IRC Logs

*******************************************************************
[03:01] kmaraas ([email protected]) left irc: "Leaving"
[03:02] smoke ([email protected]) got lost in the net-split.
[03:08] Kurosu ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[03:09] Nick change: mx|gone -> mxpxpod
[03:12] linux ([email protected]) joined #gstreamer.
[03:13] linux ([email protected]) left irc: Client Quit
[03:47] BBB ([email protected]) left irc: "Client exiting"
[03:48] iain ([email protected]) left irc: "leaving"
[03:55] <Company> nobody loves me - no i wanna show off my cool audioconvert fix, but everyone interested is gone
[03:56] jcsston` ([email protected]) joined #gstreamer.
[03:56] <mxpxpod> Company: show me :)
[03:57] <mxpxpod> even tho I'm nobody
[03:57] <bitshifter> fix for the number-of-channels issue?
[03:57] <Company> yeah
[04:08] somex1 ([email protected]) joined #gstreamer.
[04:14] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[04:14] <thaytan> Company: what's your fix?
[04:15] <ds> moo
[04:16] <Company> thaytan: a sophisticated fixation function :)
[04:17] Action: Company goes milking ds
[04:17] <Company> ds: why don't we link plugins to libgstreamer?
[04:26] <ds> because plugins are loaded _by_ gstreamer
[04:27] <ds> and it's possible to load plugins from a slightly differnt ABI, and not have things crash
[04:28] <ds> in fact, gst-register should be able to load 0.6 plugins, realize they're not 0.8, and ignore them
[04:28] <ds> if they were linked with libgstreamer-0.6, it would be hell
[04:41] <Company> the problem is that this does lead to problems for bindings
[04:42] <thaytan> shouldn't the bindings link to libgstreamer?
[04:42] <Company> because the symbols from libgstreamer might not be loaded GLOBAL
[04:43] <Company> if the bindings are opened as RTLD_LOCAL by the other languages runtime, loading the plugin won't work
[04:44] <Company> that's why i've told the java bindings guy (benow) to do a g_module_open ("libgstreamer-0.8.so", 0); before loading the first plugin
[04:44] <Company> which is ugly
[04:45] <Company> GStreamer itself should make sure the symbols are available for the plugin that is opened, even when libgstreamer was opened with RTLD_LOCAL
[04:46] <Company> IMO at least
[04:46] <ds> the bindings module should be linked with libgstreamer.so
[04:46] <ds> ie., for python, _GStreamer.so
[04:47] <Company> the bindings module is linked to libgstreamer
[04:48] <ds> then the bindings module should dlopen libgstreamer.so without RTLD_LOCAL
[04:48] <Company> isn't that rather ugly?
[04:48] <ds> so what?
[04:48] <Company> and shouldn't the core do this if it requries this?
[04:48] <ds> bindings are glue to make this kind of stuff work
[04:49] <ds> the core has absolutely no business doing that
[04:49] <Company> the core has to make sure it can load plugins
[04:50] <Company> and if that requires globally available libgstreamer symbols it should make sure it gets those
[04:50] <ds> if your application causes libgstreamer.so to be loaded strangely, it's _your_ duty to make it work
[04:50] <Company> opening a plugin with RTLD_LOCAL is not strang
[04:50] <Company> e
[04:51] <ds> yes, it is
[04:51] <Company> hm
[04:52] <Company> if you say so
[04:52] <ds> because it's not the same as, say, rhythmbox linking with libgstreamer.so
[04:54] <Company> how does linking work anyway?
[04:54] <Company> are those symbols you linked to available like with RTLD_GLOBAL or RTLD_LOCAL?
[04:54] <Company> or even different?
[04:55] <ds> huh?
[04:55] <ds> in the rb case, they're global
[04:56] <Company> yeah
[04:56] <Company> but in the dlopen a lib case, they're the same as what you opened the lib?
[04:57] <Company> the symbols from the libs you linked the dlopened lib against, that is
[04:57] <Company> (that was confusing)
[04:57] <ds> if you dlopen a lib, all its dependent libs get the same flag
[04:58] <Company> ok, got it
[05:12] bitshifter ([email protected]) left irc: "bye"
[05:15] bitshifter ([email protected]) joined #gstreamer.
[06:49] steve_b ([email protected]) joined #gstreamer.
[07:26] <thaytan> ~/o ohhhh, make a selection, a selection a selection. make a selection from the dee vee ddeeee! o/~
[07:28] Action: steve_b dances the jig
[07:30] <thaytan> hallo mr steve! 
[07:30] <thaytan> how's sunny nz?
[07:30] <thaytan> any babies yet?
[07:30] <steve_b> sunny, for once
[07:30] <steve_b> nope, just a huge wriggling bump
[07:30] <steve_b> any day now...
[07:31] <thaytan> *suspense* :)
[07:31] <steve_b> making progress with dvd?
[07:32] <thaytan> I'm recreating mpeg2subt at the moment
[07:32] <steve_b> bbs, i'm just going to reboot
[07:33] steve_b ([email protected]) left irc: Remote closed the connection
[07:36] steve_b ([email protected]) joined #gstreamer.
[08:04] <taaz> thaytan: are you making a dvd app too or just using gst-player?  vdv needs some love.
[08:05] jcsston` ([email protected]) left irc: Read error: 60 (Operation timed out)
[08:05] Uraeus ([email protected]) joined #gstreamer.
[08:05] <thaytan> taaz: I'm testing with gst-launch atm
[08:06] <thaytan> later, I want it to work with vdv and libgstplay
[08:06] <taaz> yeah! ;)
[08:08] <taaz> that will be tricky
[08:08] <thaytan> how do you mean?
[08:08] <taaz> a full featured dvd pipeline is probably more than libgstplay spidering can handle
[08:09] <thaytan> the idea is to have a special subclassed dvddemux that handles audio/subtitle channel selections
[08:10] <thaytan> the tricky part (IMO) is convincing it to put mpeg2subt at the end to overlay subtitles
[08:10] <thaytan> but for now, I want to get mpeg2subt working, and figure that other out later :)
[08:11] <Uraeus> good morning
[08:11] <thaytan> Hi Uraeus
[08:13] <taaz> good plan
[08:13] <taaz> these debs are too freaking confusing
[08:14] <taaz> i'm tempted to just tell all the users to fuck off and just make one big moster package with everyting! ;)
[08:14] <thaytan> hahaha
[08:15] <taaz> seeing that i get bug reports on the meta package that includes everything saying no one needs both oss and alsa... i'm guessing that idea won't be accepted by at least 1 user.
[08:15] <Uraeus> what is the gdb command again to get the stacktrace from all threads?
[08:16] <taaz> there are a boat load of weird deps in there.  bunch of plugins that depend on pango and x stuff and so on.
[08:16] <thaytan> thread apply all bt
[08:16] <taaz> thr<tab>a<tab><tab>bt<ret>
[08:16] <thaytan> so many <> to type ;)
[08:17] <taaz> hmm. ssems arts pluging just didn't bother to even build.  sigh.
[08:18] jcsston ([email protected]) joined #gstreamer.
[08:21] <taaz> hmm.  caca-config is called with just --libs instead of --plugin-libs.  that's no good.  on debian systems at least that means it's using libcaca vs libcaca_pic
[08:27] mathrick|sleep ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[08:43] ChrisHJW ([email protected]) joined #gstreamer.
[09:07] markey ([email protected]) joined #gstreamer.
[09:26] apoc__ ([email protected]) joined #gstreamer.
[09:34] apoc_ ([email protected]) left irc: Read error: 60 (Operation timed out)
[09:36] kmaraas ([email protected]) joined #gstreamer.
[09:49] wheels ([email protected]) left irc: "ok, making it to my plane today.  :-)"
[09:50] ChrisHJW_log ([email protected]) left irc: Read error: 60 (Operation timed out)
[09:53] kwm ([email protected]) joined #gstreamer.
[10:02] ChrisHJW ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:08] pb_ (~pb@2002:5169:e72f:1:240:95ff:fe30:aa84) joined #gstreamer.
[10:15] somex1 ([email protected]) left #gstreamer.
[10:17] Nick change: apoc__ -> apoc
[10:17] <apoc> morning
[10:17] steve_b ([email protected]) left irc: Remote closed the connection
[10:19] Shoragan ([email protected]) joined #gstreamer.
[10:20] <thaytan> hiya
[10:21] Shoragan ([email protected]) left irc: Client Quit
[10:39] hyriand ([email protected]) left irc: "bleh"
[10:39] Nick change: markey -> markey|afk
[10:52] hyriand ([email protected]) joined #gstreamer.
[11:17] apoc ([email protected]) left irc: "Leaving"
[11:23] sjoerd ([email protected]) left irc: Read error: 113 (No route to host)
[11:26] Shoragan ([email protected]) joined #gstreamer.
[11:26] sjoerd ([email protected]) joined #gstreamer.
[11:36] ChrisHJW ([email protected]) joined #gstreamer.
[11:41] ChrisHJW_log ([email protected]) joined #gstreamer.
[11:58] kmaraas ([email protected]) left irc: Client Quit
[11:58] kmaraas ([email protected]) joined #gstreamer.
[12:00] pb_ (~pb@2002:5169:e72f:1:240:95ff:fe30:aa84) left irc: "Client exiting"
[12:37] markey|afk ([email protected]) left irc: Read error: 60 (Operation timed out)
[12:39] jdahlin ([email protected]) joined #gstreamer.
[12:41] <Company> good morning angels
[12:42] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[12:43] <thaytan> hi Company :)
[13:04] markey|afk ([email protected]) joined #gstreamer.
[13:04] Nick change: markey|afk -> markey
[13:15] ensonic ([email protected]) joined #gstreamer.
[13:15] <ensonic> hi ho
[13:16] <thaytan> yo ensonic
[13:16] Action: ensonic is on the cebit2004 now and showing gst to people ;-)
[13:16] The_Company ([email protected]) joined #gstreamer.
[13:22] sublett ([email protected]) joined #gstreamer.
[13:30] <ensonic> on monday I'll probably have to submit another thousand bug reports :-(
[13:35] Company ([email protected]) left irc: Read error: 110 (Connection timed out)
[13:37] <Uraeus> hehe :)
[13:45] Nick change: markey -> markey|afk
[13:45] markey|afk ([email protected]) left irc: "leaving"
[13:55] sublett ([email protected]) left irc: "Why are all the really good songs short?"
[13:57] jdahlin ([email protected]) left irc: Read error: 60 (Operation timed out)
[14:09] teuf ([email protected]) joined #gstreamer.
[14:09] ain ([email protected]) joined #gstreamer.
[14:15] <Uraeus> hi teuf and ain
[14:17] <teuf> hi
[14:18] <ain> hi
[14:20] <Uraeus> ain: managed to get past the bugs that where holding Marlin up?
[14:28] <ain> umm, I have, but only because I've applied ds's patch
[14:28] <ain> but I don't think that patch'll make it into CVS
[14:29] <ain> well, thats the GStreamer side of the bugs
[14:29] <ain> my life has bugs too that keep holding it up as well
[14:29] <ain> but I can't really expect you lot to fix them :)
[14:29] <Uraeus> ah
[14:29] <ain> (but it would be nice if you could)
[14:30] <Uraeus> ain: if I where religious I would pray for you :)
[14:31] <ain> thanks
[14:31] <ain> I think I just need germany to disappear and all my problems will be solved
[14:32] <ain> or one of them anyway
[14:32] smoke ([email protected]) joined #gstreamer.
[14:34] <Uraeus> ain: well if you look at Germany over the last 200 years they have been gradually shrinking, but it would probably still take a while if the trend continued
[14:35] <ain> yeah...before july would be the best time really
[14:35] <ain> which gives us time to ship company somewhere safe
[14:37] Action: ensonic is german too ;-)
[14:41] <Uraeus> ensonic: don't worry we get you out too :)
[15:02] <ensonic> had you guys many problems with gst-0.8 - I've tried it on a freshly installed suse-9 pro and there is nearly nothing working :-( (manybe many things are due to suse broken devel packages)
[15:04] <thaytan> ensonic: it's working quite well for me.
[15:04] <Uraeus> ensonic: I am having some Totem issues, but apart from that things are working well
[15:05] jdahlin ([email protected]) joined #gstreamer.
[15:06] Rotty ([email protected]) joined #gstreamer.
[15:16] <Uraeus> jdahlin: find those beers yesterday?
[15:17] <jdahlin> Uraeus: indeed
[15:17] <jdahlin> Uraeus: livejournal.com/users/jdahlin/
[15:18] <Uraeus> poetry :)
[15:21] <jdahlin> yeah, is it not? :)
[15:23] <Uraeus> jdahlin: http://librsvg.sourceforge.net/images/nuvolaspatialus.png <- I updated Nuvola for spatialus and gnome 2.6 :)
[15:25] <jdahlin> Uraeus: cool, I found a bug in it the other day (for 2.6)
[15:25] <jdahlin> home icon looked a bit too transparent
[15:25] <jdahlin> hmm, maybe it was another theme
[15:27] <jdahlin> no, but a different home icon [from the one found in your screenshot]
[15:27] <Uraeus> no that was Nuvola, I fixed that to, just used the home icon
[15:27] <jdahlin> right, here it's a folder + a small home icon
[15:27] <Uraeus> the one with the bug was one I tried making myself, but due to Sodipodi issues it got borked
[15:27] foser ([email protected]) joined #gstreamer.
[15:28] <jdahlin> Uraeus: meet two lovely norwegian girls last night
[15:28] <jdahlin> Amazingly I could understand what they said :-)
[15:28] <ain> "If KDE starts using GStreamer I will probably switch to Gnome. Whats the point of KDE if it only wraps up Gnome technology?"
[15:29] <jdahlin> ain: damn gnome luser!
[15:30] <Uraeus> ain: on kde multimedia?
[15:31] <Uraeus> ain: if so mail back and ask whats the point of KDE if it only wraps X windows technology
[15:31] jdahlin ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[15:31] <ain> dot.kde.org
[15:32] <ain> another quality line "GStreamer's inclusion on freedesktop.org was _exactly_ not an endorse, although a deliberate attempt by the GNOME camp to spread general confusion and misinformation."
[15:35] <thaytan> well, looks like we can't fool them
[15:35] <thaytan> time to come clean, everyone
[15:37] <walters> ain: heh
[15:38] jdahlin ([email protected]) joined #gstreamer.
[15:38] <ain> http://dot.kde.org/1079750571/1079842421/1079843793/1079862251/1079871438/1079874852/1079875786/
[15:39] <ain> any truth in that one?
[15:39] Action: jdahlin wonders if he missed a reply to his insult
[15:45] <Uraeus> ain: well making gstreamer 'itself' network transparent might be hard, but I have still not heard what can't be accomplished with using plugins. I would think that warzaw-pakt project is proof of network transparancy
[15:50] Action: Uraeus replies to that last network transp. post
[15:51] markey ([email protected]) joined #gstreamer.
[15:55] Marsupilami23 ([email protected]) left irc: Remote closed the connection
[15:55] walters ([email protected]) left irc: "out"
[15:57] yellowcalx ([email protected]) joined #gstreamer.
[15:59] yellowcalx ([email protected]) left #gstreamer ("Leaving").
[16:00] Marsupilami23 ([email protected]) joined #gstreamer.
[16:07] <Uraeus> markey: some gstreamer flamefesting on dot.kde.org today :)
[16:07] <markey> yes, followed it
[16:07] <markey> rather stupid discussion, tho
[16:07] <markey> not on technical merits
[16:07] <Uraeus> markey: well some people want you to be the judge : For a comparison between NMM & Gstreamer, please note that the amaroK developpers are currently trying NMM. Let's wait for their feedback since they have already tried Gstreamer .
[16:08] <Uraeus> :)
[16:08] <markey> yes, we hope to bring both GST and NMM engines to on par with our aRts engine ASAP
[16:08] <markey> only then we can really judge
[16:08] <Uraeus> markey: but what does it prove? that both gst and nmm can be used to play back mp3 files?
[16:09] <markey> yes
[16:09] <markey> and more
[16:09] <markey> how easy it is to use these frameworks in a KDE environment
[16:09] <markey> maybe
[16:09] <markey> dunno
[16:09] <markey> it's a first field test ;)
[16:10] <Uraeus> heh :)
[16:10] sublett ([email protected]) joined #gstreamer.
[16:10] <markey> there's also our crossfading stuff, visualisations, effects and so on
[16:10] <markey> it will be interesting to see how well that maps to the frameworks
[16:11] <Uraeus> markey: and tag editing :)
[16:11] <markey> no
[16:11] <markey> that's handled by taglib
[16:11] <markey> (wheels tech)
[16:11] <Uraeus> doesn't need to (once we have our ogg muxer done)
[16:11] alley_cat ([email protected]) left irc: Read error: 110 (Connection timed out)
[16:11] Kaetzchen ([email protected]) joined #gstreamer.
[16:12] md` ([email protected]) left irc: 
[16:13] <markey> at the moment I'm a bit stuck with the GST engine
[16:13] <Uraeus> markey: but I hope you guys are aware that if you do go with NMM you will yourself have to do more of the basic development, like writing demuxers/muxers etc.
[16:13] <markey> stopped working when converted our engines to plugins
[16:13] Nick change: The_Company -> Company
[16:16] md` ([email protected]) joined #gstreamer.
[16:17] <Company> markey: g_mopdule_open ("libgstreamer-0.8.so", 0); before calling gst_init
[16:18] <markey> uh
[16:18] <markey> thank you
[16:18] <markey> didn't know about it
[16:18] <markey> I suspected something is fishy..
[16:18] <markey> will try right after lunch
[16:18] <Company> solved that yesterday with the java bindings
[16:19] <markey> hold on, gotta have lunch
[16:19] <markey> bbl
[16:19] Nick change: markey -> markey|lunch
[16:19] <jdahlin> Maybe something wrt to RTLD_LAZY
[16:32] <Company> more likely with RTLD_LOCAL
[16:40] <thaytan> alright! back to the future + subtitles and I AM OFF TO BED :)
[16:41] <Company> gnight and grats
[16:44] Nick change: markey|lunch -> markey
[16:44] <markey> re
[16:44] <markey> ok let's try
[16:47] <markey> Uraeus: inalid conversion from int to GModuleFlags
[16:47] <markey> what flag should I use
[16:48] <markey> |  G_MODULE_BIND_LAZY    = 1 << 0,
[16:48] <markey> |  G_MODULE_BIND_LOCAL   = 1 << 1,
[16:48] <markey>   G_MODULE_BIND_MASK    = 0x03
[16:48] <Uraeus> markey: ask Company :) I have no clue :)
[16:48] <markey> sounds good :)
[16:48] <markey> Company: any clue?
[16:49] <markey> I'll try BIND_LAZY
[16:49] <markey> suits me best
[16:51] <ain> markey: cast it to GModuleFlags no?
[16:51] <markey> and crash
[16:51] <markey> next one
[16:51] <ain> (GModuleFlags) G_MODULE_BIND_LAZY | G_MODULE_BIND_LOCAL etc?
[16:52] <markey> but which flags are correct?
[16:52] <taaz> jdahlin: you were saying?  gthread and player.py work better?  at some point i tried all the schedulers and they failed only slightly differently each time.  not enough that i could claim the scheduler had anything to do with the failures
[16:52] <ain> markey: man dlopen explains the flags
[16:52] <jdahlin> taaz: not player.py, but other tests
[16:53] <ain> markey: you probably want something to do with GLOBAL
[16:53] <ain> oh, hmmm
[16:54] lupusBE ([email protected]) joined #gstreamer.
[16:54] <ain> there is no GLOBAL in GModuleFlags
[16:54] <markey> right :)
[16:54] <Company> markey: no flag at all
[16:55] <Company> markey: use (GModuleFlags) 0
[16:55] <ain> http://developer.gnome.org/doc/API/2.0/glib/glib-Dynamic-Loading-of-Modules.html#GModuleFlags
[16:55] <markey> ok
[16:55] <Company> ain: did i mention i fixed audioconvert?
[16:55] <ain> no?
[16:55] <ain> sweet, thanks
[16:55] <ain> you're defintely a german I'll save
[16:56] <Company> what's so bad about germans?
[16:57] <ain> one of them's going out with the girl I want to go out with
[16:57] <Company> ah
[16:57] <ain> and I don't know anything more specific than "he's in germany"
[16:57] <ain> so, sorry
[16:57] <Company> understood
[16:57] <markey> hmm still won't work
[16:57] <markey> |** (process:4394): CRITICAL **: how to remove plugins?
[16:57] <markey> |
[16:57] <markey> |(process:4394): GStreamer-CRITICAL **: Factory for `mad' has no type
[16:57] <Company> probably because all german girls take the turkish guys ;)
[16:57] <markey> no difference really
[16:58] <Company> markey: could you run with env var GST_DEBUG=GST_PLUGIN*:5 plz>?
[16:58] <markey> yes
[16:59] <Uraeus> I got those factory messages when it turned out I had an old registry file laying around somewhere
[16:59] <jdahlin> taaz, http://www.gnome.org/~jdahlin/player.py
[16:59] <ain> yeah, I had to remove my ~/.gstreamer-0.8/registry.xml file to stop it
[17:00] spyder482 ([email protected]) joined #gstreamer.
[17:00] <markey> lots of debug output
[17:00] <Company> well, there are 4 different reasons why it can show up
[17:00] <markey> doesn't reveal much tho
[17:00] <Company> markey: the last 5 lines
[17:00] <jdahlin> taaz: works fine, with and without use_thread
[17:00] <markey> <flood>
[17:00] <markey> |LOG   GST_PLUGIN_LOADING( 4423) gstplugin.c(203):gst_plugin_register_func: plugin "/usr/local/lib/gstreamer-0.8/libgsttagedit.so" initialised
[17:00] <jdahlin> taaz: or were you thinking about something different?
[17:00] <markey> |INFO  GST_PLUGIN_LOADING( 4423) gstplugin.c(365):gst_plugin_load_file: plugin "/usr/local/lib/gstreamer-0.8/libgsttagedit.so" loaded
[17:00] <markey> |LOG   GST_PLUGIN_LOADING( 4423) gstplugin.c(203):gst_plugin_register_func: plugin "/usr/local/lib/gstreamer-0.8/libgstmad.so" initialised
[17:00] <markey> |INFO  GST_PLUGIN_LOADING( 4423) gstplugin.c(365):gst_plugin_load_file: plugin "/usr/local/lib/gstreamer-0.8/libgstmad.so" loaded
[17:00] <markey> |
[17:00] <markey> |** (process:4423): CRITICAL **: how to remove plugins?
[17:01] <markey> |
[17:01] <markey> |(process:4423): GStreamer-CRITICAL **: Factory for `mad' has no type
[17:01] <Company> does gst-launch work?
[17:01] <markey> yes
[17:01] <markey> I tried the test command
[17:01] <markey> sinewave to osssink
[17:01] <markey> works fine
[17:02] <Company> can you try /path/to/file.mp3 ! spider ! osssink ?
[17:02] <markey> ok
[17:02] <jdahlin> markey: there is a problem dlopening the mad plugin I guess, unresolved symbols or something
[17:02] <Uraeus> markey: and remove your .gstreamer-0.8/registry.xml file
[17:02] <jdahlin> markey: ldd it and see what happens
[17:03] <Company> we need to stabilize our registry handling, damnit
[17:04] <markey> /path/to/file.mp3 ! spider ! osssink   <-- same error as with amaroK
[17:04] <markey> maybe my GST install is b0rked?
[17:04] <Company> i'd expect that
[17:04] <markey> yes
[17:04] <Uraeus> tried removing your user registry file?
[17:04] <markey> yes
[17:04] <markey> removed it all
[17:04] <Uraeus> hmm
[17:05] <markey> maybe some stale files around from 0.7
[17:05] <Company> try running gst-register again as root
[17:05] <markey> did that 5 times already
[17:05] <Company> yeah, checking for 0.7 files might help, too
[17:05] <markey> I'll reinstall..
[17:06] <Company> someone should find a way to repreoduce that error reliably ;)
[17:06] <markey> well it installed fine
[17:06] <markey> strange
[17:07] <ain> simple
[17:07] <ain> sudo rm /usr/var/cache/gstreamer-0.8/registry.xml
[17:07] <ain> gst-register-0.8
[17:07] <ain> sudo gst-register-0.8
[17:08] <ain> oh, typical
[17:08] <ain> now it works
[17:09] <Uraeus> I think the problem occurs if you have a user registry and then updates the system one and they get out of sync
[17:10] <Uraeus> I think having a user registry only works if it is identical to the sytem one
[17:10] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[17:11] <lupusBE> are there gstreamer 0.8 packages for redhat rawhide available with all codecses enabled?
[17:11] <lupusBE> so I can play xvid and mpeg etc
[17:12] dolphy ([email protected]) joined #gstreamer.
[17:12] dolphy ([email protected]) left irc: Client Quit
[17:14] Action: ain starts looking for online "Learn to play the tin whistle"
[17:14] <Uraeus> lupusBE: yes, thomas made some, they are available from the gstreamer website
[17:17] dolphy ([email protected]) joined #gstreamer.
[17:17] dolphy ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[17:25] <lupusBE> k thx
[17:26] dolphy ([email protected]) joined #gstreamer.
[17:30] <Uraeus> hi dolphy
[17:30] <dolphy> re
[17:38] sublett ([email protected]) left irc: "Why are all the really good songs short?"
[17:40] <ensonic> cioa
[17:40] ensonic ([email protected]) left irc: "Leaving"
[17:45] apoc ([email protected]) joined #gstreamer.
[17:45] <apoc> hi
[17:46] smoke ([email protected]) got netsplit.
[17:50] jdahlin ([email protected]) left irc: "Leaving"
[17:52] smoke ([email protected]) got lost in the net-split.
[17:54] <Uraeus> hello apoc
[17:54] <Uraeus> apoc: did you manage to grab that wmv file I gave you the url to?
[17:56] ChrisHJW_log ([email protected]) left irc: Read error: 60 (Operation timed out)
[17:57] <lupusBE> k can someone tell me why xvid has patent issues ?
[17:58] <Uraeus> lupusBE: prolly cause someone has taken out a patent on something xvid does :)
[17:58] <ain> http://discomachinegun.prettypeople.org/~iain/maybe-borked.png
[17:59] <ain> I'm not sure...something might be broken somewhere
[17:59] Action: Uraeus wonders what the shot is supposed to tell me :)
[18:00] <ain> well, generally there should be something in the lower half of the wave form
[18:00] <Uraeus> ah :)
[18:00] <lupusBE>  there even is (or used to be) a patent on YUV as a color space for image compression... 
[18:00] <lupusBE> bleh 
[18:00] <Company> ain: good thing there is something at the lower half ;)
[18:01] <ain> Company: yeah...thats just weird...
[18:01] Action: ain loads the mp3 in soundforge
[18:01] <apoc> Uraeus: yes ... It works fine with ff_asfdemux but not with our native asfdemux
[18:02] ploum ([email protected]) joined #gstreamer.
[18:02] <apoc> Uraeus: I'll have a look
[18:02] <apoc> hi ploum
[18:02] <ploum> hello
[18:03] <ain> Company: wow, it is actually supposed to look like that
[18:03] <ain> Company: thats the first sample I've seen where everything seems to be positive
[18:04] <ain> oh, it's "broken"
[18:04] <ain> the centre line seems to be shifted somehow
[18:05] <ain> and I've found a large bug in marlin's drawing code too...
[18:08] mathrick ([email protected]) joined #gstreamer.
[18:09] <Uraeus> hi mathrick
[18:09] <ain> Company: thanks for doing whatever you did to audioconvert
[18:10] <Uraeus> ain: I am compiling Marlin now, ok if I commit a small change to look for 0.8 instead of 0.7 ?
[18:10] <mathrick> ho
[18:11] <ain> oh, yeah, sure
[18:11] <ain> I have that locally :)
[18:11] NFusi0n ([email protected]) left irc: Remote closed the connection
[18:18] ChrisHJW ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[18:20] ChrisHJW ([email protected]) joined #gstreamer.
[18:21] <apoc> Have someone read Miguel blog about "Audio/Video" ?
[18:24] ploum ([email protected]) left irc: "Leaving"
[18:26] Nick change: Kaetzchen -> alley_cat
[18:28] <Uraeus> apoc: just saw it, I don't think we amount to 160M
[18:29] lupusBE ([email protected]) left irc: Remote closed the connection
[18:29] <apoc> Uraeus: yep
[18:30] <Uraeus> actually that argument is just stupid. If that is what he is so worried about he should have Ximian package gstreamer with 1 plugin per RPM and he will get his small, mean and lean system
[18:31] ChrisHJW_log ([email protected]) joined #gstreamer.
[18:36] <mathrick> Uraeus: and 160MB of control files ;P
[18:38] Action: Uraeus posted counter blog
[18:38] Action: mathrick reads Miguel's entry
[18:40] <ain> I don't get miguel's problem
[18:40] <ain> how will having a standardised API mean GStreamer won't take 160megs?
[18:41] <ain> if my blog was public and on planet.gnome I'd ask him that
[18:44] <mathrick> gah, it's rants all over the place, with no real approach to issues raised by Havoc :\. Entirely unlike usual Miguel
[18:44] <mathrick> and 160M for gst, hah
[18:46] <mathrick> "I would love to see a standardized C-based interface that every one of them exposed and allow people to pick one over the other" - IMHO bullshit, as real as Java's WORA (hint, hint)
[18:57] <ain> Uraeus: did it compile that time?
[18:57] <Uraeus> haven't failed yet at least :)
[18:58] <ain> sounds promising
[18:58] ploum ([email protected]) joined #gstreamer.
[18:58] <ain> mathrick: I agree...but I don't see how having one of those will stop gstreamer taking up 160meg (which I'm not sure where he gets that number from)
[18:59] Shoragan ([email protected]) left irc: "Leaving"
[18:59] Action: ain starts getting worried about his interview
[18:59] <Uraeus> ain: the 160M is gstreamer+dependencies (including glibc and a kernel etc.)
[19:02] <mathrick> ain: neither do I, and that 160M taken out of nowhere seriously pisses me off. Now we'll have flock of people asking "is it true that gst doesn't fit on single CD?"
[19:12] <mathrick> ugh, http://yro.slashdot.org/article.pl?sid=04/03/21/0052220&mode=nested&tid=123&tid=98&tid=99
[19:13] Action: mathrick speechless
[19:17] benow ([email protected]) joined #gstreamer.
[19:17] pel ([email protected]) joined #gstreamer.
[19:17] <benow> holas.
[19:17] <pel> hola
[19:17] <benow> so, who do I talk to to check in the working gst-java? ;)
[19:18] <pel> I got things working yesterday, thanks all for  the help =)
[19:19] <pel> however, to celebrate - I borked it today =-/
[19:20] <pel> gstbasicscheduler won't compile
[19:20] <pel> benow, congratulations btw =)
[19:23] <Uraeus> benow: attach a patch to bugzilla and someone (maybe even me) will get to it very soon
[19:23] <Uraeus> benow: or you can mail thomasvs asking him to give you CVS access
[19:25] <benow> nice, I'll email him, as there'll be many changes.
[19:25] <benow> pel: thx, what did you do to bork basicscheduler? 
[19:26] <pel> basicly I recompiled it, with an apt-get upgrade in between.
[19:26] <pel> nothgin fancy
[19:27] <pel> it's strange
[19:27] <pel> i get a few warnings (use of cast expressions as lvalues is deprecated) an then 
[19:27] <benow> perhaps a lib that was upgraded... 
[19:27] <pel> make: *** [libgstbasicomegascheduler_la-gstbasicscheduler.lo] Error 1
[19:28] <pel> i'd say that's the problem...
[19:28] <benow> any other errors before the 'Error 1'?
[19:28] <pel> the cast thingie isn't a problem unless -pedantic is used , and it is not used.
[19:28] <pel> nope
[19:29] <pel> gstbasicscheduler.c: In function `gst_basic_scheduler_chainhandler_proxy':
[19:29] <pel> gstbasicscheduler.c:493: warning: use of cast expressions as lvalues is deprecat ed
[19:29] <pel> gstbasicscheduler.c: In function `gst_basic_scheduler_select_proxy':
[19:29] <pel> gstbasicscheduler.c:514: warning: use of cast expressions as lvalues is deprecat ed
[19:29] <pel> gstbasicscheduler.c: In function `gst_basic_scheduler_gethandler_proxy':
[19:29] <pel> gstbasicscheduler.c:562: warning: use of cast expressions as lvalues is deprecat ed
[19:29] <pel> gstbasicscheduler.c: In function `gst_basic_scheduler_eventhandler_proxy':
[19:29] <pel> gstbasicscheduler.c:600: warning: use of cast expressions as lvalues is deprecat ed
[19:29] <pel> gstbasicscheduler.c: In function `gst_basic_scheduler_cothreaded_chain':
[19:29] <pel> gstbasicscheduler.c:748: warning: use of cast expressions as lvalues is deprecat ed
[19:29] <pel> gstbasicscheduler.c: In function `gst_basic_scheduler_chain_remove_element':
[19:29] <pel> gstbasicscheduler.c:905: warning: use of cast expressions as lvalues is deprecat ed
[19:29] <pel> gstbasicscheduler.c: In function `gst_basic_scheduler_reset':
[19:29] <pel> gstbasicscheduler.c:1086: warning: use of cast expressions as lvalues is depreca ted
[19:29] <pel> make: *** [libgstbasicomegascheduler_la-gstbasicscheduler.lo] Error 1
[19:30] <benow> narsty, you're compiling the 0.8 src bundle?  perhaps a 'make clean'...
[19:30] <pel> it is a fresh checkout...
[19:39] <Uraeus> ain: ok, got marlin up on my screen :)
[19:40] <Uraeus> ain: some ugly messages in the console though:
[19:40] <Uraeus> [cschalle@localhost i386]$ marlin
[19:40] <Uraeus> Opening /tmp/untitled.v3CWRF
[19:40] <Uraeus> Opening /tmp/untitled.iZxWCb
[19:40] <Uraeus>  
[19:40] <Uraeus> (marlin:2193): GLib-GObject-WARNING **: value "0" of type `guint' is invalid  or out of range for property `channels' of type `guint'
[19:40] <Uraeus> Connecting pipeline
[19:40] <Uraeus>  
[19:40] <Uraeus> (marlin:2193): libmarlin-WARNING **: Caps nego fail 1
[19:40] <Uraeus>  
[19:40] Last message repeated 1 time(s).
[19:40] <Uraeus> (marlin:2193): Gtk-WARNING **: FileMenu: missing action
[19:40] <Uraeus> Channels changing to 2
[19:41] <markey> hmm I've set PKG_CONFIG_PATH=/usr/lib/pkgconfig, but gstreamer still installs stuff in /usr/local/lib/pkgconfig
[19:41] <markey> that normal?
[19:41] <Uraeus> markey: I think Gstreamer follows --prefix for where it installs pkgconfig stuff
[19:41] <markey> ah ok
[19:42] <markey> thx
[19:42] Action: markey rebuild and reinstalled it all, fingers crossed
[19:44] Action: mathrick hugs UTF-8
[19:45] <benow> pel: perhaps try the 0.8 tar.bz, it compiled fine for me, while cvs had issues.
[19:47] <pel> benow, thanks - but I'd rather try to figure this out on my spare cycles. 
[19:47] <pel> benow, who knows - I might learn something ;-)
[19:48] <ain> Uraeus: yeah, I get that too, I think I might know what the problem is...
[19:48] <ain> it means it won't be able to play anything, but it should still load stuff
[19:48] <ain> recording and saving will be broken too
[19:48] <ain> still got to update them for the new plugins
[19:48] <markey> huzzah, it works again
[19:48] Action: markey jumping around
[19:49] <Uraeus> markey: good to hear :)
[19:49] ploum ([email protected]) left irc: "Leaving"
[19:49] <Uraeus> ain: ok, I commited a spec file to marlin CVS as we talked about long ago, people can now do rpmbuild -ta on the marlin tarball :)
[19:50] <Uraeus> ain: ok if I create a desktop file and get someone to make a icon for Marlin?
[19:52] <Uraeus> mathrick: have you seen the helix player?
[19:52] <mathrick> Uraeus: nope
[19:52] <mathrick> Uraeus: anything special about it?
[19:53] <Uraeus> mathrick: well I have been using it to view some of the video interviews on news.com; it is actually rather nice. If you want something to testbed your mms stuff eventually I think looking at the helix player could give some good info on how to create a great gstreamer mozilla plugin player
[19:54] <mathrick> Uraeus: k, will keep that written down somewhere :)
[19:56] <Uraeus> mathrick: since it is done in gtk2 it can probably function very well as docs on how to make a well working gtk/mozilla plugin player
[19:57] <mathrick> Uraeus: news.com uses mms://? care to throw me URL for their streams?
[19:57] ploum ([email protected]) joined #gstreamer.
[19:58] kmaraas ([email protected]) left irc: Client Quit
[20:00] <Uraeus> mathrick: they offer both real and wmv streams, I assume the wmv streams use mms
[20:00] <Uraeus> mathrick: http://www.linuxrising.com/files/helixplayer.png
[20:01] <mathrick> Uraeus: k, and where can I get actual mms:// content? :)
[20:01] <Uraeus> gimme a sec
[20:03] <Uraeus> mathrick: think you have to dig out the URL yourself, I have preset it to use real/helix for now and trying to change it crashes mozilla :)
[20:04] <Uraeus> mathrick: news.com <- right sidebar (for instance Novell interview)
[20:04] <mathrick> yep, digging it right now
[20:06] <markey> oh, seeking finally working in 0.8
[20:06] <markey> that's good to see
[20:06] <Uraeus> :)
[20:07] <mathrick> hrmpf, looks like it's buried somewhere in JavaScript :\
[20:08] <Uraeus> mathrick: well implement the plugin player and make it work with Company's xine plugin to begin with. That way you have a testbed that is verified to work for your libmms library
[20:09] <pel> mathrick, mms://sr-wm.qbrick.com/02038_srk-wm-High
[20:09] <pel> mathrick, swedish radiochannel with classical music.
[20:09] <Uraeus> mathrick: I guess some kind of preference to wether the plugin should pretend to be real or M$ is also in order :)
[20:12] <mathrick> pel: thx
[20:13] <mathrick> Uraeus: hmm, I don't think it's really needed
[20:13] <mathrick> Uraeus: for mms:// we always feign M$, and Real is casted over rtsp:// only I think, so it's non-issue
[20:15] <Uraeus> mathrick: yes, but many webpages look for microsoft media player plugin or real media player plugin, so we need our plugin to pretend to be those in order to be detected
[20:15] <mathrick> Uraeus: ahhm that thing
[20:16] <Uraeus> mathrick: I mean I want to be able to go to news.com and watch those interviews using GStreamer
[20:16] <mathrick> Uraeus: yep, we should probably do something like that. But I have no clue how one does things like that ;)
[20:17] <Uraeus> mathrick: that is why I am suggesting to look at the code of the helix player :)
[20:17] <mathrick> Uraeus: k, I get you now :)
[20:18] sublett ([email protected]) joined #gstreamer.
[20:28] jcsston ([email protected]) joined #gstreamer.
[20:38] <markey> yay, I think I'm getting our vis to work vis GST
[20:38] <markey> identity seems to work now
[20:38] NFusi0n ([email protected]) joined #gstreamer.
[20:38] <markey> make that: "vis to work with"
[20:39] <Uraeus> yeah; those NMM guys don't stand a chance <g>
[20:39] <markey> hehe
[20:40] <markey> mxcl is quite eager to get NMM working nicely
[20:40] <markey> competition can only be good *g*
[20:40] kmaraas ([email protected]) joined #gstreamer.
[20:40] steve_b ([email protected]) joined #gstreamer.
[20:41] <Uraeus> markey: yeah, it is very good as long as we always 'win' :)
[20:48] <pel> I've found myself a problem.
[20:49] <pel> while compiling gstbasicscheduler i get a couple of warnings like warning: use of cast expressions as lvalues is deprecated
[20:49] <pel> which is no biggie, but the Makefile want's to use -Werror
[20:50] <pel> how do I fix the warning?
[20:50] <pel> and will that break any other compilers (I'm using gcc 3.3.3)
[20:52] <teuf> pel: you need to change (sometype)value = something to value = (someothertype)something
[20:52] <teuf> can't be more specific without looking at the code ;)
[20:53] Uraeus ([email protected]) left irc: "Client exiting"
[20:54] <pel> teuf, ah. that is good enough to me
[20:54] <pel> I got a bit confused what to do, but moving it to the right side is.. obvious =)
[20:55] <mathrick> ahh, looks like it's that nice time of year again :)
[20:56] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[20:57] dolphy ([email protected]) joined #gstreamer.
[21:00] Shoragan ([email protected]) joined #gstreamer.
[21:00] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[21:04] <pel> ick. not to simple inside macro-madness :-(
[21:04] Zeenix ([email protected]) joined #gstreamer.
[21:04] <Zeenix> hi
[21:04] <Company> pel: theres a patch in bugzilla that fixes that
[21:04] <mathrick> yo Zeenix
[21:05] <pel> company, thanks. 
[21:05] Action: pel feels a bit better
[21:07] <Company> we need to fix that correctly soon
[21:07] <Company> but noone of the devs uses a gcc version that triggers that
[21:07] <pel> ah.
[21:07] <pel> what is wrong with the patch?
[21:07] <teuf> Company: I think a gcc version triggering that landed in debian
[21:08] <Company> i don't like adding a macro for setting something
[21:08] <pel> teuf, sure did =)
[21:08] <Company> it looks a lot better to me to just use the assignment directly
[21:08] Action: Company goes downloading gcc 3.3.3
[21:09] <pel> company, true, but it seems to expand into something very hairy... :-/
[21:11] <Company> i'm gonna compile me a gcc 3.3.3 and look at it
[21:14] Shoragan ([email protected]) left irc: "Leaving"
[21:42] <Company> booooooring
[21:44] Action: Company looks at gcc compiling again...
[21:50] lupusBE ([email protected]) joined #gstreamer.
[22:26] asdlfjkklsj ([email protected]) joined #gstreamer.
[22:26] <asdlfjkklsj> anyone know of RPMs/SRPMs of gstreamer-player for Fedora?
[22:27] kwm ([email protected]) left irc: "Leaving"
[22:33] pel ([email protected]) left irc: "ChatZilla 0.9.52B [Mozilla rv:1.6/1]"
[22:44] <taaz> asdlfjkklsj: gstreamer.net -> download -> fedora
[22:45] Nick change: mathrick -> mathrick|bath
[22:48] teuf ([email protected]) left irc: Remote closed the connection
[22:51] <asdlfjkklsj> oh christ, how did I miss that
[22:51] <asdlfjkklsj> <--- asshat ;)
[22:55] <asdlfjkklsj> ** (gst-player:29632): WARNING **: libgstplay: failed initializing pipeline, error: Could not link video output thread (cs and balance)
[22:55] <asdlfjkklsj> doh
[22:56] <taaz> maybe you're missing videobalance and colorspace plugins
[22:59] <dolphy> asdlfjkklsj: you need ffmpeg
[22:59] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[23:00] <taaz> the player requires ffmpeg?
[23:00] <Company> the player requires ffcolorspace
[23:01] <Company> that has been but into gst-plugins though
[23:01] <taaz> that's in gst-plugins though
[23:03] <taaz> i need advice with these debs.  they are sort of a mess
[23:04] <taaz> added new speex and caca packages and threw everything else into -misc.  which makes -misc rather large and nearly a requirement
[23:04] jcsston ([email protected]) joined #gstreamer.
[23:04] <bitshifter> taaz: btw, just out of curiosity: why don't you put/maintain your debian/ dir into/in gst CVS?
[23:05] <taaz> and for lack of a better idea, i put the 3 colorspace plugins into -colorspace package
[23:05] <taaz> bitshifter: because that's too hard
[23:05] jdahlin ([email protected]) joined #gstreamer.
[23:05] <taaz> well... not hard... just annoying
[23:05] <taaz> or nearly impossible depending on how you look at it
[23:05] <taaz> do you want the long answer too? ;)
[23:06] <bitshifter> if you can be bothered, sure ;)
[23:07] <Company> taaz: you should certainly put every plugin from the gst/ dir into the same package
[23:07] <taaz> main reason is that debian packages are, in general, based off of upstream released code.  something stable such as a tarball or something named
[23:08] <Misirlou> IMHO, it's better if the only thing in the .diff.gz is the debian directory.
[23:08] <taaz> and if you actually try this sort of thing you'll find out that upstream cvs really never is quite what is released in the tarballs.  random junk slips in, generated files are missing or over-included or other differences between cvs dev code and tarballs
[23:08] <bitshifter> taaz: ah, I see. Thanks for the explanation
[23:08] <taaz> Company: from the gst-plugins/gst dir?  they mostly are.  its the ext/ and sys/ dirs
[23:09] <taaz> bitshifter: it makes cvs sort of a moving target vs just packaging what upstream has declared "stable" enough for a release
[23:09] <Company> well, those should be packaged based on dependencies
[23:09] <taaz> you can of course try to manage this with branches and other cvs evilness but that, imho, is rather messy
[23:10] <Company> that has never been a problem, has it?
[23:10] <bitshifter> taaz: yeah, I just assumed that release stuffed gets tagged accordingly in CVS, so that checking out a release-tagged version + 'make dist' gets you more or less the same as the release tarball (with different autofoo stuff of course)
[23:12] <taaz> bitshifter: more or less, it does.  but that means you need to branch from that point and do debian work.  i dunno.  maybe it's doable.  i'm very bad at groking cvs branching though.
[23:13] <taaz> bitshifter: i'm using arch now (aka tla).  i'll put it up on arch.debian.org soon.
[23:14] <taaz> Company: there are -alot- of dependencies.  a package for every (major) one would be annoying to developers, users, admins, etc etc.  yet i'm getting closer to that every day.
[23:15] jdahlin ([email protected]) left irc: "Leaving"
[23:15] <Company> I would put everything in gst/ into one package
[23:15] <taaz> Company: and what do i do with things like etc/pango/?  just because those -use- pango doesn't mean naming a package for text/timeoverlay gstreamer-pango makes any sense at all
[23:15] <Company> if that requires Hermes or whatever, so be it
[23:16] <Company> and i'd make every dir in sys/ and ext/ its own plugin
[23:16] <Company> just because that's how we do it upstream (or plan to do it)
[23:16] <taaz> that's over 50 you realize?
[23:17] <Company> no i don't, i just apt-get install gstreamer-plugins and after that remove arts ;)
[23:17] <Company> or build from source
[23:17] <Company> but if you don't wanna go that way you'd only get problems about where to draw the line
[23:17] <taaz> oh crapola... now i get lots of console spam on gst-register.  argh.
[23:18] <Company> it's gstreamer-0.8 :p
[23:19] <Company> gst-register-0.8 even
[23:19] <taaz> i actually consider it a bug that gst must have -0.8 on everything
[23:20] <Company> only on stuff where it might collide with 0.6 or 0.10
[23:20] ChrisHJW ([email protected]) left irc: Read error: 60 (Operation timed out)
[23:20] <taaz> i think we really should be using the autoconf stuff to do the naming mangling and default to -0.8 but let the user decide
[23:20] <Company> and if you know a better fix i'd happily take that fix
[23:21] sublett ([email protected]) left irc: "Why are all the really good songs short?"
[23:23] lupusBE ([email protected]) left irc: "Bezig met verlaten"
[23:29] sublett ([email protected]) joined #gstreamer.
[23:41] <markey> yay, gst engine now running with full analyzer
[23:41] <markey> works great
[23:42] <markey> slowly getting on par with arts engine
[23:43] <Company> go markey!
[23:46] Nick change: mathrick|bath -> mathrick
[23:50] asdlfjkklsj ([email protected]) left #gstreamer (".").
[23:56] Misirlou ([email protected]) left irc: "asdf"
[00:00] --- Mon Mar 22 2004
[00:04] <taaz> markey: what are you working on?
[00:04] <markey> amaroK
[00:04] <taaz> which is what?
[00:04] <markey> oh, an audio player for KDE
[00:04] <markey> XMMS like
[00:05] <markey> http://amarok.sf.net
[00:05] <taaz> i need to start up a audio player too.  i feel left out of the party ;)
[00:05] <markey> hehe
[00:05] <markey> you're invited to hack on our GST engine *g*
[00:05] <thaytan> taaz: I've got one you can borrow
[00:06] <thaytan> written in sh using gst-launch ;)
[00:06] <taaz> my ideas for audio players seem to differ from just about everyone else though
[00:08] <markey> tell!
[00:08] <taaz> thaytan: sounds like something you should be doing in python ;)
[00:08] <thaytan> markey: are you porting the visualisations to gstreamer elements, or getting hold of the audio buffers somehow?
[00:09] <markey> no, we just grab the audio sample data from gst
[00:09] <markey> via identity
[00:09] <markey> and do the processing & rendering ourselves
[00:09] <markey> FFT routine
[00:09] <markey> advantage: unified system for all frameworks
[00:09] <taaz> markey: do i have to give up my sekrets? ;)
[00:09] <markey> ARTS, NMM, GST
[00:09] <markey> hehe
[00:10] <taaz> sigh...
[00:10] <taaz> arts, nmm, gst for what?
[00:10] <markey> well, for fun, choice, and testing
[00:10] <markey> helping KDE decide 
[00:11] <taaz> right.  of course.  all projects need their own portability layer between all possible choices.
[00:11] <markey> well, there's no disadvantage
[00:11] <taaz> excuse me if i break into a rant ;)
[00:11] <markey> no performance hit or something
[00:11] <markey> np
[00:12] <taaz> disadvantage is more code
[00:12] <markey> oh it even helped getting our code cleaner
[00:12] <markey> good separation of interfaces
[00:12] <taaz> and all the systems you port to probably don't have the same capabilities.  so you end up with a layer that is the lowest common denominator
[00:12] <markey> so far it works nicely
[00:13] <markey> it's a challenge
[00:13] <taaz> i've never gotten good at this sort of thing.  i understand the reason for abstract interfaces but i've never done a good job at it.  always seems to have some sort of negative effect on my code as far as utilizing whatever is behind the interface
[00:14] <markey> basically, we didn't have much choice
[00:14] <markey> we started as a pure ARTS app
[00:14] <markey> but ARTS is dying..
[00:14] <markey> so we had to do something about it
[00:15] <taaz> which is of course a reason to have layered interfaces.  so when one dependency dies you can add another fairly easily.
[00:16] <markey> yes
[00:16] <taaz> it really seems like all these players are just front ends for the same backend concepts
[00:17] muesli ([email protected]) joined #gstreamer.
[00:17] Larson ([email protected]) joined #gstreamer.
[00:17] <markey> lol
[00:17] <markey> hi muesli, Larson 
[00:17] <Larson> hey markey  :)
[00:17] <muesli> hehe, hi
[00:17] <markey> fellows
[00:17] <mathrick> hehe: "In the beginning was the word, and the word was content-type: text/plain"
[00:17] <taaz> you know what player i use more than any other?  cplay.  curses.  clean.  easy to use.
[00:18] <taaz> tragically, it doesn't seem to do tag stuff properly, but i've lived with that for a while
[00:18] <taaz> once the python gst stuff is working i'll probably do the hack to get cplay to use gst
[00:19] <taaz> here's an interesting user interface research topic:  how to write a seek widget for large amounts of data
[00:20] <taaz> i have yet to find a player that can handle 7 hour audio recordings and let you seek at a finer resolution than pixels used for a slider
[00:20] <taaz> which might end up being 20 minutes in the case of rhythmbox's tiny slider
[00:21] <thaytan> 2 sliders for coarse/fine control, or a logarithmic control perhaps?
[00:21] <taaz> i've been thinking of some sort of dynamic zooming log scale widget... might work
[00:22] <taaz> thaytan: i did a back-of-a-napkin drawing trying to figure dual sliders out.  it didn't look usable. ;)
[00:22] <taaz> might work well for speed control though
[00:23] <taaz> ||||-|-|-|--|---|----|---*---|----|---|--|-|-|-||||
[00:23] <taaz> how about that?  as you slide it mushes ticks together like:
[00:24] <taaz> |||||||||-|-|-|--|---|----|---*---|--|-|-||||
[00:24] <taaz> eh, not quite to scale ;)  but you get the idea
[00:24] <taaz> not sure if it would work in practice though
[00:25] <taaz> seems these sorts of thigns require repositioning the user pointer (mouse) but that's not good ui design in most cases
[00:25] <taaz> since it doesn't scale to touchscreen devices
[00:26] <taaz> hmm... maybe that doesnt work at all... i dunno
[00:27] <Company> taaz: the problem with those abstractions is that you can abstract A/V really well
[00:27] <Company> taaz: at least as long as you only do playback and seeking, which is what all those players do
[00:28] <Company> taaz: it only gets harder when you do more interesting stuff
[00:28] <Company> amarok could probably do a helix backend, too
[00:32] <taaz> is that important? ;)
[00:32] <Company> apart from the last line: no
[00:33] <Company> s/no/yes/
[00:33] <Company> whoops
[00:33] <Company> it only gets interesting when you do more advanced stuff
[00:34] <Company> like tag writing or syncing ipods with auto-transform of music
[00:34] <taaz> i must be honest... i don't truest gst with my tags ;)
[00:35] <taaz> where is the tag writing test suite?
[00:35] <Company> i wouldn't either
[00:35] <sxpert> they work for me in sound-juicer :D
[00:35] <Company> but that's just because noone finished and tested it yet
[00:35] <taaz> tag handling code is almost an ideal candidate for a whole lot of unit tests
[00:37] <Company> we'll put your tests into the media testsuite
[00:38] <taaz> ok, item #4053 on my todo list
[00:38] ploum ([email protected]) left irc: "Leaving"
[00:40] <taaz> a good complement to tag editing would be a tag override system.
[00:41] <taaz> so if you don't trust the tag write code you could at least store what you would want those tags to be
[00:41] <taaz> which also is needed in the case of overrides for read-only media
[00:42] sublett ([email protected]) left irc: "Why are all the really good songs short?"
[00:42] <taaz> and could easily be exported for a tag writing program you trust to actually do the updates
[00:43] <taaz> or sent to whoever might own, say, web site with read-only streams you were using
[00:43] <taaz> or used to update a dvd archive of oggs
[00:43] <taaz> does any player do that now?
[00:44] <Company> you mean like storing the tags you want in some database?
[00:44] <mathrick> nite
[00:45] Nick change: mathrick -> mathrick|sleep
[00:45] <taaz> Company: yes.  instead of actually writing them to the source file/resource, just store the associate from resource to user tag
[00:45] <Company> that's kinda stupid
[00:46] <Company> but it's what medusa is all about
[00:46] <taaz> stupid?  i have a need for this right now
[00:46] <Company> and all those other "attach metadata to files" stuff
[00:46] <taaz> medusa?
[00:46] <Company> i think that's the name, yeah
[00:51] <taaz> looks like overkill for what i'm talking about
[00:54] Action: bitshifter likes the idea
[00:55] <taaz> and really, how is that idea "stupid"?  case study: networked multi-user system.  you load someone elses music into rhythmbox.  tags are quite broken with spelling mistakes, numbers in the track titles, missing tags, etc.  you have no write access to this media.  rhythmbox can't possibly sort it properly.  other than copy & modify how can you deal with this?
[00:56] <Company> tell the user to fix his tags
[00:56] <Company> that's just like wanting to group similarly spelled words in the artist browser in rhythmbox
[00:57] <Company> those tags are broken, you shouldn't support them staying that way
[00:57] <taaz> that sounds like a good idea
[00:58] <Company> at least it's something i won't hack on ;)
[00:58] <taaz> i also have lots of music with similar but slightly different names
[00:58] <taaz> would be nice to sort those the same
[00:58] <Company> yeah, but the correct fix is to correct the tags
[00:58] <taaz> ie, Foo J. Bar sorted along with Foo Bar  and Foo John Bar
[00:59] <taaz> some people might not give a damn and refuse to fix anything.
[00:59] <Company> yeah
[00:59] <Company> and those should have to get the ugly lists
[00:59] <ain> Company: your audioconvert fixes are only in CVS, they're not going to be in 0.8.0?
[00:59] <Company> you shouldn't encoureage laziness, stupidity and so on
[01:01] <Company> ain: 0.8.0 is out already, i don't have a time machine
[01:01] <Company> ain: 0.8.1 will have them i think
[01:01] <taaz> ok, clearly we have different opinions on this ;)
[01:01] <ain> Company: oh, I thought only a 0.8 pre-release was out. my bad
[01:01] <taaz> i'm sure we all have a long ugly list of things that are not done our way and we still have to deal with it.  i'm just trying to come up with solutions for the deal with it part.
[01:01] <ain> Company: I wonder what other bugs will mean Marlin will be permanently tracking CVS head :)
[01:02] <Company> ain: i think you can use tarballs/packages when 1.0 is out ;)
[01:02] <Company> ain: for now you're one of the testers we rely on :)
[01:03] <ain> heh :)
[01:03] <ain> yey
[01:05] <Company> taaz: there's some things I don't fix - one of them is stupidity, another is laziness - but you can of course add elements that fix that ;)
[01:08] <taaz> so you are in agreement that we should remove all support for that ugly #ifdef __cpluplus  extern "C" construct?
[01:08] <taaz> there is no excuse for that being in C code.  it's a hack so C++ code that uses the C code will work.  why is that hack not in the C++ files that include the C headers?  lazy and stupid C++ programmers?
[01:10] Action: taaz is always bitching about that nonsense ;)
[01:11] <Company> yeah, that should go
[01:12] <Company> the C++ code should do that
[01:12] <Company> ideally
[01:12] <Company> unfrotunately it's a standard to put it in C headers, and standards are important enough to follow them
[01:14] markey ([email protected]) left irc: "g'nite"
[01:16] muesli ([email protected]) left irc: Remote closed the connection
[01:17] spyder482 ([email protected]) left irc: "Leaving"
[01:20] bilboed_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[01:21] <taaz> moo
[01:22] ain ([email protected]) left irc: "scary job interview"
[01:22] bilboed_ ([email protected]) joined #gstreamer.
[01:23] <Company> bah
[01:23] <Company> standard gcc 3.3 doesn't trigger that warning
[01:23] <taaz> the "moo" warning?
[01:24] <Company> no the lvalue cast warning
[01:24] <Company> all those projects adding random warnings should just be shot IMNSHO
[01:32] sxpert ([email protected]) left irc: Remote closed the connection
[01:37] sxpert ([email protected]) joined #gstreamer.
[01:47] jcsston` ([email protected]) joined #gstreamer.
[01:48] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[01:54] <Company> ha, debian didn't add any more warning flags
[01:54] <Company> go debian!
[02:02] <taaz> yeah!  huh?
[02:03] <Company> the lvalue-cast stuff
[02:03] <Company> i thought i update my debian box to get those issues now that fedora had that
[02:03] sub_pop ([email protected]) left irc: "Client exiting"
[02:03] <Company> but turns out only fedora does weird stuff with gcc ;)
[02:08] <Zeenix> Company: what weird stuff, dont scare me more from fedora :)
[02:10] <Company> Zeenix: warnings for lvalue casts
[02:12] sub_pop ([email protected]) joined #gstreamer.
[02:24] Rotty ([email protected]) left irc: Remote closed the connection
[02:26] Action: sxpert remembers that exact same conversation a couple of weeks ago
[02:36] <Zeenix> sxpert: have you heard of 'false memory' or 'dilusions'? ;)
[02:37] <sxpert> Zeenix: possibly...
[02:37] <sxpert> lol
[02:49] judith ([email protected]) joined #gstreamer.
[02:50] <Company> damn
[02:50] Larson ([email protected]) left #gstreamer.
[02:50] <Company> my X vidmode extension is borked
[02:50] judith ([email protected]) left irc: Client Quit
[02:51] BBB ([email protected]) joined #gstreamer.


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click