[Ffmpeg-devel-irc] ffmpeg-devel.log.20120310

burek burek021 at gmail.com
Sun Mar 11 02:05:03 CET 2012


[00:12] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * r1a3d4b88a4 10ffmpeg/libavcodec/imgconvert.c: 
[00:12] <CIA-17> ffmpeg: imgconvert: add macro pixdesc_has_alpha for checking if a pixel format has an alpha component
[00:12] <CIA-17> ffmpeg: Reduce redundancy and simplify.
[00:12] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * rfa35d880aa 10ffmpeg/ (6 files in 3 dirs): lavfi: port MP swapuv filter
[00:12] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * raefd1cb725 10ffmpeg/ (5 files in 3 dirs): 
[00:12] <CIA-17> ffmpeg: lavfi: remove swapuv libmpcodecs wrapper
[00:12] <CIA-17> ffmpeg: The filter has been ported to a native libavfilter filter.
[00:15] <bcoudurier> cbsrobot, that's what happens where un-reviewed crap gets commited
[00:18] <cbsrobot> bcoudurier: you happen to have ffmbc on git somewhere ?
[00:21] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r1af9fdc3ba 10ffmpeg/libavcodec/snowdec.c: 
[00:21] <CIA-17> ffmpeg: snow: reject unsupported chroma shifts.
[00:21] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[00:21] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:48] <Compn> cbsrobot : isint it first result when you google ffmbc ?
[00:48] <Daemon404> there is no ffmbc git
[00:48] <Daemon404> only tarballs
[00:49] <Compn> ah
[00:49] <Compn> what about getting ffmbc and ffmpeg merged? :D
[00:49] <Daemon404> 0.000X% chance of that probably
[00:49] <Daemon404> <_<
[00:50] <cbsrobot> furthermore ffmbc is mostly gpl
[00:51] <Daemon404> entirely.
[00:51] <cbsrobot> didnt check it
[00:55] <michaelni> 6cb1d361 libavcodec/dnxhdenc.c (Baptiste Coudurier  2007-10-08 12:28:36 +0000  417)                 level = (1-2*level) * qscale * weight_matrix[i];
[00:56] <michaelni> havnt checked what ffmbc contains
[00:56] <michaelni> but it sounded like its about this line ...
[00:57] <Daemon404> technically you cant backport anything from ffmbc anyway
[01:06] <cbsrobot> michaelni: there are multiple lines afterwards
[01:07] <cbsrobot> ffmbc: level = (1-2*level) * qscale * weight_matrix[j];
[01:11] <michaelni> cbsrobot, does changing that improve quality (PSNR) ?
[01:12] <cbsrobot> i'll need to check ...
[01:19] <Daemon404> why do people sitll use PSNR as a metric..
[01:20] <bcoudurier> for verification it's no big deal
[01:21] <Daemon404> i suppose
[01:21] <Daemon404> bcoudurier, btw, you dont know any good plae to apply for an internship do you? (besides obv. hulu :P)
[01:21] <Daemon404> place*
[01:22] <bcoudurier> where ?
[01:22] <Daemon404> UK, or worst case, somewhere EST
[01:41] Action: Daemon404 whistles
[01:43] <Compn> Daemon404 : justin.tv is hirign
[01:43] <Compn> you want the recruiter email ?
[01:43] <Daemon404> interns?
[01:44] <Daemon404> i specifically need an internship
[01:44] <Compn> i dont know, ffmpeg programmers probably
[01:44] <ubitux> justin.tv is in the usa afaik
[01:44] <Compn> yes it is
[01:44] <Daemon404> yeah
[01:44] <Daemon404> unless it's WFH it's probably not too doable :/
[01:44] <Daemon404> depending on things.
[01:44] <Compn> you can always ask :P
[01:44] <Daemon404> true
[01:52] <CIA-17> ffmpeg: 03Martin Storsjö 07master * r6294d708b8 10ffmpeg/libavformat/rtsp.c: 
[01:52] <CIA-17> ffmpeg: rtsp: Only set the ttl parameter if the server actually gave a value
[01:52] <CIA-17> ffmpeg: Passing ttl=0 to the rtp/udp url contexts makes packets never
[01:52] <CIA-17> ffmpeg: leave the host machine.
[01:52] <CIA-17> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[01:52] <CIA-17> ffmpeg: 03Martin Storsjö 07master * r1b89bcdd7f 10ffmpeg/libavformat/udp.c: 
[01:52] <CIA-17> ffmpeg: udp: Clarify the comment about binding the multicast address
[01:52] <CIA-17> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[01:52] <CIA-17> ffmpeg: 03Diego Biurrun 07master * rffae713a5b 10ffmpeg/ (5 files in 3 dirs): Fix a bunch of common typos.
[01:52] <CIA-17> ffmpeg: 03Alex Converse 07master * r100c3fb2d1 10ffmpeg/libavformat/mpegts.c: mpegts: Always honor a registration descriptor if present and there is no other codec information.
[01:52] <CIA-17> ffmpeg: 03Gil Pedersen 07master * r49e35f497f 10ffmpeg/configure: 
[01:52] <CIA-17> ffmpeg: configure: darwin: Change dylib install names to include major version.
[01:52] <CIA-17> ffmpeg: This will cause linkers to link against the major lib names, instead of the
[01:52] <CIA-17> ffmpeg: base names, allowing multiple major versions of the libraries to co-exist.
[01:52] <CIA-17> ffmpeg: Signed-off-by: Diego Biurrun <diego at biurrun.de>
[01:52] <CIA-17> ffmpeg: 03Martin Storsjö 07master * r2bfd92b330 10ffmpeg/libavformat/udp.c: 
[01:52] <CIA-17> ffmpeg: udp: Set ttl for read-write streams, too, not only for write-only ones
[01:52] <CIA-17> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[01:52] <CIA-17> ffmpeg: 03Martin Storsjö 07master * rc700fdb00f 10ffmpeg/libavformat/udp.c: (log message trimmed)
[01:52] <CIA-17> ffmpeg: udp: Only bind to the multicast address if in read-only mode
[01:52] <CIA-17> ffmpeg: This fixes sending back RTCP RR packets if receiving RTP over
[01:52] <CIA-17> ffmpeg: multicast.
[01:52] <CIA-17> ffmpeg: If the multicast stream is sent on demand (set up and signalled
[01:52] <CIA-17> ffmpeg: via RTSP), the sender might depend on getting RTCP RR packets
[01:52] <CIA-17> ffmpeg: knowing that there are listeners, otherwise the stream can be
[01:52] <CIA-17> ffmpeg: 03Martin Storsjö 07master * r113d3e106d 10ffmpeg/libavformat/udp.c: 
[01:52] <Daemon404> Compn, do you know teh recruited at justin.tv?
[01:53] <CIA-17> ffmpeg: udp: Reorder comments
[01:53] <CIA-17> ffmpeg: When this code was added in 36b532815cb83, the new code was added
[01:53] <CIA-17> ffmpeg: between the existing comment and the existing line of code, making
[01:53] <CIA-17> ffmpeg: the old comment seem to refer to the new code. This makes it read
[01:53] <CIA-17> ffmpeg: correctly.
[01:53] <CIA-17> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[01:53] <CIA-17> ffmpeg: 03Diego Biurrun 07master * reab6968f24 10ffmpeg/libavcodec/Makefile: build: Skip compiling xvmc.h under the correct condition.
[01:53] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r744dd1d356 10ffmpeg/libavcodec/aacdec.c: 
[01:53] <CIA-17> ffmpeg: aacdec: Fix SCE parity check.
[01:53] <CIA-17> ffmpeg: An unpaired SCE preceding a CPE only makes sense for front SCEs
[01:53] <CIA-17> (17 lines omitted)
[02:09] <Daemon404> bcoudurier, so nothin? :/
[02:12] <bcoudurier> not really
[02:13] <Daemon404> ok.
[02:13] <bcoudurier> pharos maybe
[02:13] <Compn> might ask j-b , i think videolan gets a lot of offers...
[02:13] <bcoudurier> I interviewed there some time ago
[02:13] <Daemon404> pharos.com sure is enterprisey
[02:14] <Daemon404> oh. pharos.tv.
[02:14] Action: Daemon404 also pings j-b 
[08:20] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rf7b57add8e 10ffmpeg/libavcodec/cook.c: 
[08:20] <CIA-17> ffmpeg: cook: tighten the quant_index_table range further.
[08:20] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[08:20] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[08:40] <j-b> Daemon404: ?
[08:40] <Daemon404> j-b, someone said you might know of some internship opportunities in UK/Europe/WFH?
[08:41] <j-b> when?
[08:41] <Daemon404> i need to start in may
[08:43] <j-b> doable. easily. 
[08:43] <j-b> coding on libav?
[08:43] <j-b> coding on libav*?
[08:43] <Daemon404> ideally. but anything unix-y really
[08:43] <Daemon404> interesting
[08:44] <Daemon404> emebdded stuff, video stfuf, etc
[08:46] <j-b> resume available somewhere?
[08:47] <Daemon404> i can email it to you
[08:47] <j-b> please do
[08:47] <Daemon404> any specific address?
[08:48] <j-b> jb at v.o
[08:48] <j-b> as usual, for porn
[08:49] <Daemon404> sent.
[08:50] <j-b> Daemon404: do you have a list of commits on FFmpeg?
[08:50] <Daemon404> see the link at the bottom
[08:50] <j-b> Daemon404: what $$$ remuneration are you targetting?
[08:50] <j-b> I'll see that
[08:51] <Daemon404> enough to live on?
[08:51] <Daemon404> it's a requirement for uni
[08:51] <Daemon404> but i do need to eat food.
[08:51] <j-b> 500$ is enough to live in India
[08:51] <j-b> 1200£ is hard to live in London
[08:52] <Daemon404> depends on where i can find accomodations
[08:52] <Daemon404> or if it's in e.g. bristol
[08:52] <Daemon404> accommodations even.
[08:53] <Daemon404> 'enough so i dont starve or go into debt where i am located'
[08:53] <j-b> And now, you are in Bristol?
[08:53] <Daemon404> i am in canada currently.
[08:53] <Daemon404> long story.
[08:53] <Daemon404> im willing  to get my own work visa etc
[08:54] <j-b> you want an internship in Montreal?
[08:54] <Daemon404> well UK above all else, but if i mjust stay in canada
[08:54] <Daemon404> montreal is essentially ideal.
[08:54] <j-b> ok
[08:54] <j-b> I have at least 4 ideas for you :)
[08:54] <Daemon404> please ntoe though
[08:54] <Daemon404> i dont speak french very well :P
[08:54] <j-b> who the hell needs to speak French?
[08:55] <j-b> one of the ideas is to work on libavcodec in Montreal.
[08:55] <Daemon404> in youre in any city in quebec except montreal
[08:55] <Daemon404> you do
[08:55] <Daemon404> :P
[08:55] <j-b> Daemon404: how many months?
[08:55] <Daemon404> i require 8
[08:55] <Daemon404> but 4 is also an option
[08:55] <Daemon404> ill just find a new one after
[08:55] <j-b> ok.
[08:56] <Daemon404> anyway...
[08:56] <Daemon404> 3 am
[08:56] <Daemon404> i really should sleep
[08:56] <j-b> Bristol being the best place? then UK, then Canada, then US, then WFH ?
[08:56] <Daemon404> bristol is 100% ideal
[08:56] <j-b> telework too?
[08:56] <Daemon404> then UK, WFH, Canada
[08:56] <Daemon404> im used to WFH
[08:56] <Daemon404> i did it at my last internship
[08:56] <Daemon404> over vpn
[08:56] <Daemon404> for a while
[08:57] <j-b> great.
[08:57] <j-b> good night, I'll send emails.
[08:57] <Daemon404> thank you very much :)
[08:57] <Daemon404> now...
[08:57] Action: Daemon404 passes out
[10:45] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * rb52bd2a4ab 10ffmpeg/libavformat/dv.c: 
[10:45] <CIA-17> ffmpeg: DV demuxer: Stricter check for avio_read result.
[10:45] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[10:45] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r8fbf825ecc 10ffmpeg/libavcodec/dvdata.c: 
[10:45] <CIA-17> ffmpeg: DV: Use profile[1] detection hack only for stype 0.
[10:45] <CIA-17> ffmpeg: The two samples both have stype 0.
[10:45] <CIA-17> ffmpeg: Without this extra check, the code breaks 4:2:2 dvsd
[10:45] <CIA-17> ffmpeg: (stype 4), since that has the same resolution.
[10:45] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[17:20] <kwizart> hello, from which source would you take the latest libcrystalhd header to compile with ffmpeg ?
[17:20] <kwizart> this component seems unmaintained from broadcom and I don't see an alternate maintainer
[18:58] <CIA-17> ffmpeg: 03Baptiste Coudurier 07master * r1b99971dc7 10ffmpeg/libavfilter/vf_frei0r.c: vf_frei0r: set outlinks sample aspect ratio to 1
[18:59] <CIA-17> ffmpeg: 03Zalewa PL 07master * r0d2f4eedc8 10ffmpeg/libavformat/mpjpeg.c: (log message trimmed)
[18:59] <CIA-17> ffmpeg: mpjpeg: video streaming will no longer break and stop on Firefox
[18:59] <CIA-17> ffmpeg: mpjpeg video streamings would break and stop on Firefox after 1 - 30
[18:59] <CIA-17> ffmpeg: seconds.
[18:59] <CIA-17> ffmpeg: In order to fix this, two changes were made:
[18:59] <CIA-17> ffmpeg: 1. Replaced all occurrences of '\n' character in mjpeg metadata
[18:59] <CIA-17> ffmpeg:  with occurences of "\r\n".
[18:59] <CIA-17> ffmpeg: 03Tomas Härdin 07master * r2ff540a0fe 10ffmpeg/libavformat/mxf.c: 
[18:59] <CIA-17> ffmpeg: mxf: Add DNxHD UL
[18:59] <CIA-17> ffmpeg: Note that the old DNxHD UL is actually JPEG 2000 according to RP224. Leaving it
[18:59] <CIA-17> ffmpeg: as-is for now.
[18:59] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:49] <Zeranoe> Quick license question: I currently release every license file that is included with each of the external libraries that I include, but my builds are compiled for v3 GPL. Doesn't that technically mean I only need to include the v3 GPL with my builds and not every license file for each package?
[20:05] <Zeranoe> I have to go, if someone could get back to me on this I would greatl appreciate it.
[20:17] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * rc4e0d845e2 10ffmpeg/libavcodec/snowenc.c: 
[20:17] <CIA-17> ffmpeg: snowenc: mark some encode_subband arguments const.
[20:17] <CIA-17> ffmpeg: This makes it more obvious that this function only does bitstream
[20:17] <CIA-17> ffmpeg: encoding.
[20:17] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[20:17] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r6fe8cb7d70 10ffmpeg/libavcodec/ (snow.h snowenc.c): 
[20:17] <CIA-17> ffmpeg: snowenc: add no_bitstream option.
[20:17] <CIA-17> ffmpeg: This allows making e.g. MPlayer's -vf uspp filter about 20% faster.
[20:17] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[20:59] <Daemon404> huh.
[20:59] <Daemon404> this is an interesting bug
[20:59] <Daemon404> ffmpeg -i 1.png 2.png
[21:00] <Daemon404> only recompressed half the image
[21:00] <Daemon404> in this one file i have
[21:00] <Daemon404> tasty
[21:00] <durandal_1707> it does not decode correctly or what?
[21:00] <Daemon404> durandal_1707, checking.
[21:01] <Daemon404> durandal_1707, nope
[21:01] <Daemon404> output to bmp works
[21:01] <Daemon404> its a problem with png encoding
[21:01] <durandal_1707> what pix_fmt is picked for encoding?
[21:01] <Daemon404> rgb24 in
[21:01] <Daemon404> rgb24 out
[21:02] <durandal_1707> weird resolution?
[21:03] <Daemon404> https://japland.org/i/Booklet_page0001.png
[21:03] <Daemon404> its dangerous to go alone
[21:03] <Daemon404> take this sample
[21:03] <JEEB> :D
[21:10] <durandal_1707> Daemon404: that file is very big, perhaps not enough memory is allocated (after porting to new api)?
[21:10] <Daemon404> i have no idea
[21:11] <Daemon404> i havent looked
[21:11] <Daemon404> just discovered it
[21:15] <sj_underwater> hi, i didnt rlly get an answer on this in #ffmpeg, but are there resources for Mac compilers?
[21:16] <Daemon404> sj_underwater, what do you mean
[21:17] <sj_underwater> for instance, ffmpeg no longer compiles with the default choice of gcc/llvm-gcc, requires cc/clang instead
[21:17] <sj_underwater> on Lion that is
[21:17] <sj_underwater> also i had to add the libs for cdio to the Makefile manually
[21:17] <Daemon404> iirc llvm-gcc is buggy-ish
[21:17] <sj_underwater> it was an error finding MM_0_FIX_*
[21:18] <sj_underwater> at link-time
[21:18] <Daemon404> known issue
[21:18] <Daemon404> iirc it's because llvm-gcc on mac os x has a gcc-incompatibility
[21:18] <sj_underwater> i had no issue when i did this on SL
[21:19] <Daemon404> SL?
[21:19] <Daemon404> snow lion?
[21:19] <sj_underwater> snow leopard
[21:19] <Daemon404> anyway
[21:19] <Daemon404> chances are the version  of llvm-gcc change
[21:19] <Daemon404> d
[21:19] <durandal_1707> Daemon404: well it looks like it is really size issue, because IOBUF_SIZE is hardcoded for whatever reason....
[21:19] <Daemon404> and this one has a said bug.
[21:19] <Daemon404> durandal_1707, lolwat
[21:20] <durandal_1707> Daemon404: max_packet_size = IOBUF_SIZE*avctx->height + FF_MIN_BUFFER_SIZE;
[21:20] <durandal_1707> png compression can be so perfect
[21:20] <sj_underwater> so is there any documentation/support for mac where i'd expect 2 find this before i get the error?
[21:20] <durandal_1707> IOBUF_SIZE is 4096
[21:21] <durandal_1707> or single image is made of multiple packets ....
[21:23] <Daemon404> durandal_1707, either way it is wrong
[21:23] <Daemon404> sj_underwater, still not sure what you mean? you mean a page that lists if every single compiler is buggy?
[21:24] <sj_underwater> well, for instance the linked-guide is terribly old
[21:24] <Daemon404> linked-guide?
[21:24] <sj_underwater> but besides detecting darwin 10.7 and choosing clang by default, a couple of notes in the release or something
[21:25] <sj_underwater> i do actually read the documentation before ./configure --help;make && sudo make install
[21:25] <sj_underwater> on the website
[21:25] <sj_underwater> for compiling on OS X
[21:26] <Daemon404> link?
[21:26] <Daemon404> it isnt immediately obvious where it is located
[21:26] <sj_underwater> ff__.org/doc__.html, the link FFmpeg on Mac OS X HOWTO: http://stephenjungels.com/jungels.net/articles/ffmpeg-howto.html
[21:26] <sj_underwater> i realize it's off-site, but still
[21:27] <sj_underwater> and btw i wouldnt mind contributing
[21:27] <Daemon404> oh
[21:27] <Daemon404> that guide isn't ffmpeg-maintained.
[21:28] <Daemon404> if you wantto contribute to that, you should probably contact its author
[21:29] <sj_underwater> but what does ffmpeg provide, if anything?
[21:29] <sj_underwater> or is willing to provide
[21:29] <Daemon404> "provde" in what context?
[21:29] <Daemon404> in terms of mac os x specific docs? probably not much
[21:29] <sj_underwater> i mean just a couple notes in the RL or INSTALL or README
[21:29] <Daemon404> well
[21:29] <Daemon404> think of it this way
[21:30] <Daemon404> how many things do you think ffmpeg would need to add in there
[21:30] <Daemon404> to document everything buggy compiler
[21:30] <Daemon404> on eveyr platform
[21:30] <sj_underwater> the clang/cc thing is probably it
[21:30] <sj_underwater> the only thing keeping it from compiling w/ default settings
[21:30] <sj_underwater> if the developers didnt want to put that specific flag in the config script
[21:31] <Daemon404> setting the default compiler to clang probably is not a valid solution
[21:31] <sj_underwater> no?
[21:31] <Daemon404> i think tehres a different fix in tehworks
[21:31] <Daemon404> it was discussed here before
[21:31] <sj_underwater> ok
[21:31] <Daemon404> let me grep  logs
[21:32] <sj_underwater> but i would be willing to contribute to docs, if theyre ever hosted
[21:32] <Daemon404> btw are you using git or a tarball?
[21:33] <sj_underwater> the harmony tb
[21:33] <Daemon404> try git
[21:33] <sj_underwater> sorry freedom, not harmony
[21:34] <Daemon404> http://git.videolan.org/?p=ffmpeg.git;a=commit;h=1aa8f5fcfba6701eb0181ea07c070d09ab0f4efd
[21:34] <sj_underwater> actually i think i tried that
[21:34] <sj_underwater> someone had posted a patch online
[21:34] <sj_underwater> i applied it and it still didnt compile
[21:35] <sj_underwater> i remember the LOCAL_MANGLE part
[21:40] <ubitux> yes sorry that was a lamely attempt that didn't workout
[21:40] <ubitux> i reverted it quickly
[21:40] <sj_underwater> np, im sure it's actually a defect in llvm-gcc
[21:41] <sj_underwater> but choosing clang on darwin systems could be a good default going forward
[22:30] <cbsrobot> sj_underwater: I compiled it on lion recently
[22:30] <sj_underwater> oh?
[22:31] <cbsrobot> I don't remember everything but just ask if you have trouble
[22:31] <sj_underwater> i was just saying that gcc didnt work, i had to choose cc/clang
[22:33] <cbsrobot> yeah had the same issue
[22:33] <sj_underwater> i was hoping to find that info in the docs somewhere, or at some point have the config script choose clang on darwin
[22:35] <cbsrobot> yeah i was to lazy
[22:35] <sj_underwater> lol
[22:35] <cbsrobot> but  just create an account on trac and add it to the wiki ... others will appreciate it. 
[22:35] <sj_underwater> k
[22:36] <cbsrobot> maybe here http://ffmpeg.org/trac/ffmpeg/wiki/CompilationGuide
[22:38] <sj_underwater> will do. also i may have had an issue w/ the cdio libs missing from the Makefile
[22:38] <sj_underwater> but im not positive, could have been a pkg-config issue
[22:51] <cbsrobot> sj_underwater: i have following line in my ~/.profile
[22:51] <cbsrobot> export PKG_CONFIG_PATH=$PKG_CONFIG_PATH:/usr/local/lib/pkgconfig:/usr/lib/pkgconfig:/opt/local/lib/pkgconfig
[22:51] <cbsrobot> and lately I had to add another path - I don't remember what it was
[22:52] <cbsrobot> ah - of course
[22:52] <sj_underwater> i didnt have an issue using pkg-config, but the libs werent added to the LIBS line in the MF
[22:52] <sj_underwater> so i suddenly got the symbols not found error
[22:52] <cbsrobot> libopebjpeg instelled with brew was putting the openjpeg.pc file in some odd place
[22:53] <sj_underwater> this is CDIO btw
[22:53] <cbsrobot> but I think it's better to copy the pc file to /usr/local/lib/pkgconfig
[22:54] <sj_underwater> i lean towards softlinks myself
[22:54] <sj_underwater> like /usr/local/man instead of /usr/local/share/man, which happens a lot
[22:57] <cbsrobot> sj_underwater: you could add some ideas from http://wiki.videolan.org/OSXCompile to the wiki page
[22:58] <sj_underwater> did you have the cmake issue, by the way?
[22:59] <sj_underwater> where the hardcoded paths were for the old xcode?
[23:00] <cbsrobot> nope
[23:01] <sj_underwater> the path to version.plist is hardcoded and now wrong, so it generates incompatible xcode projects
[23:11] <cbsrobot> sj_underwater: tbh I rather use vim than xcode
[23:12] <cbsrobot> btw. you could also add some info how to setup ffmpeg in xcode ... :-)
[23:12] <sj_underwater> lol, now thats too much, at least 4 me
[23:12] <sj_underwater> im gonna document what it takes to get it to compile
[23:12] <cbsrobot> ok
[23:16] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r50f4f272fe 10ffmpeg/libavcodec/indeo3.c: 
[23:16] <CIA-17> ffmpeg: indeo3: Fix out of reference reading with NULL blocks.
[23:16] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[23:16] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:16] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * recc31630f9 10ffmpeg/libavcodec/mjpegdec.c: 
[23:16] <CIA-17> ffmpeg: mjpegb: Detect changing nb of planes in interlaced video.
[23:16] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[23:16] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:16] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r8e9a0a3568 10ffmpeg/libavcodec/mpc7.c: 
[23:16] <CIA-17> ffmpeg: mpc7: check subband index
[23:16] <CIA-17> ffmpeg: This fixes a overread
[23:16] <CIA-17> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[23:16] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:17] Last message repeated 1 time(s).
[00:00] --- Sun Mar 11 2012


More information about the Ffmpeg-devel-irc mailing list