[gst-devel] Daily IRC logs

[<wim...@ch...>]

[06:31] omega_ (om...@om...) joined #gstreamer.
[06:32] <iGN_> Yo, omega_ :)
[06:32] <omega_> yo
[06:32] <iGN_> What's up?
[06:32] <omega_> trying to get the rest of rh7.1
[06:33] <iGN_> Eww. ;) RH70 was such a big fiasco that it was an easy tas=
k to convert the rest of the doubtful at the office to Debian/Progeny.
[06:33] <omega_> heh
[06:33] <iGN_> Hehe, we spent a full afternoon "debugging" stuff that tur=
ned out to be compiler-bugs.
[06:33] <omega_> whee
[06:34] <iGN_> The hot tip is to check the errata early, and check it oft=
en. But you already knew that :)
[06:34] Action: omega_ is just having trouble finding an FTP server that =
isn't fragged
[06:34] <omega_> yup <g>
[06:36] <omega_> what I want to know is: how friggin hard is it to write =
an FTP server that doesn't suck!?!?!
[06:36] <omega_> just connecting to these servers is a chore, let alone c=
d'ing and ls'ing to find stuff
[06:37] <omega_> they're not bandwidth-starved....
[06:37] <omega_> well, they are, but not that starved...
[06:43] <iGN_> I dunno.
[06:43] <iGN_> I haven't tried.
[06:43] <iGN_> Being how hard a time most people have, I'm not going to s=
ay that it's easy.
[06:44] <iGN_> I've said things like that before, and the "OK, show me" i=
s always painful.
[06:44] <omega_> heh
[08:48] matth_ (ma...@qw...) got netsplit.
[08:48] omega_ (om...@om...) got netsplit.
[08:49] matth_ (ma...@qw...) returned to #gstreamer.
[08:49] omega_ (om...@om...) returned to #gstreamer.
[08:51] taazzzz (dlehn@66.37.66.32) left irc: Ping timeout for taazzzz[66=
.37.66.32]
[08:53] taazzzz (dlehn@66.37.66.32) joined #gstreamer.
[09:16] taazzzz (dlehn@66.37.66.32) left irc: Ping timeout for taazzzz[66=
.37.66.32]
[09:16] taazzzz (dlehn@66.37.66.32) joined #gstreamer.
[10:34] omega_ (om...@om...) left irc: Ping timeout for omega_[omeg=
acs.net]
[10:54] sienap (sy...@ip...) joined #gstreamer.
[10:54] <sienap> hej all
[11:11] sienap (sy...@ip...) left irc: sienap has no rea=
son
[11:41] matth_ (ma...@qw...) got netsplit.
[11:42] matth_ (ma...@qw...) returned to #gstreamer.
[12:20] matth_ (ma...@qw...) got netsplit.
[12:21] matth_ (ma...@qw...) returned to #gstreamer.
[12:22] thomas (th...@ur...) joined #gstreamer.
[12:23] <thomas> names
[12:28] Action: thomas is away =02- Automatically set away. -=02 messages=
 will be saved.
[14:12] matth_ (ma...@qw...) got netsplit.
[14:12] matth_ (ma...@qw...) returned to #gstreamer.
[14:19] ajmitch (me...@p3...) joined #gstreamer.
[14:21] Action: thomas wants the log to record that he's now getting a ve=
ry nice sounding mix of an .ogg and an .mp3
[14:53] thomas (th...@ur...) left irc: I'm outta here!
[14:53] thomas (th...@ur...) joined #gstreamer.
[14:59] Action: thomas is away =02- Automatically set away. -=02 messages=
 will be saved.
[17:37] Scav (za...@za...) joined #gstreamer.
[17:38] <ajmitch> hi=20
[17:39] <matth_> hey... (it's quiet in here)
[17:39] <ajmitch> hehe
[17:40] <ajmitch> i thought everyone else was asleep ;)
[17:40] <matth_> not me, though i'm not paying that much attention
[17:40] <matth_> ;-)
[17:41] <ajmitch> i'm just considering whether i should start doing some =
coding or go to bed ;)
[17:42] <Scav> yo
[17:43] Action: Scav was looking for hadess
[17:43] <Scav> or anyone that has or uses or has used a mac
[17:44] <ajmitch> i have used a mac (but not with a decent OS loaded)
[17:50] <Scav> i was wonderign about macintosh filesystems
[18:00] <ajmitch> dunno much, i only use them for uni when forced to ;)
[18:03] manbash (ds...@15...) joined #gstream=
er.
[18:04] steveb (st...@no...) joined #gstreamer.
[18:06] manbash (ds...@15...) left irc: Read =
error to manbash[151-MADR-XL2.libre.retevision.es]: Connection reset by p=
eer
[18:10] Nick change: taazzzz -> taaz
[18:17] ajmitch (me...@p3...) left irc: Ping timeout for a=
jmitch[p33-max7.dun.ihug.co.nz]
[18:18] matth_ (ma...@qw...) got netsplit.
[18:18] iGN_ (ig...@lo...) got netsplit.
[18:18] taaz (dlehn@66.37.66.32) got netsplit.
[18:18] iGN_ (ig...@lo...) returned to #gstreamer.
[18:18] taaz (dlehn@66.37.66.32) returned to #gstreamer.
[18:18] matth_ (ma...@qw...) returned to #gstreamer.
[18:18] ajmitch (me...@p4...) joined #gstreamer.
[18:21] taaz (dlehn@66.37.66.32) left irc: Ping timeout for taaz[66.37.66=
.32]
[18:21] taaz (dlehn@66.37.66.32) joined #gstreamer.
[18:49] thomas (th...@ur...) left irc: Leaving
[18:56] thomas (th...@ur...) joined #gstreamer.
[18:57] ChiefHighwater (pa...@te...) joined #gstreamer.
[19:00] thomas (th...@ur...) left irc: I'm outta here!
[19:01] hadess (ha...@pc...) joined #gstreamer.
[19:05] <Scav> hadess!
[19:06] <hadess> Scav: dude !
[19:06] Action: Scav has seen the new titanium powerbook and wants one....
[19:06] <hadess> i want one as well :/
[19:07] <hadess> i want a laptop for short actually
[19:07] omega_ (om...@om...) joined #gstreamer.
[19:07] <Scav> hold a sec
[19:07] <Scav> yo omega
[19:07] <hadess> hey omega_
[19:08] <omega_> yo
[19:08] Nick change: wtay-zZz -> wtay
[19:08] <wtay> dudez
[19:08] <hadess> mate !
[19:09] Nick change: omega_ -> omega-breakfast
[19:10] Action: wtay is listening to an .ogg and .mp3 mixed with gstreame=
r
[19:11] <Scav> mixer?
[19:12] <wtay> yeah, adder in plugins/filers/added
[19:12] <wtay> s/added/adder
[19:12] <Scav> how was easter guys?
[19:12] <hadess> short
[19:12] <wtay> nice vacation
[19:13] <wtay> and yours?
[19:13] <Scav> hadess: mac illustrator 8, it reads eps files?
[19:13] <hadess> iirc yes
[19:13] <Scav> mine was good, 5 rounds of golf and frantic previking codi=
ng
[19:14] <wtay> Scav: what's the previking status now?
[19:14] <Scav> with some very heavy testing, we found a major bug
[19:14] <wtay> Scav: gstreamer related?
[19:15] <Scav> nope
[19:15] <Scav> i had to shelve my release till we have finished fixing th=
e outstanding previking bug
[19:16] <omega-breakfast> ooh ooh ooh
[19:16] <Scav> unfortunately it has required a major architecture change.=
..
[19:16] <wtay> ouch
[19:16] <omega-breakfast> dell sent me a hard drive and a carrier to repl=
ace the cracked carrier they didn't replace with my laptop
[19:16] <omega-breakfast> since they aren't capable of sending just the c=
arrier
[19:17] <omega-breakfast> and guess what?
[19:17] <omega-breakfast> the carrier they shipped has a 20gb drive in it
[19:17] <omega-breakfast> I only have a 12gb drive
[19:17] <Scav> coool
[19:17] <wtay> cool
[19:17] <omega-breakfast> so guess what I'll be doing soon?  moving my fi=
les over <g>
[19:17] Action: Scav wonders if omega would donate the 12gb 2.5" hard dis=
k to gstreamer development in london :)
[19:18] <omega-breakfast> have to return it to dell...
[19:18] <Scav> thats a pity
[19:18] <omega-breakfast> esp since their packing slip says 12gb on it
[19:18] <omega-breakfast> I'd hate to disappoint them by sending back a d=
rive that doesn't match their packing slip <g>
[19:20] <wtay> hehe
[19:20] Action: omega-breakfast also has to disassemble his laptop to put=
 the keyboard in correctly
[19:21] Action: Scav sends omega his 6gb hard disk if the 12gb hard disk =
is sent to him :)
[19:21] <wtay> what's the status with glib2.0?
[19:21] <omega-breakfast> n/c
[19:21] Action: Scav has glib1.3.x working on his machine
[19:21] <Scav> but i have found a great short term solution
[19:22] <Scav> to the gtk requirement
[19:22] <wtay> Xvfb?
[19:22] <omega-breakfast> xvfb?
[19:22] <Scav> nope
[19:22] <Scav> even better
[19:22] <wtay> export DISPLAY?
[19:22] <wtay> custom gtk version?
[19:22] <Scav> someone has produced a subset of GTK that is just the GTK =
object system
[19:22] <omega-breakfast> oh, where????
[19:22] <Scav> that is stable
[19:22] <wtay> aha
[19:23] Action: matth_ perks up
[19:23] <wtay> cool
[19:23] <Scav> hold a sec, i will find it
[19:23] <Scav> i have it on my hard disk
[19:27] <Scav> on my home hard disk
[19:27] <omega-breakfast> doh
[19:27] Action: Scav is on his laptop
[19:27] <Scav> im trawling to find it
[19:29] Nick change: omega-breakfast -> omega_
[19:32] <Scav> found it
[19:32] <Scav> http://gsk.sourceforge.net/
[19:33] <Scav> if you browse their cvs repos
[19:33] <Scav> theres a dir
[19:33] <wtay> hum
[19:34] <omega_> it's 500KB...
[19:34] <Scav> src/gtk
[19:35] <Scav> and that has the gtk subset
[19:35] <omega_> cool, we could copy that into a subdir in gst/ for now
[19:35] <Scav> yah exactly
[19:35] <Scav> and use the same autoconf directive they have used
[19:35] <wtay> cool
[19:36] <Scav> to merge it with gtk if wanted or to use the subset
[19:36] <Scav> and that would allow console use of gstreamer without klud=
ges like xvfb
[19:36] <omega_> neat
[19:37] <Scav> i should have pointed it out earlier but i have been overw=
helmed with this bug thrashing
[19:37] <omega_> ick
[19:37] Action: Scav crosses his fingers coz it seems fixed
[19:38] Action: Scav doesnt wanna know the diff size between previking 0.=
4.8 and CVS HEAD for previking 0.4 tree
[19:39] <Scav> however we have found a memory leak now as a result of the=
se changes that needs to be ironed out
[19:39] <Scav> till I can go back to normal again :)
[19:41] <wtay> looks like they copied glib2.0...
[19:41] <wtay> or not completely
[19:41] <wtay> but the interfaces are there too
[19:41] <omega_> they are?
[19:42] <omega_> um...
[19:42] <omega_> src/gtk/, not src/
[19:42] <Scav> yah look in src/gtk
[19:42] <wtay> I was looking at src/
[19:43] <wtay> they even have their own xml parser...
[19:43] <omega_> sounds like glib2.0, sorta
[19:43] <omega_> but gsk seems more for 'servers'
[19:43] <Scav> gsk seems nice, but really its a bit more like a layer on =
top of glib
[19:44] <wtay> also has threadpools...
[19:44] <omega_> hmmm
[19:44] <wtay> looks like a pre glib2.0 clone
[19:44] <omega_> well, we'd just copy src/gtk/ into gst/gtk/ and patch it=
 in
[19:44] <Scav> yah
[19:45] <wtay> yup
[19:45] <omega_> well, and fix it up to use libtool and other stuff
[19:47] Action: Scav is busily investigating the memory leak
[19:47] <wtay> we should take this opportunity to wrap them in gst_* macr=
os/functions so we can prepare glib2.0 too
[19:48] <omega_> only the functions we already use by name
[19:48] <omega_> the rest is what gstobject.[ch] is for
[19:48] <wtay> yup
[19:49] <Scav> thankfully its not much of a memory leak
[19:49] <omega_> so wrap up any reference to GtkObject with our own macro=
s
[19:49] <omega_> Scav: until you run it for a year solid ,g>
[19:49] Action: omega_ hands Scav the 20gb drive for swap
[19:49] <Scav> omega_: yah we have done 10,000 calls thru it, a megabyte =
extra ram is being used now
[19:49] <omega_> hmm
[19:50] <omega_> so 100bytes per call
[19:50] <omega_> zaheer quoted much larger numbers last weekend
[19:50] <Scav> yah
[19:50] <omega_> this 20gb driver is faster, too <g>  12ms instead of 13m=
s
[19:50] <Scav> but that was for different tests
[19:50] <omega_> and it's an IBM drive instead of a fujitsu
[19:51] <omega_> ok
[19:51] <Scav> this is after all the calls are released
[19:51] <Scav> it has increased by a meg after 10,000 calls were started =
and completed
[19:52] <omega_> and a 2MB buffer to a 512KB buffer
[19:52] <omega_> Scav: hmm
[19:53] <wtay> it's not a mmap issue again, is it?
[19:53] <Scav> its definitely some stuff that hasnt been freed
[19:54] <wtay> ok
[19:56] <omega_> btw, either tomorrow or Friday I'll be going back to OGI=
 again to do a libdv and hopefully 1394src plugin for gst
[19:56] <Scav> cool
[19:56] <wtay> yes!
[19:56] <Scav> 1394src
[19:56] Action: wtay has to get a firewire card then
[19:56] <omega_> 1394sink can't happen yet because the driver isn't capab=
le of it
[19:56] <Scav> is what I want for another project
[19:56] <omega_> heh
[19:57] <wtay> back in 40 minutes...
[19:57] Nick change: wtay -> wtay-away
[19:57] <Scav> i have a situation where we want to do the following for s=
ome talks at the mosque:
[19:58] <Scav> from digital video want the video and audio teed to an mpe=
g2 encoder to a disksrc and to a sink that can be used by realencoder and=
 netshow's encoder
[19:58] matth_ (ma...@qw...) got netsplit.
[19:58] ajmitch (me...@p4...) got netsplit.
[19:58] hadess (ha...@pc...) got netsplit.
[19:58] iGN_ (ig...@lo...) got netsplit.
[19:58] steveb (st...@no...) got netsplit.
[19:58] ChiefHighwater (pa...@te...) got netsplit.
[19:58] <omega_> hmmm
[19:58] <omega_> oops
[19:58] <Scav> to be sent from the mosque to our office
[19:58] <omega_> convergence.de was using vlc to do that at GUADEC, but o=
nly to mpeg2 streaming
[19:58] <omega_> but from a bt848 analog card ;-(
[19:59] steveb (st...@no...) returned to #gstreamer.
[19:59] <Scav> i have absolutely no experience with real and video
[19:59] Action: omega_ envisions df going from 28MB to 8028MB free on /
[19:59] <Scav> what kinda input will its encoder take from?
[19:59] <omega_> which encoder?
[19:59] iGN_ (ig...@lo...) returned to #gstreamer.
[19:59] matth_ (ma...@qw...) returned to #gstreamer.
[19:59] ajmitch (me...@p4...) returned to #gstreamer.
[19:59] hadess (ha...@pc...) returned to #gstreamer.
[19:59] ChiefHighwater (pa...@te...) returned to #gstreamer.
[19:59] <Scav> realencoder
[19:59] <omega_> dunno
[20:02] thomas (th...@ur...) joined #gstreamer.
[20:02] <thomas> hi everyone
[20:02] <omega_> yo
[20:03] <thomas> omega: what should be the right file extension for an mp=
eg video ?
[20:04] <thomas> I tried playing one but it used mad
[20:04] <omega_> um...
[20:04] <thomas> it was .mpeg and .mpg
[20:04] <omega_> yeah
[20:04] <omega_> if the autoplugger used mad, then something's wrong with=
 either the file or the typefind function
[20:04] <thomas> hmmm... I just downloaded some sample movie.  I wanted t=
o test if it would mix that as well...
[20:05] <omega_> do a `file` of it
[20:05] <thomas> I have some issues with the startup of gstreamer... If I=
 use buffers ...
[20:05] <thomas> ... can I make the mix start later but have a more const=
ant output ?
[20:05] <thomas> ... esd seems  to not do too well priority wise...
[20:06] <thomas> ... so the start is a bit dodgy.  I can't tell if it's m=
y fault or my machine's
[20:06] <omega_> eh?
[20:06] <omega_> esd causes major headaches as far as latency
[20:06] <thomas> well... during startup, there's so much going on ...
[20:06] <thomas> ah ok... I'll try with oss again then.
[20:06] <thomas> but would buffers help here ?
[20:06] <omega_> all sound servers suck in that regard
[20:06] <omega_> not really
[20:07] <omega_> depends on whether you're getting mangled audio or not
[20:07] <thomas> no, I think it's just a bit patchy at the start
[20:07] <hadess> i have an issue on my own
[20:07] <thomas> if you have the time you should try it out and see how i=
t sounds for you...
[20:07] <hadess> will it be possible to speed up the startup of gstreamer=
 ?
[20:07] <omega_> hadess: how long does it take?
[20:07] <thomas> I tried oss today but it didn't work, strangely enough
[20:08] <Scav> doesnt take long to startup here
[20:08] <hadess> omega_: a couple of seconds
[20:08] <thomas> omega_: ok, file sample.mp gives : sample.mpg: MPEG syst=
em stream data
[20:08] <omega_> do you run -register?
[20:08] <hadess> yep
[20:08] <omega_> thomas: hmmm
[20:08] <omega_> hadess: then I really have no idea what would take that =
long
[20:10] <hadess> that might be the way i compile it...
[20:10] <thomas> if you delete some of the plugins, will it run faster ?
[20:10] <omega_> not if you have a registry
[20:11] <Scav> aaaargh!!!
[20:11] <omega_> Scav: don't have a cow.. <g>
[20:11] Action: Scav has realised solaris is stupid
[20:11] <omega_> well...
[20:11] <omega_> Duh!
[20:11] <Scav> i have realised we dont have a memory leak :P
[20:11] <omega_> mu
[20:12] <Scav> it just has scheduled garbage collection
[20:12] <omega_> doh
[20:12] <Scav> and it has now reclaimed the memory :P
[20:12] <Scav> from previking
[20:12] <omega_> doh
[20:13] Action: omega_ has lost 70% of his battery's charge in 1hr5min
[20:13] <Scav> eewww
[20:13] Action: omega_ is not happy
[20:13] Action: Scav suggests plugging into oregon's electricity system :=
)
[20:13] <omega_> um, sure <G>
[20:15] Action: omega_ is trying to patch in minigtk
[20:15] Action: thomas is away =02- Automatically set away. -=02 messages=
 will be saved.
[20:16] matth_ (ma...@qw...) got netsplit.
[20:16] ajmitch (me...@p4...) got netsplit.
[20:16] hadess (ha...@pc...) got netsplit.
[20:16] thomas (th...@ur...) got netsplit.
[20:16] iGN_ (ig...@lo...) got netsplit.
[20:16] iGN_ (ig...@lo...) returned to #gstreamer.
[20:16] matth_ (ma...@qw...) returned to #gstreamer.
[20:16] ajmitch (me...@p4...) returned to #gstreamer.
[20:16] hadess (ha...@pc...) returned to #gstreamer.
[20:16] thomas (th...@ur...) returned to #gstreamer.
[20:16] <Scav> -lilo- [GlobalNotice] Hi all.  You will notice some intent=
ional splits and joins as we test the new hubbing.  Sorry for the inconve=
nience and thanks for using OPN!
[20:16] <Scav> :)
[20:18] ChiefHighwater (pa...@te...) got netsplit.
[20:18] ChiefHighwater (pa...@te...) returned to #gstreamer.
[20:19] <omega_> Scav: have you tried minigtk yet?
[20:19] <Scav> nope
[20:19] <omega_> ok, we'll see how well it works
[20:19] <omega_> compiling OK so far
[20:19] <Scav> i have compiled gsk
[20:19] <Scav> on my home machine
[20:19] <Scav> using the mini gtk
[20:19] <omega_> but I have a feeling it's getting headers from /usr/incl=
ude in preference to the minigtk dir
[20:19] <omega_> so I may have to do some tricks
[20:19] <Scav> :)
[20:20] <omega_> only three files include gtk*.h, so that should be trivi=
ally solvable <g>
[20:20] taaz (dlehn@66.37.66.32) got netsplit.
[20:21] taaz (dlehn@66.37.66.32) returned to #gstreamer.
[20:21] Ow3n (ow...@ti...) joined #gstreamer.
[20:21] <omega_> yo
[20:21] <Ow3n> Hi.
[20:22] <Ow3n> How's it going?
[20:22] <Scav> hiya fraser
[20:22] <omega_> Ow3n: there were some discussion about sound servers on =
one of the gtk lists
[20:22] <omega_> I'll fwd you the two bits that referred to gstreamer
[20:22] <Ow3n> Cheers.
[20:22] <Scav> cool, how was the ride back with erik?
[20:22] <Ow3n> Are you still in Norwy?
[20:22] <omega_> me? no
[20:22] <omega_> did you find the Ximian monkeys?
[20:23] <omega_> that I left in the car ;-(
[20:23] <Ow3n> I noticed a couple lying on the back seat. I guess someone=
 else picked them up......
[20:23] <omega_> hmmm
[20:24] Action: omega_ wants them back some day <g>
[20:24] <Ow3n> Maybe the next person to rent the car gets the reminants o=
f dogs and monkeys <G>
[20:24] <omega_> hmmmm
[20:24] <Ow3n> I just got back from skiing on Monday so I havn't had much=
 of a chance to do any hacking :(
[20:25] fons (fo...@ur...) joined #gstreamer.
[20:25] <omega_> heh
[20:25] <omega_> yo
[20:25] <fons> thomas, 't eten is ver gereed
[20:25] <Ow3n> Except for this evening but I'm still gluing together bono=
bo boiler plate code.
[20:25] <fons> drinkte gij water of cola?
[20:25] <thomas> fons: coming down... fruitsap !
[20:25] <thomas> see you guys
[20:25] <omega_> someone (rdj) has managed to do a bonobo-media player al=
ready <g>
[20:25] <omega_> thomas: l8r
[20:25] thomas (th...@ur...) left irc: I'm outta here!
[20:25] fons (fo...@ur...) left irc: fons
[20:25] <Ow3n> The bonobo-media in CVS?
[20:26] <Scav> omega_: using gstreamer?
[20:26] <omega_> yup
[20:26] <Ow3n> Cool. Where is it?
[20:26] <omega_> rdj: ping?
[20:26] <Ow3n> Ah, right. I j didn't notice the (rdj)
[20:26] <omega_> Ow3n: what's your email addr?
[20:26] <omega_> (to fwd the mail in questio)
[20:26] <Ow3n> ow...@di...
[20:27] <omega_> sent
[20:27] <Ow3n> I can hear my mail server disk clicking away...
[20:27] <omega_> there's a lot more to that thread, you should check the =
archives
[20:27] <omega_> the gist of it is that a lot of people aren't pleased wi=
th artsd either
[20:28] <Ow3n> Hmmm... Interesting...
[20:28] <omega_> 40 msgs in thread
[20:29] <Scav> brb
[20:29] <omega_> I just sent you the thread contents, since it's on gnome=
-haackers and I dunno if that's archived
[20:30] Action: Ow3n is hunting through the archives now
[20:30] Action: Ow3n is having trouble finding the archives though :(
[20:30] <omega_> see your mail
[20:31] Action: omega_ hasn't read the thread yet
[20:31] <Ow3n> Wehay. Thanks.
[20:32] Ow3n (ow...@ti...) got netsplit.
[20:33] Ow3n (ow...@ti...) returned to #gstreamer.
[20:33] Nick change: wtay-away -> wtay
[20:33] <wtay> yo
[20:33] <Ow3n> Hi.
[20:33] <hadess> yo Ow3n
[20:34] <Ow3n> yo.
[20:34] <Ow3n> Did you all get back OK from GUADEC?
[20:34] <wtay> hadess: gstreamer slow at startup with the XML registry in=
 place?
[20:34] <wtay> Ow3n: yeah
[20:34] <omega_> yup
[20:35] <omega_> though I'm still very much in the wrong timezone
[20:35] <wtay> CSL (C is hell) 0.1.1 released
[20:35] <omega_> huh?
[20:36] Action: omega_ needs to fix his keyboard soon, the upper left cor=
ner is sticking up at least 5mm from where it should be
[20:36] <hadess> yep
[20:36] <hadess> gotta run
[20:37] Nick change: hadess -> hds-foot
[20:37] <Ow3n> Damn. I've got to go again. My girlfriend just got home an=
d I've got to cook dinner :(
[20:37] <omega_> oops
[20:38] <omega_> cya
[20:38] <Ow3n> Cya. I'll drop by again tomorrow so we can catch up then.
[20:38] Ow3n (ow...@ti...) left #gstreamer.
[20:40] ChiefHighwater (pa...@te...) left irc: Read error to =
ChiefHighwater[temple-baptist.com]: Connection reset by peer
[20:40] ChiefHighwater (pa...@te...) joined #gstreamer.
[20:40] Action: omega_ stares in wonder: http://mindstorms.lego.com/membe=
rs/galleryimage.asp?image=3D72bae2eca6ec50375ba001a93e6787fcimage2.jpg
[20:41] <ChiefHighwater> hehe
[20:41] <omega_> "Woah" - Neo
[20:42] <wtay> very cool
[20:42] Action: wtay could solve the cube in less then 1 minute when he w=
as young :)
[20:42] <omega_> um
[20:43] <omega_> somehow I think this one takes longer
[20:43] <wtay> heh, I was just wondering about that
[20:45] <omega_> 15-25 moves from the solver, each move probably takes up=
 to a minute
[20:45] <omega_> http://mindstorms.lego.com/members/gallery.asp?userid=3D=
72bae2eca6ec50375ba001a93e6787fc#CubeSolver
[20:47] <wtay> ah yes, lubricants are definatly needed..
[20:51] <omega_> the problem with minigtk is that it has the same namespa=
ce as gtk
[20:51] <omega_> which causes serious problems when you're using *real* g=
tk too
[20:52] <wtay> not when it has the same features I guess
[20:52] <omega_> hrm
[21:00] <omega_> I'm renaming to mgtk*
[21:03] <wtay> yeah
[21:04] <omega_> and then the joy of fixing up all the gtk reference in g=
st
[21:05] <omega_> I'll do a half job for now, shimming gtk to mgtk
[21:05] <omega_> er...
[21:05] <omega_> no, that won't work
[21:05] thomas (th...@ur...) joined #gstreamer.
[21:05] <wtay> yo
[21:05] <thomas> hi again
[21:06] <wtay> omega_: you'll need macros
[21:06] <omega_> yup
[21:06] <omega_> and we need to change some stuff
[21:06] <wtay> bleck
[21:06] <omega_> we almost need a branch to do this right
[21:06] <wtay> hmm, that dangerous?
[21:06] <omega_> pervasive, yes
[21:07] <thomas> is it possible to do something like : request pad, conne=
ct, start playing, sleep 5 seconds, request second pad, connect to alread=
y playing bin ?
[21:08] <omega_> that's what incsched is all about
[21:08] <wtay> thomas: I doubt it..
[21:09] <thomas> and what would happen if I disconnect one of the two pad=
s while playing ?
[21:09] <thomas> because the problem I have is not with the eos, it's tha=
t an eos causes the pull not to work and gstreamer hangs, taking all the =
CPU
[21:09] <omega_> you'd have to PAUSE, at least
[21:09] <thomas> well... If I add a buffer after the mixer, a pause would=
n't hurt right ?
[21:09] <thomas> as long as it's short enough
[21:10] <omega_> right
[21:10] <omega_> um, still fuzzy on the pipeline structure..
[21:10] <thomas> omega_: how so ?
[21:10] <wtay> ?
[21:10] <thomas> btw are you forced to use threads when you use buffers ?
[21:11] <omega_> queues?
[21:11] <omega_> currently, yes
[21:11] <wtay> I'd recomend to use threads in your setup to avoid hicups..
[21:11] <omega_> ok, back to original question: request pad from what?
[21:11] <thomas> omega_: from the adder plugin...
[21:12] <thomas> I create a disksrc->volenv bin with a ghost pad...
[21:12] <omega_> ok, then yes you can do that with incsched
[21:12] <thomas> then connect the ghost pad to the requested pad from the=
 adder
[21:12] <thomas> now I'd like to request the second pad a while after the=
 first
[21:12] <thomas> and disconnect pads when the disksrc is eos
[21:12] <omega_> i.e. do a dj mixer app
[21:12] <thomas> yes, basically...
[21:12] <thomas> but computer-controlled, no human interaction
[21:12] <thomas> though that would be possible as well
[21:13] <omega_> that's our major test app, since it stresses a lot of th=
e in-development features
[21:13] <wtay> adder is cothreaded so it'll fail without incsched
[21:13] <thomas> omega_: what is the major test app ? You have something =
like that already ?
[21:13] <omega_> no
[21:13] <omega_> that concept
[21:14] <thomas> wtay: why is it cothreaded ? and what causes it to fail =
then ?
[21:14] <omega_> if you write one, that will be a very useful testcase
[21:14] <thomas> omega_: ah ok.  but it's not under development right now=
 ?
[21:14] Action: Scav is back
[21:14] <thomas> omega_: well, that's the goal.  feed the mixer an xml sc=
ript, containing the input files and envelopes and describing the ways of=
 output
[21:14] <wtay> thomas: cothread setup is currently done only on NULL->REA=
DY state transition
[21:14] <thomas> then feed it to the mixer
[21:14] <Scav> solaris now reports less utilisation of previking than bef=
ore the heavy testing :P
[21:15] <thomas> wtay: so you mean for the adder to work that way, I woul=
d have to set input streams to NULL instead of PAUSE...
[21:15] <omega_> Scav: probably cleaned up some stuff used in libc init o=
r somesuch
[21:15] <thomas> ... which would probably be a bad thing
[21:15] <omega_> thomas: no
[21:15] <omega_> what he's saying is that you can't do this right now, in=
csched has to be merged first
[21:15] <thomas> ah ok... then I'll have to wait for that.
[21:15] <wtay> thomas: going back to NULL will stop all (reset the disksr=
c to offset 0 and such) which is bad
[21:16] <thomas> wtay: what if you know where it was and can rewind to th=
at position ?
[21:16] <omega_> wtay: actually, can you do the branch that abstracts out=
 from gtk ?
[21:16] <wtay> thomas: thats ugly
[21:16] <wtay> omega_: what do you mean?
[21:16] <thomas> wtay: I know.  but I'll take whatever works.  I have to =
prove that gstreamer is the right platform to do it;
[21:16] <thomas> ;)
[21:17] <omega_> the goal is to convert gstreamer over to as close to the=
 glib 2.0 API as possible, and write shims to use minigtk or gtk
[21:17] <omega_> there are a lot of symbols that really shouldn't have GS=
T_* #defines
[21:17] <wtay> omega_: you want me to try it ?
[21:17] <omega_> like GTK_RUN_LAST
[21:17] <omega_> so I'd rather go straight to G_RUN_LAST or whatever it i=
s, and have the shims
[21:18] <omega_> I should be working on incsched....
[21:18] <wtay> thomas: use threads, it'll work with the current scheduler
[21:18] <omega_> wtay: howso?
[21:18] <omega_> he still has to connect up the queue to the adder thread=
, which requires incsched
[21:19] <wtay> omega_: hmm ok, true
[21:19] <thomas> is it too soon to ask what incsched is all about ?
[21:19] <wtay> the new pad is not a src pad...
[21:20] <wtay> thomas: incremental scheduling as opposed to the static on=
e we currently have
[21:20] <omega_> hmm, I just had a 2sec startup of a gst program
[21:20] <thomas> hmmmm... here's another way I could get it to work...
[21:20] <omega_> but subsequent is fast
[21:20] <wtay> thomas: we currently setup the scheduler when going from N=
ULL->READY and never touch it after that
[21:21] <thomas> wtay: I'll wait until you've explained incsched to me ;)
[21:21] <wtay> thomas: the incremental scheduler will calculate the sched=
uling whenever a new element/pad/bin is added to the pipeline
[21:21] <wtay> omega_: too slow?
[21:22] <omega_> wtay: yeah, it was bringing in all the stuff from disk, =
lots of it
[21:22] Action: Scav wonders if hadess has realised man utd are losing 1-=
0 to bayern munich (and now need to score 2 to get through)
[21:22] <omega_> probably because everythig has debug symbols and everyth=
ing
[21:22] <thomas> not to seem too dense, but what does the scheduler do ri=
ght now ?
[21:23] <wtay> thomas: set up pointers and cothread wrappers around each =
element/pad
[21:24] <thomas> wtay: ok... can I throw something at you that I think mi=
ght work now as well ?
[21:24] <wtay> thomas: not really my field of expertise :)
[21:24] <wtay> thomas: ok
[21:24] <thomas> wtay: no, but i mean from an internal gstreamer perspect=
ive
[21:25] <hds-foot> Scav: sod off
[21:25] <thomas> wtay: the problem right now is the on-demand requesting =
of pads and connecting them
[21:25] <wtay> omega_: did it load all the plugins?
[21:25] <omega_> no
[21:25] <thomas> so if I create a few bins, as much as needed max...
[21:25] <thomas> ... and change the input plugins (the disksrc) to keep f=
eeding data as necessary
[21:25] dobey (dobey@141.154.95.104) joined #gstreamer.
[21:25] <thomas> ... with an on/off toggle in the adder input channel...
[21:25] <thomas> ... that should work.
[21:26] <thomas> since the output samples are zero-allocated anyway...
[21:26] <wtay> thomas: only if you can avoid a pull from the added someho=
w
[21:26] <omega_> yeah, create a static pipeline with plugin-specific on/o=
ff switches
[21:26] <wtay> thomas: have a arg in the added somewhere to disable a pad=
 and it'll work
[21:26] <wtay> s/added/adder
[21:26] <Scav> hadess: its 2-0 i see now :P
[21:27] <dobey> la la
[21:27] <hds-foot> Scav: S.O.D. O.F.F.
[21:27] Action: Scav smiles.....i actually want man utd to go thru :P
[21:27] <omega_> hds-foot: what does that stand for??
[21:27] <omega_> hds-foot: or do I want to know?
[21:27] <dobey> hehe
[21:27] <dobey> hey hadess
[21:28] <wtay> yo dobey
[21:28] <hds-foot> hi dobey
[21:28] <dobey> yo
[21:28] Action: Scav wonders if hadess is wearing his cantona shirt tonig=
ht
[21:28] <hds-foot> i'm wearing my manu shirt actually
[21:29] <Scav> aah ok
[21:29] <Scav> are man utd wearing the red home kit or their away kit?
[21:29] Uraeus (csc...@c2...) joined #gstreamer.
[21:30] <Scav> hiya christian
[21:30] <omega_> yo
[21:30] <wtay> hi
[21:30] <Uraeus> hi
[21:30] <omega_> Uraeus: do you have my Ximian monkeys?
[21:30] <hds-foot> Scav: away white kit
[21:30] <Scav> omega_: you want me to send you my spare one?
[21:30] <hds-foot> hey Uraeus
[21:30] <dobey> haha
[21:30] <dobey> monkey!
[21:30] <omega_> Scav: no, I just want to make sure they're still around,=
 and I can get them eventually <g>
[21:31] <omega_> at GUADEC 3 if nothing else
[21:31] <dobey> sigh
[21:31] Action: Scav wonders what omega did with the monkey he undressed =
on the sunday night
[21:31] <omega_> um
[21:31] <hds-foot> omega_: zoophil
[21:31] Action: thomas is away =02- Automatically set away. -=02 messages=
 will be saved.
[21:32] Action: omega_ decided that the t-shirt was way way way too small=
 for me to wear, and put it back on the monkey
[21:32] <dobey> hah
[21:32] <omega_> besides, they gave me one the night before at the party
[21:32] <omega_> of the right size....
[21:32] Action: Scav thinks of using his monkey as a dartboard :)
[21:32] <omega_> pretty small target
[21:33] <Scav> i dont drink so my acvcuracy is pretty high :)
[21:33] Uraeus (csc...@c2...) left irc: Ping timeout f=
or Uraeus[c224s9h5.upc.chello.no]
[21:33] <omega_> hmm, true
[21:33] <Scav> pity about my spelling
[21:33] <wtay> hehe
[21:33] <hds-foot> dobey: neither martha nor peter called me :/
[21:33] Action: wtay has his monkey on top of his monitor
[21:34] Action: omega_ has a large collection of 'things' on top of his m=
onitors: penguin, penguin, several other birds, a troll, a gargoyle...
[21:34] <wtay> cool
[21:34] Uraeus (csc...@c2...) joined #gstreamer.
[21:34] <dobey> troll =3D=3D gem in belly button?
[21:34] <Scav> wb uraeus
[21:34] <Uraeus> crappy cable modem shit
[21:34] Action: wtay doesn't have a penguin :(
[21:35] <Uraeus> omega_: yes I have you penguins
[21:35] <omega_> Uraeus: you missed the discussion about Ximian monkeys
[21:35] <wtay> Uraeus: uhm, not here...
[21:35] <Uraeus> err...monkeys
[21:35] <omega_> paengions?
[21:35] <omega_> er, paenguins?
[21:35] <Uraeus> omega_: mail me your snail mail adress and I send them t=
o you
[21:35] <Scav> wtay: i have loads...collected from linuxworld
[21:35] <omega_> too bad that doesn't extrapolate to maonkeys
[21:35] <Scav> Uraeus: London Zoo, London, United Kingdom
[21:36] <Uraeus> hehe
[21:36] <omega_> Uraeus: not that critical.... we can find some cheaper w=
ay to transfer them, at GUADEC 3 if all else fails
[21:36] matth_ (ma...@qw...) left irc: Read error to matth_[=
qwest.dsplinux.net]: EOF from client
[21:37] <Scav> organise a GST-CONF just to transfer monkeys :)
[21:37] <wtay> Uraeus: put it in a bottle and throw it in the sea :)
[21:38] <omega_> um
[21:38] <Uraeus> heh, anyway new developments on my quest for a gf, my ho=
rse riding girl still doesn't want to go out with me, but her girlfriend =
thinks she should :)
[21:38] <Scav> previking running 120 channels takes less RAM than emacs w=
ith one source file :P
[21:38] <wtay> Uraeus: heheeh
[21:38] <dobey> Uraeus: you suck :-)
[21:38] <hds-foot> Uraeus: stalk her !
[21:39] <wtay> Scav: lol
[21:39] <Uraeus> dobey: I know :)
[21:39] <hds-foot> tomorrow i'm asking the new cutey receptionist out
[21:39] <hds-foot> back to the game
[21:39] <wtay> hds-foot: yes!
[21:39] <hds-foot> bbl
[21:39] <Uraeus> hds-foot: hope you have better luck than me
[21:39] <Scav> hadess: good luck...3 goals needed
[21:39] <hds-foot> Scav: i know, i know, but they can do it =3D)
[21:40] <Scav> Uraeus: plenty of fish to fry, i mean in the sea :P
[21:40] <Uraeus> Scav: true, but when you want a seahorse you get kinda d=
isapointed when you get a salmon
[21:41] <wtay> Uraeus: uhm :)
[21:41] <Scav> salmon...eeww
[21:42] <Scav> plenty of real fish :)
[21:42] <wtay> Uraeus: like in X weights more than Y?
[21:43] <steveb> Scav: what did you think of this? http://www.vovida.org/
[21:43] <Scav> Uraeus: sea horses dont taste that good anyway, salmon tho=
ugh dont look too good, actually taste not so bad :)
[21:43] <Scav> steveb: i know the vovida guys well
[21:43] Uraeus (csc...@c2...) left irc: Ping timeout f=
or Uraeus[c224s9h5.upc.chello.no]
[21:43] <hds-foot> Scav: and here comes 1 !
[21:44] <Scav> I have seen the protocols they have put up there, however =
only recently they have released their softswitch
[21:44] <Scav> hadess: is it 2-1?
[21:44] <steveb> Scav: how would it work with your stuff
[21:44] <steveb> ?
[21:45] <Scav> steveb: we have a SIP driver in the works, so it "should" =
integrate very well with PreVikinh
[21:45] <steveb> cool
[21:46] <Scav> but they could probably use OpenTeleMedia very easily in V=
ocal, for things like music on hold servers and text2speech and speech re=
cognition servers
[21:46] <Scav> ryan giggs just scored....2 more goals needed
[21:47] ajmitch (me...@p4...) left irc: Ping timeout for a=
jmitch[p44-max6.dun.ihug.co.nz]
[21:51] <thomas> wtay: if you have the time, could you try the mixer and =
see if the startup noise I get is really latency or something else entire=
ly ?
[21:51] <Scav> bonobo 1.0.2 released already...
[21:51] <thomas> wtay: also, if I change the output to osssink, the audio=
 is completely screwed
[21:52] <thomas> wtay: don't know why
[21:52] <wtay> thomas: ok, will do that soon
[21:53] <thomas> btw: does anyone know if there's a printable manual of g=
lib ? I only find html based ones
[21:53] <omega_> glib 2.0 ?
[21:53] <thomas> whatever... just something to get me started
[21:54] <omega_> I haven't seen one, but it should be possible to generat=
e it from the tarball
[21:54] <Scav> should be able to generate a pdf from the src ta
[21:54] <Scav> ta=3Dtar
[22:00] <omega_> er...
[22:00] <omega_> when building with --enable-profiling...
[22:00] <omega_> I get a segfault in gststatusautoplug.so
[22:00] <wtay> hum
[22:01] <omega_> um, stack trace is not useful
[22:01] <Scav> that's a nice situation to have :)
[22:01] <omega_> sf0 is ??, sf1 is plugin_load_absolute
[22:03] Action: omega_ will try removing -O from the cmdline during profi=
ling
[22:03] <wtay> it extends from GtkObject...
[22:03] Action: omega_ is on incsched1
[22:04] <wtay> oh
[22:04] <omega_> no change but -pg
[22:05] <omega_> removing -o6 fixed it
[22:05] <wtay> hmm I see a problem
[22:05] <wtay> hmm, no I dont..
[22:06] <wtay> so -O6 makes it crash on plugin_init?
[22:06] <omega_> yup
[22:06] <omega_> or somewhere
[22:06] Action: omega_ thinks it's an ld.so problem
[22:07] <wtay> probably, plugin_init doesn't do a lot..
[22:08] <thomas> wtay: what extension should an mpeg video have for the a=
utoplug to work ?
[22:08] <omega_> extension should be irrelevant, typefind should do the w=
ork
[22:08] <thomas> wtay: it doesn't work when used in either the mixer or h=
elloworld since it tries the mad decoder
[22:08] <taaz> heh... cool, sf using viewcvs now.  has the neato annotate=
d code view so we know who last touched what source lines
[22:09] <wtay> thomas: so?
[22:09] <wtay> thomas: the mad decoder works for mpeg system streams too
[22:10] <thomas> wtay: oh... well it didn't.  I'll check why.
[22:10] Action: dobey waits for something that can play stuff on ppc good=
 that doesn't suck
[22:10] <wtay> thomas: it wont work with mpeg system streams 'cause the d=
emuxer dynamically creates pads...
[22:10] matth (ma...@qw...) joined #gstreamer.
[22:10] Nick change: matth -> matth_
[22:10] <taaz> wtay: i'm confused by that buffer pool update to mpeg2dec =
plugin
[22:11] <wtay> taaz: ?
[22:11] <taaz> wtay: is it supposed to get rid of the mem allocation too?
[22:11] <wtay> taaz: well, it only works for RGB images...
[22:11] <wtay> taaz: not always, only when a bufferpool can be obtained f=
rom the peer plugin
[22:12] <taaz> that code is always going to do the g_new()
[22:12] <taaz> which is the evil part
[22:12] <thomas> wtay: here's the errors I get :
[22:12] <thomas> ***** GStreamer ERROR ***** in file gstscheduler.c at gs=
t_bin_schedule_func:431
[22:12] <thomas> Element: /bin/pipeline/autoplug_bin/mad.sink
[22:12] <thomas> Error: peer is null!
[22:12] <wtay> thomas: exactly
[22:12] <thomas> wtay: how so ?
[22:13] <wtay> thomas: the mpeg system demuxer creates dynamic pads, so m=
ad isn't connected=20
[22:13] <thomas> ah ok... so you need the autoplug renderer then ?
[22:13] <wtay> thomas: incsched
[22:14] <thomas> wtay: it's starting to sound like god ;)
[22:14] <wtay> taaz: line 240 in gstmpeg2dec.c is crucial
[22:14] <wtay> thomas: it is :)
[22:14] <thomas> wtay: I'll wait for it then...
[22:14] <thomas> omega_: i'll try not to bother you too much with beginne=
r's question so you can get it finished ;)
[22:14] <omega_> hehe
[22:15] <wtay> thomas: ask me :)
[22:15] <thomas> wtay: I don't know what to ask anymore ;)
[22:16] <thomas> wtay: or maybe this... is copying an element like disksr=
c about the same kind of work as copying a plugin ?
[22:16] <taaz> wtay: line 288 gets the buffer but 300+303 will alloc new =
memory for that buffer anyway.  that seems like the expensive part
[22:16] steveb (st...@no...) left irc: [x]chat
[22:16] <wtay> taaz: oops :)
[22:16] <wtay> thomas: define copying?
[22:17] <thomas> well... cheating by looking at it.
[22:17] <thomas> like I did for passhtrough based on stereo2mono
[22:17] <wtay> thomas: oh ok, yeah
[22:17] <wtay> taaz: that's a bug :)
[22:17] <taaz> wtay: heh.. ok.
[22:18] <wtay> taaz: hmm, that might explain why it didn't work :)
[22:18] <wtay> doh
[22:18] <thomas> wtay: the helloworld2 not working, is that something I c=
ould fix ?
[22:19] <taaz> i haven't had a chance to even try the code... was just lo=
oking at a diff to see what you changed and got confused
[22:19] <wtay> thomas: yes, it's very similar to what you did
[22:19] <dobey> la la la
[22:19] <thomas> wtay: but with threads ?
[22:20] <wtay> thomas: you decide :)
[22:20] <thomas> wtay: I'll save it for tomorrow.  get some decent sleep =
tonight first.
[22:20] <wtay> thomas: tests/autoplug3.c has the code
[22:20] <wtay> thomas: together with you typefind code
[22:21] <wtay> taaz: I'm trying to fix it now
[22:21] <thomas> wtay: the typefind isn't mine, I copied it from autoplug
[22:22] <taaz> by the way... what is the best way to check for performanc=
e gains?  ie, killing off sync to do timing or whatever
[22:22] Nick change: omega_ -> omega-lunch
[22:22] <wtay> taaz: yeah
[22:25] <thomas> wtay: is someone working on a filter factory ?
[22:25] <thomas> wtay: and should it be in c
[22:25] <wtay> thomas: no idea... omega-lunch?
[22:27] <omega-lunch> no one is working on one right now, afaik
[22:27] <omega-lunch> but please please post any specific ideas on syntax=
 to hte list
[22:27] <wtay> omega-lunch: any idea how it will work?
[22:28] <omega-lunch> cpp gone mad
[22:28] <wtay> hehe, right
[22:28] <wtay> taaz: yes!
[22:29] <taaz> ?
[22:29] <wtay> taaz: now it works very fast
[22:29] <taaz> excellent!
[22:29] <wtay> taaz: retrying with Xv images
[22:30] <hds-foot> F.U.C.K
[22:30] Nick change: hds-foot -> hadess
[22:30] <dobey> heh
[22:31] <hadess> any german approaching near me will fear my anger
[22:31] <dobey> hahah
[22:32] Action: dobey sends hadess to #kde
[22:32] <hadess> second fucking year in a fucking row
[22:32] Action: omega-lunch watches hadess having a cow
[22:33] <thomas> wtay: if you have time, mail me how the mixer works for =
you.
[22:33] <thomas> wtay: i'm off now.
[22:33] <Scav> hadess: you lost?
[22:33] <wtay> thomas: I tried it, it works
[22:33] <omega-lunch> thomas: l8r
[22:33] <wtay> cya
[22:33] Action: Scav is wearing michael schumacher's cap.....
[22:33] <hadess> Scav: that was very very tense at the end
[22:33] <thomas> wtay: ok thanks... no problems with latency ? clean audi=
o ?
[22:34] <wtay> thomas: not realy
[22:34] <thomas> wtay: ok, that's good.
[22:34] <thomas> later...
[22:34] thomas (th...@ur...) left irc: I'm outta here!
[22:34] <Scav> hadess: well thats life, now you can concentrate on winnin=
g the premiership (oh whoops, i forgot man utd have already done that)
[22:35] <Scav> hows it feel to see leeds and liverpool in semi finals of =
european competition and man utd not?
[22:35] <omega-lunch> Scav: be nice now...
[22:36] matth_ (ma...@qw...) left irc: Read error to matth_[=
qwest.dsplinux.net]: EOF from client
[22:37] <Scav> omega-lunch: yah well man utd are the best team in the cou=
ntry and have been for the last 8 years....so they can handle the flak :P
[22:40] matth (ma...@qw...) joined #gstreamer.
[22:40] Nick change: matth -> matth_
[22:40] <omega-lunch> yo
[22:42] <matth_> howdy... just rebuilt my kernel a few times (got 2.4.3, =
tempting the ide dma gods)...
[22:42] <matth_> how did that gtk object replacement work out?
[22:42] <omega-lunch> other problems still
[22:42] <matth_> ?
[22:43] <omega-lunch> need to do the same think to patch that in that we =
need to do for glib2
[22:43] <omega-lunch> need to fix up gst to use the new glib2 api, and sh=
im in the minigtk stuff
[22:43] <matth_> how different is glib2?  where are the differences?
[22:43] <wtay> brb
[22:44] <omega-lunch> the API is similar, so it shouldn't be a problem, a=
t least for the parts of it that we use
[22:44] <omega-lunch> but I don't want to go overboard setting up GST_* m=
acros for all the random junk in gobject, like the marshallers and such
[22:44] <omega-lunch> that'd be insane
[22:46] <matth_> hmm... i thought the minigtk and glib2 approaches were m=
utually exclusive
[22:46] <matth_> ?
[22:46] <omega-lunch> not really
[22:46] <omega-lunch> the gtk and minigtk api are the same
[22:46] <matth_> yup
[22:46] <omega-lunch> this has other problems when you actually want to u=
se the real gtk as well
[22:46] <matth_> yup
[22:46] <omega-lunch> massive symbol conflicts
[22:46] <omega-lunch> so I changed to mgtk
[22:47] <omega-lunch> but then I still have to shim the gst usage over to=
 mgtk
[22:47] <omega-lunch> I'd rather shim it over to g_object_ and shim minig=
tk into that namespace
[22:47] <omega-lunch> then we can build on gtk, minigtk, or glib2
[22:47] <matth_> ahh
[22:47] <omega-lunch> because for what we do with it, there isn't much di=
fference except namespace between gtkobject and gobject
[22:47] <omega-lunch> afaict
[22:48] <matth_> in next version of gtk (that uses glib2) will gnome chan=
ge gtk_object to gobject?
[22:48] <omega-lunch> yes
[22:48] <matth_> for regular gtk stuff?
[22:48] <omega-lunch> afaik
[22:49] <omega-lunch> actually, no
[22:49] <matth_> then won't you face the same namespace problem?
[22:49] <omega-lunch> they're keeping gtkobject as a shim, like our gstob=
ject
[22:49] <matth_> ahh...
[22:49] <matth_> then we should be okay
[22:49] <matth_> just have to do it
[22:50] <matth_> :-)
[22:50] <omega-lunch> yup
[22:50] <omega-lunch> not too hard either
[22:50] <matth_> yup
[22:50] <omega-lunch> just look at glib2, port to it, and make sure that =
gtk can still be shimmed in
[22:51] <matth_> should be able to since that's what gnome is doing
[22:52] <omega-lunch> they're not shimming, afaik
[22:52] <matth_> ?  what did you mean when you said "they're keeping gtko=
bject as a shim"?
[22:53] maYam_bath (ma...@ca...) joined #gs=
treamer.
[22:53] <omega-lunch> just like gstobject is a small layer on top of gtko=
bject, gtkobject in gtk2 will be a thin layer over gobject
[22:54] <maYam_bath> hello
[22:54] <ChiefHighwater> ello
[22:54] <omega-lunch> yo
[22:54] <maYam_bath> :)
[22:54] Nick change: maYam_bath -> maYam
[22:54] Action: omega-lunch hands maYam a towel..
[22:54] <ChiefHighwater> was begining to wonder about your waterproof key=
board
[22:54] <omega-lunch> hehe
[22:55] <maYam> thanks omega, but i skipped the bath :)
[22:55] <maYam> *everyone run!*
[22:55] <omega-lunch> heh
[22:56] <wtay> yo maYam
[22:56] Nick change: omega-lunch -> omega_
[22:56] <maYam> ChiefHighwater: yes, baths are hell..  you can do absolut=
ely everything sitting at your comp, except for taking a bath..
[22:56] Action: dobey hands mayam some non-evil
[22:56] <maYam> they should invent something for that
[22:56] <wtay> woohoo! XvImages in a bufferpool!
[22:56] <maYam> dobey, my loving friend!
[22:56] <maYam> hey wtay ;)
[22:57] <omega_> matth_: called voice recognition..
[22:57] <omega_> er, maYam: called voice recognition
[22:57] <omega_> <grumble> stupid completion
[22:57] <maYam> omega_: ofcourse!  tisk..
[22:57] <dobey> haha
[22:57] <omega_> bash completion won't complete unless the prefix is uniq=
ue, xchat completes the first match. stupid.
[22:58] <maYam> good old lernout and hauspie ;)
[22:58] Action: omega_ search his brain for that reference....
[22:58] <omega_> EAGAIN
[22:59] Action: dobey gives mayam a hug
[22:59] <wtay> a belgian voice recognition company that is on the verge o=
f bankrupsy :=E0
[22:59] <omega_> ah
[23:00] <maYam> dobey: you sure you wanna do that?  i skipped my bath, yo=
u now ;)
[23:00] <wtay> bought by M$
[23:00] <maYam> but thanks anyway :)
[23:00] <dobey> heh
[23:01] <wtay> hmm 50% CPU usage on the THV vob...
[23:01] <omega_> compared to.. ?
[23:01] <wtay> s/THV/THX
[23:01] <wtay> omega_: sec..
[23:02] <wtay> 55%
[23:02] <wtay> bah
[23:02] <omega_> that's actually a 10% decrease
[23:03] <omega_> or 9.1111%
[23:03] <wtay> omega_: hmm I need better a better measurement to be sure.=
..
[23:03] <omega_> init 1
[23:03] <omega_> <g>
[23:04] <wtay> heh
[23:04] Action: omega_ needs to put this new drive in asap
[23:04] Action: omega_ has 60MB free, needs the extra 8gb
[23:04] <taaz> you need to kill off any sync and check the time to decode=
 X number of frames
[23:04] <omega_> it's not decode, it's transfer to X and display
[23:04] <omega_> but yes, kill sync
[23:05] <wtay> I'm not really using a Xv Pool either...
[23:05] <wtay> I create/destory them for now..
[23:06] <wtay> let's try a real pool
[23:21] <omega_> taaz: is walken around?
[23:23] <wtay> down to 30% but awfull images...
[23:23] <omega_> 30% what?
[23:23] <wtay> CPU usage with a real pool of xvimages
[23:24] <omega_> ooo
[23:24] <wtay> but it looks like crap
[23:24] <omega_> how would that make a difference?
[23:24] <wtay> somehow the images are mixed up...
[23:24] <omega_> oh, are you using the pools as mpeg reference images too=
?
[23:24] <omega_> xv might not want to cooperate there
[23:25] <wtay> not sure...
[23:25] <wtay> I hand mpeg2dec 3 XvImages
[23:25] <omega_> xv might not keep those three distinct
[23:25] <wtay> but my first attempt did work
[23:26] <wtay> that was when I created a new Xv image every time and dest=
royed it when it was rendered
[23:26] <omega_> yup
[23:27] <wtay> oh, stupid me
[23:27] <omega_> eh?
[23:27] <wtay> I insert the XvImage in front of the GSlist
[23:27] <omega_> which?
[23:27] <wtay> and I take a new Xvimage from the front
[23:27] <omega_> oh
[23:27] Action: omega_ hands wtay a clue <g>
[23:28] Action: wtay grabs the BPB
[23:29] <wtay> wow, 30% perfect
[23:29] <omega_> 85% speedup!
[23:29] Action: omega_ bows before wtay
[23:30] <wtay> just one memcpy and a few xvimage allocs !
[23:30] <omega_> told ya
[23:30] <wtay> about the same as gvlc now
[23:30] <omega_> gvlc ?
[23:30] <omega_> ah, right
[23:31] <wtay> wow, I'm amazed
[23:31] Action: omega_ knows the power of the cache
[23:31] <wtay> yeah
[23:32] <wtay> this also means the bufferpool concept works
[23:32] <omega_> cool!
[23:32] <wtay> although the API sucks :)
[23:32] <omega_> hmm
[23:32] <omega_> don't forget to include a timestamp in a buffer request =
API
[23:32] <wtay> I only get the destroy callback *after* the buffer has bee=
n destroyed
[23:33] <omega_> oops
[23:33] <wtay> no way to store the complete buffer
[23:33] <wtay> oh well, easy fix
[23:33] <hadess> gone for bed
[23:33] <hadess> cya all
[23:33] <wtay> cya
[23:33] <maYam> bye hadess
[23:33] <hadess> hello, and good bye maYam
[23:34] hadess (ha...@pc...) left irc: sleep
[23:34] Action: wtay thanks taaz for pointing him to an obvious bug
[23:40] <Scav> i am off home now
[23:40] <Scav> cya guys
[23:40] <wtay> cya Scav
[23:40] <Scav> had enough of previking coding for one day :)
[23:41] Scav (za...@za...) left irc: off
[23:43] <dobey> omega: mayam thinks we are related or something :/
[23:44] <omega_> howso?
[23:44] <wtay> yuv2rgb_mmx at 65% CPU
[23:44] <omega_> ouch
[23:44] Action: omega_ is actually hacking on libcodec right now to clean=
 up a few issues, maybe I can get yuv to not suck
[23:45] Action: ChiefHighwater thinks that osunds like a good idea
[23:45] <wtay> compared to 90% CPU without bufpools
[23:45] <omega_> dobey: I see what she means
[23:45] <omega_> modulo the car <g>
[23:46] <dobey> heh
[23:47] <wtay> maYam hates cars :)
[23:48] <omega_> wtay: hrm, ya think maybe she sabotaged your car to get =
you to fly to GUADEC? <g>
[23:48] <wtay> omega_: oh yes :)
[23:48] <omega_> "wtay, I hear something, the car sounds funny, let's tur=
n back and fly instead" - maYam
[23:49] <dobey> heh
[23:49] <wtay> almost.. :)
[23:49] <taaz> omega_: walken is not around now... he doesnt show up ofte=
n anymore.  vmware ate him.
[23:50] <omega_> taaz: suck
[23:50] <taaz> i'm sure he will respond to email though
[23:57] <wtay> any links to gtk/arts/gstreamer discussions on mailing lis=
ts?
[23:58] <omega_> nope
[23:58] Action: omega_ has used 44 of his 200 CD-R's so far
[23:59] <ChiefHighwater> 42 for pics off of his coolpix camera 8-]
[23:59] <omega_> not quite
[23:59] <omega_> though I need to burn another ASAP before I need to burn=
 2
[23:59] <ChiefHighwater> hehe
[00:00] --- Thu Apr 19 2001
[00:01] <dobey> heh
[00:14] <omega_> on /., some 'indrema vetrans' are supposedly creating Tu=
xBox, which is supposed to be what Indrema was supposed to be
[00:14] <omega_> we should contact them and get them interested in GStrea=
mer
[00:15] <wtay> what does it do?
[00:15] <omega_> indrema's gaming console?
[00:16] Action: wtay is reading the gnome-hackers mailing list about soun=
ds servers
[00:16] <wtay> ah a gaming console
[00:16] <omega_> indrema was one of the founders of that ill-fated and de=
ad open media framework thingie
[00:16] <wtay> oh
[00:17] <wtay> never heard of it..
[00:17] <dobey> hrh
[00:17] <omega_> hrm, doesn't surprise me <g>
[00:18] <omega_> doh
[00:18] <omega_> er, wrong window
[00:28] Action: omega_ listens to the 2-house distant sounds of an Indian=
 Pow-wow
[00:36] <ChiefHighwater> hrmm...if they offer, don't smokem peace pipe
[00:36] <omega_> I'll keep that in mind <g>
[00:37] Action: omega_ is an idiot
[00:39] Action: wtay is falling asleep..
[00:45] <wtay> I'm going to sleep.. cya all
[00:46] <omega_> ok, l8r
[00:46] Nick change: wtay -> wtay-sleeping
[00:56] Action: taaz wants to go home an play with faster mpeg2dec plugin=
...
[00:56] <omega_> hehehe
[00:57] <taaz> i really need to put some time into this oms thing
[00:58] <omega_> oms/gstreamer thing?
[00:58] <taaz> i would be amused (and not that suprised) if it would be f=
aster using gstreamer than what oms does now
[00:58] <taaz> yup
[00:58] <taaz> there is some scary buffer locking crap in there now
[00:59] <omega_> yeah
[00:59] <taaz> and no sync
[00:59] <taaz> any idea on how to deal with QoS issues?
[00:59] <omega_> well, /me thinks mpeg2dec needs a rev'd API
[00:59] <taaz> heh... talk to mr walken
[00:59] <omega_> yeah, but it seems that mr walken is, um, distracted
[00:59] <taaz> or better yet, accompany your suggestions with a patch
[01:00] <omega_> yeah, well, it needs discussion
[01:01] <omega_> libdv, mad, and mpeg2dec (and hopefully my own libmpeg, =
and ac3dec) need to all kinda have the same style of API, IMO
[01:02] <taaz> yup.  let us know when you're done ok? ;)
[01:02] <omega_> heh
[01:03] <taaz> ac3dec has lots of issues... cvs code has some hacked toge=
ther floating point stuff.  we're not sure how good it is, but its faster=
 than the old int version, but that version may be just fine if someone p=
ut the time into some optimization
[01:03] <omega_> someone has to understand ac3....
[01:03] <taaz> walken was going to take over ac3dec but hasn't gotten aro=
und to it yet
[01:03] <omega_> is Aaron still maintaining it?
[01:03] <taaz> aaron is mia
[01:04] <omega_> ah, and walken is sorta mia too
[01:04] <taaz> i need to email him... i might want to crash on aarons flo=
or during OLS ;)
[01:04] <omega_> doh
[01:04] <omega_> when is ols?
[01:04] <taaz> end of july i think...
[01:04] <omega_> ok, plenty of time
[01:07] sienap (sy...@ip...) joined #gstreamer.
[01:08] <sienap> hi all
[01:08] <sienap> Good evening or do i have to say morning Doctor maYam
[01:09] <taaz> what evil person decided that there had to be 'int' and 'l=
ong' and no standard on their size?
[01:09] <omega_> K&R
[01:09] Action: taaz throws a rock at K&R
[01:09] <omega_> Subject: Gnome-aRts-0.1.1 Release
[01:09] Action: omega_ needs to get that arts plugin working...
[01:10] <sienap> wtay asleep maYam idle.. i know  anough..
[01:10] <taaz> omega_ needs to get incsched working ;)
[01:10] <omega_> yeah, yea
[01:13] Action: taaz thinks gstreamer should be a part of tuxbox
[01:14] <omega_> yes
[01:22] sienap (sy...@ip...) left irc: BOe
[01:22] <dobey> oh
[01:22] <dobey> damn
[01:26] <taaz> omega_: are you poluting the source tree with C++?
[01:26] Action: omega_ hides
[01:27] <taaz> now you're going to have to add C++ checks to configure.in=
 and mask off that dir if you don't have it...
[01:27] <omega_> I already did
[01:27] <omega_> and it doesn't build automatically yet anyway
[01:28] <omega_> am searching for how to make it depend on presence of c+=
+
[01:29] <omega_> is it just me or is the gnome help browser a joke?
[01:30] Action: dobey wonders where ajmitch is
[01:30] <dobey> what's wrong with ghb?
[01:30] <omega_> ghb?
[01:31] <omega_> it's pathetic
[01:31] <dobey> why?
[01:31] <dobey> <- has a particular interest
[01:31] <omega_> navigation is substandard, you can't search for any text
[01:31] <omega_> etc.
[01:32] <dobey> oh
[01:33] <taaz> what are you looking for?
[01:34] Action: omega_ wants to search through autoconf macros for cxx or=
 cplusplus
[01:34] <dobey> he wants something like the window help browser or someth=
ing
[01:34] <omega_> but /me thinks he has what he needs
[01:34] <omega_> not exactly
[01:34] <dobey> heh
[01:34] Action: dobey hides
[01:34] <taaz> "info autoconf"
[01:34] <omega_> yeah, but I hate info
[01:34] <taaz> oh poor baby
[01:35] <dobey> info:autoconf
[01:35] <dobey> heh
[01:35] <omega_> hrm, AC_CHECK_LIB of a C++ lib
[01:35] <omega_> what kind of symbol do I give it??
[01:36] <omega_> I think I found a Cish symbol I can use
[01:39] <taaz> i'm confused.. you just trying to check if c++ is found?
[01:39] <omega_> yes
[01:39] <taaz> http://crib.lehn.org:8080/cgi-bin/info2www?(/usr/share/inf=
o/autoconf.info.gz)Particular+Programs
[01:39] <taaz> AC_PROG_CXX docs may help
[01:39] <omega_> yes, but it does not make it obvious what the result of =
that macro is
[01:40] <omega_> afaict, either CXX is defined or it's not
[01:41] <taaz> hmm.. i dunno.  i'd just look at some c++ project and see =
what it does
[01:41] <taaz> though i guess that's different
[01:41] <omega_> some c++ project will *assume* there's a c++ compiler
[01:42] <taaz> actually... what i'd do is not use c++ ;)
[01:42] <omega_> no choice if we want to get an arts plugin
[01:42] <taaz> maybe if you ask nicely they will convert arts to C
[01:42] <omega_> hah
[01:42] dobey (dobey@141.154.95.104) left #gstreamer (eh).
[01:42] <omega_> too bad you were't at guadec
[01:43] <omega_> you could have asked him yourself <g>
[01:43] <taaz> i think not... i just chose to not use such things.  damni=
t why don't people use Objective-C?
[01:43] <omega_> http://gstreamer.net/guadec-pics/dscn0005.jpg
[01:44] Action: omega_ needs to suggest a few things to stw to make using=
 arts outside of arts not suck
[01:44] <omega_> i.e. he puts headers in /usr/include/arts/*
[01:45] <omega_> such as arts/artsflow.h
[01:45] <omega_> then in his code refers to them as <artsflow.h>
[01:45] <omega_> that requires you have -I/usr/include/arts on your cmdli=
ne
[01:45] <omega_> which is wrong
[01:45] <omega_> the code should refer to <arts/artsflow.h>
[01:46] <omega_> grr, and AC_CHECK_LIB doesn't understand c++
[01:46] <taaz> AC_LANG_CPLUSPLUS is what you're supposed to use...
[01:46] <omega_> yeah, and AC_LANG_RESTORE
[01:46] <omega_> just found it
[01:47] <omega_> why is there cxx and cplusplus???
[01:47] <taaz> to provide more confirmation that c++ sucks
[01:47] <omega_> good point
[01:47] CHW (pa...@te...) joined #gstreamer.
[01:47] ChiefHighwater (pa...@te...) left irc: Read error to =
ChiefHighwater[temple-baptist.com]: Connection reset by peer
[01:48] Action: omega_ is going to put an rm -f config.cache at the top o=
f autogen.sh
[01:58] taaz_ (dlehn@66.37.66.32) joined #gstreamer.
[01:58] <omega_> um
[01:58] taaz (dlehn@66.37.66.32) left irc: Ping timeout for taaz[66.37.66=
.32]
[01:59] <omega_> hrm, two in a row
[01:59] Nick change: taaz_ -> taaz
[02:02] <taaz> ok, i don't need to knwo how seeking is going, but is it e=
asy to jump around in streams?  i'm not sure how to do that.  its somethi=
ng that needs to be done for OMS assimilation
[02:03] <omega_> right now you simply provide a new offset to the...
 
[truncated message content]