon404> so what's an example workflow here
[14:56:00 CET] <Daemon404> assumign ive already merged
[14:56:08 CET] <Daemon404> and just want to get ready ti push
[14:56:09 CET] <Daemon404> to*
[14:56:26 CET] <nevcairiel> fetch, rebase -p, add/commit, push
[14:56:39 CET] <nevcairiel> Need to enable rerere first
[14:56:43 CET] <Daemon404> so rerere is a config option
[14:57:25 CET] <Daemon404> google says it's also an option
[14:57:26 CET] <Daemon404> confusing
[14:57:32 CET] <Daemon404> s/option/command/
[14:58:07 CET] <nevcairiel> After the rebase -p it will still show you the usual conflicts messages but then apply your previous conflict resolution, so you can just add the changes and commit/ continue rebase
[14:58:24 CET] <nevcairiel> Or just merge faster :D
[14:58:32 CET] <Daemon404> so i never call 'git rerere' ?
[14:59:00 CET] <nevcairiel> No you enable the option
[14:59:06 CET] <Daemon404> it's also a command ;p
[14:59:13 CET] <Daemon404> what a terrible name and UX
[14:59:18 CET] <Daemon404> par for teh course for git i suppose
[14:59:21 CET] <nevcairiel> To clear the rerere cache I think
[14:59:30 CET] <Daemon404> ic
[14:59:30 CET] <nevcairiel> It's a hidden feature more or leas
[14:59:48 CET] <Daemon404> it's somewhat advanced for most i guess
[14:59:57 CET] <Daemon404> usually i just see 40 merge master commits
[14:59:59 CET] <Daemon404> in ugly branches
[15:00:01 CET] <Daemon404> (on github)
[15:00:17 CET] <nevcairiel> Yeah I wanted to avoid that, so I searched for solutions
[15:00:39 CET] <Daemon404> yea.
[15:01:20 CET] <nevcairiel> And some commits take a bit to figure out, fix, test, so the chance of someone pushing in between was huge
[15:01:50 CET] <Daemon404> yeah exactly
[15:01:57 CET] <nevcairiel> Note that it doesn't track changes you made to files the merge didn't originally change
[15:01:57 CET] <Daemon404> i had to redo the movenc merges like 5 times
[15:07:48 CET] <cone-326> ffmpeg 03Carl Eugen Hoyos 07master:73b03249133d: aiff: Support demuxing G.722 streams
[15:07:49 CET] <cone-326> ffmpeg 03Derek Buitenhuis 07master:c63847528219: Merge commit '73b03249133d6631520089798087cbd963eed9ca'
[15:09:44 CET] <cone-326> ffmpeg 03Vittorio Giovara 07master:e25cac50e05d: lavc: Add missing mem.h header to libxvid and screenpresso
[15:09:45 CET] <cone-326> ffmpeg 03Derek Buitenhuis 07master:8f63355f8f38: Merge commit 'e25cac50e05d29a15d7a52c01c394ba913c97aee'
[15:32:22 CET] <cone-326> ffmpeg 03Andreas Cadhalpun 07master:e6459c655e11: dds: validate source buffer size before copying
[15:32:23 CET] <cone-326> ffmpeg 03Andreas Cadhalpun 07master:29b17528612c: dds: validate compressed source buffer size
[15:32:24 CET] <cone-326> ffmpeg 03Vittorio Giovara 07master:303f931938c6: mov: Correctly store dref paths
[15:32:25 CET] <cone-326> ffmpeg 03Vittorio Giovara 07master:b5f963bfec1f: mov: Drop dref when unable to parse
[15:32:26 CET] <cone-326> ffmpeg 03Andreas Cadhalpun 07master:7b2211bfc4b0: dds: add missing newline to log messages
[15:32:27 CET] <cone-326> ffmpeg 03Martin Storsjö 07master:1fc94724f1fd: xtea: Clarify that the current API works in big endian mode
[15:32:28 CET] <cone-326> ffmpeg 03Derek Buitenhuis 07master:218f46fc99dd: Merge commit '303f931938c618668f7f83c646a1850bef84641e'
[15:32:29 CET] <cone-326> ffmpeg 03Derek Buitenhuis 07master:f7c3ec3a01e4: Merge commit 'b5f963bfec1f452c37eee900c7b11f065d10dd60'
[15:32:30 CET] <cone-326> ffmpeg 03Derek Buitenhuis 07master:443b3c03e8ef: Merge commit '7b2211bfc4b0c4568180a8db2478023c42d9ff51'
[15:32:31 CET] <cone-326> ffmpeg 03Derek Buitenhuis 07master:3d2363fbf9fa: Merge commit '1fc94724f1fd52944bb5ae571475c621da4b77a0'
[15:33:06 CET] <Daemon404> thats enough for now...
[17:11:51 CET] <ubitux> command input doesn't work anymore in ffmpeg :(
[17:12:40 CET] <ubitux> (can't type shit after opening it)
[17:14:20 CET] <Daemon404> ffmpeg had an interactive mode?
[17:15:00 CET] <ubitux> yes
[17:15:02 CET] <ubitux> press 'c'
[17:15:17 CET] <ubitux> i use it sometimes, but it appears i can't anymore
[17:15:19 CET] <ubitux> sadness.
[17:40:01 CET] <RiCON> ubitux: if on mintty+msys you probably need to run ffmpeg using winpty
[17:40:15 CET] <ubitux> i don't use windows
[17:47:52 CET] <J_Darnley> 24 hours down and I am so close to having a working cross compiler
[17:52:52 CET] <ubitux> haha
[17:53:06 CET] <ubitux> yeah it's always a mess to get that :(
[17:53:33 CET] <J_Darnley> Down to just this error when trying to compile hello world: ld: cannot find -lgcc_s
[17:58:20 CET] <jamrial> RiCON: thank you. now i can press q to abort encodings when using msys2's mintty
[18:20:32 CET] Action: Daemon404 is too scared to start reading this udp thread
[18:22:04 CET] <kierank> yeah i might stop replying
[20:09:02 CET] <cone-308> ffmpeg 03James Almer 07master:01c63e69f174: avformat/rsd: GADP files are adpcm_thp_le
[20:09:02 CET] <cone-308> ffmpeg 03James Almer 07master:183827e6894c: avformat/rsd: reindent after the last commit
[21:25:04 CET] <ubitux> libav rewrote the hevc mc x86 simd?
[21:26:04 CET] <JEEB> yeah, I was wondering if that was similar/same
[21:32:11 CET] <cone-308> ffmpeg 03Paul B Mahol 07master:9fb8c5d85344: avformat: add VPK demuxer
[21:37:29 CET] <michaelni> ubitux, pressing 'c' still brings up the "Enter command: <target>|all <time>|-1 <command>[ <argument>]"
[21:38:13 CET] <ubitux> but can you type anything?
[21:56:26 CET] <andrewrk> for retrieving album art, AVStream.attached_pic gives an AVPacket which contains the image. but how can one tell what format the data is in? and can ffmpeg decode it to a bitmap?
[21:57:00 CET] <andrewrk> perhaps the stream codec field?
[21:58:02 CET] <J_Darnley> Something in ffmpeg is certainly dumb enough to turn jpg covers into png covers without notice.
[21:58:56 CET] <andrewrk> J_Darnley, I'm not sure what you're trying to communicate
[21:59:09 CET] <J_Darnley> I'm under the impression that it just reads the picture as a video stream
[21:59:24 CET] <J_Darnley> can ffmpeg decode it to a bitmap? <-- Yes
[22:00:14 CET] <andrewrk> can an AVPacket represent more than one video frame?
[22:00:48 CET] <andrewrk> if so, then I should try to use a decoder to decode the AVPacket, and then somehow get the bitmap of the first frame?
[22:01:20 CET] <J_Darnley> Something like that but I can't give a specific answer.
[22:01:27 CET] <andrewrk> ok, thanks for the clues, anyway
[22:09:33 CET] <michaelni> ubitux, yes, works here, "all -1 ping gives some pongs
[22:20:52 CET] <wm4> andrewrk: the AVStream tells you the format, just like with normal video streams
[22:21:18 CET] <wm4> as far a sI know, libavformat will even "inject" the attached picture packet so you can read it with av_read_frame
[22:21:28 CET] <wm4> but you can just take the picture directly as well
[22:21:39 CET] <wm4> and yes, this means the attached picture can't be animated or anything
[22:21:55 CET] <wm4> (not sure what happens if the attachment is an animated gif?)
[22:28:58 CET] <ubitux> michaelni: i tried "ffmpeg -nostats -f lavfi -i testsrc -f null -" with both xterm and urxvt; nothing is printed when i type things after 'c'
[22:29:39 CET] <ubitux> it seems to take the input though (because commands work when i pressed enter)
[22:30:32 CET] <michaelni> ubitux, was that different in the past ?
[22:31:19 CET] <ubitux> my memory is defective, but i don't remember having that issue last time i tried it (a while ago now)
[22:32:18 CET] <michaelni> i tried with some old binaries and none echos the entered text, so if that worked previously iam curious what broke it
[22:32:42 CET] <ubitux> oh, so it doesn't echo on your side either?
[22:32:47 CET] <ubitux> that's expected?
[22:33:33 CET] <michaelni> no echo here no, expected, i dont remember for sure, i havnt used it for a while so i am not sure if it echoed previously
[22:33:48 CET] <ubitux> ah, alright, well i guess that's normal then
[22:34:10 CET] <michaelni> it would be nice if it did echo though ...
[22:34:35 CET] <ubitux> tty.c_lflag &= ~(ECHO|ECHONL|ICANON|IEXTEN);
[22:34:45 CET] <ubitux> i guess it's because of this ^
[22:35:15 CET] <ubitux> maybe the echo flag should be re-set temporary while inside the command prompt
[22:36:01 CET] <michaelni> if that works and has no sideeffects, then id say yes
[22:37:13 CET] <andrewrk> wm4, thanks
[22:48:34 CET] <ubitux> michaelni: http://pastie.org/pastes/10568470/text this apparently works for me
[22:49:35 CET] <ubitux> the handling of ^M is ugly though, we might want to add least fprintf \n
[22:57:57 CET] <ubitux> anyway, patch on ml
[23:48:24 CET] <kierank> grrr avdev hard drive is dying
[23:50:25 CET] <J_Darnley> ;_; #PrayForAVDev
[00:00:00 CET] --- Fri Nov 20 2015


More information about the Ffmpeg-devel-irc mailing list