[gst-devel] Daily IRC logs

wim.taymans at chello.be wim.taymans at chello.be
Wed May 16 06:28:14 CEST 2001

[06:35] <omega_> wow.  47 days gives 2.6MB of IRC logs
[06:36] <omega_> about 1K lines/day
[06:37] <omega_> is linuxvideo.org down?
[06:38] <taaz> yes
[06:38] <omega_> ok
[06:38] <omega_> eta back up?
[06:39] <taaz> dunno
[06:39] <taaz> was fixed this morning but went down again...
[06:39] <omega_> neat
[06:39] <taaz> july 2 it moves to a realy isp apprently
[06:39] <taaz> s/realy/real
[06:39] <omega_> where is it now?
[06:39] <taaz> i have no idea
[06:40] <taaz> one of the convergence guys has physical access but its like in a different town than he lives or something
[07:53] steveb (steveb at node1ee03.a2000.nl) joined #gstreamer.
[07:54] <omega_> yo
[07:54] <steveb> hi
[07:54] <omega_> bleck
[07:54] <ajmitch> hi
[07:54] <omega_> RME goes and makes neat hardware
[07:54] <ajmitch> RME?
[07:54] <omega_> and then they go and build even neater hardware using firewire cables and such, but don't use the 1394 bus protocol!!!
[07:54] <ajmitch> hehe
[07:54] <omega_> so the interface card is $300++
[07:54] <omega_> that's really stupid, guys
[07:55] <steveb> do you think it is worth having a policy of no interleaved audio channels with float type?
[07:56] <omega_> possibly
[07:56] <steveb> that makes intfloatconvert more comelete than I thought!
[07:56] <omega_> heh
[08:17] <steveb> its off to work we go
[08:17] steveb (steveb at node1ee03.a2000.nl) left irc: [x]chat
[08:18] walken (foobar at c1583255-a.smateo1.sfba.home.com) joined #gstreamer.
[08:25] Nick change: taaz -> taazzzz
[09:13] Nick change: ajmitch -> ajbusy
[09:21] steveb (stevebaker at 212186169160.chello.com) joined #gstreamer.
[10:15] Nick change: ajbusy -> ajmitch
[10:15] Action: omega_ started using galeon, and discovered GTM
[10:15] Action: omega_ is now writing a really long message listing about 50 or so things that could be changed/improved/added to GTM
[10:16] Action: omega_ is going to copy the gnome usability list on this mail
[10:16] <walken> whats gtm ?
[10:16] <omega_> gnome transfer manager
[10:16] <omega_> a download manager
[10:16] Action: walken upgraded his X today
[10:16] Action: walken happy
[10:16] <omega_> 4.0.3 ?
[10:17] Action: walken drawing funny diagrams about the ac3 bit allocation algorithms
[10:17] <walken> yeah 4.0.3
[10:17] <walken> I know I'm late :)
[10:17] <omega_> I grabbed the rh7.1 rpms and installed them on my rh7.0 machine <g>
[10:17] <omega_> then still had to go to GATOS to get Xv working
[10:17] <walken> the g450 driver is improving
[10:18] <omega_> problem is that Xv still sucks
[10:18] <walken> now it only needs the matrox hal in .so form, before it needed to remove the whole driver and install the matrox one
[10:18] <walken> you have what video ?
[10:18] <omega_> my laptop has a Rage Mobility M2
[10:18] <walken> in 4.0.3 xv is still not accurate enough horizontally
[10:18] <walken> (on matrox cards)
[10:19] <omega_> hrm, not even worried about that
[10:19] <omega_> the API is a joke
[10:19] <walken> cant do much about it though
[10:19] Action: omega_ used DirectDraw for 3mo, it's overkill, but at least it has all the features
[10:19] <omega_> walken: why not?  Xv isn't set in stone
[10:19] <walken> oh ?
[10:19] <omega_> it's about as solid as quicksand anyway <g>
[10:20] <omega_> all else fails we ditch Xv and make a new extension that doesn't suck
[10:20] <omega_> but that means getting hardware drivers updated ;-(
[10:20] <walken> what do you think is missing
[10:20] <walken> I for one hate the 4cc thing
[10:20] <omega_> ah, about a dozen features
[10:20] <omega_> fourcc is how it is, down to the hardware
[10:21] <walken> 4cc in hardware ????????????????????
[10:21] <omega_> yes
[10:21] <walken> yuck
[10:21] <omega_> what did you expect?
[10:21] <walken> something reasonable ?
[10:21] <omega_> '4cc in hardware' *is* what Xv is
[10:21] <walken> well no
[10:21] <omega_> that's all it does
[10:22] <walken> 4cc as in passing 4-char strings to inaccurately describe what you want
[10:22] <omega_> yes, but as I said, that's how the hardware does it
[10:22] <omega_> if you look at the G400 ref docs, they have #defines with the fourcc's in them
[10:22] <omega_> the chip itself keys on those constants
[10:22] <walken> like you say "yuv 420" but you're not allowed to say what exact colorspace it is and where the chroma pixels  are relative to the luma pixels and stuff
[10:23] <omega_> exactly
[10:23] <omega_> of the ~3 chipset docs I've looked at, none of the hardware even had a concept of that
[10:23] <walken> well, 4cc in hardware sucks
[10:23] <omega_> ATI may be different, but there are no docs for that
[10:23] <walken> i thought 4cc was just a crappy api
[10:24] <ajmitch> hmm, i hate computer professionals that are clueless ;)
[10:24] <omega_> regardless, that's how it all works, it's the other areas that Xv sucks horribly
[10:24] <omega_> ajmitch: ?
[10:24] Action: ajmitch just got a call from someone who thought that the only way to get around an 'invalid system disk' erro was to wipe the drive & reload windows
[10:24] <omega_> isn't it? <g>
[10:24] <omega_> wait.  reload window?  no.  install Linux.
[10:25] <ajmitch> well, i'd put linux on it (just so that i could partition the hard drive)
[10:25] <ajmitch> windows couldn't handle a 12GB hard drive ;)
[10:25] <omega_> walken: have to looked at the <chip>_video.c files?
[10:25] <ajmitch> unfortunately linux probably wouldn't be an option for these people
[10:26] <walken> where ?
[10:26] <omega_> in xfree drivers
[10:26] <ajmitch> cos i was the only one in the town who knew much about it, and i left ;)
[10:26] <omega_> ajmitch: oops.  your mistake was giving them your email addr or phone # before leaving
[10:26] <walken> not much
[10:27] <omega_> walken: here's the thing: *every* one of those does a line-by-line memcpy of every single frame going through Xv
[10:27] <ajmitch> omega_: ouch?
[10:27] <walken> oh yes, I had noticed that
[10:27] <omega_> the X hardware driver is responsible for a processor-driven copy of every pixel of data
[10:27] <omega_> now... this is what AGP is for.
[10:27] <walken> yes
[10:27] <walken> definitely
[10:27] <ajmitch> do any X drivers use AGP? ;)
[10:27] <omega_> a) allocate system RAM in the AGP apperture
[10:27] <omega_> b) application writes data to this region
[10:28] <omega_> c) driver tells chip to DMA it into video memory
[10:28] <walken> even aaron's mga driver is quite faster than X though it still does a silly memcpy
[10:28] <omega_> d) driver tells chip to display that video memory via the overlay engine
[10:28] <walken> do you know how to use the agpgart stuff at all ?
[10:28] <omega_> at least mga claims to support triple buffering
[10:28] <walken> I'd like to play with it sometime
[10:28] <omega_> no, I don't
[10:28] <omega_> it requires kernel memory region support
[10:28] <walken> nobody does
[10:28] <ajmitch> omega_: gonna write berlin support into gstreamer? ;)
[10:28] <omega_> which the SGI people have done
[10:28] <omega_> ajmitch: ummm
[10:29] <omega_> the region support is the same that's required to support big CC-NUMA machines
[10:29] <omega_> then a user-space API is needed to interface the X driver to the kernel region support
[10:29] <omega_> then XShm to get that to the application
[10:29] <omega_> and you need to support multiple buffers in the same logical overlay
[10:29] <omega_> and Xv just fails miserably at all that
[10:30] <walken> well
[10:30] <walken> xshm sucks too because the app allocs the memory and passes it to the X server, not the opposite
[10:30] <omega_> of all the stuff the XFree people have done recently, that actually surprise me that it was so poorly done
[10:30] <walken> so for agp stuff it doesnt help at all
[10:30] <omega_> ok, then that's yet another problem
[10:30] <omega_> how does DRI do it?
[10:31] <walken> no idea
[10:31] <omega_> DRI must have that problem solved, since they should be using the AGP GART as well to transfer their textures, as well as triangles
[10:31] <omega_> either that, or DRI sucks too
[10:31] <omega_> I know they plan it with DRI, but I got the impression that the AGP transfers were still very very experimental
[10:32] <walken> well I guess they dont have anything to do with xshm so they should be fine :)
[10:32] <omega_> but then they must have another way of doing it, if at all
[10:33] <omega_> some of it has to do with the fact that the textures can be stored in X pixmaps, which means that the application then has to worry about filling system RAM, and paging them in and out of X space
[10:33] <omega_> which would mean a read from disk, and a copy into X space for every re-load
[10:33] <omega_>  ;-(
[10:33] <omega_> even....
[10:33] <omega_> for some chips it's possible for the overlay engine to read from the apperture....
[10:33] <omega_> which means that even the DMA copy isn't required
[10:34] <omega_> just write to system ram, then tell it to read from system ram
[10:34] <omega_> I think the G4xx can do this
[10:35] <walken> i called aaron yesterday
[10:35] <walken> he said good luck if I want to use the agp stuff
[10:35] <omega_> he's still alive?
[10:35] <omega_> hehehe
[10:35] <walken> so I guess its still experimental :)
[10:36] <walken> yeah... hasnt been doing any c for a few months though
[10:36] <walken> he's deep into vhdl
[10:36] <omega_> eek
[10:36] <omega_> I was gonna say 'sucked into doing Java', but I was way off <g>
[10:36] <walken> doing wireless dsl modems
[10:36] <omega_> uh?
[10:36] <omega_> surely that's a marketing term...
[10:36] <walken> well no
[10:37] <walken> wireless dsl
[10:37] <omega_> hrm, I guess 'dsl' is a marketing term, not a line discipline, so I guess it doesn't matter
[10:37] <walken> hmmm probably :)
[10:39] Action: ajmitch wonders why XFree86 is so damned sloooooow in this machine
[10:40] Action: walken thinks aaron is a signal processing god
[10:40] <omega_> if he's doing vhdl for wireless dsl, he must be.
[10:40] <walken> I wasnt doing too bad in class, but I dont think I could do modem stuff
[10:41] <walken> well I could always learn
[10:41] <ajmitch> hmm, i can't even get smooth audio playback with an mpeg1 video with gstmediaplay :(
[10:41] Action: omega_ cowers in fear
[10:41] <omega_> is your proc maxed?
[10:41] <ajmitch> more or less
[10:41] <omega_> ah, well, that's why it's not smooth
[10:42] <omega_> just need faster decoders, that's all ;-)
[10:42] <ajmitch> not too bad
[10:42] <ajmitch> sometimes it skips, other times it doesn't
[10:42] <ajmitch> it's abnormal....
[10:42] Action: omega_ decides not to comment
[10:42] <ajmitch> hey, why don't we use smpeg? ;)
[10:42] Action: ajmitch hears omega_ choke
[10:42] <omega_> because its API isn't quite up to the challenge
[10:42] Action: omega_ still wants to write a libsmeg.so that uses gstreamer underneath
[10:43] <omega_> then apps that use smpeg today can play divx tomorrow <g>
[10:43] <ajmitch> that would be very useful...
[10:43] <omega_> and /me thinks Sam would not be happy
[10:43] <ajmitch> heh, overall playback is pretty smooth
[10:43] Action: ajmitch tests out whether he can playback his copy of saving private ryan... ;)
[10:44] <ajmitch> yay, gstmediaplay crashed again! ;)
[10:45] Action: ajmitch just tried to playback a standard mpeg
[10:46] <ajmitch> if i close it, reopen it, and try and play, it works fine
[10:46] <ajmitch> it just can't change files ;)
[10:46] <omega_> of course
[10:46] <omega_> that's been the case from the beginning
[10:47] <ajmitch> is that bad? ;)
[10:47] <omega_> uh, whaddya thing?
[10:47] <omega_> er, think?
[10:47] <ajmitch> nah, just tell the users it was their fault ;)
[10:48] <ajmitch> wow, watching an mpeg that's about 20sec out of sync is weird ;)
[10:48] dr88dr88 (dr88dr88 at c1220.upc-c.chello.nl) joined #gstreamer.
[10:48] <ajmitch> it caught up eventually, i think...
[10:49] <omega_> hrm
[10:49] <ajmitch> yup, definitely weird looking ;)
[10:49] <ajmitch> i think it is the video
[10:49] <ajmitch> since smpeg is horribly out of sync as well
[10:50] <omega_> oops
[10:50] <ajmitch> mpegtv could play it ok
[10:50] <ajmitch> but you're right, we need a faster decoder
[10:50] Action: ajmitch looks to see who is the victim on the roadmap
[10:51] <walken> omega. what do you use for mpeg1 ?
[10:51] <walken> mpeg2dec should be able to handle this too
[10:51] <omega_> wtay seriously hacked up mpeg_play, got it a lot faster and sucking less
[10:51] <omega_> we tend to use mpeg2dec though, these days
[10:52] <ajmitch> ok, i dunno what gstmediaply will be using, how can i find out?
[10:52] <walken> also is it true that your version of mpeg2dec does more memcopies ?
[10:52] <omega_> need to do some timings and build a merit system, so the autoplugger can make a smart decision between mpeg_play and mpeg2dec
[10:52] <walken> like does a memcpy for the frame output or something
[10:52] <omega_> probably, I haven't looked at it
[10:52] <walken> thats what I've been told
[10:52] <omega_> taazzzz is asleep, as if he knew this would come up <g>
[10:52] <walken> and then you complained that it didnt ran well on your 500 MHz machine or something :>
[10:53] <ajmitch> hehe
[10:53] <omega_> yeah, it isn't quite fast enough, or maybe ac3dec has more to do with that
[10:53] <omega_> since it's a branch of a 1+yo ac3dec
[10:53] <walken> (last time I tested on my athlon it was a 30% speed impact to memcpy the output yuv frames)
[10:53] <ajmitch> ouch
[10:53] <omega_> eeeek
[10:53] <ajmitch> i run this stuff on a 400mhz k6-2 ;)
[10:54] Action: ajmitch should test on my k6-233
[10:58] <ajmitch> hmm, i don't often see the audio lag the video by 10sec or so ;)
[11:00] <walken> OK I'm leaving....... nite nite
[11:00] <omega_> l8r
[11:00] <ajmitch> bye
[11:01] walken (foobar at c1583255-a.smateo1.sfba.home.com) left irc: l8r
[11:01] <ajmitch> omega_: something tells me that whatever decoder gstmediaplay uses is slightly broken
[11:01] <omega_> probably mpeg_play
[11:01] <ajmitch> yeah
[11:01] <omega_> check for a new .gst file after you run gstplay
[11:02] <omega_> ls -lart
[11:02] <omega_> check to see if it exists, and says which decoder you're using
[11:02] <ajmitch> hmm, in the current dir?
[11:02] <omega_> wherever you ran it from, yeah
[11:02] <ajmitch> not there
[11:02] <omega_> hrm, I forgot, that's not easy
[11:03] <omega_> grrr
[11:03] <omega_> sec..
[11:04] <omega_> cvs update gstplay.c main.c
[11:04] <omega_> make
[11:05] <omega_> heh.  1:14 to make patch <g>
[11:05] <ajmitch> retrieving revision 1.8
[11:05] <ajmitch> retrieving revision 1.9
[11:05] <ajmitch> Merging differences between 1.8 and 1.9 into main.c
[11:05] <ajmitch> M main.c
[11:05] <ajmitch> oops ;)
[11:05] <omega_> that's fine
[11:06] <omega_> just means you've touched it too
[11:06] <ajmitch> just me playing
[11:06] <ajmitch> yup
[11:06] <ajmitch> was adding an open location entry on the toolbar
[11:07] <ajmitch> oh, it does use mpeg_play
[11:07] <omega_> ick
[11:07] <omega_> try removing libmpeg_play.so
[11:07] <omega_> plugins/mpeg1/mpeg_play
[11:07] <omega_>  /.libs
[11:09] <ajmitch> and make sure to run gstreamer-register ;)
[11:10] <omega_> um, yeah <g>
[11:11] <ajmitch> oops, crashes still ;)
[11:11] <ajmitch> may i paste top of stack trace?
[11:11] <omega_> ok
[11:11] <ajmitch> ***** GStreamer ERROR ***** in file gstscheduler.c at gst_bin_schedule_func:431
[11:11] <ajmitch> Element: /main_thread/main_bin/autoplug_bin/mp3parse.sink
[11:11] <ajmitch> Error: peer is null!
[11:11] <ajmitch> ***** attempting to stack trace.... *****
[11:11] <ajmitch> #0  0x40077708 in g_on_error_stack_trace () from /usr/lib/libglib-1.2.so.0
[11:11] <ajmitch> #1  0x40029873 in gst_default_error_handler (file=0x4004a25c "gstscheduler.c", 
[11:11] <ajmitch> #2  0x40039bfd in gst_bin_schedule_func (bin=0x80a98e0) at gstscheduler.c:431
[11:11] <ajmitch> #3  0x4003b164 in gst_thread_main_loop (arg=0x80a98e0) at gstthread.c:314
[11:11] <ajmitch> #4  0x40094fc5 in pthread_start_thread () from /lib/libpthread.so.0
[11:12] <ajmitch> and it locked at 100% cpu usage there
[11:12] <omega_> did it work before you removed mpeg_play.so ?
[11:13] <ajmitch> umm, yep ;)
[11:13] <omega_> hmm, interesting
[11:13] <omega_> this is unrelated
[11:13] <omega_> ok, make -C ../plugins/mpeg1/mpeg_play;../tools/gstreamer-register;!./gstmed
[11:14] <ajmitch> just a min...
[11:23] <ajmitch> ok, friend has gone ;)
[11:25] Action: ajmitch runs make clean && make 
[11:25] <omega_> huh?
[11:25] <ajmitch> in mpeg_play dir
[11:26] <ajmitch> running make didn't do anything
[11:26] <ajmitch> didn't recreate the libs
[11:27] <omega_> odd
[11:27] <omega_> oh, remove libmpeg_play.la and then it would
[11:27] <ajmitch> ah well, back to the wonderful mpeg_play ;)
[11:28] <ajmitch> i think there are a few issues to be worked out before 0.2.0
[11:28] <omega_> run ./gstreamer-inspect mpeg2dec
[11:28] <ajmitch> how usable are you wanting the release to be?
[11:28] <omega_> uh, gstmediaplay should work
[11:29] <ajmitch> ajmitch at ajmitch:~/cvs/gstreamer/tools$ ./gstreamer-inspect mpeg2dec
[11:29] <ajmitch> INFO(14487:-1): Initializing GStreamer Core Library
[11:29] <ajmitch> INFO(14487:-1): CPU features: (808029bf) MMX 3DNOW 
[11:29] <ajmitch> no such element or plugin 'mpeg2dec'
[11:29] <omega_> ah, that would do it
[11:29] <omega_> cd into plugins/mpeg2/mpeg2dec and build it
[11:29] <omega_> that is, if you have libmpeg2dec installed
[11:29] <ajmitch> umm, it is installed tho ;)
[11:29] <omega_> ok
[11:29] <omega_> configure needs some help
[11:30] <ajmitch> hmm
[11:30] <ajmitch> maybe not...
[11:30] <ajmitch> am getting errors there
[11:30] <ajmitch> gstmpeg2dec.c:24: mpeg2dec/mm_accel.h: No such file or directory
[11:30] <ajmitch> gstmpeg2dec.c:25: mpeg2dec/video_out.h: No such file or directory
[11:30] <ajmitch> In file included from gstmpeg2dec.c:26:
[11:30] <ajmitch> gstmpeg2dec.h:27: mpeg2dec/mpeg2.h: No such file or directory
[11:30] Action: omega_ forgot that he rebuild his copy of HEAD with *all* debug turned off
[11:30] <ajmitch> where shall i get these from? ;)
[11:30] <omega_> linuxvideo.org
[11:30] <omega_> but that was down a few hours ago
[11:30] <ajmitch> ok, trying apt-get
[11:31] <omega_> I can email you a CVS snapshot if all else fails
[11:31] <omega_> 185KB
[11:31] <ajmitch> but then again i may have a cvs copy...
[11:31] steveb (stevebaker at 212186169160.chello.com) left irc: Read error to steveb[212186169160.chello.com]: Connection reset by peer
[11:32] <ajmitch> hehe, yup, i have ~/cvs/mpeg2dec
[11:32] <omega_> that's a new one
[11:32] <omega_> steveb's hostname has the *decimal* IP in it
[11:32] <omega_> with no -'s or anything
[11:34] <ajmitch> what do you mean?
[11:34] Action: ajmitch runs make
[11:34] <ajmitch> oh ;)
[11:37] <ajmitch> does configure check for libs installed into the same prefix as '--prefix' ?
[11:37] <omega_> not afaik
[11:37] <ajmitch> damn
[11:37] <ajmitch> i need somewhere to install mpeg2dec
[11:37] <omega_> just set your PATH and LD_LIBRARY_PATH
[11:38] <ajmitch> have set LD_LIBRARY_PATH
[11:38] <ajmitch> PATH already contained that dir
[11:38] <omega_> I have a prefix permanently added like that in my .bash_profile
[11:38] <ajmitch> what do you use?
[11:38] arik (arik at sdn-ar-001waseatP105.dialsprint.net) joined #gstreamer.
[11:38] <omega_> I have one for the codecs.org stuff, not gstreamer
[11:38] <ajmitch> ok
[11:38] <arik> lo all
[11:38] <ajmitch> hi
[11:38] <omega_> yo
[11:39] <arik> how goes it?
[11:39] <omega_> autoplug works <g>
[11:39] <ajmitch> hey, another victim! ;)
[11:39] <ajmitch> omega_: that is good ;)
[11:39] <arik> wow! ;-)
[11:39] <ajmitch> omega_: what will you work on now? ;)
[11:39] <omega_> not complete yet though
[11:39] <arik> ;-)
[11:40] <omega_> ajmitch: cvs update plugins/mpeg2/mpeg2dec/gstmpeg2dec.c
[11:40] <arik> yippie!
[11:40] <omega_> now its caps say it can do mpeg 1
[11:41] <ajmitch> omega_: so if i want to dump stuff in /opt/gnome like the new glib, etc, how will configure find those over the ones in /usr? ;)
[11:41] <omega_> depends on how it searches
[11:43] <ajmitch> hmm
[11:43] <arik> well gstplay still doesn't work ;-)
[11:44] <omega_> yeah, because this autoplugger is a new element, and gstplay doesn't have the first clue how to use it yet
[11:44] <omega_> (nor do I <g>)
[11:44] <arik> omega_: hehe
[11:44] <ajmitch> lol
[11:44] <ajmitch> omega_: where are the docs on it? ;)
[11:44] <arik> omega_: autoplug example seems to work although i still get tons of cothreads errors
[11:44] <omega_> yeah, it's not being nice to the scheduler
[11:44] <arik> *nod*
[11:45] <arik> it does work though
[11:45] <arik> which is nice
[11:45] <omega_> though ideally I should figure out what it's doing wrong, document it, then see if I can have the scheduler make more useful noise in those cases
[11:47] <omega_> hrm, capsnego isn't firing as often as it should
[11:47] Action: ajmitch wonders why there is now no picture
[11:47] <arik> right
[11:47] <arik> or fix it ;-P
[11:48] <omega_> there are going to be some things you just can't do to the scheduler any more
[11:48] <arik> hmm
[11:48] <omega_> they simply don't mean anything useful anymore
[11:48] <arik> autoplug still doesn't play the other test cases though
[11:48] <ajmitch> omega_: using mpeg2dec, but have no picture ;)
[11:48] <omega_> the hacks necessary to get pseudo-dynamic autoplug to work excercise most of those things to not do anymore <g>
[11:49] <omega_> ajmitch: you have Xv?
[11:49] <arik> heh
[11:49] <ajmitch> omega_: yeah, right! this is a crap card
[11:49] <ajmitch> omega_: DGA doesn't even work with this driver ;)
[11:49] <omega_> hmmm
[11:50] <omega_> is there a 'colorspace' in the .gst file?
[11:50] <ajmitch> yup
[11:50] <ajmitch> most certainly is ;)
[11:51] <omega_> hmm
[11:52] <ajmitch> using the same video as before, too
[11:52] <arik> well one video now works for me and like 200 others dont
[11:52] <arik> but that's still a start ;-)
[11:52] <omega_> hrm, odd
[11:52] <ajmitch> Using MMX for IDCT transform
[11:52] <ajmitch> Using 3DNOW for motion compensation
[11:54] Action: ajmitch might go sleep now and think about it tomorrow
[11:55] <arik> ajmitch: sleep is for the week ;-P
[11:55] <omega_> ah, heheh
[11:55] <omega_> it is the week
[11:55] <arik> ;-)
[11:55] <ajmitch> arik: i just had a psycho test before, i need my sleep ;)
[11:55] <omega_> arik: were you at GUADEC?
[11:56] <arik> omega_: no ;-( i would have been but then...
[11:56] <arik> ajmitch: ;-)
[11:56] <ajmitch> anyway...
[11:57] <omega_> I'm wondering what north-american conference would be best to try to gather people for a meet....
[11:57] <arik> omega_: i used to work at eazel
[11:57] <ajmitch> night all
[11:57] Nick change: ajmitch -> ajzzzz
[11:57] <arik> ajzzzz: night
[11:57] <omega_> arik: just recently, or a while ago?
[11:57] <arik> omega_: just before the layoffs one might say ;-)
[11:57] <omega_> I though you worked for N$/AOL on mozilla?
[11:58] <arik> i do
[11:58] <arik> now
[11:58] <omega_> ok, you worked for them a while ago
[11:58] <arik> yeah, sorry, missunderstood
[12:00] <omega_> any thoughts on a conference to try?
[12:01] <arik> hmm
[12:01] <arik> ols or als? maybe an lwe?
[12:01] <arik> there is supposed to be a gnome meetup at some conference right?
[12:01] <omega_> yeah, there are several
[12:02] <arik> maybe one of those
[12:02] <omega_> yup
[12:02] <arik> hard to do with everyone's shedules
[12:03] <omega_> yup
[12:03] <omega_> I'd say OLS would be an option, since it's fairly 'central' (even though it's way up north)
[12:03] <arik> yep
[12:03] <omega_> and it's very technically oriented
[12:03] <arik> and it's supposed to be fun ;-)
[12:03] <omega_> is ALS even still happening?
[12:03] <arik> don't know
[12:04] <omega_> I know LinuxExpo NC kinda fell flat
[12:04] <arik> hmm
[12:04] wtay-zZz (wim at cable-195-162-214-58.upc.chello.be) got netsplit.
[12:04] Dr_Zaius (jeff at h00a0244bc361.ne.mediaone.net) got netsplit.
[12:04] dr88dr88 (dr88dr88 at c1220.upc-c.chello.nl) got netsplit.
[12:04] taazzzz (dlehn at got netsplit.
[12:04] wtay-zZz (wim at cable-195-162-214-58.upc.chello.be) returned to #gstreamer.
[12:04] taazzzz (dlehn at returned to #gstreamer.
[12:04] Dr_Zaius (jeff at h00a0244bc361.ne.mediaone.net) returned to #gstreamer.
[12:04] dr88dr88 (dr88dr88 at c1220.upc-c.chello.nl) returned to #gstreamer.
[12:05] <omega_> hrm, can't find url for OLS
[12:05] <arik> must be out there
[12:07] <omega_> not finding anything ;-(
[12:07] <omega_> ah, misspelling it
[12:08] <omega_> http://www.linuxsymposium.org/
[12:09] <arik> ;-)
[12:12] <omega_> I guess I should post to the list to see who might be able to go
[12:12] <omega_> there are ~150 out of 500 spots left, it costs US$260
[12:12] <arik> wow
[12:12] <omega_> plus getting there...
[12:12] <arik> that is alot of money
[12:12] <arik> i'm still not on the list
[12:12] Action: arik is bad
[12:12] <arik> i just get so much mail...
[12:12] <arik> ah well
[12:13] <omega_> your addr?
[12:13] <arik> arik at gnome.org
[12:13] <omega_> done
[12:13] <arik> neeto
[12:13] <omega_> er, maybe not
[12:13] <arik> gstreamer-list at gnome.org?
[12:13] <omega_> 'mass subscribing has been disabled due to abuse'
[12:13] <arik> heh
[12:13] <omega_> gstreamer-devel at lists.sourceforge.net
[12:13] <arik> ah ok
[12:14] <omega_> that means you get to do it..
[12:14] <omega_> http://lists.sourceforge.net/lists/listinfo/gstreamer-devel
[12:14] <arik> neet
[12:15] <arik> done
[12:16] <arik> argh
[12:16] <omega_> ?
[12:16] <arik> i wish i had smtp working in gnus
[12:17] <arik> i have to send all my mail from a web based interface cause gnus doesn't send my mail atm
[12:17] <arik> cause i haven't set it up
[12:17] <arik> and i don't know how ;-)
[12:18] <omega_> eeew
[12:18] <arik> heh
[12:22] <omega_> ok, I need to sleep now
[12:22] <omega_> l8r
[12:23] omega_ (omega at omegacs.net) left irc: [x]chat
[12:24] arik (arik at sdn-ar-001waseatP105.dialsprint.net) left #gstreamer.
[12:48] taazzzz (dlehn at left irc: Ping timeout for taazzzz[]
[12:50] taazzzz (dlehn at joined #gstreamer.
[13:35] greg_ (greg at home.sente.pl) joined #gstreamer.
[13:38] greg_ (greg at home.sente.pl) left irc: Ping timeout for greg_[home.sente.pl]
[14:04] greg_ (greg at home.sente.pl) joined #gstreamer.
[14:42] wtay-zZz (wim at cable-195-162-214-58.upc.chello.be) left irc: Ping timeout for wtay-zZz[cable-195-162-214-58.upc.chello.be]
[14:42] dr88dr88 (dr88dr88 at c1220.upc-c.chello.nl) left irc: Ping timeout for dr88dr88[c1220.upc-c.chello.nl]
[16:05] Nick change: taazzzz -> taaz
[16:42] ajzzzz (ajmitch at p11-max6.dun.ihug.co.nz) left irc: Ping timeout for ajzzzz[p11-max6.dun.ihug.co.nz]
[17:25] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[17:25] <Uraeus> hi
[17:48] hadess (hadess at pc2-guil2-0-cust121.gui.cable.ntl.com) joined #gstreamer.
[17:49] <hadess> hi gang
[17:49] <Uraeus> hi hadess
[17:50] <hadess> heya Uraeus
[17:50] <hadess> howdy ?
[17:50] <Uraeus> ho?
[17:50] <hadess> how are you, i asked
[17:51] <Uraeus> good, if a little bored :)
[17:51] <Uraeus> and you?
[17:52] <hadess> very good
[17:52] <hadess> i'm finishing the gnome vfs src
[17:54] <Uraeus> thats nice, so it will be ready for 0.2.0?
[17:54] greg_ (greg at home.sente.pl) left irc: [x]chat
[17:54] <hadess> i hope so, i still can't test it, but it should be finished some time today i think
[17:56] <hadess> my plugin is not appearing in the registered plugins...
[17:57] Action: hadess is waiting for omega_ to appear here
[17:57] <Uraeus> well, if it is in working order in within the next 10 days I can put your name in the GNOME summary for it :)
[17:57] <taaz> heh... you can't test it but it'll be ready?  you must be a really good coder ;)
[17:58] <hadess> Uraeus: wheee
[17:58] <taaz> how did you try to regsiter it?
[17:59] <hadess> taaz: heh, i'm waiting for omega to appear to tell me what i should do to make it be recognized, shouldn't be long afterwards
[17:59] <hadess> taaz: using gst-reg
[17:59] <taaz> well, is it installed in the main plugin dir?
[18:00] <hadess> it is not installed
[18:01] <taaz> it should be just a matter of doing what the other plugins do...  make sure you have the init function set stuff up and poof
[18:08] matth (matth at qwest.dsplinux.net) joined #gstreamer.
[18:08] Nick change: matth -> matth_
[18:09] <Uraeus> hi matth_
[18:13] sienap (synap at ipc379c13f.dial.wxs.nl) joined #gstreamer.
[18:14] <sienap> Hi all :)
[18:15] <Uraeus> hi sienap
[18:15] <sienap> ureaus!
[18:17] <Uraeus> that is not how it is spellt :)
[18:18] <sienap> he
[18:18] <sienap> :)
[18:23] <matth_> Uraeus: howdy... as usual i'm distracted ;-)
[18:24] <Uraeus> matth_: what is happening at RidgeRun, any announcements/releases etc. upcoming?
[18:24] <sienap> :)
[18:26] <hadess> taaz: hmm, it doesn't seem to find the plugin, still
[18:26] sienap (synap at ipc379c13f.dial.wxs.nl) left #gstreamer.
[18:31] <matth_> Uraeus: we released the 1.0 version (first official version) of our SDK... the press release isn't on our web site yet
[18:32] <matth_> Uraeus: main real target supported is TI DSC21 (a digital still camera system-on-a-chip)
[18:33] Nick change: hadess -> hds-gone
[18:37] <Uraeus> matth_: so who are your customers? 
[18:40] steveb (steveb at node1ee03.a2000.nl) joined #gstreamer.
[18:43] <matth_> Uraeus: i'm not sure we have permission to say, and the people who know are all in a meeting
[18:43] <matth_> Uraeus: i'll let you know when i find out
[18:44] <Uraeus> matth_: i don't mean names, I was more thinking along the line of types of companies 
[18:45] <Uraeus> who the market are for a linux to run on a digital camera system isn't clear to me :)
[18:45] <matth_> not all of them are doing dsc's - which is why they want to keep it under wraps
[18:46] <matth_> the dsc21 has build in asic's that do a bunch of multimedia stuff like:
[18:46] <matth_> + mpeg 1, 2, 4 playback
[18:46] <matth_> + mpeg 2 -> 4 conversion
[18:47] <matth_> + jpeg encode/decode (obviously)
[18:47] <Uraeus> so a customer would be Kodak wanting a system for their advanced digital camera? or is it more end user oriented?
[18:47] <matth_> i've seen proposed devices that are all over the map
[18:47] <matth_> one shipping device is Kodak's MC3
[18:53] <matth_> Uraeus: people tend to want linux because 1. it's cheap ;-) and 2. it's rich in the protocol/transport layers (tcp/ip, ir, bluetooth, wap, http, ...)
[18:54] Action: Uraeus nods
[19:17] Dr_Zaius (jeff at h00a0244bc361.ne.mediaone.net) left irc: [x]chat
[19:30] wtay (wim at cable-195-162-214-58.upc.chello.be) joined #gstreamer.
[19:30] <wtay> hmm, sorry, I was disconnected...
[19:32] <Uraeus> hi wtay
[19:32] <wtay> hi
[19:32] Action: wtay has read the _gst_newt_ log
[19:33] Action: Uraeus wanted to ask wtay something but can't remember
[19:34] <wtay> Uraeus: think harder :-)
[19:35] <Uraeus> hmm, how long does it take for a dove to fly from africa to europe with a coconut?
[19:36] <Uraeus> no,that wasn't it
[19:37] <Uraeus> ah, now I remember, wtay had any time to fix the XFree86 dependency?
[19:41] BBB (BBB at ucu-105-116.ucu.uu.nl) joined #gstreamer.
[19:41] <Uraeus> hi BBB
[19:41] <BBB> congrats guys, you've just conquered OMS
[19:41] <BBB> :)
[19:42] <BBB> hi Uraeus
[19:42] <Uraeus> OMS? the meta system for documention?
[19:42] <BBB> no :P
[19:42] <BBB> Open Media System (OMS, http://www.linuxvideo.org)
[19:42] <Uraeus> Object Management Group?
[19:42] <BBB> read your e-mail :P
[19:43] <wtay> nice work taaz :-)
[19:43] <Uraeus> wtay: XF4 dep?
[19:43] <wtay> no time for that yet
[19:44] <wtay> as soon as I've plugged two memleaks, I'll look at it, I promise
[19:44] <taaz> BBB: i'm not sure what other LiViD people want to do... I'm just looking for shortest path to a DVD player and gstreamer looks like a good way to get things going quickly
[19:45] <BBB> I think it's a good idea
[19:45] <wtay> taaz: yes
[19:45] <BBB> I mean....
[19:45] <BBB> generally....
[19:45] <BBB> we're all doing double code now
[19:45] <BBB> and I also mean the MJPEG project, now
[19:46] <BBB> and having one basis framework for that would be really really good
[19:46] Action: BBB notices that taaz = David Lehn
[19:46] <taaz> no way?! ;)
[19:46] <wtay> BBB: did you see the lavsink I made?
[19:47] <BBB> you made a lavsink?
[19:47] <BBB> omg
[19:47] <BBB> we should have done that
[19:47] <wtay> yup
[19:47] <BBB> :)
[19:47] Action: BBB bows deeply
[19:47] <BBB> where can I get it?
[19:47] <wtay> I managed to use mpeg2enc to encode a dv decoded stream
[19:47] <wtay> BBB: cvs of course :)
[19:47] <BBB> just the main tree?
[19:48] <wtay> yes
[19:48] <BBB> [rbultje at tux gstreamer]$ cvs update
[19:48] <BBB> let's try
[19:48] <wtay> just wraps YUV frames into lav frames
[19:48] <BBB> that's pretty nice
[19:48] <BBB> and 'lav' is, in this case, jpeg-encoded?
[19:48] <wtay> no lavsrc yet though
[19:49] <wtay> err, no, just the LAV4MPEG header and stuff
[19:49] Action: BBB looks up what the lav4mpeg header was
[19:49] <BBB> isn't that the YUV header?
[19:49] <wtay> pretty stupid but I wanted to test mpeg2enc
[19:49] <BBB> YUV <width> <height> <encoding>
[19:50] <BBB> that's a YUV header
[19:50] <BBB> :)
[19:50] <wtay> yup
[19:50] <BBB> not a lav header :)
[19:50] <wtay> ok
[19:50] <BBB> lol
[19:50] <BBB> we'll mke a lavsrc pretty soon (I didn't hear complaints about 1.4pre1, so I think we'll have a final pretty soon)
[19:51] <wtay> BBB: I've been thinking about an editing framework around gstreamer..
[19:51] <wtay> I hope to set up a sf page for that soon to design the thing
[19:51] <BBB> I want to start thinking about that as soon as I got a nice release of my li'l LVS :)
[19:52] <BBB> but you mean a 'general editor'?
[19:52] <BBB> I think the guys from Kino would also be interested in that
[19:52] <wtay> no, an API for video editing
[19:52] <wtay> apps come after that
[19:52] <BBB> not only an API
[19:52] <BBB> API+app
[19:52] <wtay> I first want a sane lib done before starting an app
[19:53] <BBB> and is that lib going to be part of gstreamer?
[19:53] Action: BBB guesses negative
[19:53] <wtay> no
[19:53] <BBB> ok....
[19:53] <wtay> :)
[19:53] Action: BBB thinks that's a hell-of-a-good-idea
[19:53] <BBB> I'd like to help if I an
[19:53] <BBB> s/an/can/
[19:53] <wtay> we'll eventually make an app to test things of course
[19:53] Nick change: hds-gone -> hadess
[19:53] <BBB> and not only test ;)
[19:53] <wtay> sure, the more help the better
[19:53] <hadess> yo guys
[19:54] <BBB> in the ideal case, I'd like to see a sweet complete video editor in ten years
[19:54] <wtay> hi hadess
[19:54] <BBB> hi hadess
[19:54] <wtay> BBB: I'm aiming for 3 years
[19:54] <Uraeus> hi hadess
[19:54] <BBB> I saw apple's imovie lately
[19:54] <BBB> I think that's _really_ cute
[19:54] <BBB> I'd like to see something like that too
[19:54] <BBB> but then with support for any format rather than only QT
[19:54] <wtay> heh
[19:55] <BBB> but just a bit more advanced options
[19:55] <wtay> I think an API can be done in about a year
[19:55] <BBB> like combining layers etc.
[19:56] <BBB> and.... what should the API do?
[19:56] <BBB> things like the combining layers?
[19:56] <wtay> add a time dimension to gstreamer
[19:56] <BBB> for the editing lists :)
[19:56] <wtay> time dimension in the sense that the pipeline changes over time
[19:56] <wtay> yes
[19:56] <BBB> and the whole object system of a "bunch of scenes"
[19:57] <wtay> yes
[19:57] <BBB> c++ or c?
[19:57] <wtay> c
[19:57] <BBB> woohoo!
[19:57] <wtay> gobject
[19:57] <BBB> I wanna help :)
[19:57] <wtay> sure
[19:57] <BBB> glib-object..... glib-2.0
[19:57] <wtay> yup
[19:57] Action: BBB doesn't even have the beta of glib-2.0 installed yet
[19:58] <wtay> I think we can start on it by september or so
[19:58] <BBB> it'd be a _really_ great idea
[19:58] <BBB> and I would surely help
[19:58] <hadess> wtay: dude, can you help me on getting gnomevfssrc to be recognized as one of the plugins ?
[19:58] <wtay> well, you have the experience and all
[19:58] <wtay> hadess: sure
[19:58] Action: BBB wonders..... experience?
[19:59] <BBB> I'm on linux for less than a year :)
[19:59] <BBB> I've been programming for only....... six months?
[19:59] <BBB> :P
[19:59] <hadess> wtay: i have my plugin compiled in gstreamer/plugins/gnomevfs
[19:59] <wtay> hadess: that's ok
[19:59] <wtay> hadess: did you use the CVS version?
[19:59] <hadess> wtay: but my plugin doesn't appear in the reg plugins cache
[20:00] <wtay> hadess: you have the plugin_desc struct in it?
[20:00] <hadess> GstElementDetails ?
[20:00] <wtay> hadess: look at aasink.c, the last 30 lines or so
[20:00] <wtay> hadess: that's some info about the element, you'll need code to register it
[20:01] <wtay> a plugin_init and a plugin_desc
[20:01] <wtay> hadess: since you started fron fdsrc, it isn't in the code yet
[20:01] <hadess> wtay: oooh, ok
[20:02] <wtay> hadess: the code to register fdsrc is in gstelements.c, so...
[20:02] <hadess> wtay: yeah, i didn't think about that
[20:03] <wtay> hadess: when the plugin is loaded, the plugin loader will look for a struct called plugin_desc
[20:03] <wtay> from that struct it'll load the name and some version info, and it'll call the function in it
[20:03] <wtay> in the plugin_init function you'll add the plugin to the registry
[20:03] <BBB> wtay: now that I think about it - a gobject is basically like a gtk-widget, right? (I mean, concerning the basis of the coding)
[20:04] <wtay> BBB: yes
[20:04] <wtay> BBB: but without the X dependency and all the other GUI bloat
[20:04] <BBB> of course
[20:05] <BBB> that's why I added, "concerning the basis of the coding" :)
[20:05] <hadess> wtay: what for gst_elementfactory_add_padtemplate ?
[20:05] <wtay> basically a s/gtk_/g_/g will solve 95% of the migration
[20:05] <BBB> it's basically a few structs, a few defines to identify objects and function calls which don't depend on global variables but only on local variables which identify the gobject.....
[20:05] <wtay> hadess: that's when you want to add a template to the pad..
[20:05] <hadess> wtay: do i need that ?
[20:06] <wtay> hadess: depends, I would guess not
[20:06] <wtay> hadess: can gnomevfssrc detect a type somehow?
[20:06] <wtay> hadess: like video/mpeg ?
[20:06] ChiefHighwater (paul at temple-baptist.com) joined #gstreamer.
[20:06] <wtay> hello
[20:07] <ChiefHighwater> ello
[20:07] <BBB> hi CHW
[20:07] <hadess> wtay: yes, it can detect the mime-type, but i didn't set that up
[20:07] <ChiefHighwater> 8-]
[20:07] <wtay> hadess: ok, just leave the template out, typefind is probably better anyway
[20:08] <hadess> wtay: do you remember the gst-launch syntax for me to test it ?
[20:09] <wtay> hadess: ./gstreamer-launch gnomevfssrc loaction=/some/mp3 ! mad ! osssink
[20:09] <wtay> s/loacation/location
[20:09] <wtay> brb
[20:10] <hadess> oopsie, forgot about gnome_vfs_init...
[20:11] <wtay> back
[20:11] <hadess> where could i put the gnome_vfs_init ?
[20:11] <hadess> in the class init ?
[20:11] Action: BBB starts thinking about a nlelib
[20:11] <BBB> libnle.so
[20:11] <BBB> :)
[20:11] <wtay> hadess: or maybe in the state change?
[20:12] <hadess> wtay: hmm, will try that
[20:12] <wtay> hadess: does it depend on the URI you try to access?
[20:13] <hadess> wtay: no, it's just like gtk_init, or gnome_init
[20:13] <wtay> hadess: then it should go in the class init
[20:16] <hadess> pooof
[20:16] <hadess> segfaults somewhere :/
[20:19] <BBB> gdb :)
[20:19] <hadess> BBB: sure bud, you haven't tried that yourself have you
[20:20] sienap (synap at ipc379c21d.dial.wxs.nl) joined #gstreamer.
[20:20] <sienap> *sniff*
[20:20] <sienap> eazel closing doors
[20:20] <sienap> too bad..
[20:20] <wtay> yo sienap
[20:21] <sienap> hi..
[20:21] <sienap> wtayt
[20:21] <sienap> how is gstreamer going ?
[20:21] <wtay> pretty ok
[20:22] <sienap> some news lately..
[20:22] <sienap> eazel is death btw
[20:22] <sienap> maybe you didn't hear yet ..
[20:22] <BBB> eazel dead?
[20:22] <BBB> :(((
[20:23] <Uraeus> eazel is dead, but Nautilus lives on :)
[20:23] <wtay> yup, dead as a dead cow
[20:23] <sienap> uraeus yeah
[20:24] <sienap> but good people will leave.
[20:24] <sienap> like andy
[20:24] <sienap> and arlo
[20:24] <sienap> believe
[20:24] <sienap> you won't hear anything from them anymore
[20:24] <Uraeus> yes, but good people like Darin will stay
[20:24] <sienap> uraeus true
[20:24] <sienap> he got hooked :)
[20:25] <Uraeus> sienap: both Arlo and Andy said they are leaving?
[20:25] <sienap> he not that i know
[20:25] <sienap> but those people have wife  childeren and stuff
[20:25] <sienap> they need work
[20:25] <sienap> :)
[20:25] <Uraeus> yes, i know but that isn't a problem :)
[20:25] <sienap> however
[20:25] <sienap> the source is there :)
[20:25] <sienap> he
[20:25] <sienap> childeren / wife can be solved .:)
[20:26] <Uraeus> I think that there will be an announcement soon about the employment situation which will be good :)
[20:26] <hadess> wtay: how can i debug this thing ?
[20:27] <wtay> hadess: libtool gdb 
[20:27] <BBB> brb
[20:28] <hadess> hmm, memory crunching
[20:35] <BBB> poor eazel
[20:36] <BBB> :(
[20:36] <sienap> he yeah
[20:36] <sienap> ze gaan NEEEEEEERRRRRRRRR >:)
[20:36] <sienap> but the source lifes on
[20:36] <sienap> and so do we
[20:36] <sienap> :)
[20:37] <sienap> so if every lonely man.. every poor kiddo every idioticial emphyrio start hacking on it
[20:37] <sienap> so araeus
[20:37] <sienap> who is going to lead the team
[20:37] <sienap> of filemanager knights..
[20:37] <BBB> not me
[20:38] <BBB> you can do that, sienap
[20:38] <BBB> wtay just told me about his libnle plan, I'll try to be involved in that :)
[20:38] <BBB> but uhm...
[20:38] <BBB> basically
[20:38] <BBB> nautilus was nice and all....
[20:39] <BBB> but it was just faaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaar too slow
[20:40] Action: BBB will still continue hacking on linux, though, even w/o nautilus
[20:41] <BBB> sienap: why don't you start hacking on linux?
[20:42] <sienap> bbb because i am the man :)
[20:42] <BBB> which means that?
[20:42] <sienap> i only watch hackers :)
[20:42] <sienap> and claim i am one my self
[20:42] <sienap> but without hacking at all
[20:42] <BBB> uhm
[20:42] <sienap> *ooh don't take me to serious btw*
[20:42] <BBB> then you're not a true hacker
[20:42] <BBB> :)
[20:43] <BBB> :P
[20:43] <sienap> i got fucked :)
[20:43] <BBB> lol
[20:43] <sienap> actually i am allowed to do my gymastic / german test on school
[20:43] <BBB> was she beautiful? <lol>
[20:43] <sienap> 2 out of 14 tests that is..
[20:43] <sienap> *ooh and i am so happy right now*
[20:43] <sienap> i am allowed to do 1/7 of my exams
[20:43] <sienap> WIEEEEEEEEEH
[20:43] <sienap> fucking obsessed fustrated assholes on school
[20:43] <BBB> hmm.....
[20:44] Action: BBB did all 9 exams, though
[20:44] <sienap> getting mad when i plan my own vacations
[20:45] <sienap> it pretty much sucks
[20:45] <sienap> i am not allowed to do my tests
[20:45] <BBB> hmm.... sounds weird
[20:45] <sienap> because i decided to have a 2.5 week vacation some weeks ago
[20:45] <BBB> you should have joined us in amsterdam :)
[20:45] <sienap> i really had it with school that time....
[20:46] <sienap> (i am at mbo btw)
[20:46] <BBB> mbo?
[20:46] <sienap> mts
[20:46] <sienap> :)
[20:46] <sienap> ti
[20:46] <BBB> how can they possibly tell you what to do in mbo?
[20:46] <BBB> ti = easy
[20:46] <sienap> i know
[20:46] <sienap> if i can do the tests
[20:46] <sienap> everything will be ok
[20:46] <BBB> and if you cannot have a break, well, then that's just gross
[20:46] <sienap> but yeah they deny me them..
[20:46] Action: BBB needs to do a short phonecall
[20:46] <BBB> brb
[20:47] <sienap> so if i am not going to make this year
[20:48] <sienap> man life gets fun :)
[20:48] <sienap> then i quit with school
[20:48] <sienap> and start doing non-sense
[20:48] <sienap> >:)
[20:48] <sienap> let's see if that works out
[20:48] <BBB> uhm
[20:48] <BBB> don't forget to join us hacking together libnle.so
[20:48] <BBB> :P
[20:49] <BBB> hmm.....
[20:49] <BBB> I gotta do some homework
[20:49] <BBB> sorry guys
[20:49] <BBB> bbl
[20:49] Action: BBB is away: homework
[20:49] <sienap> libnle.so === ?
[20:50] <BBB> a library fro a video editor (non-linear editor, NLE) = libnle
[20:51] <BBB> and since we'll make it a library, it's gonna be libnle.a/la/so
[20:51] <BBB> libnle.so :)
[20:51] <sienap> ahjeah
[20:51] <sienap> and what will it do ?
[20:51] <sienap> suck on my nipples ?
[20:51] <BBB> it will make a video editor :P
[20:51] <sienap> :)
[20:51] Action: BBB has to go away
[20:51] <wtay> provide an API for non linear media editing
[20:51] <BBB> aaaaaaah
[20:52] <BBB> wtay: media + video+sound?
[20:52] <BBB> hmm.....
[20:52] <BBB> that first + was supposd to be a '='
[20:52] <sienap> hehe
[20:52] <wtay> e everything media related, yes
[20:52] <BBB> how can I possibly program if I cannot even type!!!
[20:52] Action: BBB likes the idea
[20:53] <wtay> omega wants to do the audio part
[20:53] Action: BBB has to go now - but I'll be back
[20:53] omega_ (omega at omegacs.net) joined #gstreamer.
[20:53] Nick change: omega_ -> omega_breakfast
[20:53] <BBB> I wanna help with the video part
[20:53] <wtay> hi omega_breakfast
[20:53] <BBB> and I think I have some other people willing to join too
[20:53] <BBB> (especially Philipp Zabel)
[20:53] <wtay> yeah, I'm sure
[20:53] <BBB> Philipp made the lavpipe idea in MJPEGtools
[20:53] Action: wtay volunteers omega_breakfast as the project leader...
[20:54] <BBB> hmm.....
[20:54] Action: BBB agrees
[20:54] Action: wtay plays an autoplugged mp3 pipeline with proper refcounting in -launch
[20:54] <BBB> how do I do that?
[20:54] Action: BBB tries a command-line thingy
[20:55] <wtay> ./gstreamer-launch disksrc location=/opt/data/mp3/files.cgi ! autoplugger ! osssink
[20:55] <BBB> where's gstreamerlaunch?
[20:55] <BBB> :)
[20:55] <wtay> the proper refcounting stuff is only on my disk though <g>
[20:55] <wtay> tools
[20:56] <BBB> ./gstreamer-launch disksrc location=/mp3/Korn\ -\ Freak\ on\ a\ Leash.mp3 ! autoplugger ! osssink
[20:56] <BBB> ***** GStreamer ERROR ***** in file gstscheduler.c at gst_bin_schedule_func:431
[20:56] <BBB> Element: /launch/osssink0.sink
[20:56] <BBB> Error: peer is null!
[20:56] <BBB> :)
[20:56] <wtay> cool
[20:57] <BBB> Couldn't create a 'Leash.mp3', no such element or need to run gstreamer-register?
[20:57] <BBB> :P
[20:57] <wtay> do you have libmad?
[20:57] <BBB> probably not
[20:57] <omega_breakfast> that's a cmdline parsing problem
[20:57] <wtay> try one without spaces in the filename :-)
[20:58] <omega_breakfast> it tries to create an element by that name
[20:58] <omega_breakfast> we need to add support for '\ '
[20:59] <BBB> Couldn't create a 'autoplugger', no such element or need to run gstreamer-register?
[20:59] <BBB> I don't have autoplugger?
[21:00] <wtay> nope
[21:00] <wtay> BBB: do you have the INCSCHED branch checked out?
[21:00] <BBB> without a space, I get lots of less errors :)
[21:00] <BBB> nope
[21:00] <BBB> only MAIN
[21:01] <BBB> how do I let it work without a autoplugger?
[21:01] <wtay> ok, then forget about what I said :)
[21:01] <wtay> ./gstreamer-launch disksrc location=/opt/data/mp3/files.cgi ! mad ! osssink
[21:01] <BBB> Couldn't create a 'mad', no such element or need to run gstreamer-register?
[21:01] <BBB> :P
[21:02] <wtay> ./gstreamer-launch disksrc location=/opt/data/mp3/files.cgi ! mp3parse ! mpg123 ! osssink
[21:02] <BBB> cool!!!
[21:02] <BBB> :)
[21:02] Action: BBB thinks this is as funny as your helloworld
[21:03] <BBB> can I use this to view video?
[21:03] <wtay> ./gstreamer-launch disksrc location=/opt/data/mp3/files.cgi ! mpeg1parse ! mp1videoparse ! mpeg_play ! aasink driver=3
[21:03] <BBB> ./gstreamer-launch disksrc location=/video/video.mpg ! mpgparse ! ....?
[21:03] <BBB> cool, let me try that one
[21:04] Nick change: taaz -> taaz-lunch
[21:04] <BBB> Segmentation fault (core dumped)
[21:05] <BBB> Couldn't create a 'aasink', no such element or need to run gstreamer-register?
[21:05] <wtay> BBB: you'll need libaa for that
[21:05] <BBB> where do I find that?
[21:06] <wtay> aalib.rpm and aalib-dev.rpm or something
[21:06] <BBB> hmm.....
[21:06] <BBB> ascii art?
[21:07] <wtay> yup
[21:07] <wtay> correction: über cool ascii ART
[21:08] <sienap> AmothafuckingAlib :)
[21:09] <BBB> and do I need to recompile gstreamer for that?
[21:09] <sienap> 10x
[21:09] <sienap> :)
[21:10] <wtay> BBB: only the aasink plugin
[21:10] <BBB> I installed the RPMs, re-ran gstreamer-register....
[21:10] <BBB> ok
[21:12] <BBB> ***** GStreamer ERROR ***** in file gstscheduler.c at gst_bin_schedule_func:431
[21:12] <BBB> Element: /launch/mp1videoparse0.sink
[21:12] <BBB> Error: peer is null!
[21:12] <BBB> ***** attempting to stack trace.... *****
[21:12] <BBB> ./gstreamer-launch disksrc location=/movies/korn.mpg ! mpeg1parse ! mp1videoparse ! mpeg_play ! aasink driver=3
[21:12] <BBB> :?
[21:13] <wtay> BBB: oh, yeah, sorry...
[21:13] <BBB> uhm...?
[21:13] <wtay> ./gstreamer-launch disksrc location=/opt/data/mp3/files.cgi ! mpeg1parse ! queue { mp1videoparse ! mpeg_play ! aasink driver=3 }
[21:14] <BBB> hmm.....
[21:14] <BBB> same error
[21:14] <wtay> ./gstreamer-launch disksrc location=/opt/data/mp3/files.cgi ! mpeg1parse video_00! queue { mp1videoparse ! mpeg_play ! aasink driver=3 }
[21:15] <omega_breakfast> you may have to escape the { }'s
[21:15] <omega_breakfast> \{ \}
[21:15] <omega_breakfast> and there's a missing ! after the queue
[21:16] <wtay> heh, yeah
[21:16] <BBB> now I don't see anything :P
[21:16] <BBB> but at least it doesn't crash
[21:16] <wtay>  ./gstreamer-launch disksrc location=/opt/data/armageddon1.mpg ! mpeg1parse video_00! queue ! { mp1videoparse ! mpeg_play ! aasink driver=2 }
[21:16] <wtay> works for me
[21:16] <BBB> ./gstreamer-launch disksrc location=/movies/korn.mpg ! mpeg1parse ! queue ! \{ mp1videoparse ! mpeg_play ! aasink driver=3 \}
[21:17] <BBB> hey, this is COOOOOOOOOOOOOOOOOOL
[21:17] <wtay> I don't use the \{
[21:17] <wtay> try driver=2 instead
[21:17] <omega_breakfast> also, it should be video_00! before the queue
[21:17] <hadess> what is the format for a double %ll ?
[21:17] <omega_breakfast> uh?
[21:17] <omega_breakfast> a guint64?
[21:18] <hadess> "double format, different type arg (arg 3)"
[21:18] ajzzzz (ajmitch at p11-max6.dun.ihug.co.nz) joined #gstreamer.
[21:18] <BBB> only difference between driver=2/3 is that =2 opens a new window, =3 plays it in the current terminal
[21:19] <ChiefHighwater> ello ajzzzz 8-]
[21:19] <BBB> ./gstreamer-launch disksrc location=/movies/korn.mpg ! mpeg1parse video_00! queue ! { mp1videoparse ! mpeg_play ! aasink driver=2 }
[21:19] <BBB> that works
[21:19] <omega_breakfast> hadess: %lf
[21:19] <wtay> BBB: oh?
[21:19] <hadess> omega_breakfast: oh, thanks
[21:20] <BBB> I prefer a normal movie window but it's still funny
[21:20] <hadess> warning: use of `l' length character with `f' type character <- not
[21:20] <BBB> how do I make it deliver me a 'normal' widnow?
[21:20] <omega_breakfast> hmmm
[21:20] <ajzzzz> crappy ISP....
[21:20] Nick change: ajzzzz -> aj_uni
[21:20] <aj_uni> internet connection goes at speed of a snail thru a tar pit ;)
[21:20] <BBB> videosink?
[21:20] <wtay> BBB: inpossible for now
[21:21] <wtay> s/in/im
[21:21] <hadess> omega_breakfast: i found my book, lemme check what this is
[21:21] <BBB> uhm..... but gstmediaplay can do that too, can't it?
[21:21] <wtay> BBB: yeah, but the command line tool can't
[21:21] <BBB> oh ok
[21:21] Action: BBB will go back to his homework for now
[21:21] <BBB> I gotta finish that
[21:21] Action: BBB is gone now
[21:22] <BBB> I liked the show, though ;)
[21:25] <hadess> omega_breakfast: it was whining because i needed a %Lu
[21:25] <hadess> omega_breakfast: unsigned long long
[21:26] <omega_breakfast> you said it was a double
[21:26] <hadess> omega_breakfast: my mistake
[21:27] <hadess> i have it working but it crashes right after the read()
[21:30] <hadess> memcpy()...
[21:30] <hadess> there's something wrong with my buffer...
[21:37] <hadess> omega_breakfast: can i send you the plugin, and you tell me what's wrong ?
[21:38] <omega_breakfast> sure
[21:38] <hadess> i fail to see what is wrong
[21:38] <hadess> it is very dirty right now, but if you can give me the hint to make this work, i'll clean it up and send in a proper patch
[21:38] <hadess> omega_breakfast: email ?
[21:39] <omega_breakfast> omega at temple-baptist.com
[21:39] Nick change: omega_breakfast -> omega_
[21:39] Nick change: taaz-lunch -> taaz
[21:40] <steveb> omega_: so what is left to do before incshed is merged?
[21:41] <hadess> sent
[21:41] <omega_> get the autoplugger built up enough so gstplay uses it
[21:42] <steveb> ah
[21:42] <hadess> making a widget certainly has helped understand the object system better
[21:43] <BBB> would you mind if I made a SDL-based videosink? :)
[21:43] <BBB> like aasink, but real
[21:43] <hadess> BBB: it would probably be slower than the xvideosink on X targets, but it would be nice indeed
[21:43] <BBB> hmm..... I could try that.....
[21:44] <omega_> ok, someone explain to me Yoann's comments? they don't make any sense
[21:44] <hadess> omega_: i didn't understand either :)
[21:44] justin_ (justin at h226-67.mahigan.albany.edu) joined #gstreamer.
[21:44] <omega_> yo
[21:44] Action: BBB will take a look at aasink as an example
[21:44] Action: justin_ lurks
[21:44] <taaz> heh.. the livid people are falling into here one by one ;)
[21:44] <omega_> intel beta compiler is available to developers
[21:45] <hadess> taaz: heh
[21:45] <taaz> omega_: does it have mmx/sse intrinsics?
[21:45] <omega_> I'd assume so
[21:46] <hadess> omega_: you received the plugin ?
[21:46] <omega_> just now
[21:47] <omega_> you're building in INCSCHED1?
[21:47] <hadess> HEAD
[21:47] <omega_> ok
[21:47] <omega_> I'll try it in incsched
[21:47] <hadess> should work equally well, i guess
[21:47] <hadess> (meaning not :P)
[21:48] <hadess> ./gstreamer-launch gnomevfssrc location=file:///home/hadess/Chuchurocket.mp3 ! mad ! fakesink <- this is an example launch
[21:48] <wtay> taaz: hmm, good, I'm not the only one having trouble with Yoann's comments...
[21:49] <taaz> wtay: I think he misunderstood something
[21:50] <omega_> ya think?
[21:50] <taaz> i'll try to clear it up...
[21:50] <hadess> </ironic>
[21:50] <omega_> good luck
[21:50] <omega_> I very much doubt you'll change his mind at all, judging from his comments
[21:51] <wtay> I think he meant to say that he"ll continue to hack on OMS until all probs are fixed...
[21:51] <omega_> is dent more, um, sane?
[21:51] <wtay> btw, what is he doing now?
[21:52] <taaz> i think there was some hacking style clashes in livid group.  different people trying to head in different directions and we ended up staying in one spot...
[21:53] <omega_> hmmmmm
[21:53] <taaz> i didnt really care how the code worked as long as it did and i could understand it
[21:53] <omega_> I would bet that you're much more likely to fork off OMS than you are to get them to move onto GStreamer, if Yoann's comments are at all representative
[21:53] <taaz> therefore i was unhappy cause i couldnt understand it ;)
[21:54] <hadess> omega_: i don't see what would really be left of oms in the end anyway
[21:54] <taaz> well, i have a large % of the vote of 'them' ;)
[21:55] <omega_> huh?
[21:55] <omega_> hadess: GST_DEBUG is your friend
[21:55] <hadess> omega_: hmm
[21:55] <hadess> omega_: what have i done wrong ?
[21:56] <omega_> dunno yet
[21:56] <taaz> hadess: i'm not sure either...  maybe it'll just be a nice front end to gst in the end... who knows.  i don't really care either as long as it works somehow.
[21:56] <hadess> taaz: and if it's a good one people will use it
[21:57] <omega_> um.... I'm not seeing gst_gnomevfssrc_open_file run
[21:57] <hadess> omega_: my change_state function is probably crappy
[21:58] Nick change: taaz -> taaz-meeting
[21:58] <hadess> omega_: i got it straight off the disksrc
[21:59] <wtay> omega_: scheduling refcounting needs a lot of work...
[21:59] <omega_> ok, I was doing something wrong too, but I fixed up the change_state while I was at it
[21:59] <omega_> wtay: which specifically?
[21:59] <hadess> cool
[22:00] <wtay> omega_: thread/create destroy has no leaks, when you add element things leak like hell
[22:00] <omega_> neat
[22:00] <omega_> doesn't surprise me
[22:00] <wtay> omega_: I'll try to track it down and try to understand a bit of it
[22:20] <hadess> omega_: you're looking at the gnomevfs ?
[22:20] <omega_> not atm, will in a sec
[22:21] <hadess> ok, just wanted to know
[22:26] <omega_> doh
[22:27] <omega_> you pass gnome_vfs_read the buf pointer
[22:27] <omega_> not the GST_BUFFER_DATA(buf) pointer
[22:27] <hadess> ooh
[22:27] <omega_> now it works perfectly <g>
[22:27] <hadess> excellent
[22:28] <hadess> you've modified something else ?
[22:28] <omega_> nothing critical
[22:29] <hadess> so i can clean up my sources, and integrate ?
[22:30] <omega_> there are some other things...
[22:30] <omega_> like at EOS, it malloc()'s until the cows come home
[22:30] <hadess> lol
[22:30] <omega_> moooo
[22:31] <hadess> ok, let me clean it up, i'll check on this when there's less code in there
[22:32] <omega_> ok, I'll mail you back what I have
[22:32] <omega_> rather, not
[22:32] <omega_> I'll email you change_state
[22:32] <omega_> addr?
[22:32] <hadess> hadess at hadess.net
[22:33] <omega_> sent
[22:33] <hadess> cool, thanks
[22:33] <omega_> oooh, /me streams a vorbis file from http and plays it live
[22:33] <hadess> using gnome-vfs ?
[22:33] <omega_> yup
[22:33] <hadess> :)
[22:34] sienap (synap at ipc379c21d.dial.wxs.nl) left irc: sienap has no reason
[22:34] <hadess> when this is fixed, do you think it would be interesting to have a sink as well ?
[22:34] <omega_> yup
[22:34] <omega_> which is why I'd suggest taking on the gstoss file structure
[22:35] <omega_> as well as the autoplugger single-source thing
[22:35] <omega_> merge the .h into the .c for each plugin, then have a single .c that is the plugin_init for them all
[22:35] <omega_> so you have a single plugin with N factories
[22:36] <hadess> huh, i don't quite understand
[22:37] <omega_> put the structures and such into gnomevfssrc.c
[22:37] <omega_> ditch gnomevfssrc.h
[22:37] <omega_> create gnomevfssink.c
[22:38] <omega_> create gnomevfs.c that does plugin_init and adds both of the src and sink factories
[22:38] <omega_> hrm, maybe oss doesn't do it that way
[22:38] <omega_> yeah, it does
[22:38] <omega_> libgstossaudio.so
[22:39] <omega_> so wtay.... thoughts on the capsnego thing?
[22:39] <wtay> hmm
[22:39] <wtay> if vorbisdec doesn do capsnego befor epushing a buffer, it is broken
[22:40] <omega_> ok, who's up for fixing that?
[22:40] <wtay> it doesn't?
[22:40] <omega_> it isn't acting like it
[22:40] <omega_> hack up your osssink like I describe, then do vorbisdec ! osssink
[22:40] Nick change: taaz-meeting -> taaz
[22:41] <BBB> omega_: when will we see your response to Yoann's weird e-mail?
[22:41] Action: BBB doesn't get that Yoann
[22:41] <omega_> am I or is taaz?
[22:42] <wtay> omega_: I know what's wrong...
[22:42] <taaz> um... i'll respond in a couple minutes
[22:42] <wtay> vorbisdec does set caps, capsnego finds out they are incompatible with osssink 
[22:43] <wtay> osssink padtemplate that is...
[22:43] trevo (trevo at joined #gstreamer.
[22:44] <BBB> ok
[22:44] <BBB> I'm kinda playing around with aasink, trying to make an sdlsink of it :)
[22:44] <wtay> so the osssink negotiate function isn't called and caps are not set, ultimatly disconnecting the pads...
[22:45] <omega_> hmmmm
[22:46] <omega_> ok, you're right
[22:46] <wtay> which reveals a problem with padtemplates btw
[22:46] <omega_> oh?
[22:47] <wtay> padtemplate compatibility check is worthless when a nego function is present
[22:47] <trevo> how well does gstreamer cvs work right now
[22:47] <trevo> ?
[22:47] <omega_> incsched is better than HEAD, but HEAD is more likely to play something with gstplay
[22:49] <trevo> has anyone had any success with head on redhat 7.1
[22:49] Action: omega_ is still rh7.0, and have been focusing on incsched, so I dunno
[22:50] <wtay> trevo: give it a try, it's supposed to work a bit
[22:50] Action: hadess hopes mp3 playing will work on ppc before 0.2
[22:51] <trevo> wtay: planning to
[22:51] <wtay> hadess: doesn't work with mad?
[22:51] <hadess> wtay: nope, plays even worse than mpg123
[22:51] <omega_> suck
[22:51] <wtay> damn
[22:51] <omega_> try madplay cmdline?
[22:52] <hadess> madplay worked, that's why i was very interested when wtay made the plugin
[22:52] <wtay> ok, then the plugin is broken
[22:52] <omega_> hmmmm
[22:53] <wtay> hadess: sound is outputted in little endian :(
[22:53] <hadess> wtay: you sure ? sndstat tells me it tries BE
[22:53] <wtay> hadess: you live in INCSCHED or HEAD now?
[22:53] <hadess> HEAD from a week ago
[22:54] <wtay> hadess: ok, I'll commit a patch in HEAD
[22:54] steveb (steveb at node1ee03.a2000.nl) left irc: [x]chat
[22:54] <hadess> wtay: cool, thanks
[22:54] <hadess> omega_: how can i fix this EOS problem ?
[22:55] <omega_> check for end of file, and set the EOS flag
[22:55] <omega_> as per the some of hte commented code
[22:55] <omega_> typically readbytes < bytes_per_read means EOS, and readbytes == 0, is guaranteed, but by then you're kinda late, and would have to send a zero-byte buffer
[22:55] <omega_> (event system will make this not be a problem, since you'd never end up sending that zero-byte buffer)
[22:56] <trevo> what is the cvs  command to try incsched
[22:56] <omega_> cvs update -r BRANCH-INCSCHED1
[22:56] <trevo> thanks
[22:56] <hadess> omega_: ok, got it
[22:57] <hadess> read: End of file, readbytes: 0 <- neat
[22:58] <omega_> but you should try to catch the 0 > readbytes < bytes_per_read case first
[22:58] <hadess> omega_: i don't need that, gnome-vfs tells me when it's EOF
[22:58] <omega_> also, we need a better name for bytes_per_read....
[22:58] <omega_> hmm, ok
[22:58] <omega_> but sending a zero-byte buffer is a bad idea, should be avoided
[22:59] <hadess> true
[22:59] Nick change: aj_uni -> ajmitch
[22:59] <omega_> if you can get and EOF indication *before* you send the last buffer, that's better
[22:59] <omega_> at least for now
[22:59] <hadess> i think it works now
[22:59] <hadess> what did you do to test it ? gst-launch, and wait til the end of the file ?
[22:59] <omega_> yup
[23:00] <hadess> and it continued "playing" ?
[23:00] <hadess> didn't exit i mean
[23:00] <omega_> right
[23:00] <hadess> so it works now :)
[23:00] <omega_> right
[23:01] <omega_> but then it sucked up all my RAM
[23:01] Action: omega_ was lucky to have caught it before it OOM'd his machine
[23:01] <hadess> true
[23:01] <hadess> sorry about that
[23:01] <omega_> heh
[23:01] <omega_> it's all your fault!!!! <g>
[23:01] ajmitch_ (ajmitch at p15-max2.dun.ihug.co.nz) joined #gstreamer.
[23:01] <BBB> how do I display errors in a sink?
[23:01] <BBB> for example, if SDL refuses to initialize....
[23:01] <BBB> GST_ERROR()
[23:01] <BBB> ?
[23:02] <omega_> for now,yes
[23:02] <ajmitch_> hey all, i'm back ;)
[23:02] <omega_> we need to rethink that stuff a bit
[23:02] <BBB> GST_ERROR("bla");
[23:02] <BBB> ok
[23:02] <omega_> everyone run for your lives!
[23:02] <omega_> dblcheck the gstinfo.h definition
[23:02] <hadess> NOOOO !!! not color codes !
[23:02] <BBB> GST_ERROR("Couldn't initialize SDL: %s\n", SDL_GetError());
[23:02] ajmitch (ajmitch at p11-max6.dun.ihug.co.nz) left irc: Ping timeout for ajmitch[p11-max6.dun.ihug.co.nz]
[23:02] <BBB> can it handle printf-like error messages?
[23:02] <omega_> it might want a element pointer first
[23:03] <omega_> yes
[23:03] <BBB> ok, thanks
[23:03] <hadess> i'm adding a check for gnome-vfs to the configure.in
[23:03] <omega_> ok
[23:03] <omega_> I put it in plugins/gnomevfs/ for now
[23:04] <BBB> some other thing.....
[23:04] Nick change: ajmitch_ -> ajmitch
[23:04] <BBB> if it refuses to initialize, is there an easy way to tell the sink to just kill itself?
[23:04] <hadess> ok, bbl
[23:04] <BBB> or do I need to keep variables to do that for me?
[23:04] Nick change: hadess -> hds-food
[23:04] <wtay> hds-food: commited a patch to try
[23:05] <hds-food> wtay: ok, thanks, will try that asap
[23:05] <omega_> BBB: just refuse to set state to READY
[23:05] <BBB> ok
[23:06] <omega_> return GST_STATE_FAILURE
[23:06] Action: omega_ thinks its time to start building the PWG in the wiki
[23:07] <wtay> ouch... the wiki? I hope it's only for making a draft...
[23:07] <omega_> yes
[23:08] <wtay> :)
[23:08] <omega_> but it's a lot easier to build it there, and for people to add to it, than to do it in db
[23:08] <wtay> I agree
[23:10] Action: BBB <--- gst-newbie
[23:10] <omega_> ok, gstreamer.net/wiki/PluginWritersGuide
[23:11] <omega_> initial material goes into the TOC as an unplaced chapter, we can reorg later
[23:11] <omega_> nodes start with PWG, must be [""]'d
[23:12] <omega_> start adding chapter titles to the TOC, we'll fill them in
[23:12] <omega_> add unformed requests at the bottom, after a -----
[23:12] <omega_> taaz: can you chmod the gstreamer.net .htaccess file to group-writable?
[23:13] <omega_> and add:
[23:13] <omega_> RewriteRule wiki /cgi-bin/wiki/moin.cgi/
[23:14] <omega_> oooh, Dan replies to Yoann <g>
[23:15] <BBB> gotta read that :)
[23:15] <trevo> is it normal for the gstreamer to eat up my memory when I run autogen.sh
[23:15] <wtay> hmm, interesting problem...
[23:15] <omega_> yeah
[23:15] <omega_> it's an autogen.sh bug
[23:16] <omega_> rather
[23:16] <omega_> it's an automake bug, sorry
[23:16] <omega_> the toplevel GStreamer README has a patch that fixes automake
[23:16] <wtay> element is destroyed, pads destroyed, pads disconnected, pad unparented, scheduler desconnected (fails because pad has no parent)
[23:16] <omega_> uh?
[23:17] <omega_> what scares me the most about Yoann's comments is his assumption that taaz would go about by forking gstreamer
[23:18] <wtay> wow, kino going gstreamer too?
[23:18] <wtay> well, some parts of it maybe...
[23:19] <wtay> omega_: destruction is not trivial it seems...
[23:20] <omega_> nope
[23:20] <omega_> now you know why I didn't complete it originally
[23:20] Action: ChiefHighwater hands wtay a bigger gun
[23:20] <ChiefHighwater> try again 8-]
[23:20] Action: wtay gets a rope
[23:20] <omega_> for it or you?
[23:21] <BBB> wtay: everyone is going gstreamer 8))
[23:21] <BBB> including MJPEG (lav?!?!?!?!?) tools
[23:21] <omega_> what does 'lav' mean anyway?
[23:21] <ajmitch> gstreamer is taking over?
[23:21] <wtay> omega_: depends if "it" can be fixed <g>
[23:21] <wtay> linux audio/video?
[23:23] <wtay> omega_: ok, in shutdown I will logically remove the element (like gst_bin_remove) while still keeping a ref to it, in destroy I unref it
[23:23] <omega_> 180 ;-)
[23:24] <omega_> Yoann figured it out
[23:24] <wtay> cool
[23:24] Action: omega_ responds
[23:25] <BBB> wtay: lav = linux audio/video or linux audio/visual
[23:26] Action: BBB wonders what a GstBuffer is
[23:26] trevo (trevo at left irc: Read error to trevo[]: Connection reset by peer
[23:26] <BBB> is that YUV data?
[23:26] Uraeus (cschalle at c224s9h5.upc.chello.no) left irc: syntax error - user imploded
[23:26] <wtay> BBB: just data, can be anything
[23:26] <BBB> hmm.....
[23:26] <BBB> can I specify it to be YUV data?
[23:26] <BBB> :)
[23:27] <wtay> BBB: you specify what it is using caps (key value pairs)
[23:27] <wtay> BBB: actually the type is determined by the pad that creates the buffer
[23:27] <BBB> hmm.....
[23:28] <BBB> so let's say I have a nice sdlsink *L* and I want YUV data from the mpegsrc - any suggestion on where to start reading? :)
[23:28] <wtay> BBB: for example: mpeg_play has a src pad with video caps (I420) so the buffers that come out of this pad are YUV frames
[23:28] <wtay> BBB: aasink only takes I420 video for example
[23:28] <BBB> that's good
[23:28] <BBB> mine is based on AA
[23:28] <wtay> ok
[23:29] <BBB> so can I just assume that it is I420 tgen?
[23:29] <BBB> g=h
[23:29] <wtay> yes
[23:29] <omega_> for the moment, yes
[23:29] <omega_> you'll want to later add caps for all the formats that SDL recognizes
[23:29] <BBB> I'll do that later on :)
[23:29] <BBB> for now, it's an exercise
[23:29] <wtay> yup
[23:30] <wtay> BBB: look at the template factory in aasink, this one is used to build a pad with the caps you want
[23:30] <BBB> ok
[23:30] <wtay> BBB: no element is allowed to connect to this pad if it cannot provide the same caps
[23:31] <omega_> or compatible
[23:31] <omega_> the autoplugger is then [optionally] responsible for building the pipeline needed to bridge if they don't
[23:32] <hds-food> hmm, i think i found a shorter way to check for gnome libs
[23:32] <omega_> gnome-config
[23:32] Nick change: hds-food -> hadess
[23:32] <omega_> or even the .m4
[23:32] <hadess> yeah, but shorter than what's in configure.in right now
[23:32] <omega_> but /me wants to use pkg-config now! /me whines
[23:32] <hadess> heh
[23:33] <omega_> I built some of pkg-config's features over 1.5yrs ago...
[23:36] x-virge (julian at adsl-151-204-74-141.delval.adsl.bellatlantic.net) joined #gstreamer.
[23:36] <hadess> hey x-virge
[23:36] <x-virge> hi
[23:36] <wtay> yo
[23:37] <x-virge> grabbing gstreamer CVS... it's been many months since the last time I played with gstreamer, decided to try it again now ;)
[23:38] <BBB> but uhm.... since YUV (I420 too?) is supposed to be a *char[3], Buf is a *Buf, how do I get the YUV char-arrays out of Buf?
[23:39] Action: BBB is really a newbie, I know
[23:39] <omega_> you use height*stride as an offset
[23:39] <omega_> see dvdec.c
[23:39] <hadess> dammit dammit !
[23:39] <omega_> hadess: ?
[23:39] <BBB> omega_: ok, dvdec.c, /me will look
[23:39] <hadess> the autogen.sh script nearly OOM my box
[23:39] <omega_> patch automake
[23:39] <hadess> omega_: it's all your fault !
[23:39] <omega_> no it isn't
[23:39] <hadess> oh, ok
[23:39] <ajmitch> hadess: hey, get more ram ;)
[23:40] <hadess> ajmitch: i have 320 megs
[23:40] <ajmitch> omega_: what version of automake has the patch?
[23:40] <omega_> it's people like you that add more checks and SUBDIRS to plugins/Makefile.am that cause automake to have a cow
[23:40] Action: ajmitch hasn't had any problems...
[23:40] <omega_> ajmitch: I dunno if even CVS HEAD has it ;-(
[23:40] <omega_> you must have applied the patch then
[23:40] <omega_> see README, check it out
[23:40] <ajmitch> nope ;)
[23:40] <ajmitch> how much ram does it usually try and use?
[23:41] <omega_> these days, hadess was looking at almost 256MB usage
[23:41] <omega_> maybe 512MB
[23:41] <ajmitch> hmm
[23:41] <omega_> it's exponential, doubles with every check you add to plugins/Makefile.am
[23:41] <ajmitch> i have 384MB ram, 270MB swap
[23:41] <ajmitch> maybe cos i'm not editing plugins/Makefile.am? ;)
[23:42] <omega_> yup
[23:42] <omega_> you'll regret not patching soon though
[23:43] <omega_> reply sent
[23:46] <omega_> ok, is anyone gonna add topics to the PWG wiki page??
[23:46] Nick change: ajmitch -> aj_uni
[23:47] <wtay> omega_: maybe we can start with the topics richard has set up in the docboook version
[23:47] <omega_> yeah, to some extent
[23:48] <omega_> but I want everyone who's writing plugins right now to post their ideas....
[23:51] Action: wtay is writing some stuff
[23:53] <x-virge> ick
[23:53] <omega_> ?
[23:53] <x-virge> jade outputs tons of errors and it stops the compile... failing to build the manual isn't supposed to stop the compile
[23:53] <omega_> however, docs are the last thing to build...
[23:54] <omega_> we still need to turn off docs-build by default.....
[23:54] <x-virge> ok
[23:54] <x-virge> well, all that needs to be done is prefix the jade commands in the Makefile.am's with "-"
[23:54] <omega_> true
[23:54] <omega_> but I'd rather not even try by default...
[23:54] <omega_> if someone wants to build docs themselves, they can turn it on <g>
[23:55] <x-virge> heh, if you say so :)
[23:55] <omega_> we have snapshots on the web that are built by a machine that's properly configured (which is really hard)
[23:57] <BBB> grrrrrrrrrr
[23:57] <BBB> sdlsink.c:253:61: not enough arguments for macro "GST_ERROR"
[23:57] <BBB> !!
[23:57] Action: BBB feels stupid
[23:57] <omega_> I warned you to check the gstinfo.h definition
[23:57] <BBB> lol
[23:58] Action: wtay is going to sleep, he has to get up early...
[23:58] <wtay> cya gys
[23:58] <omega_> bleck
[23:58] <wtay> er guys
[23:58] <omega_> ok, l8r
[23:58] Nick change: wtay -> wtay-sleeping
[23:58] <wtay-sleeping> I added something to the wiki...
[23:58] <omega_> ok
[23:58] <omega_> where?
[23:58] <wtay-sleeping> will add more :-)
[23:59] <wtay-sleeping> plugin guide
[23:59] <omega_> uh?
[23:59] <omega_> node?
[23:59] <omega_> oh
[23:59] <omega_> hrm, I wanted toc entries <g>
[00:00] --- Wed May 16 2001
[00:00] <omega_> oh well
[00:06] <BBB> woohoo! it compilet
[00:06] <BBB> :)
[00:06] <BBB> t/d
[00:11] <BBB> <-- stupid question again
[00:11] <BBB> since these are all .so files
[00:12] <BBB> it complains about unresolved symbols - although I did do -lSDL...
[00:16] <hadess> use sdl-config
[00:16] <x-virge> um
[00:16] <x-virge> why does the arts plugin hardcode the directory arts is in?:)
[00:17] <hadess> sdl-config --libs
[00:17] <omega_> because I was lazy <g>
[00:17] <x-virge> that's a way to annoy people :)
[00:17] <omega_> yup
[00:18] Action: omega_ is good at that <g>
[00:20] <BBB> grom
[00:20] <BBB> where do I define HAVE_LIBSDL (macro)?
[00:20] <BBB> plugins/Makefile.am:91: HAVE_LIBSDL does not appear in AM_CONDITIONAL
[00:20] <omega_> acconfig.h
[00:20] <BBB> I did
[00:20] <BBB> it doesn't take it
[00:20] <BBB> :|
[00:20] <omega_> rerun autogen?
[00:20] <omega_> oh, wait
[00:20] <BBB> did that too
[00:21] <hadess> BBB: edit configure.in and look for AM_CONDITIONAL
[00:21] <BBB> this happens WHEN running autogen.sh
[00:21] <BBB> :)
[00:21] <omega_> dnl # Set automake conditionals #
[00:21] <omega_> there's a whole long list of them
[00:21] <hadess> BBB: as a starter, hardcode the libs and cflags in your Makefile.am
[00:23] <BBB> I'll first try the official way :)
[00:23] Action: BBB notices his swap is full
[00:24] <hadess> hmm, i think gweather is broken, says it's -17 degrees...
[00:24] <omega_> oops
[00:24] <omega_> if it says -17K, then it's *really* broken <g>
[00:24] <hadess> BBB: make it the easy way first, why do you want to do complicated ?
[00:24] <hadess> omega_: hehe
[00:24] <omega_> complicated is fun !
[00:24] <BBB> hadess: good question ;)
[00:25] Action: BBB agrees with omega
[00:25] <BBB> if you start easy, you will never finish
[00:25] <hadess> BBB: BS, that's what i just did, I first made the plugin work before poking with configure.in
[00:26] <BBB> good idea, now that you sy it ;)
[00:27] <hadess> an autoconf check is just a nicety like documentation :P
[00:27] <BBB> creating plugins/sdlsink/Makefile <-- woohoo! :P
[00:27] <hadess> checking for Gnome VFS... found <- hehe
[00:28] <omega_> now... you two are the ones I want to get ideas on the PWG sections from.....
[00:28] <BBB> me?
[00:28] Action: BBB <-- newbie
[00:28] <hadess> huh, what kind of ideas you want ?
[00:28] <BBB> what functions do I need, what should they do?
[00:28] <omega_> yup.  the two 'new guys' building plugins
[00:29] <BBB> simple
[00:29] <omega_> you are asking the questions that need to be answered in the pwg
[00:29] <BBB> I just copied it from aasink now - but I'd rather know what I'm actually doing *L*
[00:29] <hadess> lol
[00:29] <omega_> right.  so add your comments to the pwg wiki
[00:29] <BBB> so explain what functions are needed in a plugin and what's needed
[00:29] <omega_> wiki
[00:29] <omega_> <g>
[00:30] <hadess> url ?
[00:30] <omega_> gstreamer.net/wiki/PluginWritersGuide
[00:31] <hadess> oh, i need to put everything in one file ? ok, doing that
[00:32] Action: BBB too
[00:32] <omega_> don't need to, but that's what I suggest now that I've done it a few times
[00:32] <omega_> makes things a lot easier
[00:34] <BBB> my god I have a headache
[00:34] Action: BBB needs to go to bed
[00:34] <omega_> heh
[00:34] <BBB> i'll write something down tomorrow morning, ok?
[00:35] <omega_> ok
[00:35] <omega_> taaz: there's been a jump in gst-devel subscribes afaict, in the last few hours <g>
[00:35] <taaz> heh ;)
[00:35] <hadess> hehe
[00:35] <taaz> just doing my part 
[00:36] <taaz> but you know what this means....
[00:36] <taaz> now i actually have to do this gst core for oms code ;)
[00:37] <taaz> i kinda wanted to put off even mentioning it on the livid list until i had it working
[00:39] x-virge (julian at adsl-151-204-74-141.delval.adsl.bellatlantic.net) left #gstreamer (oAo).
[00:39] <omega_> hehe
[00:40] <omega_> what you should do is write up a doc with your idea of how it would be structured
[00:40] <omega_> as far as navigation, ui, etc.
[00:43] <hadess> i wanna do a dvd player as well :)
[00:43] <hadess> maybe i'll end up just doing a skin, which would be a better use of my time :P
[00:43] <ChiefHighwater> l8rz y'all 8-]
[00:44] ChiefHighwater (paul at temple-baptist.com) left irc: 
[00:44] <hadess> ChiefHighwater: see ya l33t m4d m4n
[00:49] <taaz> heh... I think I may be getting yoann to beleive in gst ;)
[00:49] <omega_> it may be
[00:49] <omega_> what might be useful is a comparison between the oms and gst buffering systems
[00:49] <omega_> and defining the distinction between buffer management buffering/queueing
[00:50] <omega_> er, buffer management and buffering/queueing
[00:50] <hadess> everybody loves gst :)
[00:50] <taaz> well, I have no idea how the current OMS code works so I can't comment on it too well...  I just know it confuses me
[00:50] <BBB> except M$
[00:50] <BBB> :P
[00:50] <hadess> heh
[00:50] <taaz> did M$ make an official comment on gst? ;)
[00:50] <omega_> nope
[00:51] <omega_> but I was thinking: if M$ adopts gstreamer over directshow, then we can truly claim to have taken over the world <g>
[00:51] <hadess> lol
[00:52] Action: BBB awaits M$ to officially declare gst evil, opensource and cheap
[00:52] <hadess> i'm recompiling today's cvs, i'll send the patch to you
[00:52] <omega_> ok
[00:52] <BBB> M$' best comment
[00:52] <BBB> opensource software is open
[00:52] <BBB> so everyone can see the source
[00:52] <BBB> with Microsoft code, this is not the case
[00:52] <BBB> which makes M$ software much safer
[00:52] <omega_> ah, right
[00:52] <BBB> makes sense, uh? :P
[00:53] Action: omega_ thinks how that works, gets a headache
[00:53] Action: BBB already had a headache anyway
[00:53] Action: BBB gives omega_ an aspirin
[00:53] Action: omega_ hands BBB some Excedrin
[00:53] <BBB> sounds unhealthy?
[00:53] <hadess> omega_: will do a sink tomorrow, while i'm in the code
[00:53] <omega_> uh?  asprin on steroids, sorta <g>
[00:54] <omega_> hadess: ok
[00:54] <hadess> i'm wondering if i can use mmap with gnome-vfs, or even if it already does in the core
[00:55] <omega_> that's what I was thinking
[00:55] <omega_> it'd be better to use that if it's available
[00:55] <hadess> yep
[00:56] <omega_> also, it could set mime type...
[00:56] <omega_> if not any properties
[00:56] <hadess> it is possible to do that
[00:56] Action: BBB just successfully compiled gstreamer including his nice sdlsink
[00:57] <omega_> no instance of 'mmap' in the gnomevfs headers ;-(
[00:57] <BBB> now I need to wait 10 minutes for make install and gstreamer-register :|
[00:57] <hadess> BBB: but does it work ?
[00:57] <omega_> the ops doesn't look promising
[00:57] <omega_> waitasec
[00:57] <BBB> hadess: /me prays
[00:57] <omega_> why are you doing a make install??
[00:57] <BBB> I dunno - probably because I don't see a reason not to :)
[00:57] <omega_> don't
[00:57] <omega_> unless you're developing an outside app, doing a make install after autogen.sh causes nothing but trouble and a massive waste of time
[00:58] <BBB> true
[00:58] <omega_> remember to make uninstall then to clean up what's there
[00:58] <omega_> and you may have to manually clean some stuff
[00:59] Action: BBB is too lazy
[00:59] <omega_> once again, I would recommend that you check it out <g>
[00:59] <hadess> omega_: that should be on the CVS page, with the other infos, in big caps
[00:59] <BBB> woohoo!!!!!!!!!!!!!!!!!!!!!!!!!!!!
[00:59] <BBB> it works :)
[00:59] <BBB> just one thing
[00:59] <BBB> the thing is flickering
[00:59] aj_uni (ajmitch at p15-max2.dun.ihug.co.nz) left irc: Ping timeout for aj_uni[p15-max2.dun.ihug.co.nz]
[00:59] <hadess> BBB: use double-buffering
[00:59] <omega_> wow
[00:59] <BBB> it's green.... then movie.... then green again
[01:00] <omega_> hadess: good point
[01:00] <BBB> I cannot do double buffering, SDL should do that
[01:00] <BBB> :)
[01:00] <hadess> BBB: nope, you need to tell SDL to do that
[01:00] <omega_> SDL shares many of Xv's flaws ;-(
[01:01] <BBB> hadess: and do you by accident know how?
[01:01] <hadess> BBB: i have some code somewhere in a LJ issue, lemme look if i can find a url
[01:03] <BBB> ok
[01:03] <hadess> ftp.ssc.com/pub/lj/listings/issue81
[01:03] <omega_> 4401
[01:05] <BBB> which file should I get?
[01:05] <omega_> 4401
[01:05] <BBB> 4401?
[01:05] <BBB> :)
[01:05] <BBB> ok
[01:05] arik (arik at sdn-ar-001waseatP205.dialsprint.net) joined #gstreamer.
[01:05] Action: BBB didn't look
[01:05] <BBB> lol
[01:05] <arik> lo
[01:05] <omega_> yo
[01:05] <hadess> hey arik
[01:05] aj_uni (ajmitch at p15-max2.dun.ihug.co.nz) joined #gstreamer.
[01:06] <arik> hadess, omega_: hey
[01:06] <hadess> arik: you the arik i saw in the gnome summary ?
[01:07] <arik> hadess: i was in the gnome summary?
[01:07] <taaz> BBB: not sure if it will help but you can look at oms sdl driver: http://www.linuxvideo.org/cvs/viewcvs.cgi/oms/src/plugin/output/video/video_out_sdl.c?rev=1.9&content-type=text/vnd.viewcvs-markup
[01:07] Action: arik is going to go read it
[01:08] <hadess> "Arik Devens who many of you will know worked at Ximian last summer is currently working for Netscape hacking Mozilla"
[01:08] <arik> hehe
[01:08] <arik> yeah that is me
[01:09] <hadess> cool :)
[01:09] <arik> ;-)
[01:09] <arik> i wonder how that happened
[01:09] <hadess> MAKE IT FASTER !
[01:09] <omega_> amen!
[01:09] <taaz> omega_: saw your comment last night about OLS.  you going?  i am.
[01:09] Action: hadess d/ls new galeon
[01:09] <omega_> lemme send out mail
[01:10] <arik> hadess: ;-)
[01:10] <arik> hadess: we are working on it
[01:12] <omega_> I get noticable keyboard lag just typing in a new location with moz0.9
[01:13] <hadess> omega_: lat.c in the test dir contains asm, and doesn't compile gere...
[01:13] <hadess> s/gere/here
[01:13] <omega_> ok, taaz....
[01:13] <arik> omega_: there are projects to do real speedup in the works
[01:13] <omega_> you wanna put that behind an x86 define? <g>
[01:13] <omega_> my main issue with moz is the fact that it has yet another object model
[01:13] <arik> it made sense at the time
[01:14] <taaz> hadess: my bad ;)
[01:14] <omega_> is it going to change?  it should be CORBA, like everything else
[01:14] <arik> xpcom is actually pretty good
[01:14] <taaz> actually... i fixed that but havent checked it in
[01:14] <omega_> but unless there are fast bridges to/from corba, it's a pain
[01:14] <arik> omega_: *nod*
[01:14] <taaz> i'll do that later... just hack it right now
[01:16] <omega_> unfortunately, KDE chose th 'it made sense at the time' path too, and now the interop stuff is stalled due to object-model differences
[01:16] <arik> omega_: yep
[01:16] <arik> omega_: i know
[01:16] <arik> omega_: at one point i wrote a nautilus view that did kparts plugins
[01:16] <omega_> eek
[01:16] <arik> was an intresting challenege
[01:16] <omega_> the arts wrapper gst plugin is messy too
[01:16] <arik> no doubt
[01:17] <hadess> wooohoo ! mad works on PPC !
[01:17] Action: omega_ watches ryhthmbox start playing audio
[01:17] <arik> hadess: how much longer till we see your project ;-)
[01:18] <hadess> arik: will take some time, but it's getting better
[01:18] <arik> hadess: ;-)
[01:18] <arik> hadess: i am curious
[01:19] <hadess> arik: i'm still writing the core, as soon as i am sure what i'm doing will work, i'll commit to gnome cvs
[01:20] <arik> hadess: ;-)
[01:20] <hadess> and i received the first artwork for it today, it will look _good_
[01:21] <arik> sweet
[01:21] Action: arik loves ui stuff
[01:22] <omega_> I saw GTM for the first time last night.... wrote up 50-some specific points that should be changed or added to the GUI
[01:22] <omega_> sent it to the author, fwd'd to Anna to put wherever appropriate
[01:22] <arik> gtm?
[01:22] <arik> gnome transfer manager?
[01:22] <taaz> anyone tried v4lsrc lately?
[01:22] <omega_> arik: yup
[01:22] <omega_> taaz: nope
[01:23] <hadess> taaz: it is broken
[01:23] <arik> omega_: ugly and lame
[01:23] <taaz> hadess: how so?
[01:23] <omega_> arik: yeah, but I think I know how to make it not ugly or lame
[01:23] <hadess> taaz: no caps wtay told me
[01:23] <omega_> I thought he added caps
[01:23] <hadess> did he ?
[01:24] <omega_> arik: email addr?
[01:24] <arik> omega_: good deal ;-)
[01:24] <arik> omega_: mine?
[01:24] <omega_> yup
[01:24] <arik> omega_: arik at gnome.org
[01:24] <omega_> fwd'd
[01:24] <hadess> hmm, broken
[01:24] <arik> omega_: danke
[01:25] <hadess> omega_: me too
[01:25] <omega_> done
[01:25] <omega_> taaz: how long is ols?
[01:27] <taaz> I think I have an excuse to do image processing on a custom pci board with FPGAs and other stuff on it.  ;)  might as well do it with gstreamer
[01:27] <omega_> hehhe
[01:27] <taaz> omega_: i dunno... check the web site  its a few days
[01:27] <omega_> 4
[01:28] <hadess> omega_: it is not possible to mmap using gnome-vfs
[01:28] <omega_> that's what it looks like
[01:28] <omega_> that sucks
[01:28] <arik> omega_: where are you?
[01:28] <omega_> Portland, OR
[01:29] <arik> omega_: oh!
[01:29] <hadess> London, UK
[01:29] <hadess> :P
[01:29] <omega_> arik: ?
[01:29] <arik> omega_: i'm in seattle atm, live in sf though
[01:29] <omega_> cool
[01:29] <omega_> I work in Boise <g>
[01:29] <arik> omega_: i might be in portland soon for a day ;-)
[01:29] <arik> omega_: hehe
[01:29] <hadess> i love ppl who live in places that have more rain than here <g>
[01:29] <omega_> arik: when?
[01:30] <arik> omega_: sometime this month
[01:30] <omega_> cool, what for?
[01:30] <arik> omega_: a picnic actually ;-)
[01:30] <omega_> heh
[01:30] <arik> not sure yet
[01:31] <hadess> hmm, moz is eating all my cpu
[01:31] <arik> what a surprise
[01:31] <omega_> how comforting <g>
[01:31] <arik> ;-)
[01:32] <hadess> killall -9 mozilla-bin fixed it
[01:32] <omega_> I'm used to doing a killall -9 netscape-communicator a lot
[01:32] <arik> omega_: most of what you say makes sense, i would have to try out the app again to really get into it
[01:32] <omega_> I don't want to ever get used to killall -9 mozilla
[01:32] <arik> yeah me too
[01:32] <arik> heh
[01:32] <omega_> arik: re: hat?
[01:32] <hadess> i've uninstalled netscape 2 months ago
[01:32] <omega_> er, er: what?
[01:32] <arik> omega_: gtm
[01:32] <omega_> ok
[01:33] <omega_> I'm hoping it'll hit the usability or gnome-ui lists
[01:33] <arik> *nod*
[01:33] <omega_> I'd send it myself, but I dunno which, I'll let Anna deal with that
[01:33] <arik> heh
[01:33] <arik> anna
[01:33] <arik> heh
[01:33] <omega_> ?
[01:33] <arik> oh nothing
[01:33] <arik> nm
[01:33] <omega_> right...
[01:33] <arik> anna worked at ximian when i did
[01:34] <omega_> still does, afaik
[01:34] <arik> near the end
[01:34] <arik> yep
[01:34] <arik> i meant she started while i was there
[01:34] <omega_> ok
[01:34] <hadess> heh, anna <g>
[01:34] <arik> hehehe
[01:34] <omega_> hadess: down...
[01:34] <arik> ah yeah, that too
[01:35] <hadess> lol
[01:35] <omega_> from: Bart: subject: Eazel
[01:35] <omega_> I regret to inform you that Eazel is in fact shutting down.   Over the
[01:35] <arik> yep
[01:35] <arik> i knew about this
[01:35] <arik> from friends who still worked there
[01:35] <omega_> this is on gnome-hackers list
[01:35] <arik> yeah
[01:37] <hadess> omega_: can you print the whole msg ? i'm not on g-h
[01:38] <omega_> fwd'd
[01:38] <hadess> thanks
[01:38] <omega_> a good 2/3 of OLS timeline has talks I'd want to go to
[01:39] <arik> heh
[01:39] <arik> when is ols?
[01:39] <omega_> July 25-28
[01:39] <omega_> any chance you could make it?
[01:39] <arik> hmm
[01:39] <arik> maybe
[01:40] <arik> i'll see about it
[01:40] <hadess> omega_: you didn't say that the icons are ugly about gtm
[01:40] <omega_> yeah I did <g>
[01:40] <omega_> just not in so many words
[01:41] <arik> heh
[01:41] <hadess> you're so smooth <g>
[01:43] <omega_> latest on l-a-d promises to be interesting
[01:43] Action: hadess drools
[01:43] <omega_> ?
[01:43] <hadess> received more artwork :)
[01:44] <taaz> hadess: do we get to see or are you just going to tease us? ;)
[01:44] <arik> heh
[01:44] <hadess> taaz: exactly :)
[01:45] <omega_> hadess: you evil, evil person ;-)
[01:45] <hadess> actually, i can't save the icons inlined in the mail with Evo, so i can't even use them myself
[01:46] <arik> hadess: um
[01:46] <arik> hadess: evo used to do that
[01:46] <hadess> arik: it's inlined in an html mail
[01:47] Action: omega_ shudders
[01:47] <arik> hadess: ick
[01:47] <arik> hadess: gimp screengrab it out ;-P
[01:47] <omega_> bleck
[01:50] <hadess> arik: when i have the whole thing, i'll show a shot :)
[01:51] <arik> hadess: ;-)
[02:30] arik (arik at sdn-ar-001waseatP205.dialsprint.net) left #gstreamer.
[02:48] <hadess> omega_: sent patch
[03:03] kagedal (simon at smithix.cs.uoregon.edu) joined #gstreamer.
[03:10] Action: hadess off to bed
[03:10] <hadess> cya
[03:10] hadess (hadess at pc2-guil2-0-cust121.gui.cable.ntl.com) left irc: sleep
[03:32] Kuroyi (rick at ubr-35.82.223.wmelbourne.cfl.rr.com) joined #gstreamer.
[03:32] <omega_> yo
[03:32] <kagedal> hi omega
[03:32] <omega_> kagedal: enjoying the rain?
[03:32] <kagedal> omega: yeah, enough of that misplaced sunshine :)
[03:33] <omega_> heheh
[03:33] <kagedal> it's not supposed to be warm in may!  it's winter!  we have snow!
[03:33] <omega_> um
[03:33] <kagedal> in sweden, that is :)
[03:33] <omega_> it was snowing a month ago in Norway, at least
[03:33] <omega_> barely
[03:33] <kagedal> yeah
[03:34] <kagedal> i guess it's probably mostly gone now
[03:34] <omega_> yup
[03:34] <kagedal> so how was europe?
[03:34] <omega_> fu
[03:34] <omega_> er, fun
[03:34] <kagedal> i bet
[03:34] <omega_> bleagh
[03:35] <omega_> redhat seems to think that we don't want /usr/include/linux to point to the kernel tree
[04:22] <taaz> blidiblah!  how's about this:
[04:22] <taaz> ./lat 10000 100 500 simple 250
[04:22] <taaz> 0010000:00804367 min:00558390 max:04520063 avg:00588437 avg-s:0.001176874
[04:23] <taaz> yeah, that's 250 identity elements 
[04:23] <kagedal> latency testing?
[04:23] <taaz> yeah
[04:25] <taaz> but that's not the point of this test... the limit was 64 cothreads a few minutes ago
[04:25] <kagedal> ah
[04:27] <taaz> now i need to figure out why that worked
[04:33] <omega_> neat
[04:34] <omega_> what did you upgrade?
[04:34] <taaz> i fiddling with int * vs void *... just a minute... 
[04:36] <taaz> i don't quite understand how some of this code works yet
[04:36] <omega_> welcome to the club <g>
[04:37] <taaz> i wanted to set the stack size... since it looks like that capability is in the newer linuxthreads... but that only works on threads you create, not the root process
[04:38] <omega_> right
[04:38] <taaz> anyway to force stack size?  or even to query what it is?
[04:38] <omega_> and it seems that the instant you run any pthread function in the root thread, you can no longer use malloc() for stacks
[04:38] <omega_> afaict
[04:38] <omega_> even if you don't use threads at all
[04:39] <omega_> change the pthread_self() test to check against 1024 (the root thread ???) and check it out
[04:39] <Kuroyi> you know, it takes a long time to compile this
[04:39] <omega_> which?
[04:40] <omega_> and for some reason COTHREAD_STACKSIZE * COTHREAD_MAXTHREADS must < (pthreadstacksize / 4)
[04:40] <taaz> not anymore
[04:40] <Kuroyi> gstreamer from cvs
[04:40] <taaz> that's what i just fixed 
[04:40] <omega_> Kuroyi: oh, of course ;-)
[04:40] <taaz> or at least it doesnt segfault ;)
[04:40] <omega_> not anymore what?
[04:40] <taaz> that /4 part
[04:40] <omega_> what was it?
[04:41] <taaz> give me a few mins... i'm deconfusing myself
[04:43] <Kuroyi> I'm curious to see how the syncing is handled
[04:45] <omega_> so am I <g>
[04:53] <omega_> taaz: deconfused yet?
[04:53] <taaz> working on it ;)
[05:04] <omega_> wow.  livid seems to be rallying behind gstreamer <g>
[05:04] <Kuroyi> I was wondering when something would happen
[05:04] <Kuroyi> it's been pretty dead for a while
[05:04] <omega_> you came here from there?
[05:04] <Kuroyi> hmm, sortcl
[05:04] <Kuroyi> argh
[05:04] <Kuroyi> sortof
[05:05] <Kuroyi> I seem to work on the em8300 driver
[05:05] <omega_> taaz: you have your work cut out for you <g>
[05:05] <omega_> which card is that?
[05:05] <Kuroyi> dxr3
[05:05] <Kuroyi> hollywood plus
[05:05] <omega_> cool
[05:06] <Kuroyi> I'm going to look at doing a gstream plugin when I get some time
[05:06] <omega_> I just made an attempt to get my dxr2 working, but I need to either setup the bt848 to capture the composite, or get my second G400 head to display blue so there's a signal for the dxr2 to overlay on
[05:07] <Kuroyi> strangely enough the dxr2 is a completely different chipset/card
[05:07] <omega_> yup
[05:07] <omega_> and nowhere near as good, from what I hear
[05:07] <Kuroyi> yeah
[05:07] <omega_> I've never actually seen its output though <b>
[05:07] <Kuroyi> even though it does simultaneous tv out/overlay
[05:07] <omega_> hmm
[05:10] <Kuroyi> heh, I only have 5megs free in /usr now
[05:10] <omega_> oops
[05:10] <taaz> omega_: this is totally confusing me... let me send you a patch to distribute the confusion
[05:11] <omega_> doh
[05:11] <taaz> i made too many changes at once... i might not have needed them all
[05:13] <taaz> oh wait... i'm slightly less confused
[05:13] <taaz> ok, now i just dont know why it works...  that's ok ;)
[05:16] <omega_> oh, duh
[05:16] <omega_> I completely missed several *4's in there when I looked
[05:16] <taaz> which ones?  i think it worked before i removed some of them
[05:16] <omega_> *sizeof
[05:19] <taaz> is there some way to verify this stuff?  i'm not sure how to tell if its working out of luck or if it's actually correct
[05:19] <omega_> um...
[05:20] <omega_> somehow you need to allocate some amount of space on the stack, of variable size (to match the cothread stack) and write to it all
[05:20] <omega_> if a valid allocation and writing it all doesn't clobber the next cothread, you're safe
[05:22] <taaz> void *stack_end = (void *)((unsigned long)sp & ~(STACK_SIZE - 1));
[05:22] <taaz> what's the proper way to write that?
[05:22] <taaz> i don't think its 64 bit safe
[05:22] <omega_> not sure, lemme think about it
[05:28] kagedal (simon at smithix.cs.uoregon.edu) left irc: Leaving
[05:32] <taaz> well... if we use inttypes.h we can use uintptr_t
[05:33] <omega_> yeah
[05:33] <taaz> hmm... glib should really define something similar
[05:38] <taaz> i don't even understand what that line does... oh well
[05:39] <taaz> void *stack_end = (void *)((uintptr_t)sp & (uintptr_t)(~(STACK_SIZE - 1)));
[05:39] <omega_> it rounds it to the nearest unit of STACK_SIZE
[05:40] <omega_> which doesn't make any sense
[05:41] <omega_> lemme see what the latest code from OGI has
[05:48] <omega_> ah, I get it
[05:49] <omega_> the assumption is that the pthread has STACK_SIZE bytes of stack, and that it's aligned on STACK_SIZE boundaries
[05:49] <omega_> so it takes the current stack frame pointer and rounds it up to the the very end of this presumed location

More information about the gstreamer-devel mailing list