[FFmpeg-devel-irc] IRC log for 2010-04-18

[irc at mansr.com]

[00:40:39] <hyc> yay, nothing left holding up rtmp seek patches
[01:11:54] <BBB___> how do I move files around via ftp?
[01:15:10] <BBB> hm, apparently rename is the ftp-version of the shell-mv command
[01:15:11] <BBB> odd
[01:17:13] <Kovensky> IIRC there are two commands
[01:17:24] <Kovensky> IIRC they are rnfr and rnto, but don't quote me on that
[01:17:37] <mru> who the fuck uses ftp for such things nowadays?
[01:17:38] <Kovensky> but any ftp client will have a 'mv' command that sends those instructions =p
[01:17:54] <Kovensky> talking about 'any ftp client', lol @ lftp's mv only working with absolute paths on sftp
[01:18:05] <Kovensky> if I give a relative path it whines about file not found
[01:21:18] <BBB> I'm using commandline fto
[01:21:19] <BBB> ftp
[01:21:23] <BBB> and it didn't do it :)
[01:21:37] <BBB> mru: and I'm sorry for being prehistoric, but this is all silly osx comes with, I think
[01:21:48] <BBB> I'm too lazy to set up Finder to log on to upload.mphq.hu
[01:22:18] <mru> are people still using osx?
[01:22:19] <mru> seriously?
[01:22:33] <BBB> no, I'm sort of an outlier
[01:22:34] <mru> I thought only the jobs zombies touched that
[01:22:58] <BBB> it's convenient, it runs MS Word and Photoshop, so I can interact with other university people
[01:23:24] <BBB> you know, these people who are not freetards and don't know and have never heard of and don't ever want to start hearing of GIMP or OOO.org
[01:23:28] <mru> openoffice sucks, but it does read msword files
[01:23:44] <peloverde> not the ones I want to view :(
[01:23:51] <BBB> interoperability is poor enough that I stopped that silly excuse for a miserable product years ago
[01:23:53] <mru> I've never, ever been sent a photoshop file
[01:24:11] <BBB> and believe me, in science publishing, photoshop is all you see
[01:24:18] <BBB> how else do you think they make nature paper figures? :)
[01:24:31] <BBB> they could do it in gimp
[01:24:33] <mru> all I ever saw was LaTeX
[01:24:34] <BBB> but they don't have gimp
[01:24:36] <BBB> so they don't
[01:24:42] <BBB> latex is for math, I'm in biology
[01:24:45] <mru> and a bit of matlab
[01:25:28] <BBB> anyway, MS Word is a good product, I wouldn't trade it for OOO.org if I had to sell my soul to the devil
[01:25:28] <BBB> which I presumably already did by buying a license for it
[01:25:30] <BBB> :)
[01:25:38] <BBB> same for photoshop, it's a very solid product
[01:26:08] <mru> msword is not good by any imagination
[01:26:28] <BBB> fine... but it's interoperable when you interact with other Word users
[01:26:32] <BBB> which is all I care about
[01:26:34] <mru> I haven't used photoshop enough to form an opinion
[01:26:35] <Kovensky> <@mru> are people still using osx? <-- I know some art people that wouldn't trade it for anything
[01:26:49] <Kovensky> at least one of them keeps trying linux but keeps getting disappointed at the tablet support
[01:26:52] <mru> guess I don't talk go many art people
[01:27:00] <BBB> move to williamsburg
[01:27:07] * mru declies
[01:27:10] <mru> +n
[01:29:41] * BBB goes back down then
[01:57:18] <astrange> os x comes with lftp which isn't bad
[01:57:26] <astrange> certainly better than the sftp client
[01:57:42] <mru> sftp is an entirely different thing
[01:58:00] <astrange> not when your server has both
[01:59:51] <mru> still very different
[02:01:02] <astrange> what does that have to do with the client not supporting tab completion?
[02:01:15] <astrange> actually, on os x it doesn't even support backspace. i wonder whose fault that is
[02:02:13] <mru> idiot's
[09:29:07] <KotH> moin
[10:40:39] <Yuvi> hm, random internet commenter says they talked to on2 at one point about licensing vp8 and were told that part of the vp8 license fees went to the mpeg-la
[10:42:48] <mru> rotfl
[12:43:55] <lu_zero> wbs: the patch 4 seems to break seeking in feng
[12:44:13] <lu_zero> and I'm experiencing slowdowns after seeking when patch 3 is in
[12:45:41] <wbs> hmm, ok
[13:33:24] <peloverde> Should I close this as invalid? https://roundup.ffmpeg.org/issue1882
[13:34:10] <peloverde> I really have faa{c,d}
[13:34:58] <peloverde> They really should rename it to nnaa (nero nonfree aac approximation)
[13:35:44] <pJok> have or hate?
[13:36:01] <peloverde> hate, sorry
[13:36:04] <pJok> ;)
[13:36:17] <pJok> my T9 does that all the time when i use my phone for irc
[13:36:47] <peloverde> yes, lets pretend i'm sending this message from my phone, less embarrassing for me
[13:38:13] <pJok> i just pretend that i read hate
[13:38:44] <pJok> and someone should just fix ffaac ;)
[13:39:01] <peloverde> the file isn't aac
[13:39:36] <peloverde> AAC forbids two identical element instance tags on the same non fill channel element in the same frame
[13:40:09] <peloverde> This file fails to meet that criterion and thus fails to be aac
[13:40:42] <wbs> lu_zero: sure that patch #3 adds the slowdowns after seeking, btw? that one should just parse headers and populate the data structures but not actually do anything with it.. so in that case, it's #1 or #2 doing something
[13:41:05] <wbs> lu_zero: but i've seen weird slowdowns in ffplay when watching rtsp-streams, too, haven't figured out exactly what's causing it
[17:38:11] <CIA-81> ffmpeg: stefano * r22901 /trunk/libavformat/avio.c:
[17:38:12] <CIA-81> ffmpeg: Make url_seek() return AVERROR(ENOSYS) rather than AVERROR(EPIPE) in
[17:38:12] <CIA-81> ffmpeg: the case where the seek operation is not defined in the protocol
[17:38:12] <CIA-81> ffmpeg: handler.
[17:38:12] <CIA-81> ffmpeg: stefano * r22902 /trunk/libavformat/aviobuf.c:
[17:38:12] <CIA-81> ffmpeg: Make url_fsize() return AVERROR(ENOSYS) rather than AVERROR(EPIPE) if
[17:38:13] <CIA-81> ffmpeg: the seek operation is not defined in the ByteIOContext.
[18:29:17] <hyc> anyone got a free second to commit this? https://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2010-April/086781.html
[18:49:38] <saste> hyc: I'm working on that right now
[18:49:52] <hyc> cool, thanks
[18:49:54] <saste> hyc: but I wanted to put a meaningful comment
[18:50:45] <saste> hyc: for example it puzzled me that we have the implementation of that read_seek2 but ifdeffed
[18:51:45] <hyc> that implementation is mostly a copy of avformat_seek_file()
[18:52:27] <hyc> as Michael pointed out, it really doesn't support the intended semantics of read_seek2()
[18:52:34] <saste> can you comment on the if (stream_index < 0) check added in rtmp_read_seek()?
[18:52:38] <hyc> but I figured I would leave it in for future reference
[18:52:55] <saste> yes so I'm going to keep that...
[18:53:14] <hyc> sure; the documentation says stream_index must never be -1 at this API level
[18:53:47] <hyc> but ... we need to handle this case....
[18:53:52] <hyc> my mplayer patch depends on it
[18:54:00] <saste> ok
[19:10:12] <CIA-81> ffmpeg: stefano * r22903 /trunk/libavformat/librtmp.c:
[19:10:12] <CIA-81> ffmpeg: Prefer AVERROR(ENOSYS) over AVERROR_NOTSUPP.
[19:10:12] <CIA-81> ffmpeg: AVERROR_NOTSUPP is (maybe) going to be deprecated.
[19:10:13] <CIA-81> ffmpeg: stefano * r22904 /trunk/libavformat/ (librtmp.c flvdec.c): (log message trimmed)
[19:10:13] <CIA-81> ffmpeg: Implement librtmp seek support.
[19:10:13] <CIA-81> ffmpeg: Implement flv_read_seek(), add a missing check on stream_index
[19:10:13] <CIA-81> ffmpeg: and fix timestamp rounding in rtmp_read_seek().
[19:10:14] <CIA-81> ffmpeg: Also add the flv_read_seek2() function, which is not enabled but is
[19:10:14] <CIA-81> ffmpeg: useful as reference.
[19:10:15] <CIA-81> ffmpeg: To actually implement flv_read_seek2() correctly, there would need to
[19:12:30] <hyc> great, thanks
[19:15:51] <saste> hyc: np :)
[20:11:37] <CIA-81> ffmpeg: stefano * r22905 /trunk/libavfilter/avfiltergraph.c:
[20:11:37] <CIA-81> ffmpeg: Fix leak in avfilter_graph_add_filter().
[20:11:37] <CIA-81> ffmpeg: In case of reallocation failure the pointer to the original filter
[20:11:37] <CIA-81> ffmpeg: array was lost. The correct behavior seems to just keep the old array
[20:11:37] <CIA-81> ffmpeg: and count.