[Ffmpeg-devel-irc] ffmpeg-devel.log.20141018

[burek]

[00:00] <J_Darnley> Does FFmpeg have an example of increasing indentation in printed output somewhere?
[00:00] <J_Darnley> What about ffprobe's various outputs?
[00:03] <Compn> i dont think so
[00:04] <Compn> screenshots and examples would look nice in the docs
[01:32] <cone-574> ffmpeg.git 03Marton Balint 07master:9dac7977a981: ffplay: remove delays when paused from video and subtitle thread
[01:32] <cone-574> ffmpeg.git 03Marton Balint 07master:0edf3e7a28d0: ffplay: remove manual bug option handling code
[01:32] <cone-574> ffmpeg.git 03Marton Balint 07master:2c4d6a342d4f: ffplay: update documentation
[02:14] <jamrial> "<ubitux> the 7 lines of vpx hacks could be replaced with a check_pkg_config "vpx >= 0.9.7" or something" <--- I disagree
[02:14] <jamrial> the libvpx checks are for vp8 decoder and encoder, and vp9 decoder and encoder
[02:16] <jamrial> a pkg-config for vpx >= 0.9.7 will only guarantee that vp8 enc/dec is available. you still need to check if vp9 enc/dec are available and new enough
[02:17] <jamrial> in the end it will still be at least three checks
[02:19] <jamrial> we can replace the checks for pkg-config ones, though. i'm not against that
[02:19] <jamrial> it's just that the replacement will still be at least three separate checks
[02:34] <cone-574> ffmpeg.git 03Supraja Meedinti 07master:0eabbe2101f1: libavutil: Add CAST128 symmetric block cipher
[03:01] <cone-574> ffmpeg.git 03Benoit Fouet 07master:e06263ef1e0e: configure: add pkg-config support for libx264
[03:02] <jamrial> ^now we just need to remove the fallback hack :p
[03:11] <cone-574> ffmpeg.git 03Uwe L. Korn 07master:f4cd8b80b9cb: rtmpproto: Track last received timestamp
[03:11] <cone-574> ffmpeg.git 03Michael Niedermayer 07master:178764e833b1: Merge commit 'f4cd8b80b9cb2a9a4112d8abb1d3b93cd3b4e467'
[03:49] <cone-574> ffmpeg.git 03Uwe L. Korn 07master:9bec3ca2b878: rtmpproto: Add pause support
[03:49] <cone-574> ffmpeg.git 03Michael Niedermayer 07master:7891b4fa590b: Merge commit '9bec3ca2b878c83d2337bc6095bd51c3e3f8eef0'
[11:43] <ubitux> fate is a bit yellow
[11:46] <ubitux> and cast leaks
[12:21] <cone-919> ffmpeg.git 03Anton Khirnov 07master:fb238f8230c1: urlprotocol: remove unused url_interrupt_cb declaration
[12:21] <cone-919> ffmpeg.git 03Michael Niedermayer 07master:83a897656aae: Merge commit 'fb238f8230c1b0314985482207add548c68f83cf'
[12:46] <cone-919> ffmpeg.git 03Michael Niedermayer 07master:7321fc2f4304: tests/fate/mpeg4: add -sws_flags +bitexact
[13:07] <cone-919> ffmpeg.git 03Timothy B. Terriberry 07master:6cbbf0592f4f: resample: Avoid off-by-1 errors in PTS calcs.
[13:07] <cone-919> ffmpeg.git 03Michael Niedermayer 07master:17d16008befa: Merge commit '6cbbf0592f4f3940aac7f687850d1b726a2ea836'
[13:52] <cone-919> ffmpeg.git 03Anton Khirnov 07master:a6e4ce9fd508: lavc: make rc_qsquish a private option of mpegvideo encoders
[13:52] <cone-919> ffmpeg.git 03Michael Niedermayer 07master:74128352047e: Merge commit 'a6e4ce9fd50897dc6d9c2ada4b6b8090de7de5bf'
[14:03] <cone-919> ffmpeg.git 03Anton Khirnov 07master:946f95354ba7: lavc: make rc_qmod_* into private options of mpegvideo encoders
[14:03] <cone-919> ffmpeg.git 03Michael Niedermayer 07master:17a065cf204f: Merge commit '946f95354ba76ef73c9b66889d86ab5fba4fb486'
[14:17] <cone-919> ffmpeg.git 03Anton Khirnov 07master:a75c2eb25a62: lavc: make rc_eq into private options of mpegvideo encoders
[14:17] <cone-919> ffmpeg.git 03Michael Niedermayer 07master:ec8f56efad2c: Merge commit 'a75c2eb25a62105c09b48521aef429dc8a231637'
[14:28] <cone-919> ffmpeg.git 03Anton Khirnov 07master:66a68ddd1ac3: lavc: make rc_buffer_aggressivity/rc_initial_cplx into private options of mpegvideo encoders
[14:28] <cone-919> ffmpeg.git 03Michael Niedermayer 07master:fc67d937cc73: Merge commit '66a68ddd1ac376b24db8695058bc8fc28f5bada6'
[15:27] <cone-919> ffmpeg.git 03Anton Khirnov 07master:59b55c494e01: lavc: deprecate unused me_threshold field
[15:28] <cone-919> ffmpeg.git 03Michael Niedermayer 07master:c48e8d9dca55: Merge commit '59b55c494e01de09dfd7cbf9cbbe7223265ac283'
[15:37] <anshul_mahe> I have completed my closed caption decoder for ffmpeg
[15:38] <kierank> anshul_mahe: how have you done it
[15:38] <kierank> and is it 608 and 708?
[15:39] <anshul_mahe> it is 608 now using external ccextractor  lbrary
[15:39] <anshul_mahe> my problem is that my work is based on base patchof nicolas
[15:39] <kierank> libzvbi could also do it
[15:39] <kierank> and ffmpeg already uses that for 608
[15:39] <kierank> i mean teletext
[15:40] <kierank> the difficult part is how you decode captions inside h264/mpeg-2
[15:40] <anshul_mahe> using movie filter
[15:41] <anshul_mahe> Just now for giving my commits on ffmpeg-devel list, I want my patch to be multi authored 
[15:41] <ubitux> speaking of this, libzvbi decodes as text like shit
[15:42] <cone-919> ffmpeg.git 03Anton Khirnov 07master:685be3d1aecf: libvpxenc: add static-thresh private option
[15:42] <ubitux> so if anyone is willing to fix that...
[15:42] <cone-919> ffmpeg.git 03Michael Niedermayer 07master:8be1120deb2b: Merge commit '685be3d1aecf965281128f3b631ed965bcfa754d'
[15:42] <ubitux> (the zvbi wrapper i mean)
[15:43] <wm4> ubitux: teletext has a more serious problem
[15:43] <wm4> it has multiple pages at once or something
[15:43] <ubitux> yeah&
[15:44] <ubitux> it needs a "don't consume the whole packet at once" just like a/v
[15:44] <wm4> uh
[15:44] <wm4> please no
[15:44] <wm4> because fuck that
[15:44] <ubitux> then what do you propose?
[15:44] <wm4> and it should be fixed for audio
[15:45] <ubitux> and how are you going to do that?
[15:45] <wm4> it makes the data flow much more complicated and awkward and has potential to introduce infinite loops if not careful
[15:45] <wm4> dunno
[15:45] <wm4> for teletext, the abstraction is wrong
[15:45] <wm4> teletext is basically stateful as I understand it
[15:47] <anshul_mahe> I would like someone to take authorship of my patch for closed caption using ccextractor library, is someone here willing to take
[15:47] <anshul_mahe> then I would send my patch to ffmpeg-devel mailing list
[15:47] <J_Darnley> What do you mean "take authorship"?
[15:49] <anshul_mahe> I dont want to be author of those patch, i signed NDA for it
[15:50] <anshul_mahe> I tried making whole new implentation, but I found my mind is working in same way
[15:51] <wm4> anshul_mahe: ask kierank or j-b, I think they have a grasp on such issues
[15:51] <wm4> I mean, don't ask them for taking ownership, but for advice what to do
[15:51] <kierank> anshul_mahe: nda for what?
[15:53] <anshul_mahe> ok, looking in mail, did not read that carefully, will come back soon with more detail
[15:56] <anshul_mahe> Its just an statement here http://pastebin.com/HAUS5UKw
[15:57] <cone-919> ffmpeg.git 03Anton Khirnov 07master:c4f7bee0872d: lavc: deprecate unused mb_threshold field
[15:57] <cone-919> ffmpeg.git 03Michael Niedermayer 07master:1a5871d1bec6: Merge commit 'c4f7bee0872d5e2e0e91673dd4b0f18de4a3d866'
[16:08] <anshul_mahe> kierank: any idea's?
[16:08] <kierank> then you can't
[16:08] <kierank> but if that company distributes ffmpeg
[16:08] <kierank> then they would have to send patches
[16:08] <kierank> ianal
[16:09] <anshul_mahe> they are using it, in there server they dont distributes
[16:12] <kierank> ah then you can't
[16:13] <anshul_mahe> ok, just now i am using whole ccextractor library for decoding, If I decompose ccextractor with small library of different decoder, then integrate it
[16:14] <anshul_mahe> will that work
[16:14] <J_Darnley> IANAL but I would side with kierank's opinion on this.  If XYZ own the patch you've written then it is up to them whether they submit it.  They are under no obligation to release source code if they don't distribute FFmpeg.
[16:15] <J_Darnley> From that one sentence you posted, any work you do while being paid by XYZ is owned by XYZ.
[16:21] <anshul_mahe> ok, I would rather implement decoder over here, I have its RFC, that was not I did when  I was getting paid. 
[16:22] <anshul_mahe> and rather stay away from signing such document 
[16:24] <anshul_mahe> I am in dilemma, should I use decomposed ccextractor library or write a new decoder for ffmpeg
[16:26] <anshul_mahe> or  may be libzvbi
[16:36] <BBB> omg more inline assembly
[16:44] <cone-919> ffmpeg.git 03Anton Khirnov 07master:15ec053c4c0b: lavc: make border_masking into private options of mpegvideo encoders
[16:44] <cone-919> ffmpeg.git 03Michael Niedermayer 07master:709ab39cb9a4: Merge commit '15ec053c4c0b198a2e93eb8e60c8f41e091e0c40'
[16:48] <J_Darnley> BBB: what now?
[16:48] <BBB> that vf_noise patch
[16:48] <BBB> inline asm
[16:48] <BBB> :(
[17:02] <ubitux> BBB: it's just moved somewhere else
[17:07] <BBB> I know, but its still inline asm
[17:07] <BBB> lets convert it and make msvc users happy
[17:07] <BBB> they get such a degenerate experience :(
[17:11] <cone-919> ffmpeg.git 03Giorgio Vazzana 07master:b48c95b6c92e: libavutil: Add missing av_free() to cast5 selftest
[17:13] Action: michaelni wonders if inline->yasm really fixes "MS users getting a degenerate experience"
[17:17] <michaelni> wouldnt it be needed to replace the OS too ...
[17:19] <kierank> sure, probably ~1 billion reinstalls needed (chrome+vlc)
[17:33] <ubitux> i think vlc doesn't care about libavfilter
[17:33] <ubitux> and chrome doesn't need it
[17:34] <ubitux> that said, i'm all for having it in yasm
[17:34] <ubitux> and that commit eases this transition
[17:34] <ubitux> it shouldn't be that much work afaict
[17:34] <ubitux> and this asm was just ported from mplayer, and it was kept as is to avoid the risk of bugs during the port
[17:51] <hima> ubitux : hi
[17:53] <hima> here is the modified code 1. removed get_pts_end 2. include event {  .please have a look http://fpaste.org/143128/41364750/
[17:55] <ubitux> does it work with "event" with AND without "{"?
[17:55] <ubitux> also, what if there is more spaces between "event" and "{"
[17:55] <ubitux> what about a line break between "event" and the timestamp?
[17:57] <hima> in all the examples i saw there was just a single space between event and { 
[17:57] <hima> ohh right i forgot that case 
[17:57] <hima> making that one 
[17:58] <hima> and line break...it does skip \n and spaces 
[18:05] <hima> ubitux : http://fpaste.org/143129/36482931/
[18:05] <ubitux> does it work?
[18:07] <ubitux> there is nothing that prevent anyone from having spaces or \n between the "event" and {
[18:09] <hima> yes succesfully did make it 
[18:09] <ubitux> can you make a patch with git so i can test?
[18:09] <hima> okayy will add spaces and \n between event and { then 
[18:11] <hima> ok after this change it doesnt work again showing segmentation fault :|
[18:11] <hima> ./ffmpeg -f kate -i sample_kate.kate out.srt 
[18:11] <hima> with make it did not give any error 
[18:11] <hima> though 
[18:12] <ubitux> i can help you with ffmpeg stuff, but you know you are supposed to know C and deal with pure C problems yourself
[18:12] <ubitux> i'm not going to debug your code
[18:12] <ubitux> i will just make sure it's not broken and point you out issues with it if there are
[18:14] <hima> http://fpaste.org/143131/64875014/ there are no syntax problems ubitux logically i guess its fine. can you just help me how to read the output of the valgrind as to figure out the problem 
[18:14] <hima> thisis the valgrind output
[18:14] <ubitux> at kate_read_header (katedec.c:60)
[18:14] <ubitux> you are reading a byte you're not supposed to
[18:15] <ubitux> that's all i can say
[18:17] <hima> okay i will figure it out.
[18:28] <hima> ubitux: do i have to make a change in libavcodec/avcodec.h too besides allformats.h and Makefile?
[18:29] <ubitux> not you're exporting as text
[18:29] <ubitux> so no need to touch libavcodec
[18:29] <ubitux> you're supposed to add an entry in the Changelog and doc/general.texi, but you probably want to do that at the end, don't bother about it for now, it will brings problem when rebasing
[18:30] <hima> okay 
[18:53] <hima> ubitux: on ./ffmpeg -f kate -i sample_kate.kate out.txt  it gives http://fpaste.org/143139/51048141/ but the cursor just keeps on waiting there same for valgrind.no segmentation fault though 
[18:53] <hima> what could be the error?
[18:54] <ubitux> infinite loop
[19:08] <arwa> I tried writing a very naive code for neighborhood scaling. I am not getting the desired image. Can anyone help me debug it? http://pastebin.com/NMie9LNi
[19:15] <J_Darnley> arwa: "This is a private paste. If you created this paste, please login to view it."
[19:24] <arwa> The paste is public.
[19:24] <arwa> Are you not able to view it?
[19:24] Action: wm4 suggests http://www.sprunge.us
[19:25] <wm4> pastebin.com is incredibly annoying slow and shitty, don't know why anyone would use it
[19:29] <anshul_mahe> wm4, what you suggest show 404 in India
[19:31] <wm4> oh, it's actually served on http://sprunge.us/
[19:34] <arwa> Actually, its giving me proxy error. How do I do the proxy settings? 
[19:35] <arwa> I tried the command line --> "cat try  | curl -F 'sprunge=<-' http://sprunge.us" 
[19:36] <wm4> here's another random one that's IMO better that pastebin.com: http://privatepaste.com/
[19:38] <arwa> My code -> http://privatepaste.com/80b17a285f
[19:38] <arwa> thanks :)
[19:40] <wm4> you should restrict the pixel formats (PIXFMTS) it takes to the formats which work
[19:41] <wm4> the (int) cast in the filter function are pointless... if j is int, then j/2 is also int
[19:42] <arwa> okay.
[19:42] <arwa> It wont work without restricting the pixelformat?
[19:43] <wm4> it will, but there's the chance it gives you a pixelformat you didn't expext, and then random things will happen
[19:43] <wm4> s/expext/expect
[19:47] <arwa> okay
[19:51] <arwa> I added the query formats
[19:52] <arwa> still its not giving me the desired image
[19:52] <ubitux> what pix fmt did you request?
[19:54] <arwa> AV_PIX_FMT_RGBA, AV_PIX_FMT_BGRA, AV_PIX_FMT_ARGB, AV_PIX_FMT_ABGR, AV_PIX_FMT_RGB24, AV_PIX_FMT_BGR24, AV_PIX_FMT_RGB565BE, AV_PIX_FMT_BGR565BE, AV_PIX_FMT_RGB555BE, AV_PIX_FMT_BGR555BE,AV_PIX_FMT_RGB565LE, AV_PIX_FMT_BGR565LE, AV_PIX_FMT_RGB555LE, AV_PIX_FMT_BGR555LE,AV_PIX_FMT_NONE
[19:54] <ubitux> do you really support all of these?
[19:54] <ubitux> it looks like to me that you only support AV_PIX_FMT_RGB24
[19:57] <arwa> no
[19:58] <arwa> but i have returned an avlog if we get any format diff than rgb24
[20:00] <ubitux> i mean your code seems to be able to cope with rgb24 only
[20:00] <ubitux> and bgr24
[20:00] <ubitux> not so sure about the others
[20:01] <wm4> the others are 4 bytes  and 2 bytes per pixelformats, so they definitely won't
[20:02] <arwa> yeah i am working on rgb24 only
[20:02] <wm4> then don't allow the other formats
[20:04] <arwa> i just requested all the pixel formats. But I am giving out a error message via avlog informing that the other pixel formats wont work
[20:07] <arwa> and now when I am removing all the pixel formats, I am getting segmentation fault. Does this mean that the format I am giving does not belong to either rgb24 or bgr24?
[20:13] <cone-105> ffmpeg.git 03Anton Khirnov 07master:530c1441fd14: lavc: make lmax/lmin into private options of mpegvideo encoders
[20:13] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:ae6ba2db8ff8: avcodec/mjpegenc: add priv_class to encoders
[20:13] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:25f4e9387d0e: Merge commit '530c1441fd1426b6a4bb33485ff3226e1ae0ad45'
[20:13] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:bf301258b49f: avcodec/snowenc: add FF_MPV_COMMON_OPTS
[20:13] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:f4445c7be45c: avcodec/wmv2enc: add priv_class
[20:14] <wm4> arwa: then try to find out where it crashes, by using gdb or valgrind
[20:19] <arwa> I was trying to do that. But I am not getting how to use it with ffmpeg executable file?
[20:20] <arwa> if i am doing normal gdb, it is not able to compile the file itself.
[20:22] <wm4> how are you trying to use it?
[20:27] <arwa> I am doing gdb, then its giving the file to be compiled as input
[20:27] <arwa> * I am giving the file to be compiled as input 
[21:43] <cone-105> ffmpeg.git 03Luca Barbato 07master:f401792595dd: vf_drawtext: Do not leak the mmapped textfile
[21:43] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:0b2f860db53e: Merge commit 'f401792595dd7760f531e8a3bd2336e9033bd45a'
[22:01] <cone-105> ffmpeg.git 03Vittorio Giovara 07master:6d96c7ea8960: vf_showinfo: Forward the av_image_get_linesize error
[22:01] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:3a4c22f8067f: Merge commit '6d96c7ea89600655b1f18a3b12f3f041fa714152'
[22:13] <cone-105> ffmpeg.git 03Vittorio Giovara 07master:0d989dbfc4bc: af_resample: check av_opt_set_dict return value
[22:13] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:733db64ce0c6: Merge commit '0d989dbfc4bc5bc1d563e967449116a7a9865258'
[22:20] <cone-105> ffmpeg.git 03Vittorio Giovara 07master:088eca28164c: avresample: prevent theoretical division by zero
[22:20] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:9bcc430481d7: Merge commit '088eca28164c8cd3b72b0c3d3f9e3fe5ee5cb28f'
[22:53] <cone-105> ffmpeg.git 03Luca Barbato 07master:963f76144897: avresample: Make sure the even check does not overflow
[22:53] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:85cf049d5092: Merge commit '963f76144897d3f7684d82ec21e51dd50ea1106e'
[23:37] <cone-105> ffmpeg.git 03Timothy Gu 07master:b7c77912b621: oss_audio: use a macro to simplify ioctl() error checking
[23:37] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:837605704daf: Merge commit 'b7c77912b62163b3b46ce93fe42fff3c83604c82'
[23:47] <cone-105> ffmpeg.git 03Vittorio Giovara 07master:fc1eda543c6e: swscale: fix sign extensions in yuv planar conversion
[23:47] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:fa040c224703: Merge commit 'fc1eda543c6ef043300612db90da5cfd972af650'
[23:56] <cone-105> ffmpeg.git 03Alexander Drozdov 07master:0034314a69e7: rtmp: Always call rtmp_close() on rtmp_open() failure
[23:56] <cone-105> ffmpeg.git 03Michael Niedermayer 07master:52bf4ad6746a: Merge commit '0034314a69e76a53534a74cceef865cfcb7b42cc'
[00:00] --- Sun Oct 19 2014