From: IRC <wt...@us...> - 2004-01-22 06:42:50
|
******************************************************************* [03:04] Rotty_ (~an...@ch...) left irc: Remote closed the connection [03:15] Nick change: Jaramir -> Jara[zZ] [03:27] sublett (~rv...@21...) left irc: Remote closed the connection [03:48] <thaytoo> ds-work: which wireless chipset have you got? [03:49] <thaytoo> I've had success using ndiswrapper for my centrino, and on my workmate's broadcom one [03:56] thomasvz (~th...@91...) left irc: Read error: 110 (Connection timed out) [03:58] thomasvz (~th...@15...) joined #gstreamer. [04:14] Company (~Company@pD958B357.dip.t-dialin.net) left irc: Read error: 104 (Connection reset by peer) [04:28] <ds-work> thaytoo: don't know, but it apparently works with the ndis wrapper as well [04:36] Company (~Company@pD958B357.dip.t-dialin.net) joined #gstreamer. [04:40] <thaytoo> ndiswrapper is working too well ... one of those double-edged sword projects [04:42] Action: wheels is away: sleep [04:43] <ds-work> perhaps. but that's how NDIS was always supposed to work. Linux should have started using NDIS drivers years ago, imo [04:44] <ds-work> but only recently did ethernet drivers get to be difficult to write [04:48] <thaytoo> it's encouraging manufacturers to only provide binary modules that I don't like [04:50] <thaytoo> I heard unofficially at LCA that Intel might be quite close to releasing a source driver for centrino, at which point I will bid ndiswrapper farewell [04:54] Nick change: harshy -> harshyStudying [05:01] <ds-work> hmmm... [05:01] <ds-work> I'd like to rewrite the swfdec mozilla plugin to use gstreamer [05:02] <ds-work> but I need one particular feature -- I need to be able to get specific errors back from libswfdec [05:02] <ds-work> so that I can offer to file a bug report for bad swf [05:25] <jimmy_dean> ds-work: I'm at a standstill on my equalizer plugin for right now...I've got the same algorithm into my GStreamer plugin as the LADSPA mbeg eq uses and it compiles with no warnings but it's segfaulting and very difficult to figure out...I've emailed the author of the LADSPA mbeq equalizer for help. [05:25] <ds-work> where does it segfault? [05:25] <thaytoo> ds-work: you mean more of an error than just a string? [05:25] <ds-work> have you run it using valgrind? [05:26] <thaytoo> otherwise I'm not sure why even the old error stuff would not work [05:26] <ds-work> thaytoo: yes [05:26] <jimmy_dean> ds-work: actually it successfully completes several iterations sometimes but then segfaults on an assignment to one of my element's members [05:26] <ds-work> thaytoo: I need to get swfdec internal errors out to the app, essentially [05:26] <jimmy_dean> ds-work: but right now it's segfaulting in the FFTW library on a malloc internally (reported by gdb's bt function) [05:27] <ds-work> jimmy_dean: then use valgrind. You are overwriting heap memory if you see errors in malloc [05:28] <jimmy_dean> ds-work: ok, I forgot about valgrind actually [05:28] <jimmy_dean> ds-work: that's a good idea...I still need some clarification from the author though...I need some explanation about the upper limits of some key data arrays [05:29] <thaytoo> ds-work: fire a custom signal ? [05:30] <jimmy_dean> ds-work: it's quite frustrating to be this close to having it work on a basic level and being stuck in segfaults :) [05:34] <ds-work> thaytoo: yeah, probably [06:13] thaytoo (~ja...@ad...) left irc: Remote closed the connection [06:30] thaytoo (~ja...@ad...) joined #gstreamer. [06:34] hallibaby (~hallibaby@pD9511DC4.dip.t-dialin.net) left irc: "Bye bye" [06:52] <ds-work> thaytoo: did you ever commit the navigation event fixes? [07:13] thaytoo (~ja...@ad...) left irc: Remote closed the connection [08:08] apoc (~ap...@dy...) left irc: "Leaving" [08:37] sub_pop (~link@129.210.180.84) left irc: "Client exiting" [08:49] bluejay (~jya...@cs...) left irc: "Leaving" [08:49] kmaraas (~km...@22...) left irc: "Leaving" [09:10] sub_pop (~link@129.210.180.84) joined #gstreamer. [09:10] <steve_b> anyone awake [09:10] <steve_b> ? [09:11] <steve_b> specifically, anyone who knows why the vorbisfile interface isn't liked? [09:11] Nick change: harshyStudying -> harshy [09:13] Nick change: mathrick|sleep -> mathrick [09:13] <mathrick> g'mornin [09:14] <mathrick> steve_b: because it doesn't allow you to cope w/ anything else that Ogg/Vorbis, ie. demultiplexes and decodes in one step. No Ogg/Mp3 possible for ex [09:15] <mathrick> s/thet/than/ [09:15] <mathrick> s/thet/that/ [09:16] <ds> steve_b: because we need oggdemux for other reasons, and vorbis is often stored in AVI and QT [09:18] thaytan (~ja...@ad...) joined #gstreamer. [09:19] <mathrick> yo thaytan [09:19] <thaytan> evening! [09:20] <steve_b> ok, so the virtual io approach is ok then? i'm going to recommend that libsndfile gets something similar [09:20] <steve_b> yop [09:20] <thaytan> yo, steve [09:20] <thaytan> steve_b: http://jaime.hemmett.org/gallery/ [09:21] <mathrick> thaytan: you're horribly wrong as usual. It's 9:21AM, 12 hours to evening left :) [09:22] Action: steve_b checks it out [09:26] <steve_b> thaytan: do you have any of those in higher resolution? [09:33] BBB (~rbultje@213.160.215.2) joined #gstreamer. [09:33] <thaytan> steve_b: only the res that's on there [09:33] <BBB> --- Cannot join #gnome-hackers (User limit reached). [09:33] <BBB> ? [09:33] Action: BBB kicks all gnome people [09:34] <thaytan> ouch! [09:34] <ds> there was a spambot issue earlier [09:34] <thaytan> steve_b: click on the pic to get to the higher resolutions [09:34] <thaytan> (if you didn't already) [09:34] <BBB> ds, could you please try to look at wavparse? it should work, but the explicit caps make it break... removing that and using getcaps/trysetcaps manually makes it work... really weird... I don't know what's wrong with it [09:34] <thaytan> oh! [09:35] <thaytan> steve_b: I'll look at it after dinner [09:35] <mathrick> yo BBB [09:35] <BBB> thaytan: I'll look at your email/code in a few hours [09:35] <BBB> hi mathrick [09:37] <ds> BBB: can you give me an example of a broken pipeline [09:37] <BBB> gst-launch filesrc location=file.wav ! wavparse ! spider ! osssink [09:37] <BBB> with, e.g., an mp3-in-wav file [09:37] <BBB> if you need a sample, kick me and I'll DCC you one, ok? [09:38] <ds> er, /kick you? [09:38] <BBB> no, the other kick [09:38] <BBB> as in 'let me know' :p [09:38] Action: ds kicks BBB [09:38] Action: BBB tries finding a mp3wav file [09:39] <mathrick> BBB: just discussed that with Company yesterday, isn't it that wavparse reports audio/x-raw-int amongst caps and that confuses spider? [09:40] <BBB> don't think so [09:40] <BBB> it does set_explicit_caps() with the actual caps after a while [09:40] <ds> it would depend if the wavparse is inside spider or not [09:40] <BBB> [rbultje@shrek mpeg1videoparse]$ grep RIFF /media/music/* [09:40] <BBB> ... [09:40] <ds> if it's before, spider will do dumb things with it [09:41] <BBB> it's before [09:41] <BBB> but it should still work [09:41] <ds> it's a bug, but not with wavparse :) [09:42] <BBB> ok, so I just leavt it until the new autoplug is done? [09:42] <steve_b> so would a vorbisfile-like interface for libsndfile be a good thing? [09:43] Nick change: harshy -> harshyZ [09:46] <ds> libsndfile should ideally be a demuxer [09:47] <ds> perhaps a demuxer+decoder for some formats [09:47] <steve_b> i'm asking the author to make the most minimal changes possible so that we can use it in autoplug pipelines [09:48] <steve_b> and the only thing holding us back is that it does direct file io - so a virtual io interface seems like the best thing to suggest [09:50] <thaytan> eric is worried about file formats that require a lot of seeking around to decode [09:51] <ds> doesn't matter to us [09:51] <ds> audiofile does that [09:51] <ds> seeking on files is cheap [09:51] <thaytan> I told him that [09:51] <thaytan> and he said he'd give it a go [09:56] Action: mathrick mumbles "StrMod... I/O... Gotta try StrMod" [09:57] <thaytan> steve_b: turns out I only uploaded 640x480 and didn't keep the originals [09:57] <thaytan> stupid mistake :( [09:58] <steve_b> hoho [10:01] <mathrick> steve_b: learned about StrMod recently, it sounds like something that would be of use for virtual I/O, especially if you're not afraid of wrapping templated C++ ;) [10:05] <BBB> ds: still want a wavmp3 sample file? [10:05] <BBB> I've found one [10:05] <ds> yeah [10:06] <BBB> I got this sample file from uraeus, I think [10:06] Action: BBB wonders why ds doesn't accept the dcc [10:06] <ds> ok, I have that [10:06] <ds> it's not working here [10:07] <BBB> you have the file already? [10:07] <BBB> or... [10:07] <ds> I have it on another machine [10:07] <BBB> I can upload it somewhere if you wish [10:07] <BBB> ah, ok [10:08] Action: ds renames it mp3-in-wav.wav [10:12] sxpert_work (~sx...@ra...) left irc: "Leaving" [10:15] sxpert_work (~sx...@ra...) joined #gstreamer. [10:17] swentel (~sw...@d5...) joined #gstreamer. [10:22] jeroen (~je...@hm...) joined #gstreamer. [10:22] <jeroen> hi guyus [10:22] <jeroen> guys even [10:22] <mathrick> hi [10:22] <jeroen> i'm trying to get gstreamer to play an mp3, but i think my LD_LIBRARY_PATH isn't set properly (gst-launch can't find alsasink or mad elements) [10:22] <jeroen> any tips? [10:22] <jeroen> $prefix/lib/gstreamer-0.7/ is in my LD_LIBRARY_PATH [10:23] <mathrick> jeroen: did gst-register-0.7 ? [10:23] <jeroen> mathrick: probably not, thanks :) [10:23] <mathrick> jeroen: you only need $prefix/lib in LD_LIBRARY_PATH [10:23] <jeroen> ok [10:24] <jeroen> so it didn't build the mad element. is there another mp3 decoding element available (ffmpeg)? [10:24] <BBB> ffdec_mp3? [10:24] <BBB> might work [10:24] <mathrick> jeroen: ffdec_mp3 [10:24] <jeroen> sigh [10:24] <jeroen> anybody here using ALSA? [10:24] <BBB> company, leif, I think iain [10:25] <jeroen> Company: around? [10:27] <BBB> doesn't it work? [10:27] Action: BBB heard that more than once :( [10:27] <jeroen> no [10:27] <jeroen> but i think it's related to my kernel 2.6 alsa setup [10:27] <jeroen> xmms can't open the alsa mixer [10:27] <BBB> there's some notes in bugzilla on things you can try to make it work [10:27] <jeroen> i want to know what its trying to open [10:28] <jeroen> i read somewhere that alsa should have a /dev/snd/ dir with devices [10:28] <jeroen> but i don't have that [10:28] <jeroen> cat /proc/asound/devices does list a playback device (among others) [10:29] <mathrick> jeroen: startup script should link /dev/snd/* to /proc/asound/ [10:29] <mathrick> jeroen: or you can use devfs, but that's probably deprecated in 2.6 [10:30] <jeroen> mathrick: the first part doesn't sound logical. what's there to link in /proc? [10:30] <jeroen> mathrick: i'm already using sysfs & udev [10:30] <mathrick> jeroen: lemme check what debian does in its scripts [10:33] <thaytan> on debian for me /dev/snd is a real dir with char devices in it [10:33] <thaytan> you may need a MAKEDEV /dev/snd [10:34] <thaytan> (although I've never needed that) [10:36] <mathrick> jeroen: hmm, it appears to not really link from /proc/asound/, except for sndstat in case of oss emulation [10:36] <jeroen> mathrick: right. same here [10:36] <jeroen> thaytan: i doubt that's the correct way to do it [10:37] <mathrick> jeroen: it's been a while since I used alsa, and always had devfs, so no clue really [10:37] <thaytan> jeroen: yeah, I just don't know who created the device nodes [10:37] <thaytan> one of the alsa userspace packages, I guess [10:38] <mathrick> jeroen: /usr/share/doc/alsa? does it say anything? [10:38] dolphy (~do...@po...) joined #gstreamer. [10:39] <dolphy> morning [10:40] <mathrick> y0 d01phy [10:43] <ds> cool [10:43] <ds> cpu Mhz : 529.927 [10:43] <ds> no more friggin fan [10:44] <jeroen> mathrick: no luck with the docs [10:48] <mathrick> jeroen: "4) If you are using a kernel that contains devfs support, you need to enable [10:48] <mathrick> it, and mount it under /dev to use ALSA, or you can create the ALSA device [10:48] <mathrick> files under /dev/snd/ by using mknod." [10:48] <mathrick> That's from README.Debian. I guess you need to get major-minor pairs from aliases and then try to mknod / udev it [10:48] <jeroen> hmm [10:48] <mathrick> jeroen: no clue however how to make it work with udev [10:48] <mathrick> jeroen: i'd guess it should be done automagically [10:49] <jeroen> mathrick: udev gets its data from sysfs (/sys) [10:50] <mathrick> jeroen: well, if there's /sys/snd, then it should be easy [10:50] <jeroen> mathrick: there isn't [10:50] thomasvz (~th...@15...) left irc: Read error: 110 (Connection timed out) [10:50] <mathrick> jeroen: but if not, then I'm not qualified to help you, I'm still on 2.4 [10:50] <jeroen> mathrick: there's /sys/cdev/major/sound/ (but its empty like all major character devices) [10:51] <mathrick> jeroen: are /proc/asound device files? [10:51] <mathrick> jeroen: /proc/asound/* even [10:51] <jeroen> mathrick: no [10:51] <jeroen> mathrick: /proc is only device info [10:54] <mathrick> jeroen: then I'm out of ideas. Maybe you're better off asking on #<your distro> about that particular matter. Then you can come back when it's alsasink that doesn't work ;) [10:54] <jeroen> perhaps i'll try fedora-devel [10:54] <jeroen> mathrick: thanks anyway :) [10:55] <mathrick> jeroen: np [11:05] tlg (~tl...@21...) joined #gstreamer. [11:08] rk (~rk...@ip...) joined #gstreamer. [11:08] thomasvs (~th...@po...) joined #gstreamer. [11:09] Company (~Company@pD958B357.dip.t-dialin.net) left irc: Remote closed the connection [11:10] <jeroen> mathrick: seems i need a new dev package (all the files/dirs in /dev are part of the dev rpm) [11:12] <mathrick> jeroen: entirely possible [11:13] <jeroen> mathrick: i've got a /dev/snd/ dir now :) (118 files) [11:14] <mathrick> jeroen: wow, that's huge [11:14] Action: mathrick wonders why udev doesn't do its magic in there [11:15] <jeroen> gst-launch-0.7: pcm_plug.c:882: snd_pcm_plug_hw_params: Assertion `err >= 0' failed. [11:15] <jeroen> Aborted [11:17] <mathrick> jeroen: try with ' alsasink device="hw:0" ' [11:17] <jeroen> mathrick: yeah, that works [11:19] <mathrick> jeroen: it's nasty in alsasink that it requires that, but IIRC it was supposed to be alsa's fault, not alsasink [11:19] <jeroen> aha [11:19] <jeroen> xmms just tries to open hw:0 by default [11:19] <jeroen> (i know because it used to give an error msg with that :) [11:21] ack (~sa...@po...) joined #gstreamer. [11:21] Nick change: ack -> sack [11:24] Rotty (~an...@ch...) joined #gstreamer. [11:27] <mathrick> i'm off [11:27] Nick change: mathrick -> mathrick|Uni [11:29] jeroen (~je...@hm...) left irc: "Client exiting" [11:41] <thaytan> interesting [11:41] <thaytan> DEBUG default( 5717) gstmpegparse.c(415):gst_mpeg_parse_loop: have chunk 0xBE [11:41] <thaytan> DEBUG default( 5717) gstmpegdemux.c(742):gst_mpeg_demux_parse_pes: in parse_pes [11:41] <thaytan> DEBUG default( 5717) gstmpegdemux.c(751):gst_mpeg_demux_parse_pes: got packet_length 342 [11:41] <thaytan> DEBUG default( 5717) gstmpegdemux.c(809):gst_mpeg_demux_parse_pes: headerlen is 5, datalen is 339 [11:41] <thaytan> DEBUG default( 5717) gstmpegdemux.c(866):gst_mpeg_demux_parse_pes: we have a unkown packet [11:41] <thaytan> DEBUG default( 5717) gstmpegparse.c(544):gst_mpeg_parse_loop: size: 348, total since SCR: 2048, next SCR: 11212253 [11:41] <thaytan> anyone got a spec handy? [11:43] <thomasvs> thaytan: nope - but if you know the iso code, a handy trick is to look for that in google, along with typical apache strings for file listings [11:43] <thomasvs> thaytan: turns up a *lot* of files in general that you would otherwise have to buy :) [11:43] <thaytan> I was just trying that [11:44] <thaytan> heh [11:44] <thaytan> "Padding Stream" [11:45] Rotty (~an...@ch...) left irc: Read error: 113 (No route to host) [11:45] Rotty (~an...@ch...) joined #gstreamer. [11:49] sublett (~rv...@21...) joined #gstreamer. [11:51] Rotty (~an...@ch...) left irc: Remote closed the connection [11:54] Shoragan (~rid...@d0...) joined #gstreamer. [12:26] <BBB> thaytan: I have the specs here... what do you need? [12:29] <BBB> oh, padding/BE, sorry, missed that [12:40] <thaytan> at the moment, I'm trying to do still frames [12:40] <thaytan> but mpeg2dec doesn't seem to have enough data to output the actual frame [12:40] <thaytan> or something, haven't figured it out yet [12:40] Rotty (~an...@ch...) joined #gstreamer. [12:42] jeroen (~je...@hm...) joined #gstreamer. [12:43] <jeroen> hi [12:43] <jeroen> what's the story with gst-mixer & alsa? [12:43] <jeroen> it says it can't find any devices [12:44] <jeroen> "Sorry, no mixer elements and/or devices found" [12:46] gheet (~gt...@ro...) joined #gstreamer. [12:46] <thaytan> yo [12:47] <gheet> hi [12:49] ChrisHJW_log (~ch...@p5...) joined #gstreamer. [12:56] Shoragan (~rid...@d0...) left irc: Read error: 60 (Operation timed out) [12:57] Shoragan (~rid...@d0...) joined #gstreamer. [12:59] <thaytan> is it invalid for a pad _get operation to return NULL? [12:59] <thomasvs> jeroen: ask Company/BBB [13:01] alley_cat (AlleyCat@pD9519B67.dip.t-dialin.net) joined #gstreamer. [13:02] <jeroen> BBB: any idea what's up with gnome-volume-control not working with alsa? [13:11] <BBB> thaytan: no [13:11] <thaytan> no, it should be valid? [13:12] <BBB> jeroen: company says that ALSA mixing works for him... leif says it doesn't work for him, though, so we know something's bad there... however, it's not fixed yet [13:12] <BBB> thaytan: erm, no, it's invalid [13:12] <BBB> sorry for the confusion [13:12] <thaytan> ok :) [13:28] <jeroen> is there a gstreamer tool to get the metadata from an mp3? [13:28] <jeroen> nautilus crashes (hard, without an error dialog) when i open the Properties dialog of an mp3 file [13:29] <BBB> jeroen: we're working on nautilus-media [13:29] <BBB> rather, thomasvs is working on that [13:29] <thomasvs> BBB: I would if someone would the fuck reply to my mails :) [13:29] Company (~as...@rz...) joined #gstreamer. [13:29] <thomasvs> speak of the devli [13:29] Action: swentel can't even open nautilusmedia audioview [13:29] <thomasvs> devil [13:29] <BBB> er? [13:29] <jeroen> thomasvs: it's a known problem? [13:29] <BBB> company will ;) [13:30] <Company> huh? [13:30] <jeroen> hey Company [13:30] <jeroen> Company: do you know what's up with gnome-volume-control & ALSA? [13:30] <Company> no idea [13:30] <thomasvs> jeroen: yes. the whole system of signalling this info has been deprecated and replaced, and n-m hasn't been ported to use the new one yet [13:30] <jeroen> it just says "Sorry, no mixer elements and/or devices found" [13:30] <jeroen> and refuses to start [13:30] <thomasvs> jeroen: I'm not sure yet though if the new one actually can handle it [13:31] <jeroen> thomasvs: ok. weird though that it crashes without even showing a crash dialog [13:31] <Company> i didn't do any mixer coding, so i have no idea [13:31] <jeroen> Company: people tell me you're the one to talk to :) [13:31] <jeroen> s/people/BBB :) [13:34] <Company> jeroen: he doesn't know what he's talking about :p [13:34] <Company> leif did the coding and i have no clue how far he's come [13:34] <thomasvs> jeroen: mail the mailing list, should be easier to get an answer [13:35] <thomasvs> Company: had any time to check my metadata mail ? [13:35] <jeroen> thomasvs: ok [13:35] <Company> thomasvs: i'm reading it atm [13:35] <jeroen> gstreamer-properties is cool btw [13:35] <jeroen> i like the test-image you see when testing a videosink [13:36] Action: thaytan goes quiet :) [13:36] <thomasvs> thaytan: you see, you need to clean it up, people like it [13:37] <swentel> thomasvs: can't find gnomevfssrc element, is that a bug in nautilus media or something else ? [13:37] <thomasvs> swentel: something else [13:37] <thomasvs> swentel: you're missing an element that you should have [13:38] <thomasvs> swentel: where did you get gst from ? [13:38] <swentel> mmm [13:38] <swentel> standard redhat 9 install i guess [13:38] <thomasvs> don't think so [13:38] <thomasvs> try gst-inspect gnomevfssrc [13:39] <swentel> no suck element [13:39] <swentel> err such [13:39] <thomasvs> so it's not a standard rh9 install [13:39] <thomasvs> rpm -qi gstreamer-plugins ? [13:40] <swentel> lots of ouput .. [13:40] <thomasvs> what version is it ? [13:40] <thomasvs> I'm assuming you're on freshrpms [13:40] <swentel> 0.6.0 [13:40] Action: thomasvs tries to remember way back to rh 9 [13:41] <thomasvs> hm [13:41] <thomasvs> you should have it though [13:41] <thomasvs> swentel: dialog me privately with output of rpm -ql gstreamer-plugins [13:52] mathrick|Uni (~mathrick@Zietka-18.a-inter.net) left irc: Remote closed the connection [14:05] <thaytan> urk [14:05] <thaytan> gstbasicscheduler.c(464):gst_basic_scheduler_chainhandler_proxy: [14:05] <thaytan> (internal error) basic: maximum number of switches exceeded [14:07] LeRoutier (~ler...@dy...) joined #gstreamer. [14:08] <LeRoutier> hello [14:13] Shoragan (~rid...@d0...) left irc: Remote closed the connection [14:14] hyriand (~hy...@no...) left irc: Read error: 60 (Operation timed out) [14:14] <LeRoutier> BBB: seen my .asf problems are partly solved ? [14:14] <BBB> yes [14:14] <BBB> I expected something similar [14:14] <BBB> but still an error, so apparently still a bug [14:15] <Company> thaytan: i dunno if src ! something ! sink src.pad2 ! somethingother ! sink.pad2 is supported without queues [14:16] <thaytan> I'm using queues [14:16] <Company> thaytan: that's what you do with the subtitling... [14:16] <LeRoutier> well. audio works. just video doesn't. the other issues might be gst-player specific (crash at end of play, slider finishing early, ...) [14:16] <thaytan> oh, right [14:17] <thaytan> well, that seems to work, but I can see it could have racing issues [14:18] <LeRoutier> thaytan, working on generic subs or only DVD ones ? [14:19] <BBB> LeRoutier: I've noticed... I'll look at it [14:19] <BBB> (some day...) [14:19] <thaytan> LeRoutier: DVD ones, atm [14:19] <LeRoutier> BBB: no hurry, i can play them within mplayer for now [14:22] dennis_ (~de...@ne...) left irc: Remote closed the connection [14:24] <thaytan> g'night all [14:28] dennis_ (~de...@ne...) joined #gstreamer. [14:31] iain (~ia...@us...) joined #gstreamer. [14:33] <jeroen> gst-plugins no longer builds: [14:33] <jeroen> vorbisfile.c: In function `gst_vorbisfile_update_streaminfo': [14:33] <jeroen> vorbisfile.c:437: `GST_TAG_ENCODER_VERSION' undeclared (first use in this function) [14:33] <jeroen> also GST_TAG_SERIAL and GST_TAG_NOMINAL_BITRATE [14:34] <BBB> hm... I suppose thomasvs edited those [14:34] <BBB> not sure [14:34] <jeroen> * ext/vorbis/vorbisfile.c: (gst_vorbisfile_update_streaminfo), [14:34] <jeroen> (gst_vorbisfile_new_link): [14:34] <jeroen> signal streaminfo through tags [14:35] <jeroen> that's thomasvs' changelog entry [14:36] <BBB> he might have forgotten to commit core changes [14:36] <BBB> or so [14:36] <BBB> do you have latest core? [14:37] <jeroen> lemme check [14:37] <jeroen> doesn't look like it [14:37] <jeroen> i'll rebuild gstreamer first [14:42] Nick change: khalek -> kh_zZz [14:42] hyriand (~hy...@no...) joined #gstreamer. [14:43] <Company> thomasvs: what's the difference between nominal and average bitrate? [14:43] <Company> thomasvs: and why is a serial needed as a tag? [14:44] foser (d0...@22...) joined #gstreamer. [15:01] mxpxpod (~bryan@65.243.233.2) left irc: "Guns kill people as much as spoons made Rosie O'Donnell fat.." [15:11] dilinger (ir...@ws...) left irc: "Terminated with extreme prejudice - dircproxy 1.0.5" [15:16] Company (~as...@rz...) left irc: Remote closed the connection [15:20] <jeroen> this is very weird [15:20] <jeroen> if i try and play an mp3 with gst-player, it segfaults [15:20] <jeroen> when i run it under gdb, it doesn't segfault, but i hear no sound either [15:20] <jeroen> the mp3 plays fine with gst-launch [15:26] <BBB> with spider or mad? [15:26] <jeroen> gst-launch uses ffdec_mp3 [15:26] <jeroen> i'm now building mad [15:26] <jeroen> then rebuilding gst-plugins [15:28] <jeroen> BBB: what's the spider pipeline? filesrc ! spider ! alsasink ? [15:35] <BBB> yes [15:35] <BBB> or .. ! typefind ! spider ! .. [15:36] jeroen (~je...@hm...) left irc: Nick collision from services. [15:37] jeroen_ (~je...@hm...) joined #gstreamer. [15:41] <jeroen_> BBB: i'm getting negotiation errors when using spider [15:41] <jeroen_> gst-launch-0.7 filesrc location="Its_My_Life.mp3" ! typefind ! spider ! alsasink device=hw:0 [15:41] <jeroen_> RUNNING pipeline ... [15:41] <jeroen_> ERROR: from element /pipeline0/spider0/id3types0: Internal GStreamer error: pad problem. File a bug. [15:41] <jeroen_> Additional debug info: [15:41] <jeroen_> gstpad.c(2208):gst_pad_set_explicit_caps: [15:41] <jeroen_> failed to negotiate (try_set_caps returned REFUSED) [15:41] <jeroen_> ERROR: from element /pipeline0/alsasink0: Internal GStreamer error: negotiation problem. File a bug. [15:41] <jeroen_> Additional debug info: [15:41] <jeroen_> gstalsasink.c(346):gst_alsa_sink_loop: [15:41] <jeroen_> ALSA format not negotiated [15:41] <jeroen_> Segmentation fault [15:41] <jeroen_> BBB: this is still without the mad plugin though [15:42] <BBB> hm... [15:42] <BBB> ffdec_mp3 should be used if it's available [15:42] <jeroen_> i'm now building libid3tag (seems gst-plugins wants that too in order to build the mad element) [15:43] <BBB> I think mad depends n it [15:43] <BBB> yes [15:46] Action: BBB cheers as ffmpeg seems to basically compile using autotools [15:47] <jeroen_> yep [15:49] <LeRoutier> well, libid3tag and mad have the same version number in gnome-pkg-tools. so i suppose those are 2 parts of the same lib [15:50] <BBB> thye're the same lib here [15:51] <LeRoutier> so, it has 2 .pc files for pkg-config, with names 'id3tag' and 'mad' and both with version 0.15.0b [15:51] <BBB> id3tag can be used alone [15:52] <BBB> w/o mad [15:52] <LeRoutier> but not the other way [15:52] <jeroen_> they're separate libs (on freshrpms at least) [15:58] <jeroen_> BBB: with the mad element present, the pipeline works fine [15:58] Action: jeroen_ tries gst-player again [16:00] <BBB> ok [16:00] <jeroen_> well, it doesn't segfault anymore [16:00] <BBB> apparently, ffmpeg wrapping still needs more work [16:00] <BBB> we knew that ;) [16:00] <jeroen_> but it doesn't play the mp3 file either [16:00] <jeroen_> seems to hang when i press play [16:00] Action: BBB doesn't know about gst-player [16:00] <jeroen_> BBB: who do i talk to about that? [16:01] <BBB> dolphy? [16:01] <jeroen_> dolphy: around? [16:01] <BBB> but I'm really delaying all capsnego and media-playback errors until the new autoplugger is finished [16:01] <jeroen_> ok [16:01] <BBB> (and that's ot disinterest, but I consider it rather useless to do it differently) [16:01] <jeroen_> ok :) [16:02] <jeroen_> hmm, spider is b0rked [16:02] <jeroen_> mp3 stops playing after ~20 seconds [16:03] Action: BBB is busy enough updating ffmpeg [16:03] <jeroen_> BBB: i'm not asking you to fix them :) [16:03] <jeroen_> just noticing a bug [16:03] <BBB> :) [16:04] <BBB> 3 things left for ffmpeg [16:04] <BBB> 1) networking libs [16:04] <BBB> 2) amr_[wn]b* [16:04] <BBB> 3) ppc profiling [16:04] <BBB> then it's converted [16:04] <jeroen_> what's 2? [16:04] Action: BBB cheers [16:04] <BBB> dunno [16:04] <BBB> weird support libs [16:05] <jeroen_> seems to run fine on ppc btw (i'm on ppc) [16:05] <BBB> yeah [16:05] <BBB> but without optimizations [16:05] <jeroen_> aha [16:05] <BBB> and it takes a lot of time [16:05] <BBB> (to wrap it) [16:05] <BBB> I'm trying to improve that [16:05] <BBB> so we can update more oftne [16:07] <jeroen_> cool [16:09] <jeroen_> could somebody give me a pipeline to playback a quicktime movie? [16:11] rk (~rk...@ip...) left irc: "leaving" [16:13] <BBB> gst-launch filesrc location=file.mov ! qtdemux name=demux .video_00 ! { queue ! ... ! ffcolorspace ! ximagesink } demux.audio_00 ! { queue ! .. ! osssink } [16:14] <jeroen_> thanks [16:14] <jeroen_> BBB: what goes in the "..." part? [16:15] <BBB> depends on the movie [16:15] <BBB> try spider ;) [16:15] <jeroen_> ok :) [16:16] <jeroen_> ERROR: from element /pipeline0/thread0/spider0/sink_ident: Could not determine type of stream. [16:16] <jeroen_> ERROR scheduler(23252) gstoptimalscheduler.c(2074):gst_opt_scheduler_iterate: [GstOptScheduler@0x1024da10] in error state [16:16] <jeroen_> ERROR: from element /pipeline0/thread1/spider1/sink_ident: Could not determine type of stream. [16:16] <jeroen_> ERROR scheduler(23252) gstoptimalscheduler.c(2074):gst_opt_scheduler_iterate: [GstOptScheduler@0x10263610] in error state [16:16] <jeroen_> gst-launch-0.7 filesrc location="SacredPlanet_T1_9471_0300.mov" ! qtdemux name=demux .video_00 ! { queue ! spider ! ffcolorspace ! xvimagesink } demux.audio_00 ! { queue ! spider ! alsasink device=hw:0 } [16:17] <BBB> ehm... [16:17] Action: BBB desperately waits for the new autoplugger [16:17] <BBB> if you know the type, I can give the exact elements [16:18] <jeroen_> dunno [16:18] <jeroen_> it's a trailer i downloaded from apple.com :) [16:18] <BBB> hexedit it and search for strings [16:18] <BBB> 'mp4v', 'mp4a', these sort [16:19] sjoerd (sj...@be...) got netsplit. [16:19] alley_cat (AlleyCat@pD9519B67.dip.t-dialin.net) got netsplit. [16:19] tlg (~tl...@21...) got netsplit. [16:19] sub_pop (~link@129.210.180.84) got netsplit. [16:19] wheels (~sc...@ds...) got netsplit. [16:19] sxpert (~sx...@sx...) got netsplit. [16:19] taaz (~dlehn@66.37.66.32) got netsplit. [16:19] <jeroen_> BBB: how do i hexedit it? [16:20] <thomasvs> I am getting sick of benjamin's attitude [16:20] sjoerd (~sj...@si...) joined #gstreamer. [16:20] <jeroen_> BBB: http://bvim-qt.vitalstream.com/SacredPlanet/SacredPlanet_T1_9471_0300.mov (that's the one i downloaded (only 2MBs)) [16:20] <thomasvs> is it THAT strange to expect us to fix regressions ? [16:22] sjoerd (~sj...@si...) left irc: Killed by weber.freenode.net () [16:22] alley_cat (AlleyCat@pD9519B67.dip.t-dialin.net) returned to #gstreamer. [16:22] tlg (~tl...@21...) returned to #gstreamer. [16:22] sub_pop (~link@129.210.180.84) returned to #gstreamer. [16:22] wheels (~sc...@ds...) returned to #gstreamer. [16:22] sjoerd (sj...@be...) joined #gstreamer. [16:22] sxpert (~sx...@sx...) returned to #gstreamer. [16:22] taaz (~dlehn@66.37.66.32) returned to #gstreamer. [16:22] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by brunner.freenode.net [16:22] sjoerd_ (~sj...@si...) joined #gstreamer. [16:23] hallibaby (~hallibaby@pD9511DC4.dip.t-dialin.net) joined #gstreamer. [16:23] <BBB> jeroen_: checking, hold on [16:23] sub_pop (~link@129.210.180.84) left irc: Excess Flood [16:23] sjoerd (sj...@be...) left irc: Network is unreachable [16:23] Action: thomasvs thinks about giving up on gstreamer in general [16:23] <jeroen_> thomasvs: don't do that just because of one person [16:23] <jeroen_> thomasvs: and don't you have a job hacking on gstreamer now? :) [16:24] sub_pop (~link@129.210.180.84) joined #gstreamer. [16:24] <BBB> ugh [16:24] <thomasvs> jeroen_: nothing prevents us from forking internally [16:24] <BBB> thomasvs: ... please tell me you're kidding? [16:24] <thomasvs> jeroen_: I know he means well, but ... [16:24] <thomasvs> jeroen_: he keeps trying to make things personal and I have no idea why [16:24] <jeroen_> thomasvs: i can't comment. i don't follow the discussions on the list too well [16:24] <thomasvs> jeroen_: all I want is to fix regressions and make GStreamer look good in GNOME [16:24] <BBB> thomasvs: in general, I think pretty much all of us are against regressions... one or two (company, maybe ds) cause a lot of them [16:24] <BBB> should we all suffer? [16:25] <thomasvs> BBB: I don't know. I have the feeling nobody cares if we have regressions or look bad in GNOME [16:25] <thomasvs> well, no one on our team :) [16:25] <BBB> I do [16:25] <BBB> it's not fair to say that [16:25] <thomasvs> well, ok [16:25] <BBB> I've spent many hours on gnome-media to fix stuff and make mixer/GSR look good [16:25] <thomasvs> yeah, like I did before [16:25] <thomasvs> but the end result is still [16:25] <thomasvs> some apps still don't work [16:25] <BBB> GSR is almost working right now [16:25] <BBB> the only thing that didn't work was flac encoding last time I tried [16:25] <thomasvs> and now I get attacked because I dind't complain even more loudly before :) [16:25] <BBB> and it didn't crash once [16:26] <thomasvs> BBB: yeah, I had to ask you some questions about that btw [16:26] <BBB> jeroen_: the movie is zlib encoded, I'm not good at that... ;) [16:26] <BBB> thomasvs: go ahead [16:26] <jeroen_> BBB: heh [16:26] <thomasvs> BBB: because the record/convert cycle is stupid :) [16:26] <BBB> I told you before, I fully agree but I won't fix it [16:26] <thomasvs> BBB: I wanted to put in the profiles stuff, and then I found out there's a two step system in saving files [16:26] <BBB> I'd rather work on gst-rec and replace GSR on the long term [16:26] <thomasvs> BBB: no, I know. but it means I can't integrate the profile stuff just yet. [16:27] <BBB> you can...? [16:27] <thomasvs> BBB: I just have to bite the bullet and go through the code completely I guess [16:27] <BBB> the profile stuff only matters for saving, right? [16:27] <thomasvs> yeah [16:27] <BBB> recording is pretty much osssrc ! wavenc ! filesink [16:27] <thomasvs> but it assumes that recording is done directly to disk [16:27] <BBB> saving is filesrc ! spider ! <profile_pipeline_part> ! filesink [16:27] <thomasvs> anyway, that's solvable [16:27] <thomasvs> this metadata stuff is a lot harder [16:28] <BBB> putting metadata in GSR? [16:28] <BBB> that's not hard [16:28] <BBB> set them on wavenc while recording [16:28] <thomasvs> tell me how to deal with him so we can agree on what to do [16:28] <thomasvs> no, metadata in nautilus-media [16:28] <BBB> and keep tags while transcoding/encoding [16:28] <BBB> oh, that [16:28] <BBB> ues [16:28] <BBB> that sucks [16:28] <BBB> I'll try to help fixing tagging when ffmpeg is done, ok? [16:28] <thomasvs> because right now I am very much considering putting in the old stuff back next to his tagging stuff, so at least something works [16:28] Action: BBB actually hoped to finish gst-rec before going to the US, but that won't work out... :( [16:31] <BBB> thomasvs: I think I can finish this tomorrow, wanna help testing then? [16:31] <thomasvs> BBB: sure [16:31] <BBB> and fixing up, finetuning, perfectizing and then I'll send a patch [16:32] <BBB> two things left [16:32] <BBB> and I don't know how to fix one of the two [16:34] Shoragan (~rid...@d0...) joined #gstreamer. [16:38] Shoragan (~rid...@d0...) left irc: Client Quit [16:39] <BBB> and I just fixed the othwe [16:39] Action: BBB happy [16:44] <BBB> from 2012 lines to 1387 [16:44] <BBB> acceptable? :) [16:44] <BBB> (configure.ac + Makefile.am's) [16:44] <BBB> and from 48,5kB to 33,5kB [16:45] <BBB> thomasvs: can I make shared libs without using libtool? [16:45] <BBB> I can only do .a libs without libtool, it seems [16:48] <thomasvs> BBB: yeah, you can [16:48] <thomasvs> BBB: the target to use is whatever_LIBRARIES = ...so [16:49] <thomasvs> iirc [16:51] Action: BBB will try in a sec [16:51] <BBB> I'm now simply usign libtool [16:52] <BBB> but they won't like that [16:52] <thomasvs> BBB: check the autobook to make sure, it's explained [16:52] <thomasvs> you may have to make configure prefer shared libs by default though [16:53] <BBB> ok [16:53] <BBB> btw, regarding this tagging... [16:53] <BBB> and all [16:53] <BBB> my worst fears did come true [16:54] <thomasvs> which is ? [16:54] <BBB> far after the intended API freeze, we started changing core subsystems, and we end up fixing issues caused by it and fixing just the most-needed plugins so at least something works [16:54] <BBB> and most stuff is just broken [16:54] Action: BBB hopes he's wrong [16:54] <thomasvs> yeah, I know [16:55] <thomasvs> what do you want me to do ? [16:55] <BBB> nothing [16:55] <BBB> just fix stuff, like me ;) [16:55] <thomasvs> it's not like I haven't a) asked him a lot "what are you oging to do" and b) I'm not sure this solves the issues [16:55] <BBB> patch seems to work [16:55] <BBB> wanna test/ [16:55] <BBB> it's against latest CVS of ffmpeg [16:55] <thomasvs> what does it patch ? [16:55] <thomasvs> autotool it ? [16:55] <BBB> ffmpeg CVS HEAD [16:55] <BBB> autotoolificaiton [16:56] <thomasvs> why are you doing this anyway ? :) [16:56] <thomasvs> but sure, I don't mind testing [16:56] <BBB> they'll never do it themselves [16:56] <BBB> maybe they'll accept it if we do it [16:56] <thomasvs> maybe ? [16:56] <BBB> they don't mind automake/stuff [16:56] <thomasvs> :) [16:56] <BBB> and we can always use it ourselves [16:56] <thomasvs> I read the thread you started :) [16:56] <thomasvs> but sure, send me the patch [16:56] <BBB> don't forget that AM_CONFIGURE() is a lot easier than the rest [16:56] <LeRoutier> BBB: they prefer keeping their shity shell code you talked about ? [16:56] <BBB> LeRoutier: yes [16:57] <BBB> thomasvs: bugzilla? or dcc? [16:57] <LeRoutier> i'm not the only braindamaged people so [16:57] <thomasvs> mail [16:57] jeroen_ (~je...@hm...) left irc: "Client exiting" [16:58] <BBB> oh, and about that thread... I'm kind of sick of how michael responds... but on the other hand, he is a genious, so I'll have to cope with it [16:58] yippi (~br...@nw...) joined #gstreamer. [16:59] <BBB> hi brian! [16:59] <BBB> thomasvs: YGM [17:01] <thomasvs> YGM ? [17:01] <BBB> you got mail [17:01] <BBB> or CYM [17:01] <BBB> check your mail [17:03] Action: BBB tries make distcheck [17:05] TD (~mi...@bd...) joined #gstreamer. [17:05] <TD> does libavcodec have a stable abi? [17:06] <BBB> no [17:07] <alley_cat> hmm, i can't play any audio with gst-launch-0.7 anymore [17:07] <alley_cat> ERROR: from element /pipeline0/vorbisfile0: Internal GStreamer error: pad problem. File a bug. [17:08] Company (~Company@pD958BDBE.dip.t-dialin.net) joined #gstreamer. [17:08] <TD> suckage [17:08] <TD> why do they offer an API if it's not stable? [17:09] <Company> huh? [17:09] <BBB> Company: not gst ;) [17:09] <BBB> TD: dunno... they odn't really offer shared libs anyway [17:10] <TD> ok [17:10] <TD> Company: ffmpeg :) [17:13] <BBB> thomasvs: don't bother make distcheck yet, I'm working on that [17:16] TD (~mi...@bd...) left #gstreamer. [17:29] sublett (~rv...@21...) left irc: Read error: 54 (Connection reset by peer) [17:31] sublett (~rv...@21...) joined #gstreamer. [17:39] hyriand (~hy...@no...) left irc: "leaving" [17:41] BBB (~rbultje@213.160.215.2) left irc: "Client exiting" [17:42] ChrisHJW (~chr...@p5...) joined #gstreamer. [17:44] Shoragan (~rid...@d0...) joined #gstreamer. [17:47] sub_pop (~link@129.210.180.84) left irc: "Client exiting" [17:53] markov (~me...@po...) joined #gstreamer. [18:00] Nick change: LeRoutier -> LeRaway [18:03] Action: thomasvs just figured out vorbistag handles ogg metadata [18:12] kmaraas (~km...@22...) joined #gstreamer. [18:14] whitehat4 (~whi...@st...) joined #gstreamer. [18:17] tlg (~tl...@21...) left irc: "Client exiting" [18:22] <Company> vorbistag doesn't handle bitrates yet though [18:23] <thomasvs> hm, I don't see why id3 has bitrate tags [18:23] <thomasvs> there's no point, really [18:23] <Company> it doesn'T have bitrate tags [18:23] <Company> vorbis has them ;) [18:23] <thomasvs> oh, I thought you said it did [18:23] Nick change: harshyZ -> harshyWork [18:24] <Company> and some mod like formats have that [18:24] <thomasvs> ah, length [18:24] <thomasvs> why is length an id3 tag ? [18:24] <Company> they set a "desired decoding method" [18:24] <Company> because length is hard to figure out for vbr [18:25] <thomasvs> ok, so people can change it and have it be wrong too I assume ? [18:25] <Company> sure [18:26] <thomasvs> hm [18:26] <Company> just like the vorbis bitreate tags [18:26] <thomasvs> so if I have a CBR mp3, that mad can figure out the length of, but it also has an id3 length tag, and they get both signaled as tags [18:26] <thomasvs> there's not really a way to distinguish them ? [18:26] <Company> well, mad figures out the length via queries [18:26] Action: thomasvs dislikes id3 length tags by design [18:27] <thomasvs> ok, but the general method of asking gstreamer "how long does this file play for" is still through a query ? [18:28] trow (~tr...@ds...) joined #gstreamer. [18:28] <Company> if you want to be extra sure, that is the preferred method [18:29] <Company> btw: why the nominal/average bitrate distinction? [18:29] <Company> and why needs a serial number be a tag [18:29] <Company> ? [18:29] <thomasvs> I've wondered too why they want to distinguish between the two [18:29] <thomasvs> but there is a difference [18:29] <thomasvs> nominal is what you asked it to aim for during encoding [18:30] <thomasvs> it doesn't necessarily, and mostly doesn't, amount ot the same as average [18:30] <thomasvs> and serial is a tag because vorbis has it [18:30] <thomasvs> don't really see how else to export it ? [18:30] <Company> you don't want to export it [18:30] <Company> the serial is just a random number that distinguishes multiple ogg streams [18:30] <thomasvs> why not ? [18:30] Action: wheels is back [18:31] <Company> it's an ogg think, not a vorbis thing [18:31] <thomasvs> right, and ? [18:31] <Company> ogg thing even [18:31] <Company> that's a format thing, not a tag [18:31] <thomasvs> so you're saying oggdemux should export it ? [18:31] <Company> no, i'm saying noone should export it [18:31] <Company> it's totally useless outside ogg [18:32] <thomasvs> ogginfo reports it, and if gstreamer can't report what ogginfo reports, it's not a good enough multimedia framework imo :) [18:32] <Company> uhm [18:32] <thomasvs> anyway, I don't see why serial is any less valid than replaygain_albumgain [18:32] <Company> what is serial? [18:33] <thomasvs> a number that has to be unique inside an ogg file/stream, to identify each logical stream [18:33] <Company> replaygain-albumgain is the volume adjustment you should do for a given album [18:33] Action: wheels reaches for his big notebook-o-media-standards. [18:33] <Company> yeah, and this is essential for anything outside oggdemux where? [18:33] <thomasvs> no, replaygain-albumgain is a broken hack that someone thought of at one point [18:34] <Company> (i don't think we should support albumgain, but it has a use) [18:34] <thomasvs> Company: the point is, gstreamer should be able to provide what media-specific tools do [18:34] <Company> we shouldn't support serial numbers as much as we don'T support avi flags [18:34] <thomasvs> Company: otherwise we can't replace them [18:34] <thomasvs> Company: and as for albumgain, no, it's a broken way of looking at it [18:34] <Company> thomasvs: in that case it should probably be a property on oggdemux [18:34] <thomasvs> Company: you don't say "play this album this much louder than normal" [18:34] <thomasvs> Company: you say "this album is about this loudness", choose a good default setting, and leave it there. [18:35] <wheels> (It's used as a unique identifier for multiplexed or chained logical streams within the same physical stream -- it's not used at the Vorbis level.) [18:35] <thomasvs> Company: it is broken because it is an absolute value for something that isn't absolute [18:35] <Company> thomasvs: agreed [18:35] <Company> thomasvs: can we get rid of serial now? [18:35] <thomasvs> Company: sure, I don't mind it being in oggdemux. I just implemented it in vorbisfile because the vorbisfile API has it. If someone puts it in oggdemux, then it can go from vorbisfile [18:35] <Company> thomasvs: and either nominal or average bitrate? [18:36] <thomasvs> Company: nominal and average are different [18:36] <Company> thomasvs: for a user of tags they're not [18:36] <thomasvs> here's ogginfo on a random track for me [18:36] markov (~me...@po...) left irc: "bbl" [18:36] <thomasvs> Nominal bitrate: 128,000000 kb/s [18:36] <thomasvs> Average bitrate: 81,085953 kbps [18:36] <thomasvs> also, bitrates shouldn't be UINTS [18:36] <thomasvs> they probably should be FLOATS [18:37] <Company> you know that tags are supposed to give the user informations, not extract everything from inside the file [18:38] hallibaby_ (~hallibaby@pD9511B3C.dip.t-dialin.net) joined #gstreamer. [18:38] <Company> there's a reason why theres no "first-byte-of-data", "second-byte-of-data" etc tags... ;) [18:38] <thomasvs> that's one way of looking at it. the other way is "gstreamer should be a framework that can provide the information current tools have in a generic way" [18:38] <thomasvs> also, why does it matter ? [18:38] <thomasvs> an app is free to not display it [18:38] <thomasvs> why limit the info we're giving if it's available ? [18:38] <Company> because the app developer is confused about what to display [18:39] <thomasvs> why is he confused ? [18:39] Action: wheels notes that when doing the call for TagLib it doesn't report the serial numbers of streams, but does report the various bitrates... [18:39] <thomasvs> you sure you're not arguing just because I did the work ? [18:39] <Company> if you have nominal, average, targetted, corrected, minimum and maximum bitrate [18:39] Action: thomasvs gets slightly annoyed at trying to do the right thing [18:39] <Company> yes, i'm sure :) [18:39] <Company> gstreamer should provide a useful abstraction in its tags [18:40] <Company> not export everything that might be useful as a tag [18:40] <thomasvs> that's an opinion, not a fact :) [18:40] <thomasvs> anyway, feel free to ignore the tag in the apps you're about to write. [18:40] <wheels> Company: Well, I think it really makes sense to have two APIs -- one for people that give a shit and one for people that don't. Have a "bitrate()" and decide what you want that to mean; you can provide the other stuff for the few that care. [18:40] <Company> that has been a fact since i wrote the tagging stuff [18:41] <thomasvs> Just like I have to add custom checking to differentiate between metadata and format data [18:41] <thomasvs> in the apps I did write [18:41] <Company> wheels: yeah, that fine, query the oggdemux element :) [18:41] <Company> thomasvs: i'd like to differentiate between them, but there's so much formats that it doesn't work [18:42] <thomasvs> Company: as I said, I'm fine with oggdemux reporting it, just as long as there is some way of getting it. [18:42] walters (wa...@ve...) joined #gstreamer. [18:42] <thomasvs> Company: so as soon as there is a standard documented way of getting nominal bitrate and serial I'll remove it from vorbisfile [18:43] <Company> we should change average bitrate to nominal bitrate [18:43] <thomasvs> change what specifically ? [18:43] <Company> average bitrate is length_in_bytes/length_in_time [18:43] <Company> both can be queried [18:43] <thomasvs> yep. [18:43] <Company> we should change the meaning of the "bitrate" tag to nominal [18:43] <thomasvs> hm, that would be nonintuitive [18:44] <thomasvs> people asking for "bitrate" would get something not remotely close to what they expect to get [18:44] <thomasvs> we should drop BITRATE and keep NOMINAL_BITRATE [18:44] sub_pop (~link@129.210.180.84) joined #gstreamer. [18:44] <Company> that's fine with me, too [18:44] <thomasvs> only, we don't have an API that we are going to keep to get length in bytes and length in time, do we ? [18:45] <thomasvs> or are we sticking with the event stuff until 1.0 ? [18:45] <thomasvs> ehm, query stuff ? [18:45] <Company> until next version at least [18:45] <Company> and after that queries will be replaced by query events [18:46] <thomasvs> gst-launch -t filesrc location=/home/audio/songs/Elbow\ -\ Red.ogg ! vorbistag ! fakesink [18:46] <thomasvs> this needs oggdemux, right ? [18:46] <Company> the method of asking for length or position will be kept [18:46] <Company> right [18:46] <thomasvs> it throws an error about the format of the comment though [18:46] <thomasvs> wonder where it got there [18:46] <thomasvs> Company: did you see ds's idea for a double printf ? [18:46] <thomasvs> Company: I'll switch to that if it's ok with you [18:47] <Company> it just takes the second buffer coming in and assumes it's the format :) [18:47] <Company> thomasvs: yeah, i like that [18:48] <Company> thomasvs: the srial in an ogg file IMO is the same as what you get when you enable debugging output in qtdemux btw (which is about 500 lines for setup only): it's stuff that's interesting for the element, but not for tags [18:54] hallibaby (~hallibaby@pD9511DC4.dip.t-dialin.net) left irc: Read error: 110 (Connection timed out) [19:03] Uraeus (~csc...@3j...) joined #gstreamer. [19:05] <Uraeus> ello [19:05] <Uraeus> thomasvs: around? [19:08] <thomasvs> yes [19:08] <thomasvs> hm, hang on [19:08] sublett (~rv...@21...) left irc: "I like food, food is good!" [19:08] <thomasvs> his stuff is still f1, f2, ... [19:08] <thomasvs> not really sure how that is better [19:08] <Uraeus> thomasvs: I think I come up with the perfect solution to our rpm packaging problems :) [19:08] gheet (~gt...@ro...) left irc: "Download Gaim [http://gaim.sourceforge.net/]" [19:09] <Uraeus> thomasvs: if you get GStreamer running on mathias machine we can just let him maintain the full set of GStreamer RPMS :) [19:09] <thomasvs> Uraeus: a) he already packages gst b) he doesn't do it fedora.us style [19:09] <thomasvs> Uraeus: but nice try [19:09] <thomasvs> Uraeus: I already came up with the perfect solution though [19:09] <thomasvs> Uraeus: you help like you said you would :) [19:10] <Uraeus> thomasvs: as for the request to not depend on libcaca, well afaik so does videolan depend on it like we do, and videolan is the libcaca authors own project [19:10] <Uraeus> thomasvs: actually I was on the fedora.us site the other day looking for the QA packages, they are supposed to be in the bugzilla right? [19:11] <Uraeus> thomasvs: I know he does packages, but only a subset and I guess his motivation to expand them and keep them up to date is small as long as it is borked on his system :) [19:11] <thomasvs> Uraeus: so that doesn't solve anything then, just move a problem. [19:11] <thomasvs> Uraeus: seriously, help out or not mention it :) otherwise you just get me depressed [19:12] <Uraeus> ok ok :) [19:12] <Uraeus> hmm, wonder if I should move the computer into the middle of the livingroom, cold air litterally reaches for me from the walls [19:14] hyriand (~hy...@no...) joined #gstreamer. [19:16] <wheels> Uraeus: You could just get an Itanium and heat the whole building with it. :-) [19:16] <walters> what's the debug argument of a gst error for? [19:17] Nick change: LeRaway -> LeRoutier [19:19] <Uraeus> wheels: heh, that might work :) these old buildings done right after ww2 doesn't really handle -15 celsius very well, think around -10 is their limit :) [19:19] markov (~me...@po...) joined #gstreamer. [19:19] <wheels> Uraeus: Ah, student housing in Oslo? [19:19] <wheels> Uraeus: I always wonderd what students there did for housing since it seemed prohibitively expensive for, well, non-students. [19:20] <Uraeus> wheels: nope, standard appartment :) and I haven't been a student in 7 years :) [19:20] <wheels> Uraeus: Oh, right. I think I knew that. [19:21] Action: wheels just tends to assume that all OSS developers are either students or not too far from it. :-) [19:21] <Uraeus> wheels: well many students rent houses toghether etc. [19:21] markov (~me...@po...) left irc: Client Quit [19:24] <Uraeus> wheels: personally I have been feeling that the average age has gone up over the years since so many people who where with us from the start is still around; Linux, Alan, Miguel, Havoc, Mattias Ettrich and so on [19:25] <Company> we're missing the youngsters! [19:25] <wheels> Uraeus: Yeah, that's definitely true, but well -- out of those Miguel, Havoc and Ettrich I think are all 30 or less... [19:25] <Uraeus> wheels: yeah, 30 seems to be the golden age for a Linux person these days :) [19:26] <wheels> And well, there are a lot of 30 year old students in KDE. :-) [19:26] Action: wheels notes that Ettrich is aparently 32 now. [19:27] <Uraeus> heh, well if I had still been a student I would be very worried about how I was ever to pay down my 1 million student loan :) [19:27] Action: wheels still has about $15,000 to go. [19:28] markov (~me...@po...) joined #gstreamer. [19:28] <Uraeus> wheels: you mean that you owe close to 1 million USD? [19:29] <wheels> Uraeus: Oh, no -- that in paying off my student loans I still have about $15,000 left (which means I've already payed for about $14,000) [19:29] markov (~me...@po...) left irc: Client Quit [19:30] <wheels> Though for PhD students in the US it's not uncommon to owe ~$200,000 at the end. [19:31] <Uraeus> wheels: ah, well I have around 15 000 USD left to pay on my student loan [19:31] <Uraeus> too [19:31] Action: wheels notes that he's loving the current exchange rate. :-) [19:32] <Uraeus> I am not that big a fan, the low dollar is hurting the value of my Oracle shares [19:33] <wheels> It's made my students loans go down by 20% relative to my salary in the last year... [19:35] <Uraeus> wheels: guess you are happy the currencies didn't go the other way :) [19:36] <Uraeus> thomasvs: what should I do on fedora? just update the bugzilla entries saying 'Built and tested, works for me' ? [19:36] Action: wheels nods [19:36] <wheels> Hmm, need to go for groceries. [19:38] Action: Uraeus votes to lobotomize whoever came up with the fedora QA system [19:40] <wheels> Redhat? [19:40] Action: wheels reaches for a scapula. [19:40] <Uraeus> wheels: nope, the fedora system was put together by volunteers [19:40] markov (~me...@po...) joined #gstreamer. [19:41] <Uraeus> wheels: the problem is that they try and immitate a company in having a very heavy QA process, which basically kills their productivity and makes potentional contributers like me demotivated every time I visit their site [19:42] <wheels> Hold that thought. The grocery store closes in 20 minutes. [19:42] <Uraeus> hehe :) [19:42] Action: Uraeus tries to hold on to the demotivation [19:43] markov (~me...@po...) left irc: Client Quit [19:44] <Company> go debian! :) [19:47] <thomasvs> Uraeus: just read the QA process [19:47] <thomasvs> Uraeus: also, debian's system is quite more elaborate, so be thankful :) [19:49] <walters> well... [19:49] <walters> it's pretty hard to get in debian, but once you do there aren't many constraints on what an individual developer can do. [19:49] <thomasvs> walters: "marrilat" ) :) [19:50] <walters> heh [19:51] <Uraeus> thomasvs: I just read it, the fact that each friggin package needs 3 sigs, with one being from a short list of people, in order to be published is a system made for dying in its own overhead [19:51] <thomasvs> Uraeus: look. this is the LAST time I'm going to say this to you. [19:51] <thomasvs> Uraeus: either you help out, or you don't. [19:52] <thomasvs> Uraeus: but PLEASE STOP COMPLAINING ABOUT THE SYSTEM :) [19:52] <thomasvs> Uraeus: I love you too much for me to bash your head in because you keep complaining about it, but I will if I have to :) [19:53] Action: Uraeus bites in the rest of his remarks about fedora.us [19:54] <Company> edora doesn't have a way to install unstripped libs by default, does it? [19:54] <Company> and fedora doesn't either? [19:55] iain (~ia...@us...) left irc: "leaving" [19:55] <thomasvs> I don't think libs are stripped [19:55] <thomasvs> and yo ucan install debug packages to get source code in gdb [19:56] <Company> just install the debug packages and the next time you run gdb the debugging symbols are there? [19:56] <thomasvs> yep [19:56] <Company> cool [19:56] Nick change: LeRoutier -> LeRaway [20:00] datadevil (~maa...@da...) joined #gstreamer. [20:00] <datadevil> eveuhning [20:00] hallibaby_ (~hallibaby@pD9511B3C.dip.t-dialin.net) left irc: "Bye bye" [20:00] markov (~me...@po...) joined #gstreamer. [20:09] whitehat4 (~whi...@st...) left irc: "Client exiting" [20:15] <wheels> Mmm. Cola. [20:15] ploum (~pl...@19...) joined #gstreamer. [20:17] <Uraeus> great seahorse doesn't compile with the latest version of gpgme [20:18] Nick change: LeRaway -> LeRoutier [20:20] ploum_ (~pl...@19...) joined #gstreamer. [20:20] <ploum_> well, but I still have a big problem with gstreamer : [20:21] <ploum_> http://gnomesupport.org/forums/viewtopic.php?t=5212 [20:23] ploum (~pl...@19...) left irc: Read error: 104 (Connection reset by peer) [20:24] iain (~ia...@us...) joined #gstreamer. [20:25] <Uraeus> thomasvs: sorry, helping with fedora is simply not my piece of pie [20:25] <Uraeus> hi iain [20:25] <iain> hello [20:26] Action: ds-work suggests filing a bug [20:26] <Uraeus> ploum_: is the sound better for you using madplay? [20:32] <Misirlou> Company: pingxx0r [20:33] <Company> Misirlou: ponx [20:33] dolphy (~do...@po...) left irc: "Network down, IP Packets delivered via UPS" [20:34] <Company> afk [20:35] Action: iain tries to work out why mozilla jumps desktops after a page has loaded and why that is considered a desirable thing [20:36] <ploum_> Uraeus, I will try [20:48] <ploum_> Uraeus, yes, the sound is better with mpg321 than madplay [20:49] thomasvs (~th...@po...) left irc: Read error: 113 (No route to host) [20:51] Nick change: datadevil -> ddtv [20:54] <ploum_> I have this on my both systems.. [20:55] <ploum_> but on this, I must admit that I have a strange problem : no way to control the volume. Applications don't care about the master volume and care about PCM volume only if it's muted [21:04] somex1 (~Ab...@d2...) joined #gstreamer. [21:04] Zeenix (Zeeshan@203.135.11.75) joined #gstreamer. [21:05] <Zeenix> hi [21:06] sxpert (~sx...@sx...) left irc: "Client exiting" [21:07] sxpert (~sx...@sx...) joined #gstreamer. [21:12] sublett (~rv...@21...) joined #gstreamer. [21:30] thomasvs (~th...@24...) joined #gstreamer. [21:34] <ds-work> moo [21:36] <thomasvs> biab [21:36] thomasvs (~th...@24...) left irc: "Client exiting" [21:37] thomasvs (~th...@24...) joined #gstreamer. [21:37] <thomasvs> sweet. it works. [21:37] Action: thomasvs just shelled out 19.95 for linuxant's driverloader [21:37] <thomasvs> does that make me a bad fosser now ? [21:39] <ddtv> wb thomasvs [21:39] Nick change: ddtv... [truncated message content] |