IRC Logs

*******************************************************************
[03:01] ramon ([email protected]) joined #gstreamer.
[03:04] <ramon> Hi.
[03:05] <BBB> hi
[03:08] <ramon> Do you already have a stable internet connection?
[03:12] <BBB> nah
[03:12] <BBB> not really
[03:12] <BBB> but it's ok
[03:13] <BBB> my laptop just turned itself off because of low batery...
[03:13] <BBB> ohwell
[03:14] <ds-work> BBB, meet power cord.  Power cord, meet BBB
[03:14] <BBB> I just got that ;)
[03:14] <BBB> thanks anyway
[03:14] <ds-work> that's happened to me several times
[03:14] #gstreamer: mode change '+o BBB' by ChanServ!ChanServ@services.
[03:14] <ramon> :)
[03:14] Topic changed on #gstreamer by [email protected]: GStreamer "In A Bar" 0.7.5 and GStreamer Plugins "Under The Sea" 0.7.5 are out - see http://gstreamer.freedesktop.org/
[03:14] #gstreamer: mode change '-o BBB' by [email protected]
[03:15] Action: BBB likes this better
[03:15] <ds-work> "where the hell are those beeping sounds coming from?"
[03:15] <BBB> heh :)
[03:17] <ramon> I guess that if a pad has several caps, the first has got higher priority and is tried first, is it?
[03:18] <ds-work> no
[03:18] <ds-work> there's no prioritization in caps
[03:19] <ramon> So if two elements are connected and support several languages for communication, how is one of them choosen?
[03:21] <ds-work> fixate functions
[03:21] <BBB> either can "prefer" one, and choose the preferred one using fixate functions
[03:21] <BBB> applications can do that too
[03:21] <BBB> or you can use filtered caps
[03:21] <ds-work> usually the app will connect to a fixate signal and decide
[03:21] <BBB> and at last, the capsnego core decides
[03:21] Action: BBB is slow
[03:23] <ramon> Thanks.
[03:25] Rotty ([email protected]) left irc: Remote closed the connection
[03:47] <ramon> What demux are most maintained? qtdemux? asfdemux?
[03:47] <ds-work> avidemux is probably the best right now
[03:49] <ramon> Thanks. I just want it to use as a pattern source for RTP.
[03:51] jimmy_dean ([email protected]) left irc: "Carrier Signal Lost"
[03:53] Action: BBB wants to implement seeking in qtdemux before 0.8.0
[03:53] Action: BBB is first going to mess around with ffmpeg
[03:53] Action: BBB kicks exchange for being such an ass
[03:53] <BBB> I just decided I'd use my school's exchange server
[03:53] <BBB> so I buy a connector license
[03:53] <BBB> and yes, I can send email
[03:54] <BBB> from my university's email addy
[03:54] <BBB> but not from [email protected]! exchange checks the from: address!
[03:54] <BBB> :'(
[03:55] <arana> :/
[03:55] <arana> hrmmm
[03:55] Nick change: arana -> spyder482
[03:57] <ramon> Exchange is doing the right thing. You should not forge email.
[03:58] <BBB> that's wrong-side security ;)
[03:58] <ramon> A GstStructure is a "single" caps, that is, the specification of a single format, is it?
[03:58] <ds-work> that's not forged email
[03:58] <BBB> yes
[03:58] <BBB> you "hope" that the caller checks... if it doesn't, your whole system is messed
[03:58] <BBB> that's also why Gates' proposal is bullocks
[03:58] <BBB> it'll never work
[03:58] <BBB> receivers are resoonsible for their own security
[03:58] <BBB> not senders
[03:59] <ramon> I find it reasonable that the server of a domain checks that you are using that domain for sending.
[03:59] <BBB> maybe, but it means I cannot access external mail resources using that domain
[03:59] <BBB> and that sucks
[04:00] <ramon> The right thing seems to connect the SMTP server of bitfreak.net and send using that server.
[04:01] <ramon> My most important concern with forging is spam. On the other hand, with spambayes my spam is practically zero. Thus the issue is not that important.
[04:01] <BBB> that's called "open relay" and is the worst thing around :)
[04:02] lupusBE ([email protected]) left irc: "Bezig met verlaten"
[04:02] <walters> BBB: not if you use authenticated SMTP
[04:02] <walters> BBB: i use postfix/tls on verbum.org for authenticated relaying, works great
[04:03] <BBB> bitfreak.net isn't mine ;)
[04:03] <walters> nag them to install postfix/tls then :)
[04:03] <ds-work> I only send mail from the mail server itself
[04:03] Action: ds-work decides to start disliking the RFB protocol
[04:04] <walters> yeah, but then you have to either use some sucky tty-only mail client, or install X and stuff on the server...
[04:04] <ds-work> mutt serves me well
[04:05] Action: BBB just did as walters suggested
[04:05] Action: ds-work wonders when the last time dxr3 was compiled
[04:06] Action: BBB compiles dxr3 once in a while
[04:06] <BBB> I didn't test it for a long time, though
[04:06] <BBB> it'll take me another half year before I go there
[04:06] <ds-work> where do the headers come from?
[04:06] <BBB> from the kernel?
[04:07] <BBB> or, well, actually
[04:07] <BBB> it's a kernel patch
[04:07] <BBB> it's not part of the vanilla kernel yet
[04:07] <BBB> http://dxr3.sf.net/
[04:11] <ds-work> how are we supposed to describe keys in navigation events?
[04:11] <ds-work> right now, we're using the XKeysym string
[04:12] <ds-work> anyway, s/-work//
[04:13] <BBB> keys?
[04:15] <ramon> See you.
[04:15] <ramon> Going to bed.
[04:15] ramon ([email protected]) left irc: "Terminando cliente"
[04:17] <BBB> ds-work: would the 'don't export symbols' also fix the symbol conflict between wav_init, dv_init etc. between ffmpeg and libdv and others?
[04:26] <thaytoo> BBB: I don't think so - because it's the .so's that are being collected that then export those
[04:26] <thaytoo> so libdv.so exports dv_init regardless
[04:27] <BBB> yes, but libffmpeg.so or libgstffmpeg.so wouldn't
[04:27] <BBB> so wouldn't that fix it?
[04:27] <thaytoo> if libffmpeg.so stops exporting it, then yes
[04:27] <BBB> that's the idea ;
[04:27] <BBB> ;)
[04:28] <thaytoo> it's not going to fix clashes between libraries we depend on - they should be being nicer in the namespace they occupy
[04:28] <thaytoo> but it will fix it in libraries we control
[04:29] <BBB> we use a local ffmpeg, so that's cool ;)
[04:30] steveb_ ([email protected]) joined #gstreamer.
[04:31] <thaytoo> hi steve
[04:32] <BBB> Error while performing operation:
[04:32] <BBB> MAIL FROM response error: Client does not have permission to Send As this sender. <- aaaaaaaaaaaaaa
[04:32] <BBB> stupid @$^#^#!&#!%&#!&#!
[04:40] sub_pop ([email protected]) joined #gstreamer.
[04:40] Nick change: spyder482 -> spyder_homework
[04:46] Action: BBB throws his ffmpeg tree inside gst-ffmpeg and starts compilign
[04:48] steveb__ ([email protected]) left irc: Read error: 110 (Connection timed out)
[04:49] Action: BBB notes that it compiles so far
[04:52] <crazney> hrm, so i have this gstreamer stuff working using its threading thing.. what i'm doing is very basic, just playing a mp3 in a different thread, using a pipe with the fdsrc source.. anyhow, i have a question about the signals - is there a tick signal of some sort, so i can find out where in the audio file the play is up to, and also when its finished etc?
[04:54] <BBB> there is a tick signal, yes...
[04:55] Action: BBB looks up
[04:55] <taaz> you could use the fine GstPlay element
[04:55] <BBB> 'iterate'
[04:55] <BBB> and return FALSE :)
[04:55] <BBB> it's on the pipeline, thread or bin
[04:56] <BBB> whee... gst-ffmpeg compiles using my autotoolized ffmpeg :)
[04:58] <crazney> BBB: ok, i'll look it up in the docu. thanks.
[04:58] <crazney> hrm, i can't find it in the documentation..
[04:59] <BBB> depends on where you look
[04:59] <BBB> :)
[05:00] <crazney> ah, its not a signal, i attach with gst_bin_set_post_iterate_function instead?
[05:00] <BBB> but taaz is right, the 'gstplay' object (in gst-plugins: gst-libs/gst/play/play.h) is a good object for simple playback
[05:00] <BBB> hm? no
[05:00] <BBB> it *is* a signal
[05:00] <BBB> g_signal_connect (pipeline, "iterate", G_CALLBACK (cb_iterate), NULL);
[05:01] <BBB> static gboolean cb_iterate (GstBin *bin, gpointer data) { g_print ("iterate\n"); return FALSE; }
[05:01] <crazney> in the stable version of gsteamer?
[05:01] <crazney> gstreamer..
[05:02] ramon ([email protected]) joined #gstreamer.
[05:03] <crazney>  GLib-GObject-WARNING **: gsignal.c:1893: signal `iterate' is invalid for instance `0x81765f0'
[05:03] <ramon> I think that I heard in the mailing list that gst_try_set_caps is deprecated?
[05:03] <ramon> is it?
[05:04] <BBB> crazney: what object?
[05:04] <BBB> GstElement?
[05:04] <BBB> GstBin?
[05:04] <BBB> GstThread?
[05:04] <BBB> GstPipeline?
[05:04] <crazney> GstThread i guess
[05:04] <BBB> that should work...
[05:04] <crazney>     pipeline_thread = gst_thread_new ("pipeline");
[05:04] <crazney>     g_signal_connect (G_OBJECT(pipeline_thread), "iterate",
[05:04] <crazney>                       G_CALLBACK (cb_iterate), NULL);
[05:04] <crazney> and it gives me that warning.
[05:05] <BBB> hm... I do that too, and it works for me
[05:05] <BBB> do you use 0.6.x or 0.7.x?
[05:05] <crazney> 0.6.x
[05:05] <BBB> hm, I use 0.7.x...
[05:06] <crazney> i guess they changed it.
[05:06] <BBB> 'they' includes me ;)
[05:07] <crazney> oh :)
[05:07] <crazney> using that cb function i posted before works, however
[05:07] <crazney> <crazney> ah, its not a signal, i attach with gst_bin_set_post_iterate_function instead?
[05:08] spyder_homework ([email protected]) left irc: "too interesting sysKin... must do homework"
[05:09] <ramon> BBB: can you tell me if gst_try_set_caps is deprecated?
[05:10] <BBB> it is not
[05:10] <BBB> company tried to make us deprecate it
[05:10] <BBB> but we didn't ;)
[05:10] <BBB> crazney: funny... must be some weird bug in 0.6.x then
[05:10] <ramon> That makes it much easier for me.
[05:11] <crazney> BBB: well i couldn't find it in the 0.6.x docu, either.
[05:11] <BBB> the 0.6.x docs suck? :)
[05:11] Action: BBB cheers as he was just able to get a fresh tarball of his ffmpeg to build "mostly" in gst-ffmpeg
[05:12] Nick change: khalek -> kh_out
[05:12] <BBB> with almost no modification! :)
[05:18] <ramon> I am placing working files of rtp under gst-sandbox.
[05:18] <ramon> I hope there is no problem with that.
[05:19] jcsston` ([email protected]) joined #gstreamer.
[05:19] <ds> um, that's what I _told_ you to do
[05:20] <crazney> BBB: so how about when playing finishes?
[05:20] jcsston` ([email protected]) left irc: Client Quit
[05:23] <BBB> connect to the 'eos' signal
[05:23] <BBB> ramon: sounds good
[05:23] <BBB> ramon: do your development in the sandbox, make your own subdir in that module ("gst-rtp" or so) and have fun
[05:24] <crazney> BBB: ok, ta.
[05:24] <BBB> (did I mention that signals are important? ;) )
[05:25] <BBB> added plugin ffmpeg with 54 feature(s)
[05:25] <BBB> whee
[05:25] Action: BBB happy
[05:25] <BBB> tomorrow, I will work on adding new codecs to the ffmpeg plugin and integrating all this
[05:28] jcsston` ([email protected]) joined #gstreamer.
[05:38] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[05:39] ramon ([email protected]) left irc: Read error: 60 (Operation timed out)
[05:48] BBB ([email protected]) left irc: "Client exiting"
[05:55] Nick change: kh_out -> khalek
[06:14] Marsupilami23 ([email protected]) joined #gstreamer.
[07:23] thaytoo ([email protected]) left irc: "leaving"
[07:29] pmfp ([email protected]) joined #gstreamer.
[07:36] steveb__ ([email protected]) joined #gstreamer.
[07:38] steveb_ ([email protected]) left irc: Read error: 60 (Operation timed out)
[07:44] steveb_ ([email protected]) joined #gstreamer.
[07:45] ChrisHJW ([email protected]) joined #gstreamer.
[08:02] steveb__ ([email protected]) left irc: Read error: 110 (Connection timed out)
[08:03] steveb__ ([email protected]) joined #gstreamer.
[08:16] ChrisHJW ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[08:21] steveb_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[08:21] steveb_ ([email protected]) joined #gstreamer.
[08:35] steveb__ ([email protected]) left irc: Read error: 110 (Connection timed out)
[08:35] steveb__ ([email protected]) joined #gstreamer.
[08:39] markey ([email protected]) joined #gstreamer.
[08:50] Company ([email protected]) joined #gstreamer.
[08:53] steveb_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[09:11] Nick change: mathrick|sleep -> mathrick
[09:12] <mathrick> mornin
[09:12] alley_cat ([email protected]) left irc: Read error: 60 (Operation timed out)
[09:15] walters ([email protected]) left irc: Client Quit
[09:23] alley_cat ([email protected]) joined #gstreamer.
[09:29] <Company> weee, this chunklib idea is cool
[09:39] markey ([email protected]) left irc: "bbl"
[09:41] Nick change: mathrick -> mathrick|Uni
[09:43] Company ([email protected]) left irc: Remote closed the connection
[09:43] jcsston` ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:09] markey ([email protected]) joined #gstreamer.
[10:29] jdahlin ([email protected]) joined #gstreamer.
[10:35] thomasvs ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:39] markey ([email protected]) left irc: "leaving"
[10:42] wheels ([email protected]) left irc: "work"
[10:42] Action: bilboed is back (gone 09:28:57)
[10:49] apoc_ ([email protected]) joined #gstreamer.
[10:55] thomasvs ([email protected]) joined #gstreamer.
[11:04] Shoragan ([email protected]) joined #gstreamer.
[11:05] apoc ([email protected]) left irc: Read error: 110 (Connection timed out)
[11:07] apoc_ ([email protected]) left irc: "Leaving"
[11:16] Action: bilboed is away: vive l'eau qui nous rend propre...
[11:21] Rotty ([email protected]) joined #gstreamer.
[12:13] Action: bilboed is back (gone 00:56:38)
[12:13] Action: bilboed is away: au tél
[12:33] md` ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[12:34] <jdahlin> taaz: just wrote a filesrc element in python
[12:37] md` ([email protected]) joined #gstreamer.
[12:38] <jdahlin> quite interesting, because initial benchmarks shows it's only around 20% slower than the real file source
[12:40] Action: bilboed is back (gone 00:27:54)
[12:42] <KoRnouille> lo
[12:42] <bilboed> hi KoRnouille
[12:43] <KoRnouille> :)
[12:43] <KoRnouille> lo bilboed.
[12:43] <KoRnouille> hellaw jdahlin !
[12:43] <bilboed> jdahlin: you can write elements in python ?
[12:44] <KoRnouille> write  elements ?
[12:44] <jdahlin> bilboed: yes
[12:44] <bilboed> cool :)
[12:44] <KoRnouille> write more elements that the ones that already exist.
[12:44] <jdahlin> hello KoRnouille
[12:44] <jdahlin> KoRnouille: it's just testing
[12:45] Action: thomasvs wonders if the gst python bindings should follow the versioning of the core
[12:45] <thomasvs> jdahlin: how does pygtk versioning work ?
[12:46] <jdahlin> thomasvs: slowly getting there, haven't really tried to do some real stuff
[12:46] <KoRnouille> me neither
[12:46] <jdahlin> basic bin/element/pad functions work though
[12:51] <jdahlin> do you think it makes sense to add keyword like arguments for GstElement?
[12:51] <jdahlin> so you can add properties in the constructor, like g_object_new
[12:51] <jdahlin> gst.Element('filesrc', 'src', location='...')
[12:58] <thomasvs> yeah, sure
[12:59] <bilboed> definitely, it makes python programming easier (and shorter)
[12:59] <jdahlin> but it's not api compatible
[12:59] <jdahlin> with the C API
[12:59] <bilboed> yes it is
[12:59] <bilboed> if you keep the regular functions
[13:00] <bilboed> you must be able tu use python paradigms IMO
[13:01] <bilboed> the problem is that it must be coded in the python bindings... add the necessary calls to regular functions in __init__()
[13:02] <bilboed> Am I talking bs or does it make sense ?
[13:02] <jdahlin> it makes sense
[13:02] <bilboed> python programmer will love it
[13:02] <bilboed> s/programmer/devs
[13:03] <thomasvs> jdahlin: it only extends the C api
[13:03] <thomasvs> jdahlin: ie, if you don't specify args, it still works like the C api
[13:03] <bilboed> but it does require a little thinking (which functions/parameters can we call at init?)
[13:05] <bilboed> like the difference between gtkmm (only c->c++ bindings) and Inti (with c++ paradigms)
[13:10] ain ([email protected]) joined #gstreamer.
[13:42] jcsston ([email protected]) joined #gstreamer.
[13:52] <ain> thomasvs: is nautilus-media still the main place for a working media-info?
[14:01] teuf ([email protected]) joined #gstreamer.
[14:01] md` ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[14:01] <teuf> hi
[14:02] <teuf> is gnome-volume-control working with alsa for anyone ?
[14:03] Action: bilboed is away: gone to school
[14:04] md` ([email protected]) joined #gstreamer.
[14:11] <ain> teuf: no
[14:13] <thomasvs> ain: none of the two work, I'm updating right now in gst-plugins
[14:14] <thomasvs> ain: I have to rewrite big chunks because with ogg, it's not just "one decoder element" anymore, and I was using a hash table for decoder elements
[14:14] <ain> ok
[14:41] <teuf> yay!!! gnome-volume-control works with alsa oss emulation here ;)
[14:41] bilboed_school ([email protected]) joined #gstreamer.
[14:48] <teuf> thomasvs: bugzilla doesn't have a 0.7.6 target milestone
[14:50] <thomasvs> teuf: yeah, should add one
[14:53] jcsston ([email protected]) left irc: "VDubMOD watchout TCME is coming!"
[15:02] Nick change: khalek -> kh_zZz
[15:05] markey ([email protected]) joined #gstreamer.
[15:28] mxpxpod ([email protected]) left irc: ""If ever time should come, when vain and aspiring men shall possess the highest seats in Government, our country will stand i
[15:30] smoke ([email protected]) joined #gstreamer.
[15:32] ChriHJW_log ([email protected]) left irc: Read error: 110 (Connection timed out)
[15:37] ramon ([email protected]) joined #gstreamer.
[15:39] <ramon> Hi.
[15:39] ramon ([email protected]) left irc: Client Quit
[15:45] Nick change: markey -> AFK
[15:52] <taaz> jdahlin: do the examples work again too?
[15:53] <taaz> i don't think i ever figured out some performance issues
[15:53] <taaz> like the py identity element being faster than the C one ;)
[15:55] <taaz> lat.py or ilat.py tests that i think
[16:01] <jdahlin> taaz: yes, they seem to work
[16:01] <jdahlin> I think :-)
[16:01] <jdahlin> I'm just adding GstTagList.get_tags()
[16:02] bilboed ([email protected]) left irc: Read error: 110 (Connection timed out)
[16:03] bilboed ([email protected]) joined #gstreamer.
[16:08] Nick change: mathrick|Uni -> mathrick
[16:09] <taaz> jdahlin: which does what?
[16:09] <mathrick> re
[16:09] <jdahlin> taaz: returns a dictionary with tags
[16:10] <taaz> is there a morron common name for such functions?  like .as_dict()?
[16:10] <jdahlin> nop
[16:10] <taaz> maybe thats java code.. 
[16:10] <jdahlin> we could add a __getitem__ attribute maybe
[16:10] <taaz> you know there is a cooler way
[16:10] <taaz> yeah
[16:10] <jdahlin> so dict(tag_list)
[16:10] <jdahlin> hmm.
[16:10] <taaz> adding functions so it looks like a dict
[16:11] Nick change: AFK -> markey
[16:11] <taaz> i wasn't sure how to go about doing that
[16:12] <taaz> in a nice OO world you could just implement a few functions, do multi inheritance from dict too and it would all just work ;)
[16:12] <jdahlin> we need to add keys/__getitem__
[16:12] <jdahlin> there is a UseDict, just for that
[16:12] <jdahlin> UserDict
[16:12] <jdahlin> but we can subclass dict if we like
[16:13] <taaz> anyway, this was a TODO item that is more like just details once other things are working too ;)
[16:14] <taaz> is it just me or is the taglist api overkill?
[16:14] <taaz> all those type specific functions are probably worthless and should be in an override ignore section
[16:14] <taaz> the GValue ones work fine i think
[16:15] foser ([email protected]) joined #gstreamer.
[16:15] jcsston ([email protected]) joined #gstreamer.
[16:17] <jdahlin> it's very bad imho
[16:17] <jdahlin> at least for gst-python
[16:19] <jdahlin> taaz: I'm planning on removing 0.6 stuff from CVS HEAD
[16:20] <jdahlin> since it's no point supporting it
[16:20] <jdahlin> and for 0.9 we can just branch
[16:23] jcsston ([email protected]) left irc: "VDubMOD watchout TCME is coming!"
[16:25] mathrick ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[16:26] mathrick ([email protected]) joined #gstreamer.
[16:27] <taaz> why?
[16:29] jcsston ([email protected]) joined #gstreamer.
[16:34] <taaz> i guess it gets a bit complex supporting both at same time
[16:34] <jdahlin> 0.6 is practically dead now
[16:34] <taaz> it's just been easy enough to do and ignore some issues like examples that might not work in one ver or the other
[16:34] <taaz> hey, i use 0.6 more than 0.7
[16:35] <jdahlin> well, in terms of support, patches that goes upstream etc
[16:35] <taaz> rb runs many hours a day on 0.6 code ;)
[16:35] <taaz> and don't branch for 0.9.  branch for 0.8 and head tracks 0.9
[16:35] <jdahlin> right
[16:36] <taaz> i don't know what to do about versioning
[16:36] <taaz> following gst ver seems silly
[16:37] <taaz> it took great control for me not to release 0.1.0 as 1.0.0 in the first place ;)
[16:38] <mathrick> taaz: silly or not, it makes it easy for ppl to tell what gst version given gst-python ver wraps
[16:38] <taaz> i guess most people don't agree with my view on versioning though
[16:38] <mathrick> taaz: I tend to hate wrappers that do not state clearly what exact version they support
[16:39] <mathrick> in fact, I hate everything not stating its version clearly
[16:39] Action: thomasvs agrees they should follow core's major/minor, but not micro
[16:39] <taaz> mathrick: how about 2.3.0.0.7.5?  that way you know its for 0.7.5? ;)
[16:39] <mathrick> taaz: yeah, please enter your 12-digit pin ;)
[16:39] <jdahlin> taaz: why not do like in pygtk?
[16:40] <mathrick> s/pin/PIN/
[16:40] <jdahlin> taaz: gst.gst_version, gst.pygst_version ?
[16:40] <jdahlin> or you mean the releases?
[16:41] <taaz> jdahlin: following the major/minor of gst means that gst-py only has 1 digit to fiddle with.  which tells uses nothing of it's own state.  how do i add info like "this is a major upgrade incompatible with the last release" with just the micro?
[16:42] apoc ([email protected]) joined #gstreamer.
[16:42] <jdahlin> taaz: well, that can be true for 0.7, but not for 0.8
[16:42] <taaz> ?
[16:43] <thomasvs> taaz: until the bindings are more or less "finished", you should assume they break on each release anyway
[16:43] <thomasvs> taaz: ie, it's not worth considering until they really are in some sort of "this is what it'll be like" state
[16:43] <thomasvs> which will only happen after you've pythonified everything
[16:44] <mathrick> taaz: you could emule's versioning scheme, where 0.29a tends to be major upgrade over 0.29 ;)
[16:45] <taaz> mathrick: people who add letters to a numeric scheme are asses
[16:45] <taaz> it breaks everyones version checking code that assumed numbers
[16:45] <mathrick> taaz: yeah, I know, debian policy clashes with that
[16:47] <taaz> while i'm insulting people, i'd also like to express my disgust at those who release 1.2.3 then 1.2.3.1 as a minor update rather than just going to 1.2.4
[16:48] <taaz> i still don't understand.  what scheme should gst-python type projects use?
[16:49] <taaz> 0.7.x?  i don't like that limitation.  0.7.x.y.z?  that's long.  x.y.z?  that doesn't show what it's tracking.  x.y.z.0.7?  that's long and strange. ??
[16:50] <mathrick> taaz: follow whatever you wrap version, and append micro if you need to 
[16:50] <taaz> mathrick: so 0.7.M.x.y.z?
[16:51] <mathrick> taaz: like, say 2 digits, should be enough to be expressible, and stay short enough so people don't laugh at it ;)
[16:51] <taaz> so make up arbitrary constraints?
[16:52] Action: taaz sighs... computers suck.  i want gst-python-(0.8.0,1.2.3)
[16:53] <taaz> i feel i'm going to be overrulled on this nonsense ;)
[16:53] <thomasvs> taaz: I'm pretty sure that for now 0.7.x (where you increase x for each python release) is more htan enough
[16:54] <thomasvs> taaz: I'm sure people would prefer having actually working python bindings over perfectly versioned but crap python bindings :)
[16:55] <taaz> probably
[16:55] <taaz> in that case, someone please fix my segfaults and crashes.  i've failed to figure out the cause.
[16:56] <taaz> now is the exact time that we NEED more versioning numbers.  since it's changing fast enough that minor will be bumped on every release and micro will probably stay at 0  (in non-tracking versioning)
[16:57] <jdahlin> hmm, it's actually quite complicated to provide a dict interface
[16:57] <jdahlin> __getitem__/__setitem__/__len__/keys/has_key/get/copy
[16:59] <taaz> jdahlin: i was considering takign the route of wrapping them as gst._Structure and gst._TagList and making a python only hierarchy dict->Structure->TagList that call the appropriate backing store C impl object funcs
[16:59] <taaz> haven't had time to try that though
[17:00] mathrick ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[17:00] <taaz> or maybe UserDict is more appropriate for that
[17:00] Action: taaz runs off for a bit
[17:00] jcsston ([email protected]) left irc: "VDubMOD watchout TCME is coming!"
[17:06] yippi ([email protected]) joined #gstreamer.
[17:12] mxpxpod ([email protected]) joined #gstreamer.
[17:17] ChrisHJW_log ([email protected]) joined #gstreamer.
[17:19] mathrick ([email protected]) joined #gstreamer.
[17:33] <jdahlin> taaz: maybe later :)
[17:33] <jdahlin> So, the question is now
[17:34] <jdahlin> how can I find out how long an ogg is, in seconds?
[17:36] <mathrick> jdahlin: it depends on which stream you're asking about
[17:37] <mathrick> jdahlin: remember that ogg is container, and that's it's mainly used with vorbis audio is purely coincidence
[17:38] <jdahlin> mathrick: right, for a vorbis stream then
[17:38] sub_pop ([email protected]) left irc: "Client exiting"
[17:38] <mathrick> jdahlin: well, now it depends on wether it's CBR or VBR
[17:38] <mathrick> jdahlin: CBR is easy, size / framesize, VBR is (in general) impossible ;)
[17:39] <jdahlin> length on mp3 files would be nice too
[17:39] <mathrick> jdahlin: so for VBR, you can either guesstimate based on some kind of average, or use one of various tags used to give some clue about it
[17:40] <mathrick> jdahlin: same applies to mp3
[17:40] <jdahlin> right
[17:40] <mathrick> jdahlin: I think peek at XMMS code may be good, somehow it manages to get accurate VBR lengths
[17:41] <mathrick> jdahlin: although it may be it just cheats, reading whole file onl load
[17:41] <mathrick> s/onl/on
[17:42] <jdahlin> mathrick: I'd rather avoid looking at xmms' source
[17:42] <jdahlin> I've been there before
[17:42] <jdahlin> (unfortunately)
[17:42] <mathrick> jdahlin: aesthetics reason?
[17:42] <teuf> jdahlin: you can have a look at rhythmbox code too, it's pretty accurate
[17:42] <thomasvs> jdahlin: you do a query using length_total
[17:42] <teuf> (for mp3)
[17:42] <thomasvs> jdahlin: check media-info for example
[17:43] <jdahlin> thomasvs: where can I find media-info?
[17:44] <ain> gst-plugins/gst-libs/gst/media-info
[17:45] <jdahlin> thanks
[17:45] Nick change: ain -> iain
[17:57] <apoc> Can someone try a pipeline for me ?
[17:57] <apoc> gst-launch filesrc location=foo.mp3 ! mad ! wavenc ! filesink location=foo.wav
[17:57] <apoc> it doesn't work here
[17:58] <teuf> ERROR         default(19689) gstwavenc.c(591):gst_wavenc_chain:<wavenc0> encoder not initialised (input is not audio?)
[17:58] <teuf>  ?
[17:58] <taaz> most vorbis files have the length encoded
[17:58] <taaz> or at least ogginfo shows the total time
[17:58] <apoc> hmm same error here ... thx teuf
[18:00] <thomasvs> apoc: that's a bug then - apparently mad could not talk to wavenc, check the debug log
[18:00] sub_pop ([email protected]) joined #gstreamer.
[18:01] <thomasvs> hm, here it worked ?
[18:01] <thomasvs> yep
[18:01] <thomasvs> file plays fine
[18:04] teuf ([email protected]) left irc: "Client exiting"
[18:04] jong ([email protected]) joined #gstreamer.
[18:10] ChrisHJW ([email protected]) joined #gstreamer.
[18:19] <jong> hello, does anyone know the status of gstmm, the C++ bindings for gstreamer?
[18:21] <thomasvs> someone just updated them I think
[18:21] <thomasvs> not sure if that patch is commited yet
[18:21] <thomasvs> check bugzilla, open bugs
[18:22] <jong> thanks, i'll go check it out immediately
[18:22] <mathrick> jong: they are at least compilable, but mxpxpod is the one you want to ask
[18:22] <mxpxpod> I know for sure that that patch in bugzilla works with gstreamer-0.7.4
[18:23] <mathrick> mxpxpod: works as in it compiles, or actually is usable?
[18:23] <mxpxpod> mathrick: it should be actually useable
[18:24] <mxpxpod> not sure to what extent, but the helloworld example works
[18:24] <mxpxpod> and plays
[18:24] <jong> and does someone know when this could enter debian unstable? because it's still gstreamer 0.6.4...
[18:25] <mxpxpod> jong: I think you can d/l gstreamer-0.7.4 from unstable... I think it's called gstreamer-0.7
[18:27] <mathrick> mxpxpod: does gstmm cover plugin writing, or user api only?
[18:27] <mxpxpod> mathrick: I'm not sure, right now... sorry, I'm trying to program in VB and it's messing with my head :(
[18:28] <mathrick> mxpxpod: stop right now, when you still can!
[18:28] <mxpxpod> heh
[18:28] <mathrick> mxpxpod: why do are you doing that?
[18:28] <mxpxpod> mathrick: class
[18:28] <mathrick> mxpxpod: what kind of freaking class is it?
[18:28] <mxpxpod> Advanced VB.NET
[18:29] <mathrick> mxpxpod: huh?
[18:29] <taaz> jong: i haven't gotten around to doing the 0.7 debs yet.  sorry.  i'll try to get to that soon.
[18:29] <mxpxpod> mathrick: that's the title of the class
[18:29] <mathrick> mxpxpod: so you had basic one before?
[18:29] <mxpxpod> yeah
[18:29] <jong> just to be sure, I suppose I can connect slots and/or signals from gstmm to gtkmm?
[18:29] <jong> taaz: cool!
[18:30] <jdahlin> is it okay to send NULL as the second argument to gst_element_factory_make?
[18:30] <mathrick> mxpxpod: hmm, give one good reason why one would do something disgusting like that?
[18:30] <mxpxpod> mathrick: to graduate
[18:31] <mathrick> mxpxpod: every teacher trying to force VB should be lynched immediately
[18:31] <mxpxpod> mathrick: I agree
[18:31] <mathrick> mxpxpod: get some normal guys from your year, and just do it
[18:31] <mathrick> mxpxpod: noone will miss him probably
[18:31] <mxpxpod> or atleast the advisory board for this community college
[18:32] <taaz> jdahlin: isn't that how auto-naming works?  (i say without looking...)
[18:33] <jdahlin> taaz: dunno
[18:34] <thomasvs> jdahlin: you don't use devhelp ?
[18:34] <jdahlin> thomasvs: not really
[18:34] <mathrick> mxpxpod: anyway, why didn't you oppose? I know many sick ideas wrt education, but this is just too much
[18:35] <jdahlin> thomasvs: or rather, I don't have the documentation built
[18:35] <mxpxpod> mathrick: I dunno... I'm so sick of this program, I'll do just about anything to graduate
[18:35] <thomasvs> jdahlin: well, the latest version of the docs are always online :)
[18:35] <mathrick> Create a new element of the type defined by the given element factory. If name is NULL, then the element will receive a guaranteed unique name, consisting of the element factory name and a number. If name is given, it will be given the name supplied.
[18:35] <mxpxpod> mathrick: why do you think I'm doing open source projects? because this program is a joke and I'll never get anywhere with the degree I get here
[18:36] <mathrick> mxpxpod: ugh, I pity ya pal, I really pity you :\
[18:36] <mxpxpod> mathrick: I keep my sanity by doing open source projects
[18:36] <mxpxpod> man, they the fuck would M$ give you 18 million ways to do something and not tell you which is the best
[18:36] <mathrick> mxpxpod: but question comes to mind, why not quit if it's pointless anyway?
[18:36] <mxpxpod> oh wait
[18:36] <mxpxpod> because M$ is retarded
[18:37] <mxpxpod> mathrick: because I started it, I might as well finish it... I don't want to be a quitter
[18:37] <mxpxpod> ok, I'm done working on that crap
[18:37] <mxpxpod> time to go do some real programming
[18:37] <mxpxpod> bbiab after I move to the cyber cafe
[18:38] mxpxpod ([email protected]) left irc: "Leaving"
[18:40] <thomasvs> jdahlin: Create a new element of the type defined by the given element factory. If name is NULL, then the element will receive a guaranteed unique name, consisting of the element factory name and a number. If name is given, it will be given the name supplied.
[18:43] <jdahlin> thomasvs: thanks
[18:43] mxpxpod ([email protected]) joined #gstreamer.
[18:44] <jong> taaz: could you please make debs for gstmm once it gets updated? It would so cool :)
[18:46] <mathrick> mxpxpod: nah, i don't like your philosophy. You finish it and then they can brag about how many graduates they have
[18:50] <mathrick> mxpxpod: I hope you don't pay for it, do ya?
[18:52] mxpxpod ([email protected]) left irc: Remote closed the connection
[18:53] mxpxpod ([email protected]) joined #gstreamer.
[18:57] <mathrick> mxpxpod: finally here? ;)
[18:57] <mxpxpod> mathrick: yeah
[18:57] <mxpxpod> something was giving me major lag :(
[18:57] <mathrick> mxpxpod: I suppose you didn't get my previous msgs?
[18:57] <mxpxpod> mathrick: which ones?
[18:57] <mathrick>  mxpxpod: nah, i don't like your philosophy. You finish it and then they can brag about how many graduates they have
[18:57] <mathrick> mxpxpod: I hope you don't pay for it, do ya?
[18:58] <mathrick> mxpxpod: these ones
[18:58] <mxpxpod> mathrick: oh, right
[18:58] <mxpxpod> I got em
[18:58] <mxpxpod> did you get my response?
[18:59] <mathrick> mxpxpod: nope
[18:59] <mxpxpod> mathrick: I said, yeah, my philosophy is probably flawed, yes I pay for the classes
[19:00] <mathrick> mxpxpod: uhmm, you really should find some better use for your money, fund a bounty if you don't know what to do with cash ;)
[19:00] <mxpxpod> hey! I'm mentioned on planet.gnome.org
[19:01] Action: mxpxpod points to murray cumming's entry
[19:01] kwm ([email protected]) joined #gstreamer.
[19:02] <mathrick> mxpxpod: nice :)
[19:07] yippi ([email protected]) left irc: Remote closed the connection
[19:14] apoc ([email protected]) left irc: Nick collision from services.
[19:14] apoc ([email protected]) joined #gstreamer.
[19:14] apoc ([email protected]) left irc: Nick collision from services.
[19:15] apoc_ ([email protected]) joined #gstreamer.
[19:28] markey ([email protected]) left irc: "leaving"
[19:34] smoke ([email protected]) got netsplit.
[19:40] smoke ([email protected]) got lost in the net-split.
[19:40] jong ([email protected]) left irc: "bye everybody"
[19:43] <KoRnouille> is there a help somewhere to understand better how mpegdemux and mpeg2dec can connect video & audio streams between each other ?
[19:44] jdahlin ([email protected]) left irc: "Klienten avslutas"
[19:45] Uraeus ([email protected]) joined #gstreamer.
[19:49] mxpxpod ([email protected]) left irc: ""If ever time should come, when vain and aspiring men shall possess the highest seats in Government, our country will stand i
[19:51] markey ([email protected]) joined #gstreamer.
[19:51] mxpxpod ([email protected]) joined #gstreamer.
[19:52] yippi ([email protected]) joined #gstreamer.
[19:54] ChrisHJW ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[20:01] sublett ([email protected]) joined #gstreamer.
[20:02] markey ([email protected]) left irc: "leaving"
[20:16] <taaz> eh... jong left.  i wonder if i should mention that i have a strong dislike for all that is related to c++.  getting me to make debs for the bindings could prove difficult ;)
[20:18] crazney_ ([email protected]) joined #gstreamer.
[20:28] thomasvs ([email protected]) left irc: Read error: 113 (No route to host)
[20:29] <ds-work> taaz: you should get over your dislike
[20:33] crazney ([email protected]) left irc: Connection timed out
[20:34] markey ([email protected]) joined #gstreamer.
[20:35] <KoRnouille> how do I connect xvimagesink to my window ?
[20:35] markey ([email protected]) left irc: Remote closed the connection
[20:36] markey ([email protected]) joined #gstreamer.
[20:36] <bilboed_school> give it the xid of your window
[20:36] <KoRnouille> okay, I'm bad at that yet.
[20:36] <KoRnouille> maybe give me a syntax, and I'll go out and see what it does. But I need a starting point
[20:37] wheels ([email protected]) joined #gstreamer.
[20:37] <KoRnouille> i've seen stuff about sockets and so...
[20:37] <bilboed_school> I'm checking right now, I haven't used it yet :)
[20:37] <KoRnouille> :) okay
[20:39] Shoragan ([email protected]) left irc: "Leaving"
[20:40] <bilboed_school> it has to do with the xwindowlistener
[20:42] <mathrick> gst_x_overlay_set_xwindow_id (overlay, xid);
[20:43] Shoragan ([email protected]) joined #gstreamer.
[20:43] <KoRnouille> that's C I guess?
[20:43] <mathrick> you should really learn to search mailing list first :)
[20:43] <mathrick> KoRnouille: yep
[20:43] <bilboed_school> :)
[20:43] thomasvs ([email protected]) joined #gstreamer.
[20:43] <KoRnouille> thanks mathrick
[20:43] kwm ([email protected]) left irc: "Leaving"
[21:06] nemoglobine ([email protected]) left irc: Remote closed the connection
[21:23] sublett ([email protected]) left irc: "I like food, food is good!"
[21:25] ChrisHJW ([email protected]) joined #gstreamer.
[21:25] <KoRnouille> what's the syntax for self.demux.connect ?
[21:25] <KoRnouille> what am I supposed to put in connect() ? (python)
[21:33] <taaz> foo.connect(signame, function[, args])
[21:34] <taaz> function can be a method like self.dosomething
[21:34] <KoRnouille> ok
[21:34] <KoRnouille> "signame" ?
[21:35] <taaz> like "eos" or "handoff" or whatever
[21:35] <KoRnouille> :/
[21:35] <taaz> it's just the GObject signal system.  read the C docs if you want
[21:38] <KoRnouille> in my case, I need the demuxer (mpegdemux) to connect it's src pad to the mpeg2dec sinkpad.
[21:38] <KoRnouille> right ?
[21:38] <taaz> uh, you mean to use link()
[21:39] <KoRnouille> well, Im a little lost I guess
[21:39] <KoRnouille> okay, I guess it's link
[21:39] <KoRnouille> so before, I need to know what pad mpegdemux creates, right ? (because it dynamic)
[21:40] jcsston ([email protected]) joined #gstreamer.
[21:48] <ds-work> thomasvs: could you add me as a dev on autostars?
[21:49] <thomasvs> ds-work: sure
[21:49] Company ([email protected]) joined #gstreamer.
[21:51] <Misirlou> Company: GUTEN TAG
[21:52] <thomasvs> ds-work: done
[21:52] <thomasvs> man, do I ever hate docbook tools
[21:52] <Company> Misirlou: going german now?
[21:52] <Company> or:
[21:52] <Company> Misirlou: going german now??????????????????????????????
[21:52] <Uraeus> thomasvs: refering to the bug report from Alex? :)
[21:54] <thomasvs> Uraeus: no, huh ?
[21:54] <thomasvs> Uraeus: I'm referring to the fact that I hate docbook tools
[21:55] <Uraeus> thomasvs: ah ok :) It is just that Alex filed a bug report about switching some doc m4 macros or similar in gstreamer
[21:56] <thomasvs> Uraeus: probably the bug asking to make our gtk-doc build the same as in glib/gtk
[21:56] <thomasvs> which I'm not yet sure of, but anyway
[21:56] <Misirlou> Company: I only know "guten tag," and "hallo," which really isn't even German.
[21:56] <thomasvs> WHY can't I just run docbook2html on an xml file and have it generate proper output
[21:56] <thomasvs> or xmlto
[21:56] <thomasvs> either one of the two, working reliably on more than one distro, would be heaven for me
[21:57] <Misirlou> thomasvs: Because Normal Walsh is an evil horse.
[21:57] <taaz> when you figure it out, let me know
[21:57] <Company> yeah, me too
[21:57] <taaz> nw is cool
[21:57] <thomasvs> it must be the tools
[21:57] <Company> i'll update my LFS box then ;)
[21:57] <thomasvs> but they are so obtuse
[21:57] <thomasvs> the error messages are completely nonrelated to the problem I would bet
[21:58] <thomasvs> I should make a simple tarball with a oneline docbook/xml file, and have it run on as many platforms as possible, and then send the success rate to all the people involved
[21:58] <thomasvs> and then I'd say "fix this for me damnit"
[21:58] <thomasvs> and then I'd stop crying
[21:58] <taaz> what problems are you having?
[21:59] <thomasvs> docbook2html faq.xml on rh9 doesn't work
[21:59] <thomasvs> -bash-2.05b$ docbook2html faq.xml
[21:59] <thomasvs> Using catalogs: /etc/sgml/xml-docbook-4.2-1.0-17.cat
[21:59] <thomasvs> Using stylesheet: /usr/share/sgml/docbook/utils-0.6.12/docbook-utils.dsl#html
[21:59] <thomasvs> Working on: /usr/src/rpm/BUILD/gstreamer-0.7.5/docs/faq/test/faq.xml
[21:59] <thomasvs> jade:/usr/src/rpm/BUILD/gstreamer-0.7.5/docs/faq/test/faq.xml:4:0:E: DTD did not contain element declaration for document type name
[21:59] <thomasvs> and then a whole slew of errors, one for each tag in there
[21:59] <thomasvs> it's using the correct catalog and stylesheet, afaict
[22:00] <thomasvs> if I change it to 4.1.2 instead of 4.2, I get the same, plus a whole bunch of iso entity errors thrown in for free
[22:00] <thomasvs> now, xmlto works mostly too, but is ten times slower, AND has a tendency to crash horribly and leave you in a critical tex error
[22:01] <Uraeus> thomasvs: that is the same errors I get on Fedora 2
[22:01] <taaz> fwiw, i still can't build gst docs on my debian box :(
[22:01] <taaz> probably related
[22:02] <thomasvs> Uraeus: I don't know, I've never seen the errors you get :)
[22:03] <Uraeus> thomasvs: well I see the one you get now at least :) (and you did see it the other day when I asked about it :)
[22:05] <thomasvs> Uraeus: I saw only part of it, you never sent me the complete build log I asked
[22:05] <thomasvs> and I don't autosave all my irc chats
[22:06] <Uraeus> I am not complaining :)
[22:06] <thomasvs> anyway, time to blow off steam
[22:06] <thomasvs> and if xmlto works on more platforms, I'm switching back to it, whether or not it's slow
[22:08] <KoRnouille> is there a way with python to see what's inside a thread ? I've added elements to the thread, now I'd like to see it.
[22:08] ibsteveog ([email protected]) joined #gstreamer.
[22:11] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[22:13] Action: Company fears people don't read his mails :o
[22:14] <Company> ds-work: should boxed GstData subtypes copy or ref on g_boxed_copy?
[22:15] <bitshifter> Company: just out of curiosity, why would one _not_ want to just ref on g_boxed_copy?
[22:16] <Company> because he wants a copy that he can modify
[22:16] <ds-work> Company: ref
[22:16] <Company> k
[22:16] <Company> that's what i did
[22:16] <ds-work> we really need to implement read/write locks on buffers
[22:17] <Company> why would we want that?
[22:17] <ds-work> because we play fast and loose with the idea that buffers with refcount > 1 are read-only
[22:17] <Company> they are read-only
[22:17] <ds-work> writeablity is only partly related to refcount
[22:18] bilboed_school ([email protected]) left irc: "Leaving"
[22:18] <ds-work> think of the case of downstream-allocated buffers
[22:18] <Company> and?
[22:18] <ds-work> ximagesink creates a buffer (refcount 1), and passes it upstream (refcount == 2)
[22:18] <Company> why would ximagesink keep a reference?
[22:19] <ds-work> actually, in that case, it wouldn't
[22:19] <ds-work> hrm
[22:19] <Company> you only keep a reference to a buffer if you need the data inside the buffer
[22:19] <Company> unmodified
[22:20] <ds-work> what about subbuffers?
[22:20] <Company> subbuffers need the data unmodified, too
[22:20] <pmfp> why do I get a segmentation violation on this? http://dev.pmfp.net/tmp/gst-support.c
[22:20] <pmfp> it's almost the same as in the docs
[22:20] <ds-work> Company: only if they're being passed downstream
[22:21] <Company> ds-work: yeah, if you know what you're doing, then not
[22:22] smoke ([email protected]) joined #gstreamer.
[22:22] Action: mathrick shoots gdk-pixbuf loaders
[22:22] <Company> but the refcount of a buffer really is all about the data
[22:22] Action: Company reloads for mathrick
[22:23] <Company> and we wouldn't want to use locks on buffers now, would we?
[22:23] <bitshifter> pmfp: because you forgot a   test_item = g_new0(struct file_entry, 1);  after line 65  I'd say
[22:26] <pmfp> bitshifter: that didn't make it work... if you're sure that it has something to do with that though, I'd be glad to research it if you say which part of the docs
[22:27] <bitshifter> pmfp: you should use gdb and/or valgrind first to find out where exactly it segfaults IMHO
[22:27] <pmfp> bitshifter: roger
[22:27] <Company> but you definitely need to allocate memory for the test_item :)
[22:27] <pmfp> Company: heh :)
[22:28] <bitshifter> pmfp: oh, the other thing is that you do NOT want to pass &test_item, and in your function you want a struct file_entry *item_source
[22:28] <bitshifter> pmfp: and you only want to use  item_source->foo then, not  *item_source->foo
[22:29] <bitshifter> pmfp: (unless I'm missing something)
[22:29] <pmfp> hmmm, I think I made a booboo
[22:29] <Company> you want to get used to pointers more i guess :)
[22:29] <pmfp> aw, crap, sorry guys, we're looking at different sources
[22:30] jcsston ([email protected]) joined #gstreamer.
[22:30] <pmfp> bitshifter: could you check the same place again... switched the source
[22:31] <pmfp> Company: I'm really just playing around so far... not an expert, but I got somewhat of an understanding :)
[22:32] <pmfp> bitshifter: seg faults when gst_main () is supposed to run
[22:34] Action: mathrick wonders what a genius has chosen Ctrl+G as "next match" in mozilla's typeahead find when modelling that feature _after Emacs_
[22:35] <Company> what do you prefer? F3?
[22:35] <mathrick> Company: no, Ctrl+S actually :)
[22:35] <Company> S as in nekSt?
[22:36] <mathrick> Company: but anyway, Ctrl+G is *worst* one could ever choose
[22:36] <mathrick> Company: no, s as in C-s in Emacs :)
[22:36] <Company> dunno
[22:36] <Company> i'm confused why typeahead needs no ctrl, but next does
[22:36] <mathrick> Company: Ctrl+G means interrupt in Emacs, and is used to _end_ incremental seacrh
[22:37] <Company> i don't use emacs
[22:37] <mathrick> Company: because there isn't good way to bind it w/o ctrl ?
[22:37] <Company> dunno
[22:37] <Company> right arrow?
[22:37] KoRnouille ([email protected]) left irc: Client Quit
[22:37] <mathrick> Company: I know, but it's advertised as "get your emacsish incremental search in mozilla"
[22:37] <Company> oh
[22:38] <mathrick> Company: no, right arrow's already used
[22:38] <ds-work> mathrick: ctrl-G is what mac and windows have traditionally used
[22:38] <Company> windows used ctrl-g ?
[22:38] <ds-work> iirc, yes
[22:38] <mathrick> ds-work: windows?
[22:38] <Company> windows uses F3
[22:38] <Company> or used at least
[22:38] <ds-work> in addition to ctrl-g
[22:39] <mathrick> ds-work: what windowish programs have incremental search, if I may ask?
[22:39] <ds-work> mathrick: Word
[22:39] <mathrick> ds-work: really? full incremental functionality?
[22:40] <ds-work> "find next"
[22:40] <mathrick> ds-work: not to mention that there *isn't* any single, consistent "Find" keystroke in windows, and that drives me crazy when I must use windoze
[22:41] <mathrick> ds-work: ahh, that's not incremental search, you know
[22:41] <mathrick> ds-work: it's just dumb find next, IS is much more useful
[22:42] <bitshifter> pmfp: sorry, I have no idea how threads work in gst. however, is it possible that you either need to put your GstThread into a pipeline or give it a scheduler or so (just a random guess really)?
[22:42] <ds-work> mathrick: so how is ctrl-G used for 'next match
[22:42] <ds-work>  different than 'next match'?
[22:43] <mathrick> ds-work: IS means you can alter your search term between matches
[22:43] <pmfp> bitshifter: I'll look into it
[22:43] Action: bilboed is back (gone 06:40:10)
[22:43] <ds-work> mathrick: and?  it still means 'next match'
[22:43] <mathrick> ds-work: no
[22:44] <Company> it's like ctrlk-f in word and not closing the window between matches :o
[22:45] <mathrick> ds-work: next match is next match, but you can't (easily) say in Word "now, let's search for something little bit specific" in the middle of search
[22:45] <pmfp> bitshifter: I tried with a direct copy of the example from the documents, just switching to a file sink... same problem
[22:45] <ds-work> mathrick: that's because Word is dumb
[22:45] <pmfp> though I'm looking at head-docs, just noticed... checking if 0.6.5 docs says anything differently
[22:45] <mathrick> Company: yep, it may kinda work similarly, but is incredibly painful if you have extensive search to do
[22:46] <mathrick> ds-work: that's one thing, but IS also needs slightly different UI
[22:46] <ds-work> mathrick: but ctrl-G has a (relatively) well known meaning of 'find next'.  So why would that be an inappropriate keystroke to mean 'find next'?
[22:47] <mathrick> ds-work: very relatively. I haven't ever heard of it, and I think I learned about 10 different keystrokes for that in various windoze apps
[22:48] <ds-work> that's windows's problem.  mac and other platforms _do_ use ctrl-g consistently
[22:48] <Company> considering i didn't know it existed until 5 minutes ago...
[22:48] <mathrick> ds-work: anyway, just why advertise it being emacsish when it actually clashes in worst possible way with emacs user expectations?
[22:48] <mathrick> Company: it's kinda cool
[22:49] <Company> ok, gst_{buffer|data}_free is gone
[22:49] <mathrick> Company: just type to search in links, and / and type to search whole text
[22:49] <ds-work> why did they make it emacsish at all?  emacs has almost as stupid key bindings as vi
[22:49] <mathrick> Company: in mozilla, that is, not in Emacs ;)
[22:49] <Company> mathrick: i knew that one
[22:49] <Company> mathrick: i didn't know ctrl-g
[22:50] <Company> and build.fluendo.com is the coolest thing since quite a long time
[22:50] <mathrick> Company: ahh, so did I, but finalyy got fed up with being unable to match next, and googled for it :)
[22:50] <mathrick> ds-work: no, it's just matter of preference, and getting used
[22:50] <ds-work> funny, I looked at Edit/Find Next, and saw that the binding was ctrl-G
[22:51] <mathrick> ds-work: and Emacs has had its bindings _long_ before M$ ever started considering windows
[22:51] <taaz> NeXT used command-g, i assume macs use squiggle-g?
[22:51] <Company> thomasvs: build.fluendo.com rules, that's the greatest thing since a long time
[22:52] <Company> eta 700 seconds...
[22:53] <mathrick> is it only me, or that table doesn't have cell edges drawn? kinda unreadable it gets
[22:54] Nick change: kh_zZz -> khalek
[22:54] ds ([email protected]) left irc: Read error: 110 (Connection timed out)
[23:09] danb_ ([email protected]) joined #gstreamer.
[23:10] danb ([email protected]) left irc: "Reconnecting"
[23:13] Nick change: danb_ -> danb
[23:18] <mathrick> damn, I'm unable to view 600K GIF because it causes every app I try allocates 140+ MB of mem and ends up killed by OOM :(
[23:19] <Company> lol
[23:19] <Company> url?
[23:20] <mathrick> it's moments like that when it shows linux apps are still not entirely ready for desktop use, I'm actually pretty sure ACDSee would just chew it and display immediately
[23:20] <Company> url? :p
[23:20] <Company> the problem is that noone ever had that problem
[23:20] <Company> (and that animations in gdk-pixbuf suck like nothing else)
[23:21] <Company> you could probably get it to display something by slowly feading the loader data...
[23:21] <mathrick> Company: nope, it's in some weird tarball with anime art :), i can give you bt link or just file itself
[23:21] <Company> mail it to [email protected] please
[23:21] <ds-work> !
[23:21] Action: ds-work wants an @gnome.org address
[23:22] <mathrick> Company: k
[23:22] <Company> ds-work: every foundation member gets one by default
[23:22] Action: pmfp wants [email protected]
[23:22] <pmfp> ;)
[23:22] <Company> nah
[23:22] <Company> [email protected]
[23:22] <Company> that one might be cool
[23:25] <ds-work> Company: isn't whitehouse.com porn?
[23:26] <Company> ds-work: yeah, that's why :)
[23:27] Action: mathrick wonders if in-kernel OOM killer was good idea
[23:27] <mathrick> I'm getting more and more convinced it was rather dumbass to add that 'feature' to linux kernel
[23:27] <Company> 2.6 doesn't have that
[23:28] <mathrick> Company: yeah, I heard it's optional now, too bad 2.6 gives nice stacktrace on boot for me :\
[23:28] <mathrick> I think I should do something about it, but for now I'm sticking with 2.4
[23:29] <Company> your choice :p
[23:30] <mathrick> Company: ok, so what quick fixes you have for that? it's kinda hard to do anything with kernel coredump printed on your screen, ya know? ;)
[23:31] <Company> get rid of everything inside the kernel
[23:31] <Company> only compile the stuff you must
[23:31] <mathrick> Company: define everything
[23:32] <Company> everything you don't need to get to runlevel 2
[23:32] <mathrick> Company: I think I'm compiling only needed stuff, everything in modules anyway
[23:32] Action: mathrick should probably try with distro built kernel to have some measure
[23:32] ChrisHJW ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[23:33] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[23:33] markey ([email protected]) left irc: "leaving"
[23:34] <mxpxpod> mathrick: dude, guess what I did today
[23:35] <mxpxpod> mathrick: (it has nothing to do with gstreamer ;)
[23:38] <mathrick> mxpxpod: no clue, what was it?
[23:38] <wheels> mxpxpod: discovered the meaning of life?
[23:38] <mxpxpod> wheels: almost
[23:38] ChrisHJW ([email protected]) joined #gstreamer.
[23:38] <mxpxpod> mathrick: I wrote the STL interface for the new Gtk::Toolbar API for gtkmm
[23:39] <mathrick> wheels: nah, there is no such thing ;)
[23:40] <mathrick> mxpxpod: woah, you probably have too much time on your hands ;)
[23:40] <wheels> mathrick: other than 42, of course.
[23:40] <mxpxpod> mathrick: you're right
[23:40] <mxpxpod> :)
[23:40] <mxpxpod> ok, gotta go home
[23:40] Action: mathrick never got around learning STL enough to actually use it
[23:40] <mathrick> but I have some basics
[23:40] <mxpxpod> it's pretty cool
[23:40] <wheels> mxpxpod: Well, it's a start -- it has been said that C++ is the true way to enlightenment.  :-)
[23:40] <mxpxpod> ok, later
[23:40] mxpxpod ([email protected]) left irc: "*sigh* I don't want to come up with a quit message..."
[23:41] Action: wheels actually isn't a big STL fan though...
[23:41] <mathrick> Company: hrmpf, looks like there's some world-size cabal determined to not allow me to send you that gif :\
[23:41] <mathrick> Company: none of my accounts seem to work for it
[23:42] <mathrick> Company: do you have some ftp or something like that ?
[23:43] <Company> mathrick: nope
[23:43] <Company> only gst cvs ;)
[23:43] <mathrick> Company: hmm, wanna BT link?
[23:44] <taaz> ds-work: speaking of stl... why should i get over my c++ dislike?  i've managed to mostly avoid it for many years and consider that good fortune ;)
[23:45] <Company> i don't use BT
[23:45] <mathrick> Company: it's only 60MB, should be fast enough
[23:45] <mathrick> Company: hmm, that's too bad
[23:45] <wheels> taaz: I'm guessing you don't want my answer.  :-)
[23:46] <taaz> wheels: well... probably best not to turn this to a language war channel ;)  we can all agree to disagree.
[23:46] herzi ([email protected]) left irc: Read error: 113 (No route to host)
[23:47] <wheels> taaz: Well, I'm not enough of a true zealot to get into a war.  Suffice it to say that it's only ok to not like languages that you are familiar with.  :-)
[23:47] <ds-work> taaz: it depends if you are obstructing the use of GStreamer
[23:47] <wheels> taaz: (And I work with C about as much as C++)
[23:49] <taaz> i really wish Objective-C got more use.  when the py bindings are working maybe i'll start obj-c ones
[23:50] <taaz> though can probably go nuts and use python apps with pygtk+pyobjc+pygst
[23:51] ibsteveog ([email protected]) left irc: "quitting"
[23:56] Action: mathrick have never seen any Obj-C
[23:56] bilboed ([email protected]) left #gstreamer ("Client exiting").
[23:57] <wheels> mathrick: OS X is actually written in Obj-C IIRC
[23:57] <Uraeus> and OpenGroupware.org is Obj-C
[23:57] <khalek> not all of it
[23:58] <mathrick> wheels: yeah, I know, unfortunately I don't have its source handy ;)
[23:58] <wheels> mathrick: Well, parts of it are.  :-)
[00:00] --- Sat Feb 28 2004
[00:02] walters ([email protected]) joined #gstreamer.
[00:08] Nick change: bitshifter -> bitshifter|phone
[00:11] crazney_ ([email protected]) left #gstreamer ("Client exiting").
[00:13] <taaz> GNUstep would be an amazing piece of work if the effort that has gone into gnome and kde was even partially focused there.  oh well.  slow and steady progress...  i should help them out.
[00:14] <Uraeus> I used to wish for Gnustep when I started using Linux thinking it looks so cool. Today I think it looks like something from 20 years ago and the devs seem to like it that way
[00:15] <taaz> i'm torn.  i don't know what i wish for ;)
[00:15] <ds-work> the problem with GNUstep is that it settled into a single vision of how the desktop should be
[00:16] <taaz> it's partially a catch up game though.  which is unfortunate
[00:16] <thomasvs> ah
[00:16] <thomasvs> nothing beats some physical exercise
[00:16] <ds-work> instead of focussing on usability, and developing your vision around that
[00:16] <taaz> if they had started with, say, NeXT releasing all code GPLed or something then the platform would have taken off
[00:16] Action: thomasvs checks buildbot
[00:17] <taaz> thomasvs: how is one supposed to hack on buildbot?
[00:17] <thomasvs> looking good
[00:17] <taaz> ie, to add stuff like gst-python
[00:17] mxpxpod ([email protected]) joined #gstreamer.
[00:17] <thomasvs> taaz: in general, or on this one ?
[00:17] <Company> mh
[00:17] <thomasvs> taaz: well, we're going to add the other modules as soon as we're fairly sure this set is running sort of stable
[00:17] <thomasvs> it's been running well for a week I think, so it's getting time
[00:17] <taaz> the fluendo one
[00:18] <thomasvs> taaz: then to hook up johan's media test suite
[00:18] <Company> gdb goes bonkers if you pass functions pointers to functions
[00:18] <taaz> thomasvs: ok, no prob.  i just want to see it be even more cool ;)
[00:18] <thomasvs> taaz: as for the fluendo buildbot version, johan is keeping patches and sending them upstream
[00:18] <thomasvs> taaz: if you want to hack on it too locally or smth, we can do that
[00:19] <thomasvs> taaz: but for now I guess it's simpler to let johan hack on it, he's good at python :)
[00:19] Action: Uraeus wonders if GNOME and KDE would even have come to pass if NextStep had been released as free software
[00:19] <thomasvs> goes a lot faster when I ask him something then when I try doing it myself
[00:21] <taaz> Uraeus: writing an article about that has been on my todo list for a long time ;)
[00:21] <thomasvs> taaz: I bet you have a big todo list :)
[00:21] <thomasvs> <cough>vdv</cough>
[00:21] <taaz> thomasvs: as does everyone else, of course
[00:21] <taaz> ahem...
[00:22] <taaz> yeah yeah... i'll work on that at some point
[00:22] <taaz> N problems exist before it'll work well though
[00:22] <thomasvs> taaz: well, make a list of them, get others to hack on it :)
[00:22] <thomasvs> taaz: I already set johan lose on the bindings
[00:23] <thomasvs> taaz: you think he's doing a good job on it ?
[00:23] <taaz> yeah ;)
[00:23] <taaz> hmm... i dunno
[00:23] <taaz> i'm not sure i agree with taking an axe to 0.6 support
[00:24] <thomasvs> hm, me neither
[00:24] <thomasvs> but I guess if you care we should either branch now, or someone should do the work to keep 0.6 support
[00:24] <thomasvs> we'll have to figure out multiple versions anyway
[00:24] <taaz> speaking of branches...
[00:25] <taaz> i've got half an idea of using arch to do gst-python dev
[00:25] Action: wheels notes that it's probably good practice for the future where there will likely be branches that are used by the desktops / distros that have to stay stable for a while.
[00:26] <thomasvs> taaz: uh oh :)
[00:26] <taaz> i'm annoyed that cvs branching confuses me so much.  bk, svn, arch... all have been easy to figure out.
[00:29] <thomasvs> well, it's not that bad
[00:29] <thomasvs> just need a cheat sheet
[00:29] <thomasvs> docs/random/release :)
[00:29] <taaz> yeah, but i shouldn't need one.
[00:32] <Uraeus> ok, night
[00:32] <Company> GLib-ERROR **: gmem.c:140: failed to allocate 1000000 bytesGLib-ERROR **: gmem.c:140: failed to allocate 1000000 bytes
[00:32] Uraeus ([email protected]) left irc: "Client exiting"
[00:32] <Company> weenie :/
[00:34] <mathrick> k, going to sleep now
[00:34] <mathrick> nite
[00:34] <Company> hum, why can'T glib alloc 4MB for me?
[00:34] Nick change: mathrick -> mathrick|sleep
[00:35] yippi ([email protected]) left irc: "Client exiting"
[00:38] <ds-work> thomasvs: does your build system work by polling, or do you push stuff out the build daemons?
[00:39] <taaz> email
[00:39] <thomasvs> ds-work: there's a server daemon, and clients connect to the daemon and wait for work
[00:39] <thomasvs> taaz: no, the email goes to the server
[00:39] <thomasvs> then the server tells the slaves to build
[00:39] <thomasvs> over the open link
[00:39] <ds-work> thomasvs: so I could set up a Solaris build client behind a firewall?
[00:40] <thomasvs> ds-work: yep, that's the point
[00:40] <thomasvs> ds-work: that's exactly what we wanted someone to do
[00:40] <ds-work> thomasvs: where's the code?
[00:40] <thomasvs> ds-work: first you need to install Twisted
[00:40] <thomasvs> (a python protocol framework)
[00:40] <thomasvs> and then our buildbot code
[00:40] <thomasvs> we'll tar it up monday
[00:40] <taaz> thomasvs: you should really have a buildbot.sf.net url on that page somewhere
[00:40] <taaz> spread the love
[00:40] <thomasvs> it might take some tweaking to work for solaris
[00:41] <thomasvs> taaz: yeah, we will
[00:41] <thomasvs> taaz: as soon as we make it official I guess
[00:41] <thomasvs> we still need to do sort of a press release anyway
[00:41] <ds-work> Solaris is painful
[00:41] <thomasvs> how do you like the favico on that site btw ?
[00:41] <ds-work> I really need apt-get in order to stay sane
[00:41] bilboed ([email protected]) joined #gstreamer.
[00:41] <thomasvs> ds-work: yeah, solaris is painful that way
[00:41] <thomasvs> ds-work: someone was working on a port btw, using rpm
[00:42] <thomasvs> ds-work: and I think there's a sun channel on gimpnet
[00:42] <taaz> icon is ok i guess.  kind of an exploding slugs idea? ;)
[00:42] <thomasvs> they were helpful when I had one
[00:42] <thomasvs> taaz: no, a cartwheel
[00:42] <thomasvs> taaz: the main site makes it slightly clearer
[00:43] <taaz> oh hey, you have a big logo now
[00:43] <taaz> that's kinda cool
[00:43] <thomasvs> it'll be nicer when we animate it
[00:44] <taaz> hmm  less space above the co name than below
[00:44] <thomasvs> so, do these lawsuit outcomes mean it's legal to distribute css ?
[00:45] <taaz> does it make sense to have buildbot connect to machines over the net?  or is that asking for trouble?
[00:45] <thomasvs> what do you mean ?
[00:45] <thomasvs> from server to slaves ?
[00:46] <taaz> like adding my box so debian gets some respect ;)
[00:46] Action: bilboed is away: I'm busy
[00:46] <taaz> yeah, just build slaves
[00:46] <walters> thomasvs: btw..this gconf versioning strikes me as a bit of overkill
[00:47] <walters> and shouldn't it be gstreamer-0.8?
[00:48] <walters> although i guess now that gconf supports uninstallation it's ok
[00:48] ChriHJW ([email protected]) joined #gstreamer.
[00:49] <thomasvs> walters: yeah, it's overkill, but it's needed for parallel installability
[00:49] <Company> why can such a simple program fail to malloc 4MB?
[00:49] <thomasvs> walters: I wish I could find a better way around it
[00:49] <thomasvs> walters: good that gconf supports uninstall now, didn't know that
[00:50] <foser> still it's odd.. you probably want to keep settings from 0.6 -> 0.8
[00:50] <thomasvs> foser: a lot of elements have changed completely, as well as the parse syntax
[00:50] <thomasvs> foser: ie, filtered links didn't exist in the 0.6 syntax
[00:51] <walters> ideally, gconf would support some sort of symlink
[00:51] <Company> ha, infinite loop :o
[00:51] <Company> i suck
[00:51] <walters> so we could have the gstreamer-0.7/audiosink "point" to gstreamer/audiosink, unless you set gstreamer-0.7/audiosink directly
[00:52] <thomasvs> walters: yeah, that'd be nice
[00:52] ChriHJW_log ([email protected]) joined #gstreamer.
[00:52] <Company> gconf entries are forwards compatible adccording to gconf docs
[00:52] <thomasvs> could be sort of possible already on install, maybe
[00:55] <Company> having multiple gconf entries is a bad idea
[00:55] <Company> because it means it erases your configuration when you update from 0.6 to 0.8
[00:56] <Company> and it means you can't use gstreamer-properties with rb if the versions don't use the sam,e gst version
[00:57] <walters> yep
[00:59] <Company> the only thing you could do (which i strongly advise against) is using 2 gconf keys - first check gst-0.7/audiosink and if that isn't set check gstreamer/audiosink
[01:00] <Company> and unset all gstreamer-0.7 things by default
[01:00] <Company> and let all apps only operate on the gstreamer/* keys
[01:04] ChrisHJW ([email protected]) left irc: Connection timed out
[01:06] ChrisHJW_log ([email protected]) left irc: Read error: 110 (Connection timed out)
[01:09] <Company> chunklib seriously starts to rule more than when i had the idea
[01:09] foser ([email protected]) left irc: "[ I want to believe ]"
[01:12] <iain> Company: audioconvert goes though caps nego for times for my pipeline, the first 3 times it gets the right number of channels (2), the final time it gets 1, which is wrong
[01:12] <iain> Company: any idea where I should look to fix it?
[01:13] <walters> chunklib?
[01:13] Action: desrt listens to the little mermaid
[01:15] <taaz> ds-work: cicl script an excellent idea
[01:15] Action: taaz listens to bass drive, again
[01:16] <desrt> taaz; digitally imported.
[01:16] <taaz> walters: rb should have an auto-play option so i don't have to bother clicking on bass drive when i start up ;)
[01:16] <walters> taaz: i just implemented a lot of session-saving stuff last night
[01:17] <walters> taaz: restoring the selection shouldn't be too bad...
[01:17] <walters> right now it restores the selected source and the search text
[01:18] <Company> iain: nope, no idea
[01:18] <iain> Company: :(
[01:18] <Company> walters: chunklib was an idea i had about how to parse chunk based data streams (.AVI, .MOV, .ASF)
[01:18] <walters> ah
[01:18] <walters> Company: it had better have a function chunklib_blow
[01:19] <Company> which calls g_blow_chunks?
[01:19] <walters> yes :)
[01:20] <iain> maybe I should email wingo, I vaguely remember him saying something about audioscale needing a fixate function of something for int2float translation
[01:20] Action: ds-work prefers liboil's "oil_spill()" function
[01:21] <Company> what does that do?
[01:21] Action: desrt assumes by the relative unfunniness of that that that function actually exists
[01:21] <ds-work> dumps a bunch of internal state to stdout
[01:21] Action: desrt should try to use 3 'that's in a row more often
[01:22] <Company> go desrt!
[01:23] <walters> desrt: that was impressive.
[01:27] <desrt> hey... random C question
[01:27] <desrt> you know how you can fill structs by name like... = { .foo = 3;  .str = "bar"; };
[01:27] <walters> in C99, yeah
[01:27] <desrt> is something similar available for static arrays?
[01:28] <walters> sure
[01:28] <desrt> [] = { 4: 3;  5: 2; };
[01:28] <desrt> like that
[01:28] <walters> ,
[01:28] <desrt> 4:3,5:2}?
[01:28] <desrt> awesome :)
[01:28] <walters> ah
[01:29] <walters> you have to use nested braces for each struct iirc
[01:30] <desrt> ok... say i wanted to create an array of ints and i wanted it to contain { 0, 0, 0, 0, 3, 0, 4, 0, 0 };
[01:30] <desrt> so i say int array[9] = { what? }
[01:31] <walters> just like you gave
[01:31] <desrt> ok.  cool.
[01:31] <desrt> the nested braces statement confused me :)
[01:32] <walters> that was for structs
[01:32] <walters> desrt: like in widgets/rb-query-creator.c if you happen to have the rb src handy
[01:35] <desrt> hmm
[01:35] <desrt> syntax error at :
[01:35] Action: desrt wonders of c99 is on
[01:36] <desrt> that's an array of structs...
[01:39] <desrt> : still breaks it with -std=gnu99
[01:39] <desrt> hmm
[01:39] Action: desrt googles
[01:39] m_wheels ([email protected]) joined #gstreamer.
[01:39] <walters> desrt: can you paste?
[01:39] <desrt> er
[01:39] <desrt> it's on my laptop
[01:39] <desrt> lemme hop on from there
[01:40] <desrt> ok
[01:40] <desrt> evdev.c:43: parse error before ':' token
[01:40] <desrt> static const KeySym evdev_to_x[256] = {
[01:40] <desrt>         KEY_ESC:        XK_Escape,
[01:40] <desrt> 43 is the 2nd line, obviously
[01:40] <desrt> KEY_ESC and XK_Escape are both constants #defined in headers
[01:42] wheels ([email protected]) left irc: Nick collision from services.
[01:42] Nick change: m_wheels -> wheels
[01:43] <desrt> i only want this if it can be evaluated at compiletime
[01:43] <desrt> otherwise i'll just write it out manually
[01:43] <desrt> and it has to be direct index... a table lookup for every keypress/release is too expensive
[01:45] <walters> desrt: hm, try . KEY_ESC = XK_Escape
[01:45] <desrt> invalid lvalue in assignment
[01:46] <desrt> hmm
[01:46] <walters> .KEY_ESC ?
[01:46] <desrt> [KEY_ESC] = ... appears to be working
[01:46] <desrt> tried .KEY_ESC first :)
[01:46] <desrt> ya
[01:46] <desrt> syntax is { [index0] = value0, [index1] = value1 };
[01:46] <desrt> thanks :)
[01:48] <Company> gnight
[01:48] Company ([email protected]) left irc: Remote closed the connection
[01:56] sub_pop ([email protected]) left irc: "Client exiting"
[02:00] Nick change: bitshifter|phone -> bitshifter
[02:10] ChriHJW ([email protected]) left irc: Read error: 60 (Operation timed out)
[02:14] sub_pop ([email protected]) joined #gstreamer.
[02:22] smoke ([email protected]) got netsplit.
[02:24] Nick change: Rotty -> rotty-beer
[02:28] smoke ([email protected]) got lost in the net-split.
[02:48] iain ([email protected]) left irc: "leaving"
[02:51] jcsston ([email protected]) joined #gstreamer.
[02:54] <walters> hm
[02:55] <walters> what does gst 0.7.5 require glib 2.3 for?
[02:55] <ds-work> it doesn't.  it's a bug
[02:55] <walters> and it seems to work ok without it...
[02:55] <walters> ok
[02:55] <ds-work> (which has been fixed)
[02:58] pmfp ([email protected]) left irc: "Leaving"


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click