[Ffmpeg-devel-irc] ffmpeg-devel.log.20111129

burek burek021 at gmail.com
Wed Nov 30 02:05:04 CET 2011


[00:02] <michaelni> thx anyway  :)
[00:37] <CIA-41> ffmpeg: 03Alex Converse 07master * r999e7ebd23 10ffmpeg/libavcodec/dca.c: dca: Replace oversized unused get_bits() with skip_bits_long().
[00:37] <CIA-41> ffmpeg: 03Alex Converse 07master * r2b45222b6a 10ffmpeg/libavformat/md5proto.c: md5proto: Fix order of operations.
[00:37] <CIA-41> ffmpeg: 03Diego Biurrun 07master * rc88ebdb42c 10ffmpeg/ (18 files in 2 dirs): Eliminate pointless 0/NULL initializers in AVCodec and similar declarations.
[00:37] <CIA-41> ffmpeg: 03Mans Rullgard 07master * r60084a1723 10ffmpeg/libavutil/timer.h: 
[00:37] <CIA-41> ffmpeg: timer: fix misspelling of "decicycles"
[00:37] <CIA-41> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[00:37] <CIA-41> ffmpeg: 03Mans Rullgard 07master * rb94a3b288e 10ffmpeg/doc/APIchanges: 
[00:37] <CIA-41> ffmpeg: APIchanges: fill in some blanks
[00:37] <CIA-41> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[00:37] <CIA-41> ffmpeg: 03Alex Converse 07master * ra27805189b 10ffmpeg/libavcodec/txd.c: txd: Fix order of operations.
[00:37] <CIA-41> ffmpeg: 03Alex Converse 07master * rac47e014bb 10ffmpeg/libavformat/ (adtsenc.c mpegtsenc.c): 
[00:37] <CIA-41> ffmpeg: adtsenc: Check frame size.
[00:37] <CIA-41> ffmpeg: Inspired by work from: Michael Niedermayer <michaelni at gmx.at>.
[00:37] <CIA-41> ffmpeg: Signed-off-by: Alex Converse <alex.converse at gmail.com>
[00:37] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r4dcd1a3145 10ffmpeg/: (log message trimmed)
[00:37] <CIA-41> ffmpeg: Merge remote-tracking branch 'qatar/master'
[00:37] <CIA-41> ffmpeg: * qatar/master:
[00:37] <CIA-41> ffmpeg:  adtsenc: Check frame size.
[00:37] <CIA-41> ffmpeg:  txd: Fix order of operations.
[00:37] <CIA-41> ffmpeg:  APIchanges: fill in some blanks
[00:37] <CIA-41> ffmpeg:  timer: fix misspelling of "decicycles"
[00:37] <CIA-41> ffmpeg: 03Kostya Shishkov 07master * r1469f943ad 10ffmpeg/libavcodec/indeo3.c: 
[00:37] <CIA-41> ffmpeg: indeo3: cosmetics
[00:37] <CIA-41> ffmpeg: Signed-off-by: Diego Biurrun <diego at biurrun.de>
[00:52] <kcm1700> thanks michaelni
[01:19] <michaelni> kcm1700, by evaluating the center point and considering its value and the dx and dy it should be possible to skip a larger class of cases
[01:24] <kcm1700> What would be the maximum value of 'shift' variable?
[01:25] <kcm1700> if it's not that large, it'll be possible to use LUT.
[02:02] <j-b> violet: ask michaelni or burek 
[02:02] <violet> thank you
[05:05] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rfc09bf57a6 10ffmpeg/ (17 files in 5 dirs): 
[05:05] <CIA-41> ffmpeg: movenc: Write file with minimal number of chunks for the given interleaving.
[05:05] <CIA-41> ffmpeg: Reviewed-by: Baptiste Coudurier <baptiste.coudurier at gmail.com>
[05:05] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[08:02] <ubitux> why fill_char=0x80 when sample format is u8 and not 0 like all the others?
[08:02] <ubitux> in generate_silence()
[08:03] <Tjoppen> .. because it's u8?
[08:05] <ubitux> i don't see the point
[08:05] <ubitux> sorry if it is obvious :p
[08:05] <Tjoppen> u8 = s8+128
[08:07] <ubitux> oh i'm stupid, i was thinking we had U16, U32 and such and was surprised of to have it only in U8
[08:07] <ubitux> thanks, makes sense obviously.
[08:07] <Tjoppen> yeah, u8 is basically only old formats
[08:08] <Tjoppen> if you mix up u8 and s8 when playing out on a sound card you end up with.. fun
[08:08] <ubitux> i guess i'll have to fix the map channel silence then
[09:16] <ubitux> huh, got a really strange bug; i have a vfilter which i called like this: -vf foobar=10,scale=300:200, and one field in the private struct of foobar gets updated with the height of the scale parameter
[09:17] <ubitux> quite fun, i wonder what i'm doing wrong here :)
[09:17] <ubitux> with the height parameter* of the scale filter*
[10:27] <ubitux> btw, http://ffmpeg.org/doxygen/trunk/index.html  why no lavfi? :(
[12:59] <j-b> good morning
[13:39] <burek> morning :)
[13:40] <burek> i have a lot of these spell checkings done, so I'll try to setup git now and update those docs :)
[14:58] <burek> michaelni or pasteeater, on what e-mail should I send patch for spelling errors?
[15:02] <cbsrobot> burek: ffmpeg-devel
[15:02] <burek> that's an email? :)
[15:03] <cbsrobot> heres the form: https://lists.ffmpeg.org/mailman/listinfo/ffmpeg-devel/
[15:04] <cbsrobot> :)
[15:04] <burek> oh ok
[15:04] <burek> :)
[15:04] <burek> but
[15:04] <burek> wouldn't it go public and stuff? am i supposed to make it publicly visible ? :)
[15:17] <iive> burek: are the spelling errors already publicly visible?
[15:17] <burek> i've just sent the patch to the ml
[16:10] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r6ffdc262b0 10ffmpeg/libavcodec/ac3enc_template.c: 
[16:10] <CIA-41> ffmpeg: ac3enc: clenaup project name in one comment
[16:10] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:10] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r313d30c180 10ffmpeg/libavutil/ (avutil.h opt.h): 
[16:10] <CIA-41> ffmpeg: avutil: revert project name messing
[16:10] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:14] <michaelni> burek, you should split the spelling fixes so that each students work is one commit and give credit to each student in each commit
[16:14] <michaelni> and subscribe to ffmpeg-dev :)
[16:15] <michaelni> the list is subscribers only as a spam prevention
[16:19] <burek> oh man..
[16:19] <burek> they will already get payed for that :)
[16:19] <burek> and changes are really small
[16:19] <burek> so its more work for me to split all that and stuff
[16:19] <burek> then it is their benefit of credits for the same thing
[16:21] <michaelni> its less than 5min work
[16:22] <michaelni> just a git reset HEAD^ and a few git commit thisfile.texi
[16:23] <michaelni> about credit, you and them should get credit
[16:24] <michaelni> about work, its true gci is alot of work
[16:24] <michaelni> maybe you should ask the student to send a applyable patch to the texi next time ...
[16:28] <burek> i never used git for comitting and editting and stuff
[16:28] <burek> so its not 5 min for me, unfortunatelly :S
[16:29] <michaelni> burek, we will see :)
[16:29] <michaelni> does git log show "that" commit as last ?
[16:30] <michaelni> last is the topmost
[16:30] <burek> er.. can you ask me something about beer? :D
[16:30] <burek> im more likely to know the answer :)
[16:31] <michaelni> whats your favorite beer? ;)
[16:31] <burek> the cold one! :D
[16:31] <burek> that was easy ^^
[16:35] <michaelni> burek, now try "git log -1" typed in the terminal where your git stuff is and confirm thats the commit with all the spelling correction
[16:35] <michaelni> "git log -1 -p" would show the actual diff
[16:35] <burek> commit fc09bf57a60d4c4a6d339b204b3282337067c06d
[16:35] <burek> Author: Michael Niedermayer <michaelni at gmx.at>
[16:37] <michaelni> burek, did you commit it to another branch ?
[16:37] <burek> burek i think
[16:37] <michaelni> git checkout "that  branch"
[16:37] <burek> yup it's the last
[16:37] <michaelni> good, now just to make sure do a "git diff" and make sure theres no output
[16:38] <burek> there is no
[16:38] <michaelni> good, now a "git reset HEAD^" that will kill the last commit but leave the files as tehy are
[16:38] <burek> ok
[16:39] <burek> yes, it showed me resetted files i guess
[16:39] <burek> all texi
[16:39] <burek> 12 of them actually
[16:40] <michaelni> and now just a few "git commit -s doc/thisfile.texi" for each commit you want to make and mention in each the student upon whos submission its based, you will be credited as author of the commit
[16:40] <michaelni> automatically
[16:41] <burek> i see, but some documents are spanned across several texi files
[16:42] <burek> which are included with @include or something in the ffmpeg.texi
[16:42] <michaelni> you can specify multiple files for commit if its all the work of the same student
[16:42] <burek> so i guess i need to check which one is whose
[16:59] <CIA-41> ffmpeg: 03Geek.Song 07master * r9cdf048ad2 10ffmpeg/libavformat/ (movenc.c movenc.h): 
[16:59] <CIA-41> ffmpeg: movenc: Remove unneeded chunkSize field from MOVIentry
[16:59] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:12] <vivienschilis> abou
[17:14] <CIA-41> ffmpeg: 03Clément BSsch 07master * r81a65b82fb 10ffmpeg/libswresample/audioconvert.c: swr: handle correctly muted channel with u8 sample fmt.
[17:34] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r4b4a02b847 10ffmpeg/libavcodec/utils.c: 
[17:34] <CIA-41> ffmpeg: lavc: dont call set_dimensions() on h264 codec init.
[17:34] <CIA-41> ffmpeg: This fixes ffprobe showing an incorrect width with
[17:34] <CIA-41> ffmpeg: http://panda-test-harness-videos.s3.amazonaws.com/panda.mp4
[17:34] <CIA-41> ffmpeg: Idea-by: Joakim Plate <elupus at ecce.se>
[17:34] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:57] <burek> michaelni, i've sent 4 patches (as attachments) to the ml :)
[17:57] <burek> can you check if thats all ok?
[18:07] <burek> also, how can i now continue working, but to be able to know where did i stop now, to distinguish new changes from these, already made
[18:15] <ubitux> burek: you can git commit --amend --author=...
[18:15] <ubitux> instead of writing the credits in the commit message
[18:16] <burek> ok, but i already started this way, so i'll just finish like that, and next time use --author :)
[18:16] <burek> thx :)
[18:18] <ubitux> well, if you don't do it, michaelni might do it instead :p
[18:19] <ubitux> maybe you could give a hand ;)
[18:19] <michaelni> ill reword the commit messages a little as they should contain the "what/where" in the first line
[18:20] <michaelni> iam way too lazy to swicth author/credit around, both burek and the student should be credited
[18:20] <ubitux> hehe ok :)
[18:20] <michaelni> the students submited changed html we needed a patch to texi ...
[18:21] <burek> i dont need to be credited :)
[18:21] <burek> just please lets get this done with :D
[18:21] <michaelni> exactly my thought
[18:21] <burek> it takes too much time for simple tasks
[18:21] <burek> :)
[18:21] <michaelni> ill do it
[18:21] <burek> thanks :D
[18:21] <burek> :beer: :)
[18:22] <michaelni> and btw --author needs full emails so i cant use it anyway ...
[18:23] <michaelni> next time i think its better to let the student do all work and then credit just them
[18:29] <michaelni> i mean at least require a changed texi file, its easy to make a patch but hard to convert between formats
[18:34] <CIA-41> ffmpeg: 03root 07master * r1c212a6465 10ffmpeg/doc/general.texi: 
[18:34] <CIA-41> ffmpeg: general.texi: fix spelling errors
[18:34] <CIA-41> ffmpeg: credits to: Samuel M (from Google Code-in)
[18:34] <CIA-41> ffmpeg: Signed-off-by: burek <burek021 at gmail.com>
[18:34] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:34] <CIA-41> ffmpeg: 03root 07master * raadbf9f74c 10ffmpeg/doc/developer.texi: 
[18:34] <CIA-41> ffmpeg: developer.texi: fix spelling errors
[18:34] <CIA-41> ffmpeg: credits to: KayC (from Google Code-in)
[18:34] <CIA-41> ffmpeg: Signed-off-by: burek <burek021 at gmail.com>
[18:34] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:34] <CIA-41> ffmpeg: 03root 07master * rb0a90c2004 10ffmpeg/doc/faq.texi: 
[18:34] <CIA-41> ffmpeg: faq.texi: fix spelling errors
[18:34] <CIA-41> ffmpeg: credits to: Philip (from Google Code-in)
[18:34] <CIA-41> ffmpeg: Signed-off-by: burek <burek021 at gmail.com>
[18:34] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:35] Last message repeated 1 time(s).
[18:35] <Compn> root is an odd name for a committer :P
[18:48] <burek> fck.. i checked out master and made changes and now i have to do it again in burek branch..
[18:50] <burek> i knew i would crew up something.. :)
[18:51] <burek> so, first i created burek branch, made all changes and sent an email to the ml
[18:51] <burek> after that, i was doing those git commands you told me, and in all those cmds there was git checkout master
[18:52] <burek> after that i made new changes (which should have been made in burek branch), because texi files overlap
[18:52] <burek> and some changes have alredy been fixed
[18:52] <burek> so, now i ended up with diff from master instead from burek
[18:52] <burek> fun never stops :D
[19:56] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * r7076967786 10ffmpeg/libavcodec/x86/imdct36_sse.asm: 
[19:56] <CIA-41> ffmpeg: Consistently use %ifdef ARCH_X86_64
[19:56] <CIA-41> ffmpeg: One out of 3 places used ifndef, which is needlessly confusing.
[19:56] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[22:08] <bcoudurier> michaelni, stop procrastinating about avfilter flush
[22:09] <bcoudurier> I sent you patch ages ago to fix this
[22:24] <pasteeater> what is "coverage support"?
[22:25] <Compn> code coverage, check melansons blog
[22:27] <gnafu> http://multimedia.cx/eggs/using-lcov-with-ffmpeg/
[22:27] <Compn> that
[22:34] <pasteeater> ah, thanks.
[22:38] <michaelni> bcoudurier, do you remember the subject of the mail with the patch or something so i can find it ?
[23:15] <CIA-41> ffmpeg: 03Clément BSsch 07master * rb6ffe441cd 10ffmpeg/ffmpeg.c: ffmpeg: do not use a negative total_size (AVERROR) in bitrate estimation.
[00:00] --- Wed Nov 30 2011


More information about the Ffmpeg-devel-irc mailing list