[Ffmpeg-devel-irc] ffmpeg-devel.log.20140603

burek burek021 at gmail.com
Wed Jun 4 02:05:02 CEST 2014


[00:13] <cone-105> ffmpeg.git 03Michael Niedermayer 07fatal: ambiguous argument 'refs/tags/n2.2.3': unknown revision or path not in the working tree.
[00:13] <cone-105> Use '--' to separate paths from revisions
[00:13] <cone-105> refs/tags/n2.2.3:HEAD: avformat/utils/avformat_find_stream_info: Move max_analyze_duration setup out of the packet loop
[00:22] <llogan> kierank: RE: swift..."yes, when we get accepted to GSoC".
[01:01] <cone-105> ffmpeg.git 03Diego Biurrun 07release/0.10:7b00340f97f2: configure: Support preprocessor macros as header names
[01:01] <cone-105> ffmpeg.git 03Luca Barbato 07release/0.10:ec772cca6042: drawtext: Drop pointless header
[01:02] <cone-105> ffmpeg.git 03Luca Barbato 07release/0.10:bf0cb89a8d11: configure: Update freetype check to follow upstream
[01:02] <cone-105> ffmpeg.git 03Reinhard Tartler 07release/0.10:aedf1a2996e7: Update Changelog for 0.8.11
[01:02] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:e771425e985e: Merge commit 'ec772cca60423b9994fe00c7cef239f93eae6112' into release/0.10
[01:02] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:2f894deeda18: Merge commit 'aedf1a2996e70d322220d2704d5a615c4f7b1b23' into release/0.10
[01:21] <cone-105> ffmpeg.git 03Lou Logan 07master:c17973f19a33: doc/bsf: document noise bitstream filter
[01:26] <cone-105> ffmpeg.git 03Martin Storsjö 07release/0.10:f1a8885ae9e7: doc: Point to the correct, actually maintained gas-preprocessor repo
[01:26] <cone-105> ffmpeg.git 03Keiji Costantini 07release/0.10:16f0f97eec79: ituh263: reject b-frame with pp_time = 0
[01:26] <cone-105> ffmpeg.git 03Johan Andersson 07release/0.10:ecf21ab0ec79: cmdutils: update copyright year to 2014.
[01:26] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:53dcd772aa1f: Merge commit 'f1a8885ae9e7f281b597d1f9188fc16fb2ab7832' into release/0.10
[01:26] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:bca2ebbeeeac: doc: switch github urls to https
[01:26] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:2b40afae7004: Merge commit 'ecf21ab0ec798eea96f862333568336098b4610c' into release/0.10
[01:30] <kierank> llogan: it was more of a joke about the people on hacker news who wanted ffmpeg to be written in haskell or rust or some language
[01:32] <Daemon404> plebs
[01:32] <Daemon404> lisp master race
[01:32] <llogan> kierank: i was going to make a "go" joke, but i couldn't make it funny.
[01:33] Action: Daemon404 calls libav* from Go code ;)
[01:33] Action: Daemon404 runs
[01:33] <kierank> llogan: you should announce that ffmpeg has moved to swift
[01:33] <J_Darnley> Yeah!
[01:33] <J_Darnley> Rewrite in a memory safe language!
[01:33] <J_Darnley> Javascript!
[01:33] <J_Darnley> Everyvalue is a float!
[01:33] <Daemon404> WRONG
[01:33] <Daemon404> a double.
[01:33] <llogan> if ; are optional in swift can i do ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;?
[01:34] Action: J_Darnley commit suicide in shame
[01:34] <kierank> 1=1 = false
[01:34] <kierank> bah
[01:34] <kierank> 1==1 = false
[01:34] <kierank> in js
[01:35] <J_Darnley> Who needs threads?  Who needs SIMD?
[01:35] <Daemon404> kierank, you forgot ===
[01:35] <Daemon404> which is a thing
[01:35] <kierank> it's a thing in python
[01:36] <llogan> i'm a boring twitterer
[01:36] <kierank> llogan: tweet "Breaking: FFmpeg developers to rewrite FFmpeg in swift"
[01:37] <J_Darnley> Don't you mean "#swift"?
[01:37] <Daemon404> #yolo
[01:37] <cone-105> ffmpeg.git 03Mans Rullgard 07release/0.10:afab4c422b60: configure: use utilities from /usr/xpg4/bin if it exists
[01:37] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:2cbc8dfeddcb: h264: check buffer size before accessing it
[01:37] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:e0d8a17402b9: h264_refs: make sure not to write over the bounds of the default ref list
[01:37] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:3ee26080d6b3: h264: reset data_partitioning if decoding the slice header for NAL_DPA fails
[01:37] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:ef076d3bd65c: Merge commit '2cbc8dfeddcbe00ef5f112162912bb49c1dd6623' into release/0.10
[01:37] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:1ab94468dd99: Merge commit 'e0d8a17402b934b8fba7b86c6c990abf1257901b' into release/0.10
[01:37] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:f2275400fe52: Merge commit '3ee26080d6b3e777992b4b4124e62e1bf0ac0a65' into release/0.10
[01:38] <Daemon404> in theory the twitter account could be used to troll
[01:38] <Daemon404> @company y u no gpl comply?
[01:39] <llogan> "FFmpeg 1000000000 is released!"
[01:39] <kierank> troll elemental
[01:40] <llogan> i don't know enough emojis for swift which i hear is a requirement
[01:41] <Daemon404> http://www.iemoji.com/
[01:41] <Daemon404> sorry about the eye cancer
[01:42] <J_Darnley> Thank god for my paranoid browser!
[01:45] <Daemon404> i started drawing an 8x8 zigzgag using the arrow emjoi
[01:45] <Daemon404> but then i got bored
[01:46] <J_Darnley> :)
[01:46] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:0f71a5df4bc9: h264: do not use 422 functions for monochrome
[01:46] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:7f33a24e824c: h264: check that execute_decode_slices() is not called too many times
[01:46] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:c4033cd4eb92: h264: reject mismatching luma/chroma bit depths during sps parsing
[01:46] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:97a5b4ba53a1: Merge commit '0f71a5df4bc913f17a53c7ac66d4957914fa1d3d' into release/0.10
[01:46] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:69db5f2779f8: Merge commit '7f33a24e824c6d20cb941e6b20c5382becfbc923' into release/0.10
[01:46] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:1e83191d5c83: Merge commit 'c4033cd4eb921a9cc8deb513efc6d6a6ba3b2163' into release/0.10
[01:49] <kierank> smarter: ping
[01:49] <smarter> kierank: pong
[01:50] <kierank> smarter: I have a 4K file (can't share) that fails to play unless I comment out this block: http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavcodec/hevc_ps.c;h=647f83eec6d84fc8949d50deb107ffc48ba42b79;hb=HEAD#l378
[01:50] <smarter> hmmf
[01:51] <smarter> I guess the decoder could be more tolerant, but that's not a bug is it?
[01:51] <kierank> afaik no
[01:51] <smarter> was that file produced by the HM?
[01:51] <kierank> no
[01:52] <kierank> i'm not sure what produced it
[01:54] <smarter> kierank: feel free to send a patch to both projects to condition the goto err on "if (s->avctx->err_recognition & AV_EF_EXPLODE)", that shouldn't break anything anyway
[01:54] <smarter> otherwise I'll take a look once I get the time to do so
[01:54] <kierank> yeah i will do at some point
[01:54] <kierank> thanks
[01:55] <kierank> it's annoying because i can't open the file in any analyser as a result
[01:56] <smarter> your analyzer should have a .so of fmpeg you can replace, shouldn't it?
[01:56] <kierank> not sure about that
[01:57] <kierank> I need to dig into the code a bit as well
[01:57] <kierank> I want to find out if the stream uses wpp
[01:57] <smarter> you can always hack the bitstream to fix that I guess :p
[01:59] <kierank> yeah I guess
[02:11] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:51ae8e26af8f: h264: limit allowed pred modes in ff_h264_check_intra_pred_mode() to 3
[02:11] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:a7cce9ebf3ae: h264: reset first_field if frame_start() fails for missing refs
[02:11] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:35ba079fbf28: h264: reset ref count if decoding the slice header fails
[02:11] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:27ac9585c97d: h264: reset data partitioning at the beginning of each decode call
[02:11] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:ced5c8e6b36f: Merge commit '51ae8e26af8f5b26efb41edc0fe4812368d16ae9' into release/0.10
[02:11] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:0558832796b0: Merge commit 'a7cce9ebf3ae3b9678970236c964900393603a73' into release/0.10
[02:11] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:8d3273562932: Merge commit '35ba079fbf281a066f3ac1e1271f3caa402dcd74' into release/0.10
[02:11] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:d076ee521647: Merge commit '27ac9585c97d35b809382be5634c8e5f7211243a' into release/0.10
[02:13] <jamrial> https://pbs.twimg.com/media/BpETkwAIYAALPUR.png
[02:17] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:07558d0b9fa0: mpegvideo: set reference/pict_type on generated reference frames
[02:17] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:d0ecfe32492b: adx: check that the offset is not negative
[02:17] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:71b8c8430cf3: sgidec: fix buffer size check in expand_rle_row()
[02:17] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:7c520e5cd61a: Merge commit '71b8c8430cf3f7056849257324fc39b423075ba1' into release/0.10
[02:17] <llogan> jamrial: it's missing "please include your command and complete console output" = "We are not psychic"
[02:22] <cone-105> ffmpeg.git 03Janne Grunau 07release/0.10:9938e450c857: mpeg12: check scantable indices in all decode_block functions
[02:22] <cone-105> ffmpeg.git 03Luca Barbato 07release/0.10:1dce4a031f8d: avfilter: Add missing emms_c when needed
[02:22] <cone-105> ffmpeg.git 03John Stebbins 07release/0.10:330c18032464: matroskaenc: allow override of "writing application" tag
[02:22] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:fe30351d528f: Merge commit '330c18032464a4e1f3da907e54db2e69a6fbfcda' into release/0.10
[02:32] <cone-105> ffmpeg.git 03John Stebbins 07release/0.10:db52f056c3d2: movenc: allow override of "writing application" tag
[02:32] <cone-105> ffmpeg.git 03Paul B Mahol 07release/0.10:b473fdcde329: bytestream: add functions for accessing size of buffer
[02:32] <cone-105> ffmpeg.git 03Justin Ruggles 07release/0.10:fa60904ebd58: bytestream: add bytestream2_copy_buffer() functions
[02:32] <cone-105> ffmpeg.git 03Justin Ruggles 07release/0.10:a0a90b1a1116: tiffdec: use bytestream2 to simplify overread/overwrite protection
[02:32] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:b827189c6f8a: Merge commit 'fa60904ebd58da33abf10b05e9933d24619cf096' into release/0.10
[02:32] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:c437ab3c4ea2: Merge commit 'a0a90b1a1116250a2494021da810cc5da89ea36f' into release/0.10
[02:37] <UtUser> I like the part about legacy code
[02:41] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:42dcfe32a867: avi: Improve non-interleaved detection
[02:41] <cone-105> ffmpeg.git 03Anton Khirnov 07release/0.10:079758e49a4d: h264: reset next_output_pic earlier in start_frame()
[02:41] <cone-105> ffmpeg.git 03Baptiste Coudurier 07release/0.10:5463a2b0566b: movdec: handle 0x7fff langcode as macintosh per the specs
[02:41] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:18b233428a98: Merge commit '079758e49a4d6b3e7cf2e22bed71d34c46712242' into release/0.10
[02:41] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:0be41c65862c: Merge commit '5463a2b0566b34b9e3847db9ceb1ef1d2a6004fc' into release/0.10
[02:49] <cone-105> ffmpeg.git 03Mark Himsley 07release/0.10:b7b798a1afb1: isom: lpcm in mov default to big endian
[02:49] <cone-105> ffmpeg.git 03Rafaël Carré 07release/0.10:65c3593792a9: apedec: do not buffer decoded samples over AVPackets
[02:49] <cone-105> ffmpeg.git 03Martin Storsjö 07release/0.10:110680c5a209: alac: Limit max_samples_per_frame
[02:49] <cone-105> ffmpeg.git 03Luca Barbato 07release/0.10:7fa727002981: swscale: Fix an undefined behaviour
[02:49] <cone-105> ffmpeg.git 03Janne Grunau 07release/0.10:6f4404b24bcf: h264: set parameters from SPS whenever it changes
[02:49] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:d499685d8020: Merge commit '7fa72700298107fe756311ecb4dee5270ff12d35' into release/0.10
[02:49] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:952738d0af74: Merge commit '6f4404b24bcf59ab29cd4b57995d374a578f51a7' into release/0.10
[02:54] <cone-105> ffmpeg.git 03Reinhard Tartler 07release/0.10:516ea2dccd1b: Prepare for 0.8.12 Release
[02:54] <cone-105> ffmpeg.git 03Sean McGovern 07release/0.10:d75b149757fc: Update Changelog for 0.8.12
[02:54] <cone-105> ffmpeg.git 03Reinhard Tartler 07release/0.10:9552b37e2604: Add some bug references
[02:54] <cone-105> ffmpeg.git 03Michael Niedermayer 07release/0.10:20883401735a: Merge commit '9552b37e2604552d5ff210175d6baf28ccc2bb80' into release/0.10
[03:11] <cone-105> ffmpeg.git 03Carl Eugen Hoyos 07master:763e714442e0: Do not limit blocksize when reading PCM from aiff.
[04:36] <Rafajafar_> would this be a good place to talk about compression
[04:37] <Rafajafar_> Yes I know it's probably eye rolling b/c of Silicon Valley, but I actually want to learn more
[04:37] <Rafajafar_> I have this video similarity tech, and I think ... maybe... it can be used for lossless compression pretty well, but I don't want to spend time on a fool's errand
[04:38] <Rafajafar_> In case you're curious what that's all about: https://www.youtube.com/watch?v=EFoMTaDab9U
[04:43] <Rafajafar_> specifically I'm trying to figure out how I-Frames are determined
[05:01] <Rafajafar_> alrighty maybe it's a bit late for that question
[05:01] <Rafajafar_> I'll pop on tomorrow and see if someone knows
[05:21] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:4e3fe65610c9: avcodec/utils/ff_init_buffer_info: Favor color information from AVFrame if available
[11:00] <cone-105> ffmpeg.git 03Carl Eugen Hoyos 07master:ad7de82218e9: Fix fate-opus and fate-aac-ln-encode with --target-path.
[15:31] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:72e4ceeb7977: avformat/asfdec: Use a parser for mpeg4-asp to find has_b_frames
[15:31] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:a0b2de77ed30: ffmpeg: remove "has_b_frames" backpassing
[15:31] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:768ef0f19e4f: ffmpeg: request sample if the decoder & demuxer has_b_frames mismatch
[15:37] Action: Daemon404 wonders what there could be to do possibly in dublin on the friday
[15:37] <Daemon404> only thing i know of is guiness
[15:59] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:3e7b4b9fbd46: Add direct link to ML in missing_feature_sample()
[16:12] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:2422a7ffe441: Loose mplayer in MPlayer/incoming/
[16:13] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:576654302ba3: ffmpeg: use av_log_ask_for_sample() instead of duplicating avpriv_request_sample()
[17:30] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:90b26d63ba0e: avcodec/vp3: factorize out free_tables()
[17:30] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:a4ed51ab4087: avcodec/vp3: free tables before allocating new ones
[18:03] <cone-105> ffmpeg.git 03Janne Grunau 07master:fd2981ea92d9: aarch64: add darwin style PAGE/PAGEOFF relocations
[18:03] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:ab9afcdf0483: Merge commit 'fd2981ea92d9a776fcb1a13377dce1c8a7db7b5e'
[18:08] <cone-105> ffmpeg.git 03Janne Grunau 07master:6a0fa4d86f2b: aarch64: remove optional :pg_hi21: for adrp instruction
[18:08] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:40285d265906: Merge commit '6a0fa4d86f2b3e9060a1153b39fa3bfe36f1b617'
[18:14] <cone-105> ffmpeg.git 03Janne Grunau 07master:68a06b3a639e: aarch64: use '#' for whole line asm comments
[18:14] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:bcbd7dbce546: Merge commit '68a06b3a639ee21c78532ee4c50c3366bf890ff7'
[18:41] <cone-105> ffmpeg.git 03Janne Grunau 07master:896a5bff6426: arm: check if AS supports .dn
[18:41] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:83e8650f7703: Merge commit '896a5bff64264f4d01ed98eacc97a67260c1e17e'
[18:47] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:8d4d12a490b1: doc/muxers: Document problems of omitting nut syncpoints as well as alternatives
[19:56] <llogan> yet another user wanting LFE in their stereo downmix... Dolby Digital decoder discards LFE, but what are "the specs" that indicate that it should be dropped?
[19:59] <JEEB> I remember the specs being linked around the LAV Filters thread @ D9
[19:59] <JEEB> but yes, many users want it there for one reason or another
[20:03] <llogan> JEEB: thanks, but i'll probably be too lazy to root around for the files.
[20:07] <ruggles> llogan: specs definitely do recommend 0 for the default lfe downmix level
[20:08] <llogan> ruggles: which specs should i refer to in particular?
[20:09] <ruggles> A/52:2010
[20:09] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:137c5e115b0f: ffmpeg: try to use the more fitting of the 2 encoder contexts
[20:10] <ruggles> llogan: actually nevermind, i was wrong
[20:10] <ruggles> llogan: "Downmixing of the lfe channel is optional. An ideal downmix would have the lfe channel reproduce at an acoustic level of +10 dB with respect to the left and right channels. Since the inclusion of this channel is optional, any downmix coefficient may be used in practice. Care should be taken to assure that loudspeakers are not overdriven by the full scale low frequency content of the lfe channel."
[20:12] <llogan> interesting. is that from A/52:2010?
[20:12] <ruggles> yes
[20:17] <ruggles> llogan: you probably know this already, but fwiw both swresample and avresample have an lfe mix level option. only the in-decoder downmixer always uses 0.
[20:27] <llogan> ruggles: thanks. lfe_mix_level, right? some users are playing around with pan to try to do what they think they want.
[20:35] <ruggles> llogan: yeah
[21:02] <cone-105> ffmpeg.git 03Lou Logan 07master:88f2586adbcd: fix various typos
[21:23] <cone-105> ffmpeg.git 03Dave Rice 07master:d63d964fb3c9: avcodec/libopenjpegenc: Add frame-threading capability
[21:36] <kurosu> michaelni, as we already discussed, I don't think I'll continue with the hevc_mc improvements
[21:37] <kurosu> I'll reply on the mailing list in that regard, and that has nothing to do with the reviews (ie it's not any of us' fault)
[21:41] <Daemon404> fffs kurosu's email crahed my email clien
[21:41] <Daemon404> wtf
[21:41] <JEEB> :D
[21:41] <Daemon404> thunderbird is a piece of crap
[21:42] <kurosu> probably smashed your stack
[21:42] <Daemon404> yet i havent found anything i like better
[21:42] <kierank> outlook ftw
[21:44] <ubitux> kurosu: try pasting shellcode in your signature next time
[21:44] <ubitux> Daemon404 will probably enjoy
[21:45] <JEEB> :D
[21:45] <kurosu> each time I smash a stack, a canary dies
[21:45] <JEEB> kierank, I'm using outlook @ work, and it's quite nice indeed
[21:45] <kierank> google calendar still can't get british summer time right
[21:45] <JEEB> wow
[21:45] <JEEB> either your summer time is super special, or google is failing badly
[21:47] <kierank> It says GMT+00 London
[21:47] <kierank> which is wrong
[21:47] <kierank> we are in BST
[21:47] <kierank> GMT+01
[21:50] <JEEB> sounds like fun
[21:50] <JEEB> I guess it's at least showing the correct time
[21:51] <kierank> yes it's showing the  correct time but in the wrong timezone
[21:51] <JEEB> aye
[21:53] <UtUser> hi JEEB
[21:55] <JEEB> UtUser, my quota for multimedia hacking is up for tonight
[21:55] <UtUser> just saying hello ;)
[22:02] <iive> daylight saving is horrible
[22:04] <JEEB> yes, just like switching tires twice a year on cars because of hard winter
[22:04] <JEEB> I enjoyed my life in east japan nicely thanks to not having to deal with either vOv
[22:05] <Daemon404> without dailight savings time here
[22:05] <Daemon404> sun would rise at 3:30 am
[22:05] <Daemon404> on the 21st of june
[22:05] <Daemon404> :|
[22:07] <llogan> where?
[22:07] <Daemon404> england
[22:07] <iive> and why is that a problem?
[22:08] <Daemon404> because i dont like sun blazing at 3:30am when i want to sleep?
[22:08] <alpa_astero> what Daemon404 says
[22:08] <iive> or do you propose making DST 3 hours, so sun rises at 6:30 ?
[22:08] <JEEB> we're already at sunrise 0405, sunset 2233 here
[22:09] <Daemon404> the extra hour is nice
[22:09] <JEEB> and this isn't even the most northern point
[22:09] <llogan> sunrise was at 4 am DST here this mroning. not that i saw it.
[22:09] <JEEB> http://en.ilmatieteenlaitos.fi/weather/helsinki
[22:09] <iive> i've heard about some inventions. something like doors, but for windows. I don't know how are they called...
[22:10] <JEEB> iive, those are something I've wanted since 2003 or so, but I've always ended up just getting used to it
[22:10] <JEEB> meanwhile rovaniemi has some nice figures http://en.ilmatieteenlaitos.fi/weather/rovaniemi
[22:11] <JEEB> sunrise at 0200, sunset at 038
[22:11] <JEEB> *0038
[22:11] <Daemon404> yeah but thats because scandinavi is north
[22:11] <Daemon404> england is shitty like this because of their stupid timezone
[22:12] <Daemon404> e.g. madrid is west of london, but in CET
[22:12] <llogan> JEEB: we're only ~2 degrees diff
[22:12] <Daemon404> as usual, England is *too good* for <thing from europe>
[22:12] <Daemon404> hurr hurr hurr
[22:14] <llogan> (latitude i mean)
[22:28] <jamrial> kurosu: maybe you could look at re-arranging the hevc_mc init macros so not every function is defined as sse4 when like half of them are only ssse3 (and some even sse2)
[22:28] <jamrial> that is, if you were looking into fixing some hevc_mc issues without actually having to fix the asm
[22:54] <kurosu> jamrial, it's a lot of ungrateful work
[22:55] <kurosu> and actually, I think it's mostly like mmx: it's unlikly you'll use xmm6+ in regular compiler code (unless float)
[22:56] <kurosu> so you could just save them on entry, ignore always in functions, and restore on exit
[22:57] <kurosu> case in point: hevc_mc code doesn't crash, yet incorrectly handles regs (unless x264asm does some magic as I haven't paid that much attention to the disassembled code)
[23:16] <Navion> Does anyone know where I can get libavformat.55.dylib for an Intel Mac?
[00:00] --- Wed Jun  4 2014


More information about the Ffmpeg-devel-irc mailing list