IRC Logs

*******************************************************************
[03:00] <thaytoo> um, actually
[03:00] <BOFH> am thinking that'd be a rhythymbox prob, not gst..
[03:01] <thaytoo> you need the id3tag element
[03:01] <thaytoo> which is in mad, yeah
[03:02] <BOFH> which is...   where?
[03:02] <thaytoo> in debian, in the gstreamer-mad package
[03:02] <BOFH> meh..  distro's ;)  
[03:02] <BOFH> don't run onee.
[03:03] <thaytoo> then you need to have libmad installed when you build gst-plugins
[03:03] <BOFH> ahh..  oki then..  that makes more sense..
[03:05] <BOFH> t/y - will install and test.
[03:07] mxpxpod ([email protected]) left irc: "To quit or not to quit.. wait... was that a question?"
[03:09] <Company> you need libid3tag, too
[03:12] <BOFH> where's that one found?
[03:12] <Company> grah
[03:13] <Company> like libmad
[03:13] <BOFH> libmad I found on freshmeat...
[03:13] <Company> they were the same packet once...
[03:13] <BOFH> libid3tag though...  no luck.
[03:13] <Company> libid3tag should be on the same ftp
[03:13] <BOFH> yah just noticed.
[03:13] <BOFH> MyBad(tm)
[03:14] <Company> what idiot invented ASF?
[03:14] <BOFH> Some f00 on the Microshaft payroll
[03:15] <BOFH> there..  id3tag installed.
[03:15] <BOFH> t/y
[03:15] <Company> there's bunches of asf parsers, but i still don't understand it :/
[03:15] <bitshifter> is ASF still used anywhere these days?
[03:15] <Company> np
[03:16] <Company> it's the default MS format
[03:16] <Company> asf = wm[av]
[03:16] <BOFH> isn;t asf kind of like and XML for MS[A/V] files or something?
[03:16] <bitshifter> ah, I thought wma/wmv were newly-designed from scratch
[03:16] <BOFH> s/and/an
[03:17] <Company> nope, asf is nothing like xml
[03:17] <Company> they went the extra way to make it as unreadable as possible
[03:18] <Company> like using GUIDs instead of readable names
[03:18] <Company> yay!
[03:49] ct__ ([email protected]) joined #gstreamer.
[03:57] chrisa ([email protected]) joined #gstreamer.
[03:58] ct_ ([email protected]) left irc: Read error: 60 (Operation timed out)
[03:58] Marsupilami23 ([email protected]) left irc: Remote closed the connection
[04:00] walters ([email protected]) joined #gstreamer.
[04:05] Marsupilami23 ([email protected]) joined #gstreamer.
[04:06] ct__ ([email protected]) left #gstreamer ("Client exiting").
[04:12] Marsupilami23 ([email protected]) left irc: Remote closed the connection
[04:16] <thaytoo> Company: I suspect you're impugning GUIDs
[04:17] <thaytoo> you want to be careful with that sort of behaviour, or MS's black suit men come a-chasing
[04:17] chrisa ([email protected]) left #gstreamer ("Client exiting").
[04:17] <Company> GUIDs are one of the worse inventions by MS
[04:17] <Company> you can quote me on that
[04:26] <ds-work> http://en.wikipedia.org/wiki/Universally_Unique_Identifier
[04:26] <ds-work> not exactly a MS invention
[04:27] <ds-work> yay.  I have oss rate probing working
[04:28] <ds-work> it sucessfully detects the list of 8 rates one of my sound cards supports in 16 probes
[04:33] <BOFH> well, now I get more annoying errors..  can;t create player.. did I run gst-register..
[04:33] <BOFH> <sigh>
[04:36] <Company> it's MS' invention that they used such a crap
[04:36] <Company> BOFH: use a distro ;)
[04:36] <BOFH> bah - that goes against everything I have built ;)
[04:37] Marsupilami23 ([email protected]) joined #gstreamer.
[04:38] <Company> you're using LFS?
[04:39] <BOFH> There we go..  working now.
[04:39] <BOFH> reloaded Rhythymbox...  (uninstall, update, make distclean, config, build. install)
[04:39] <BOFH> yah some demented spin off of LFS...
[04:40] <BOFH> looked at LFS and decided 80% of the apps were too old...
[04:40] <BOFH> don;t ask, I;m one of the really WIERD ones...
[04:40] <Company> oh
[04:40] <Company> the machine i'm sitting on is LFS
[04:41] <Company> at least the base system - i built the rest myself
[04:41] <BOFH> (running a CVS pull of X, Gnome, Apache, and php, as well as MySQL 5, etc. etc.)
[04:41] <Company> but after doing that once i don't need it again
[04:41] <BOFH> well, I wanted the 2.6 kernel...
[04:41] <BOFH> and I wanted to give NPTL a shot..
[04:41] Marsupilami23 ([email protected]) left irc: Client Quit
[04:41] <Company> your fault :p
[04:42] <BOFH> and since there was no distro that made it easy, and I had loads of time to spare...
[04:42] trow ([email protected]) left irc: "Leaving"
[04:42] <Company> well, that means you'll figure most problems out yourself atleast :)
[04:43] <Company> so i can go to bed now, gnight
[04:43] <BOFH> night.
[04:45] <thaytoo> *coughsandmumblessomethingabouthisdebianunstablesystemwith2.6.3kernelpackages*
[04:47] <BOFH> ahh, but is it pure 2.6?
[04:47] <bitshifter> even 2.6.4 ones for some archs
[04:48] <thaytoo> BOFH: my laptop is
[04:48] <thaytoo> 2.6.3 kernel tree used to build my own debian packages
[04:48] <BOFH> 2.6 kernel headers in /usr, glibc built against 2.6 headers with an up to date gcc etc. etc.
[04:48] <thaytoo> 2.6.4, rather
[04:48] <thaytoo> my workstation at home is running 2.6.3 from the debian archive
[04:48] <ds-work> sid has had 2.6 headers in /usr/include since 2 October
[04:49] <BOFH> Some reason I never liked debian...  I installed it years back and it just left a bad taste...
[04:49] <thaytoo> 2.6.0test7 headers
[04:49] <BOFH> same as slack..
[04:49] Action: bitshifter is bored and ponders fetching glib-2.4/gtk-2.4 from incoming.debian.org
[04:49] <BOFH> recently only ran Mandrake and RedHat..  didn;t like em, but I referred e over Deb...
[04:50] <BOFH> s/referred/preferred/
[04:50] mxpxpod ([email protected]) joined #gstreamer.
[04:51] Action: BOFH running 2.4.0 ATM
[04:51] <BOFH> (glib/gtk that is)
[04:51] <BOFH> anyhow..   gone to bed..  is late, am tired.
[04:54] <ds-work> thaytoo: do you ever have issues with OSS sound card rates?
[04:58] joem_ ([email protected]) joined #gstreamer.
[04:58] joem_ ([email protected]) left irc: Client Quit
[04:59] <thaytoo> ds-work: nope, but I try not to use OSS
[04:59] <thaytoo> I have a crappy intel8x0 soundcard
[05:00] <ds-work> do you use alsa?
[05:00] <thaytoo> yeah
[05:00] <thaytoo> or esdsink, depending on the time of day and the app
[05:01] <thaytoo> alsa needs period-size=[2048,4096] on this sound card
[05:01] <thaytoo> I don't know how to get it to figure that out on its own and JFDI
[05:01] <ds-work> does it work with osssink, too?
[05:01] <thaytoo> yeah, it does
[05:02] <ds-work> could you update gst-plugins/sys/oss/, recompile that dir, and run ./oss_probe and email me the output?
[05:02] <thaytoo> not until I can get my laptop on the net again tonight
[05:02] <thaytoo> 3 hours
[05:02] <ds-work> ok, nevermind then
[05:02] <thaytoo> oh, I can do it at home
[05:03] <thaytoo> on the workstation with an Audigy
[05:03] <ds-work> I'll be home in 30 minutes, in which case I can check myself
[05:04] <ds-work> with alsa, that is
[05:08] <thaytoo> you'll probably beat me
[05:09] <thaytoo> I have to rebuild this source tree from scratch for 0.8
[05:09] <thaytoo> hadn't updated it yet
[05:28] <thaytoo> ds-work: sent
[05:29] <ds-work> hrm, odd
[05:30] <ds-work> does alsa have automatic rate changing?
[05:30] <ds-work> it claims to support 100kHz
[05:40] walters ([email protected]) left irc: Client Quit
[05:42] jcsston ([email protected]) joined #gstreamer.
[05:45] <thaytoo> ds-work: pretty sure the Audigy does 96KHz
[05:45] <thaytoo> but I don't know why it would say 100
[05:47] <thaytoo> "24-bit/96KHz 5.1 surround or 24-bit/192KHz stereo"
[06:16] sxpert ([email protected]) got netsplit.
[06:17] taaz ([email protected]) got netsplit.
[06:17] steve_b ([email protected]) got netsplit.
[06:17] apoc ([email protected]) got netsplit.
[06:17] dilinger ([email protected]) got netsplit.
[06:17] jimmy_dean ([email protected]) got netsplit.
[06:17] md` ([email protected]) got netsplit.
[06:17] Misirlou ([email protected]) got netsplit.
[06:17] bilboed ([email protected]) got netsplit.
[06:17] shawarma ([email protected]) got netsplit.
[06:17] joem ([email protected]) got netsplit.
[06:19] sxpert ([email protected]) returned to #gstreamer.
[06:19] Misirlou ([email protected]) returned to #gstreamer.
[06:19] bilboed ([email protected]) returned to #gstreamer.
[06:19] taaz ([email protected]) returned to #gstreamer.
[06:19] dilinger ([email protected]) returned to #gstreamer.
[06:19] jimmy_dean ([email protected]) returned to #gstreamer.
[06:19] steve_b ([email protected]) returned to #gstreamer.
[06:19] apoc ([email protected]) returned to #gstreamer.
[06:19] md` ([email protected]) returned to #gstreamer.
[06:19] joem ([email protected]) returned to #gstreamer.
[06:19] shawarma ([email protected]) returned to #gstreamer.
[06:28] crazney ([email protected]) joined #gstreamer.
[06:28] <crazney> hey, if i'm not subscribed to gstreamer-devel and i send an email, does it just get silently ignored or go into a queue for aproval?
[06:31] <ds> Uraeus will approve it 
[06:32] <crazney> ok, ta
[06:36] <crazney> also, does anyone happen to know when gstreamer 0.8 will hit debian sid?
[06:37] bilboed ([email protected]) left irc: Read error: 110 (Connection timed out)
[06:40] <thaytoo> crazney: when taaz gets the packages finished
[06:40] <thaytoo> he's working on them
[06:41] <crazney> cool, thanks.
[06:45] <taaz> yeah, and it sucks
[06:45] <taaz> i'm trying to figure out how to make their use easier
[06:50] <taaz> after i finish a few changes i can put them up for people to test if you all want
[06:51] <crazney> taaz: that'd be cool, howcome it sucks?
[06:52] <taaz> it's hard to know how to package large numbers of plugins
[06:52] <dilinger> taaz: i'll test
[06:52] <crazney> taaz: whats wrong with the way its done for 0.6?
[06:53] <taaz> crazney: i don't really like the small packages that just have 1 tiny wrapper plugin in them.  seems like a waste of resources to deal with packages at that level
[06:54] <taaz> crazney: but all-in-one packages would have too many external deps
[06:54] <dilinger> taaz: perhaps you could break them up by general category?
[06:54] <bitshifter> what about something like -plugins-audio-basic -plugins-audio-all -plugins-video-basic -plugins-video-all etc. ?
[06:54] <dilinger> yea
[06:55] <crazney> the category thing would have problems.. what about things that are kinda cross-category, and user might not know whats installed.. or something
[06:55] <crazney> with lotsa small packages at least you know whats installed.
[06:55] <dilinger> video, audio, sound drivers, cd handling plugins, etc
[06:57] <ds> taaz: why not just make one large gst-plugins package, with Recommends: that specify dependencies?
[06:57] <dilinger> taaz: honestly, though, i'd rather see them split out, and if any groupings are done, it's by virtual package
[06:57] <taaz> ds: i do have gstreamer-plugins that depends on all other plugins.  i'm not sure if recommends would be all that useful?
[06:58] <ds> change all the Depends to Recommends
[06:58] <taaz> at the moment it's roughly all packages with external deps are their own.  everything else is in -misc
[06:58] <taaz> where do i put ext/pango/ plugins?
[06:58] <taaz> -pango is not where users would look for gsttextoverlay
[06:59] <ds> why not just make one large package?
[06:59] <crazney> eww recommend instead of depend is gross imho..
[06:59] <ds> it's how it's supposed to be done
[06:59] Action: bitshifter isn't particularly convinced of the Recommends idea either
[06:59] <crazney> ds: no, its a terrible terrible terrible idea.
[06:59] <dilinger> crazney: why?  there's no need to have alsa plugins installed on my system if i don't use alsa..
[07:00] Nick change: mxpxpod -> mx|sleep
[07:00] <crazney> dilinger: say i know an app needs the mad plugin.. so i install gstreamer-plugins.. then i have to and manually install 'libid3tag0 (>= 0.15.0b), libmad0 (>= 0.15.0b), zlib1g (>= 1:1.2.1)'.
[07:00] <taaz> dilinger: you must like this guy: #213176
[07:00] <crazney> most people don't really give a stuff about recommend/suggests, ti just causes heaps of problems
[07:00] <ds> crazney: according to you.  according to experienced Debian developers, it's not
[07:00] <crazney> s/ti/it/
[07:01] <dilinger> crazney: no, then you add a dep for gst-plugin-mad (or whatever the plugin is called)
[07:01] <crazney> ds: maybe so, but i'd argue i'm more of a user than they :)
[07:01] <crazney> dilinger: but then you have the small packages again, so whats the point?
[07:01] <dilinger> ds: ah, tv, my hero ;)
[07:02] <dilinger> s/ds/taaz/
[07:03] <dilinger> taaz: i agree w/ him, though.  if apps need specific functionality provided by a plugin, they should depend upon that plugin specifically.  otherwise, leave it up to the user to choose what plugins to install
[07:03] <taaz> dilinger: you know who that is?
[07:03] <taaz> dilinger: no one is forcing you to install gstreamer-plugins package
[07:03] <crazney> dilinger: unless you meant have individual packages, and the meta package just depend on them. in that case i'm wrong and i agree :)
[07:03] <dilinger> taaz: he was the original intermezzo ITP'er (i later took over the ITP).  i've also run into him in a couple other places
[07:04] <dilinger> taaz: right, unless i need the functionality provided by one of the plugins.  in that case, i'm forced to install _all_ of them.  that's a waste of my system resources
[07:04] <dilinger> it also makes dist-upgrade take forever, as gst-register is run in each postinst
[07:04] <taaz> ds: do you really mean one large large package?  long ago i ran the dependency trace and it was rather large for all gst packages.
[07:06] <taaz> dilinger: lack of incremental register/unregister is an old bug.  feel free to fix it please please.
[07:06] <ds> taaz: yes.  and change all the depends to suggests
[07:07] <dilinger> taaz: heh, even with that fixed, it's still not necessary to have all those plugins taking up space on my machine.   and then when plugins (or libs that plugins depend upon) break, my dist-upgrade is also affected.
[07:08] <taaz> ds: ?  do you mean in addition to lots of small packages?  or instead of one monster plugin bucket?
[07:08] <ds> taaz: _one_ package
[07:09] <dilinger> taaz: i don't see a way to fix the issue w/out a pre-depends, and that'd suck
[07:10] <dilinger> actually, no, i take that back, a normal depends might work
[07:12] <taaz> ds: eh?  that will be very large and have a ton of dependencies.  you saying i should do some magic to conver the automagic library deps to suggests and let plugins fail to load?
[07:12] <ds> <ds> taaz: yes.  and change all the depends to suggests
[07:12] <taaz> dilinger: pre-depends? huh?
[07:12] <dilinger> taaz: can't you just have a single gst-register in the postinst of gstreamer-plugins?   that depends on all the individual plugins, so they should be unpacked and configured by the time its postinst gets called
[07:13] <taaz> dilinger: you're asking for the "hooks" capabilities that are not yet in dpkg
[07:13] <dilinger> taaz: no..
[07:14] <taaz> it's the same sort of thing with wanting to run ldconfig once after a large update rather than for every lib package installed
[07:14] <dilinger> taaz: if gstreamer-plugins is going to depend upon all plugins, then it will be configured after all those plugins have been configured
[07:14] <taaz> dilinger: what if i don't install gstreamer-plugins?
[07:15] <taaz> just the individual plugins
[07:15] <dilinger> heh, i was thinking gstreamer-plugins was necessary for any to be installed
[07:16] <taaz> ds: how would that work?  how is a user supposed to know which libs in a huge suggests list would enable a certain plugin they want?
[07:16] <taaz> dilinger: no, other way around.  it was initially just a lazy way for me to install all plugins.
[07:16] <ds> taaz: it's approximately the same problem as now, just with different names
[07:17] <taaz> ds: i disagree.  if i don't know about ldd and wanted gsttextoverlay to work, how would i know to install pango?
[07:18] <ds> how would you know to install gstreamer-pango?
[07:18] <ds> it's the same situation
[07:19] <bitshifter> if you add a Provides: gstreamer-textoverlay to the gstreamer-pango package, does it show up in the package list?
[07:19] <ds> in either case, it would be wise to have a script/program that tells you what to install to enable specific plugins
[07:19] <ds> bitshifter: no
[07:19] <bitshifter> shame
[07:19] <taaz> not really... you could search for gsttextoverlay in the package descriptions or contents (via the web) to know which package contained it and install that and would assume that would make it work and include all the required deps
[07:20] <ds> taaz: that's the same as knowing how to use apt-cache to see the library suggests and figure out which one you need
[07:21] <taaz> ds: how would you figure it out without using ldd?
[07:21] <ds> for most, it would be easy
[07:22] <ds> Suggests: libmpeg2-3
[07:22] <ds> hmm, I probably need to install that to get mpeg
[07:22] Action: taaz not sure we're talking about the same thing
[07:22] <ds> yes, we are
[07:22] <taaz> it'll be a Suggests list with about 100 items
[07:23] <ds> you could even put the list of packages to install in the gstreamer-plugins description
[07:23] <ds> "for mpeg2dec, install libmpeg2-3"
[07:24] <taaz> i've been dealing with smarter-than-average users recently and i doubt this idea would work
[07:26] <ds> I'm just saying that if you really want to make it convenient and easy, it doesn't matter how the internals work, you must provide a user-friendly interface like a script
[07:26] <taaz> a similar approach is a ton of gstreamer-* packages and gstreamer-plugins with Recomments/Suggests for them and "for mpeg2 decoding, install gstreamer-mpeg2dec [or ...]"
[07:28] <ds> in which case, you still need a user-friendly interface
[07:28] arana ([email protected]) joined #gstreamer.
[07:30] Shoragan ([email protected]) joined #gstreamer.
[07:31] apoc_ ([email protected]) joined #gstreamer.
[07:36] spyder482 ([email protected]) left irc: Read error: 60 (Operation timed out)
[07:37] markey ([email protected]) joined #gstreamer.
[07:38] <markey> moin
[07:41] thaytoo ([email protected]) left irc: "hometime"
[07:48] apoc ([email protected]) left irc: Read error: 110 (Connection timed out)
[08:01] Marsupilami23 ([email protected]) joined #gstreamer.
[08:02] Marsupilami23 ([email protected]) left irc: Client Quit
[08:03] Marsupilami23 ([email protected]) joined #gstreamer.
[08:04] arana ([email protected]) left irc: "Leaving"
[08:06] Marsupilami23 ([email protected]) left irc: Remote closed the connection
[08:07] Marsupilami23 ([email protected]) joined #gstreamer.
[08:07] Marsupilami23 ([email protected]) left irc: Remote closed the connection
[08:13] Marsupilami23 ([email protected]) joined #gstreamer.
[08:17] bilboed ([email protected]) joined #gstreamer.
[08:43] <markey> does GstThread use g_main_loop?
[08:43] Action: markey wondering how to a g_idle_add for my eos callback
[08:43] <markey> while I'm not using a g_main_loop
[08:44] crazney_ ([email protected]) joined #gstreamer.
[08:47] <ds> markey: if you don't use g_main_loop(), you need to use a different idle handler
[08:48] <markey> ds: what kind of?
[08:53] thaytan ([email protected]) joined #gstreamer.
[08:54] <markey> I've googled for "idle handler", but the search onle yields g_main_loop API docs
[08:54] <markey> s/onle/only
[09:08] crazney ([email protected]) left irc: Read error: 110 (Connection timed out)
[09:12] nvrpunk ([email protected]) joined #gstreamer.
[09:12] nvrpunk ([email protected]) left #gstreamer ("Leaving").
[09:20] steve_b_ ([email protected]) joined #gstreamer.
[09:20] steve_b ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[09:31] jcsston ([email protected]) left irc: Read error: 60 (Operation timed out)
[09:46] ChriHJW_log ([email protected]) joined #gstreamer.
[09:57] ChrisHJW_log ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:13] Rotty ([email protected]) joined #gstreamer.
[10:13] Nick change: mathrick|sleep -> mathrick
[10:13] <mathrick> g'morning
[10:14] <mathrick> markey: maybe try "idle func(tion)" ?
[10:14] <markey> moin mathrick 
[10:14] <mathrick> moin
[10:14] <markey> ok let's see
[10:16] <mathrick> taaz: what exactly is the pb with gst-plugins package suggesting all other plugins, each in its own package that would depend on appropriate libs?
[10:17] <markey> mathrick: can you explain to me what exactly I need to do inside of my eos callback, so I can set the thread to STATE_NULL?
[10:18] <markey> it's really hard for me to grok that, coming from KDE land
[10:19] <mathrick> hmm, I don't remember the exact issues that come from doing it directly from callback, I think it was that possibly some elements could be in inconsistent state or something. Lemme grep for a while
[10:19] <mathrick> markey: and I'm almost positive that Qt has some notion of idle handler, just under different name
[10:19] <mathrick> markey: it's extremely common thing to do
[10:20] <markey> is it a matter of thread unsafetyness?
[10:20] <taaz> mathrick: lots of packages
[10:21] <mathrick> taaz: well, that's the smallest issue, isn't it? compared to all deps magic necessary with any other approach
[10:21] <markey> strangely, back with gst 0.7.3 I could manipulate the thread element from within the callback
[10:21] <markey> no problem
[10:22] <mathrick> taaz: btw, it's kinda standard way to do it, so ppl would expect something like that
[10:22] <mathrick> markey: I wouldn't count on it working
[10:22] <markey> KDE GST bindings don't seem to care for this either
[10:22] <markey> I guess it breaks them 
[10:26] <mathrick> markey: http://lists.trolltech.com/qt-interest/2003-01/thread00792-0.html
[10:26] <markey> heh, been looking at the same page
[10:26] <markey> isn't much help, tho
[10:26] <markey> but I've got an idea, I'll try the QTimer stuff
[10:27] <mathrick> I guess Qt doesn't do it explicitly as idle handler
[10:27] <mathrick> which is strange enough
[10:27] <markey> might work, let's see
[10:28] <mathrick> I wonder if timer == 0 means "as often as possible", or "once per jiffy" ?
[10:28] <mathrick> if the latter, then it isn't the exact equivalent of idle handler
[10:29] <markey> As a special case, a QTimer with timeout 0 times out as soon as all the events in the window system's event queue have been processed. 
[10:30] <mathrick> ok, so it looks like a true idle handler then
[10:32] dolphy ([email protected]) joined #gstreamer.
[10:32] <mathrick> yo dolphy
[10:36] <dolphy> mornign
[10:37] <markey> BINGO
[10:37] <markey> :)
[10:37] <markey> works nicely
[10:38] <markey> thx mathrick 
[10:38] <mathrick> markey: np
[10:44] jdahlin ([email protected]) joined #gstreamer.
[10:44] <mathrick> hi jdahlin
[10:44] <jdahlin> morning
[10:58] markey ([email protected]) left irc: "bbl"
[11:16] kmaraas ([email protected]) joined #gstreamer.
[11:36] harshy ([email protected]) left irc: Remote closed the connection
[11:37] The_Company ([email protected]) joined #gstreamer.
[11:45] jonathang ([email protected]) joined #gstreamer.
[11:46] Company ([email protected]) left irc: Read error: 60 (Operation timed out)
[11:48] foser ([email protected]) joined #gstreamer.
[12:00] thaytan ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[12:02] thaytan ([email protected]) joined #gstreamer.
[12:05] Nick change: mathrick -> mathrick|Uni
[12:17] <jdahlin> For a pipeline like this; avidemux name=demux ! { queue ! spider ! ffmpegcolorspace ! xvimagesink } { demux. ! queue ! spider ! osssink }
[12:17] <jdahlin> How can I express the "demux." without using a string?
[12:17] <jdahlin> eg, { } are GstThreads etc
[12:23] <thaytan> I don't think you can make both links without referencing the demuxer twice like that
[12:23] <thaytan> why do you care?
[12:25] <jdahlin> I'd like to do that in a program, instead of using gst-launch
[12:26] sublett ([email protected]) joined #gstreamer.
[13:08] <The_Company> good morning angels
[13:08] <thaytan> hi Benjamin
[13:09] <jdahlin> lovely, www.gnome.org hacked.
[13:09] <thaytan> awr crap
[13:09] Nick change: The_Company -> Company
[13:10] <thaytan> jdahlin: how do you know?
[13:11] <Company> #gnome-hackers
[13:29] jdahlin ([email protected]) got netsplit.
[13:29] dolphy ([email protected]) got netsplit.
[13:29] taaz ([email protected]) got netsplit.
[13:29] jimmy_dean ([email protected]) got netsplit.
[13:29] md` ([email protected]) got netsplit.
[13:29] dilinger ([email protected]) got netsplit.
[13:29] shawarma ([email protected]) got netsplit.
[13:29] joem ([email protected]) got netsplit.
[13:29] Misirlou ([email protected]) got netsplit.
[13:30] sublett ([email protected]) got netsplit.
[13:30] Company ([email protected]) got netsplit.
[13:30] danb ([email protected]) got netsplit.
[13:30] bluejay ([email protected]) got netsplit.
[13:30] Rotty ([email protected]) got netsplit.
[13:30] wtay_ ([email protected]) got netsplit.
[13:30] ds ([email protected]) got netsplit.
[13:30] lilo ([email protected]) got netsplit.
[13:30] ChriHJW_log ([email protected]) got netsplit.
[13:30] bilboed ([email protected]) got netsplit.
[13:30] Marsupilami23 ([email protected]) got netsplit.
[13:30] jonathang ([email protected]) got netsplit.
[13:30] apoc_ ([email protected]) got netsplit.
[13:30] mx|sleep ([email protected]) got netsplit.
[13:30] bitshifter ([email protected]) got netsplit.
[13:30] steve_b_ ([email protected]) got netsplit.
[13:30] sri ([[email protected]) got netsplit.
[13:30] hyriand ([email protected]) got netsplit.
[13:30] grub_booter ([email protected]) got netsplit.
[13:30] tromey ([email protected]) got netsplit.
[13:30] ds-work ([email protected]) got netsplit.
[13:30] sjoerd ([email protected]) got netsplit.
[13:30] Jara[zZ] ([email protected]) got netsplit.
[13:30] Shoragan ([email protected]) got netsplit.
[13:30] foser ([email protected]) got netsplit.
[13:30] thaytan ([email protected]) got netsplit.
[13:30] mathrick|Uni ([email protected]) got netsplit.
[13:30] je4d ([email protected]) got netsplit.
[13:30] sxpert_work ([email protected]) got netsplit.
[13:30] aldug ([email protected]) got netsplit.
[13:31] Misirlou ([email protected]) returned to #gstreamer.
[13:31] joem ([email protected]) returned to #gstreamer.
[13:31] shawarma ([email protected]) returned to #gstreamer.
[13:32] jdahlin ([email protected]) returned to #gstreamer.
[13:32] dolphy ([email protected]) returned to #gstreamer.
[13:32] md` ([email protected]) returned to #gstreamer.
[13:32] jimmy_dean ([email protected]) returned to #gstreamer.
[13:32] dilinger ([email protected]) returned to #gstreamer.
[13:32] taaz ([email protected]) returned to #gstreamer.
[13:33] sublett ([email protected]) returned to #gstreamer.
[13:33] thaytan ([email protected]) returned to #gstreamer.
[13:33] foser ([email protected]) returned to #gstreamer.
[13:33] jonathang ([email protected]) returned to #gstreamer.
[13:33] Company ([email protected]) returned to #gstreamer.
[13:33] Rotty ([email protected]) returned to #gstreamer.
[13:33] ChriHJW_log ([email protected]) returned to #gstreamer.
[13:33] steve_b_ ([email protected]) returned to #gstreamer.
[13:33] bilboed ([email protected]) returned to #gstreamer.
[13:33] Marsupilami23 ([email protected]) returned to #gstreamer.
[13:33] apoc_ ([email protected]) returned to #gstreamer.
[13:33] Shoragan ([email protected]) returned to #gstreamer.
[13:33] mx|sleep ([email protected]) returned to #gstreamer.
[13:33] lilo ([email protected]) returned to #gstreamer.
[13:33] je4d ([email protected]) returned to #gstreamer.
[13:33] mathrick|Uni ([email protected]) returned to #gstreamer.
[13:33] danb ([email protected]) returned to #gstreamer.
[13:33] bluejay ([email protected]) returned to #gstreamer.
[13:33] aldug ([email protected]) returned to #gstreamer.
[13:33] sxpert_work ([email protected]) returned to #gstreamer.
[13:33] sri ([[email protected]) returned to #gstreamer.
[13:33] grub_booter ([email protected]) returned to #gstreamer.
[13:33] Jara[zZ] ([email protected]) returned to #gstreamer.
[13:33] tromey ([email protected]) returned to #gstreamer.
[13:33] ds-work ([email protected]) returned to #gstreamer.
[13:33] bitshifter ([email protected]) returned to #gstreamer.
[13:33] hyriand ([email protected]) returned to #gstreamer.
[13:33] sjoerd ([email protected]) returned to #gstreamer.
[13:33] ds ([email protected]) returned to #gstreamer.
[13:33] wtay_ ([email protected]) returned to #gstreamer.
[14:00] iain ([email protected]) joined #gstreamer.
[14:12] <Company> that doesn't sound nice
[14:13] <Company> howmuch delay do we expect for gnome 2.6? :/
[14:14] <thaytan> enough time to rebuild and verify all source :-/
[14:21] <Company> looking at debian, that might result in more than a month :(
[14:29] <foser> afaik only widget got hacked (?)
[14:30] <thaytan> so far, yes
[14:30] <jdahlin> gnome is not as big as debian
[14:31] <Company> gnome is not as organized as debian
[14:32] <jdahlin> that's not necessarily bad
[14:32] <Company> what's on widget btw?
[14:33] <jdahlin> (www|bugzilla|planet|art|foundation)gnome.org
[14:34] <Company> so not the really critical stuff
[14:36] <thaytan> bugzilla, man BUGZILLA!
[14:36] <thaytan> how are we gonna know our bugs?!
[14:36] <thaytan> what if they inserted NEW ONES?
[14:36] <jdahlin> then we might just need to INVALIDate them
[14:36] <Company> i'd check those bugs for validity and either close them or fix them? :p
[14:37] <thaytan> ber, how am I going to start a panic with attitudes like that to work with?
[14:37] <jdahlin> deal, Company gets to fix all non-bogus bug reports reported by h4xx0rs
[14:37] <Company> would result in nice changelog entries ;)
[14:37] teuf ([email protected]) joined #gstreamer.
[14:38] <teuf> hi
[14:38] <Company> ho
[14:39] <teuf> I was wondering if there was anything planned for upcoming 0.8.x releases, ie are you focusing on bug fixing, will new features (I'm thinking about all these dvd patches) go ?
[14:40] <Company> as far as i'm concerned new features are fine anytime
[14:40] <teuf> Company: I could have guessed that ;)
[14:40] <jonathang> hello, am using cvs and always get a SIGFPE while using the adder element and a scheduler different than opt
[14:41] <jonathang> with opt, i get no FPE and no sound :(
[14:41] <Company> that sounds a lot like stack corruption
[14:41] <Company> but i remember someone saying adder was kind of broken, though that was quite a while ago
[14:42] <thaytan> teuf: our goal is to maintain backwards compat 
[14:42] <jdahlin> jonathang: I think adder is broken
[14:42] <thaytan> so new features are fine if they don't break old apps
[14:42] <jdahlin> jonathang: try to poke thomasvs on monday (when he comes back)
[14:42] <jonathang> do you have any ideas how i could repair it? :)
[14:42] <jdahlin> nope
[14:42] <jdahlin> I'm not plugin guru
[14:43] <jonathang> jdahlin, on MONDAY?
[14:43] <Company> we don'T have "plugin gurus" - we only have people that know a partcular plugin :)
[14:43] <Company> and i don't know adder
[14:43] <teuf> thaytan: modifying existing code, especially in gst-plugins is bound to break old apps in subtle ways imo
[14:44] <jdahlin> jonathang: right
[14:44] <thaytan> teuf: hopefully not
[14:44] <Company> teuf: we don't guarantee your app breaking the same way after an update
[14:44] <jdahlin> Company: adder should be reasonably easy to fix, no?
[14:44] <Company> jdahlin: _no_ idea, i have never looked at what it does and how
[14:45] <jdahlin> just mixes two different streams into one
[14:45] <teuf> thaytan: if you are still "hopeful", that means you haven't been programming long enough ;)
[14:45] <jonathang> yes, sounds simple
[14:45] <jonathang> :)
[14:45] <jonathang> but isn't (for me)
[14:45] <jdahlin> Company: this is the golden oppertunity to do so!
[14:45] Action: jdahlin runs
[14:45] <Company> jdahlin: go ahead
[14:45] <Company> (as i'm sure you hadn't looked either :p)
[14:46] <Company> i'm busy enough maintaining close to all rb plugins :/
[14:48] smoke ([email protected]) joined #gstreamer.
[14:49] md` ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[15:05] <Company> if i had read the ASF specs, i'd now have complained that MS has absolutely no clue about writing specs either
[15:09] <Company> at least i understand why it's so hard to write correct code for handling this data
[15:10] Action: Company wonders how many exploits can be found in MS' asf handling code
[15:14] md` ([email protected]) joined #gstreamer.
[15:26] mathrick|Uni ([email protected]) left irc: Remote closed the connection
[15:30] <jdahlin> rb plugins?
[15:31] <teuf> jdahlin: I guess he means "plugins used by rb" (ie flac, mad, vorbis)
[15:31] <jdahlin> ok
[15:31] <Company> vorbis, ogg, autoplugger, alsa, id3tag
[15:32] <Company> mad, oss, filesrc, gnomevfsssrc are not mine, but i do a lot of work there, too
[15:32] <Company> oh, and audioconvert
[15:32] <Company> those have kept me busy quite enough
[15:34] <jdahlin> ok
[15:34] <jdahlin> anyway, time for some demonstration!
[15:34] <jdahlin> import gst, gst.play
[15:34] <jdahlin>     play = gst.play.Play()
[15:34] <jdahlin>     play.set_data_src(gst.Element('filesrc'))
[15:34] <jdahlin>     play.set_audio_sink(gst.Element('osssink'))
[15:34] <jdahlin>     play.set_video_sink(gst.Element('xvimagesink'))
[15:34] <jdahlin>     play.set_location(sys.argv[1])
[15:34] <jdahlin>     play.set_state(gst.STATE_PLAYING)
[15:34] <jdahlin>     gst.main()
[15:35] <jdahlin> voila, a movie player in 8 lines of code
[15:36] <teuf> 8 lines ? this should be I.want.a.fucking.movie.player(movie) ;)
[15:36] <Company> the beauty of gst: it takes more code to query the number of channels than it takes to setup a player
[15:36] <jdahlin> I know, I'm working on that
[15:36] <jdahlin> Company: not really:
[15:36] <jdahlin>         pad = sink.get_pad('sink')
[15:36] <jdahlin>         for structure in pad.get_caps():
[15:36] <jdahlin>             if not structure.has_field('channels'):
[15:36] <jdahlin>                 continue
[15:36] <jdahlin>             channels = structure['channels']
[15:37] <jdahlin> but almost
[15:37] <Company> teuf: the problem with that function is people doing I.want.a.fucking.movie.player(something)
[15:38] markey ([email protected]) joined #gstreamer.
[15:38] <Company> teuf: if you give me a Is.this.a.fucking.movie(something), i'll write you the other function :p
[15:38] <iain> Company: surely if you watch it you'll know if its a fucking movie or not
[15:38] Action: teuf writes a function which always return False, and doesn't even need the function to play the movie ;)
[15:39] <Company> yeah, but unfortunately gst doesn't watch it beforehand
[15:39] <jdahlin> gst sucks, it should watch the movie before showing it telling me if it's good or not
[15:39] <Company> oh, rb will do that for audio some day
[15:39] <teuf> don't forget a buffy detector to make micke happy ;)
[15:40] Action: Company is pretty sure it'll be possible to query other people's ratings
[15:40] <Company> ... some day
[15:40] <iain> hmm, I want to go and ride my bike, but I dunno where to go
[15:41] <Company> that would rock btw
[15:41] <Company> to be able to query other people's song ratings
[15:42] <Company> if you combined that with the shuffle algorithm, you could then write an apache mpodule that automatically creates charts
[15:43] <Company> just query all rbs available and apply an algorithm
[15:43] <Company> go p2p!
[15:43] Action: Company gets back to earth
[15:47] kmaraas ([email protected]) got netsplit.
[15:47] crazney_ ([email protected]) got netsplit.
[15:47] sxpert ([email protected]) got netsplit.
[15:50] kmaraas ([email protected]) returned to #gstreamer.
[15:50] crazney_ ([email protected]) returned to #gstreamer.
[15:50] sxpert ([email protected]) returned to #gstreamer.
[15:51] sublett ([email protected]) left irc: Read error: 60 (Operation timed out)
[15:59] walters ([email protected]) joined #gstreamer.
[16:09] <jdahlin> Company: do you know how I am supposed to query for the current values for a GstIntRange?
[16:09] <jdahlin> g_value_get () seems to return the minimum value
[16:11] trow ([email protected]) joined #gstreamer.
[16:16] <Company> jdahlin: there's API for that in gstvalue.h
[16:17] <jdahlin> Right.
[16:17] <jdahlin> Company: or, not really
[16:17] <jdahlin> Company: there is one to retreive the max value and another one for the min
[16:17] <jdahlin> but what about the *current* ?
[16:18] <Company> a range has no current value
[16:18] <jdahlin> or maybe I'm just confused
[16:18] <jdahlin> so how can I figure out how many channels a specific stream has?
[16:18] <iain> you can only get a "current" value from a fixed caps
[16:19] <iain> I had the same problem when I started on marlin :)
[16:19] <jdahlin> okay, how can I retrieve this magical fixed caps?
[16:19] <Company> get_negotiated_caps
[16:19] <iain> wish I knew :)
[16:19] <iain> oh, thats how
[16:20] <Company> that returns NULL if there are no negotiated caps yet
[16:20] <Company> and otherwise the fixed caps that are negotiated
[16:21] <jdahlin> are there any functions I can connect to when they are negoitated?
[16:21] <jdahlin> s/functions/signals/
[16:26] <Company> it does a notify::caps
[16:30] <jdahlin> so I need to connect to deepnotify to retrieve it?
[16:30] <Company> depends on what you want
[16:31] <Company> if you know the pad, connect to the pad
[16:31] <Company> notify::caps on the pad
[16:31] <Company> otherwise connect to deepnotify::caps on the element
[16:33] jcsston ([email protected]) joined #gstreamer.
[16:37] <jdahlin> Company: thanks, seems to work just fine
[16:37] markey ([email protected]) left irc: "bbl"
[16:43] <thaytan> Company: thank you for the debugging subsystem. it is cool.
[16:44] yippi ([email protected]) joined #gstreamer.
[16:51] jcsston ([email protected]) left irc: Read error: 60 (Operation timed out)
[16:53] <Company> thaytan: you're the third person to tell me that
[16:53] <Company> thaytan: thanks
[16:53] Action: Company should suggest that for glib :)
[16:58] kmaraas ([email protected]) left irc: Remote closed the connection
[16:59] sublett ([email protected]) joined #gstreamer.
[17:06] <iain> hey BBB is right...with basicgthread even my lene marlin asf file plays
[17:07] <Company> uh
[17:07] <Company> asfdemux kills opt
[17:07] <jdahlin> crash?
[17:07] <Company> opt can't handle loopbased elements that loop infinitely
[17:07] <jdahlin> oh
[17:07] <Company> "kill" in the sense of "make unusable"
[17:07] <iain> well, opt is broken with avi,asf and mov
[17:07] <jdahlin> plays avis fine here
[17:08] <iain> here it only plays audio
[17:09] <iain> hmm, nothing seems to play wmv's
[17:09] <iain> but they make nice patterns on the screen :)
[17:09] <iain> and movs don't play, but at least give errors why
[17:10] <Company> i'm currently trying to write a nice asf demuxer
[17:11] <Company> but the asf format makes that harder than it needs to be :/
[17:11] <iain> write a nice scheduler :)
[17:11] <Company> it's on my list
[17:11] kmaraas ([email protected]) joined #gstreamer.
[17:13] Action: iain debates writing a cd burner in c# using gst# and libburn
[17:13] <jonathang> Company, so that's why adder crashes opt too
[17:16] <Company> weeeee
[17:16] <Company> i get output :)
[17:19] <Company> step 2: get useful output
[17:21] kmaraas ([email protected]) left irc: Client Quit
[17:22] kmaraas ([email protected]) joined #gstreamer.
[17:23] <thomasvs> whee
[17:23] <jdahlin> lo thomasvs
[17:23] <jdahlin> already back? :)
[17:23] <thomasvs> no, hacked the local wireless network
[17:23] <thomasvs> well
[17:24] <thomasvs> I think I am supposed to pay for it
[17:24] <jdahlin> :)
[17:24] <jdahlin> so how's the snow in France?
[17:24] <thomasvs> well, first two days too much sun and slushy snow
[17:24] <thomasvs> and now we have a blissard, but the snow is great :)
[17:24] <thomasvs> how's work ? :)
[17:24] <jdahlin> bummer
[17:24] <thomasvs> no, it's great
[17:25] <jdahlin> just nice :-P
[17:25] <thomasvs> I mean, I don't see much, but the snow makes up for it
[17:25] <thomasvs> and tomorrow, nice sun, good tracks
[17:25] <jdahlin> continuing with hacking gst-python
[17:25] <jdahlin> let's hope it's not a blizzard tomorrow
[17:26] <thomasvs> well, one more day is fine, better snow afterwards
[17:26] <thomasvs> there's enough to do in the evening anyways
[17:29] kwm ([email protected]) joined #gstreamer.
[17:29] <dolphy> thomasvs: huhu :)
[17:29] <dolphy> thomasvs: hi :)
[17:30] BOFH ([email protected]) left irc: Read error: 60 (Operation timed out)
[17:31] <thomasvs> dolphy: evening :)
[17:32] <dolphy> thomasvs: theora is in alpha 3
[17:33] Uraeus ([email protected]) joined #gstreamer.
[17:33] <jdahlin> it even made it to /.
[17:33] BOFH ([email protected]) joined #gstreamer.
[17:33] <dolphy> thomasvs: they didn't mention fluendo ye
[17:33] <dolphy> yet
[17:33] <dolphy> thomasvs: will be in the next step
[17:33] <Uraeus> hi guys
[17:33] <dolphy> Uraeus: hi christian
[17:35] <Uraeus> jdahlin: 18 degrees eh? here it si snowing today
[17:36] <thomasvs> dolphy: it was in the announcement mail
[17:36] <jdahlin> Uraeus: only 15 today
[17:36] <thomasvs> dolphy: but anyway, we need to coordinate a bit for the next one
[17:38] <thomasvs> anyways, time for apres-ski
[17:39] <Uraeus> jdahlin: sounds cold ;)
[17:39] <dolphy> thomasvs: hehe
[17:39] <Uraeus> so have any of you put in bids on my appartment?
[17:39] smoke ([email protected]) left irc: Remote closed the connection
[17:40] <iain> I'll give you 50quid for it?
[17:40] <Uraeus> hmm, just the floor tiles are worth more than that :)
[17:41] <thaytan> 55, here
[17:41] smoke ([email protected]) joined #gstreamer.
[17:41] <iain> damn
[17:41] <iain> if I'd have known there was porn I'd have brought more money
[17:42] <thaytan> interesting segue, where are you going with it?
[17:42] <iain> nowhere...just quoting the simpsons
[17:43] <iain> when I run out of things to say, I quote the simpsons...
[17:43] <iain> just like everyone on /.
[17:43] lilo ([email protected]) left irc: "brb"
[17:44] lilo ([email protected]) joined #gstreamer.
[17:52] <jdahlin> dolphy: you're maintaining GstPlay right?
[17:53] <dolphy> jdahlin: yeah i try
[17:54] <jdahlin> dolphy, can I commit http://www.mdstud.chalmers.se/~jdahlin/gst-play-fakesink.diff?
[17:54] <jdahlin> dolphy: so I can be able to use it with a fakesink for audio or video
[17:55] <jdahlin> would be nice to have filesrc instead of fakesrc as default source too
[17:57] spyder482 ([email protected]) joined #gstreamer.
[17:57] <dolphy> jdahlin: go ahead i read it
[17:58] <jdahlin> dolphy: I will, thanks
[18:10] <Uraeus> jdahlin: managed to look at my gsr bug report yet?
[18:10] <jdahlin> Uraeus: oh, I forgot!
[18:10] Action: Uraeus spanks jdahlin
[18:11] Nick change: mx|sleep -> mxpxpod
[18:12] <iain> martin soto wrote a scheduler
[18:12] <iain> sweet
[18:13] thomasvs ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[18:14] <Uraeus> iain: is it any good?
[18:14] <iain> dunno...see the list post
[18:16] kmaraas ([email protected]) left irc: Client Quit
[18:17] <Company> commit it
[18:21] thomasvs ([email protected]) joined #gstreamer.
[18:32] markey ([email protected]) joined #gstreamer.
[18:41] <ds> strangely, I couldn't get anything working with any scheduler last night
[18:44] BBB ([email protected]) joined #gstreamer.
[18:44] <BBB> dolphy: screenshot of totem will come tomorrow, I wasn't at hoime yesterday
[18:44] <Uraeus> hi BBB
[18:45] <BBB> as for the scheduler: I really appreciate the investment, but I'd seriously like some discussion on the topic before everyone goes do random things on it
[18:45] <BBB> thinking before acting, you know? :)
[18:45] <BBB> hi Uraeus!
[18:45] <iain> awww muine compiles but won't run :/
[18:45] <BBB> I thought it was called marlin?
[18:45] <Uraeus> BBB: well it was always the plan to have multiple scheduler with different atributes right?
[18:46] <BBB> Uraeus: yes
[18:46] <iain> muine is jorn's new c# audio player
[18:46] <BBB> but if we want a good default scheduler, we'd better do some talking on that
[18:46] <Uraeus> BBB: muine is Jorn's C# music player using gst
[18:46] <iain> marlin both compiles AND runs
[18:46] <BBB> ah, allrigth
[18:46] <iain> on more than one person's box
[18:46] <BBB> whee
[18:46] <BBB> we have a pro audio editor ;)
[18:48] Action: Uraeus raise his hand as the second person to have been able to compile and run Marlin
[18:48] <BBB> heh :)
[18:48] Action: BBB has some working gst-rec code at home
[18:48] <BBB> gst-rec is really cute, you know that?
[18:49] <BBB> oops
[18:49] <BBB> class
[18:49] <BBB> bbl!
[18:49] BBB ([email protected]) left #gstreamer.
[18:49] <Uraeus> well until gstreamer supports by webcam I don't really have anything to record with gst-rec, but I guess I can try building it anyway :)
[18:50] <iain> I don't even have a webcam...
[18:50] <Company> someone could write an X window recorder
[18:50] <Uraeus> well ds vnc src could do that job I would think
[18:51] <iain> I really want cdrsink
[18:51] <Company> thewn use that ;)
[18:52] <Uraeus> iain: isn't our cdrecord element actually a cdrsink ?
[18:52] <iain> we have one?
[18:52] <teuf> I fear a gapless cdrsink would be tricky to do :-/
[18:53] <iain> teuf: probably :/ but it would be cool as
[18:55] <ds> gapless playback is really easy
[18:55] <ds> it's just that nobody does it
[18:56] ploum ([email protected]) joined #gstreamer.
[18:57] <teuf> ds: in the cd recording case, the problem is that you need to know the complete cd layout before starting to burn it if you want to use dao to get a gapless cd
[19:09] Nick change: apoc_ -> apoc
[19:10] teuf ([email protected]) left irc: "Client exiting"
[19:10] <apoc> hi
[19:16] kmaraas ([email protected]) joined #gstreamer.
[19:16] <taaz> does arts still build for people?  it's not working on my debian box
[19:17] <taaz> doesn't add the /usr/include/kde/arts/ dir but tries to include artsflow.h
[19:18] <taaz> in the m4/as-arts.m4 check code i think
[19:19] <Company> you should try poking the KDE people, though I don't think they use arts output either ;)
[19:19] <Company> i wouldn't be surprised if it didn't work
[19:19] <taaz> KDE people: help
[19:19] <Company> ds has sent me 3 viruses today
[19:20] <Company> bad David
[19:20] <taaz> that's actually a good sign.  probably means windows users are reading gstreamer mailing lists ;)
[19:24] <Company> or it's the matroska guys ;)
[19:25] mathrick ([email protected]) joined #gstreamer.
[19:25] <apoc> yo mathrick
[19:26] <mathrick> yo apoc & all
[19:26] <Company> anyone want to write a 100 word bio about me?
[19:27] <Company> iain? ;)
[19:28] <iain> yes?
[19:28] Blaasvis ([email protected]) joined #gstreamer.
[19:29] <iain> only if you can answer my "what are your greatst achievments"
[19:29] <taaz> tell me how to fix arts crap and add that to you achivements list
[19:30] <Uraeus> The Man, The Myth, The Legend - The Story of Benjamin
[19:30] <jdahlin> taaz: did you see the GstPlay example?
[19:30] <Company> Uraeus: 90 more to go
[19:30] <Uraeus> He was born in Hamburg on a cold winters night while the wolves where howling outside
[19:31] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[19:32] <taaz> jdahlin: yeah.  you forgot to import sys and i don't know what gst.main() would do to know about the player pipeline
[19:32] jonathang ([email protected]) left irc: "Leaving"
[19:32] <Uraeus> In the crying face of that little todlet few signs could be seen of the man he would grow up to become, but the world learn, oh yes it would learn
[19:33] <jdahlin> taaz: it works however, gst.main makes sure the GThread continues to run
[19:34] <Company> Uraeus: the setting is nice, but it could be a little bit more on topic ;)
[19:34] <jdahlin> taaz: but maybe I'll wrap it and make it work a bit smoother, so you can just do Player(file).run()
[19:34] <Uraeus> Company: So this isn't for your Evil Genious bio?
[19:34] <taaz> jdahlin: oh yeah... i forgot player adds itself to event loop.
[19:35] <jdahlin> taaz: I don't think it's player specific, I'm quite sure it's GThread that does that
[19:35] <bitshifter> Company: if you did Zivildienst you could dig out that essay you had to write ... ;)
[19:35] <Company> bitshifter: diabetics don't need to do that
[19:35] <bitshifter> ah well
[19:35] <Company> Uraeus: does guadec count as evil?
[19:36] <taaz> jdahlin: my old player tests all did some sort of manual pipeline iteration.  i can't remember what gstplay does but that method didn't work iirc.  whatever... as long as it works ;)
[19:36] Action: taaz runs off for a bit
[19:37] apoc_ ([email protected]) joined #gstreamer.
[19:37] <Company> yay, gimp 2.0 is out
[19:37] Action: Company is a #commits junkie
[19:39] <Company> apoc: you should now the cicl script for commits :)
[19:39] <Company> s/now/use/
[19:44] apoc ([email protected]) left irc: Read error: 60 (Operation timed out)
[19:44] Nick change: apoc_ -> apoc
[19:44] <apoc> Company: cicl ?
[19:46] <Company> apoc: a script ds sent to the mailing list a month or so ago that automatically checks in the files you put in the changelog
[19:47] <ds-work> iain: around?
[19:47] <apoc> Company: oh, i'll have a look
[20:02] jdahlin ([email protected]) left irc: "Leaving"
[20:07] <trow> Anybody seen thomasvs around today?
[20:08] <Company> trow: he's on holidays this week
[20:10] <trow> Company: ah, OK.  Thanks.
[20:16] <iain> ds-work: ?
[20:16] <ds-work> iain: you still have the "special" sound card?
[20:16] <iain> yes
[20:16] <ds-work> could you update gst-plugins/sys/oss and run oss_probe?
[20:17] <iain> its not connected or compiled in
[20:17] <ds-work> oss_probe doesn't depend on gstreamer, so you can scp the binary over
[20:18] <ds-work> actually, enable the '#if 0'd block on line 59, too
[20:18] <iain> no, I mean I don't have the extigy connected to any machine
[20:18] <ds-work> oh, I see
[20:18] <iain> I'll update my kernel/alsa stuff tonight and see if I can do it tomorrow
[20:19] <ds-work> using alsa doesn't work, since alsa does resampling internally
[20:19] <ds-work> (poorly, too)
[20:19] <iain> oh, I use alsa
[20:20] <Company> they closed my bug btw
[20:20] Action: ds-work wonders if it was iain with the strange sound card that didn't like 44100 hz
[20:20] <Company> i think i'm gonna close the gstreamer bug about that failed assertion, too
[20:20] <iain> ds-work: yes it was
[20:20] <Company> yeah, it was iain
[20:20] <iain> ds-work: the extigy
[20:20] <Company> jdahlin claimed to have such a card, too
[20:21] <iain> ds-work: but alsa never worked on it (although they claim it does now) but I gave up and went back to using my SB Live
[20:21] <iain> Company: how/why did they close it?
[20:23] <Company> iain: noone followed up on the bug report :)
[20:23] <ds-work> that sounds like half the gstreamer bug reports
[20:24] <Company> yeah
[20:24] <Company> i ran after iain and jdahlin and whom else i could time, but noone cared
[20:24] <Company> *shrug*
[20:25] Action: iain hates alsa
[20:26] lupusBE ([email protected]) joined #gstreamer.
[20:26] <Company> i start liking it more again
[20:26] <Company> now that they fix bugs and have a bugtracker
[20:31] teuf ([email protected]) joined #gstreamer.
[20:31] <Uraeus> http://www.osnews.com/story.php?news_id=6460 <- Ok, my SVG article is up :)
[20:33] <sri> greetings folks
[20:33] <sri> hi teuffy
[20:34] <Company> hi master of the summary
[20:45] Shoragan ([email protected]) left irc: "Leaving"
[20:47] <Company> YES!
[20:47] <Company> my asf demuxewr works :)
[20:48] <Company> die asf, die
[20:48] <Uraeus> hehe :)
[20:56] jdahlin ([email protected]) joined #gstreamer.
[21:05] coqRe ([email protected]) joined #gstreamer.
[21:18] <taaz> Uraeus: nice article
[21:18] <Uraeus> taaz: thanks :)
[21:19] <jdahlin> yeah, it was a nice read
[21:27] <taaz> well... except that like a fool i follow links and end up spending an hour browsing neat sites ;)
[21:28] <Company> our wmv ffmpeg plugin needs to be updated
[21:28] <Company> someone update ffmpeg!!!
[21:30] thomasvs ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[21:33] BBB ([email protected]) joined #gstreamer.
[21:36] jdahlin ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[21:36] <Company> wmv2 looks cool at least
[21:38] <BBB> video?
[21:38] Action: BBB agrees
[21:38] <mathrick> taaz: it's common problem. It eats almost all of my free time :>
[21:39] <BBB> too bad totem-opt cannot playback anything... it's even worse than I thought in opt
[21:39] <BBB> didn't test opt for a while
[21:39] <Company> BBB: how old is our ffmpeg snapshot?
[21:39] <BBB> a few weeks
[21:39] <BBB> 2, I think
[21:39] <BBB> maybe 3
[21:40] <BBB> I'm hoping for thomasvs to document me how to update the ffmpeg mirror
[21:40] <Company> hrm
[21:40] <BBB> then, I'll update every few days or so
[21:40] <Blaasvis> BB
[21:40] <Company> why does my wmv look good in totem but sucks in gstreamer?
[21:40] <Blaasvis> BBB
[21:40] <BBB> ey visje!
[21:40] <Company> (totem on ppc)
[21:40] <BBB> Company: totem-xine?
[21:40] <Blaasvis> ey bubje... i was weating for you ;)
[21:40] <BBB> it looks great on my totem-gst
[21:40] <Company> BBB: yeah
[21:41] <BBB> #define sucks?
[21:41] <BBB> Blaasvis: XPlain!
[21:41] <Blaasvis> i got a major problem with dvd playback and gstreamer... 
[21:41] <BBB> we're currently working on that :)
[21:41] <Blaasvis> it goes shock shock ethx.. 
[21:41] <Company> mplayer plays it via native dlls
[21:41] <Blaasvis> etc
[21:41] <BBB> ah right
[21:41] <BBB> in totem?
[21:41] <Company> at least on my x86
[21:41] <BBB> totem --gst-scheduler=basicgthread
[21:41] <BBB> and enjoy
[21:42] <Blaasvis> didn't help... 
[21:42] <BBB> hm, that's odd
[21:42] <Blaasvis> tried that yesterday already 
[21:42] <BBB> Company: wmv2 looks great locally, so might be ppc specific
[21:42] <BBB> please provide screenshots and a sample file for me
[21:43] thomasvs ([email protected]) joined #gstreamer.
[21:43] <Company> BBB: it plays on PPC
[21:43] <Blaasvis> BBB, you mean me ?
[21:43] <Company> BBB: it doesn't play on gstreamer though ;)
[21:44] <Blaasvis> BBB, are you running totem cvs ?
[21:44] <BBB> Blaasvis: no
[21:44] <BBB> I meant company
[21:45] <BBB> and yes I'm running totem-cvs
[21:45] <BBB> Company: that's what I mean
[21:45] <Blaasvis> let's tru gst-player 
[21:45] <BBB> Company: give me a sample file that doesn't playback on totem-ppc
[21:45] <BBB> then give me a ppc
[21:45] <BBB> ;)
[21:45] <BBB> or test it on x86 yourself
[21:45] <BBB> totem-gst is able to playback wmv2 here fine
[21:45] <Company> BBB: it doesn't playback on gstreamer no matter where
[21:45] <BBB> oh
[21:45] <BBB> that's wierder
[21:46] <BBB> isn't it wmv9 then?
[21:46] <Blaasvis> cool a big error in gst-player :)
[21:46] <Blaasvis> Additional debug info: gstpad.c(3096): gst_pad_pull: /work_thread/autoplugger/src_0:
[21:46] <Blaasvis> pull on pad src_0:sink but the peer pad mpegdemux0:private_stream_2 has no gethandler
[21:46] Action: BBB is so annoyed by opt...
[21:46] <BBB> Blaasvis: --gst-scheduler=basicgthread :)
[21:47] <Blaasvis>  gst-player --gst-scheduler=basicgthread
[21:47] <foser> the solution to everything (TM)
[21:47] <Blaasvis> this way i started it :)
[21:47] <Company> BBB: ed2k://|file|Silbermond Rockt Video Berlin Machs Dir Selbst Deutsch German.wmv|11657308|dd5979de61364c62ff50c70e840defe2|/
[21:47] <Company> ;)
[21:47] <BBB> foser: if opt is your problem, then yes
[21:47] <BBB> Blaasvis: hm...
[21:48] <BBB> Blaasvis: ok, well, I know Martin Soto is working on dvd playback, so... I'd wait for his commits to HEAD
[21:48] <BBB> Company: tsk, DCC me the file tonight ;)
[21:48] <Blaasvis> so your are saying try the cvs in a few day's ;)
[21:48] <Company> BBB: tell me when
[21:48] <BBB> I'll be online tonight
[21:48] <BBB> I gotta work
[21:49] <BBB> Blaasvis: I fear so
[21:49] <BBB> Blaasvis: I never tried DVD playback myself
[21:49] <BBB> I've focussed on playback of regular movies (like kazaa MPG,ASF/WMV,AVI,etc.)
[21:49] <Blaasvis> that's no way for a release with totem and gstreamer
[21:49] <Blaasvis> well they will be stuck with xine-lib for a while 
[21:50] <BBB> Blaasvis: if all those people working on the quick workarounds for using xine-ib would work on totem-gst, it'd be so much easier
[21:50] <BBB> but there's one single person working on media playback issues here
[21:50] <BBB> and that sucks
[21:50] <BBB> because I don't even want that
[21:50] <BBB> I want to work on transcoding and recording
[21:51] <Blaasvis> well it great someone does :)
[21:53] thomasvs ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[21:55] <BBB> Blaasvis: yes, but why me? ;)
[21:55] <BBB> Company: is chunklib what I said sime time ago
[21:55] <BBB> :?
[21:55] <Blaasvis> because i still do not have the skills :)
[21:56] <BBB> "we should separate parsing from interpreting"?
[21:56] <BBB> work harder ;)
[21:56] <Blaasvis> i am trying... 
[21:56] <BBB> ;')
[21:56] <BBB> don't worry :)
[21:56] <Blaasvis> first need to learn C++ for school... do i am learning gtkmm as a start 
[21:56] <Blaasvis> do == so 
[21:57] <Company> BBB: something like that
[21:57] yippi ([email protected]) left irc: "Client exiting"
[21:57] <BBB> Company: interesting... so some sort of generalized, automated version of rifflib that I wrote?
[21:58] <Company> i've never looked at rifflib
[21:59] _gst_newt_ joined #gstreamer.
[21:59] Action: BBB gets so frustrated from everyone here just doing something nonsensical
[22:00] <Company> nonsensical?
[22:00] <Uraeus> BBB: well we could be doing something malecious so it could be worse :)
[22:01] <ds-work> BBB: that would be GStreamer being nonsensical.  We're just trying to fit in
[22:01] <BBB> Company: dolphy planning on paying people to fix opt, you making a chunklib from scratch whereas you could just as well get some ideas from rifflib to see how you can make demuxers easy, etc.
[22:02] <BBB> but nobody ever fixed quicktime-pcm, quicktime-mpeg4, all of which are so straightforward that it's a minute to fix those...
[22:02] Action: BBB frustrated
[22:02] <BBB> I should do this as a fulltime job and totem-gst would rock in a month or two
[22:02] Action: BBB very frustrated
[22:02] <ds-work> BBB: I would, but I really can't get qtdemux working at all on my home system
[22:05] <Company> rifflib does a lot of things that i don't want
[22:05] <Company> and it does a lot of other things chunklib does automagically
[22:08] earthworm ([email protected]) joined #gstreamer.
[22:09] <earthworm> ello lads, any word on freshrpms getting you latest and greatest packaged yet?
[22:09] <earthworm> s/you/your
[22:17] coqRe ([email protected]) left irc: "Leaving"
[22:18] <mathrick> Uraeus: nice article, gave me a better view on svg support's history. Few typos slipped in, though
[22:19] <earthworm> i just read that, i thought it was ok too
[22:22] <Uraeus> earthworm: thomasvs has put some rpms up on gstreamer.net (with apt repo), Matieu is supposed to make similar rpms for freshrpms, but not sure when
[22:22] <Company> muahaha
[22:22] <Uraeus> thanks guys :)
[22:22] Action: Company just searched for "wmv" on kazaa
[22:23] <Company> i wanted more testfiles
[22:23] <Company> s/kazaa/emule/
[22:23] <Company> guess what the hit with the most sources was :)
[22:23] Action: Uraeus bets Company found a bit of porn on that search
[22:23] <earthworm> yeah, i was going to use the ones on your site, but it caused dependancy hell because i used dag/freshrpms rather than fedora.us/livina
[22:23] <Company> Uraeus: Paris_Hilton_Sextape.wmv
[22:24] <Uraeus> hehe :)
[22:24] Action: Uraeus has actually seen that and it was booooring
[22:25] <ds-work> interesting how much you can screw up programs by changing or deleting gconf keys
[22:26] <ds-work> Uraeus: who wants to see grainy video of a guy having sex with a girl that looks like a little boy?
[22:26] <Company> everything is porn
[22:26] <Company> i was expecting at least some music videos
[22:26] <Company> Je Baise Mon Ex Copine Sandy Tel 0676602477.wmv
[22:27] <Company> some filenames are at least a bit funny
[22:27] <teuf> hehe
[22:27] <Uraeus> ds-work: those of use who like having sex with little boys? <g>
[22:27] Action: teuf dials the phone number
[22:29] <ds-work> Uraeus: to the connoisseur, a little boy != a girl who looks like a little boy
[22:30] <Uraeus> I saw 10 minutes of that farm/reality-show thing with Paris Hilton, think being around a girl like that could potentially turn me into a woman beater
[22:34] earthworm ([email protected]) left #gstreamer (""I allege that SCO is full of it." - Linus Torvalds").
[22:39] thomasvs ([email protected]) joined #gstreamer.
[22:45] <dilinger> i'm suprised i don't see more bug reports along the lines of "Paris_Hilton_Sextape.wmv crashes <video player>!"
[22:46] Action: mathrick wonders why he doesn't know who is Paris Hamilton
[22:47] Action: Company wonders where mathrick lives
[22:47] <Uraeus> mathrick: she is the heir to the founder of the Hilton hotel chain (the founder was actually a norwegian who emigrated to the US afaik)
[22:48] <Company> and she's busy partying
[22:50] <Uraeus> actually her grandpa founded it, not her father afaik
[22:52] Action: ds-work idly dreams of a future where all GStreamer elements are nice state machines
[22:53] <Company> people could start by writing them that way
[22:53] <Company> or alternatively writing a testsuite that breaks them
[22:56] <ds-work> I've been thinking about wrapping loop/chain/get functions in a state machine function
[22:56] <Company> asf is not supposed to have A/V sync, is it
[22:56] <Company> ?
[22:56] <ds-work> probably "iterate"
[22:57] <Company> write a scheduler that does that
[22:57] <ds-work> where the iterate function provides hints to the scheduler
[22:57] <ds-work> like, "wait for time X" or "wait for buffer on pad sink"
[22:58] thaytan ([email protected]) left irc: No route to host
[22:58] <ds-work> loop functions are harder to wrap, but still possible
[22:59] <ds-work> ideally, we'd start converting loop functions to iterate functions
[22:59] walters ([email protected]) left irc: Client Quit
[22:59] <ds-work> I basically am looking for methods to decrease the number of stack contexts required to schedule elements
[23:04] <ds-work> ok, I need to log out, because GNOME is going seriously wacky
[23:04] <ds-work> like, no text
[23:05] ds-work ([email protected]) left irc: "Leaving"
[23:07] ds-work ([email protected]) joined #gstreamer.
[23:08] <BBB> Company: asf works with a/v in sync using basicgthread
[23:10] arana ([email protected]) joined #gstreamer.
[23:10] <BBB> and btw, paris hilton looks good... who compared her to a little boy?
[23:10] <BBB> (not that I've seen *that* specific video (yet! ;), but from all magazines that I've seen her in)
[23:11] kwm ([email protected]) left irc: "Leaving"
[23:14] <ds-work> she has the body of a little boy -- no shape.  Compare to, say, Britney Spears
[23:14] <markey> huhu BBB 
[23:16] spyder482 ([email protected]) left irc: Read error: 113 (No route to host)
[23:20] <ds-work> iain: do you know anything similar to gegl (gegl.org) for audio?  Might be nice for Marlin
[23:21] Blaasvis ([email protected]) left irc: "Leaving"
[23:22] ploum ([email protected]) left irc: "Leaving"
[23:23] Uraeus ([email protected]) left irc: Read error: 113 (No route to host)
[23:28] <taaz> any arts people here?
[23:28] <taaz> i don't understand how this arts test is supposed to work
[23:29] <taaz> uses artsc-config to get -I paths but then includes something from libarts1-dev debian package in another dir
[23:29] <iain> ds-work: what is gegl? I never really understood it
[23:30] <ds-work> iain: it's a framework for applying filters to static images
[23:31] <ds-work> I've written a similar framework for scientific data here at work, but I'm leaving
[23:31] <ds-work> soon
[23:31] lupusBE ([email protected]) left irc: "Bezig met verlaten"
[23:32] <ds-work> so I'm trying to figure out how to float this project as an open source project
[23:33] <iain> hmm, sort of like you have pipeline elements?
[23:34] <ds-work> well, sort of
[23:34] <ds-work> except the pipeline doesn't "flow"
[23:34] <iain> sounds sort of like beast?
[23:34] <ds-work> if I change a parameter on one filter, it triggers a cascade that eventually causes a plot display to update, for example
[23:35] <iain> ok, I think I understand...
[23:36] Action: ds-work thinks he's a little disgusted by the lyric "Shoot some love into my vein"
[23:36] <iain> like effects that dont affect the original data
[23:36] <ds-work> iain: right
[23:37] <ds-work> I'm thinking of it mainly for visualization, really
[23:37] <BBB> ds-work: you're kidding me? her face is brilliant
[23:37] <ds-work> for example, a downsampler that takes N seconds of data and downsamples it to a number of points that are reasonable to plot
[23:37] <BBB> her eyes compare to annemiek!
[23:38] <BBB> and she's blond
[23:38] <BBB> ;)
[23:38] <ds-work> BBB: bleached
[23:38] <BBB> who cares
[23:38] <BBB> it's blond
[23:38] <iain> I dont think shes all that nice either
[23:38] <BBB> you guys just don't have any taste ;)
[23:39] <iain> must be
[23:39] Action: ds-work has the taste to prefer women that are visually different from little boys
[23:39] Action: ds-work ducks
[23:39] <BBB> d'oh
[23:40] <BBB> ohwell... I guess I should see the video to know how her body compares to her face
[23:40] <BBB> you guys are too obvious :-p
[23:41] <mathrick> BBB: blonde is boring
[23:41] <ds-work> rather strange, really, since I seem to have an affinity for women that look like butch lesbians
[23:41] <mathrick> now, red is _something_ :)
[23:42] teuf ([email protected]) left irc: Remote closed the connection
[23:42] <BBB> pff...
[23:42] <BBB> blond is nice :)
[23:43] <BBB> anyway
[23:43] <BBB> bbl
[23:44] Action: BBB will do some "media testing" tonight
[23:44] BBB ([email protected]) left #gstreamer.
[23:48] Action: Misirlou thinks dirtyblondred is interesting
[23:48] yippi ([email protected]) joined #gstreamer.
[23:51] Action: mathrick thinks copper-like red, with fiery shade in it is uber sweet
[23:58] joem ([email protected]) left irc: Read error: 113 (No route to host)
[23:58] <Company> shit
[23:58] <Company> i missed an interesting discussion
[23:59] Action: Company goes looking up pictures of miss hilton
[00:00] --- Wed Mar 24 2004
[00:03] Action: mathrick stares at her pics, but fails miserably to find any way how someone could possibly mistake her for a little boy
[00:08] Action: mathrick suspects his Galeon is posessed
[00:10] <mathrick> 1 of 3 times, it would crash when _closing_ these annoying XXX popups
[00:11] <Company> uh, why does it even open them?
[00:11] <mathrick> because I don't have any blockers set, just "open everything in background tab"
[00:12] <Misirlou> Britney Spears > Paris Hilton
[00:12] <mathrick> that usually works fine, unless it crashes when closing them, that is -_-'
[00:13] <mathrick> Misirlou: in what unit? ccm ?
[00:13] <Company> heh
[00:13] <Company> usefullness for computer building probably
[00:14] <Company> britney has more silicon
[00:16] <mathrick> the best part is, that when some popup crashes, it just crashes. And on session-recovery, it will open it again, and crash when closing, and so on... seems like solution is to split the tab into separate window, close it, click "Close" on gnome crash dialog, start again and it will be gone
[00:16] <Misirlou> mathrick: Yeah, m³. :)
[00:17] Rotty ([email protected]) left irc: Remote closed the connection
[00:17] <Misirlou> Company: I'm not sure that silocon == silocone.
[00:18] <ds-work> or even silicone
[00:18] <Misirlou> 'tis what I meant :)
[00:18] <mathrick> sillycone :)
[00:19] <Company> though britney isn'T that female looking either (if you count her pre-knee-operation)
[00:19] <Misirlou> ????
[00:20] <Company> though i'm not a J.Lo fan
[00:20] <Company> you could probably mistake her for a boy if you removed her implants and cut her hair
[00:20] <mathrick> britney isn't female-like at all
[00:20] <Misirlou> Silly furriners. :P
[00:21] Action: mathrick thinks talking about britney at all is bad symptome
[00:21] <ds-work> at least Britney has hips
[00:22] <ds-work> mathrick: you think?
[00:22] <Company> well
[00:22] <Company> i guess you can count that as hips
[00:23] <mathrick> ds-work: yep
[00:28] Rotty ([email protected]) joined #gstreamer.
[00:40] thaytoo ([email protected]) joined #gstreamer.
[01:03] crazney__ ([email protected]) joined #gstreamer.
[01:03] smoke ([email protected]) got netsplit.
[01:03] md` ([email protected]) left irc: Read error: 60 (Operation timed out)
[01:04] md` ([email protected]) joined #gstreamer.
[01:06] Rotty ([email protected]) left irc: Remote closed the connection
[01:08] Rotty ([email protected]) joined #gstreamer.
[01:09] smoke ([email protected]) got lost in the net-split.
[01:09] BOFH ([email protected]) left #gstreamer ("Leaving").
[01:15] crazney_ ([email protected]) left irc: Connection timed out
[01:29] Rotty ([email protected]) left irc: Remote closed the connection
[01:30] Rotty ([email protected]) joined #gstreamer.
[01:31] <Company> sometimes i just hate glib
[01:32] Action: Company adds a note that gst_element_class_get_pad_template should not be used in _init functions
[01:38] <mathrick> g'night
[01:39] Nick change: mathrick -> mathrick|sleep
[01:39] spyder__ ([email protected]) joined #gstreamer.
[01:40] <dilinger> Company: i find singling out one of the glib team and yelling at them helps
[01:43] <iain> what signal is emitted after each iteration of a bin?
[01:44] <iain> I'm currently using the handoff signal, but thats not right
[01:48] <Company> iain: iterate?
[01:48] <Company> shouldn't be that hard to use gst-inspect bin
[01:49] <iain> ohhh, I didn't know you could do that :)
[01:49] <iain> what is the gboolean return value signify?
[01:50] sublett ([email protected]) left irc: "Why are all the really good songs short?"
[01:51] <Company> you can gst-insect every GstElement
[01:51] <Company> if not, file a bug ;)
[01:51] <Company> the return value tells you if something happened
[01:51] <Company> it's what's returned from gst_bin_iterate
[01:51] <iain> no
[01:51] <iain> the return value is from the callback
[01:52] <Company> it tells if something happened ;)
[01:52] <iain> so it iterates, I get a signal
[01:52] <iain> and I'm supposed to return something
[01:52] <Company> if you return TRUE
[01:52] <Company> _iterate will always return TRUE
[01:53] <iain> I'll worry about it tomorrow
[01:53] <iain> g'night
[01:53] iain ([email protected]) left irc: "!"
[01:53] <Company> just return FALSE
[01:53] <Company> gnight
[01:53] <Company> ds-work: did i mention i'm thinking about killing gst_bin_iterate?
[01:56] arana ([email protected]) left irc: Read error: 110 (Connection timed out)
[01:56] yippi ([email protected]) left irc: "Client exiting"
[02:10] <ds-work> Company: why get rid of gst_bin_iterate?
[02:12] <Company> ds-work: i think the concept of iterating is confusing to the outside
[02:12] <Company> ds-work: i want to force usage of mainloops instead
[02:13] <ds-work> yeah, probably
[02:13] <ds-work> like gtk
[02:13] <Company> yeah
[02:14] <Company> threads already do this
[02:14] <ds-work> I was thinking in the sense of bins not being in charge of scheduling
[02:14] <Company> yeah, that's another point
[02:14] <ds-work> or rather, not all bins
[02:14] <ds-work> just top-level bins/pipelines
[02:14] <Company> let the scheduler decide if and when it wants to be scheduled again
[02:15] <Company> and allow elements to register main loop callbacks (mixers or ximagesink for handling X events)
[02:15] <ds-work> that's part of what I was thinking of with element iteration functions
[02:16] <Company> i'm not sure if i want to allow elements to register such callbacks though
[02:16] <ds-work> main loop callbacks, or scheduler callbacks
[02:16] <ds-work> I think it must be done through the scheduler
[02:16] <Company> main loop callbacks via the scheduler
[02:17] <ds-work> right
[02:17] <Company> we need a mainloop when we want to support async callbacks in clocks
[02:18] <ds-work> we need element iteration if we want single-threaded playback
[02:18] <Company> what do you mean by "element iteration"?
[02:18] <ds-work> the good thing is that our concepts are very similar (i think)
[02:19] <ds-work> basically, like a loop function, but with more restrictions and gives the scheduler hints at the end of ever function
[02:22] <ds-work> ervery
[02:22] <Company> heh
[02:22] <ds-work> er, nm
[02:24] <ds-work> for the ximagesink case, you could tell the scheduler that you are waiting on the Xlib file descriptor, for example
[02:24] <Company> an interesting thing for scheduling would be to flag pads as "can take input" or "can give output" and not use _push or _pull at all
[02:24] <ds-work> and get woken up every time an X event comes in
[02:24] <ds-work> huh?
[02:25] <Company> make all elements work like queue
[02:26] <Company> but instead of blocking when you can't take more input you just tell it to the scheduler
[02:26] <ds-work> I've thought about that, too
[02:26] <ds-work> if we had a deeper class structure, it would be easier to implement
[02:27] <ds-work> do it once in GstAudioFilter, etc.
[02:27] <Company> we could support elements that do that and rewrite everything slowly to that model
[02:28] <Company> would be nice to visualise in gst-editor and step through it ;)
[02:28] <ds-work> same for getting rid of get/chain/loop
[02:28] <ds-work> btw, why don't we use the element mutex for every call to an element?
[02:29] <Company> because we're supposed to be at least a bit fast?
[02:29] <Company> and because it leads to nasty deadlocks when you try to lock the second time from the same thread
[02:30] <ds-work> it makes more sense in a model where elements don't call back to the core and get lost
[02:31] <Company> yeah
[02:31] <ds-work> we are really disgusting in that respect.  In a loop/chain function, you can call the core, and the core goes off and does _completely unrelated_ stuff
[02:33] <Company> yeah
[02:33] <Company> that was the idea
[02:33] <Company> same as a process reading and writing to a pipe
[02:33] <ds-work> I think gdb would be kinder to us
[02:34] <Company> the core (called kernel in the other model) goes of and does other unrelated stuff ;)
[02:35] <ds-work> we do it poorly
[02:35] <ds-work> if we really want that model, we should do one thread per element
[02:35] <Company> i don't want that model
[02:35] <Company> 1 thread per element is a hell to debug
[02:37] <Company> the "signalling model" (where pads tell the scheduler if they are ready) would allow multiple threads in one scheduler
[02:39] <Company> but doing that model would be hell to do before 1.0
[02:39] <Company> we'd piss off everyone because we'd need to rewrite all elements _again_
[02:40] <ds-work> maybe
[02:41] <ds-work> perhaps only loop elements
[02:41] <ds-work> get/chain elements are easy to wrap in any model
[02:41] <Company> not really
[02:42] <Company> an element that takes 1 buffer and outputs 5 would not be easy to wrap in my model, because you can only signal "i'm ready" once
[02:44] <ds-work> true
[02:44] <ds-work> loop elements + wacky chain elements
[02:44] BBB ([email protected]) joined #gstreamer.
[02:44] <ds-work> I already consider those broken, though
[02:45] Action: BBB aves
[02:45] <BBB> +w
[02:45] <Company> i think it'd be possible to keep the current model
[02:45] <Company> if we get a non-broken cothread implementation
[02:45] <BBB> gthread? ;)
[02:45] <Company> and then gradually move to a better model
[02:45] <BBB> nptl is pretty cool
[02:46] <Company> BBB: want the wmv?
[02:46] <BBB> yes please
[02:46] <Company> it's 11MB
[02:46] <BBB> put it online and I'll download
[02:46] <BBB> or maybe DCC works
[02:46] <BBB> that's fine
[02:46] <Company> i'll try dcc (considering i still don't have webspace)
[02:47] <Company> tell me if it works
[02:48] <BBB> --- DCC RECV connect attempt to Company failed (err=Connection refused).
[02:48] <BBB> :()
[02:48] <BBB> try scp'ing it to gstreamer.fdo.net/media/
[02:52] <foser> Company: send it to me and ill put it up
[02:52] <BBB> please do
[02:54] Action: BBB is watiign for more test movies for his collection
[02:54] <BBB> specifically the hilton one ;)
[02:55] <foser> you can get that one everywhere.. 
[02:55] <BBB> I'm not gonna do any actual effort ;)
[02:55] <Company> it's uploading to gstreamer.freedesktop.org/media/large
[02:55] <foser> typing a line in gift is not mucho effort


-------------------------------------------------------
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