IRC Logs

*******************************************************************
[03:06] <taaz> what did you want to do with it?
[03:06] <sethn> Looking for the path of least resistance to get data from the soundcard and into the frequency domain
[03:07] <taaz> but why?  visualization?  eq?
[03:08] <sethn> ah, I'm building a number of throwaway experiments for doing computer music stuff
[03:08] camh_ ([email protected]) joined #gstreamer.
[03:08] <sethn> Just looking for a simple setup that I can use to hack out some ideas in python
[03:09] <taaz> the short answer is that i don't think anyone has really done much with freq domain data
[03:10] <sethn> I can imagine it being a useful plugin for doing visualization stuff though, now that you mention it
[03:10] Action: sethn nods
[03:10] camh ([email protected]) left irc: Read error: 110 (Connection timed out)
[03:10] <taaz> it needs some thought to get a proper description of the data so plugins can be hooked together in some sane way
[03:11] <taaz> maybe gst-sci has some stuff for this?
[03:11] Action: taaz looks at ds
[03:11] <sethn> afaik there isnt' any sort of standard format for storing frequency data
[03:11] <sethn> there's also a question of "how good" a ft you want
[03:11] <taaz> somehow plugins have to agree on data passed between them
[03:12] <taaz> the python part might be hard too
[03:12] <sethn> Something that you could build an eq atop would be a much more difficult task
[03:12] <taaz> the current code is kind of crap.  it needs to be updated to support doing data processing with Numeric
[03:13] <sethn> ah, I hadn't looked closely at the python bindings yet
[03:13] <sethn> I mean, in principle I can just do the transform in python
[03:21] sethn ([email protected]) left irc: "Leaving"
[03:23] seth ([email protected]) joined #gstreamer.
[03:24] Nick change: seth -> sethn
[03:36] aeyakovenko ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[03:47] <sethn> taaz: I take it from your comment that there's no convenient way to get an array of values in python out of a gst pipeline today without writing C code?
[03:55] BBB ([email protected]) joined #gstreamer.
[04:07] <taaz> eh... there sort of is
[04:08] <taaz> i wrote a simple version long ago.  i think it still works
[04:08] <sethn> got it still?
[04:08] <taaz> but i think it does alot of buffer copying
[04:08] <sethn> eh, I could live with that ;-)
[04:09] <taaz> it's in gst-python/examples/gst/
[04:10] <sethn> taaz: the identity filter?
[04:10] <taaz> rot13.py is a better example
[04:10] <taaz> gives a string of bytes.  very lame.  sorry ;)
[04:11] <sethn> heh
[04:11] <taaz> the proper thing to do would be to add support for a Numeric array backed by the raw buffer
[04:11] <taaz> so then you could just call whatever fancy methods you want
[04:12] <sethn> taaz: yeah
[04:12] <sethn> taaz: I was thinking about it, then I felt lazy again
[04:12] <sethn> :-)
[04:12] <taaz> if you intend to do lots of math you should realize that python is single threaded so you won't get any speedup on parallel processing machines (ie, SMP)
[04:12] <sethn> yeah, I'm accutely aware of python's threading shortcomings :-P
[04:13] <taaz> you probably also don't want to do long processing in python since that would also lock up the interpreter which might be needed to complete other tasks
[04:13] <sethn> I'm just writing a dozen or so throwaway experiments, so its not a big deal
[04:13] <taaz> make sure you don't bring this up in front of wingo.  he'll laugh and talk about how this isn't a problem in his guile bindings ;)
[04:13] <sethn> ha ha ha
[04:14] <sethn> taaz: the problem with his guile bindings is... well...
[04:14] <sethn> taaz: ...guile
[04:14] <sethn> ;-)
[04:15] <taaz> functional programming languages are one of the few programming styles i've never quite groked
[04:15] <taaz> i should try again.  i'm sure i'd like it if i could get into it.
[04:18] <sethn> Guile is a disfunctional programming language :_)
[04:19] <sethn> Actually, I mostly just like to abuse it
[04:19] <sethn> taaz: I feel like if you are going to write in a functional language, go all the way and use Haskell
[04:20] <sethn> I've actually found haskell useful in places, whereas I've rarely felt a desire to excercise, say, scheme
[04:22] <taaz> i'm not familiar enough with any fp languages to compare them
[04:23] <taaz> but i'm pretty sure all are better than c++ or perl ;)
[04:23] <sethn> eh, better is as better does
[04:24] <sethn> http://haskell.org/aboutHaskell.html
[04:24] <sethn> See the quicksort implementation down on the page ;-)
[04:26] <taaz> typical my-language-is-better-than-yours info
[04:27] <sethn> yes indeedy
[04:27] <taaz> those code snippets are obviously not comparable
[04:27] <taaz> if you tried to make the C version more functional you could
[04:29] <sethn> I feel like the code snippets pretty fairly illustrate both a straightforward C implementation and a straightforward Haskell implementation
[04:33] <taaz> maybe.  need to look at efficiency and so on too
[04:33] <sethn> see, that's the thing
[04:33] <sethn> Haskell isn't necessarily the most efficient way to implement it
[04:34] <sethn> Undoubtedly the C version will run faster
[04:34] <taaz> so when can we expect the Haskell Gst bindings? ;)
[04:34] <sethn> Some poor fuck did GTK haskell bindings
[04:34] <sethn> Actually, gst wouldn't work so badly....
[04:36] <sethn> But in general, the more functional the language (and haskell is sort of toward an extreme) the trickier it is to deal with I/O
[04:56] jcsston ([email protected]) joined #gstreamer.
[05:15] BBB_ ([email protected]) joined #gstreamer.
[05:15] sethn ([email protected]) left irc: "Leaving"
[05:16] Action: BBB_ waves
[05:16] <BBB_> oops
[05:16] <BBB_> forgot to turn off the other
[05:16] BBB ([email protected]) left irc: Nick collision from services.
[05:17] Nick change: BBB_ -> BBB
[05:24] <taaz> eh... why isn't 0.7.2 src online?
[05:38] <taaz> i guess it doesn't exist or something... hmm
[05:51] spyder482 ([email protected]) left irc: "Leaving"
[06:09] Misirlou ([email protected]) got netsplit.
[06:10] Misirlou ([email protected]) joined #gstreamer.
[06:22] Marsupilami23 ([email protected]) left irc: "I don't wanna grow up, I'm a Toys R Us Kid. There's a million toys at Toys R Us that I can play with!"
[06:27] steveb__ ([email protected]) joined #gstreamer.
[06:34] BBB ([email protected]) left irc: "Client exiting"
[06:44] mathrick ([email protected]) left irc: Read error: 60 (Operation timed out)
[06:44] steveb_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[06:45] steveb_ ([email protected]) joined #gstreamer.
[06:50] jcsston` ([email protected]) joined #gstreamer.
[06:51] mathrick ([email protected]) joined #gstreamer.
[06:53] steve_b ([email protected]) joined #gstreamer.
[06:54] jcsston ([email protected]) left irc: Nick collision from services.
[06:54] Nick change: jcsston` -> jcsston
[06:54] steveb__ ([email protected]) left irc: Read error: 60 (Operation timed out)
[06:59] steveb_ ([email protected]) left irc: Read error: 60 (Operation timed out)
[07:07] steve_b ([email protected]) left irc: Read error: 60 (Operation timed out)
[07:08] bitshifter ([email protected]) left irc: "bye"
[07:14] thaytoo ([email protected]) left irc: "hometime!"
[07:17] bitshifter ([email protected]) joined #gstreamer.
[07:25] Nick change: camh_ -> camh
[07:54] Nick change: mathrick -> mathrick|Uni
[08:01] apoc_ ([email protected]) left irc: "Leaving"
[08:02] somex1 ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[08:13] LeRoutier ([email protected]) joined #gstreamer.
[08:13] <LeRoutier> hello
[08:15] markey ([email protected]) joined #gstreamer.
[08:17] markey ([email protected]) left irc: Client Quit
[08:31] thaytan ([email protected]) left irc: "See y'all"
[08:34] thaytan ([email protected]) joined #gstreamer.
[09:30] jcsston ([email protected]) left irc: "VDubMOD watchout TCME is coming!"
[09:39] <sxpert> LeRoutier: hello
[09:44] danb_ ([email protected]) joined #gstreamer.
[09:44] danb ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[10:05] markey ([email protected]) joined #gstreamer.
[10:08] <LeRoutier> bbl
[10:08] LeRoutier ([email protected]) left irc: "Leaving"
[10:20] lilo ([email protected]) got netsplit.
[10:26] lilo ([email protected]) got lost in the net-split.
[10:31] jdahlin ([email protected]) joined #gstreamer.
[10:55] thomasvs ([email protected]) left irc: Read error: 110 (Connection timed out)
[11:14] pb_ ([email protected]) joined #gstreamer.
[11:15] alley_cat ([email protected]) left irc: Read error: 60 (Operation timed out)
[11:20] thomasvs ([email protected]) joined #gstreamer.
[11:20] wheels ([email protected]) left irc: "using sirc version 2.211+KSIRC/1.3.9"
[11:24] sublett ([email protected]) joined #gstreamer.
[11:26] alley_cat ([email protected]) joined #gstreamer.
[11:34] sweet_elise83 ([email protected]) joined #gstreamer.
[11:35] sweet_elise83 ([email protected]) left #gstreamer.
[11:43] PhotoBob ([email protected]) joined #gstreamer.
[11:52] lilo ([email protected]) joined #gstreamer.
[11:59] teuf ([email protected]) joined #gstreamer.
[12:22] smoke ([email protected]) joined #gstreamer.
[12:24] Nick change: markey -> afk
[12:26] foser ([email protected]) joined #gstreamer.
[12:31] danb_ ([email protected]) left irc: "leaving"
[12:37] danb ([email protected]) joined #gstreamer.
[13:05] nsty_belle15 ([email protected]) joined #gstreamer.
[13:06] nsty_belle15 ([email protected]) left #gstreamer.
[13:11] Nick change: afk -> markey
[13:21] thomasvs ([email protected]) left irc: Read error: 110 (Connection timed out)
[13:23] jdahlin ([email protected]) left irc: Read error: 110 (Connection timed out)
[13:28] danb ([email protected]) left irc: Read error: 110 (Connection timed out)
[13:31] iain ([email protected]) joined #gstreamer.
[14:03] thomasvs ([email protected]) joined #gstreamer.
[14:03] jdahlin ([email protected]) joined #gstreamer.
[14:13] sublett ([email protected]) left irc: "Nice of you to put me on the money"
[14:24] sack ([email protected]) joined #gstreamer.
[14:24] <sack> Hi everyone :-)
[14:25] nsty_rose31 ([email protected]) joined #gstreamer.
[14:26] nsty_rose31 ([email protected]) left #gstreamer.
[14:35] sublett ([email protected]) joined #gstreamer.
[14:39] sjoerd ([email protected]) got netsplit.
[14:39] sublett ([email protected]) got netsplit.
[14:39] bitshifter ([email protected]) got netsplit.
[14:39] mathrick|Uni ([email protected]) got netsplit.
[14:39] ds ([email protected]) got netsplit.
[14:41] sublett ([email protected]) returned to #gstreamer.
[14:41] bitshifter ([email protected]) returned to #gstreamer.
[14:41] mathrick|Uni ([email protected]) returned to #gstreamer.
[14:41] sjoerd ([email protected]) returned to #gstreamer.
[14:41] ds ([email protected]) returned to #gstreamer.
[14:41] pb_ ([email protected]) left irc: "bbl"
[14:44] mathrick|Uni ([email protected]) left irc: Ping timeout: 14400 seconds
[14:51] <alley_cat> hmm, i though basicomega scheduler is supposed to work on non i686 glibc ...
[14:56] kmaraas ([email protected]) joined #gstreamer.
[15:02] Action: sxpert_work has his sound player running under valgrind..... lessee
[15:19] bforbes ([email protected]) left irc: "*sigh* I don't want to come up with a quit message..."
[15:37] PhotoBob ([email protected]) left irc: "ChatZilla 0.9.52B [Mozilla rv:1.6/20040116]"
[15:50] mathrick ([email protected]) joined #gstreamer.
[16:02] <sxpert_work> mathrick: any idea how to get rhythmbox to update the library on startup ?
[16:03] <mathrick> sxpert_work: i think it does that automagically, at least it's supposed to IIRC
[16:03] <sxpert_work> ah... (doesn't)
[16:06] Action: sxpert_work forgets about rhb, it's still segfault central
[16:15] <teuf> if you added new files to your library, you need to tell it explicitly to reload it
[16:15] Nick change: markey -> markey|afk
[16:16] <sxpert_work> teuf: 
[16:17] <sxpert_work> oops
[16:17] <sxpert_work> teuf: itunes doesn't need that...
[16:18] <teuf> sxpert_work: so what? itunes also does tag editing and cd writing
[16:18] <teuf> that's not like rhythmbox was declared 1.0 material with a complete featureset
[16:18] <sxpert_work> teuf: and coffee :D
[16:19] <sxpert_work> teuf: just saying :D
[16:32] yippi ([email protected]) joined #gstreamer.
[16:46] <mathrick> bbl
[16:46] mathrick ([email protected]) left irc: "Client exiting"
[16:59] kwm ([email protected]) joined #gstreamer.
[17:08] Nick change: markey|afk -> markey
[17:10] Action: sxpert_work suspects spider from leaking it's insides when removed from a pipeline
[17:24] kmaraas ([email protected]) left irc: Client Quit
[17:28] virus ([email protected]) joined #gstreamer.
[17:30] <virus> I wonder wether GStreamer is useable right now for building solid applications.
[17:31] <virus> I'm looking for some nice sound backend for my Voice-over-IP project and while having tried to work my way tru PortAudio and other crappy audio libraries I finaly found my way back to the godfather of media pipelines...
[17:33] <virus> I'm asking this dumb sounding question as I remember a lot gstreamer applications just crashing in random manner in the past.
[17:43] <thomasvs> yeeha
[17:43] <thomasvs> I think I found a nasty hack in oss that I can drop
[17:43] Action: thomasvs wishes someone qualified was around to verify though
[17:43] Nick change: iain -> a-doctor
[17:44] <a-doctor> As a doctor I am eminently qualified to verify anything at the right price
[17:44] Nick change: a-doctor -> iain
[17:50] <thomasvs> heh :)
[17:50] <thomasvs> anyways, seems to be correct to me
[17:50] <thomasvs> and, playing back mpeg seems to want to fixate to 1000 hz
[17:50] <thomasvs> strangeness
[17:50] <thomasvs> might explain why mpeg is playing back too slow
[18:12] mxpxpod ([email protected]) joined #gstreamer.
[18:24] <thomasvs> yikes
[18:24] Action: thomasvs figures out why it fixates to 1000 Hz, but has NO idea how to fix it
[18:24] <thomasvs> shouldn't fixate functions be proxied somehow ?
[18:24] BBB ([email protected]) joined #gstreamer.
[18:25] <BBB> thomasvs: could you delay gst-ffmpeg until tomorrow?
[18:25] <teuf> BBB: totem + gst worked fine for me yesterday ;)
[18:25] <thomasvs> BBB: all I'm going to do to ffmpeg is rerelease the 0.7.1
[18:25] <BBB> nice :)
[18:25] <thomasvs> BBB: I had just come to the same fix for oss as you yesterday
[18:25] <thomasvs> BBB: too bad I lost two hours figuring it out :)
[18:25] <BBB> thomasvs: heh... there's some blocker bug reports first ;)
[18:25] <thomasvs> BBB: release-device is a bit harder than I thought it would be
[18:25] <BBB> d'oh :D
[18:26] <thomasvs> BBB: as for mpeg playback
[18:26] <BBB> anyway, I'm happy I finally fixed it
[18:26] <thomasvs> yeah, me too
[18:26] <thomasvs> that guint cast was really dumb
[18:26] <BBB> how about libgspltay (the blocker bug from my email)?
[18:26] <BBB> yup
[18:26] <BBB> but that's C
[18:26] <thomasvs> I was looking at the values in the docs, and it totally threw me 
[18:26] <BBB> invisibly stupid errors
[18:26] <thomasvs> yeah
[18:26] <BBB> :/
[18:26] <BBB> anywya
[18:26] sub_pop ([email protected]) left irc: "Client exiting"
[18:26] <BBB> let's settle for 0.8.0 soon
[18:26] <thomasvs> so
[18:26] <thomasvs> next bug
[18:26] <BBB> before you release gst-plugins 0.8.0, please please please fix switch or remove it from libgstplay
[18:26] <thomasvs> mpeg playback fixates on 1000 Hz
[18:27] <BBB> libgstplay does *not* work a single bit for me
[18:27] <thomasvs> BBB: hang on
[18:27] <BBB> ds fixed that, according to a recent commit
[18:27] <thomasvs> BBB: ok, step by step
[18:27] <BBB> I just don't see any improvement ;)
[18:27] <thomasvs> BBB: your switch problem, I don't know what problem you have.  all I know is "don't turn on vis"
[18:27] <BBB> (as in: I've noticed the same bug)
[18:27] <thomasvs> other than that, libgstplay works fine
[18:27] <thomasvs> what exactly are you seeing ?
[18:27] <BBB> 16x16 video playback
[18:28] <BBB> and if I remove switch from the libgstplay pipeline and replace it with identity, it works fine
[18:28] <BBB> and it's obvious
[18:28] <thomasvs> (note, I can call dolphy for specific questions, but not for just anything, and removing switch when it works for a lot of people might not be the best solution)
[18:28] <BBB> *look* at the switch code
[18:28] <BBB> it doesn't do capsnego
[18:28] <thomasvs> so how come it works for me and teuf ?
[18:28] <BBB> _at_ _all_
[18:28] <BBB> that's what I'm wondering, too
[18:28] <thomasvs> it just proxies caps nego
[18:28] <BBB> :p
[18:28] <thomasvs> which is what it should do
[18:28] <BBB> it fails to do that here :/
[18:28] <thomasvs> switch shouldn't do any caps nego, should it ?
[18:29] <thomasvs> it should just proxy, and use the default caps handler
[18:29] <BBB> not internally
[18:29] <BBB> but it doesn't proxy at all here
[18:29] <thomasvs> well, how come ?
[18:29] <BBB> that's what I'm not sure about
[18:29] <BBB> I just now that s/switch/identity/ in libgstplay fixes libgstplay for me
[18:29] <BBB> maybe vis. is turned on here, I don't know
[18:29] <thomasvs> turn it off
[18:29] <thomasvs> try it
[18:29] Action: BBB not at home so cannot check
[18:29] <thomasvs> heh
[18:29] <BBB> ;)
[18:29] Action: thomasvs turns on vis to chekc
[18:30] <BBB> it'll take me another half a year before I can do realtime debugging with you guys again
[18:30] <BBB> sorry :/
[18:30] <BBB> for you: turn it on, and restart player
[18:30] <BBB> and see if it gives a 16x16 window
[18:30] <BBB> oh, nm, /me late ;)
[18:30] <thomasvs> hm, not sure how vis is turned on actually :)
[18:30] <thomasvs> BBB: how long are you online for right now ?
[18:31] <thomasvs> BBB: btw, are you using xvimagesink or ximagesink ?
[18:31] <BBB> few more minutes
[18:32] <BBB> try in gconf-editor
[18:32] <thomasvs> ok, so vis works fine, that's not the issue
[18:32] <BBB> hm
[18:32] <thomasvs> BBB: afaik you're the only one having this problem btw
[18:32] <BBB> I'm really confused then
[18:32] <thomasvs> BBB: but, other question
[18:32] <thomasvs> libgstplay has this problem where for mpeg, it fixates to 1000 Hz
[18:32] <thomasvs> because the core fixates
[18:32] <BBB> yes
[18:32] <teuf> BBB: I'm not sure which codec totem used for video decoding if that matters
[18:32] <teuf> the video quality was bad
[18:32] <teuf> (it was a divx)
[18:32] <thomasvs> between audioconvert and tee, they both don't have a fixate function
[18:32] <BBB> teuf: "bad"? define "bad"
[18:33] <thomasvs> BBB: should fixate functions be proxied for elements that don't affect fixating ?
[18:33] <BBB> I proposed that to company
[18:33] <BBB> but he saw issues if we do that
[18:33] <teuf> some block artifacts appear when people move in the movie
[18:33] <thomasvs> BBB: ie, how are either tee or audioconvert able to decide what to fixate to ?
[18:33] <teuf> it doesn't happen with mplayer afair
[18:33] <BBB> I personally feel that only src/sink *elements* (so end/start of pipeline) should be allowed to fixate
[18:33] <BBB> but that's probably something for 0.9.x
[18:34] <thomasvs> well, it's hugely annoying right now
[18:34] <thomasvs> it settles on 1000 Hz :)
[18:34] <BBB> i KNOW
[18:34] <BBB> oops
[18:34] <BBB> caps
[18:34] <BBB> sorry
[18:34] Action: thomasvs wonders why this is for MPEG and not for AVI
[18:34] Action: BBB kicks stupid keyboard
[18:34] <thomasvs> there MUST be something about that
[18:34] <BBB> I don't really know why...
[18:34] Action: thomasvs starts comparing debug logs
[18:34] <BBB> look at pipelines
[18:34] <thomasvs> anyways, this fixating sucks atm :)
[18:34] <BBB> look at which elements it uses
[18:35] <thomasvs> do you remember where dolphy put his image of the pipeline ?
[18:35] <BBB> yes, and I told you when we decided that we'd use it for 0.8.0 that it'sd give us some issues sooner or later
[18:35] <BBB> so we'll have to live with it and fix them as well as possible
[18:35] <thomasvs> BBB: anyway, when I make the core fixate to 44100, I still have slowed down playback
[18:35] <thomasvs> BBB: but again, only for MPEG
[18:35] <BBB> make sure you declare an API/ABI freeze, but don't split off a 0.8.x branch yet
[18:35] Action: thomasvs is really confused what is going on
[18:35] <BBB> I want to keep ds/company hacking on it
[18:35] <thomasvs> BBB: why not split off a branch ?
[18:35] <thomasvs> ah, yeah
[18:35] <thomasvs> but I am splitting of a release branch
[18:35] <BBB> MPEG might have different caps then avi
[18:36] <BBB> avi uses audio/mpeg,mpegversion=1,layer=2
[18:36] Action: thomasvs wonders how he could make tee fixate
[18:36] <BBB> maybe mpeg uses something different
[18:36] <BBB> or layer=3
[18:36] <thomasvs> why would that matter for the fixation ?
[18:36] <BBB> dunno... 
[18:37] <BBB> ion mpegdemux' audio_%02d caps template, remove the layer properties
[18:37] <BBB> and see if that fixes anything
[18:37] <BBB> I'm currently just guessing
[18:37] <thomasvs> that makes me feel really confident :)
[18:37] Action: thomasvs wonders if he can release as we are right now
[18:38] <BBB> I'm online tonight, until very late
[18:38] <BBB> make sure you're at work EARLY tomorrow and I can help discussing while it's night here
[18:38] <BBB> if you want
[18:38] <teuf> thomasvs: can't you start with prereleases to find out which issues are still there ?
[18:38] <thomasvs> how early ?
[18:38] <thomasvs> teuf: yeah, planning to do another 0.7
[18:38] <BBB> 1am, 2am, 3am here
[18:38] <BBB> so 7/8/9am for you
[18:38] <thomasvs> I can't get here that early
[18:38] <thomasvs> I can send mails though
[18:38] <BBB> can you be at work at 8am for once?
[18:38] <BBB> hm
[18:38] <BBB> at home is fine
[18:38] <thomasvs> BBB: spain doesn't work that way :)
[18:38] <BBB> can you get up at 8am?
[18:38] <thomasvs> I'll try
[18:38] <jdahlin> thomasvs: isn't WTC open at 8?
[18:39] <thomasvs> should work
[18:39] <jdahlin> that means going home at 5
[18:39] <thomasvs> jdahlin: dude, have you ever gotten up before 8 here ?
[18:39] <BBB> 2am is really late here, too
[18:39] <jdahlin> thomasvs: no :)
[18:39] <thomasvs> this country makes it impossible :)
[18:39] <BBB> don't forget I gotta be at the university at 9am
[18:39] <BBB> so I need a bit of sleep in between
[18:39] teuf ([email protected]) left irc: "Client exiting"
[18:40] <BBB> for tonight, I'm planning on fixing up gst-ffmpeg more
[18:40] <BBB> I hope that'll work
[18:40] <BBB> I want it to be able to playback all movies I have on my HD one day
[18:41] <BBB> btw, why wouldn't you get up at 8am?
[18:41] <BBB> here, people get up at 6am to be at memorial at 7:30am
[18:41] <BBB> !
[18:41] <BBB> now *that's* devoted students ;)
[18:41] <thomasvs> ok, removing that layer property didn't help
[18:41] <BBB> crap
[18:41] <thomasvs> seriously, we need to fix that fixating stuff
[18:41] Action: thomasvs thinks how he can make tee fixate
[18:41] <BBB> compare avidemux' capsnego audio code with mpegdemux'
[18:41] <BBB> any difference, how subtle, might be the issue
[18:42] <sack> thomasvs, WTC is open always , you can be at office before 8 if you want but you need to break the fucking security door :-D
[18:42] <BBB> :D
[18:42] Action: sxpert_work has an apparent memleak using spider
[18:42] <BBB> sack: are you working on streaming sex videos for dolphy again? ;)
[18:42] <sack> BBB : no ... 
[18:42] <BBB> :p
[18:42] <BBB> (given your v4l bug report ;) )
[18:42] <sack> BBB . i stop working with dolphy
[18:43] <thomasvs> BBB: btw are you using ximagesink or xvimagesink ?
[18:43] <jdahlin> is it possible to run a pipeline for a limited amount of time easily?
[18:43] <jdahlin> like a property on the source element or something
[18:43] <jdahlin> I_WANT_TO_DIE_AFTER_N_SECONDS
[18:43] <sack> BBB : well did you see the new vp3 streaming based codec from nullsoft ? 
[18:43] <iain> I_WANT_TO_DIE_NOW
[18:43] sub_pop ([email protected]) joined #gstreamer.
[18:44] <jdahlin> iain: is that an environment variable
[18:44] <BBB> thomasvs: ximagesink
[18:44] <BBB> sack: no
[18:44] <BBB> jdahlin: I have an element that does that
[18:44] <jdahlin> ah, num-buffers!
[18:44] <iain> jdahlin: its local to my environment only,sorry
[18:44] <jdahlin> iain: gross
[18:44] <BBB> those are called "managers"
[18:44] <sack> BBB : seems nice ;-) well just to keep in mind .
[18:45] lupusBE ([email protected]) joined #gstreamer.
[18:45] <BBB> a manager is some sort of a bin that contains a pipeline with a ghost src pad forwarded to  the output of that bin
[18:45] <thomasvs> BBB: maybe that's the problem
[18:45] <BBB> maybe
[18:45] <thomasvs> BBB: can you try xvimagesink ?
[18:45] <BBB> try it out yourself
[18:45] <BBB> no
[18:45] <BBB> xvimagesink crashes my display
[18:45] <sack> jdahlin, i'm missing your hand squeezing mine :-D
[18:45] <BBB> and I'm not at home
[18:45] <thomasvs> weird, for avi it also fixates to 1000 hz
[18:45] <BBB> you broke it!
[18:46] Action: BBB kicks thomasvs
[18:46] <BBB> ;)
[18:46] <BBB> jdahlin: I'll post the code online in the PWG sometime soon, ok?
[18:46] <BBB> I use it to emulate EOS behaviour in various elements such as v4lsrc, v4l2src, osssrc and such after a given amount of time
[18:46] <alley_cat> hi BBB, got the chance to valgrind my hangs in avidemux, no result :(
[18:47] <BBB> and I want to use it (far future) in something like gnonlin to emulate SEEK_SEGMENT (doing a SEEK and then emulate EOS after a given amount of stream time)
[18:47] <BBB> makes plugins lots easier
[18:47] <BBB> it works well
[18:47] <BBB> I just never posted it online
[18:47] <BBB> :)
[18:47] <BBB> I'll stuff it in the PWG ;)
[18:47] <BBB> alley_cat: so, any chance where it's freezing?
[18:47] <BBB> alley_cat: oh, and how did you run valgrind?
[18:48] <BBB> you need to run valrind ../gstreamer/tools/.libs/lt-gst-launch-0.7 ....
[18:48] <BBB> not the libtool wrapper scripts
[18:48] <BBB> those have no memleaks
[18:48] <BBB> ;)
[18:48] <alley_cat> i ran valgrind on the installed binary
[18:48] <BBB> oh
[18:48] <BBB> on gst-launch or gst-launch-0.7?
[18:48] <BBB> you need the second
[18:48] <BBB> the first is a wrapper script, too
[18:48] lilo ([email protected]) left irc: "Lost terminal"
[18:48] <alley_cat> it hangs in a g_malloc inside gst_avi_demux_stream_index
[18:48] <alley_cat> -0.7
[18:49] <BBB> hangs in g_malloc are always memory management issues
[18:49] <BBB> off
[18:49] <alley_cat> valgrind --skin=memcheck gst-launch-0.7 filesrc location="Salope.avi" ! avidemux name=demux ! fakesink
[18:49] <BBB> odd
[18:49] lilo ([email protected]) joined #gstreamer.
[18:50] <BBB> I'll try to help later tonight
[18:50] <BBB> can't do it now
[18:50] <BBB> university :/
[18:50] <BBB> (hey, I want a good grade! ;) )
[18:52] <alley_cat> oh, now i get something from valgrind
[18:54] <BBB> whee
[18:55] <alley_cat> though i also get lots of glib warnings i don't get when i run it without valgrind
[18:58] <BBB> that means you're doing really wrong memory accesses
[18:58] <BBB> :)
[18:58] <alley_cat> should i dcc you the log?
[18:59] <alley_cat> interesting is that it doesn't hang when run with valgrind :P
[19:00] <jdahlin> sack: ah, Angel!
[19:00] <jdahlin> sack: you're welcome anyday to get your hand squeezed again
[19:00] <jdahlin> :-
[19:00] <jdahlin> :-)
[19:00] Action: sxpert_work installs valgrind and hopes it work
[19:01] <sack> jdahlin, hehehe nice :-) i'm training my hand :-P
[19:01] <jdahlin> sack: why did you end here btw?
[19:01] Action: jdahlin forgot to ask
[19:02] <BBB> ehm
[19:03] <BBB> alley_cat: no, just bugzilla
[19:03] <BBB> I cannot do anything with the log here
[19:04] <sack> jdahlin, sorry i can't explain you ... it too sad 
[19:05] <thomasvs> DEBUG        GST_PADS(19731) gstpad.c(1253):gst_pad_link_negotiate: negotiating
[19:05] <thomasvs> link from pad mpeg2dec0:user_data to pad src_0:sink
[19:05] <thomasvs> DEBUG        GST_PADS(19731) gstpad.c(1082):gst_pad_link_intersect: intersecting link from mpeg2dec0:user_data to src_0:sink
[19:05] <thomasvs> OMG !
[19:05] <thomasvs> why is this user_data pad getting intersected ?
[19:05] <thomasvs> this is where the core decides it needs to be 1000 Hz
[19:05] <BBB> interesting
[19:05] <thomasvs> because the user_data is being intersected
[19:05] <thomasvs> crap crap crap
[19:05] <BBB> eh?!?
[19:05] <thomasvs> yeah
[19:05] <BBB> user_data?!?!?!?!?
[19:05] <thomasvs> yeah
[19:05] <BBB> user_data is a PAD?!?
[19:05] <thomasvs> probably for DVD :)
[19:06] <thomasvs> yeah
[19:06] Action: thomasvs removes it
[19:06] <BBB> WHO BROKE THAT! &^#@&@
[19:06] <BBB> [/troll]
[19:06] <BBB> now this makes sense ;)
[19:07] Action: BBB waits for thomasvs to fix it and tell him how all his MPEGs suddenly play fine
[19:07] Action: BBB would also love to see warcraft3 movies to play
[19:07] <thomasvs> MPEG PLAYS FINE !!!
[19:08] Action: thomasvs rocks
[19:08] Action: thomasvs rocks hard
[19:08] <thomasvs> it's slightly jerky, but I'm sure that's something else
[19:09] <BBB> "jerky"?
[19:09] <BBB> that's the protein that my colleague here is researching
[19:09] <BBB> anyway
[19:09] <BBB> cool!
[19:09] <sxpert_work> BBB: beef jerky ?
[19:09] <thomasvs> so, someone willing to test my fix ? teuf ?
[19:10] <BBB> post the patch please?
[19:10] <BBB> I know mpegdemux quite well, so I can probably give a nice theoretical comment
[19:10] <thomasvs> I'm going to commit it so I can prerelease
[19:11] <thomasvs> too bad about my oss problem though
[19:11] <jdahlin> thomasvs: what problem?
[19:11] <thomasvs> release-device
[19:11] <thomasvs> I have it locally, but there's something still missing
[19:11] <thomasvs> I think I'd better redo it now that I understand the problem better
[19:12] <BBB> thomasvs: for OSS, try some "more relaxed" values for before it drops frames or pauses
[19:12] <BBB> it currently is extremely strict w.r.t. sync
[19:12] <jdahlin> I have the same problem as richard is about to report in the mailing list
[19:12] <BBB> it could be somewhat loosened
[19:12] <jdahlin> noises and small disortion in osssink
[19:12] Action: jdahlin is about to blame BBB
[19:12] <BBB> hey
[19:12] <BBB> without my fixes, osssink doesn't work at all
[19:13] <BBB> with my fixes, at least it does what it's supposed to do
[19:13] <BBB> the fact that it worked for audio-only was mere luck
[19:13] <BBB> osssink was completely broken
[19:13] <BBB> as soon as one buffer would arrive too late (underflow/underrun), it'd pause (guint64)-1ns
[19:13] <BBB> which is a few millenia
[19:14] <BBB> that's what you guys call noise
[19:14] <sxpert_work> BBB: sounds more like silence to me :D
[19:14] <BBB> for now, it "mostly" picks up sound again after an underrun
[19:14] <jdahlin> BBB, I can hear a difference when using alsasink
[19:14] <BBB> it's not perfect, but it's about as good as I can get it
[19:14] <jdahlin> sounds cleaner
[19:14] <BBB> I can't help you with that
[19:15] <jdahlin> I don't know enough about osssink to tell why
[19:15] <alley_cat> hmm,  i think i'll recompile everything gstreamer related with my usual optimizations instead of no optimizations
[19:15] <alley_cat> that already fixed several other programs i had problems with
[19:15] <thomasvs> jdahlin: I'll have a listen in ten minutes
[19:15] <jdahlin> I'll first check if it's broken by the last commits
[19:16] <BBB> try some more relaxed sync values
[19:16] <BBB> that's my guess
[19:16] <jdahlin> doh, I removed the makefile
[19:16] Action: jdahlin reruns autogen
[19:18] <BBB> http://bugzilla.gnome.org/show_bug.cgi?id=135925
[19:18] <BBB> eh
[19:18] <BBB> crap
[19:18] <BBB> if (MAX (buftime, soundtime) - MIN (buftime, soundtime) > GST_SECOND / 10) {
[19:18] <BBB> that line in gstosssink.c
[19:18] <BBB> make that GST_SECOND / 2 or so
[19:18] <BBB> play with the values
[19:18] <BBB> 0.10 is *very little*
[19:19] <BBB> and we should use llabs() there
[19:19] pb_ (~pb@2002:5160:45ef:0:240:95ff:fe30:aa84) joined #gstreamer.
[19:21] <sack> BBB : did you ckeck again v4l stuff ? :-P
[19:23] <BBB> not yet
[19:23] <BBB> but I'm quite sure you can change with qcset to get the cam to work
[19:23] <BBB> I'll get back to that
[19:23] <BBB> first food
[19:24] <BBB> bbl!
[19:24] BBB ([email protected]) left #gstreamer.
[19:24] teuf ([email protected]) joined #gstreamer.
[19:30] <thomasvs> teuf: were you able to play back MPEG using totem yesterday ?
[19:31] <teuf> I only tried with an avi file
[19:31] <teuf> I can try mpeg
[19:31] <thomasvs> teuf: can you try without updating ?
[19:32] <teuf> do you have a sample file I should use ?
[19:32] <thomasvs> teuf: ie, try with your current code
[19:32] <teuf> yeah sure
[19:32] <thomasvs> teuf: yeah, gstreamer.freedesktop.org/media
[19:32] <thomasvs> take medium/alien.mpg
[19:33] <teuf> thomasvs: it runs properly for ~1sec, and then becomes really really slow
[19:34] <thomasvs> teuf: ok
[19:34] <thomasvs> teuf: now update mpeg2dec element code, rebuild and reinstall, then run gst-register
[19:34] <thomasvs> then try again
[19:34] <sxpert_work> any idea what "** (process:28808): WARNING **: opt: unlink elements in same group: implement me" means ???
[19:35] <jdahlin> sxpert_work: means we're all too lazy to fix opt
[19:36] <iain> s/lazy/scared/
[19:36] <teuf> thomasvs: mpeg2dec doesn't seem to be used here
[19:36] <teuf> I don't seem to have mpeg2.h on my system
[19:36] <thomasvs> teuf: how can you tell ?
[19:36] <thomasvs> teuf: ah, ok
[19:36] <thomasvs> teuf: now, update oss element
[19:36] <thomasvs> teuf: and reinstall
[19:36] <thomasvs> then try again
[19:37] <teuf> works fine for 2-3 secs, and then becomes slow again
[19:38] <thomasvs> with updated oss ?
[19:39] <teuf> yeah
[19:39] <thomasvs> hm, that's weird :)
[19:39] <teuf> but maybe the element which is used here to decode mpeg is also borked ?
[19:39] <thomasvs> is there a way you can try with mpeg2dec instead ?
[19:39] <teuf> I didn't install mpeg2dec
[19:39] <thomasvs> teuf: it's possible
[19:39] <teuf> which package does mpeg2.h come from ?
[19:39] <thomasvs> mpeg2dec-devel
[19:39] <thomasvs> what dist are you on ?
[19:40] <teuf> debian, I guess that will be libmpeg2-4-dev
[19:42] <sack> teuf , yes (libmpeg2-4-dev) at least on sid :-)
[19:43] <sack> teuf : there're also this package gstreamer-mpeg2dec 
[19:43] <teuf> sack: I need to test gstreamer from cvs ;)
[19:44] <sack> ups sorry , LOL !
[19:44] <teuf> thomasvs: still the same issue :-/
[19:45] <thomasvs> teuf: it probably didn't switch decoders though
[19:45] <thomasvs> teuf: what is your audiosink ?
[19:45] <thomasvs> teuf: can you delete all your ffmpeg decoders to test ?
[19:45] <teuf> thomasvs: I think it uses mpeg2dec since at first it crashed because it couldn't find mpeg2_accel
[19:45] <teuf> (I hand edited the Makefile to avoid rerunning configure)
[19:45] <sack> thomasvs, Do you have the pipeline to test that stuff or must be test it with totem ?
[19:46] <thomasvs> sack: it can be tested using gst-plugins/examples/gstplay/player
[19:46] <teuf> thomasvs: if I remove all plugins and rerun make install in gst-plugins, I shouldn't have any ffmpeg plugin since hey are in a separate package now, right ?
[19:46] <sack> thomasvs, ok then i gonna test it as soon as plugins are built
[19:52] <KoRnouille> hello everyone
[19:52] <KoRnouille> I have sent my bug, but I'm not sure I understand the answer :
[19:52] <KoRnouille> http://bugzilla.gnome.org/show_bug.cgi?id=136334
[19:54] <teuf> KoRnouille: ds pasted a changelog entry, that means it should fix your problemù
[19:56] <thomasvs> teuf: yeah
[19:57] <teuf> I'm doing that atm, since gstreamer didn't seme to like my rm gstffmp* in the plugin dir
[20:02] Action: thomasvs uncovers an irritating bug in libtool 1.5.2
[20:05] <thomasvs> tromey: ping ?
[20:09] <teuf> thomasvs: I get Could not link video output thread (cs and balance)
[20:10] <sack> thomasvs, me too
[20:11] <sack> libgstplay: failed initializing pipeline, error: Could not link video output thread (cs and balance)
[20:11] <thomasvs> hm
[20:11] <ds-work> moo
[20:11] <thomasvs> since you need ffcolorspace
[20:11] <thomasvs> makes sense
[20:11] <ds-work> anything interesting happen lately?
[20:11] <thomasvs> ds-work: lots of painful stuff
[20:11] <thomasvs> ds-work: check my mail on MPEG, there's a nice fixating problem for you if you feel up to it
[20:12] <sack> thomasvs, yeah could be 
[20:12] <thomasvs> ds-work: how's the girl ?
[20:12] <KoRnouille> teuf: okay, thanks
[20:12] <thomasvs> sack,teuf: do you get a nice error dialog saying so?
[20:12] <sack> dialog no , just console 
[20:12] <KoRnouille> this has nothing to do with gstreamer, but I just wanted to share the "fun"
[20:12] <KoRnouille> http://www.richsalter.btinternet.co.uk/index.html
[20:13] <teuf> thomasvs: yep, even though I can see totem main window quickly appearing and disappearing before getting the error dialog
[20:13] BBB ([email protected]) joined #gstreamer.
[20:13] <teuf> would be better not to show it at all, or to show it until the user clicks on ok
[20:13] <thomasvs> teuf: hm, interesting
[20:13] <BBB> alley_cat: see your bug report, there's some hints in there
[20:13] <teuf> but that's a detail :)
[20:13] <thomasvs> sack: you're running the cmdline player, no ?
[20:13] <sack> thomasvs, yes
[20:13] <thomasvs> teuf: it's probalby a totem bug, run_dialog vs show
[20:13] <thomasvs> sack: ok, then that's normal
[20:13] <thomasvs> so, we still don't have a yuv->rgb colorspace el
[20:13] <thomasvs> that's too bad
[20:13] jdahlin ([email protected]) left irc: "Leaving"
[20:13] <BBB> ffcolorspace
[20:14] <BBB> we do
[20:14] <sack> thomasvs, ahh i can fix that adding and recompile it
[20:14] <BBB> and I really don't see why you guys don't like ffcolorspace... it's highly optimized and works well
[20:14] <BBB> :)
[20:15] Action: BBB also mumbles something about LCS and code from him in there that is slow
[20:17] <thomasvs> BBB: you know why it's not the right solution, you really don't want me to repeat why
[20:17] <taaz> why was user_data support removed?  what does that have to do with fixating?  (not that i really know what fixating is anyway...)
[20:17] <BBB> read the logs
[20:17] <BBB> :p
[20:17] <taaz> yeah, it doesn't explain anything
[20:18] <BBB> thomasvs: as long as people don't give me a good reason to work on LCS I won't
[20:18] <BBB> like someone assuring me that novell will ship totem if we make it "legally shippable" for them
[20:18] <BBB> and assuring me that there is no other way
[20:18] <BBB> like a partial inclusion of gst-ffmpeg
[20:18] Action: BBB wants to fix gst-rec
[20:19] <taaz> it looks like you cured the symptom of some other problem.
[20:19] Action: BBB agrees with taazs
[20:19] <alley_cat> BBB: will try to fiddle around with those as soon as it finishes the current compile run
[20:20] <thomasvs> BBB: there is no way *anyone* can ship ffmpeg
[20:20] <thomasvs> BBB: even if you pay the license fees included
[20:20] <thomasvs> taaz: of course I cured the symptom of another problem
[20:20] <BBB> alley_cat: sure
[20:20] <thomasvs> taaz: feel free to join in fixing the ACTUAL problem :)
[20:20] <thomasvs> BBB: no one ever said a partial inclusion of ffmpeg wouldn't work
[20:21] <BBB> thomasvs: you're not answering my questions
[20:21] <BBB> thomasvs: you're just repeating what you said before
[20:21] <thomasvs> BBB: in fact, I've suggested ripping out the colorspace code and making one element of it
[20:21] <BBB> - why can't they ship a partial ffmpeg
[20:21] <BBB> - why don't we rip out csp conv. functions from ffmpeg
[20:21] <thomasvs> BBB: like I said
[20:21] <BBB> - will they ship player if we do either of those?
[20:21] <thomasvs> BBB: I suggested that more than once already, ripping out the code for colorspaces
[20:21] <thomasvs> BBB: I even tried myself, that's how desperate I was :)
[20:21] <BBB> so...?
[20:21] <thomasvs> BBB: well, go ahead
[20:22] <thomasvs> BBB: novell isn't going to do it
[20:22] <kwm> does somebody knows how a pipeline for cdplayer looks likes? google only turns up packages ...
[20:22] <BBB> it's the file libavcodec/imgconvert.c
[20:22] <thomasvs> BBB: novell is going to take what works the best to them
[20:22] <BBB> default compilation options
[20:22] <thomasvs> BBB: yeah, I now zilch about colorspaces
[20:22] <BBB> it needs some small changes for types
[20:22] <BBB> it depends on inttypes.h, we rather have glib.h
[20:22] <BBB> that's all
[20:22] <thomasvs> BBB: I think it's a halfhour job for you, and it would rock
[20:22] <thomasvs> BBB: I don't know why my suggestion for it got ignored every time :)
[20:22] <BBB> answer my last question
[20:22] <KoRnouille> hummmmf !
[20:22] <KoRnouille> 
[20:22] <KoRnouille> checking for libtoolize >= 1.5.0 ... found 1.4.3, not ok !
[20:22] <BBB> - will they ship player?
[20:22] <BBB> s/player/totem/
[20:22] <thomasvs> it's pretty simple.  they won't ship the player if we don't :)
[20:23] <KoRnouille> I tried installing 1.5.2, but it's apparently broken
[20:23] <BBB> i9f they don't, I don't see why I'd do it
[20:23] <BBB> d'oh
[20:23] <thomasvs> other than that, they're a company, so they will ship what works
[20:23] <BBB> youknow what?
[20:23] <KoRnouille> I still get these "../../libtool: line 4040: test: : integer expression expected" errors
[20:23] <thomasvs> BBB: i'm not novell, why are you asking me ? :)
[20:23] <BBB> I need to get into touch with those guys
[20:23] <thomasvs> KoRnouille: yeah, just checked the source for libtool, it's a bug there
[20:23] Action: BBB will go to boston one of the next few weekends
[20:23] <thomasvs> BBB: I don't see why only novell would change your opinion
[20:23] <KoRnouille> shoudl I get an other libtool version ?
[20:23] <BBB> I need a checklist what they expect from us
[20:23] <thomasvs> BBB: it's pretty simple - nobody can ship ffmpeg code, so we need to rip out the colorspace code
[20:24] <BBB> thomasvs: novell or redhat, any company which is a commercial model for selling free software
[20:24] <thomasvs> KoRnouille: yeah, 1.5
[20:24] <BBB> I don't care whichi
[20:24] <thomasvs> BBB: even gnome cannot ship ffmpeg
[20:24] <KoRnouille> okay thomasvs, thanks, I'll give it a try
[20:24] <BBB> I need the *checklist*
[20:24] <BBB> once I've got that, I'm fine with anything
[20:24] <BBB> and GNOME can, because GNOME ships source
[20:24] <thomasvs> it's pretty simple.  1) stability.  2) workingness.  3) shippability.
[20:24] <BBB> I need an *exact* checklist
[20:24] <BBB> ;)
[20:24] <thomasvs> BBB: for the last time, YOU'RE WRONG.  GNOME CANNOT SHIP IT.
[20:24] <thomasvs> BBB: you can restate that you are not a lawyer but you think they can.
[20:25] <BBB> I'll get into touch with the peepz at novell
[20:25] <thomasvs> They say they can't
[20:25] <BBB> who is they?
[20:25] <thomasvs> the gnome foundation
[20:25] <BBB> make them proof it to me and I'll take my words back
[20:25] <thomasvs> why do you have such a hard time accepting the fact
[20:25] Nick change: iain -> a-doctor
[20:25] <thomasvs> BBB: no, it works the other way around.
[20:25] <BBB> I've proven before that there's a source-only clause which allows it
[20:25] <thomasvs> if gnome says they cannot ship it, you need to tell them why they can
[20:25] <BBB> hm
[20:25] <BBB> nevermind
[20:25] <thomasvs> BBB: the ffmpeg code INFRINGES on MULTIPLE patents
[20:25] <a-doctor> I will once again remind people that as a doctor I can verify anything for the right price
[20:25] Nick change: a-doctor -> iain
[20:26] <thomasvs> BBB: the license has zero to do with it :)
[20:26] <BBB> is trow around?
[20:26] <thomasvs> iain: what's your price ?
[20:26] <BBB> then what is?
[20:26] <thomasvs> BBB: how come you don't understand that code that infringes patents is not shippable by a project such as gnome ?
[20:26] <BBB> patents?
[20:26] <BBB> hm
[20:26] <iain> a bag of peanuts
[20:26] <thomasvs> BBB: next you're going to tell me mplayer and xine are legal projects that are shippable by any distro :)
[20:26] <iain> and a 2 litre bottle of coke
[20:26] <thomasvs> hm, that's not too bad a price
[20:26] <BBB> thomasvs: in case you didn't notice, xine is shipped by mandrake in earlier versions (didn't check mdk9)
[20:27] <iain> thomasvs: I'm cheap
[20:27] <BBB> mplayer isn't legal, mplayer violates the GPL in several ways
[20:27] <BBB> plus their code is stolen from all around, including from GStreamer and mjpegtools, but that's not really a GPL violation on itself
[20:27] <ds-work> thomasvs: the girl is a bit, er, difficult
[20:27] <thomasvs> BBB: think why mdk removed it !
[20:28] <BBB> thomasvs: I just said - I didn't know *whether* they removed it
[20:28] <thomasvs> ds-work: making it hard to concentrate ?
[20:28] <BBB> so I don't know that
[20:28] <ds-work> mplayer is quite possibly 100% GPL now
[20:28] <thomasvs> BBB: they did
[20:28] <Misirlou> Eh? Debian ships Xine in main.
[20:28] <BBB> Misirlou: without lavc, without any functionality whatsoever
[20:28] <thomasvs> Misirlou: since when ?
[20:28] <thomasvs> right, heavily patched
[20:28] <BBB> debian ships a "crippled" xine
[20:28] Action: ds-work wonders why he can't meet girls that aren't broken in some way
[20:28] <thomasvs> ds-work: because there aren't any :)
[20:29] <BBB> ("crippled" is the word mplayer devs associate with xine in debian, I didn't make that up myself)
[20:29] <thomasvs> ds-work: or, there are, but they have a tendency to be boring
[20:29] Action: iain wonders why he can't meet girls that aren't already taken
[20:29] <alley_cat> BBB: commenting the gst_tag_list_free fixes it =)
[20:29] <ds-work> thomasvs: that's not true, because I broke up with one about 6 months ago
[20:29] <alley_cat> finally i can watch avi again
[20:29] <BBB> alley_cat: which of the two solved it? :)
[20:29] <BBB> ah
[20:29] <BBB> nice
[20:29] <BBB> please attacjh patch there so thomasvs can apply it
[20:29] <alley_cat> doing that
[20:29] <thomasvs> ds-work: so why did you break up ?
[20:30] <BBB> and sorry for the bug, which is apparently my fault
[20:30] <ds-work> thomasvs: because *I* was broken :)
[20:30] Action: thomasvs thinks aobout a prerelease
[20:30] <thomasvs> ds-work: well, then you know what to fix :)
[20:30] <Misirlou> thomasvs: Since 2001-09-18 it looks like.
[20:30] <thomasvs> ds-work: so what are you complaining about then ? the girls are fine, you're the problem :)
[20:31] <BBB> can anyone here play ogg/theora?
[20:31] Action: iain knows how to fix the problem too...nuke germany...
[20:31] <BBB> I want to write ogg muxer and theora encoder
[20:31] <BBB> and release gst-rec or so
[20:31] <Misirlou> BBB: Not Ogg/Theora!
[20:31] <ds-work> thomasvs: I'm mostly fixed now.  It's just too late
[20:31] virus ([email protected]) left #gstreamer ("Leaving").
[20:31] <Misirlou> BBB: Ogg Theora!
[20:31] Action: BBB is bored and will bring a ;laptop tomorrow
[20:31] LeRoutier ([email protected]) joined #gstreamer.
[20:31] <BBB> Misirlou: ?
[20:31] <Misirlou> BBB: There's no solidus. :)
[20:31] <BBB> whatta?
[20:32] <Misirlou> BBB: Anyway, it would be fantastic if you wrote a Theora encoder. :)
[20:32] <BBB> talk english dude, I'm only a poor dutchman ;)
[20:32] <LeRoutier> hello
[20:32] <Misirlou> Solidus/slash = slash.
[20:32] <Misirlou> er
[20:32] <BBB> aha
[20:32] <Misirlou> Solidus/slash = /.
[20:32] <BBB> a slash
[20:32] <thomasvs> teuf: what's your audio sink set to ?
[20:33] <BBB> thomasvs: did you try several audio pipelines?
[20:33] <BBB> and did you play with the sync value?
[20:33] <BBB> and did it help anything?
[20:33] <teuf> thomasvs: osssink sync=false
[20:33] <teuf> I can't compile gst-ffmpeg though
[20:33] <thomasvs> teuf: remove sync=false and try again
[20:33] <thomasvs> BBB: try to do what ?
[20:33] <teuf> well, I'm stuck with no gstffmpeg plugin
[20:33] Action: BBB will give gst-ffmpeg some love tonight
[20:33] <BBB> how does it not compile?
[20:33] <BBB> you guys seriously need some verbosity added to your ~/.defaults
[20:34] <teuf> BBB: same issue as last time i386/cputest.c:5:21: dsputil.h : Aucun fichier ou répertoire de ce type
[20:34] <teuf> (no such file or directory)
[20:34] <thomasvs> BBB: *PLEASE* *please* *PLEASE* rip out colorspace conversion code from ffmpeg
[20:34] <alley_cat> so, my last problem now is that gst-player and totem both abort with an error from the opt scheduler
[20:34] <BBB> the patch is not applied properly
[20:34] <teuf> and I retried with a fresh checkout
[20:34] <thomasvs> BBB: what do you want me to give you for it
[20:34] <thomasvs> alley_cat: which one ?
[20:34] <BBB> thomasvs: a date with ilse delange?
[20:34] <thomasvs> BBB: argh
[20:34] <BBB> *g*
[20:34] <alley_cat> ** ERROR **: file gstoptimalscheduler.c: line 723 (destroy_group): assertion failed: (group->group_links == NULL)
[20:34] <BBB> teuf: I'll work on it tonight
[20:34] <BBB> it'll take me a few hours
[20:34] <alley_cat> for gst-player cvs and totem 0.99.9
[20:35] <teuf> BBB: any suggestion for a quick hack to get it to compile ?
[20:35] iain ([email protected]) got netsplit.
[20:35] smoke ([email protected]) got netsplit.
[20:35] <thomasvs> BBB: anything slightly more realistic ?
[20:35] <thomasvs> btw, is there anyone objecting to me doing a release today/tomorrow ?
[20:36] <BBB> teuf: cd gst-libs/ext/ffmpeg && patch -p0 < patch/matroska.diff && patch -p0 < patch/autotools.diff && patch -p0 < patch/disablemmx.diff && patch -p0 < patch/disabletools.diff && patch -p0 < patch/disableinstalllibs.diff && cd ../../..
[20:36] <BBB> thomasvsehm... let me think
[20:36] <BBB> how about a digital photocamera? :p
[20:36] <thomasvs> BBB: how about I fix gst-ffmpeg proprely ?
[20:36] iain ([email protected]) returned to #gstreamer.
[20:37] <BBB> I've seen a real beauty for $180
[20:37] <BBB> thomasvs: please do
[20:37] <BBB> I suck at autotools
[20:37] <BBB> even though I tried a real nice hack
[20:37] <thomasvs> BBB: I meant, "in exchange" :)
[20:37] <BBB> uhm
[20:37] <BBB> ok
[20:37] <BBB> deal
[20:37] <BBB> :)
[20:38] <BBB> but then you have to delay the release of gst-ffmpeg until tomorrow
[20:38] <BBB> so I can work some more on it tonight
[20:38] <thomasvs> what do you mean ?
[20:38] <BBB> I'll discuss with you tomorrow morning your time
[20:38] <BBB> I'll stay up for a few hours tonight
[20:38] <thomasvs> yeah, but I can't untangle gst-ffmpeg in one day
[20:38] <BBB> it's not that hard
[20:38] <BBB> I'm just doing some things in a very clumsy way
[20:38] <BBB> if you know what to do, it's very easy, I think
[20:39] <thomasvs> my plan was to rerelease gst-ffmpeg 0.7.1 as 0.8, then in the next week after we released for gnome fix it properly
[20:39] <thomasvs> because we have tarballs due
[20:39] <BBB> you just said GNOME won't ship gst-ffmpeg?
[20:39] <thomasvs> anyways
[20:39] <BBB> and tarballs are only for GNOME beta2
[20:39] <BBB> not GNOME rc1
[20:39] <thomasvs> BBB: yeah, so I'm saying "I need to focus on releasing 0.8" instead of gst-ffmpeg
[20:40] <BBB> oh,
[20:40] <BBB> ok
[20:40] <thomasvs> only one release candidate, yikes
[20:40] <thomasvs> gnome's release schedule is too tight
[20:41] smoke ([email protected]) got lost in the net-split.
[20:41] Action: BBB thinks GNOME will slightly slip
[20:41] <BBB> gtk was too late
[20:41] <BBB> anyway
[20:41] Action: BBB off
[20:41] <teuf> BBB: wait :)
[20:41] <BBB> will finish stuff early so I can work long tonight
[20:41] <BBB> grmbl
[20:41] <BBB> :P
[20:41] <teuf> BBB: I think my compile issue has to do with locale settings
[20:41] <teuf> ah no
[20:41] <teuf> it still fails :)
[20:41] <teuf> sorry
[20:41] <BBB> told you
[20:41] <BBB> patch manually
[20:42] <BBB> and re-run configure
[20:42] <BBB> that's all
[20:42] <teuf> yeah, but I wanted to try that first
[20:42] <BBB> for PATCH in patch/*.diff; do
[20:42] <BBB> that line in configure.ac
[20:43] <BBB> that's the error
[20:43] <BBB> it patches in the wrong order
[20:43] <BBB> stupid that I didn't see that before
[20:43] <BBB> put the patches in the order that I just gave you
[20:43] <BBB> and then re-run configure
[20:43] <BBB> then it'll work
[20:44] <BBB> so for PATCH in matroska autotools disableinstalllibs disablemmx disabletools; do
[20:44] <BBB> patch -p0 -N -r rejects < patch/${PATCH}.diff > /dev/null || true
[20:44] <BBB> done
[20:44] <BBB> then rerun ./configure
[20:45] <KoRnouille> thomasvs: I still get these kind of errors : ../../libtool: line 3972: test: : integer expression expected
[20:45] <KoRnouille> with 1.5.0
[20:45] <KoRnouille> any ideas ?
[20:46] <thomasvs> KoRnouille: yeah, I have the same thing
[20:46] <thomasvs> KoRnouille: export max_cmd_line=419844
[20:46] <thomasvs> then rerun make
[20:46] <thomasvs> will fix it
[20:46] <thomasvs> not sure yet where the problem is coming from though
[20:46] <KoRnouille> hummf... ok
[20:47] <KoRnouille> when I "You should update your `aclocal.m4' by running aclocal.", I get a few errors, is this a big deal ?
[20:48] <thomasvs> depends on the errors
[20:48] <thomasvs> but if it keeps going, then no
[20:48] <BBB> teuf: I've added this to #136366, please post updates there
[20:48] <KoRnouille> stuff like aclocal: configure.ac: 174: macro `AM_GCONF_SOURCE_2' not found in library
[20:49] smoke ([email protected]) joined #gstreamer.
[20:52] <thomasvs> KoRnouille: yeah, that's bad
[20:52] <KoRnouille> lol
[20:52] <thomasvs> KoRnouille: doesn't it stop after that ?
[20:53] <KoRnouille> I remake && reinstalled libtool with your variable set
[20:53] <KoRnouille> aclocal looks good now
[20:53] <KoRnouille> :/
[20:53] <KoRnouille> Im remaking && reinstalling gstreamer now. I keep getting these errors though
[20:53] spyder482 ([email protected]) joined #gstreamer.
[20:53] <KoRnouille> the make breaks when I wanna install gst-plugins
[20:54] <thomasvs> "these errors"
[20:54] <KoRnouille> gstreamer goes all the way through
[20:54] <thomasvs> you already mentioned two :)
[20:54] <thomasvs> so I don't know which you mean
[20:54] <KoRnouille> thomasvs: errors, I mean libtool: line XXXX : test: : interger expression expected
[20:55] <thomasvs> KoRnouille: sorry, my mistake, it's max_cmd_len
[20:55] <KoRnouille> hummmf, thomasvs, in fact, I get aclocal errors (when running aclocal) when I run it from gst-plugins only
[20:55] <thomasvs> KoRnouille: you shouldn't run aclocal yourself
[20:55] <thomasvs> KoRnouille: run autoregen.sh
[20:55] <KoRnouille> hummmf...
[20:56] <KoRnouille> do I have to remake && reinstall libtool with export max_cmd_len=419844 ?
[20:56] <thomasvs> nope, just run make
[20:57] <KoRnouille> for gstreamer ? or libtool :/
[20:57] <thomasvs> sigh, you're mixing two things
[20:57] <thomasvs> a) for your aclocal problem, run autoregen.sh, nothing else
[20:57] sublett ([email protected]) left irc: "Nice of you to put me on the money"
[20:57] <thomasvs> b) for your make problem, do the export, then run make, nothing else
[20:57] <KoRnouille> ok
[20:57] <KoRnouille> :)
[20:57] <thomasvs> in neither case you have to fix/change/reinstlal libtool
[20:58] <KoRnouille> okay okay
[20:58] <thomasvs> hm, /me starts getting the clue
[20:58] <KoRnouille> aah ?
[20:58] <thomasvs> wrong libtool.m4 being picked up
[20:58] <thomasvs> damn
[20:58] <thomasvs> KoRnouille: just do what I said :)
[20:59] <thomasvs> KoRnouille: let me think loudly, don't mind me
[20:59] <kwm> ok second try. Does somebody knows how a pipeline for cdplayer looks likes? google only turns up packages ...
[20:59] sublett ([email protected]) joined #gstreamer.
[20:59] <KoRnouille> thomasvs: let me be curious, even if I don't understand. :)
[21:00] <thomasvs> kwm: don't think you can pipeline it - it's a bin
[21:00] <thomasvs> kwm: maybe check who wrote it in cvs and ask him directly
[21:00] <KoRnouille> thomasvs: should I export this variable to my .rc_profile for future installs (not just gstreamer, all appllications)
[21:01] danb ([email protected]) joined #gstreamer.
[21:02] <thomasvs> KoRnouille: no, just for now, until I fixed the problem correctly
[21:02] <KoRnouille> ooh, so it's a gstreamer problem then... not libtool (so once this is fixed, it should compile with 1.5.2 I guess, huh ?)
[21:03] <LeRoutier> aclocal: configure.ac: 547: macro `AM_PATH_IMLIB' not found in library  (missing m4/imlib.m4 for gst-ffmpeg ?)
[21:05] <KoRnouille> LeRoutier: aaah I got the same errors
[21:05] <teuf> ok, so I actually also have those AM_PATH_IMLIB and AM_PATH_SDL not found in library, that might be what is causing problems in the build later on
[21:05] <thomasvs> KoRnouille: no, it's not a gst problem.
[21:07] <LeRoutier> KoRnouille, what's your problem with libtool 1.5 ? mkinstalldirs related ?
[21:07] <thomasvs> LeRoutier,Kornouille: the problem with libtool is "if you build your libtool in a different prefix, aclocal doesn't pick up the new libtool.m4"
[21:07] <KoRnouille> cvs question : when I cvs ... co module, should I "cd module && cvs update -dPA" ?
[21:07] <thomasvs> later guys
[21:08] <KoRnouille> later thomasvs
[21:09] <KoRnouille> LeRoutier: no, I had errors like "libtool: line XXXX : test: : interger expression expected"
[21:10] <LeRoutier> which distrib ?
[21:10] <KoRnouille> when "make" my gst and plugz
[21:10] <KoRnouille> RH9.0 upgraded with the default apt-get repository
[21:10] <KoRnouille> it finally worked
[21:10] <KoRnouille> just now
[21:14] lupusBE ([email protected]) left irc: "Bezig met verlaten"
[21:21] lupusBE ([email protected]) joined #gstreamer.
[21:22] lupusBE ([email protected]) left irc: Client Quit
[21:28] thomasvs ([email protected]) left irc: Read error: 113 (No route to host)
[21:46] mathrick ([email protected]) joined #gstreamer.
[21:49] <mathrick> re
[21:51] <alley_cat> hmm, gst-ffmpeg needs lots of optimizations
[21:52] <LeRoutier> well, let it compile first
[21:53] <alley_cat> i got several files for which gstreamer can't keep up with video decoding
[21:55] <LeRoutier> bbl
[21:55] LeRoutier ([email protected]) left irc: "Leaving"
[22:04] Shoragan ([email protected]) left irc: "Leaving"
[22:06] thomasvs ([email protected]) joined #gstreamer.
[22:15] dany ([email protected]) joined #gstreamer.
[22:16] <dany> hi all, I have a very stupid question : is it possible to use gst-launch when gstreamer is compiled with the option --disable-registry (I disable this because I working on a embedded environment) ?
[22:17] <thomasvs> dany: might very well be, you need to check :)
[22:17] <thomasvs> dany: it used to be, but nobody has tried that we know of recently
[22:17] <dany> I seems to not working correctly
[22:17] <dany> this because it can't find any plugins
[22:18] thaytan ([email protected]) left irc: No route to host
[22:18] <dany> I do not brownse into the the code to see how it is actually implemented
[22:21] <thomasvs> I suggest you do.  It should be relatively easy to fix, but none of us have tried
[22:22] Uraeus ([email protected]) joined #gstreamer.
[22:22] <Uraeus> good evening
[22:24] <dany> Ok thanks, I will have a look ; but you are agree that this behaviour is not really normal
[22:24] <thomasvs> well, it's not how it should be
[22:24] <thomasvs> but it's quite normal given that onbody tests it or works on it :)
[22:25] <alley_cat> evening Uraeus
[22:25] <alley_cat> got the nas update?
[22:25] <dany> thomasvs, oki :-)
[22:25] <Uraeus> alley_cat: yes! :)
[22:26] markey ([email protected]) left irc: "leaving"
[22:29] <alley_cat> after reading about nas a little more it seems i won't be able to provide clocking easily, but mixer and source will definitely come when i have some free time
[22:30] <Uraeus> alley_cat: so you think NAS is better than Esound?
[22:31] <alley_cat> Uraeus: i have less problems with it, but i haven't really tested either in networked playback
[22:31] <alley_cat> getting esound to use dsp1 has always been a problem for me, it didn't seem to pick up any configuration i tried
[22:33] <bitshifter> alley_cat: excuse my ignorance, but what does it mean if you can't "provide clocking" for NAS? 
[22:33] <thomasvs> how do we do gst_caps_get_int these days ?
[22:34] <alley_cat> bitshifter: i will be able to do it, but not as easily as i first thought
[22:34] <alley_cat> i though nas would provide a stream clock, but it turned out it doesn't
[22:34] <alley_cat> so i have to do some more complicated stuff to find out how far the server is in processing the buffers i sent it
[22:37] Action: Uraeus wonders if everyone else is still able to playback mp3 and ogg with gstreamer with latest CVS? is it just me who have problems?
[22:37] <ds-work> thomasvs: gst_structure_get_int()
[22:37] <thomasvs> Uraeus: like i said, you have duplicate plugins
[22:37] <thomasvs> ds-work: yeah, but I can't just call this on caps->structs
[22:37] <thomasvs> ds-work: I guess that's by design ?
[22:37] <ds-work> thomasvs: yes, since there may be multiple structures
[22:37] thaytoo ([email protected]) joined #gstreamer.
[22:38] <thomasvs> ds-work: well, in my case I know they're fixed and only one
[22:38] <thomasvs> annoying to jump through so many layers, but I guess I have to suck it up
[22:38] <ds-work> gst_caps_get_structure(caps, 0)
[22:38] yippi ([email protected]) left irc: "Client exiting"
[22:38] <ds-work> "never make things more complicated, nor more simple, than they need to be"
[22:39] <thomasvs> with the bit open for interpretation being "need" :)
[22:39] <thomasvs> it's an interesting exercise to port stuff from 0.6 to 0.8
[22:40] <thomasvs> ds-work: gave the avi/mpeg logs a go ?
[22:40] <ds-work> ?
[22:40] <ds-work> I didn't actually see them anywhere
[22:40] <thomasvs> ds-work: the logs in my link in my mail
[22:40] <thomasvs> ds-work: they point to a site :)
[22:41] <thomasvs> hm, G_PTR_FORMAT doesn't exist for apps - should we provide a caps_debug function ?
[22:42] <thaytoo> good morning, all
[22:42] <ds-work> thomasvs: doesn't gstconfig.h get installed?
[22:43] Action: ds-work sees that it does
[22:43] <thomasvs> hm, yeah, but I guess we don't have it in our pkg-config flags
[22:43] <ds-work> I think you're looking for GST_PTR_FORMAT, anyway
[22:44] <thomasvs> ah, that could be the case too :)
[22:44] <thomasvs> hm
[22:44] <thomasvs> we're installing gstconfig.h in the wrong place, no ?
[22:44] <thomasvs> glib has it in /usr/lib
[22:44] yippi ([email protected]) joined #gstreamer.
[22:45] <ds-work> glib is just trying to be über-clever
[22:46] <thomasvs> hm, no gstconfig.h in the packages
[22:46] <teuf> glib puts it in /usr/lib because it's system dependant I think
[22:46] <thomasvs> yeha, exactly
[22:47] <teuf> that's probably required to work properly on platforms with lib and lib64 dirs
[22:48] <ds-work> no, they need separate include directories anyway
[22:50] <Uraeus> thomasvs: didn't have extra packages installed, but thinks work now, so I think it was a non-versioned .gstreamer directory under my home dir causing the issue
[22:52] ramon ([email protected]) joined #gstreamer.
[22:52] ramon ([email protected]) left irc: Client Quit
[22:55] <thaytoo> thomasvs: how can I suppress the warning messages from gst_parse_launch?
[22:55] <thaytoo> I'm giving it a GError pointer to store its error in, but it still prints a warning to to console
[22:55] <thaytoo> or is that only in the CVS build?
[22:56] <thomasvs> only in cvs build
[22:56] <thomasvs> I'd sitll like it gone, but company throws up a fit if we remove it
[22:56] <Uraeus> trying to use oggdemux and vorbisdec instead of vorbisfile I get: (process:3063): GStreamer-WARNING **: push on peer of pad vorbisdec0:src but peer is not active - anyone else with the same issue?
[22:57] <thaytoo> thomasvs: ok. I've got a gstreamer-properties patch to send off then
[22:57] <thaytoo> I left out adding audioconvert ! audioscale
[23:01] <thomasvs> Uraeus: like I said before, use audioconvert after vorbisdec :)
[23:02] <Uraeus> thomasvs: you seem to be repeating yourself recently :)
[23:04] <Uraeus> should vorbisfile be disabled in 0.8? 
[23:04] Zeenix ([email protected]) joined #gstreamer.
[23:04] <Uraeus> hi Zeenix
[23:04] <Zeenix> hello
[23:07] <thomasvs> Uraeus: yep
[23:15] <thomasvs> ds-work: found the logs now ?
[23:16] <ds-work> yes
[23:17] dany ([email protected]) left irc: "Leaving"
[23:17] Action: BBB reads logs
[23:17] Action: BBB finds how people complain about gst-ffmpeg
[23:17] Action: BBB doesn't get that, given that divxcore and xvidcore are about 3x slower
[23:17] <BBB> literally
[23:18] Action: teuf wishes he could complain about gst-fmpeg speed ;)
[23:18] Action: Uraeus points out he didn't complain this time no matter what thomasvs says
[23:18] <BBB> teuf: :)
[23:19] Action: Uraeus is not a complainer
[23:19] <alley_cat> BBB: ffmpeg is a lot faster than xviddec, but still not fast enough to decode some xvid videos here
[23:19] <alley_cat> i know i'm not on a current box, but other players play it, so i have to push a bit ;)
[23:20] <thomasvs> ds-work: find anything interesting in them ?
[23:20] <alley_cat> is there some major difference in optimizations between gst-ffmpeg and stock ffmpeg?
[23:21] <BBB> yes
[23:21] <BBB> we disable MMX
[23:21] <ds-work> thomasvs: haven't looked yet
[23:21] <BBB> I'm hoping to remove that within a few days
[23:21] <alley_cat> hmm, that would be it then
[23:21] <BBB> but I couldn't get MMX to compile
[23:21] <BBB> I was hoping for ds to help me fixing that up
[23:21] <alley_cat> why? no register for blabla?
[23:21] <BBB> I couldn't get it done
[23:21] <BBB> I suck at assembler
[23:21] <BBB> yes
[23:21] <ds-work> BBB: disable it until I get a chance.  And remind me frequently, please
[23:22] <BBB> sure :)
[23:22] <BBB> and it is disabled right now
[23:22] <BBB> it's just rather slow ;)
[23:22] <ds-work> no, disable the _function_
[23:22] <alley_cat> i got that when i manually upgraded ffmpeg to test something in gst-ffmpeg, switching to higher optimization settings made it compile
[23:22] <BBB> oh, yes
[23:22] <ds-work> like how it used to be
[23:22] <BBB> will try
[23:22] <BBB> but when I tried last time, it still failed to compile
[23:22] <BBB> I probably did it wrong
[23:22] <alley_cat> specifically from -O2 to -march=athlon -O3
[23:22] <BBB> I applied the same patch
[23:22] <BBB> ...
[23:22] <BBB> anyway
[23:22] <BBB> I will try
[23:22] <BBB> give me a few days
[23:22] <BBB> gst-ffmpeg isn't exactly *easy*
[23:23] <alley_cat> i know
[23:23] <ds-work> alley_cat: don't bother with compiler flags, it's impossible to make it faster that way
[23:23] <alley_cat> ds-work: well, but it's possible to compile it with mmx in gst-ffmpeg that way
[23:23] <BBB> ?
[23:23] <BBB> that's interesting
[23:23] <alley_cat> without compiler flags i got the message "unable to find a register to spill in ...", with compiler flags it compiled
[23:24] Action: thomasvs released nautilus-media, enough for today
[23:24] <BBB> weird
[23:24] <ds-work> alley_cat: probably a compiler bug
[23:24] <BBB> anyway
[23:24] <BBB> I'm off
[23:24] <alley_cat> i deal with ffmpeg in lots of different packages, i know how weird this thing is :)
[23:24] Action: BBB goes home to fix ffmpeg the whole night
[23:24] <BBB> :/
[23:24] <alley_cat> ds-work: quite possible
[23:24] BBB ([email protected]) left #gstreamer.
[23:25] Shoragan ([email protected]) joined #gstreamer.
[23:26] <alley_cat> wouldn't be the first one i encounter (and probably not the last either ;)
[23:26] <Uraeus> alley_cat: ok, NAS patch commited
[23:29] Zeenix ([email protected]) left irc: "The whole problem with the world is that fools and fanatics are always so certain of themselves, but wiser people so full of 
[23:33] kwm ([email protected]) left irc: "Leaving"
[23:41] wheels ([email protected]) joined #gstreamer.
[23:52] Uraeus ([email protected]) left irc: "Client exiting"
[23:53] bilboed ([email protected]) joined #gstreamer.
[23:55] Action: sxpert recompiles gstreamer. will launch sound-server for the night and see if any leaks are still there tomorrow morning
[00:00] --- Tue Mar  9 2004
[00:03] sack ([email protected]) left irc: "Leaving"
[00:11] teuf ([email protected]) left irc: "Client exiting"
[00:13] ramon ([email protected]) joined #gstreamer.
[00:14] <ramon> Hi.
[00:18] <ds-work> thomasvs: aournd?
[00:18] <ds-work> thomasvs: what file was that log from?
[00:25] <sxpert> the memory usage on last night's incarnation is much better. it seems that there is yet another memory leak somewhere ;(
[00:29] foser ([email protected]) left irc: "[ I want to believe ]"
[00:31] <mathrick> nite
[00:31] Nick change: mathrick -> mathrick|sleep
[00:32] <ramon> Hi.
[00:33] <ramon> What is the difference between finalize and dispose?
[00:35] <sxpert> ramon: uh, finalize will allow for the object to remove subobjects, dispose is supposed to instantaneously get rid of the object
[00:36] <ramon> So I guess that if I need to cleanup resources managed by an element, I should use dispose.
[00:36] <sxpert> guess so...
[00:36] <ramon> Thanks. 
[00:36] <ramon> I am going to bed.
[00:36] <ramon> See you.
[00:36] <sxpert> no idea though, was just giving generic usage
[00:36] <sxpert> see ya
[00:37] <ds-work> ramon: see http://le-hacker.org/papers/gobject/
[00:37] <ds-work> ramon: the part about object destruction
[00:45] <ramon> Thanks. Somehow I didn't find it in the GObject tutorial. Now I can see it.
[00:45] <ramon> And now going to bed.
[00:45] ramon ([email protected]) left #gstreamer (""No war for oil"").
[00:46] dap ([email protected]) joined #gstreamer.
[00:50] harshy ([email protected]) left irc: "I quit for now"
[00:52] jcsston ([email protected]) joined #gstreamer.
[00:53] md` ([email protected]) left irc: 
[00:53] harshy ([email protected]) joined #gstreamer.
[00:57] md` ([email protected]) joined #gstreamer.
[00:59] <sxpert> ok, sound server starts with 2.2% ram...
[01:02] LeRoutier ([email protected]) joined #gstreamer.
[01:02] <LeRoutier> re
[01:03] <sxpert> re
[01:04] <sxpert> LeRoutier: ava ?
[01:06] <LeRoutier> yep, ça va bien
[01:06] <LeRoutier> j'espérais voir ronald mais semble qu'il est parti
[01:07] <sxpert> oue, il est rentre allah casa a 23h24
[01:07] <sxpert> LeRoutier: de mon cote, je fais des tests d'endurance ;D
[01:08] <sxpert> y'a un gros mieux semble t'il par rapport aux leaks...
[01:08] <LeRoutier> hum, physiques ou logiciels ? :)
[01:08] <LeRoutier> valgrind ?
[01:08] <sxpert> logiciels, je teste gstreamer sur plusieurs jours...
[01:08] <sxpert> LeRoutier: non, valgrind marche pas sur ma box...
[01:08] <wheels> sxpert: Aren't you supposed to guard against using words like "leaks" in French?  :-)
[01:09] <sxpert> wheels: why would that be ?
[01:09] <wheels> sxpert: "language polution"?
[01:09] <sxpert> wheels: who the f*** cares ?
[01:09] <LeRoutier> French academia hadn't found a stupid word to replace it yet
[01:09] <wheels> sxpert: The French, apparently.  :-)
[01:09] <sxpert> wheels: and this is not quebec ;D
[01:09] <sxpert> LeRoutier: yeah ;D
[01:09] Action: wheels will actually be back in Franch later this week...
[01:09] <wheels> err, France...
[01:09] <sxpert> LeRoutier: like c\'ed\'erom...
[01:10] <sxpert> LeRoutier: where tf they found that one ???
[01:10] <LeRoutier> nan, z'ont osé faire une merde phonétique N
[01:10] <sxpert> oue...
[01:10] <sxpert> LeRoutier: c'est partout a la fnac...
[01:10] Action: sxpert trouve ca d'une debilite...
[01:10] Nick change: harshy -> harshyAWAY
[01:10] <LeRoutier> putain, je te les clouerais sur une porte de grange ces cons, z'ont que ça à foutre
[01:11] <sxpert> LeRoutier: oue...
[01:11] <sxpert> faut croire ;D
[01:11] <sxpert> faut dire, ils ont VGE parmi eux ;D
[01:11] <LeRoutier> yep, faut être sénile pour y rentrer, c'est l'effet Chamalière
[01:12] <sxpert> lol
[01:12] <sxpert> bref, pour en revenir a gstreamer, je fais tourner mon sound server pendant des heures, voir si ca crache et tout...
[01:12] <LeRoutier> valgrind marche pas sur ta box ? à cause du noyau 2.6 ? de ta glibc+nptl ?
[01:13] <sxpert> atta ;D
[01:13] <sxpert> valgrind.so: When searching for client's argc/argc/envp:        suspiciously many (1000) argv[] entries; giving up
[01:14] <sxpert> sais pas d'ou ca sort ;D
[01:14] <LeRoutier> ça te fait ça aussi bien avec la dernière release qu'avec le CVS current ?
[01:14] <sxpert> tiens, mon soft est a http://sxpert.esitcom.org/sound-server.c
[01:14] <sxpert> LeRoutier: oue
[01:17] <sxpert> et puis la 2.1 sur le laptop, elle a segfaulte au moment de trouver les 9800 leaks...
[01:17] <LeRoutier> faut dbus 0.20 ou CVS ?
[01:17] <sxpert> cvs
[01:18] <sxpert> mais le dbus, tu peux le desactiver entierement ;D
[01:18] <LeRoutier> DISABLE_DBUS=1 ?
[01:18] <sxpert> -DDISABLE_DBUS dans le make :D
[01:19] <sxpert> il te faux le .h aussi :  http://sxpert.esitcom.org/sound-server.h
[01:19] <LeRoutier> 404
[01:19] <sxpert> reessaye :D
[01:20] Action: sxpert est curieux de savoir si ca marche chez un autre ;D
[01:21] <sxpert> euh, il faut un petit xml de configuration aussi ;D
[01:22] <thaytoo> aussie! You rang? ;)
[01:23] iain ([email protected]) left irc: "it was the perfect end to a perfect life"
[01:24] <sxpert> thaytoo: no :D
[01:24] <sxpert> thaytoo: "aussi" in french is "too"
[01:25] <thaytoo> ah :)
[01:25] <thaytoo> well, I understood 'xml', I think ;)
[01:26] <sxpert> lol
[01:26] <sxpert> thaytoo: I meant that he also needed a little xml config file ;D
[01:27] <thaytoo> ah, je comprehende. Merci
[01:27] <thaytoo> is that right?
[01:28] <ds-work> perhaps you people didn't get the memo.  The new official language of #gstreamer is español
[01:29] <thaytoo> iie! nihongo desho!
[01:30] <LeRoutier> no ablar espanol
[01:30] <spyder482> ablar?
[01:30] <ds-work> hablo
[01:30] <spyder482> i think you messed that up ;)
[01:30] <spyder482> yeah
[01:31] <LeRoutier> that's possible, i never learned spanish
[01:31] jcsston ([email protected]) left irc: Read error: 60 (Operation timed out)
[01:32] <spyder482> Hablo un poco.
[01:33] <sxpert> thaytoo: "ah, je comprends. Merci."
[01:33] pb_ (~pb@2002:5160:45ef:0:240:95ff:fe30:aa84) left irc: "later"
[01:33] <sxpert> ds-work: no hay ningun problema.
[01:33] <thaytoo> sxpert: ta
[01:34] <thaytoo> 1 year of high school french ... 10 year gap :)
[01:34] <sxpert> thaytoo: heh
[01:40] harshyAWAY ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[01:41] danb ([email protected]) left irc: "Lost terminal"
[01:43] danb ([email protected]) joined #gstreamer.
[01:47] sublett ([email protected]) left irc: Read error: 113 (No route to host)
[02:03] arana ([email protected]) joined #gstreamer.
[02:18] spyder482 ([email protected]) left irc: Read error: 113 (No route to host)
[02:21] <ds-work> it's so strange to read a bug that is >= 6 months old that I honestly cannot ever remember seeing before
[02:27] <sxpert> lol
[02:28] dap ([email protected]) left irc: "Client exiting"
[02:28] <sxpert> if anyone is daring enough, could that person correct this :
[02:28] <LeRoutier> bonne nuité
[02:28] <sxpert> * (process:24080): WARNING **: opt: unlink elements in same group: implement me
[02:28] <LeRoutier> so, good night in english
[02:28] LeRoutier ([email protected]) left irc: "Leaving"
[02:28] <sxpert> thank you ;D
[02:35] Action: ds-work SCREAMS
[02:35] <ds-work> GstCaps has no structure padding
[02:39] BBB ([email protected]) joined #gstreamer.
[02:41] <ds-work> BBB: GstCaps has no structure padding
[02:42] <ds-work> BBB: does 'fakesrc ! { queue ! identity ! fakesink }' work for you?
[02:42] <ds-work> I'm getting crashes
[02:42] <BBB> should it?
[02:42] <BBB> GstCaps is only a GstStructure
[02:43] <BBB> seems to run fine
[02:43] <ds-work> right _now_, GstCaps is only a list of GstStructures
[02:43] <BBB> this is yesterday's CVS, btw
[02:43] <BBB> oh, yes, right
[02:43] <BBB> well, go ahead and add it
[02:43] <BBB> I don't care about ABI much right now
[02:43] <BBB> I'm supposed to get gst-ffmpeg to work tonight :/
[02:44] <BBB> yup, pipeline works perfectly fine here
[02:44] <BBB> any other testcases that you want?
[02:44] <ds-work> no
[02:44] <ds-work> I'm just questioning my sanity
[02:45] <arana> BBB: what version of ffmpeg sources does this matroska patch work on?
[02:45] Nick change: arana -> spyder482
[02:47] <BBB> ds-work: as in?
[02:47] <BBB> :)
[02:48] <BBB> arana: er, the one in gst cvs is slightly bugged
[02:48] <BBB> one of the few small things that I still don't like about quilt
[02:48] <BBB> I'll fix it later
[02:48] <BBB> it'll fail partially, but it doesn't matter
[02:48] <ds-work> BBB: why should a simple pipeline suddenly start failing?
[02:48] <BBB> but it operates on current CVS
[02:48] <BBB> ds-work: hardware issue? check dmesg
[02:48] <BBB> memory issue?
[02:48] <BBB> CPU heated?
[02:48] <BBB> did you make a typo in your new scheduler?
[02:49] <BBB> maybe a weird electron bump during your latest compile?
[02:49] <ds-work> it happens with every scheduler
[02:49] <BBB> once every while (say: after a few years or months) my computer just stops working
[02:49] Action: ds-work scrutinizes wingo's last change
[02:49] <BBB> the thing is 8 years old
[02:49] <spyder482> quilt?
[02:49] <BBB> I suppose that's normal
[02:49] <BBB> quilt
[02:49] <thaytoo> ds-work: works here
[02:49] <BBB> so I reboot
[02:49] <spyder482> what's quilt?
[02:49] <BBB> and then everything's back to normal
[02:49] <BBB> a patchset manager
[02:50] <spyder482> oh
[02:50] <thaytoo> ds-work: oh, but I might be behind wingo's latest
[02:50] <BBB> same here
[02:50] <spyder482> i mean the patch you posted on our ML
[02:50] <ds-work> thaytoo: do you have 2004-03-07 David Schleef?
[02:50] <BBB> spyder482: uh, that one should work
[02:50] <spyder482> i tried with 0.4.8 but the patching fails
[02:50] <spyder482> the makefile isn't patched
[02:53] <spyder482> some special way to patch besides patch -p0/-p1
[02:54] <ds-work> BBB: http://www.schleef.org/~ds/patch
[02:54] <ds-work> BBB: this seemed to be needed on my system
[02:59] <thaytoo> ds-work: 2004-03-07 David Schleef on which file?


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