[gst-devel] Daily IRC logs

[wim.taymans at chello.be]

[07:49] Nick change: aj_uni -> ajmitch
[07:54] steveb (steveb at node1ee05.a2000.nl) joined #gstreamer.
[07:54] <omega_> yo
[07:54] <steveb> that was quick!
[07:54] <omega_> 2sec.. not bad
[07:54] <omega_> considering I had to move the mouse too <g>
[07:54] <ajmitch> hi
[07:55] <steveb> yo
[07:55] <omega_> 6sec ;-(
[07:55] <steveb> i typed that with my pinky - coffee in one hand, toast in the other
[07:55] <omega_> oh
[07:55] <omega_> you're excused, then ;-)
[07:56] <steveb> :)
[07:57] <ajmitch> so what're ppl up to tonight? ;)
[07:58] Action: omega_ still debugging various autoplugger scenarios
[07:58] Action: steveb is going to work
[08:02] Action: steveb steveb-work
[08:02] Nick change: steveb -> steveb-work
[08:02] <steveb-work> er, too early
[08:03] <ajmitch> hehe
[08:04] <ajmitch> hmm, gstreamer-devel list has been busy...
[08:04] <omega_> yup
[08:06] <omega_> brb
[08:18] <taaz> omega_: should I check this cothread stuff in?  I'm leaving the 64bit issue till later.  having inttypes.h dependency is another little configure detect nightmare for platforms without it
[08:19] <omega_> identical to the patch you sent me?
[08:20] <taaz> yeah plus a GST_DEBUG line in the max cothread failure part
[08:20] <ajmitch> taaz: you working hard to convince OMS developers to move to gstreamer? ;)
[08:20] <omega_> ok, did you outline that case?
[08:20] <taaz> outline?
[08:21] <omega_> move it to a goto failure, put the return and DEBUG at the end of the function
[08:21] <omega_> to compact the normal code path in memory
[08:21] <taaz> ajmitch: trying...  i'll just do the conversion and let the results speak for themselves ;)
[08:21] <omega_> ajmitch: they seem to be convincing themselves, even...
[08:22] <ajmitch> yeah, i noticed that ;)
[08:22] <ajmitch> gstreamer just keeps on absorbing everything in it's path....
[08:23] <taaz> omega_: the compiler doesn't optimize that for us?
[08:23] <omega_> nope
[08:23] <omega_> gcc 2.x doesn't outline, I doubt 3.0 does either
[08:23] <omega_> there need to be hints to do it effectively anyway
[08:23] <taaz> omega_: i usually just use that style with multiple fallthroughs with redundant cleanup code... 
[08:23] <omega_> there's a cosource proposal relating to this, I pasted the URL here for you once
[08:24] <omega_> taaz: yes, but it's relevant here too
[08:24] <omega_> oh wait
[08:24] <omega_> I'm thinking of the switch() case
[08:24] <omega_> switch() is heavily outlined
[08:25] <taaz> well.. cothread_create returning NULL isn't handled properly in gstschedule anyway...  it should really propagate out to avoid segfaults
[08:25] <omega_> right
[08:25] <omega_> to be fixed
[08:25] <omega_> if you want to add a // FIXME in those locations while you're at it, so we don't forget...
[08:27] <omega_> then, I'd change the STACKSIZE to 16384 and MAXTHREADS to 128
[08:27] <omega_> and then commit the patch
[08:27] <omega_> it'll be strange to have someone else showing in the cvs annotate for cothreads.c <g>
[08:27] <taaz> has anyone hit the 8k stack size yet?
[08:27] <omega_> dunno
[08:27] <omega_> could trace that at cothread_switch()
[08:27] <omega_> get some stats
[08:28] <taaz> yeah you could
[08:28] <omega_> hehee
[08:28] <omega_> it's simple math: compare top_sp (which I think I removed) with sp
[08:28] <taaz> top_sp is still there...
[08:28] <omega_> ok
[08:29] <omega_> then remove the is this needed comment from that while you're at it <g>
[08:32] <taaz> ok, dealing with the cothread_create failure touchs a bunch of code... the call stack to taht point is all functions with no return value.  not sure what to do on a failure in there
[08:32] <omega_> just FIXME it for now
[08:32] <taaz> ie, that an omega_-problem ;)
[08:33] <omega_> that reminds me, matth has been very impressed with the development model we have <g>
[08:33] <taaz> 224 FIXME lines in HEAD ;)
[08:33] <omega_> he tells me that at Microsoft (he worked there for HP doing laserjet drivers)...
[08:33] <omega_> they have exactly two comments:
[08:33] <omega_> BUG: probably wrong, likely to cause problems, should be revisited
[08:33] <omega_> BUG BUG: a bug
[08:34] <omega_> you can imagine how readable this makes code
[08:34] <omega_> he would have been shot at with nerf weapons at M$ for leaving comments in the code such as what he's put in INCSCHED1 <g>
[08:35] <ajmitch> hmm, ideas on adding volume slider to gstmediaplay?
[08:35] <omega_> it's a good idea, the trick is implementing it with the staticrenderautoplugger
[08:36] <omega_> dynamic autoplugger done right will make that much easier, see my next mail to -devel
[08:36] <ajmitch> i think arik was wanting to add support in, and i am apparantly meant to know how to get the volume plugin up to scratch ;)
[08:36] <omega_> hehehe
[08:41] <taaz> so, 16k stack, 128 cothreads?
[08:41] <omega_> yup
[08:42] <omega_> in the long run, we should make it support selectable stack size, with some base unit (say, 4kb)
[08:42] <omega_> and the context_init would check the pthread stack size
[08:42] <omega_> the Thread would have an arg that would let you set the stacksize attr before _create()
[08:43] <omega_> but we need cothread destruction and re-use before then
[08:43] <omega_> the scheduler isn't so good about that just yet
[08:51] <omega_> dangit
[08:51] <omega_> taaz: you have a SMP machine, right?
[08:52] <taaz> there's a possibility of breakage in gsti386.h i think... perhaps that GST_ARCH_SETUP_STACK should cast to unsigned long * like the other arch headers
[08:52] <taaz> yeah i do
[08:52] <omega_> can you run the program I just sent you?
[08:52] <omega_> gcc -o atomic-test atomic-test.c -lpthread
[08:54] <taaz> don't have it yet...
[08:54] <taaz> err.. got it
[08:55] Action: ajmitch hits check mail to see what else has unfolded on the gstreamer lists
[08:56] <taaz> what should this program do?
[08:56] <omega_> fail utterly
[08:56] <taaz> addthread ran 30585312 times, delthread ran 34128787 times
[08:56] <taaz> counter should be -3557301, it is -3474888
[08:56] <omega_> good
[08:56] <omega_> on a SMP machine those number should almost never be equal
[08:56] <omega_> I can't test on my machine because the threads never run at the same time
[08:56] <omega_> now I need to write atomic_lock and atomic_lock somehow, sometime
[08:57] <taaz> huh?
[08:57] <omega_> use an atomic var (LOCK prefix) as a barrier
[08:57] <omega_> a spinlock, basically
[08:58] <omega_> the problem is that asm/spinlock.h brings in several kernel headers I don't want (and may not be needed)
[08:58] <taaz> mutexes and so on not good enough?
[08:58] <omega_> too slow
[08:58] <omega_> have you ever traced through the mutex code in linuxthreads?
[08:58] <omega_> uck
[09:04] <omega_> taaz: how much later will you be up?
[09:04] <omega_> I might have a test for you soon
[09:04] <taaz> not too long ;)
[09:04] <omega_> can I log into your machine? <g>
[09:05] <taaz> would have to go indirect though my firewall...
[09:05] <omega_> not hard <g>
[09:05] <omega_> that's what wtay has to do to get to my laptop
[09:05] <ajmitch> hehe
[09:05] <ajmitch> taaz: i guess you trust him...
[09:12] steveb (stevebaker at 212186169160.chello.com) joined #gstreamer.
[09:12] <steveb> yo
[09:12] <ajmitch> hi steveb ;)
[09:12] <ajmitch> bye all ;)
[09:13] Nick change: ajmitch -> ajbusy
[09:17] <omega_> ok, I think I have it worked out in my head, I can code it in 5min for you to test it
[09:17] <taaz> ok
[09:19] <omega_> grr, I dunno how this algo is gonna work with more than one contender
[09:19] <omega_> but it'll work for 2
[09:21] <omega_> ok, in the mail
[09:22] <omega_> i might have to flip it over somehow to get it to work in the >2 contenders case
[09:23] <taaz> i thought locking issues were taken care of?
[09:24] <omega_> they are, but these are faster spinlocks, which I want to use for the per-cothread locks
[09:24] <omega_> with zero contention, mutexes take a long time
[09:24] <omega_> with zero contention, spinlocks are very very fast
[09:24] <taaz> you should add some quick rdtsc timing between your fast locks and pthread locks
[09:24] <omega_> I was just starting to do that <g>
[09:26] <taaz> addthread ran 3213238 times, delthread ran 3187651 times
[09:26] <taaz> counter should be 25587, it is 25587
[09:26] <taaz> addthread ran 3767134 times, delthread ran 3201718 times
[09:26] <taaz> counter should be 565416, it is 565416
[09:26] <omega_> wheee ;-)
[09:26] <omega_> it worked the first time!
[09:26] <taaz> of course i'm still not sure of rdtsc works accross cpus
[09:26] <omega_> you wouldn't happen to have a quad handy, would you?
[09:26] <omega_> doesn't need to for my experiment
[09:27] <taaz> no quad here, sorry
[09:29] <omega_> ok, timing code running
[09:29] <omega_> and it fails utterly
[09:30] <omega_> ok, at least for unlock, it takes 61 cycles (???) for atomic, 141 for mutex
[09:31] <omega_> lemme calibrate rdtsc first <g>
[09:34] <omega_> hrm
[09:34] <omega_> ah, -O6 helps quite a bit <g>
[09:34] <BBB> still awake?
[09:35] <taaz> jeez.. i had over 600 messages since i went to bed last night.
[09:35] <omega_> ok, atomic gives ~103 cycles for lock+unlock
[09:35] <taaz> i think that's a record for me
[09:35] <omega_> mutexes take 518 cycles for lock+unlock
[09:36] <omega_> those numbers are high because they include a lot of switch time
[09:36] <taaz> is that just a raw lock+unlock?  i assume there's some overhead when it has to wait
[09:36] <omega_> yes, just lock+unlock
[09:36] <omega_> the timing version is in the mail
[09:37] <omega_> it's SMP-safe from the rdtscll point of view
[09:37] <omega_> for a 10sec run I get 106:655
[09:38] <omega_> you should get worse numbers because you'll have significant contention
[09:38] <omega_> I'd expect more on the order of the equiv of 160 cycles for atomic
[09:38] <omega_> er, 170
[09:38] matth_ (matth at qwest.dsplinux.net) left irc: Ping timeout for matth_[qwest.dsplinux.net]
[09:38] ajbusy (ajmitch at p15-max2.dun.ihug.co.nz) left irc: Ping timeout for ajbusy[p15-max2.dun.ihug.co.nz]
[09:38] wtay-sleeping (wim at cable-195-162-214-58.upc.chello.be) left irc: Ping timeout for wtay-sleeping[cable-195-162-214-58.upc.chello.be]
[09:38] wtay-sleeping (wim at cable-195-162-214-58.upc.chello.be) joined #gstreamer.
[09:38] matth_ (matth at qwest.dsplinux.net) joined #gstreamer.
[09:38] ajbusy (ajmitch at p15-max2.dun.ihug.co.nz) joined #gstreamer.
[09:39] <taaz> heh... the first test with -O6 gives this:
[09:39] <omega_> btw, use -O6 when you compile <g>
[09:39] <taaz> addthread ran 0 times, delthread ran 0 times
[09:39] <taaz> counter should be 0, it is 0
[09:39] <omega_> ergh
[09:39] <omega_> that's odd
[09:40] <omega_> um...
[09:40] <omega_> that means they both wedged immediately
[09:40] <omega_> that shouldn't happen
[09:40] <omega_> try with PTHR defined
[09:41] <omega_> oh wait, it was
[09:41] <omega_> uh, something's wrong here ;-(
[09:41] <taaz> ok, v3 no O:
[09:41] <taaz> addthread ran 2584921 times, delthread ran 2881331 times
[09:41] <taaz> counter should be -296410, it is -296410
[09:41] <taaz> base for rdtscl is 34 cycles
[09:41] <taaz> avg cycles per lock/unlock: 1808
[09:41] <omega_> ouch
[09:41] <taaz> -O6:
[09:41] <taaz> addthread ran 2848603 times, delthread ran 2805754 times
[09:41] <taaz> counter should be 42849, it is 42849
[09:41] <taaz> base for rdtscl is 32 cycles
[09:41] <taaz> avg cycles per lock/unlock: 1736
[09:41] <omega_> neat
[09:41] <omega_> now undef PTHR
[09:42] <taaz> is this good or bad? ;)
[09:42] <omega_> it's high, but not unexpected
[09:43] <taaz> addthread ran 498346 times, delthread ran 498309 times
[09:43] <taaz> counter should be 37, it is 37
[09:43] <taaz> base for rdtscl is 34 cycles
[09:43] <taaz> avg cycles per lock/unlock: 10037
[09:43] <omega_> eeeek
[09:43] <taaz> addthread ran 484196 times, delthread ran 416063 times
[09:43] <taaz> counter should be 67919, it is 67919
[09:43] <taaz> base for rdtscl is 32 cycles
[09:43] <taaz> avg cycles per lock/unlock: 11113
[09:43] <taaz> Segmentation fault
[09:43] <taaz> i like the last line ;)
[09:43] <omega_> that sucks
[09:43] <omega_> major contention
[09:43] <omega_> but that's what this program is all about
[09:44] <omega_> add #include <stdlib.h> up top
[09:44] <taaz> uh... that will do what?
[09:44] <omega_> and usleep(random()>>16); at the bottom of the while (spinning) block
[09:45] <omega_> hmm, that's not quite right
[09:45] <omega_> oh, yeah it is
[09:45] <omega_> ok, revise the code a little....
[09:46] <omega_> right after the first #endif
[09:46] <omega_> add
[09:46] <omega_>     rdtscll(stop);
[09:46] <omega_>     ts->cycles += stop - start;
[09:46] <omega_> and add rdtscll(start) befure the second #ifdef
[09:46] <omega_> wait
[09:46] <omega_> that won't change anything either ;-(
[09:46] <taaz> do you just want an account? ;)
[09:46] <omega_> yes, but not for this <g>
[09:47] <omega_> this is such a pathological test case anyway....
[09:47] <omega_> what matters is the no-contention case, because that's the most common
[09:47] <omega_> and with contention, it'll spin anyway
[09:47] <omega_> but I need a quad to test with
[09:48] <omega_> and did you figure out what that first run with all zeros was all about?
[09:49] <omega_> although... I know it's wrong because of another test
[09:50] <omega_> hrm, I need to rethink this ;-(
[09:50] <taaz> so at some point is gstreamer going to negotiate caps with a system object?  choose locking system and how to best use threads based on UP/SMP and so on?
[09:50] <omega_> possibly
[09:50] <omega_> though I kinda doubt it
[09:50] <taaz> well... i think it's needed
[09:51] <omega_> yeah, it should probably be aware, though I'm not sure how atm
[09:51] <taaz> if you have 4 cpus you may be able to optimze where decoding and rendering happen and so on
[09:51] <omega_> yup
[09:51] <omega_> pset
[09:52] <omega_> I can simulate 3-way contention by setting the lock to 2 (locked) up front
[09:53] <omega_> actually, can you do that too?  atomic_set(&alock,2); instead of 1
[09:53] <omega_> then run it, you'll get the wrong results
[09:54] <taaz> i'm too sleepy to figure out why v1 -O6 does nothing...
[09:54] <taaz> addthread ran 482360 times, delthread ran 441532 times
[09:55] <taaz> counter should be 40828, it is 40828
[09:55] <taaz> base for rdtscl is 32 cycles
[09:55] <taaz> avg cycles per lock/unlock: 10824
[09:55] <omega_> hmm, interesting
[09:55] <taaz> looks like correct results
[09:55] <omega_> I'll have to reason that case through
[09:56] <omega_> but I'm certain the true >2 contender case will fail
[09:56] <taaz> are you meaning to set spinning to 0 after the sleep?
[09:56] <omega_> yes
[09:57] <omega_> that shuts down the threads
[09:57] <omega_> the usleep is an attemp to make sure they shut down <g>
[09:57] <omega_> really should be a sleep(1) to make sure, though
[09:57] <omega_> esp in your case
[09:57] <taaz> all my results were off then...
[09:57] <omega_> ?
[09:58] <taaz> your code set it to 1 again
[09:58] <omega_> where?
[09:58] <omega_> oh, so it does
[09:58] <taaz>   sleep(10);
[09:58] <taaz>   spinning = 1;
[09:58] <taaz>   usleep(1000);
[09:58] Action: omega_ hides under a large BPB
[09:59] <omega_> shouldn't change the results, though.  if you got a match with that set wrong, you'll get a match with it set right
[09:59] <omega_> the timings might show differently though
[10:01] <omega_> netscape's css support SUCKS!
[10:01] <BBB> <hint>mozilla</hint>
[10:02] <omega_> yes, but somehow what I do stresses mozilla well past its breaking point for normal functioning
[10:02] <BBB> even 0.9?
[10:02] <BBB> 0.9 is quite good, actually
[10:02] <omega_> it doesn't *break*, it just doesn't work near as well as I need it to
[10:02] <omega_> yeah, 0.9, even latest galeon has some UI foibles
[10:03] <BBB> hmm....
[10:03] <BBB> opera?
[10:03] <BBB> :)
[10:03] <omega_> um
[10:04] <BBB> my stupid SDLsink still has green striped - that basically sucks quite hard....:|
[10:04] <omega_> cool
[10:04] <BBB> that is - it works, but not as I would like it to work
[10:05] <taaz> i crashed moz 0.9 twice.  and it seemed sluggish.  back to netescape for me too
[10:05] <BBB> I420 = 4:2:2 data ratio, right?
[10:05] <omega_> nope, 4:2:0
[10:05] <BBB> no V?
[10:05] <omega_> see www.webartz.com/fourcc/
[10:05] <omega_> 4:2:0 is odd
[10:05] <omega_> hrm, I forgot, webartz isn't too useful in describing what the YUV formats actually mean
[10:07] <BBB> lol
[10:07] <taaz> speaking for four... its four am here.  nap time. omega_: let me know if/when you need a SMP testing account 
[10:07] Nick change: taaz -> taazzzz
[10:07] <omega_> I'd like one, but not time critical (so to speak <g>)
[10:08] <BBB> it says that I420 = 2:1:1 vertical and 2:1:1 horizontal
[10:08] <BBB> that is 4:1:1
[10:08] <omega_> no, 4:1:1 is 1:1:1 vertical and 4:1:1 horizontal
[10:09] <BBB> 4:2:0 = 1:2:2 horizontal and vertical?
[10:09] Action: omega_ writes a wiki page
[10:09] <BBB> lol
[10:09] <omega_> no, 2:1:1 h+v
[10:09] Action: BBB is a shameless newbie
[10:11] <BBB> woohoo!
[10:11] <BBB> it works :)))))
[10:11] Action: BBB fixed the bug
[10:11] <BBB> should I send it? (although it is just video, no audio, it is still funny
[10:12] <omega_> send which?
[10:12] <BBB> the sdlsink.[ch]
[10:12] <BBB> (and probably the Makefile)
[10:13] <BBB> brb - gotta go to toilet
[10:15] <BBB> ok, back
[10:16] <omega_> http://codecs.org/cgi-bin/wiki/moin.cgi/YUV
[10:16] <omega_> each cell is a pixel, each letter is a sample
[10:17] <BBB> hmm.... Too bad that if I SIGINT it, it doesn't stop
[10:17] <BBB> that works for aasink....
[10:17] Action: BBB will copy-paste some more code
[10:17] <omega_> yeah, that's a major flaw I see in SDL, it traps Ctrl-C *and* SEGV
[10:17] <omega_> there's a way to turn that off, I'd strongly suggest doing that right after SDL_Init in your plugin
[10:17] <BBB> ok
[10:18] Action: BBB kicks SDL's sig* trapping
[10:18] <omega_> does that wiki page make sense?
[10:18] <BBB> yup, it does
[10:18] <omega_> good
[10:18] <BBB> better than the fourcc one :)
[10:19] <omega_> heh
[10:19] <omega_> I should put the fourcc codes in there just for fun
[10:19] <omega_> eventually
[10:19] Action: BBB wonders what the use of 4:1:1 is
[10:19] <BBB> 4:2:0 seems quite reasonable, we use 4:2:2 ourselves in mjpeg.....
[10:19] <BBB> (I think?)
[10:20] <omega_> 4:1:1 is used in DV, the main reason is that the eye is for some reason less sensitive to horizontal chroma than it is to vertical chroma
[10:20] <omega_> same reason 4:2:2 is used over 4:4:4
[10:21] <BBB> hmm...... makes sense, then
[10:22] <BBB> 		SDL_EventState(SDL_KEYDOWN, SDL_ENABLE);
[10:22] <BBB> 		SDL_EventState(SDL_MOUSEMOTION, SDL_IGNORE);
[10:22] <BBB> is that how to make it ignore the sigint?
[10:22] Action: BBB RTFMs
[10:22] <omega_> something else
[10:23] <BBB> hmm... no that is to grab keyboard commands (makes no sense) :)
[10:23] <omega_> SDL_INIT_NOPARACHUTE
[10:23] <omega_> pass that |'d with SDL_INIT_VIDEO to SDL_Init()
[10:24] <BBB> ok
[10:26] <BBB> hmm... it still doesn't catch it :|
[10:26] <omega_> catch which?
[10:27] <BBB> ctrl-C
[10:27] <omega_> hmm
[10:28] Action: BBB continues to RTFM
[10:31] walken (foobar at c1583255-a.smateo1.sfba.home.com) joined #gstreamer.
[10:31] <omega_> yo
[10:31] <omega_> you going to OLS?
[10:31] <walken> yup
[10:32] <walken> you ?
[10:32] <omega_> cool. maybe
[10:32] <walken> that'd be nice
[10:32] <walken> you *have* to go there :)
[10:32] Action: walken happy - my ac3 hacking is progressing
[10:33] <omega_> that's the impression I get <g>
[10:33] <walken> the impression - about ols or about ac3 :)
[10:34] <omega_> ols
[10:35] <walken> ok :)
[10:35] <walken> yeah it was quite nice last year
[10:35] <walken> damn vmware wouldnt send me there too
[10:35] <walken> had to take unpaid vacation
[10:35] <walken> grumpf
[10:36] <walken> not that I have money problems but I think its unfair :)
[10:36] <omega_> heh
[10:37] <BBB> I can catch the event now ;)
[10:37] <BBB> look:
[10:37] <BBB> RUNNING pipeline
[10:37] <BBB> Segmentation fault (core dumped)
[10:37] <BBB> [rbultje at tux tools]$ 
[10:37] <BBB> :P
[10:37] <omega_> neat <g>
[10:37] <BBB> Ctrl-C will make it SIGSEGV :P
[10:38] <omega_> uh
[10:38] <BBB> I just call 'sdlsink_close()' on the quit event....
[10:39] <BBB> I guess that's not good? :P
[10:39] <omega_> can you just have Ctrl-C do what it otherwise would?
[10:39] <BBB> I'm trying to do that :P
[10:39] <BBB> just didn't find how yet
[10:39] <omega_> hmm
[10:40] Action: BBB joins #sdl
[10:43] <BBB>   /* Event handling */
[10:43] <BBB>   while(SDL_PollEvent(&(sdlsink->event))) {
[10:43] <BBB>     switch(sdlsink->event.type) {
[10:43] <BBB>       case SDL_QUIT:
[10:43] <BBB>         gst_sdlsink_close (sdlsink);
[10:43] <BBB>         break;
[10:43] <BBB>     }
[10:43] <omega_> hmm, ick
[10:43] <BBB>   }
[10:43] <BBB> hmm.....
[10:43] <BBB> gst_sdlsink_close() should probably be something else - what function do I call in case of a signal?
[10:43] <omega_> not safe at all, and the correct solution is to somehow have SIGINT do whatever it normally does (i.e. not install a handler at all)
[10:44] <omega_> right now there is no good solution
[10:44] <BBB> ok
[10:50] <BBB> I cannot find it :?
[10:51] <BBB> the peepz @ #sdl tell me to simply make a signal handler for SIGINT *bleeh*
[10:51] <omega_> grrr
[10:52] <BBB> who said SDL was annoying? :P
[10:56] <BBB> what if I simply call SDL_quit on the SDL_QuitEvent?
[10:56] <omega_> the problem is that we want Ctrl-C to interrupt GStreamer, not SDL
[10:57] <BBB> what I could do is simply SDL_quit() and then re-emit SIGINT
[10:57] <omega_> maybe
[10:57] <BBB> would be evil but it might work
[10:58] Nick change: ajbusy -> ajmitch
[10:58] Action: omega_ goes to sleep now
[10:59] <ajmitch> hmm
[10:59] <omega_> l8r
[10:59] <ajmitch> bye ;)
[10:59] omega_ (omega at omegacs.net) left irc: [x]chat
[11:00] <BBB> signal(SIGINT, SIG_DFL);
[11:06] <BBB> woohoo!
[11:06] Action: BBB is less a newbie now
[11:07] Action: BBB is totally happy
[11:07] <BBB> who wants to see my cool SDLsink? :)
[11:09] <walken> ok going to sleep too
[11:09] <walken> cya
[11:09] walken (foobar at c1583255-a.smateo1.sfba.home.com) left irc: l8r
[11:10] <ajmitch> sleep? sounds good...
[11:10] <ajmitch> BBB: what ya making?
[11:10] <BBB> sdlsink
[11:10] <BBB> like aasink
[11:10] <BBB> to see video
[11:10] <ajmitch> hmm, sounds useful
[11:10] <BBB> but it's SDL-based now :)
[11:10] <ajmitch> probably wouldn't like my computer ;)
[11:10] <BBB> and it basically works
[11:10] <BBB> why not?
[11:11] <ajmitch> does it require dga in X? ;)
[11:11] <BBB> I don't think so
[11:11] Action: ajmitch has crappy video card drivers
[11:11] <BBB> it's not fullscreen (I could make it fullscreen, though)
[11:11] <ajmitch> i740 isn't supported too well
[11:11] steveb (stevebaker at 212186169160.chello.com) left irc: quit has steveb, yes
[11:11] Action: BBB recommends a matrox
[11:11] Action: ajmitch needs cash
[11:11] <BBB> lol
[11:12] Action: BBB too
[11:12] Action: BBB tries to make it work fullscreen now
[11:13] <ajmitch> heh ;)
[11:13] <BBB> aasink can understand command-line options (aasink driver=X)
[11:13] <BBB> do you know how?
[11:13] <ajmitch> nope
[11:14] Action: BBB is too lazy to find out
[11:14] <BBB> I'm gonna take a shower
[11:14] <ajmitch> hehe
[11:15] Action: BBB is away: shower
[11:15] Action: ajmitch should sleep
[11:15] <BBB> bbl
[11:15] <BBB> :)
[11:15] <BBB> I wont stop ya ;)
[11:15] Nick change: ajmitch -> ajzzzz
[11:32] beeble (pix at ipfpc32.e-technik.uni-stuttgart.de) joined #gstreamer.
[11:45] ajzzzz (ajmitch at p15-max2.dun.ihug.co.nz) left irc: Ping timeout for ajzzzz[p15-max2.dun.ihug.co.nz]
[11:49] ajzzzz (ajmitch at p15-max2.dun.ihug.co.nz) joined #gstreamer.
[11:54] beeble (pix at ipfpc32.e-technik.uni-stuttgart.de) left #gstreamer.
[12:11] holger (holger at 212.84.236.149) joined #gstreamer.
[12:12] holger (holger at 212.84.236.149) left irc: Client Exiting
[12:14] ajzzzz (ajmitch at p15-max2.dun.ihug.co.nz) left irc: Ping timeout for ajzzzz[p15-max2.dun.ihug.co.nz]
[12:14] ajzzzz (ajmitch at p10-max5.dun.ihug.co.nz) joined #gstreamer.
[12:31] hadess (hadess at pc2-guil2-0-cust121.gui.cable.ntl.com) joined #gstreamer.
[12:33] ajzzzz (ajmitch at p10-max5.dun.ihug.co.nz) left irc: Ping timeout for ajzzzz[p10-max5.dun.ihug.co.nz]
[12:38] ajzzzz (ajmitch at p10-max5.dun.ihug.co.nz) joined #gstreamer.
[13:07] BBB (BBB at ucu-105-116.ucu.uu.nl) left irc: Read error to BBB[ucu-105-116.ucu.uu.nl]: EOF from client
[14:20] sienap (synap at ipc379c2a3.dial.wxs.nl) joined #gstreamer.
[14:23] <sienap> hi
[14:51] Nick change: hadess -> hds-afk
[14:52] <sienap> hades!
[15:27] Nick change: hds-afk -> hadess
[15:27] <hadess> hey sienap
[16:03] BBB (BBB at ucu-105-116.ucu.uu.nl) joined #gstreamer.
[16:03] <hadess> yo BBB
[16:03] <BBB> hey guys, how do I submit pieces of code?
[16:03] <BBB> my sdlsink is fully-functional now :)
[16:03] Action: BBB thinks it's cute
[16:03] <hadess> make a patch
[16:04] <BBB> uhm.....
[16:04] <BBB> a patch......
[16:04] <BBB> how?
[16:04] <BBB> :)
[16:04] <BBB> cvs -Q diff -c?
[16:04] <hadess> diff -urN gstreamer.old gstreamer.new
[16:04] Action: BBB has no gstreamer_old :)
[16:05] <sienap> BBB with caps nego ?
[16:05] <hadess> use the cvs diff then, and clean up the patch
[16:05] <BBB> sienap: just based on aasink (ascii-art)
[16:05] <BBB> and it has caps nego, I think
[16:05] <sienap> i know aa
[16:05] <sienap> :)
[16:05] <sienap> bbb you saw the bb demo
[16:05] <sienap> from aa ?
[16:05] <sienap> :)
[16:06] <sienap> it is damn cute
[16:06] <sienap> :)
[16:06] <BBB> bb demo?
[16:06] <BBB> no
[16:06] <BBB> where?
[16:06] <sienap> at the aa hoempage :)
[16:07] Action: hadess sends BBB A clue
[16:08] Action: BBB needs many clues :)
[16:12] Action: BBB tries to hack Makefile/configure stuff
[16:50] matth_ (matth at qwest.dsplinux.net) left irc: Ping timeout for matth_[qwest.dsplinux.net]
[16:52] Action: BBB knows nothing about patches
[16:52] <BBB> can I send it to gst-devel at ..?
[16:52] <hadess> gzip'ed
[16:54] sienap (synap at ipc379c2a3.dial.wxs.nl) left irc: sienap has no reason
[16:54] <BBB> ok... can you test it for me?
[16:55] <BBB> just to be sure and such
[16:55] <hadess> i'm pretty busy right now
[16:55] <BBB> hmm...ok
[16:56] <hadess> send it to gst devel with a warning, they'll include it if it works correctly
[16:56] <BBB> good idea.....
[16:57] Action: BBB thinks it's all much easier if you have CVS access ourself
[16:58] <hadess> i don't have cvs write access
[17:14] <BBB> hmm.....
[17:14] Action: BBB neither
[18:13] Nick change: steveb-work -> steveb
[18:23] <BBB> okay, e-mail sent
[18:36] <hadess> BBB: ?
[18:36] <BBB> yes?
[18:36] <BBB> you should receive an e-mail on gstreamer-devel now
[18:36] <hadess> the HAVE_LIBSDL check in plugins/sdlsink/Makefile.am is useless, as you already skip it the plugins/sdlsink dir if it's not there
[18:37] <BBB> oh wait, that was an old one :)
[18:37] <hadess> that's the one you sent to gst-devel...
[18:37] thomas (thomas at 212-100-172-175.adsl.easynet.be) joined #gstreamer.
[18:37] <BBB> that HAVE_LIBDSL is indeed useless
[18:37] <hadess> then the bit about one .c file only
[18:37] <thomas> hi all
[18:37] <thomas> anyone know if gstreamer would compile on mac osx ?
[18:37] <BBB> thomas: it should
[18:38] <thomas> anyone ever tried ?
[18:38] <hadess> nope
[18:38] <thomas> ok, I'll try that
[18:38] <hadess> BBB: EXTRA_DIST = README is useless, there's no README file :)
[18:39] <hadess> BBB: the SDL cflags and libs you should get from the configure.in, see how i did in the plugin i sent yesterday
[18:39] Action: BBB didn't apply that patch yet
[18:39] Action: BBB <-- newbie :)
[18:39] <BBB> let's see
[18:40] <BBB> check_PROGRAMS = imagetest too I suppose?
[18:40] <hadess> no, not needed
[18:41] <hadess> open the patch i sent yesterday, and see what the change i made to configure.in look like
[18:41] <BBB> Hunk #2 FAILED at 113.
[18:41] <BBB> 1 out of 2 hunks FAILED -- saving rejects to file plugins/Makefile.am.rej
[18:41] <BBB> :)
[18:42] <hadess> heh, normal, we both patched plugins/Makefile.am
[18:43] <BBB> yup
[18:43] <BBB> true
[18:44] <BBB> ok, I patched it
[18:44] <BBB> (hand-opatch)
[18:44] <BBB> I'll look at it in a few minutes
[18:58] matth_ (matth at qwest.dsplinux.net) joined #gstreamer.
[19:09] <hadess> bbl
[19:13] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[19:13] <Uraeus> hi
[19:19] <BBB> hmm...
[19:19] <BBB> Makefile.am bug fixed
[19:19] Action: BBB apologizes for the crappy Makefile he sent
[19:21] Action: Uraeus puts BBB's SDLsink into the roadmap
[19:21] <BBB> hold it, wait, the Makefile is crappy - please take a new Makefile instead
[19:27] <hadess> heya Uraeus
[19:27] <Uraeus> hi hadess
[19:27] <Uraeus> hadess: I am putting your gnome-vfs plugin into the roadmap :)
[19:27] <hadess> Uraeus: wheee :)
[19:29] <BBB> okay, my second patch is on it's way
[19:29] <BBB> this should make the Makefile/configure stuff work better
[19:29] <BBB> *proud*
[19:29] Action: BBB is crappy in configure stuff
[19:32] Action: hadess curses nautilus
[19:32] <BBB> hmm?
[19:32] <BBB> nautilus 1.03 should be quite nice
[19:33] <hadess> try renaming a file, and use 'Del' to remove a character, the file is now in your trash
[19:33] <hadess> and that's 1.0.3
[19:35] <BBB> ?
[19:35] <BBB> :P
[19:35] <BBB> del doesnt work? :PPPPP
[19:37] <Uraeus> hadess: write a bug report :)
[19:38] <hadess> Uraeus: yeah, to bugzilla.eazel.com ? sigh
[19:38] <Uraeus> yup bugzilla.eazel.com
[19:38] <hadess> BBB: remove the CFLAGS line, it's not useful
[19:39] <BBB> it is
[19:39] <BBB> without it, it won't compile
[19:39] <BBB> really
[19:39] Action: BBB tried it
[19:39] <BBB> try it yourself
[19:39] <BBB> it failed to compile without the CFLAGS line here
[19:40] <hadess> dirty sdl-config
[19:40] <hadess> but in any case you end up having $(SDL_CFLAGS) twice
[19:40] <Uraeus> BBB: does SDL sink take both audio and video?
[19:41] <BBB> for now, only video
[19:41] <hadess> BBB: oooh, that's an SDL _video_ sink you made
[19:41] <BBB> I'm not sure how to get it to do audio
[19:41] <BBB> basically, audio is not the task of SDL, I suppose
[19:41] <Uraeus> BBB: there is some audio support in SDL isn't there?
[19:41] <BBB> I'm not exactly sure
[19:41] <hadess> there is
[19:41] <BBB> I only played with SDL-video until now
[19:41] Action: BBB will look into SDL-audio too
[19:42] <BBB> it would be nice if you (gstreamer) could be a framework which can cooperate 100% with SDL
[19:42] <BBB> I think gstreamer and SDL are really the leaders on both their areas
[19:42] <BBB> gstreamer+SDL = nice
[19:42] <Uraeus> BBB: ok, but if it is video only you probably should call it sdlvideo or something
[19:43] Action: BBB will integrate audio too :))
[19:43] <BBB> I'll work on audio
[19:43] <BBB> don't worry
[19:43] <BBB> I'm just a newbie :)
[19:45] <BBB> what's a good example of how to get audio in a sink?
[19:46] <BBB> or can I 'include' osssink/alsasink?
[19:46] <hadess> you can use code from them, but don't use include any .h file from them
[19:47] <BBB> hmm.....
[19:47] <BBB> I'll try to see if SDL can do sound itself
[19:47] <BBB> that would be pretty good
[19:47] <hadess> of course it does sound itself
[19:48] <hadess> it has ESD, aRTS, alsa and oss outputs, at least
[19:48] Action: BBB reads the SDL audio docs
[19:48] <BBB> this is why SDL is so nice
[19:48] <BBB> it's easy and complete
[19:50] Action: hadess has an interview tomorrow
[19:51] <BBB> wow, whatfor?
[19:51] <BBB> sounds like you're a big man :P
[19:51] <hadess> heh, no, for a job
[19:52] <BBB> lol
[19:52] <BBB> *totally disappointed*
[19:53] <BBB> hmm.....
[19:53] <BBB> difficult issue
[19:53] <BBB> for gstreamer-launch, I just give pipes for mpeg-decoding
[19:53] <BBB> not for mp2-decoding (music)
[19:54] <hadess> check mp3mad for example
[19:59] <BBB> the problem is more, how should the command on the commandlin look like? :)
[20:00] <BBB> ./gstreamer-launch disksrc location=/movies/korn.mpg ! mpeg1parse video_00! queue ! { mp1videoparse ! mpeg_play ! sdlsink }
[20:00] <BBB> that's how it looks like now
[20:00] <BBB> what should it look like if I also want to do mp2->wav and the wav in the sdlsink into SDL to the /dev/dsp?
[20:08] Nick change: wtay-sleeping -> wtay
[20:08] <wtay> yo
[20:08] <BBB> hey wtay
[20:08] <BBB> try out my cool sdlsink :)
[20:08] <hadess> yo wtay
[20:09] <wtay> BBB: you can't do audio *and* video with -launch yet... :(
[20:09] <wtay> yeah, I will :)
[20:09] <wtay> hi hadess
[20:09] <BBB> :((
[20:09] <wtay> BBB: ?
[20:09] <BBB> wtay: can I do it in any other way? I'd like to try out doing audio as well as video using SDL
[20:09] <BBB> would be a good exercise, I suppose
[20:10] <hadess> BBB: write a small C program...
[20:10] <wtay> BBB: you'll need to write a little app
[20:10] <BBB> hmm..... ok
[20:10] <wtay> mpeg2parse is a good start
[20:10] <wtay> mpeg2parse2 actually
[20:10] <hadess> wtay: btw, your "mad" fix made it work
[20:11] <wtay> hadess: cool, nice to hear that :)
[20:11] <wtay> BBB: is it a patch for HEAD or INCSCHED?
[20:11] <BBB> HEAD
[20:11] <BBB> I still didn't download INSCHED - no time yet
[20:11] <wtay> ok
[20:15] <hadess> wtay: tested my gnomevfs src ?
[20:15] <Uraeus> here is an updated version of my roadmap, feedback please: http://www.linuxrising.com/files/gstreroadmap2.html
[20:15] <wtay> hadess: where? in CVS?
[20:15] <hadess> wtay: nope, sent patch to ML
[20:16] <wtay> ok
[20:17] <hadess> Uraeus: one 'm' to incoming...
[20:18] <Uraeus> hadess: fixed
[20:18] <hadess> wtay: against HEAD, omega_ tested it, and it works, i was hoping he'd put it in CVS last night...
[20:18] <Uraeus> try and reload
[20:18] <hadess> Uraeus: sdlsink doesn't depend on gnome-vfs but on sdl =)
[20:19] <wtay> hadess: I'll put it in CVS
[20:19] <hadess> wtay: cool, thanks
[20:19] <BBB> gnomevfsink misses a 's'
[20:19] <BBB> and SDL indeed depends on SDL, has nothing to do with gnome-vfs ;)
[20:19] <hadess> fuck, that's an awful lot of plugins =)
[20:20] <BBB> of which half is broken ;))
[20:20] <hadess> yeah, well...
[20:20] Action: wtay seriously fucked up his tree with the patches...
[20:20] <BBB> wtay: lol
[20:21] <hadess> Uraeus: "Developmer comments" <- *cough*
[20:22] <hadess> wtay: you have a sec ?
[20:22] <wtay> yeah
[20:22] <hadess> about the gnome-vfs plugin, there is no way to use mmap with it
[20:23] <wtay> not surprisingly
[20:23] <hadess> so i wanted to make the plugin not use gnome-vfs if the file was local
[20:23] <wtay> oh , ok
[20:23] <hadess> and cut'n'paste code from fdsink (iirc)
[20:23] <wtay> ouch
[20:24] <hadess> yeah, that's what i thoyght as well 
[20:24] <wtay> maybe you should fire a signal and let the user app decide to add a disksrc instead
[20:24] <hadess> that's a good idea
[20:24] <wtay> not sure it'll work though...
[20:24] <hadess> yep
[20:24] Action: wtay should apply the automake patch now...
[20:25] <hadess> where's the patch ?
[20:25] <Uraeus> hadess: typo in developer now fixed
[20:25] <wtay> in the README in HEAD
[20:25] <hadess> ok
[20:38] Action: wtay compiles INCSCHED and HEAD at the same time
[20:42] <hadess> anybody knows how to have gdb launch a command as soon as it's started ?
[20:42] <wtay> dunno
[20:42] <hadess> i want it to do as if i typed "run"
[20:43] <wtay> hadess: it's football on TV <g> (liverpool I think)...
[20:43] <hadess> ooh, true
[20:44] <hadess> i was supposed to have a date tonight but she's in bed with a fever :/
[20:44] <wtay> hadess: you need to 'soigner' her..
[20:45] <hadess> <h>
[20:45] <hadess> <g> even :P
[20:45] <wtay> what's that in english?
[20:45] <wtay> confort?
[20:45] <hadess> heal
[20:45] <wtay> hmm, not really
[20:45] Action: hadess plays Sexual Healing
[20:46] <wtay> doh
[20:46] <hadess> comfort ?
[20:46] <wtay> liek the french 'se soigner'
[20:46] <hadess> take care
[20:46] <wtay> yeah, that's it
[20:46] <wtay> :)
[20:47] <hadess> heh, took me a while :P
[20:47] <wtay> hmm the build stuff is screwed up now :(
[20:47] <hadess> wtay: you applied the 2 patches ?
[20:47] <wtay> m -lSM -lICE -ldl -lXi -lXext -lX11 -lm -lz -lm -ldl -lSM -lICE -ldl -lXi -lXext -lX11 -lm -ldl -lXi -lXext -lX11 -lm -ldl -lXi -lXext -lX11 -lm -ldl -lz -lm -lz -lm -lm -lm -lm -lm -lm -ldb -ldl -ldl -ldl -ldl -ldl -ldl -ldl -lXi -lXext -lX11 -lm -lz -lz -lz -lz -lz -ldl -ldl -ldl -lXi -lXext -lX11 -lm -ldl -lXi -lXext -lX11 -lm -ldl -ldl -lXi -lXext -lX11 -lm -lz -lz -lz -lz -lz /usr/lib/libgnomeui.so -lm -lm -ldb -ldl -ldl -lXi -lXext -lX
[20:47] <wtay> 1 -lm -lSM -lICE -ldl -lXi -lXext -lX11 -lm -lz -lm 
[20:47] omega_breakfast (omega at omegacs.net) joined #gstreamer.
[20:47] <steveb> yo
[20:47] <hadess> oh my
[20:47] <wtay> probably, yeah ;)
[20:48] <wtay> yo
[20:48] <hadess> yo steveb
[20:48] <hadess> heya omega_breakfast
[20:50] <hadess> zaheer is probably in front of the telly, or in germany :)
[20:50] <hadess> and he's happy now
[20:50] Uraeus (cschalle at c224s9h5.upc.chello.no) left irc: syntax error - user imploded
[20:50] Nick change: hadess -> hds-busy
[20:50] <wtay> hadess: the patch didn't recurse into the subdirs... I screwed up by applying twice or more <g>
[20:50] <omega_breakfast> hds-busy: ?
[20:53] matth_ (matth at qwest.dsplinux.net) left irc: Ping timeout for matth_[qwest.dsplinux.net]
[20:56] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[21:01] Uraeus (cschalle at c224s9h5.upc.chello.no) left irc: syntax error - user imploded
[21:03] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[21:19] ChiefHighwater (paul at temple-baptist.com) joined #gstreamer.
[21:19] <wtay> whooohoooho!
[21:19] <wtay> ./gstreamer-launch gnomevfssrc location=http://gstreamer.net/media/south.mp3 ! mad ! osssink
[21:19] <wtay> coool
[21:20] <omega_breakfast> um
[21:20] Nick change: omega_breakfast -> omega_
[21:21] <omega_> osssink (as hacked by me) causes it to be really slow
[21:21] <wtay> "south of the border..."
[21:21] <omega_> esdsink causes it to run dblspeed
[21:21] <omega_> then it oom's my machine
[21:21] <ajzzzz> hehe
[21:21] <wtay> oops
[21:21] Nick change: ajzzzz -> aj_uni
[21:21] <wtay> esdsink doesn't adapt nicely
[21:21] <omega_> capsnego isn't happening as it should, it seems
[21:21] <omega_> even osssink isn't
[21:21] <aj_uni> oh dear, oom? ;)
[21:21] <wtay> nope, not implemented for esdsink
[21:21] <omega_> or maybe it's because I crippled the padtemplate <g>
[21:21] <wtay> that's the case in osssink :)
[21:22] <omega_> shouldn't it at least try to set the params to the new caps, at a minimum?
[21:22] <wtay> no
[21:22] <omega_> why not?
[21:22] <omega_> what if it works?
[21:22] <wtay> it's incompatible with the template
[21:22] <omega_> it is?
[21:22] Nick change: taazzzz -> taaz
[21:22] <omega_> then the template is invalid
[21:22] <wtay> yes, it's mono 22Khz
[21:23] <omega_> hmm, ok
[21:23] <wtay> yes, very invalid
[21:25] <wtay> "... that's where I fell in love... when the stars came out to play ..."
[21:26] <omega_> hmm
[21:26] <omega_> nice timing though <g>
[21:27] <wtay> howso?
[21:28] <omega_> I just finished playing it <g>
[21:28] <wtay> doh
[21:30] CHW (paul at temple-baptist.com) joined #gstreamer.
[21:30] ChiefHighwater (paul at temple-baptist.com) left irc: Read error to ChiefHighwater[temple-baptist.com]: Connection reset by peer
[21:30] <CHW> goodness...Win 98 disconnects all network devices when I plug my CF reader in my PCMCIA slot
[21:31] <omega_> neat
[21:31] Nick change: CHW -> ChiefHighwater
[21:31] <ChiefHighwater> stupid.
[21:31] Action: omega_ struggles to act surprised and shocked
[21:32] <ChiefHighwater> maybe it just ditches the dock for sec?
[21:32] <wtay> never say stupid to user friendly <g>
[21:33] <omega_> wtay: ?
[21:33] <wtay> hmm, nothing..
[21:33] Nick change: hds-busy -> hadess
[21:33] <hadess> half-time
[21:33] <omega_> ok, now to try again to get the autoplugger to do what it's supposed to
[21:33] <hadess> omega_: what was your oom problem ?
[21:33] <wtay> hadess: ./gstreamer-launch gnomevfssrc location=http://gstreamer.net/media/south.mp3 ! mad ! osssink
[21:33] <omega_> lack of EOS
[21:34] <omega_> I have old copy
[21:34] <Uraeus> omega_: http://www.linuxrising.com/files/gstreroadmap2.html
[21:34] <wtay> Uraeus: reading
[21:34] <hadess> omega_: should be fixed now, i'll had some more checks for things like 404, or not readable, etc.
[21:34] <Uraeus> wtay: small error in your name a couple of places :)
[21:35] <wtay> Uraeus: incsched1 is not a component
[21:35] <omega_> Incremental Scheduling is
[21:35] <Uraeus> wtay: what should I call the core parts then?
[21:35] <wtay> hmm, sorry I thought you meant a plugin...
[21:36] <omega_> add my name to autoplugging
[21:36] Action: wtay reads before complaining
[21:36] <wtay> yes
[21:36] <Uraeus> omega_: add or replace?
[21:37] <omega_> add Matt Howell to interlock and interruptibility
[21:37] <omega_> add
[21:37] <omega_> and incsched
[21:38] <wtay> remove mpeg1encoder, it's obsolete...
[21:38] <omega_> make it red, rather
[21:38] <omega_> add GObjet to core, red in cvs, 0.2.0, green in 0.3.0
[21:38] <omega_> er, GObject
[21:38] <omega_> change cross boundary to 'Cross-process'
[21:39] <wtay> sdlsink depends on libSDL, not gnome-vfs
[21:39] <omega_> wtay: should we have separate entries for autoplug and dynamic autoplug?
[21:39] <wtay> is there a future for the static autoplugger?
[21:40] <omega_> not in its current form, I don't think
[21:40] <omega_> I think the autoplug core should be reworked to fit into the dynamic autoplugger
[21:41] <wtay> I would then vote for throwing it away
[21:41] <omega_> ok
[21:41] <omega_> Uraeus: jpegenc/dec depends on libjpeg
[21:41] <omega_> vorbis depends on both libogg and libvorbis
[21:42] <omega_> dv1394src depends on libraw1394, not libdv
[21:42] <omega_> dvdec depends on libdv
[21:42] <hadess> soundbox is now rhythmbox
[21:42] <omega_> dvdec CVS should be yellow, it has 1394-specific memory management problems
[21:43] <omega_> change ladspa CVS to orange, comment is: Does everything but actual data-flow.
[21:43] <omega_> and full name is 'Zaheer Merali'
[21:45] <omega_> ok, url posted on gstreamer.net front page
[21:46] <hadess> back to the game
[21:46] Nick change: hadess -> hds-busy
[21:46] <omega_> Uraeus: you're making all these changes?
[21:46] <Uraeus> omega_: try and reload :)
[21:47] <omega_> ok, getting there
[21:47] <omega_> you missed some
[21:47] <omega_> GObject, cross-process, incremental scheduling
[21:47] <omega_> add matth to interlock
[21:47] <aj_uni> damn i know what i'll be doing this afternoon ;)
[21:47] <Uraeus> omega_: they are all in
[21:47] Action: aj_uni will actually go to uni now ;)
[21:47] <omega_> not correctly
[21:48] <omega_> syntax, not semantics/meaning
[21:48] <omega_> and GObject should be red,red,green, not yellow,red,red
[21:48] <omega_> cross-process interruptibility should be yellow,green,green
[21:51] <omega_> also, I'd make incsched, autoplug, interlock, cothread locking all yellow
[21:51] <omega_> for CVS
[21:51] <wtay> Uraeus: quite a few "Tayman" where it should be "Tayman4s1"
[21:52] <Uraeus> I know, fixing
[21:53] <Uraeus> wtay: it must have 'fallen' of during cut'n paste :)
[21:53] <BBB> gnomevfsink should be gnomevfsSink
[21:54] <omega_> and for deps... autoplug depends on incsched
[21:54] <omega_> interlock depends on cothread locking
[21:55] <wtay> GObject on glib2.0
[21:55] <omega_> another one would be 'Threadsafe element control', i.e. doing a set_arg from one process on an element in another
[21:55] <omega_> which depends on cothread locking
[21:56] <omega_> and that may be red,yellow,green
[21:57] <wtay> plugins assigned to "wtay" depends on "wtays mood"
[21:57] <omega_> hehehe
[21:57] <omega_> same with me
[21:57] <wtay> :)
[21:58] <Uraeus> hmm, you are undermining the roadmap with that attitude ;(
[21:58] <wtay> BBB: can SDLsink be tested with -launch?
[21:58] <Uraeus> ok, everything fixed
[21:58] <omega_> Uraeus: move threadsafe element control to after interruptibility
[21:59] <omega_> add matth to that, and interlock
[21:59] <omega_> and you're missing a lot of deps
[22:00] <Uraeus> like glibc? :)
[22:00] <omega_> no
[22:00] <omega_> threadsafety -> cothread locking
[22:01] Action: omega_ tries to figure out the others
[22:01] <omega_> interruptibility depends on interlock
[22:03] <omega_> comment on threadsafety is 'Ensure safety of cross-process element manipulation'
[22:03] <omega_> comment on cothread locking is 'Primitives to ensure single-threading of cothread-specific code'
[22:04] <omega_> oops, I missed a letter in your name ;-(
[22:04] <omega_> you've still got glib 2.0 as a dep for threadsafety
[22:05] <omega_> cothread locking does not depend on interlock
[22:05] <wtay> glib2.0 a dep for threadsafety?
[22:05] <omega_> no
[22:06] <omega_> not that threadsafety
[22:06] <wtay> oh, that's what's wrong..
[22:07] <omega_> Uraeus: you're skipping comments above.  threadsafe element control depends on cothread locking, *not* glib 2.0
[22:07] Action: wtay whishes 0.3.0 was out with all things fixed :)
[22:09] <Uraeus> wtay: well, you have the power to make that happen :)
[22:09] <Uraeus> omega_: fixed
[22:09] <omega_> except it's a capital C Cothread <g>
[22:10] <wtay> loaded 72 plugins with 96 elements and 10 types
[22:10] <omega_> wtay: what world are you in?
[22:10] <omega_> INFO ( 2303:-1)gst_plugin_load_all:239: loaded 67 plugins with 135 elements and 10 types
[22:10] <wtay> uh
[22:10] <wtay> LADSPA
[22:10] <omega_> yup <g>
[22:10] <wtay> ?
[22:11] <wtay> grr
[22:11] Action: wtay apt-gets all he can get
[22:14] <omega_> also, gnomevfs is now no longer an incoming plugin
[22:15] <wtay> hmm, sdlsink doesn't use Xv
[22:16] <omega_> rather gnomevfssrc is no longer incoming, but in CVS
[22:16] <omega_> wtay: huh?
[22:16] <wtay> nope, plain X and SW color conversion
[22:16] <omega_> neat
[22:17] <omega_> he might be missing a flag
[22:17] <wtay> weird
[22:17] <omega_> Uraeus: you have mpeg2parse twice
[22:17] <omega_> second one is unclaimed, remove that one
[22:19] <BBB> wtay: yes
[22:19] <BBB> ./gstreamer-launch disksrc location=/movies/korn.mpg ! mpeg1parse video_00! queue ! { mp1videoparse ! mpeg_play ! sdlsink }
[22:19] <BBB> that's what I used to test it
[22:19] <wtay> BBB: ok, got that working too
[22:19] <taaz> oh my... since when did I become OMS maintainer? ;)
[22:19] <omega_> guess <g>
[22:19] <BBB> now I gotta think of how to get the sound in there working too ;-)
[22:19] <Uraeus> going out for a minute, brb
[22:20] <BBB> omega_: how about this idea? Everyone is gonna use gstreamer as backend multimedia library - so gstreamer is just supposed to deliver the decoders, encoders and datastream. Why don't you drop your own player sand such and just use SDL for that?
[22:20] <BBB> That would be a perfect example of re-use of code
[22:21] <omega_> huh?
[22:21] <BBB> and I personally think that although SDL is in some ways crappy, it is in generally a very-well programmed library
[22:21] <omega_> because SDL's APU is not suitable
[22:21] <omega_> er, API
[22:21] <taaz> does SDL support 24 channel audio cards?
[22:21] <omega_> hardly
[22:21] <omega_> it doesn't support 2 separate cards, afaict
[22:21] <BBB> but isn't gstreamer supposed to not care about that? :)
[22:21] <omega_> um?
[22:21] <BBB> I mean
[22:22] <BBB> gstreamer is just the library
[22:22] <BBB> the OMS kinda applications can care about that more advanced stuff
[22:22] <BBB> right?
[22:22] <BBB> I mean, gstreamer doesn't need to be the application itself
[22:22] <omega_> yes, but what does SDL have to do with that?
[22:22] <BBB> well....... yes
[22:22] <wtay> BBB: what do you want to see ditched then?
[22:22] <BBB> you can then use small SDL-based plugins to test the library
[22:22] <BBB> for example, drop gstmediaplay
[22:22] <omega_> 'SDL-based plugins'?
[22:22] <wtay> define SDL based?
[22:23] <BBB> oops
[22:23] <BBB> not SDL-based plugins
[22:23] <BBB> SDL-based frontend to test the library
[22:23] <BBB> :P
[22:23] <omega_> SDL has nothing to do with multimedia in the sense of gstreamer
[22:23] <omega_> smpeg is as close as it gets, and that's a very very MPEG-1-centric API
[22:23] <wtay> why? I wanna play using aalib :)
[22:23] <BBB> lol
[22:23] Action: BBB cries - I hoped you would use sdllib :P
[22:23] <omega_> I'm very confused by what you think SDL is relevant to
[22:23] <taaz> SDL supports aalib output
[22:24] <BBB> it does?
[22:24] <BBB> omg
[22:24] <BBB> omega_: look, gstreamer is supposed to be a streaming framework, right?
[22:24] <omega_> yes
[22:24] <wtay> SDLsink is a nice plugin to show an image, that's all, nothing more
[22:24] <BBB> but the fact is that gstreamer basically also has a player etc - all not based on other people's libs
[22:24] <BBB> this is not about sdlsink :)
[22:25] <taaz> BBB: set SDL_VIDEODRIVER=aalib
[22:25] <wtay> can you easily embed SDL output into your own app?
[22:25] <omega_> SDL is of no use to us except as an output method
[22:25] <BBB> taaz: ohyeah, that's true :)
[22:25] <omega_> SDL doesn't do embedding
[22:25] <BBB> wtay: yes, I already did :)
[22:25] <BBB> I embedded SDL in my Gtk app
[22:25] <omega_> that's a hack
[22:25] <wtay> can you open it N times?
[22:25] <BBB> yes, but it works :)
[22:26] <omega_> as are all other forms of running SDL in something else
[22:26] <omega_> no, only one instance of SDL
[22:26] <BBB> I think only one instance, yes
[22:26] <omega_> it uses a global state
[22:26] <omega_> == bad
[22:26] <wtay> I want a 8 tv screens!
[22:26] <omega_> !SDL
[22:26] <BBB> lol
[22:27] <omega_> let's put it this way: outside of sdlvideosink and sdlaudiosink, no SDL code is of use to us
[22:27] <BBB> true
[22:27] <omega_> that leaves the API, which is very 2D graphics centric
[22:27] <BBB> but that's exactly what I'd suggest
[22:27] <omega_> so what else is there that's relevant?
[22:27] <BBB> drop gstmediaplay in favour of a new SDL-based player
[22:27] <omega_> there is no such thing as an SDL-based player
[22:27] <wtay> BBB: it's just a one line change in gstplay?
[22:27] <omega_> SDL does not have an API that has anything to do with playing media
[22:28] <omega_> wtay: btw, if you haven't committed it already, change sdlsink to sdlvideosink, so BBB can add an audiosink
[22:28] <omega_> for the element, at least
[22:28] <BBB> oh that's true :)
[22:28] <wtay> ok, 
[22:28] <BBB> the dir can be called sdlsink
[22:28] Action: wtay abouts the CVS commit
[22:28] <omega_> already is
[22:28] <wtay> s/abouts/aborts
[22:29] <BBB> I'll make a sdlsinkvideo.c and sdlaudiosink.c
[22:29] <omega_> videosink
[22:29] <BBB> oops
[22:29] <BBB> sorry
[22:29] <BBB> I type too fast
[22:29] <omega_> hehehe
[22:29] <wtay> sdlsink dir is fine IMO
[22:29] steveb (steveb at node1ee05.a2000.nl) left irc: [x]chat
[22:29] <omega_> yup
[22:29] <wtay> SDL would be better
[22:29] <omega_> we may change it later, during restructuring
[22:29] <wtay> shall I change it too?
[22:30] <omega_> we need a plan for that
[22:30] <wtay> ok, I'll leave it for now then
[22:31] <omega_> wtay: we need to have debug routines that print out caps and such, like -inspect
[22:31] <taaz> so with the new plugins today what is the automake process size up to? ;)
[22:31] <omega_> but if we make it a core function, it has to be either DEBUG or g_print
[22:31] <omega_> how should we do that?
[22:31] Action: BBB stupidly asks for the gstreamer-launch command to play the sound of a mpeg movie
[22:32] <BBB> ./gstreamer-launch disksrc location=/mp3/korn.mp3 ! autoplugger ! osssink
[22:32] <BBB> that?
[22:32] <omega_> ./gstreamer-launch disksrc location=video.mpg ! mpegNparse audio_00! mad ! osssink
[22:32] <wtay> omega_: like GST_DEBUG_CAPS?
[22:32] <omega_> yeah
[22:32] <wtay> print out a tree or something?
[22:32] <omega_> yup, like -inspect
[22:33] <omega_> but then we should share the code, but -inspect doesn't use DEBUG
[22:34] <BBB> whoops - I forgot, I don't have mad
[22:34] <omega_> mp3parse ! mpg123
[22:36] <BBB> dammit - it all segfaults
[22:36] <omega_> neat
[22:36] <BBB> ./gstreamer-launch disksrc location=/movies/korn.mpg ! mpeg2parse audio_00 ! mpg123 ! osssink
[22:36] <BBB> ./gstreamer-launch disksrc location=/movies/korn.mpg ! mp3parse ! mpg123 ! osssink
[22:36] <BBB> ?
[22:36] <omega_> it's a system stream?
[22:36] <omega_> you must parse out the audio
[22:36] <omega_> disksrc ! mpegNparse video_00! mp3par...
[22:36] <omega_> where N is 1 or 2
[22:37] <omega_> er, audio_00
[22:37] <taaz> audio_00 ! -> audio_00! ... crazy spacing syntax ;)
[22:37] <BBB> that's what the first one does
[22:37] <omega_> you're missing mp3parse in the first one
[22:37] <omega_> and if it's mpeg2, it's probably ac3, not mp3
[22:38] <BBB> ac3parse?
[22:38] <BBB> [rbultje at tux tools]$ ./gstreamer-launch disksrc location=/movies/korn.mpg ! mpeg2parse audio_00! ac3parse ! mpg123 ! osssink
[22:38] <BBB> INFO(20870:-1): Initializing GStreamer Core Library
[22:38] <BBB> INFO(20870:-1): CPU features: (0c040843) MMX 
[22:38] <BBB> RUNNING pipeline
[22:38] <BBB> ***** GStreamer ERROR ***** in file gstscheduler.c at gst_bin_schedule_func:431
[22:38] <BBB> Element: /launch/ac3parse0.sink
[22:38] <BBB> Error: peer is null!
[22:38] <BBB> ***** attempting to stack trace.... *****
[22:38] Action: Uraeus is back
[22:38] <BBB> [rbultje at tux tools]$ 
[22:38] <BBB> lol
[22:38] <omega_> is it mpeg1 or mpeg2?
[22:38] <BBB> I think mpeg2
[22:39] <omega_> disksrc location=.... ! mpeg2parse private_stream_1.0! ac3dec ! osssink
[22:39] <BBB> ac3dec keeps complaining about peer=NULL
[22:40] <omega_> then it's not an ac3 stream
[22:40] <BBB> what's the decoder for mp1 then? :P
[22:40] <omega_> ok.  can plaympeg play it?
[22:40] <BBB> plaympeg korn.mpg? yes
[22:40] <omega_> then it's mpeg1
[22:41] <omega_> mpeg1parse audio_00! mp3parse ! mpg123
[22:41] <BBB> ./gstreamer-launch disksrc location=/movies/korn.mpg ! mpeg1parse audio_00! mp3parse ! mpg123 ! osssink
[22:41] <BBB> ***** GStreamer ERROR ***** in file gstscheduler.c at gst_bin_schedule_func:431
[22:41] <BBB> Element: /launch/mp3parse0.sink
[22:41] <BBB> Error: peer is null!
[22:41] <BBB> ***** attempting to stack trace.... *****
[22:42] Action: BBB is in huge support for the autoplugger
[22:42] <Uraeus> taaz: shouldn't your name be on the OMS field?
[22:42] <omega_> it is
[22:43] <Uraeus> omega_: I just saw taaz's comment about being OMS maintainer
[22:44] <omega_> it's there, it's correct
[22:47] Action: BBB thinks this whole gstreamer-launch stuff is just too complicated
[22:48] <taaz> Uraeus: i was just being silly...  trying to avoid responsibility or something ;)
[22:48] <BBB> omega_: what about making a wiki page on what all plugins do, input, output and how to make gstreamer-launch commands for that?
[22:48] <BBB> would be a huge help for me
[22:48] <BBB> I cannot even play a simple mpeg movie sound :|
[22:48] <taaz> Uraeus: for full overkill you could add a section on packaging... ie, debian stuff is me
[22:49] <taaz> BBB: http://gstreamer.net/wiki/GstPlugins
[22:49] <omega_> BBB: go for it <g>
[22:49] <BBB> woohoo!
[22:49] <BBB> I found it :)
[22:50] <BBB>  ./gstreamer-launch disksrc location=/movies/korn.mpg ! mpeg1parse audio_00! queue ! { mp3parse ! mpg123 ! osssink }
[22:50] <BBB> lol
[22:50] <BBB> :)))))))
[22:50] Action: BBB doesn't know what it does, but at least it works :)
[22:50] <wtay> BBB: actually, it's easy...
[22:50] <taaz> space/time tradoff... you could move the queue to right before the osssink
[22:51] <omega_> Alan weighs in, sorta <g>
[22:51] <wtay> yeah, sorta :)
[22:51] <wtay> just to tell us that shit happens <g>
[22:52] <BBB> haha, Alan speaks doom on you guys :P
[22:52] Action: BBB kinda likes Alan's way of thinking
[22:52] <wtay> xmms plugins are so lame... now the cdread plugin wants to access the playlist_mutex
[22:53] <omega_> mu
[22:54] <taaz> omega_, wtay: you two should point out you've been working on this for quite a long time before funding came along
[22:54] <omega_> that's what I'm doing <g>
[22:55] <omega_> sent
[22:55] <wtay> I'm not leaving, ever..
[22:56] <BBB> haha, if this project turns out to become a mess later on (which I doubt, of course, but just in case), it wouldn't be of much use anymore, would it?
[22:56] Action: BBB hopes gstreamer will lay out an open road for Linux Video Studio to conquer the world >:)
[22:57] <wtay> it will, trust me.. :)
[22:57] <BBB> lol
[22:57] <BBB> someday
[22:58] <omega_> my manager: "Is this 'the Alan Cox?" <g>
[22:59] <BBB> hmm........ do you guys think writing a sdlaudiosink is actually useful?
[22:59] <omega_> maybe....
[22:59] <BBB> I mean, osssink basically does the job for m
[22:59] <BBB> e
[22:59] <omega_> if sdlvideosink can be made to somehow cooperate with existing sdl apps, then having an audiosink do the same could be very valuable
[22:59] Action: BBB will start writing it tomorrow, based on.....osssink! :P
[22:59] <omega_> heheh
[22:59] <justin_> whats this about alan cox?
[22:59] <wtay> what would be nice is audio/video sink 
[22:59] <BBB> justin_: he wrote something to gst-devel@
[22:59] <omega_> justin_: see gst-devel
[22:59] <wtay> s/sink/sync
[22:59] <justin_> ahh
[22:59] <taaz> yes it's valuable... SDL does help with portability to win32 platform
[23:00] <BBB> hmm..... true
[23:00] <omega_> at the application layer, yeah
[23:00] <justin_> sink=sync? now you tell me
[23:00] <BBB> I'll work on it tomorrow
[23:00] <omega_> though I used SDL on win32.... <shudder>
[23:00] <BBB> I'd need to look deeped into gst to see how to get something like sync
[23:01] <BBB> I mean, basically I cannot do much more than seperate audio and video threads yet
[23:01] <BBB> to get sync I'd need to look more into it
[23:01] <justin_> sync is easy:) just have the audio and the video play at the same time;)
[23:01] <omega_> yeah, the syncing stuff works, but needs to be extended
[23:01] <BBB> doh!!!
[23:01] Action: justin_ goes back to learning opengl
[23:01] <omega_> justin_: hehehe
[23:01] Action: BBB looks @ justin_ - you go and do it then thankyou! :P
[23:02] <justin_> i could probably make it in sync, it may be at 1fps, but it would be in sync
[23:02] <BBB> lol
[23:02] matth_ (matth at 63.228.189.94) joined #gstreamer.
[23:02] <omega_> yo
[23:02] <BBB> that's not the point :P
[23:02] <wtay> hi matth_
[23:02] <BBB> it needs to be at full-speed and in sync :P
[23:03] <BBB> whoops
[23:03] <BBB> I crashed galeon with a segfault :}
[23:03] <omega_> don't do that!
[23:03] Action: BBB is up to crashing gstmediaplay with a segfault too now
[23:04] <BBB> although that's not too difficult >:)
[23:06] <omega_> matth_: how's the move prep going?
[23:07] Action: BBB still segfaulted his mpeg-music-command
[23:07] <wtay> omega_: any news from krasics dv->mpeg1 transcoding?
[23:07] <BBB> if it ends:
[23:07] <BBB> Gtk-WARNING **: invalid cast from (NULL) pointer to `GstPad'
[23:07] <BBB> Segmentation fault
[23:07] <BBB> :)
[23:07] <omega_> wtay: nope
[23:07] Nick change: aj_uni -> ajmitch
[23:07] <wtay> yeah, EOS is hosed
[23:07] <ajmitch> hi all
[23:08] <ChiefHighwater> ello
[23:08] <BBB> wtay: oh ok
[23:08] <omega_> wtay: I'm getting problems with the autoplugger
[23:08] <wtay> omega_: oh
[23:09] <omega_> scanario: mad ! autoplugger ! esdsink(crippled to mono padtemplate)
[23:09] Nick change: hds-busy -> hadess
[23:09] <hadess> european cup to england biatches !
[23:09] <wtay> omega_: no elements found to connect this?
[23:09] <BBB> poor Alaves
[23:09] <BBB> :(
[23:09] <omega_> autoplugger is set to negoproxy, nego fails
[23:10] <omega_> call autoplugger, it finds stereo2mono (with some fixups on my copy)
[23:10] <Uraeus> ok, roadmap updated with more apps and a table with platform support
[23:10] Action: ajmitch hits reload
[23:11] <wtay> lavtools?
[23:11] <omega_> arm is substantially untested
[23:11] <ajmitch> Uraeus: one day, you may be able to add an app of mine to the list ;)
[23:11] <omega_> 'mozillasink' is odd, it's really an app that's a component on top of libgstplay
[23:11] <omega_> and it's gnomevfssink, not gnomevfSink
[23:12] <wtay> Uraeus: yup, bonobo-media is like mozilla/nautilus plugin
[23:12] <omega_> and it doesn't exist yet, so it can't work well <g>
[23:12] <BBB> omega_: I said that 10 times already
[23:12] <omega_> but there is no such thing as mozillasink
[23:12] <omega_> BBB: I know ;-(
[23:12] <Uraeus> ajmitch: let me know when :)
[23:12] <omega_> Uraeus isn't catching everything
[23:12] <omega_> more likely he's flooded
[23:12] <omega_> <g>
[23:13] <BBB> lol
[23:13] <ajmitch> Uraeus: when == when i get time to do something on it ;)
[23:13] <wtay> ajmitch: so what are you doing now, appart from chatting? <g>
[23:14] <omega_> so, gst_static_autoplug_to_caps says 'attempting to create ghostpad for stereo2mono', and goes into check_compatibilty, and segfaults.....
[23:14] Action: wtay pokes into ajmitch's mind
[23:14] <ajmitch> wtay: hey, i just came back from a very boring maths lecture, give me a break ;)
[23:14] <omega_> 2+2 = ?
[23:14] <wtay> ajmitch: oh yes, university
[23:14] Action: taaz just now realizes he needs sdl and gnomevfs debian plugin packages now... argh
[23:14] <omega_> hehehehe
[23:14] <wtay> omega_: whoa
[23:15] <ajmitch> seems like there's quite a bit of work on gstreamer at the moment ;)
[23:15] <wtay> taaz: I was going to mention is...
[23:15] <BBB> and that's good, ajmitch :)
[23:15] Action: BBB loves this new mp3 player
[23:16] <BBB> ./gstreamer-launch disksrc location=/movies/korn.mpg ! mpeg1parse audio_00! queue ! { mp3parse ! mpg123 ! osssink }
[23:16] <BBB> mp2, even
[23:16] <BBB> or mp1
[23:16] <BBB> whatever ;)
[23:16] <wtay> BBB: a bash script and you have a playlist for free :)
[23:16] <omega_> doh
[23:16] <BBB> lol
[23:17] <omega_> #0  gst_static_autoplug_to_caps (autoplug=0x8172310, srccaps=0x807e928, sinkcaps=0x807e298, args=0xbfe3faac)
[23:17] <wtay> omega_: trying to recreate the error
[23:17] <omega_>     at gststaticautoplug.c:448
[23:17] <omega_> #1  0x40020185 in gst_autoplug_to_caps (autoplug=0x8172310, srccaps=0x807e928, sinkcaps=0x807e298)
[23:17] <omega_>     at gstautoplug.c:137
[23:17] <omega_> you'll need my autoplugger and some changes to esdsink
[23:17] <wtay> can you commit it?
[23:17] <Uraeus> is my plattform listing correct?
[23:17] <ajmitch> does everything new get committed to INCSCHED1 now? ;)
[23:18] <omega_> done
[23:18] <omega_> ajmitch: atm, yes ;-(
[23:18] shitowax (yann at Mix-Annecy-101-1-112.abo.wanadoo.fr) joined #gstreamer.
[23:18] <omega_> that's why were' frantically trying to get incsched to work, so we can merge down to HEAD ASAP
[23:18] <ajmitch> omega_: so should i drop HEAD and grab that branch?
[23:18] <shitowax> yo
[23:18] <ajmitch> or just hack on plugins in HEAD?
[23:18] <omega_> not yet.  I (once again) hope it'll be merged soon
[23:18] <omega_> hack on plugins in HEAD
[23:19] <ajmitch> what's holding up incsched? ;)
[23:19] <wtay> yo shitowax
[23:19] <taaz> 'soon' == -3 weeks or so... ;)
[23:19] Action: taaz ducks
[23:19] <hadess> shitowax: sympa comme nick...
[23:19] <omega_> wtay: update gstautoplugcache too
[23:19] <wtay> ok
[23:20] <shitowax> I've just a remark about the roadmap ...
[23:20] <omega_> then hack out 16-bit stereo caps from esdsink
[23:20] <omega_> oh, and you need stereo2mono updates too, committing
[23:20] <Uraeus> shitowax: yes
[23:20] Action: ajmitch looks for a plugin to rip caps nego code example from...
[23:20] <omega_> wtay: suggestions for ajmitch?
[23:21] <wtay> ajmitch: audio/video?
[23:21] <shitowax> the 3ivx codec is not compatible with divx at all ...
[23:21] <omega_> ok, stereo2mono committed
[23:21] <ajmitch> wtay: i guess, i'm looking at the volume plugin first ;)
[23:21] <omega_> isn't it mpeg4?
[23:22] <wtay> ajmitch: osssink has a bit of capsnego for audio
[23:22] <shitowax> yes it is, the original divx isn't
[23:22] <omega_> right
[23:22] Action: ajmitch tests autogen.sh to see how much RAM it takes up...
[23:22] <taaz> Uraeus: I don't actually have OMS+GST code at all right now... maybe that box should be blank or red
[23:22] <omega_> ajmitch: patch your automake!
[23:22] <wtay> shitowax: how much of mpeg4 is implemented?
[23:22] <Uraeus> taaz: the first box is, you will not have any code ready for 0.2.0?
[23:23] <shitowax> simple profile mpeg4v3
[23:23] <taaz> Uraeus: I have no idea... this isn't a "deadline" sort of project for me ;)
[23:23] <Uraeus> shitowax: updated
[23:24] <taaz> Uraeus: who knows... maybe i'll have it working tomorrow...
[23:24] <shitowax> cool ;)
[23:24] <shitowax> another thing, will I have access to you CVS or maybe you prefer a patch ?
[23:24] <omega_> taaz: now that OMS seems all gung-ho about it, you sure do have a deadline: ASAP <g>
[23:25] <omega_> probably just a set of files for the quicktime plugin at first
[23:25] dobez (dobey at dreadnought.ximian.com) joined #gstreamer.
[23:26] <taaz> omega_: that's what i wanted to avoid ;)
[23:26] <dobez> *bbep*
[23:26] <justin_> quicktime?
[23:26] <Uraeus> hi dobez?
[23:26] <dobez> hey
[23:26] <shitowax> ok ... I will give you an URL where to download a first beta version probably at the end of the week
[23:26] Nick change: dobez -> dobey
[23:26] <omega_> cool
[23:27] <omega_> wtay: reproducible?
[23:28] <BBB> ximian.com - wow guys, you got lots of interested people around
[23:28] <BBB> :)
[23:28] <dobey> hrmm
[23:28] <dobey> eh
[23:29] <BBB> ?
[23:29] <dobey> "interested"
[23:29] <wtay> sorry, was away
[23:29] <BBB> dobey: well..... why are you here if you're not? :P
[23:29] <dobey> BBB: because i want it to work on my laptop ;-)
[23:30] matth_ (matth at 63.228.189.94) left irc: Ping timeout for matth_[63.228.189.94]
[23:30] Action: BBB doesn't see that connection
[23:30] <hadess> dobey: mp3 playing works through mad now
[23:30] <dobey> hadess: damn, i need to update more often then
[23:30] <dobey> heh
[23:30] <dobey> BBB: powerbook
[23:30] <ajmitch> hmm, unpatched automake does tend to use a little bit of RAM, no? ;)
[23:30] <dobey> BBB: i work at ximian, yes
[23:30] <wtay> omega_: ok, got the same segfault
[23:30] <omega_> ok
[23:30] Action: ajmitch watches the swap space get chewed up
[23:30] <dobey> BBB: i'd also like the bonobo control for it to not suck as well
[23:31] <hadess> omega_: btw, you have an idea for byapssing gnome-vfs if the uri is a local file ?
[23:32] <dobey> BBB: do you know who i am? (out of curiosity)
[23:32] <omega_> yeah, use disksrc <g>
[23:32] <Uraeus> hadess: ask Darin
[23:32] <hadess> omega_: you think this should be up to the app ?
[23:32] <ajmitch> hehe, automake is over 300MB now...
[23:32] <omega_> hadess: not necessarily.  we may want to have a mechanism for dealing with different URIs at the gstreamer level
[23:32] <omega_> maybe a smart source meta-element
[23:33] <ajmitch> stupid kernel still has 192MB used for caching
[23:33] <ajmitch> omega_: maybe i should have patched? ;)
[23:33] <omega_> ya think?
[23:33] <hadess> ajmitch: sounds like a good idea
[23:33] <wtay> ajmitch: I patched today :)
[23:34] <ajmitch> omega_: yeah, it causes xmms to skip :(
[23:34] <dobey> dudes, i'm gonna go home i guess
[23:34] <ajmitch> cya dobey
[23:34] <hadess> ajmitch: xmms skips when i change workspaces...
[23:34] dobey (dobey at dreadnought.ximian.com) left irc: later
[23:35] <omega_> hadess: xmms native or plugged into gstreamer?
[23:35] <BBB> oops
[23:35] <BBB> I missed something
[23:35] <hadess> omega_: both, but it's because of the drivers/kernel
[23:35] <BBB> who is dobey then?
[23:35] <BBB> :)
[23:35] <BBB> is he important/famous?
[23:35] <hadess> BBB: a ximian employee, a friend of mine
[23:36] <BBB> ok, cool...:)
[23:36] <BBB> <dobey> BBB: do you know who i am? (out of curiosity)
[23:36] <BBB> ...?
[23:36] <Uraeus> he, I thought he was the ORBit-cpp developer :)
[23:36] <omega_> wtay: any idea?
[23:36] <wtay> omega_: not yet
[23:37] <hadess> Uraeus: he's not that mad :P
[23:37] <hadess> wtay: you commited gnome vfs ?
[23:37] <omega_> Uraeus: alpha asm is made, and tested, but not recently.  tested by me, maintained by me
[23:37] <omega_> to HEAD, yeah
[23:37] <Uraeus> hadess: I know, he told me 
[23:38] <hadess> omega_: ok, thanks
[23:38] <wtay> omega_: looks like stereo2mone caps are incompatible with esdsink caps somehow...
[23:38] Action: hadess goes to make the sink
[23:38] <omega_> wtay: that's not the problem, because it's not nego'ing them yet
[23:38] <omega_> er, maybe it is
[23:38] <omega_> but it shouldn't segv, regardless
[23:38] <wtay> omega_: loook at the _init function of stereo2mono and find waldo
[23:39] <omega_> ?
[23:40] <wtay> stereo->srcpad is assigned twice
[23:40] <omega_> oh, that init
[23:40] <wtay> right after the chain function is set
[23:40] <omega_> checking
[23:41] <omega_> hmm, ok
[23:41] <omega_> whoops
[23:41] <shitowax> cu
[23:41] <omega_> l8r
[23:41] <omega_> how do we catch that with the validator?
[23:41] <wtay> cya
[23:41] shitowax (yann at Mix-Annecy-101-1-112.abo.wanadoo.fr) left #gstreamer.
[23:42] <omega_> committed
[23:43] <hadess> Uraeus: is darin around ?
[23:43] <wtay> omega_: and then it crashes somewhere else :-)
[23:43] <omega_> mu
[23:44] <ajmitch> automake is still chugging away ;)
[23:44] <wtay> Can't resolve host name "unix"!
[23:44] <omega_> kill it, patch it, restart it
[23:44] <wtay> what's that?
[23:44] <Uraeus> hadess: think he is in #nautilus
[23:44] <omega_> um?
[23:44] <hadess> anybody thought of porting smpeg to gstreamer ? it seems to have a better error recovery than whatever we're using now
[23:44] <Uraeus> hadess: he was at least not long ago
[23:44] <ajmitch> omega_: have to find the source of it first
[23:44] <omega_> hadess: yes, I thought of that a while ago
[23:44] <omega_> ajmitch: of which? automake? /usr/bin/automake
[23:45] <omega_> patch is in gstreamer/README
[23:45] <omega_> hadess: oh, you mean porting smpeg as a plugin? that won't bring anything but the codec
[23:45] <hadess> omega_: it's much faster after having applied the patch
[23:45] <omega_> what I want to do is implement the smpeg API with a gstreamer lib, so smpeg-based programs can use any codec we support <g>
[23:45] <ajmitch> omega_: ah, true, it's not a compiled program ;)
[23:46] chillywilly (baumannd at d4.as13.nwbl0.wi.voyager.net) joined #gstreamer.
[23:46] <ajmitch> taaz: still around? ;)
[23:46] <hadess> omega_: hmm, that wasn't really what i was thinking of, but it seems interesting
[23:46] <Uraeus> why do we get two copies of all the livid/gstreamer cross-postings?
[23:46] <omega_> hrm, not sure.  ChiefHighwater mentioned that too
[23:46] <chillywilly> where's the debs of gst at man?
[23:46] <chillywilly> :P
[23:47] <omega_> I get two because I'm on both lists
[23:47] <wtay> I'm not but I get two too
[23:47] <omega_> odd
[23:47] <Uraeus> I am on only gstreamer-devel
[23:47] <omega_> ENOCLUE
[23:47] <chillywilly> heh
[23:48] <omega_> wtay: you get a segv still?
[23:48] <wtay> omega_: hmm, it keeps on, typefinding
[23:48] <omega_> ok, that's what I get
[23:48] Action: omega_ made several mistakes
[23:48] <omega_> trying to settle on the best way to keep track of some dozen pads and caps
[23:49] <wtay> PAUSED->PLAYING: element "autoplugger0" has parent "launch" and sched 0x8072790
[23:49] <wtay> PAUSED->PLAYING: element "esdsink0" has parent "launch" and sched 0x8072790
[23:49] <wtay> ** CRITICAL **: file gstpad.c: line 626 (gst_pad_connect): assertion `GST_RPAD_PEER(realsink) == NULL' failed.
[23:49] <omega_> yup
[23:49] <wtay> ok
[23:49] Action: omega_ losing track
[23:50] <ajmitch> chillywilly: taaz has yet to become a debian maintainer, iirc ;)
[23:50] <omega_> wtay: so, any idea how to catch that plugin mistake?
[23:50] <wtay> omega_: the final segv is just EOS here
[23:50] <taaz> i get 5 copies of livid/gst posts ;)
[23:50] <ajmitch> hmm, patched automake seems to only use about 3 MB instead of 400
[23:51] <omega_> would have to keep track of what pads are instantiated within which context, and see if they stick around or get leaked
[23:51] <wtay> omega_: you can't add pads to a plugin that violate the template
[23:51] <omega_> but which template?
[23:51] <wtay> any
[23:51] <omega_> hmm
[23:51] <taaz> i'll get the debs done sometime...
[23:51] <wtay> matching the template padname
[23:51] <omega_> ok, that'll help, but it could still leak, if both pads match, or the first one is the bogus one and is never added
[23:51] <ajmitch> taaz: chillywilly is just lazy, so don't worry about him ;)
[23:51] <taaz> i'm writing msg to debian-metnors about package explosion right now
[23:52] <omega_> taaz: cc gst-devel
[23:52] <Uraeus> hadess: darin is there now
[23:52] <hadess> Uraeus: i know, i'm talking to him =)
[23:52] <taaz> omega_: why?
[23:52] <omega_> because it's relevant
[23:52] <wtay> omega_: the plugin also misbehaves if it doesn't create a pad for the ALWAYS template
[23:52] <taaz> ok
[23:52] <omega_> ok, you added these to the validation node?
[23:52] <taaz> i'm going to post about a few other issues too
[23:52] <wtay> will do now
[23:53] <taaz> like the lintian PIC errors
[23:53] <omega_> who's gonna build the validator?
[23:54] <omega_> and maintain the public error log on the web <g>
[23:54] <wtay> hmm, validating a new pad against a padtemplate requires every plugin to have a template...
[23:54] <omega_> Uraeus: did you get my comment on alpha?
[23:54] <wtay> omega_: I want to give the validator a try
[23:54] <omega_> ok
[23:55] <Uraeus> omega_: it is in, reload and see
[23:55] <wtay> I'm like doing lots of testing now anyway
[23:55] <omega_> it is?
[23:55] <Uraeus> yes
[23:55] <omega_> Uraeus: did you forget to save?
[23:55] <Uraeus> no
[23:55] <omega_> um, cacheless reload shows no changes
[23:56] <omega_> alpha cvs still red, listed as unknown
[23:56] <Uraeus> omega_: ok, alpha+linux works well?
[23:56] <omega_> yeah, see above
[23:56] <Uraeus> omega_: or Alpha+True64?
[23:56] <omega_> unknown
[23:56] <omega_> but if it's got gcc it should work
[23:57] <hadess> would it be possible to add a "wanted plugins" table ?
[23:57] <omega_> that's the GstWishList
[23:58] Action: hadess joins the wiki
[23:59] <omega_> Uraeus: I'd move bonobo and mozilla to applications, not plugins
[23:59] <Uraeus> omega_: ok
[23:59] <hadess> hmm, could somebody add me to the sourceforge list of developers ?
[23:59] <omega_> also, s/Eric/Erik/
[00:00] --- Thu May 17 2001
[00:00] <wtay> omega_: I have some local object destroy updates (reducing the thread leaks, passing regression tests)
[00:00] <omega_> you don't need that for the wiki...
[00:00] <omega_> but you have code to maintain now, so <g>
[00:00] <omega_> ok
[00:00] <omega_> committable?
[00:01] <wtay> just doing the test again to dblcheck
[00:02] maYam (mayam at cable-195-162-214-58.upc.chello.be) joined #gstreamer.
[00:02] <wtay> yo
[00:02] <maYam> helloo :-
[00:02] <maYam> bye!
[00:02] <ajmitch> hi maYam
[00:03] <wtay> omega_: thread destruction fails (segv)... dunno it that's bad enough not to commit it.. 
[00:03] <omega_> if it helps other stuff, commit, and make a note of it somewhere (-devel?)
[00:04] <wtay> omega_: it in the children/ghostpads case
[00:04] <omega_> ok
[00:04] <hadess> hey maYam
[00:05] <wtay> if you plan on using threads with ghostpads, I better not commit it
[00:05] <omega_> I don't, just yet
[00:05] <ajmitch> omega_: are you planning to try and get gstreamer into GNOME 2.0?
[00:05] <wtay> omega_: it solves the the thread with elements case without leaks though
[00:05] <omega_> if they want it....
[00:05] <ajmitch> i do ;)
[00:05] chillywilly (baumannd at d4.as13.nwbl0.wi.voyager.net) left irc: Philosophers and plow men, each must know his part, to sow a new mentality closer to the heart...
[00:06] <omega_> wtay: ok
[00:06] <wtay> omega_: ok, you know who to complain to :-)
[00:06] <omega_> and complain I will <g>
[00:06] CHW (paul at temple-baptist.com) joined #gstreamer.
[00:06] BBB (BBB at ucu-105-116.ucu.uu.nl) got netsplit.
[00:06] thomas (thomas at 212-100-172-175.adsl.easynet.be) got netsplit.
[00:06] Uraeus (cschalle at c224s9h5.upc.chello.no) got netsplit.
[00:06] ChiefHighwater (paul at temple-baptist.com) got netsplit.
[00:06] maYam (mayam at cable-195-162-214-58.upc.chello.be) got netsplit.
[00:06] <wtay> omega_: be warned: I'll tkae my time to fix it  :)
[00:07] <omega_> better not ;-)
[00:09] <hadess> "v4lsrc, not claimed" <- what is that boys ?
[00:09] <omega_> you claim it?
[00:10] <hadess> god no
[00:10] <omega_> darn
[00:10] <hadess> i first came to gstreamer because there would be somebody else maintaining this code
[00:10] <omega_> doh
[00:11] <hadess> but i must say that this is a hell of a pita
[00:11] <omega_> which?
[00:11] <hadess> v4lsrc
[00:11] <omega_> heh
[00:11] <omega_> yeah, complex API that's not well designed, it has its moments
[00:11] <hadess> dodgy et all
[00:11] Action: omega_ much prefers v4l2
[00:11] <omega_> not only because he helped define bits of it....
[00:12] <hadess> you're biased
[00:12] <omega_> so?
[00:12] <hadess> hehe
[00:13] <omega_> <gasp> someone is complaining about an application crashing his OS
[00:13] <omega_> he's running w2k.  DUH!
[00:14] <hadess> gstreamer-launch gnomevfssrc location=http://www.eavel.com/gnome.mp3 ! mad ! osssink
[00:14] <omega_> it's an SDL app using direct3d.  what on earth does he expect??
[00:14] <omega_> s/v/z/ ??
[00:14] <hadess> no no
[00:14] <omega_> uh, ok
[00:14] <hadess> try it out
[00:14] <ajmitch> hadess: over a 33.6k modem? ;)
[00:14] Action: omega_ switches to HEAD
[00:14] <omega_> ergh
[00:15] <omega_> no gnomevfssrc
[00:15] <ajmitch> omega_: how much is different between HEAD & INCSCHED1?
[00:15] <omega_> um, you already know the answer to that
[00:15] <omega_> I'll have to write up a whitepaper on it soon
[00:15] <hadess> the incremental scheduler
[00:16] <omega_> mu, I can't get enough bandwidth for it
[00:16] <hadess> ajmitch: actually, yes, if you could
[00:16] <ajmitch> hadess: plus a whole lot more, no doubt ;)
[00:16] <hadess> what it does ?
[00:16] <hadess> omega_: what does it do ?
[00:16] <omega_> uh, it works better <g>
[00:16] <omega_> let me write something up to the list, it'll make more sense
[00:16] <hadess> ok
[00:17] BBB (BBB at ucu-105-116.ucu.uu.nl) got lost in the net-split.
[00:17] thomas (thomas at 212-100-172-175.adsl.easynet.be) got lost in the net-split.
[00:17] Uraeus (cschalle at c224s9h5.upc.chello.no) got lost in the net-split.
[00:17] ChiefHighwater (paul at temple-baptist.com) got lost in the net-split.
[00:17] maYam (mayam at cable-195-162-214-58.upc.chello.be) got lost in the net-split.
[00:17] <ajmitch> hadess: you'll have to wait until gstreamer has compiled ;)
[00:17] Action: hadess needs to fix the problem when open failed...
[00:17] <omega_> is this your domain?  I can't get much bandwidth off it
[00:17] <hadess> this is not my domain
[00:17] <hadess> i can get enough bandwidth from here
[00:19] <wtay> are this the smurfs or what? <g>
[00:20] <hadess> ftp://ftp.dkuug.dk/pub/GUADEC2001 <- videos from guadec
[00:20] <omega_> sounds like southpark
[00:20] <hadess> it is south park
[00:20] <wtay> ok
[00:20] <omega_> ah, there we go
[00:20] Action: wtay tries to add buffering to it
[00:20] Uraeus (cschalle at c224s9h5.upc.chello.no) joined #gstreamer.
[00:21] <Uraeus> is it Thomas or Ajmitch who will make the new OSS plugins?
[00:21] <omega_> um, anyone know which one is the gstreamer talk?
[00:21] <wtay> omega_: what about transcoding your video?
[00:21] <ajmitch> Uraeus: you want new ones now? ;)
[00:22] <Uraeus> ajmitch: I was just refering to the recent thread on gstreamer-devel ;)
[00:22] <ajmitch> Uraeus: well, thomas will probably have more knowledge & time
[00:22] <ajmitch> Uraeus: i have exams in a few weeks ;)
[00:23] <wtay> what time was that gst talk again? <g>
[00:23] BBB (BBB at ucu-105-116.ucu.uu.nl) joined #gstreamer.
[00:23] Kuroyi (rick at ubr-35.82.223.wmelbourne.cfl.rr.com) left irc: Read error to Kuroyi[ubr-35.82.223.wmelbourne.cfl.rr.com]: EOF from client
[00:24] sienap (synap at ipc379c2a3.dial.wxs.nl) joined #gstreamer.
[00:24] <ajmitch> sienap!
[00:24] <sienap> Hi all
[00:24] <sienap> :)
[00:24] Action: ajmitch runs
[00:24] <sienap> run human
[00:24] <omega_> 15:15
[00:24] <sienap> i am superior
[00:24] <sienap> you are infinior
[00:24] <sienap> i am the alpha and the omega
[00:25] <omega_> no, I'm omega
[00:25] <sienap> he
[00:25] <sienap> :)
[00:25] <sienap> lol
[00:25] <sienap> maar goed
[00:25] <sienap> wtay awake ?
[00:25] <sienap> hi all btw ;)
[00:25] <wtay> yo
[00:25] <hadess> stream-4-2001-04-07-15:16:46.mpg <- from day2 i think
[00:26] <hadess> the 15:13 one is gnome office
[00:26] <ajmitch> sienap: quoting Revelations now? ;)
[00:27] <hadess> 1.89 GB <- gah !
[00:27] <ajmitch> hadess: quite big? ;)
[00:27] <hadess> sorry but 2G are better used with mp3s...
[00:28] omega_ (omega at omegacs.net) left irc: Read error to omega_[omegacs.net]: Connection reset by peer
[00:30] <ajmitch> hadess: why so big? ;)
[00:31] <hadess> 1 Mbps
[00:31] Action: ajmitch cannot believe an mpeg could be that big....
[00:32] Action: hadess laughs in face of adversity and takes on the challenge of making the gnomevfs sink work in the next hour
[00:33] <sienap> SUCCES uber slut :)
[00:34] <ajmitch> hehe, i think i'll try out caps negotiation on the volume plugin this afternoon then ;)
[00:34] <wtay> ./gstreamer-launch --gst-mask=-1 gnomevfssrc location=ftp://ftp.dkuug.dk/pub/GUADEC2001/day2/stream-4-2001-04-07-15%3A16%3A46.mpg ! mpeg1parse video_00 ! queue ! { mp1videoparse ! mpeg_play ! sdlvideosink }
[00:35] omega_ (omega at omegacs.net) joined #gstreamer.
[00:35] <sienap> lol
[00:35] <hadess> wtay: nice one =)
[00:35] <wtay> ./gstreamer-launch --gst-mask=-1 gnomevfssrc location=ftp://ftp.dkuug.dk/pub/GUADEC2001/day2/stream-4-2001-04-07-15%3A16%3A46.mpg ! mpeg1parse video_00 ! queue ! { mp1videoparse ! mpeg_play ! sdlvideosink }
[00:35] <omega_> Incsched do that?
[00:35] <wtay> HEAD
[00:35] <ajmitch> wtay: how fast is your connection? :)
[00:35] <wtay> waaay oo slow for that ;:)
[00:36] <ajmitch> hehe
[00:36] Action: ajmitch would expect to have that video downloaded in about a month ;)
[00:36] Action: omega_ runs autogen.sh;make on HEAD
[00:36] <wtay> lol
[00:36] <omega_> um, HAVE_LIBSDL does not appear in AM_CONDITIONAL
[00:36] <wtay> sdlsink doesn't show yet... maybe it does after decoding a frame..
[00:37] <wtay> 80k/sec downloading...
[00:37] <omega_> what rate is the video?
[00:37] <wtay> 1Mbps
[00:37] <omega_> ok
[00:38] <wtay> 150K/sec should do
[00:38] <omega_> barely
[00:38] <wtay> VCD is 150K/sec
[00:38] <omega_> yup
[00:38] <omega_> um, where's the check for sdl?
[00:38] <wtay> grep SDL?
[00:39] <omega_> I didn't see anything from configure
[00:39] <omega_> maybe I missed it
[00:39] <omega_> no, no check
[00:39] <wtay> my fault, commiting
[00:39] <omega_> and the makefile.am
[00:39] <omega_> not listed in AM_OUTPUT
[00:39] Action: ajmitch wonders why windows media player was faster playing mpeg videos than gstmediaplay ;)
[00:40] Action: ajmitch ducks, runs, & hides ;)
[00:40] <omega_> because intel wrote those codecs
[00:40] <wtay> and because DirectX doesn't suck as much as Xv
[00:40] <ajmitch> wtay: i can't use Xv ;)
[00:40] <omega_> DirectDraw
[00:40] <wtay> hmm, still no sdlvideosink
[00:40] <wtay> Direct{whatever}
[00:41] <omega_> not in acconfig.h
[00:42] <wtay> fixing
[00:43] <wtay> no sdlsink... trying with aasink
[00:43] <ajmitch> ugh
[00:43] <ajmitch> why aasink? ;)
[00:43] <hadess> what is the "handoff" signal in the disksink ?
[00:44] <omega_> if you want to capture the buffers into your application
[00:44] <hadess> ok
[00:45] <ajmitch> bbl
[00:45] Nick change: ajmitch -> aj_uni
[00:46] Action: omega_ waits for HEAD to build
[00:48] <Uraeus> sleeptime, night
[00:48] <omega_> l8r
[00:48] Uraeus (cschalle at c224s9h5.upc.chello.no) left #gstreamer.
[00:49] <wtay> hmm, weird mpeg, plaympeg fails on it and so does gstreamer
[00:49] <wtay> xine works
[00:50] <omega_> it's mpeg2
[00:51] <omega_> now /me knows why he doesn't watch SouthPark
[00:51] <wtay> hmm, mpeg2parse has problems too
[00:53] Action: wtay rebuilds with debugging
[00:59] <hadess> what can i use to test the gnomevfs sink ?
[00:59] <wtay> need to sleep now
[00:59] <wtay> cya guys
[00:59] <hadess> cya
[00:59] Nick change: wtay -> wtay-zZz
[00:59] <omega_> l8r
[01:00] <hadess> omega_: ?
[01:00] <omega_> hadess: test what part of it?
[01:00] <wtay-zZz> hadess:  ./gstreamer-launch gnomevfssrc location=http://gstreamer.net/media/south.mp3 ! mad ! osssink
[01:00] <hadess> wtay-zZz: the _sink_
[01:00] Action: wtay-zZz shuts up and goes to bed
[01:00] <omega_> oh, um
[01:00] <hadess> wtay-zZz: hehe
[01:01] <taaz> oops... my cc of that msg to gst-devel may confuse people ;)
[01:01] <taaz> oh well
[01:01] <omega_> why?
[01:01] aj_uni (ajmitch at p10-max5.dun.ihug.co.nz) left irc: Ping timeout for aj_uni[p10-max5.dun.ihug.co.nz]
[01:07] <hadess> gnome-vfs doesn't seem to be able to open() a file it doesn't already exist...
[01:07] <hadess> that could be a major pb
[01:08] <omega_> rather fundamental...
[01:08] <hadess> indeed
[01:08] <hadess> open: Success <- weirf
[01:09] <hadess> i have an open success but it doesn't work...
[01:09] sienap (synap at ipc379c2a3.dial.wxs.nl) left irc: sienap has no reason
[01:26] aj_uni (ajmitch at p61-max6.dun.ihug.co.nz) joined #gstreamer.
[01:26] <hadess> omega_: success !
[01:27] <hadess> and i'm about 6 minutes early :P
[01:29] <omega_> howso?
[01:29] <hadess> * hadess laughs in face of adversity and takes on the challenge of making the gnomevfs sink work in the next hour
[01:29] <omega_> oh
[01:29] Action: omega_ wasn't here for that
[01:30] <hadess> and i just finished it =)
[01:30] <hadess> i'm mailing you the new Makefile.am and the source for this file
[01:30] <omega_> ok
[01:31] <hadess> should it default to erasing the file or not ?
[01:31] <omega_> no, an option
[01:31] <omega_> and a signal when that's the case
[01:31] <hadess> hmm, need a couple more time then
[01:33] <omega_> Filesystem           1k-blocks      Used Available Use% Mounted on
[01:33] <omega_> grrr
[01:33] <omega_>  /dev/hda1              2064144   1882216     77048  97% /
[01:33] <omega_>  /dev/hda2              2028771   1906598     17291 100% /home
[01:33] <omega_>  /dev/hda6             10831520   8208988   2622532  76% /scratch
[01:33] <omega_>  /dev/hde1             78786704  70585788   4198756  95% /scratch2
[01:33] <omega_>  /dev/hdg1             78786704  59777872  15006672  80% /scratch3
[01:33] <omega_>  /dev/hdh1             80040784    507132  79533652   1% /scratch4
[01:38] matth_ (matth at qwest.dsplinux.net) joined #gstreamer.
[01:38] <omega_> yo
[01:41] <hadess> omega_: huh, i have a stupid question... how do i launch a signal ?
[01:41] <omega_> gtk_signal_emit
[01:42] <omega_> look for examples
[01:42] <omega_> gstpad.c has several
[01:50] arik (arik at sdn-ar-003waseatP258.dialsprint.net) joined #gstreamer.
[01:51] <omega_> yo
[01:51] <arik> hey
[01:54] <hadess> done
[01:54] <hadess> hey arik
[01:54] <arik> hey hadess 
[01:55] <hadess> omega_: have i delivered what you expected master ?
[01:56] <omega_> uh, I think so, slave <g>
[01:56] <hadess> try it out please
[01:56] <omega_> is it committed?
[01:56] <hadess> no, but i sent you the tarball
[01:56] <omega_> ok
[01:56] <arik> hadess: dude!
[01:56] <arik> tarball?
[01:56] <omega_> oops.  autoplug kinda worked <g>
[01:57] <hadess> hehe
[01:57] <hadess> arik: si ?
[01:57] <hadess> que tal ?
[01:57] <hadess> or something like that
[01:57] <arik> hadess: tarball?
[01:57] <omega_> tmbg at 4x is funny <g>
[01:58] <hadess> arik: a tarball is a tar.gz file
[01:58] <arik> hadess: ;-P
[01:58] <arik> hadess: duh
[01:58] <arik> hadess: did you send omega_ your app?
[01:58] <hadess> or a tar file...
[01:58] <hadess> arik: no, i sent him the gnomevfs sink
[01:58] <omega_> you have cvs commit access you know
[01:59] <arik> ah
[01:59] <omega_> just have to switch to a !anon working copy
[01:59] <hadess> do i ?
[01:59] <omega_> should
[01:59] <hadess> cool, i need to update my tree then, and learn the sourceforge way
[01:59] <omega_> you need a new tree
[01:59] <hadess> i know
[01:59] <omega_> unless you have a good script to convert
[02:00] <omega_> pretty easy, actually
[02:00] <omega_> just replace all the Root files, afaict
[02:00] <hadess> whatever, i just rm -rf'ed it :P
[02:00] <omega_> hope you didn't have significant changes
[02:00] <omega_> like your gnomevfs stuff <g>
[02:00] <omega_> which I don't have yet
[02:00] <hadess> nothing in the tree i use for CVS
[02:01] <hadess> i always work on a copy
[02:01] <omega_> ok
[02:01] <omega_> got it
[02:01] <omega_> ooooooh, cp(1) implemented with gstreamer!
[02:01] <hadess> <g>
[02:02] <omega_> wtay loses
[02:02] <hadess> that was the only test i could think of...
[02:03] <hadess> if you could come up with something like the same with a tee inside and sound output, that would be cool (stream saving)
[02:03] <omega_> cool, ok, I'll worry about it a little later though <g>
[02:08] matth_ (matth at qwest.dsplinux.net) left irc: Ping timeout for matth_[qwest.dsplinux.net]
[02:09] <hadess> shall i commit ?
[02:09] <omega_> yes, you should be committed <g>
[02:10] <hadess> cvs [server aborted]: "add" requires write access to the repository <- hmmm
[02:10] <omega_> hmm, maybe you have to wait a little while
[02:11] <omega_> you're set up for ssh cvs and all that?
[02:11] <hadess> i just checked out the whole gst using my sf account...
[02:11] <omega_> hmm
[02:12] <hadess> export CVSROOT=':pserver:hadess at cvs.gstreamer.sourceforge.net:/cvsroot/gstreamer' <- does yours look like this ?
[02:12] <hadess> except the bit with hadess of course
[02:12] <omega_> omegahacker at cvs.gstreamer.sourceforge.net:/cvsroot/gstreamer
[02:12] <omega_> not pserver
[02:12] <hadess> hmm, k
[02:13] <omega_> and I don't set the CVSROOT, just let CVS figure it out based on where I am
[02:13] <omega_> because I have so many cvs working copes
[02:13] <hadess> i set cvs root only for the first checkout
[02:13] <omega_> cvs -d omegahacker at cvs.gstreamer.sourceforge.net:/cvsroot/gstreamer
[02:13] <omega_>  co gstreamer
[02:14] <hadess> ok, it asked me about ssh key and stuff this time
[02:14] <omega_> right
[02:14] <hadess> thanks
[02:14] <omega_> details <g>
[02:15] thomas (thomas at 212-100-172-175.adsl.easynet.be) joined #gstreamer.
[02:15] Action: taaz laughs... cp with gstreamer hehe
[02:16] <omega_> it even reads from http and other locations, which cp doesn't do
[02:17] <taaz> does it just write to files?
[02:17] <omega_> I dunno, hadess: can gnomevfs upload to ftp and http?
[02:18] <hadess> maybe ftp, webdav for sure on http
[02:18] <hadess> committed =)
[02:19] <omega_> good, the world feels safer now with you under someone's watchful eye <g>
[02:19] <hadess> haha
[02:20] <arik> omega_: ;-)
[02:23] <hadess> any reasons v4l is the only plugin under capture ?
[02:29] <omega_> yup
[02:30] Action: omega_ will be back
[02:49] <arik> dum de dum
[02:49] <arik> hadess: does gstmediaplay work for you?
[02:49] <hadess> does indeed
[02:50] <arik> interesting
[02:50] <arik> on incshed1?
[02:50] <arik> to play video?
[02:52] <hadess> no, i use HEAD, INCSCHED1 is broken and broken by omega_ and wtay
[02:52] <arik> heh
[02:52] <hadess> you shouldn't use incsched1 if you don't hack on core gst
[02:52] <arik> ok
[02:52] <arik> maybe you're right
[02:52] <arik> omega_ was the one that told me to use it though
[02:54] <hadess> you have better chances of being able to play something if you use HEAD
[02:54] <arik> hadess: *nod*
[02:54] <arik> and my full-screen stuff is done i just can't test it
[02:54] <arik> maybe i should go back to head
[02:55] <hadess> full-screen for the xvideosink ?
[02:56] <arik> hadess: no full-screen just for the player for now
[02:56] <arik> everntually it should go into the sink
[02:56] <hadess> should... i will need it :)
[02:57] <arik> ;-)
[02:57] <arik> the current way of doing it is a bit of a hack
[03:11] Nick change: aj_uni -> ajmitch
[03:13] <ajmitch> hi all
[03:13] <arik> hey
[03:15] <ajmitch> what's been happening since i left? ;)
[03:15] <arik> nothing much and i need to go ;-)
[03:15] arik (arik at sdn-ar-003waseatP258.dialsprint.net) left irc: leaving
[03:16] <ajmitch> ok
[03:19] Action: omega_ goes away for a while
[03:19] omega_ (omega at omegacs.net) left irc: [x]chat
[03:40] ajmitch (ajmitch at p61-max6.dun.ihug.co.nz) left irc: Ping timeout for ajmitch[p61-max6.dun.ihug.co.nz]
[03:45] hadess (hadess at pc2-guil2-0-cust121.gui.cable.ntl.com) left irc: sleep
[03:46] ajmitch (ajmitch at p7-max5.dun.ihug.co.nz) joined #gstreamer.
[04:17] matth_ (matth at qwest.dsplinux.net) joined #gstreamer.
[05:01] omega_ (omega at temple-baptist.com) joined #gstreamer.
[05:02] <ajmitch> hi omega_
[05:02] <omega_> yo
[05:02] <omega_> taaz: I agree with Jaldhar
[05:04] <ajmitch> split it?
[05:05] <omega_> yeah.  I can't see any sane way to group them
[05:06] Nick change: omega_ -> omega_away
[05:06] Kuroyi (rick at ubr-35.82.223.wmelbourne.cfl.rr.com) joined #gstreamer.
[06:11] <ajmitch> hmm
[06:11] Action: ajmitch yells to noone in particular