[Ffmpeg-devel-irc] ffmpeg-devel.log.20141221

burek burek021 at gmail.com
Mon Dec 22 02:05:03 CET 2014


[00:46] <tim_g> cd ffmpeg
[00:46] <tim_g> wrong tab whoops
[01:02] <iive> at least you didn't type your password :)
[01:03] <wm4> what if this ws his password?
[01:14] <cone-170> ffmpeg.git 03Michael Niedermayer 07master:7cdce8cb55e7: avfilter/vf_boxblur: factor rounder offset out
[01:14] <cone-170> ffmpeg.git 03Michael Niedermayer 07master:f622ff16cde3: avfilter/vf_boxblur: avoid one addition per line
[01:22] <pross> at least nothing incriminating..
[01:35] <tim_g> pross: lucky me
[02:03] <cone-170> ffmpeg.git 03Michael Niedermayer 07master:6706a2986c48: avfilter/vf_spp: Fix overflow in 8bit store slice
[04:15] <cone-170> ffmpeg.git 03Diego Biurrun 07release/1.1:9920a8a83f57: doc: Fix syntax and logical errors in avconv stream combination example
[04:15] <cone-170> ffmpeg.git 03Diego Biurrun 07release/1.1:2cd5640bce0e: Prepare for 9.17 release
[04:15] <cone-170> ffmpeg.git 03Diego Biurrun 07release/1.1:606bbd50b1eb: Update Changelog for v9.17
[04:15] <cone-170> ffmpeg.git 03Katerina Barone-Adesi 07release/1.1:56ac2cbd0464: apetag: Fix APE tag size check
[04:15] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:2b8effcda3f9: Merge commit '9920a8a83f578821e00212da6b79c8997a725fa8' into release/1.1
[04:15] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:8a9feeff9e31: Merge commit '2cd5640bce0ee8f515ff05040ce1c18d6af2694c' into release/1.1
[04:15] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:c7b8f02940ec: Merge commit '606bbd50b1ebaa3c040fc7cab84f1d8bb5d3d248' into release/1.1
[04:15] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:8eaecab5f401: Merge commit '56ac2cbd0464e0146e62c91843e2b1f5e0908504' into release/1.1
[04:27] <cone-170> ffmpeg.git 03Diego Biurrun 07release/1.1:489c180451b1: Add some bug references to the changelog
[04:27] <cone-170> ffmpeg.git 03Anton Khirnov 07release/1.1:ce424e57d629: mp3enc: fix a triggerable assert
[04:28] <cone-170> ffmpeg.git 03Anton Khirnov 07release/1.1:93f919d0b4c4: mov: avoid a memleak when multiple stss boxes are present
[04:28] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:74e52afe8f8a: Merge commit '489c180451b1cde469f6aeea44f8bdb11cd941d0' into release/1.1
[04:28] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:7a6907fcc334: Merge commit 'ce424e57d629072652935501b439c1a6310c2089' into release/1.1
[04:28] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:5e16809d493c: Merge commit '93f919d0b4c4341ccee366c98ac9af813f8fe622' into release/1.1
[04:34] <cone-170> ffmpeg.git 03Anton Khirnov 07release/1.1:50cb695bf124: jvdec: check frame dimensions
[04:34] <cone-170> ffmpeg.git 03Anton Khirnov 07release/1.1:03dba25a4001: mmvideo: check frame dimensions
[04:35] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:6e5dd43bf909: Merge commit '50cb695bf124b0bd4d9e2b3c1bfdd08b35b14438' into release/1.1
[04:35] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:9bbe24d455a3: Merge commit '03dba25a4001495226651068232b4c6b1e75fd02' into release/1.1
[04:40] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:02de44073a8e: gifdec: refactor interleave end handling
[04:40] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:306ee9508824: smc: fix the bounds check
[04:40] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:0ee7505e63b0: Merge commit '02de44073a8e116ea177b53081219d32ef135ad8' into release/1.1
[04:40] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:c633427e927e: Merge commit '306ee95088243fefa2dfcb5c355d439db75e2d2a' into release/1.1
[04:50] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:8df5b0250d10: avformat/swfdec: Use side data to communicate w/h changes to the decoder
[04:50] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:857eaa45ba94: avformat/swfdec: Do not change the pixel format
[04:50] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:a2933aa59a46: avcodec/mpegvideo: Use "goto fail" for all error paths in ff_mpv_common_frame_size_change()
[04:50] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:91c3a829a4ed: avcodec/mpegvideo: check that the context is initialized in ff_mpv_common_frame_size_change()
[04:50] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:66b1f5894a1d: avcodec/mpegvideo: Set err on failure in ff_mpv_common_frame_size_change()
[04:50] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:a76198ed288b: avformat/m4vdec: Check for non startcode 00 00 00 sequences in probe
[04:50] <cone-170> ffmpeg.git 03lvqcl 07release/1.1:1eb59e4d5402: avutil/x86/cpu: fix cpuid sub-leaf selection
[04:50] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:135a09c46ee6: avcodec/ac3enc_template: fix out of array read
[04:50] <cone-170> ffmpeg.git 03Reimar Döffinger 07release/1.1:0872359f9fff: configure: add noexecstack to linker options if supported.
[04:50] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:d44d87e17264: avcodec/jpeglsdec: Check run value more completely in ls_decode_line()
[04:50] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:29abba35723a: avcodec/mjpegdec: check bits per pixel for changes similar to dimensions
[04:50] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:37800352bba2: avcodec/utils: Add case for jv to avcodec_align_dimensions2()
[04:50] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:d68ae96ea8d8: avcodec/mmvideo: Bounds check 2nd line of HHV Intra blocks
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:aa17021e812e: avcodec/tiff: more completely check bpp/bppcount
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:0b99a18ebc41: avcodec/pngdec: Check bits per pixel before setting monoblack pixel format
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:d37e539d5ef0: avcodec/pngdec: Calculate MPNG bytewidth more defensively
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:86e57695257f: avcodec/qpeg: fix off by 1 error in MV bounds check
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:5e8b8e4b9d15: avformat/mpegts: Check desc_len / get8() return code
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:f6cfd160ca45: avcodec/h264: Check mode before considering mixed mode intra prediction
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:03ea3982275e: postproc/postprocess: fix quant store for fq mode
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:26dddec4665a: postproc: fix qp count
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:251d66d746ca: avcodec/dxa: check dimensions
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:e43872c3a93c: avcodec/dnxhddec: treat pix_fmt like width/height
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:961bbb98cf14: avcodec/utils: Align dimensions by at least their chroma sub-sampling factors.
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:6c5a57db94b5: avcodec/svq1dec: zero terminate embedded message before printing
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:3535f0867e41: avcodec/h264_slice: Clear table pointers to avoid stale pointers
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:ce35c10b88d1: avcodec/options_table fix min of audio channels and sample rate
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:c1e048dffea1: avcodec/utvideodec: fix assumtation that slice_height >= 1
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:b4f08f34a852: avcodec/wmaprodec: Fix integer overflow in sfb_offsets initialization
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:07b98ea3963f: avcodec/mjpegdec: Fix context fields becoming inconsistent
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:46dfd59ae199: avcodec/utils: Check that the data is complete in avpriv_bprint_to_extradata()
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:2fcfcd5c8787: avcodec/flacdec: Call ff_flacdsp_init() unconditionally
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:33d74e62c25a: avcodec/pngdec: Check IHDR/IDAT order
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:ed75a9a60ecb: avformat/rmdec: Check codec_data_size
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:b3e98715c0e5: swscale/x86/rgb2rgb_template: fix crash with tiny size and nv12 output
[04:51] <cone-170> ffmpeg.git 03wm4 07release/1.1:437e50b3d65f: avformat/matroskadec: fix handling of recursive SeekHead elements
[04:51] <cone-170> ffmpeg.git 03Rob Sykes 07release/1.1:7520f1fb3aca: swresample/soxr_resample: fix error handling
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:9189ae8ff4b7: avformat/aviobuf: Check that avio_seek() target is non negative
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:24fefa73dba4: avcodec/vmdvideo: Check len before using it in method 3
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:b9d2c2e98701: avcodec/utvideodec: Fix handling of slice_height=0
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:2c5898c1d303: avformat/mov: check atom nesting depth
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:dc1cf5febb7f: swscale: increase yuv2rgb table headroom
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:e78e2dc6e15d: avcodec/indeo3: use signed variables to avoid underflow
[04:51] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:45509d0730e3: avcodec/indeo3: ensure offsets are non negative
[05:14] <cone-170> ffmpeg.git 03Michael Niedermayer 07release/1.1:c5f52a9440e7: Update for 1.1.15
[13:20] <cone-666> ffmpeg.git 03Michael Niedermayer 07master:34f103b17115: avfilter/vf_boxblur: add gbrp pixfmt
[14:45] <ubitux> dammit this subtitles text stuff drive me crazy :P
[14:45] <ubitux> there are so much to change at once...
[14:46] <ubitux> it seems i'm really going to make an AVSubtitle2 but... :(
[14:49] <Compn> ubitux : sounds like a huge redesign :\
[14:49] <ubitux> i don't have much choice
[14:49] <Compn> char detection, probing, character encodings, demuxers, decoders, muxers, sub encoders, intermediate formats, native formats, 
[14:50] <ubitux> my main concern is that i'll have to think about the bitmap subs
[14:50] <ubitux> which i'm not comfortable with
[14:50] <Compn> closed captions, bitmap subs, ass subs
[14:50] <ubitux> so i will either have to copy what we currently have, or overthink it
[14:50] <ubitux> Compn: you forget mixed bitmaps & text
[14:50] <Compn> i dont want to think of that :)
[14:51] <ubitux> anyway, there is also the problem we have with dvb subs
[14:51] <Compn> gotta run, bbl
[14:51] <Compn> oh teletext
[14:51] <ubitux> with 2 packets: one for the subs, and one to end the previous one
[14:51] <ubitux> http://git.videolan.org/?p=ffmpeg.git;a=blob;f=ffmpeg.c;h=9f29eac7b1673480771dc4d0b099515050f27167;hb=HEAD#l828
[14:51] <ubitux> this crap that's here forever
[14:52] <ubitux> but again yeah that's related to stuff i don't know yet properly
[14:52] <ubitux> i guess i'll have to dive into that first... this shit is going to take forever
[14:53] <ubitux> first, i think i'll end up putting AVFrames instead AVPicture for the bitmap ones, but people will hate me for that
[14:54] <ubitux> but that way, i'm pretty sure to be able to represent any insane picture types (such as PNG, GIF, or YUV based ones)
[15:03] <ubitux> anyway, does anyone know why we have a {start,end}_display_time?
[15:04] <ubitux> it sounds like it's useful to have all packets stuck together
[15:04] <ubitux> (typically to clear the previous one)
[15:05] <nevcairiel> those times are relative to the frames PTS, presumably because some codec codes them like this, ie. dvd uses them
[15:06] <ubitux> yes, but what's the motivation here?
[15:06] <ubitux> i mean, why do they like that?
[15:07] <nevcairiel> back then, the decoder didnt know the timebase of the pts it was handling, so it was unable to actually modify it to offset it by start_display_time, so it was stored in a separate field for the calling code to sort out, i imagine
[15:08] <nevcairiel> it was likely modeld after some subtitle codec, maybe dvd which has them coded just like this in the bitstream
[15:09] <nevcairiel> you need a pts and duration anyway, so the only discrepancy is the start time, which could be explained by the missing timebase information i mentioned
[15:10] <ubitux> but why would they ever need to have this information in the decoder?
[15:10] <ubitux> i mean, why didn't they just assume a start time of 0 all the time?
[15:10] <ubitux> i mean the decoder doesn't need to work with timestamps
[15:11] <ubitux> fuck i'd really love to just have access to the specs to understand this
[15:11] <nevcairiel> if you require strict interleaving on the pts, but you want the subtitles to arrive earlier than their video counter-parts, you could use such a mechanism to reduce the pts of the subtitle frame, but still show it a later time only
[15:13] <ubitux> nevcairiel: why would you have the subtitles earlier?
[15:14] <ubitux> i mean, decoding subtitles is certainly faster than decoding video, you wouldn't need such delay
[15:14] <ubitux> what am i missing?
[15:14] <nevcairiel> some playback chains have such requirements
[15:15] <ubitux> why? :p
[15:16] <nevcairiel> who knows, but they do
[15:16] <nevcairiel> bad design? what do i know :)
[15:20] <nevcairiel> i dont know if start_display_time is actually != 0 in any dvd media i've ever seen, but the format at least has this concept
[15:20] <iive> ubitux: e.g. dvb subtitles could be big and passed in a number of packets in PID that have lower bitrate
[15:37] <ubitux> wouldn't it make more sense to have that start/end time for each rect?
[15:37] <ubitux> iive: not sure what's your point
[15:38] <ubitux> iive: also, dvb doesn't actually uses start/end time apparently
[15:38] <ubitux> only dvd and xsub
[15:38] <iive> it does
[15:38] <ubitux> not our decoder then
[15:38] <ubitux> i mean, it uses the end time, yes
[15:39] <iive> well, it sends empty sub, i think... i don't remember.
[15:39] <ubitux> yes, which make sense assuming the next one is stuck to the current one
[15:39] <ubitux> (pts+duration)
[15:39] <iive> there is end time, that is more like an timeout, that is to hide the sub if nothing else comes (e.g. transmission problem)
[15:39] <ubitux> and you want the current to stop displaying text before the next one appears
[16:32] <cone-443> ffmpeg.git 03Michael Niedermayer 07master:c73b28892748: avfilter/vf_boxblur: Support 10bit planar formats
[17:15] <cone-443> ffmpeg.git 03Michael Niedermayer 07fatal: ambiguous argument 'refs/tags/n1.1.15': unknown revision or path not in the working tree.
[17:15] <cone-443> Use '--' to separate paths from revisions
[17:15] <cone-443> refs/tags/n1.1.15:HEAD: avfilter/vf_boxblur: Support 10bit planar formats
[19:33] <cone-443> ffmpeg.git 03Michael Niedermayer 07master:47b9481d9db8: avformat/mpegenc: Use av_freep() to avoid leaving stale pointers in memory
[19:34] <cone-443> ffmpeg.git 03Michael Niedermayer 07master:f7a4589b3670: avformat/mpegtsenc: Use av_freep() to avoid leaving stale pointers in memory
[21:30] <cone-443> ffmpeg.git 03Michael Niedermayer 07master:b7e506b3b9ca: avformat/swfdec: Check frame size rectangle in probe()
[21:48] <Daemon404> ...
[21:49] <Daemon404> "This is a (very) bad regression and we should do the same as for x264."
[21:49] <Daemon404> what fuckign drugs is carl on
[21:52] <iive> pkg-config again...
[21:53] <Daemon404> carl seems to have some sort of blood fued with pkg-config
[21:53] <Daemon404> despite it eign defined as the defacto and only way supported by the libraries' authors
[21:53] <Daemon404> being*
[21:55] <BtbN> Well, a hard dependency on pkg-config would be quite annoying, specialy when cross compiling.
[21:56] <JEEB> you get pc files even if you cross-compile, and you can use PKG_CONFIG_whateveritwas to tell pkg-config where to look for them
[21:57] <Compn> Daemon404 : iirc carl came from mplayer , you know mplayer hates pkg-config...
[21:57] <jamrial> who doesn't have pkg-config anyway? it's on every linux distro, and even on msys/cygwin environment
[21:57] <Compn> cant remember if carl started on ffmpeg or mplayer first
[21:58] <Daemon404> BtbN, there is absolutely nothing wrong with pkg-config for corss compiling
[21:58] <BtbN> It forces you to actualy install stuff, building from a local build tree gets very hard
[21:58] <Compn> probably pkg-config pulls in other deps...
[21:58] <Daemon404> BtbN, you should ALWAYS install to a sysroot
[21:58] <JEEB> BtbN, set the prefix
[21:58] <Daemon404> *always(
[21:58] <Daemon404> JEEB, he wants to not do make install to a sysroot
[21:59] <Daemon404> which si defacto wrong.
[21:59] <JEEB> lörs lärä
[21:59] <BtbN> Well, there are systems, like windows, which simply don't work like that.
[21:59] <Daemon404> mingw does work like that.
[21:59] <JEEB> but the build systems do :P
[21:59] <Compn> Daemon404 : you know the amount of people you say are wrong... is it possible that they just like doing things a different way and that you should respect that ?
[21:59] <Daemon404> Compn, no not really
[21:59] <Daemon404> in this case it is purley wrong
[21:59] <Compn> purley :)
[22:00] <Daemon404> purely even
[22:00] <Daemon404> the only system which doesnt work with a sysroot is msvc for arm afaik
[22:00] <Daemon404> which is a dead platform
[22:00] <Compn> winrt surface ?
[22:00] <ubitux> well, the main problem is that carl makes a lot of efforts into not understanding a stupid env variable
[22:00] Action: Compn dont remember arm
[22:01] <Daemon404> export PKG_CONFIG_LIBDIR is too hard?
[22:01] <Compn> ask to vnc to his build box and show him how to fix it ?
[22:01] <Compn> if cant explain in english...
[22:01] <BtbN> So far i never had issues building without pkg-config, by just installing in a prefix and then adding the lib/include dirs of that prefix to extra c/ldflags.
[22:01] <JEEB> yeah, sure
[22:01] <JEEB> that's correct as well
[22:01] <JEEB> pkg-config is just made to automatize that for you
[22:01] <Compn> time to reinvent pkg-config :D
[22:02] <Daemon404> or jus tadd the prefix to pkg configs path
[22:02] <BtbN> But i need pkg-config to use pkg-config.
[22:02] <Daemon404> no you dont
[22:02] <wm4> Compn> Daemon404 : iirc carl came from mplayer , you know mplayer hates pkg-config... <- the mplayer configure script is 10kloc, the mpv one derived from it (not used by default anymore) is 1kloc, and pkg-config helped a lot to reduce it
[22:02] <Daemon404> or well what teh fuck is teh big deal with installign pkg-config
[22:02] <iive> that's not the point here. There is absolutely no reason to break configure if pkg-config is missing
[22:02] <Daemon404> OH NO 400 KB
[22:02] <wm4> Daemon404: non-open source people hate it
[22:02] <JEEB> I'm not saying you shouldn't be able to set the CFLAGS/LDFLAGS yourself, of course
[22:02] <iive> when it have been working before.
[22:02] <wm4> Daemon404: windows people too
[22:02] <JEEB> I'm a windows person
[22:02] <JEEB> yet I don't hate pkg-config
[22:02] <JEEB> :V
[22:03] <Compn> wm4 : why not autoconf too ?
[22:03] <wm4> cmake, the "awesome" build system can't even use pkg-config correctly
[22:03] <Compn> get rid of mplayer configure entirely
[22:03] <Daemon404> wm4, windows people also embed entire source trees
[22:03] <Daemon404> because msvc.
[22:03] <ubitux> iive: warn "$pkg_config not found, library detection may fail."
[22:03] <BtbN> wm4, hm? You cmake uses pkg-config as additional hint for its own library finding.
[22:03] <BtbN> *usualy
[22:03] <ubitux> ’ no pkg-config = shitstorm => expected behaviour :)
[22:03] <Daemon404> i absolutely love how api break -> Sure fine deprecate it
[22:03] <Daemon404> configure "break" -> carl loses his shit
[22:04] <iive> ubitux: was it working before?
[22:04] <Compn> carl fixes configure a lot...
[22:04] <Daemon404> "fixes"
[22:04] <Daemon404> mostly i see hacks
[22:04] <Compn> and tests.
[22:04] <ubitux> iive: "maybe" :)
[22:04] <Compn> and looks at fate logs (i assume)
[22:05] <iive> well, hacks is what ffmpeg does.
[22:05] <Compn> Daemon404 : i'd actually like to see the list of devs you dont hate in ffmpeg :D 
[22:05] <ubitux> iive: like, it was working if you were lucky that "-lspeex" was enough
[22:05] <Compn> wonder how short the list would be :)
[22:05] <Daemon404> Compn, ubitux is my soul mate
[22:05] <ubitux> :o
[22:06] <iive> ubitux: why it doesn't work now? Let me guess... you broke it?
[22:06] <ubitux> i didn't change anything, but depending on the libspeex you have it might not be enough
[22:06] <ubitux> and you can't tell without looking at the .pc file
[22:06] <Daemon404> it's almost as if libspeex authors knew this and told people to use pkg-config
[22:07] <ubitux> we should encourage people from other projects to randomize the name of the libraries
[22:07] <ubitux> to force the use of pkg-config
[22:08] <Daemon404> e.g. for x265, it is absolutely required to use pkg-config
[22:08] <Daemon404> otherwise you cant knwo which c++ rt to use
[22:08] <ubitux> i mean, even the idiots from cmake are starting to rely of pkg-config
[22:09] <BtbN> Allmost every cmake find module uses pkg-config.
[22:09] <BtbN> Just as additional hints, but it does use it, when available.
[22:09] <ubitux> BtbN: not all of them, but yeah
[22:09] <ubitux> there are too many of them that are totally broken because they don't use pkg-config
[22:10] <ubitux> anyway...
[22:10] <ubitux> [~]- pkg-config --libs --cflags --static x265
[22:10] <ubitux> -lx265 -lstdc++ -lm -lrt 
[22:10] <ubitux> [~]- pkg-config --libs --cflags --static sdl
[22:10] <ubitux> -D_GNU_SOURCE=1 -D_REENTRANT -I/usr/include/SDL -lSDL -lpthread -lSDL -lpthread -lm -ldl -lpthread 
[22:10] <ubitux> this is indeed typically why you use pkg-config for linking libraries
[22:10] <Compn> bi-yearly pkg-config troll session
[22:10] <ubitux> [~]- pkg-config --libs --cflags --static x264
[22:10] <ubitux> -lx264 -lpthread -lm -ldl 
[22:10] <ubitux> or even this
[22:10] <ubitux> with speex you're lucky though
[22:10] <Daemon404> wtf does x264 use libdl for
[22:10] <JEEB> opencl
[22:10] <Daemon404> olol
[22:10] <Daemon404> how "useful"
[22:11] <ubitux> i guess it depends on your build
[22:11] <Daemon404> i always hard disable opencl
[22:11] <BtbN> Would you prefer it linking against it, so you have to have it installed?
[22:11] <Daemon404> probably why i never saw it
[22:11] <ubitux> ...which is why you have a .pc file; you don't want to have all this crap linked against if your x264 doesn't have this
[22:11] <BtbN> You don't realy need to link against it. Only if you use static libs.
[22:11] <ubitux> (in the case of x264, it might not make a difference though)
[22:12] <ubitux> BtbN: which is what many people do :)
[22:12] <Daemon404> there's even a configure flag to do this iirc
[22:12] <ubitux> yes
[22:12] <ubitux> it's not lib-specific, but there is one indeed
[22:13] <BtbN> static libs on linux are extremely annoying. And there is like a dozen diffrent tools which try to deal with them.
[22:13] <Daemon404> which is why prefixes and sysroots with pkg-config are your freind.
[22:13] <Daemon404> friend*
[22:13] Action: Daemon404 has spent extensive time working on cross-compilign envs
[22:13] <BtbN> Which is why i don't link stuff staticaly.
[22:14] <Daemon404> "dont do X" is never a solution, usually.
[22:18] <BtbN> pkg-config should have flags for that
[22:19] <BtbN> So it only shows those additional libs when you are actualy linking staticaly
[22:31] <Daemon404> BtbN, yes, and it does
[22:31] <Daemon404> and out configure handles it as such.
[22:31] <Daemon404> our*
[22:38] <wm4> BtbN: it ignores several pkg-config aspects
[22:38] <wm4> even those which just work with gcc file.c `pkg-config --libs --cflags package`
[22:39] <j-b> what is nnedi3
[22:40] <Daemon404> j-b, upscale and deint
[22:40] <Daemon404> why
[22:40] <j-b> what's the use ?
[22:41] <j-b> just curious, mostly
[22:41] <j-b> still on the same quest as ever
[22:41] <Daemon404> people use it in deint scripts, such as QTGMC
[22:42] <Daemon404> madvr allows it for playback upscsaling
[22:42] <Daemon404> (e.g. fullscreen)
[22:46] <nevcairiel> its just a fancy upscaler, well, a bit more limited, it can only double the resolution, not scale to arbitrary sizes
[22:46] <nevcairiel> but it does look nice on SD anime content
[22:46] <nevcairiel> live-action content benefits a bit less
[22:47] <Daemon404> ~neural networks~
[22:47] <iive> ENHANCE!
[22:49] <nevcairiel> it might be nice once 4K screens become a bit more common, as 1080px2 is a common use-case then :d
[22:50] <Daemon404> nevcairiel, yeah currnetly my zenbook is in a weird place wrt upscaling
[23:16] <kierank> Daemon404: is the zenbook any good?
[23:17] <kierank> I need a new laptop ideally with a i7-4700MQ and < £1k
[23:17] <j-b> ux303
[23:17] <nevcairiel> i have a previous gen zenbook, the UX31A, and I love it
[23:17] <Daemon404> kierank, i dont recall it being <1k
[23:17] <Daemon404> it might be
[23:18] <Daemon404> and yeah i love it except for the weird-ass power key placement
[23:18] <kierank> j-b: Core" i7 4510U
[23:18] <kierank> only dual core
[23:18] <Daemon404> i7-4558U
[23:18] <Daemon404> on mine
[23:19] <kierank> might end up getting http://www.hotukdeals.com/deals/hp-envy-touchsmart-15-j144na-laptop-intel-core-i7-16gb-ram-1tb-15-6-touch-screen-2065160
[23:19] <kierank> in the sales
[23:19] <kierank> assuming they drop the price to 699 post-christmas
[23:20] <nevcairiel> i dont think they have zenbooks with quad cores, its too low power profile for that
[23:20] <nevcairiel> dual with HT is the best it offers
[23:30] <kierank> ok
[23:34] <cone-443> ffmpeg.git 03Michael Niedermayer 07master:41ee459e8809: avformat/img2dec: check w/h in dpx_probe
[00:00] --- Mon Dec 22 2014


More information about the Ffmpeg-devel-irc mailing list