IRC Logs

*******************************************************************
[03:00] <pippin> yes,,. and gegl will allow you to request a region of interest ROI in the output image,. and only do the (re)calculations needed to get the requested data,.. merging affine transforms,. etc.
[03:00] <pippin> which I think goes beyond the scope of gst
[03:01] <Company> dunno
[03:01] <Company> it requires support from the elements
[03:02] <jcsston> why does gst-register set the extern gboolean _gst_registry_auto_load to FALSE? is it required?
[03:03] <Company> i would betz that this variable controls if gst_init loads its registries...
[03:04] <Company> and i'd have won that bet :)
[03:04] <jcsston> heh :)
[03:04] <pippin> Company: I bet it might get interesting at gvadec at least,. to get people working on somewhat similar things talking with each other :)
[03:05] <jcsston> I get an Access Violation (under Win32) on that line so I'm having to skip it.
[03:06] <Company> jcsston: you probably need to fix some stuff with how windows exports variables in DLLs
[03:06] <Company> pippin: yeah
[03:07] <Company> pippin: as long is gegl is lgpl so i can grab all the code i like :)
[03:07] ChrisHJW ([email protected]) left irc: Read error: 110 (Connection timed out)
[03:08] <pippin> Company: my stuff is currently GPL,. but if any of it seems worth stealing,. other licensing agreements could probably be made,.
[03:10] Zeenix ([email protected]) left irc: Read error: 110 (Connection timed out)
[03:12] <Company> hm yeah, gegl works quite different from gstreamer
[03:13] <Company> gegl doesn't work in the time domain afaik
[03:13] <Company> afaics
[03:13] <pippin> nope
[03:13] <jcsston> hmm, under Linux does each app get it's own loaded copy of a shared lib? because under Windows exporting variables from a DLL so that they can be changed is not vey clean.
[03:13] <Company> so you can access an element or data in it multiple times
[03:14] <Company> jcsston: yes, every app gets its own copy
[03:14] <pippin> but adding something on top of gegl that works in the time domain is very feasible,. (bauxite which I'm tinkering with, will do that)
[03:14] <jcsston> ouch, under Windows all apps share the same loaded copy of a .dll
[03:15] <Company> well, on linux they share until you write to it
[03:15] <Company> pippin: yeah, but you have no constraints on how often you can access some particular data
[03:16] <pippin> Company: do you have that in gst?
[03:16] <Company> pippin: well, when a buffer is send out, the buffer is out :)
[03:16] <Company> pippin: you need to seek to get it again - and that might not work
[03:18] <Company> pippin: gegl is pull-based i assume?
[03:18] <pippin> Company: yes
[03:18] <Company> yeah
[03:19] <Company> gstreamer is push-based or loop-based (elements decide what they want)
[03:19] <pippin> both makes sense in different contexts
[03:20] <mathrick> jcsston: .dll copy == .data section?
[03:20] <mathrick> jcsston: it doesn't make much sense
[03:21] <Company> pippin: yeah
[03:21] <Company> pippin: gst and gegl could share a lot of stuff though, but they are quite different in other contexts
[03:22] Action: mathrick thinks of marvelous subtiltles possible to achieve with gegl and gst together
[03:23] <mathrick> and not only subtitles
[03:23] <Company> mathrick: writing a gegl filter should be as hard as subclassing videofilter ;)
[03:23] <Company> writing a gegl compositor would be a bit harder ;)
[03:23] <mathrick> Company: as hard == very hard or very easy? :)
[03:24] <Company> mathrick: very easy, as long as you do 1 src => 1 sink
[03:24] <pippin> Company: 1 src => 1 sink isn't very interesting
[03:24] <mathrick> Company: writing gegl compositor should be as hard as subclassing <insert proper class name> ;)
[03:25] <mathrick> Company: at least, if they have gotten to having that template class already :)
[03:25] <pippin> GeglNode  (not element),.. probably GeglFilter
[03:25] <Company> that's why it's very easy :p
[03:25] <mathrick> pippin: it's filter, 1 -> 1 by definition
[03:26] <mathrick> Company: but you'll need at least 2 sinks
[03:26] <Company> yeah
[03:26] <Company> that's doable, too :)
[03:26] <pippin> mathrick: my definitions: http://schweden.mine.nu/bauxite/src/libgggl/Terminology.html
[03:27] <Company> that's nice
[03:27] <Company> it's close to gst :)
[03:27] Action: Company wonders if (de)muxers are (de)compositors
[03:28] <pippin> quite similar,. but this is only about image data
[03:28] <Company> is gegl supposed to be acyclic?
[03:28] <mathrick> Company: are timing and various other subtitle things doable with DParams? or are they not really expressible with that?
[03:28] <pippin> Company: cycles doesn't make sense for compositing
[03:28] <Company> mathrick: dparams is about controlling elements, not about controlling a stream
[03:29] Action: mathrick had Electronics exam today, his head is still full of feedback loops and such ;)
[03:29] <Company> pippin: that's indeed correct (I'm still doing timedomain thinking here ;))
[03:29] <mathrick> Company: hmm, right
[03:30] <mathrick> Company: is gst supposed to be acyclic? ;)
[03:30] <pippin> Company: the reason it is quite similar to gst, is that I took the terminology we were confusing ourselves with when discussing gegl,. and looked to gst where there was doubt,. e.g. pad vs port,.. I decided (and noone opposed),. should be pad
[03:31] <Company> mathrick: no, i don't think
[03:31] <Company> mathrick: though currently you would burn yourself a lot by making it cyclic
[03:32] <mathrick> eos must behave funny in cyclic pipelines :)
[03:33] <Company> mathrick: you might need special elements for that
[03:33] <Company> mathrick: that problem is not solved yet :)
[03:40] Action: ds-work has used acyclic pipelines in gstreamer
[03:41] <ds-work> at the very least, we should work on a common set of widgets for manipulating graphs, i.e., like gst-editor
[03:41] <ds-work> I need something like gegl, but for generic scientific data
[03:42] <mathrick> ds-work: it's not particularly spectacular to use acyclic ones, you know :)
[03:42] walters ([email protected]) joined #gstreamer.
[03:42] <ds-work> er, I meant cyclic
[03:42] <mathrick> ds-work: hmm, really?
[03:42] <mathrick> lo walters
[03:43] <walters> yo
[03:43] <mathrick> ds-work: what did you do?
[03:43] <ds-work> apparently wingo has done it, too
[03:44] <mathrick> ds-work: any example of such a pipeline?
[03:45] <ds-work> it required a few special elements
[03:45] <ds-work> basically changing timestamps
[03:48] <mathrick> ds-work: and where are such pipelines useful?
[03:49] <eventide> can anyone familiar with v4lsrc confirm if it works for webcams (as opposed to tv card)?  ("pipeline doesn't want to play": gst-launch v4lsrc ! sdlvideosink)
[03:49] <ds-work> mathrick: i was just testing it to see if it worked
[03:50] <mathrick> ds-work: ah, ic
[03:50] <ds-work> it did, sort of.  It seemed rather fragile
[03:50] <ds-work> but then, 0.6 seems rather fragile anyway
[03:51] <mathrick> eventide: there were some problems, ie. not all webcams did work due to missing some read methods, I don't know if these issues are solved already
[03:52] <Company> BBB claimed to have them solved, though he did only test with 1 webcam afaik
[03:53] <eventide> so it's supposed to work in theory...  I guess I should get 0.7 before I start hacking
[03:53] <mathrick> eventide: you talking about 0.6?
[03:53] <eventide> yes.. 0.7 isn't in debian :-P
[03:54] <mathrick> eventide: dude, then webcams are almost definitely no go
[03:54] <mathrick> eventide: it builds fine and easily, debian here
[03:54] <mathrick> eventide: and now, when xml catalog issue is (hopefully) solved, even docs should build :)
[03:55] Action: mathrick hasn't tested with docs yet
[03:55] <eventide> ok, thanks
[03:56] Action: Company fixes "easy" bugs
[03:57] <mathrick> eventide: and one thing, if you want to use webcams, grab HEAD
[03:57] <mathrick> eventide: AFAIR, 0.7.3 was way before serious v4l work was done
[03:57] <eventide> have you heard of anyone writing a gstsrc for digital still cams?
[03:58] <Company> na, it's just that 0.7.3 is just ~300 bugfixes behind cvs HEAD
[03:58] <mathrick> Company: details ;)
[03:59] <Company> # wc -l ChangeLog
[03:59] <Company>    2691 ChangeLog
[03:59] <Company> ;)
[03:59] <mathrick> eventide: what exactly would that thing do?
[04:00] <mathrick> Company: was CL set up right after 0.7.3? or some time later?
[04:00] <Company> mathrick: right after iirc
[04:00] <mathrick> hmm, wasn't it on fdo move?
[04:00] <Company> well, right after caps merge which happened right after fdo which happened right after 0.7.3 :)
[04:01] <mathrick> ah, right ;)
[04:01] <eventide> mathrick: grab frames from a digital still cam...  I want it for doing stop-motion.  I could probably write one, but I've a lot to learn about gstreamer first.  And then I have to port my stop-motion app to gstreamer
[04:01] <eventide> which is the plan
[04:03] <Company> man, that animal frightened me
[04:03] <Company> there's a catr sitting outside my window looking in
[04:03] <mathrick> Company: huh?
[04:04] <mathrick> heh, my cat is probably sleeping now :)
[04:04] <Company> and cats have big eyes at night
[04:05] Action: Company closes curtains
[04:06] <mathrick> Company: sissy :P
[04:06] <Company> yeah
[04:06] <Company> better than dieing from shock
[04:42] Action: Company fixed 4 or 5 bugs today, including 2 marked as blockers
[04:58] <eventide> Company: are you the same person from #gnome the other day when I asked about gtk widget sizes?
[05:02] <Company> eventide: yeah
[05:05] <Company> eventide: it's for a video widget - you want 1x size as default, but be able to shrink it
[05:05] <eventide> well, I found how gimp sets widgets to some size
[05:06] <eventide> and it's not exactly pretty, but it's simple
[05:06] <eventide> if you're still curious
[05:07] <eventide> or maybe you weren't curious yesterday, and I just misunderstood
[05:09] <Company> just tell me, might be useful :)
[05:09] pippin ([email protected]) left irc: Remote closed the connection
[05:11] <Company> though it probably won't work for widgets that you put into libs
[05:12] <eventide> take gtk_window_get_size (window, width, height); from those subtract widget.allocation_width, height; resize the window to the answer + desired dimensions;  yeah, I guess that only works for widgets you're already shown
[05:13] <eventide> and I suppose the widget in question should be the only one to expand with window resizing
[05:13] jcsston ([email protected]) left irc: Read error: 60 (Operation timed out)
[05:14] <eventide> I believe resizing the entire window with gtk_window_resize () is the only way to do it and still allow shrinking
[05:14] <Company> yeah
[05:14] <Company> that'd be interesting for gst-player though
[05:20] taf2 ([email protected]) left irc: Read error: 60 (Operation timed out)
[05:28] Action: ds-gromit wonders who forgot _stdint.h
[05:30] <Company> forgot?
[05:30] <ds-gromit> it doesn't exist
[05:30] jcsston ([email protected]) joined #gstreamer.
[05:30] <Company> configure nuilds it
[05:30] <Company> builds, too
[05:31] <ds-gromit> ah, ok
[05:35] herzi ([email protected]) left irc: Remote closed the connection
[05:36] herzi ([email protected]) joined #gstreamer.
[05:40] herzi ([email protected]) left irc: Remote closed the connection
[05:41] herzi ([email protected]) joined #gstreamer.
[05:44] Rotty ([email protected]) left irc: Remote closed the connection
[06:28] _gst_newt_ joined #gstreamer.
[06:34] ChrisHJW_log ([email protected]) left irc: Read error: 110 (Connection timed out)
[06:49] thaytoo ([email protected]) left irc: "leaving"
[07:47] kmaraas ([email protected]) left irc: "Leaving"
[07:47] mathrick ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[07:49] ChrisHJW ([email protected]) joined #gstreamer.
[07:49] mathrick ([email protected]) joined #gstreamer.
[08:03] hadley ([email protected]) left irc: "Leaving"
[08:09] <walters> mmm.  it's annoying that the .sgml files are kept in cvs
[08:09] ChrisHJW ([email protected]) got netsplit.
[08:11] ChrisHJW ([email protected]) returned to #gstreamer.
[08:13] ChrisHJW ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[08:13] <jcsston> I'm getting an error when I run bison, $ bison.exe gst/parse/grammar.y -v
[08:13] <jcsston> gst/parse/grammar.y contains 42 shift/reduce conflicts and 5 reduce/reduce conflicts.
[08:13] <jcsston>  Any ideas what is wrong?
[08:14] <jcsston> err, nm. that's not an error msg :/ it made the .c file
[08:20] ChrisHJW_log ([email protected]) joined #gstreamer.
[08:37] <ds-gromit> yeah, ignore that
[08:39] <thomasvs> walters: yeah, but we were told that is the correct way
[08:39] <thomasvs> walters: it pisses me off royally too
[08:40] jcsston ([email protected]) got netsplit.
[08:41] jcsston ([email protected]) returned to #gstreamer.
[08:42] somex1 ([email protected]) joined #gstreamer.
[08:42] <walters> correct?
[08:42] <walters> i don't see how that's possible.
[08:43] <walters> thomasvs: btw, could you ping me when you've got the gst release ready?  it'd be sweet to do a coordinated release
[08:43] <thomasvs> walters: sure.  I hope today is the day.
[08:44] <thomasvs> walters: I'm still bummed about mpeg video behaving weird though
[08:44] Action: thomasvs thumbnailed svq3 with gst last night
[08:44] <thomasvs> walters: what is the correct way to manage gtk-doc stuff then ?
[08:44] <walters> generate it at dist time?
[08:45] <walters> dist-hook: ?
[08:45] <thomasvs> oh, you mean only build/update docs at dist time ?
[08:45] <thomasvs> that would be nightmare too
[08:45] <thomasvs> then each time we want to release we find out the docs are screwed :)
[08:45] <walters> or do it at build time, if it's a cvs checkout
[08:45] <walters> test -d CVS
[08:46] <thomasvs> you are working from a cvs checkout right ?
[08:46] <walters> me? yeah
[08:46] <thomasvs> if it's not a cvs checkout there's no problem with the docs, that's the point iirc
[08:46] <walters> yes, but if you generate them at build time you don't check them into cvs
[08:48] <jcsston> is there an sample or example gstreamer registry file somewhere?
[08:49] <thomasvs> walters: so how do you keep track of changes then ?
[08:49] <thomasvs> ie, documenting the stuff that's not documented in the source files ?
[08:49] <walters> thomasvs: ah, a separate file?
[08:49] <walters> i haven't actually looked at gtk-doc in detail
[08:50] <thomasvs> walters: well, everything that's not documented in the source but is documented has to be put in the tmpl/ files
[08:51] <walters> are the .sgml files in tmpl/ generated or not?
[08:52] <ds-gromit> jcsston: do you just want to see what it looks like?
[08:52] <ds-gromit> jcsston: you should be able to function without it
[08:53] <thomasvs> walters: gtk-doc scans them, adds other stuff to them from the docs, keeps the stuff that was hand-put there, and removes the functions that are no longer part of the API
[08:53] <thomasvs> walters: so it's all just big messy voodoo
[08:54] <thomasvs> walters: it might be though that the complaints about that have caused gtk-doc to find a better way to manage them, and I just don't know about it
[08:54] <walters> yeah, that's evil.
[08:54] <jcsston> ds-gromit: gst-launch is failing with an unknown plugin error with 'fakesrc ! fakesink'. so I'm trying to run gst-register. however gst-register is crashing after it fails to find the gst registry
[08:54] <thomasvs> walters: and I'd love to change it so that it's not as brittle, so if you know there's a better way ...
[08:54] <walters> you basically should never mix human and machine-generated source
[08:54] <walters> so it looks like a gtk-doc bug
[08:54] <thomasvs> walters: agreed :)
[08:54] <thomasvs> jcsston: chances are good we use unix path separators and the like
[08:55] <thomasvs> jcsston: so step 1 would be to figure out what it is doing
[08:55] <thomasvs> jcsston: I don't suppose you have strace ?
[08:55] <thomasvs> (jcsston: what platform are you on ?)
[08:55] <jcsston> Win32
[08:55] <thomasvs> jcsston: and you're not robux4 ?
[08:55] <jcsston> no, I'm someone else from the Matroksa/TCME team
[08:55] <thomasvs> ah, ok
[08:56] <thomasvs> well, you could try running with --gst-debug=*:5 as an argument, but make sure you pipe the stderr output to a file
[08:56] <thomasvs> then we can check that file for what it's doing
[08:56] Action: thomasvs is already surprised compiling actually went this well
[08:59] <jcsston> hmm, with --gst-debug=*:5 the gst_init() kills gst-register
[08:59] <jcsston> gst_init() call
[08:59] somex1 ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[09:01] <jcsston> well, I'm going to sleep now. I'll work on it tomorrow
[09:01] jcsston ([email protected]) left irc: "zZz"
[09:01] <mathrick> i think robux4 stripped all debug to be noop / return null
[09:01] <mathrick> dyh
[09:01] <mathrick> he left
[09:13] swentel ([email protected]) joined #gstreamer.
[09:13] swentel ([email protected]) left #gstreamer.
[09:23] pippin ([email protected]) joined #gstreamer.
[09:37] walters ([email protected]) left irc: "out"
[09:51] wereHamster ([email protected]) joined #gstreamer.
[09:55] <taaz> thomasvs: looked at that patch yet? ;)
[10:01] Nick change: KA_afk -> KA
[10:01] jdahlin ([email protected]) joined #gstreamer.
[10:01] thomasvs ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:16] Shoragan ([email protected]) joined #gstreamer.
[10:16] thomasvs ([email protected]) joined #gstreamer.
[10:17] nzkoz ([email protected]) joined #gstreamer.
[10:17] <nzkoz> steve_b: you around?
[10:21] aflinta ([email protected]) joined #gstreamer.
[10:21] <aflinta> hi
[10:21] alley_cat ([email protected]) left irc: "May the Source be with you!"
[10:21] <aflinta> I have problem during gst-plugins compillation
[10:22] <aflinta> gstmpeg2dec.c:622: error: `PIC_FLAG_PTS' undeclared (first use in this function)
[10:22] alley_cat ([email protected]) joined #gstreamer.
[10:22] <aflinta> I have mpeg2dec 0.4.0b
[10:22] <aflinta> what's wrong?
[10:25] <taaz> works here
[10:25] <taaz> latest cvs?
[10:26] <taaz> btw, why do we have code that does AC_SUBST(FOO_CFLAGS/LIBS) after pkgconfig checks?  the pkg config macro does that automagically
[10:27] <aflinta> no, im compiling 0.7.3 - is this fixed in cvs?
[10:28] <thomasvs> taaz: it probably didn't at some point in the past
[10:28] <taaz> aflinta: yes
[10:29] nzkoz ([email protected]) left #gstreamer ("Client exiting").
[10:29] <aflinta> thx
[10:31] <wereHamster> What does this mean: error #1419: external declaration in primary source file ?
[10:37] dolphy ([email protected]) joined #gstreamer.
[10:39] <thomasvs> wereHamster: -ENOCONTEXT
[10:40] <dolphy> morning
[10:42] Action: thomasvs wonders if he should go through bugzilla before trying to do a release
[10:42] <thomasvs> dolphy: when vis is turned on in the player, it locks up on playing sound files
[10:42] <thomasvs> dolphy: I'm assuming that's normal ?
[10:43] thaytan ([email protected]) joined #gstreamer.
[10:44] <dolphy> thomasvs: yup
[10:44] <dolphy> thomasvs: gst_pad_pull in switch...
[10:45] <taaz> thomasvs: feel free to comment on my patch before release ;)  i just threw more gst-python stuff in cvs that needs the includedir parts for unintalled builds.
[10:48] <KoRnouille> hello
[10:49] <KoRnouille> I have already asked that question, but maybe asking it again might get things going. I have a project I'd like to acompilsh. And I need gstreamer for that
[10:49] <KoRnouille> in fact, i need the swfdec plugin
[10:49] <KoRnouille> but that's not the problem, the swfdec plugin is included I think now
[10:50] <KoRnouille> what I need is to be able to read flash files on my TV out.
[10:50] <wereHamster> thomasvs: I tried to compile gstreamer with the intel compiler (icc)
[10:50] <taaz> the litte GstPlay player.py example i checked in has serious issues for me.  dies in <10s almost always.  gets stuck printing one time value.  no clue why.  will do video somewhat ok with a few lines adjustment
[10:50] <KoRnouille> this TVout is a DVB card (linuxtv.org)
[10:50] <wereHamster> lex._gst_parse_yy.c(118): error #1419: external declaration in primary source file
[10:50] <wereHamster>   extern int yyleng;
[10:51] <KoRnouille> many players have a pluging to use the application using the DVB TV out... mplayer for example
[10:51] <KoRnouille> so this way, everything I read on gstreamer, we could all TVout on a DVB out card
[10:52] <KoRnouille> so I could read flash files with gstreamer out on my TV out. :)
[10:52] <KoRnouille> isn't that cool ?
[10:56] markey ([email protected]) joined #gstreamer.
[11:03] markey ([email protected]) left irc: "leaving"
[11:03] <thomasvs> KoRnouille: yep.  it would need a plugin for the dvb tv out though
[11:04] jdahlin ([email protected]) left irc: Read error: 60 (Operation timed out)
[11:09] <NFusi0n> using gstreamer with rhythmbox, is it possible to make it play "gaplessly"? Is that a gstreamer problem or a rhythmbox problem anyways?
[11:09] <NFusi0n> i asked on the rb mailing list and never got a straight answer
[11:11] <thomasvs> it's a rhythmbox problem, but a design issue
[11:12] <thomasvs> rhythmbox is designed to play one file at a time
[11:14] <NFusi0n> for example, musicpd www.musicpd.org plays "one file at a time" but it transitions from one to the next with no gap at all
[11:14] Company ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[11:18] <mathrick> bbl
[11:18] Nick change: mathrick -> mathrick|Uni
[11:18] <thomasvs> NFusi0n: which means, internally, it has a concept of two tracks at the same time playing
[11:18] <thomasvs> NFusi0n: which rhythmbox doesn't have
[11:20] <NFusi0n> oh, ok, i gotcha
[11:20] <NFusi0n> i hate gaps so much O:-)
[11:21] <thomasvs> NFusi0n: I understand why you'd like that, though :)
[11:21] <NFusi0n> I listen to a lot of trance mixes
[11:22] <NFusi0n> split from a 700 meg wav with a perl script and encoded losslessly
[11:22] <NFusi0n> gaps *ruin* it
[11:23] <NFusi0n> one of the main reasons i avoid portable MP3 players like the plague
[11:33] ensonic ([email protected]) joined #gstreamer.
[11:33] <ensonic> hi hi & re
[11:40] sublett ([email protected]) joined #gstreamer.
[11:50] pb_ ([email protected]) joined #gstreamer.
[11:59] <ensonic> shit upd plugin is in the same broken state as tcp :-(
[12:01] Action: ensonic is away: I am nicht hier at the computer
[12:13] aflinta ([email protected]) left irc: "brb"
[12:15] jdahlin ([email protected]) joined #gstreamer.
[12:17] <thomasvs> do we want our tools i18n-ized as well ?
[12:17] <thomasvs> say, output of gst-inspect ?
[12:29] <thaytan> there's not that many strings in gst-inspect, I don't think
[12:29] <thaytan> but is it worth the effort? They're not really user tools
[12:30] Action: ensonic is back (gone 00:29:10)
[12:30] <thomasvs> thaytan: well
[12:31] <thomasvs> thaytan: everyone translates options from --help
[12:31] <thomasvs> thaytan: so, I started doing that for the two options in gst-inspect
[12:31] <thomasvs> thaytan: then the rest starts sticking out like sore thumbs :)
[12:31] <thomasvs> thaytan: for example, in the editor, should we translate stuff ?
[12:32] <jdahlin> What's the correct way of checking if a certain plugin is installed?
[12:33] <thaytan> thomasvs: I don't need it, so no ;)
[12:36] <thomasvs> jdahlin: gst-inspect returning 0 on it
[12:37] markey ([email protected]) joined #gstreamer.
[12:39] <jdahlin> thomasvs: okay, thanks
[12:39] <jdahlin> I was just looking for something like --plugin-exists which would only return and not print anything...
[12:44] <thomasvs> maybe we have or should add a -q --quiet flag ?
[12:45] <jdahlin> nope
[12:45] <thomasvs> we shouldn't ?
[12:45] <jdahlin> perhaps we should, but we don
[12:45] <jdahlin> don't have :)
[12:45] Action: thomasvs struggles with popt
[12:46] markey ([email protected]) left irc: "bbl"
[12:51] wereHamster ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[12:51] <thaytan> what was the thing earlier about _stdint.h ?
[12:51] <thomasvs> I don't know either
[12:52] Action: thomasvs fixes his popt problem
[12:53] <thaytan> oh, it needs a reconf
[13:00] <thomasvs> thaytan: why does dvdnavsrc include dvdread headers ?
[13:02] Action: thomasvs fixes that
[13:03] <thaytan> it was using navPrint from that header
[13:03] <thaytan> but it shouldn't, yeah
[13:03] <thaytan> I hadn't noticed it
[13:03] <thomasvs> wel, dvdnav has it too apparantly
[13:05] <thaytan> probably because it's based on dvdread
[13:07] Action: thomasvs weeds out changes for popt and docs stuff from the rest
[13:14] <KoRnouille> thomasvs, you want me to test the repository for RH 9?
[13:14] <KoRnouille> im just about to install it now
[13:14] <KoRnouille> and play around with gstreamer
[13:14] <KoRnouille> im following the howto "GStreamer apt for rpm repository"
[13:23] <KoRnouille> i followed the instructions
[13:23] <KoRnouille> but
[13:23] <KoRnouille> apt-get install gstreamer-universe
[13:23] <KoRnouille> E: Couldn't find package gstreamer-universe
[13:33] Company ([email protected]) joined #gstreamer.
[13:37] sublett ([email protected]) left irc: "I like food, food is good!"
[13:46] sublett ([email protected]) joined #gstreamer.
[13:55] <bitshifter> If I create a sub buffer with gst_buffer_create_sub(),  the parent buffer/data gets ref'ed. However, when I unref that sub buffer again later, the parent buffer/data doesn't seem to get unrefed as far as I can see.
[13:56] <bitshifter> is that how it's supposed to be? Or am I reading the code wrong?
[13:57] <ensonic> can the one who checked in the colorbalance module please commit the correspondeding headerfile as well
[13:57] Action: ensonic is only trying to build the damn thing
[13:58] <Company> bitshifter: that would be an evil bug...
[13:58] sublett ([email protected]) left irc: "I like food, food is good!"
[13:59] <thaytan> it does seem to do an unref of buffer_private in _gst_buffer_sub_free
[13:59] <thaytan> which is where it stashes the reference to the parent buffer
[13:59] <Company> bitshifter: it's unreffed in gstbuffer.c:68
[13:59] <ensonic> colorbalanceneumtypes.h to be precise
[14:00] <ensonic> and commit it now please before someone else get a chance to breake a different plugin :-(
[14:00] <bitshifter> ah yes. I missed the virtual function there, sorry
[14:00] Action: Company bets ensonic should ping dolphy
[14:01] Action: ensonic thanks Componay for enlightenment who is responsible for what
[14:02] <thaytan> that should be built by glib-mkenums
[14:03] <Company> probably
[14:03] <Company> and it should be built by make
[14:03] <Company> it's possible that makefile.am was updated though
[14:04] <thomasvs> ensonic: you have to rebuild gst-plugins
[14:04] <thomasvs> ensonic: it's generated automatically
[14:04] <thaytan> oh taaz did that today
[14:04] <Company> isn't it built by make automatically?
[14:04] <thomasvs> yep
[14:04] <thomasvs> it is
[14:04] <thomasvs> he needs to rebuild, is all
[14:04] <Company> then just doing "make" should be enough, no?
[14:04] <Company> ah, ok
[14:05] <Company> rebuild involves running autogen in my book
[14:05] <Company> and tyhat takes a looo....ooong time
[14:05] <ensonic> I have updated cvs and just types make, the autotools should detect changes
[14:06] <ensonic> anyway I rebuild gstreamer packages as this has changes to and maybee the file is there ?
[14:06] <ensonic> Company: you're so right
[14:07] <Company> if you didn't run with --enable-m,aintainer-mode changes in Makefiles aren't picked up
[14:07] <Company> autogen does that automagically, configure does not...
[14:07] <ensonic> unfortunately the sun is the only halfway free machine atm., in a month or so I have a good internet connection at home, then I can develop on a recent linux and on a faster machine
[14:08] <ensonic> Company: I run autogen ;-)
[14:09] <Company> then i'll throw -ENOCLUE now
[14:19] Company ([email protected]) left irc: Remote closed the connection
[14:25] <dolphy> ensonic: what s wrong with colorbalance?
[14:26] <ensonic> the include is missing here, just have checked out the cvs, yesterday this part compiled, is the include in gstreamer module?
[14:30] <dolphy> ensonic: which file is missing ?
[14:30] <dolphy> gst-plugins/gst-libs/gst/colorbalance/ ?
[14:31] <ensonic> dolphy: yup
[14:31] <dolphy> but which file
[14:31] <dolphy> that's a complete directory 
[14:31] <ensonic> dolphy: colorbalanceneumtypes.h 
[14:33] <dolphy> colorbalanceenumtypes.h you mean ?
[14:34] <thomasvs> ensonic: like i said, it's autogenerated
[14:34] <thomasvs> ensonic: run make in gst-plugins/gst-libs/gst/colorbalance/
[14:35] markey ([email protected]) joined #gstreamer.
[14:35] <thomasvs> ensonic: unless you see a different error ?
[14:35] <ensonic> thomasvs: by the makefile in gst-plugins? or during the autogen run?
[14:36] <ensonic> cant be make as there is no target for it in the makefile
[14:36] <thomasvs> ensonic: there is
[14:36] <thomasvs> ensonic: are you *SURE* you have recent cvs update ?
[14:36] <thomasvs> Makefile.am has it clearly in that dir
[14:37] <ensonic> thomasvs: cvs is from today and 3 hours ago
[14:38] <thomasvs> ensonic: cvs stat Makefile.am in colorbalance/ and paste please
[14:38] <ensonic> strage, Makefile is newer that Makefile.am, but Makefile.am has the target and Makefile not (I'll rerun config.status)
[14:38] <thomasvs> check Makefile.in first
[14:39] <thomasvs> to make sure that has it
[14:39] <thomasvs> otherwise config.status won't help
[14:40] <ensonic> irks, only Makefile.am has it, seems like automake failed to run,seems like I have to wait another 30 minutes for trying to agian to build
[14:41] foser ([email protected]) joined #gstreamer.
[14:41] <dolphy> ensonic: you kidding ! gst-plugins is very fast to autogen :-D
[14:41] <thomasvs> ensonic: run autoreconf instead
[14:42] <ensonic> dolphy: I am the exception (sun/sparc 366 Mhz) - needs 25 minutes for autogen, and 1:20 hours to build :-(
[14:42] <ensonic> thomasvs: already doing that ;-)
[14:43] <dolphy> ensonic: same here :)
[14:45] <ensonic> dolphy: sun/sparc or doing autoreconf ?
[14:51] <dolphy> ensonic: well gst-plugins is very long to autogen on slow machines
[14:51] <dolphy> ensonic: i have a G4 1Ghz ppc
[14:53] Shoragan ([email protected]) left irc: Remote closed the connection
[14:56] thaytan ([email protected]) left irc: Remote closed the connection
[15:00] mathrick|Uni ([email protected]) left irc: Remote closed the connection
[15:06] Rotty ([email protected]) joined #gstreamer.
[15:07] thaytan ([email protected]) joined #gstreamer.
[15:12] Shoragan ([email protected]) joined #gstreamer.
[15:14] markey ([email protected]) left irc: "bbl"
[15:17] AlexCONRAD ([email protected]) joined #gstreamer.
[15:21] arkaitz ([email protected]) joined #gstreamer.
[15:21] <arkaitz> hello
[15:22] <AlexCONRAD> lo
[15:23] <arkaitz> I have a problem, I can listen the sound when I use gst-player in gnome (I can if I disconnect esd), so, I can use rhythmbox... etc... what can I do?
[15:24] <thomasvs> arkaitz: could you explain it a bit more ? I didn't understand completely
[15:25] <KA> uh. are somewhere dpman process functions explained? what they internaly do?
[15:27] <arkaitz> I'm using Gnome. When I try to open a file (e.g. a mpg film) I get a message like: /dev/dsp in use. XMMS works fine, other media players (like Totem) works fine, but gst-player can't deal with sound when esd is working.
[15:29] <thomasvs> arkaitz: open up gconf-editor, and tell me what /system/gstreamer/audiosink is set to
[15:29] <thomasvs> KA: in the manual and PWG they're well documented
[15:31] KoRnouille ([email protected]) left irc: Read error: 113 (No route to host)
[15:34] <KA> thomasvs: manual explains how dparam can be registered, I've read the chapt. 10 mins ago. I've found nothing about GST_DPMAN_PROCESS for examples
[15:35] sublett ([email protected]) joined #gstreamer.
[15:35] <arkaitz> ther was something like osssink... and I have changed right now to esdsink. Is it ok?
[15:35] <thomasvs> arkaitz: yep
[15:35] <arkaitz> sink=true or sink=false?
[15:36] <KA> I can think it is a function to process a/synchronous data
[15:37] <KA> well, sycn, since async aren't implemented
[15:38] <KA> ah, in pwg. there's a chapt I've not read!
[15:38] <arkaitz> so, I erase sync=foo?
[15:40] <arkaitz> osscommon: Unable to open /dev/dsp (in use ?)
[15:41] <thomasvs> arkaitz: restart your app
[15:42] <arkaitz> now it's working... I had to disconnect esd
[15:44] BBB ([email protected]) joined #gstreamer.
[15:44] <BBB> howdy
[15:45] <dolphy> BBB: hi
[15:46] <arkaitz> if I have esd working, is it possible to use gst-player? or I have to switch it off?
[15:49] <thomasvs> arkaitz: if esd is running, and your gconf key is set to esdsink, then yes, you can just use gst-player
[15:50] <arkaitz> thanks
[15:50] <arkaitz> I'll try again
[15:50] arkaitz ([email protected]) left irc: "Leaving"
[15:53] Nick change: KA -> KA_afk
[16:01] <BBB> <-- arkaitz heeft  verlaten ("Leaving") <- oops, small translation bug ;)
[16:02] <BBB> the "%s" is missing :p
[16:03] <BBB> I noticed jeremy fixed svq3 playback?
[16:03] <BBB> or didn't he?
[16:03] <BBB> I'm intending to fix faad once I'm in the US
[16:03] <BBB> for now, I'm not really planning anything
[16:03] <BBB> :)
[16:09] BBB ([email protected]) left irc: "Bezig met afsluiten"
[16:32] ploum ([email protected]) joined #gstreamer.
[16:34] ploum ([email protected]) left #gstreamer ("Leaving").
[16:41] trow ([email protected]) joined #gstreamer.
[16:44] <jdahlin> anyone around with experience with the swfdec plugin?
[16:49] <dolphy> jdahlin: maybe ds
[16:50] <jdahlin> It looks pretty broken, at least to me
[16:52] Company ([email protected]) joined #gstreamer.
[16:54] markey ([email protected]) joined #gstreamer.
[16:57] <bitshifter> gah, there are people who put whole readme files between the ID3v2 tag and the first mpeg audio sync
[16:58] <Company> sure
[16:58] <Company> there's also people who use id3v2 footers
[16:58] <Company> or xing headers
[16:59] markey ([email protected]) left irc: "bbl"
[17:06] <ensonic> yeah, build are complete ;-)
[17:07] yippi ([email protected]) joined #gstreamer.
[17:12] <Company> ensonic: virgin cvs or loads of little fixes?
[17:14] <ensonic> Company: all fixes are in bugzilla, major problem was tcp/upd plugin as solaris has no inet_aton, rest is nitpicking due to -Wall
[17:14] jcsston ([email protected]) joined #gstreamer.
[17:15] <ensonic> Company: currently pluins/make install still run and then I'll run gst-register
[17:15] <ensonic> Company: i'll then see what 'make check' does
[17:16] Manny ([email protected]) joined #gstreamer.
[17:16] <Company> ensonic: good luck
[17:16] <Manny> hi
[17:16] kmaraas ([email protected]) joined #gstreamer.
[17:17] <Manny> does anybody of you know a rm->wav pipeline?
[17:17] <Manny> gst-inspect didn't return anything like rmparse
[17:17] <Manny> which would be analouge to wavparse
[17:17] apoc ([email protected]) joined #gstreamer.
[17:17] <apoc> hi
[17:18] <Manny> apoc: hi
[17:18] <ensonic> Company: am i going to need it for make check ?
[17:20] <Company> ensonic: luck? yeah, even I need it when running make check sometimes ;)
[17:20] <Company> Manny: i don't think we do real audio yet - i'm not sure though
[17:21] <Manny> Company: how can I do mp3->wav?
[17:23] <Company> Manny: filesrc location=file.mp3 ! mad ! wavenc ! filesink location=file.wav
[17:23] <Manny> thanks
[17:23] <jdahlin> mpeg1videoparse seems to be pretty broken
[17:25] <Company> jcsston: you might need something like http://cvs.gnome.org/bonsai/cvsblame.cgi?file=glib%2Fglib/gtypes.h&rev=&root=/cvs/gnome#398 for the exported variables in gst btw
[17:28] <jcsston> Company: thanks, I'll try that
[17:41] BBB ([email protected]) joined #gstreamer.
[17:42] Nick change: KA_afk -> KA
[17:45] <apoc> hi BBB
[17:45] <bitshifter> Is there by any chance some kind of debugging hook for a GstBuffer/GstData to track when it gets freed? 
[17:46] Manny ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[17:47] <BBB> hi apoc... nice job in the svq3 decoding! :)
[17:47] <BBB> bitshifter: see --gst-debug-level= documentation
[17:47] Action: BBB doesn't know exactly how it works, but it's possible
[17:48] <thomasvs> BBB: does it work with spider now ?
[17:48] <thomasvs> ok, just commited some changes to those glib-mkenum Makefile commits from taaz
[17:48] <thomasvs> can someone check out and run make to test ? works fine here
[17:49] <apoc> BBB : thx :)
[17:49] <apoc> thomasvs: you mean SVQ3, qtdemux with spider ?
[17:50] <thomasvs> apoc: yep
[17:51] <apoc> thomasvs: it doesn't here
[17:51] <thomasvs> apoc: any idea why ?
[17:51] <apoc> thomasvs: qtdemux doesn't support seek events too
[17:51] <thomasvs> apoc: no seek ? why ?
[17:52] <dolphy> qtdemux is just not implementing src event handler
[17:52] <apoc> thomasvs: we should add padtemplate to qtdemux i think
[17:53] <ensonic> can anyone please tell me what you have in gst-plugins/ext/gdk_pixbuf/Makefile in the line defining 'mkdir_p' - I have an invalid command here
[17:54] <BBB> thomasvs: no clue
[17:54] <BBB> thomasvs: I don't have a computer right now
[17:54] <thomasvs> ensonic: I don't have that var in there
[17:54] <dolphy> BBB: typing on a mobile phone ?
[17:54] <BBB> my parents'
[17:55] <BBB> no laptop = no development for me
[17:55] <ensonic> thomasvs: ok, I uncommented it, touched Makefile.am and rerun
[17:57] ploum ([email protected]) joined #gstreamer.
[17:57] <BBB> hi ploum
[17:57] <ploum> hi !
[17:57] <ploum> I'm just learning wxwindows in order to work on TCME :)
[17:58] <BBB> wxwindows?
[17:58] <BBB> yikes
[17:58] <ploum> I see your matroska gstreamer is done ? 
[17:58] <BBB> you're not really gonna write UIs in that pre-ancient UI toolkit, are you?
[17:58] <BBB> yes
[17:59] <ploum> well, ash ChrisHJW_log ...
[17:59] <ploum> (ask
[17:59] <BBB> I did
[17:59] <ploum> but it seems that wxWindows will be used.
[17:59] <BBB> he wants a portable UI
[17:59] <BBB> gtk is portable ;)
[17:59] <BBB> but anyway, I only care about gnome
[17:59] <BBB> so I want a gnome UI
[18:00] <Company> rb on windows would be cool
[18:00] <Company> i could uninstall winamp then :)
[18:00] <ploum> BBB, I'm also a gnome fan. And as far as I can, the UI will be HIG compliant
[18:01] <ploum> for me, gtk2 is the best, but it's not my decision here.. 
[18:01] <BBB> I'll only help - at best - for the technical backend
[18:01] <BBB> as long as it's gnonlin alike
[18:01] <BBB> I'll write my own UI ;)
[18:01] <ploum> lol
[18:01] <ploum> I'm very intersted in the backend (but as a spectator for now)
[18:02] <BBB> :)
[18:02] <ploum> Well, I hope to see wxWindows ported to GTK2 soon
[18:02] <BBB> ???
[18:03] <jcsston> <ploum> Well, I hope to see wxWindows ported to GTK2 soon <-- ?
[18:03] <apoc> ploum: I think it is 
[18:03] <ploum> really ?
[18:03] <jcsston> ./configure --with-gtk2 IIRC
[18:03] taf2 ([email protected]) left irc: Read error: 110 (Connection timed out)
[18:03] <ploum> well.. I wasn't aware about that.
[18:04] <ploum> I use the debian package..
[18:04] Zeenix ([email protected]) joined #gstreamer.
[18:04] <ploum> I will check this ! thks for the tip :)
[18:04] <Zeenix> hi
[18:05] <ploum> hello
[18:06] <ploum> but there's a good news for you BBB : the two hundreds pages of Gnome HIG are printed and standing on my desk :)
[18:06] <BBB> omg you printed them?
[18:06] <BBB> I just use them as a reference once in a while
[18:08] <ploum> Well, I ask my father to print it at his work ;)
[18:09] pippin ([email protected]) left irc: "disappear  v 1: get lost, esp. without warning or explanation;  [syn: {vanish}, {go away}] [ant: {appear}]"
[18:11] <bitshifter> BBB: thanks. Looks like the debugging section in the manual is not quite up-to-date though :)
[18:11] markey ([email protected]) joined #gstreamer.
[18:11] <BBB> hm :/
[18:15] taf2 ([email protected]) joined #gstreamer.
[18:15] <bitshifter> also, if I run something like    int main (...)  {   gst_init(...);   while(1) ; }   through memprof, it shows ca. 1.4 MB in use, of which ca. 1MB is allegedly leaked
[18:16] <Company> "allegedly leaked" ?
[18:17] <bitshifter> "allegedly" because I don't know if it's a memprof fsck-up or if gstreamer is leaking all that
[18:17] jcsston ([email protected]) left irc: "VDubMOD watchout TCME is coming!"
[18:18] <Company> dunno - gstreamer loads the registry
[18:19] <bitshifter> yeah, but all that memory should still be referenced somewhere, shouldn't it? 
[18:20] <Company> it should
[18:20] <dolphy> Company: what was your hack to work around alsasink clicks ?
[18:21] <dolphy> Company: device="hw:0" ?
[18:21] <bitshifter> valgrind says the same thing more or less (856kB definitively lost in ...)
[18:22] <Company> dolphy: yeah
[18:22] <dolphy> dolphy@dolphy:~/workdir/totem$ /opt/gstreamer/bin/gst-launch-0.7 sinesrc ! alsasink device="hw:0"                                                                                                 
[18:22] <Company> uhm, no
[18:22] <dolphy> RUNNING pipeline ...
[18:22] <dolphy> ERROR: pipeline doesn't want to play.
[18:23] <Company> that was to get around the assertion failure
[18:23] <Company> use audioconvert
[18:23] <Company> maybe alsasink doesn't want mono
[18:24] <dolphy> well that's not removing the clicks :)
[18:24] <jdahlin> Is there a utility for displaying information about an mpeg file, like streams and so?
[18:24] <jdahlin> more than "file" does at least...
[18:24] <Company> dolphy: no idea where that clicks even come from
[18:25] <apoc> thomasvs: qtdemux works with spider for svq1 but not for svq3
[18:26] <ensonic> thomasvs: in ext/gdk_pixbug some variable like DESTDIR,loaderdir,plugindir is empty and thus the make install fails
[18:26] <apoc> thomasvs: with SVQ3 files, spider chooses ffdec_svq1 ... strange
[18:27] <Company> make qtdemux set the right caps :o
[18:28] <Company> btw, why are svq 1 and svq 3 both video/x-svq and not video/x-svq-1 and video/x-svq-3
[18:28] <Company> ?
[18:29] <apoc> oh i thought the svqversion property was enough ...
[18:29] <ensonic> or there is some clock skew going on and after the cvs update not everything gets remaked ...
[18:29] <Company> the svqvewrsion should be enough
[18:29] <dolphy> Company: i guess it simply comes from kernel scheduling
[18:30] <apoc> Company: bug ? ;)
[18:30] <BBB> Company: svqversion is a property
[18:31] <BBB> it's a property to make things understandable
[18:31] <BBB> like mpegversion, etc.
[18:31] <Company> yeah
[18:31] Action: ensonic slowly starts hating it, building a cvs snapshot of gnome is easier
[18:31] <Company> but it makes no sense to use a property there, because those 2 formats are so different
[18:32] <BBB> I disagree... but anyway, feel free to propose it to the others and see what they think
[18:32] <Company> which plugin would use the same code path for processing svq1 and svq3?
[18:33] <BBB> it's not for code paths... it's for clarity... svq formats were created by the same company, so we use the same mimetype... versions don't belong in mimetypes (IANA guidelines, I suppose)
[18:34] <Company> we don't use iana guidelines here
[18:34] <Company> we use what makes sense
[18:34] <BBB> I created these things one or two years ago... and I thought this made sense
[18:34] <Company> i dunno
[18:35] <BBB> version numbers don't belong in mimetypes was one of the things that I took into account by then
[18:35] <BBB> but as I said: feel free to propose new mimetypes if you want
[18:35] <BBB> even tohugh I don't see how it improves this
[18:35] <Company> it never made sense to me to put two totally different formats under the same mimetype umbrella
[18:36] <Company> best way to differentiate: do they use the same properties?
[18:36] <BBB> the one is the sequel for the other
[18:36] <Company> 2nd best way: are there apps that read out those properties and store them for processing (not for sweitching code paths)
[18:36] <BBB> I don't think so, but that's only because ffmpeg requires some data for svq3 which it doesn't for svq1... as for your second: no
[18:37] <BBB> (did I ever mention that ffmpeg is evil?)
[18:37] <Company> that's why we split int and float audio for example
[18:37] <BBB> I know
[18:37] <BBB> and I agreed there
[18:37] <BBB> and I this doesn't make as much sense... but again: feel free to propose something different
[18:37] <BBB> I don't see how it helps
[18:38] <Company> nah, it doesn't help
[18:38] <Company> i was just wondering why we did this
[18:38] <Company> and it helps for clarity
[18:38] <Company> you don't accidently accept svq3 because you forgot to add the svqversion
[18:41] <BBB> then, both wouldn't accept it
[18:41] <BBB> (both decoders)
[18:41] <BBB> both require the version set
[18:42] <BBB> (unless I did that wrongly in gstffmpegcodecmap.c)
[18:44] ploum ([email protected]) left irc: "Leaving"
[18:46] <thomasvs> so, we can fix it so svq3 will work ?
[18:46] hadley ([email protected]) joined #gstreamer.
[18:47] <jdahlin> gst-launch-0.7 filesrc location="media-data/medium/barrage.vob" ! mpegdemux name=demux ! { queue ! mpeg2dec ! ffcolorspace ! xvimagesink }  { demux. ! queue ! a52dec ! osssink }
[18:47] <jdahlin>  
[18:47] <jdahlin> RUNNING pipeline ...
[18:47] <jdahlin> No accelerated IMDCT transform found
[18:48] <apoc> jdahlin: I think that "No accelerated IMDCT transform found" is a liba52dec warning/error
[18:48] <jdahlin> apoc: nothing is displayed
[18:50] <apoc> jdahlin: no error ?
[18:50] <apoc> jdahlin: and with ximagesink ?
[18:51] <jdahlin> hmm, better
[18:53] <thomasvs> apoc: so the svq3 problem is an easyfix then ?
[18:54] tsiar ([[email protected]) left irc: "Client exiting"
[18:55] hadley ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[18:57] <apoc> thomasvs: if the fix is change video/x-svq to video/x-svq-1 and video/x-svq-3 ... i think yes
[18:58] <thomasvs> Company,BBB: is that the only right way to fix it ?
[18:58] water ([email protected]) joined #gstreamer.
[18:58] <thomasvs> I'm assuming video/x-svq-1 is not an official mime type
[18:58] water ([email protected]) left #gstreamer ("The water has parted.").
[19:00] Shoragan ([email protected]) left irc: Read error: 113 (No route to host)
[19:01] yoann ([email protected]) joined #gstreamer.
[19:01] <yoann> hi
[19:02] <jdahlin> How can I simplify a pipeline like this mpegdemux name=demux ! { .demux ! queue ! md5sink } ?
[19:02] <jdahlin> mpegdemux .audio_00 ! md5sink does not seem to do the trick
[19:03] Shoragan ([email protected]) joined #gstreamer.
[19:04] <apoc> jdahlin: .audio_00! md5sink
[19:04] <apoc> jdahlin: no space
[19:04] <jdahlin> apoc: why is it important not to have spaces?
[19:05] <ensonic> <offtopic> has anyone an idea what values one can pass to the --class option (window manager class) the gtk-apps do have </offtopic>
[19:06] <apoc> jdahlin: duno ... the grammar needs no space in this case
[19:07] <jdahlin> apoc: I don't care, it looks better with spaces on both sides of the exclamation mark :-)
[19:08] Action: jdahlin pokes Company
[19:09] <ds-work> the 0.7 grammar ignores whitespace
[19:10] <jdahlin> name=demux ! { .demux ! queue ! a52dec ! md5sink } works perfectly, .audio_00 ! a52dec ! md5sink fails
[19:11] <ds-work> which element has a .demux pad?
[19:11] <jdahlin> mpegdemux
[19:12] <ds-work> it does?
[19:14] <Company> it ignores the .demux probably
[19:15] <jdahlin> hmm, yes
[19:15] <Company> and since sometimes pads do lazy linking you never know if the md5 you get is actually the correct one ;)
[19:15] hadley ([email protected]) joined #gstreamer.
[19:15] <jdahlin> right :/
[19:16] <Company> it's probably the same as the output of eossrc ! md5sink would be
[19:16] <jdahlin> mpeg2dec and a52dec seems to accept the stream
[19:16] <Company> most likely the file is not ac3
[19:16] <jdahlin> mplayer said so
[19:17] <Company> according to gst-inspect ac3 streams are private streams
[19:19] <jdahlin> xine says something about liba52:a52 frame failed crc16 checksum.
[19:19] ChrisHJW ([email protected]) joined #gstreamer.
[19:19] robUx4 ([email protected]) joined #gstreamer.
[19:20] <robUx4> lo
[19:21] <Company> robUx4: <Company> jcsston: you might need something like http://cvs.gnome.org/bonsai/cvsblame.cgi?file=glib%2Fglib/gtypes.h&rev=&root=/cvs/gnome#398 for the exported variables in gst btw
[19:21] <ensonic> lets see if plugins compile though this night - ciao and see you tomorrow
[19:22] ensonic ([email protected]) left #gstreamer.
[19:26] <apoc> Company: if spider don't use the svqversion property. is this a bug ?
[19:28] <Company> apoc: yes
[19:28] <Company> though i would be quite surprised if it doesn'T - but who knows
[19:29] <apoc> Company: can i send you a gst-log ?
[19:30] <Company> i can probably reproduce it
[19:30] <robUx4> Company: I used a .def file instead
[19:30] <robUx4> as done in GLib already
[19:31] <robUx4> it's easier and no need to change any code :D
[19:31] <apoc> Company : can spider work with plugin with no padtemplates like qtdemux
[19:32] <Company> apoc: yes
[19:36] <Company> apoc: fix will hit cvs in a minute
[19:37] mxpxpod ([email protected]) joined #gstreamer.
[19:38] <apoc> Company: cool :)
[19:39] <apoc> hm colorbalance don't build anymore
[19:39] <thomasvs> apoc: you have to rerun autogen.sh I'm afraid
[19:39] <apoc> Makefile:618: .deps/libgstcolorbalance_la-colorbalance-enumtypes.Plo: No such file or directory
[19:40] <thomasvs> apoc: or autoreconf from the top
[19:40] <thomasvs> apoc: or, remove the four lines around 618 from the Makefile would probably work too
[19:40] md`` ([email protected]) joined #gstreamer.
[19:40] <thomasvs> later people
[19:42] md` ([email protected]) left irc: Nick collision from services.
[19:42] Nick change: md`` -> md`
[19:43] <Company> apoc: fixed
[19:43] <Company> apoc: gst-plugins
[19:43] alley_cat ([email protected]) left irc: "May the Source be with you!"
[19:45] <apoc> Company: ok thx
[19:46] hadley ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[19:58] thomasvs ([email protected]) left irc: Read error: 113 (No route to host)
[20:03] alley_cat ([email protected]) joined #gstreamer.
[20:03] mathrick ([email protected]) joined #gstreamer.
[20:04] <mathrick> re
[20:06] <mxpxpod> Company: fixed in cvs?
[20:06] jdahlin ([email protected]) left irc: "Klienten avslutas"
[20:09] alley_cat ([email protected]) left irc: "May the Source be with you!"
[20:09] alley_cat ([email protected]) joined #gstreamer.
[20:09] <Company> mxpxpod: what?
[20:10] <mxpxpod> Company: is that colorbalance thing fixed in cvs?
[20:10] <mxpxpod> I'm getting the same error
[20:10] <Company> i think so
[20:11] <Company> you need to reautogen though
[20:12] <BBB> the svq-3 change doesn't make it easier
[20:12] <BBB> it's something else
[20:12] <BBB> (late)
[20:13] <Company> BBB: it was qtdemux
[20:22] pb_ ([email protected]) left irc: "Client exiting"
[20:23] somex1 ([email protected]) joined #gstreamer.
[20:23] <BBB> Company: oh... sorry, missed that
[20:23] Action: BBB slow
[20:24] <BBB> bbl, gotta run for something
[20:24] BBB ([email protected]) left irc: "Bezig met afsluiten"
[20:24] Action: dolphy really doubts the gst_pad_select implementation of the basic scheduler works
[20:25] thomasvs ([email protected]) joined #gstreamer.
[20:27] <dolphy> thomasvs: this time i saw you quit :)
[20:32] <robUx4> anyone saw this ? http://www.theinquirer.net/?article=13813
[20:34] <ds-work> that's like, last week's news
[20:35] <robUx4> dunno
[20:35] <robUx4> I just saw it
[20:44] Uraeus ([email protected]) joined #gstreamer.
[20:46] <yoann> just to know, is there any audio application musicbrainz aware ?
[20:47] <apoc> hi Uraeus
[20:47] <Uraeus> yoann: like RB you mean?
[20:47] <Uraeus> hey apoc
[20:48] <yoann> Uraeus: afaik, support in rhythmbox is disabled
[20:49] <Zeenix> Uraeus: sam did'nt like me changes to libcaca & he proposed a different mechanism which MAY take some time (a week or two) to implement :(
[20:49] <Zeenix> s/me/my
[20:49] <Uraeus> Zeenix: that is a maintainers right
[20:50] <Zeenix> Uraeus: no, his objection was quite legitimate
[20:52] <Zeenix> Uraeus: i was just sad on this delay BOSS :)
[20:52] <ds-work> how is gst-libs/gst/colorbalance supposed to build?
[20:53] <taaz> thomasvs: your glib-gen build code looks nicer than what i did. ;)
[20:54] <thomasvs> taaz: thanks :)
[20:54] <thomasvs> taaz: I was kinda annoyed at the mess, too bad I couldn't get it to run maintainer mode
[20:54] <thomasvs> ds-work: either touch the .Plo files manually, or run autoregen from the top
[20:54] <taaz> it doesn't build though
[20:54] <ds-work> autogen from the top doesn't fix it
[20:54] <thomasvs> taaz: sure it does, just rerun autogen.sh
[20:54] <taaz> i just did
[20:54] <ds-work> no, it really doesn't
[20:54] <thomasvs> ds-work: it doesn't ? you sure ? it does here
[20:54] <taaz> i think...
[20:54] <thomasvs> crap
[20:55] <taaz> those .deps things scare me
[20:55] <thomasvs> libtool hackery
[20:55] <taaz> the auto* code is dumb sometimes and puts unexpanded vars in the deps which fail easily
[20:55] <thomasvs> well, I can add some things to autogen.sh to do that automatically, or put it in the makefile snippets
[20:55] <thomasvs> taaz: exactly
[20:56] <taaz> how do you fix that?  i was trying to do the same sort of thing for gst-python but couldn't figure it out
[20:56] <taaz> top level @FOO@ stuff is a lame way to fix it but that's ugly
[20:57] <thomasvs> taaz: just add rules to touch the .Plo files
[20:57] <ds-work> why not just specify the names of the files instead of using $(glib-moo)
[20:58] <thomasvs> I don't think that's the problem, is it ?
[20:59] <ds-work> yes
[20:59] <taaz> yeah, in _SOURCES
[20:59] <taaz> it's autocraptitude i say
[21:01] <taaz> in .deps it creates foo-$(glib_root).blah files
[21:01] <thomasvs> argh, ok :)
[21:01] <thomasvs> give me a few secs, i'll fix it then
[21:01] <taaz> let me know how, i want to do the same thing for gst-python
[21:08] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[21:10] Zeenix ([email protected]) left irc: "The whole problem with the world is that fools and fanatics are always so certain of themselves, but wiser people so full of 
[21:12] yippi ([email protected]) left irc: "Client exiting"
[21:17] <thomasvs> taaz: what I don't get is why the includes for .Plo files arent sinclude so the build doesn't fail, or why there aren't autogened rules for recreating them
[21:18] <thomasvs> ds-work: ok, update and run autoregen.sh and try again
[21:19] Action: thomasvs does make clean; make to make sure
[21:19] <taaz> yeah, it's a serious pain to deal with since you have to reautotool the thing once its broken
[21:21] yippi ([email protected]) joined #gstreamer.
[21:21] <thomasvs> I wish I could fork myself so another thread of me could go look how to add that in libtool :)
[21:21] <Uraeus> good evening yippi
[21:22] <taaz> thomasvs: you could at least check for related bug reports and add one if one one else has complained
[21:22] <ds-work> you'd have to add a lot more rules in the Makefiles
[21:23] walters ([email protected]) joined #gstreamer.
[21:24] mxpxpod ([email protected]) left irc: "Guns kill people as much as spoons made Rosie O'Donnell fat.."
[21:25] <Uraeus> ello walters
[21:25] <yippi> good evening!  :)
[21:26] <thomasvs> ok, so last item on my list is writing a gst-run wrapper that executes versioned tools
[21:26] <thomasvs> I'm probably not going to do that before releasing though
[21:26] <thomasvs> are there any major issues left right now ?
[21:26] Action: thomasvs is sad about mpeg video not playing
[21:26] <thomasvs> walters: how's seeking going in rhythmbox ? stuff works nicely in player/totem now
[21:27] <Uraeus> yippi: don't know if you already know, but we have gotten a nice set of Solaris build fixes over the last few days
[21:27] Nick change: KA -> KA_afk
[21:28] <walters> thomasvs: seems to work fine
[21:30] <taaz> thomasvs: my minor patch still pending if you have time
[21:36] <thomasvs> taaz: the pkgconfig one ?
[21:36] <thomasvs> taaz: I have an idea to make check that
[21:39] <yippi> Uraeus: i didn't know about that.  tell me about them.
[21:44] thaytan ([email protected]) left irc: Read error: 113 (No route to host)
[21:47] <Uraeus> apoc: http://bugzilla.gnome.org/show_bug.cgi?id=104361 <- your plugin fix this wishlist bug right?
[21:47] <Uraeus> yippi: not sure what to tell you trying to find the bugzilla reports with the patches, but haven't yet. I think Company or ds-gromit commited them
[21:54] <Uraeus> yippi: http://bugzilla.gnome.org/buglist.cgi?product=GStreamer&bug_status=UNCONFIRMED&bug_status=NEW&bug_status=ASSIGNED&bug_status=NEEDINFO&bug_status=REOPENED&bug_status=RESOLVED&bug_status=VERIFIED&bug_status=CLOSED&email1=kost%40imn.htwk-leipzig.de&emailtype1=substring&emailreporter1=1&emaillongdesc1=1&email2=&emailtype2=substring&emailreporter2=1&changedin=10&chfieldfrom=&chfieldto=Now&chfieldvalue=&short_desc=&short_desc_type=substring&lon
[21:54] <Uraeus> g_desc=&long_desc_type=substring&bug_file_loc=&bug_file_loc_type=substring&status_whiteboard=&status_whiteboard_type=substring&keywords=&keywords_type=anywords&op_sys_details=&op_sys_details_type=substring&version_details=&version_details_type=substring&cmdtype=doit&newqueryname=&order=Reuse+same+sort+as+last+time&form_name=query <- think this list has them all
[21:54] <Uraeus> ugh
[21:54] <Uraeus> that URL was a tab bigger than I thought
[21:55] <Uraeus> s/tab/tad/
[21:56] Action: Company hits ds-work with a cluestick taken from gststructure.c:240
[21:57] <Company> that took a while to find
[21:59] <ds-work> heh.  that's looks particularly dumb
[22:00] <ds-work> fixing that should decrease memory leakage a bit
[22:01] <walters> wow, using conditional breakpoints in gdb is...slow :)
[22:01] <Company> ds-work: i'm currently debugging mem leakage in gst-typefind
[22:03] <Company> biggest memleak was in string parsing
[22:03] <Company> caps string parsing
[22:03] <Company> 800k leaked on registry load ;)
[22:04] <ds-work> yes
[22:04] <ds-work> I obviously made no attempt to worry about memory leaks
[22:04] <Company> you can fix gst-player memleaks ;)
[22:06] <ds-work> could you fix the pause bug in rb then?
[22:08] <Company> pause bug?
[22:08] <Company> ah
[22:08] <ds-work> if I pause in rb, then unpause, it never recovers
[22:08] jobi (~jobi@3ffe:b80:3:2fbc:0:0:0:2) joined #gstreamer.
[22:08] <jobi> hi
[22:08] <Company> no, i can't until we have agreed to send discont events after an element goes READY=>PAUSED
[22:09] <ds-work> http://www.comedi.org/librfb/librfb-0.1.1.tar.gz
[22:09] <Company> ds-work: explicit caps are never reset
[22:09] <ds-work> why did it work before?
[22:09] BBB ([email protected]) joined #gstreamer.
[22:09] <ds-work> Company: set_explicit_caps (pad, NULL)
[22:09] <Company> it worked until we set the sink explicitly to NULL
[22:09] <jobi> I'm experiencing something strange with gstreamer 0.6.4, since I moved from linux-2.6.1 to linux-2.6.2-rc3-mm1
[22:09] <Company> ds-work: i mean after I set them and then unref the element
[22:10] <jobi> When I play an mp3 with rhythmbox or gst-player, the system becomes extremely slow, and I get no sound
[22:10] <ds-work> btw, that was the link for the librfb release
[22:11] <Company> jobi: i'd say it's a kernel issue
[22:12] <jobi> Company: yes, could very well be, though gstreamer is the only application that triggers this
[22:12] <Company> ds-work: in theory PAUSED=>READY should automatically unset explicit caps
[22:12] bitshifter ([email protected]) left irc: "(null)"
[22:12] bitshifter ([email protected]) joined #gstreamer.
[22:12] <ds-work> mmm, true
[22:13] <ds-work> that should be up to the element
[22:13] <Company> jobi: GStreamer was never "optimized" for smooth playback, so it's always the first to show buggy kernel behaviour
[22:13] <ds-work> pads don't know about element states
[22:13] <jobi> Company: ok
[22:13] <Company> we could autoclear them on PAUSED=>NULL
[22:13] <Company> we could autoclear them on PAUSED=>READY
[22:14] <ds-work> how?
[22:14] <Company> iterate over all pads and set them to NULL?
[22:14] <ds-work> in gstelement.c?
[22:14] <ds-work> gross
[22:14] <Company> yeah
[22:14] <Company> yea
[22:15] <Company> but there's probably a bunch of bugs because we don't reset the explicit caps
[22:15] <ds-work> perhaps
[22:15] <ds-work> seems like a resonable solution
[22:16] <taaz> ds-work: neat, comedi supports some hardware in my lab.  i should try it out sometime.
[22:16] <walters> ds-work: hm, i never noticed the pause issue :)
[22:17] <walters> ds-work: that is bad
[22:18] <ds-work> for a long time I ignored it, because I only pause to play with gst-launch
[22:18] <ds-work> and I assumed that gst-launch was somehow affecting it
[22:18] <ds-work> usually by being associated with a make install
[22:19] <Company> the problem is that the sound sinks sync to the timestamps
[22:19] <Company> on READY=>PAUSED they are set to 0
[22:19] <Company> if the next timestamp says 1:30, it waits until 1:30
[22:19] <Company> so if you pause after 10 seconds, unpausing causes 10 seconds of silence
[22:20] <Company> this is necessary because some videos start sound some time into the stream, so the first timestamp is not 0 and we make the audiosink wait until then
[22:20] Action: ds-work checks
[22:20] <ds-work> yup
[22:21] <Company> the correct solution is to send a discont automatically from the core after a link is initiated
[22:21] <Company> which is when one pad is (re)activated
[22:21] ChriHJW ([email protected]) joined #gstreamer.
[22:21] <Company> or when two active pads are linked
[22:22] <Company> but i don't know where discont events would start breaking stuff
[22:22] <ds-work> sending discont at pad link is bad for an autoplugger
[22:22] <Company> s/breaking/showing bugs/
[22:22] <Company> why?
[22:23] <ds-work> since it may want to have a wedge element (i.e., like spideridentity) that it wants to remove later
[22:23] <Company> doesn't matter
[22:23] <ds-work> without disrupting the stream
[22:23] <Company> relinking causes a discont, too
[22:23] <Company> (or should at least)
[22:24] <ds-work> we should implement that and test it out
[22:24] ChrisHJW ([email protected]) left irc: Read error: 60 (Operation timed out)
[22:25] <Company> where's the best place to put it?
[22:25] <Company> gst_pad_pull?
[22:25] <Company> schedulers?
[22:26] <ds-work> hrm, good question
[22:26] <ds-work> src elements call peer's chain functions effectively directly
[22:26] <Company> yeah
[22:27] <ds-work> you'd need a chainhandler that decides if it has sent a discont since being linked
[22:27] <Company> yup
[22:27] <Company> (oggdemux sends them explicitly which surprised me when listening to a webstream ;)
[22:33] jcsston ([email protected]) joined #gstreamer.
[22:33] <Company> what's the shortcut for %P
[22:33] <Company> GST_POINTER ?
[22:33] <ds-work> GST_PTR_FORMAT
[22:34] <ds-work> sorry for the abbr
[22:36] ploum ([email protected]) joined #gstreamer.
[22:36] jcsston ([email protected]) got netsplit.
[22:39] <Company> ds-work: is it ok to clear explicit caps on unnegotiation?
[22:39] <Company> hm, probably not
[22:42] jcsston ([email protected]) got lost in the net-split.
[22:43] sjoerd ([email protected]) left irc: Read error: 113 (No route to host)
[22:44] jcsston ([email protected]) joined #gstreamer.
[22:44] jcsston ([email protected]) got netsplit.
[22:45] sjoerd ([email protected]) joined #gstreamer.
[22:45] dolphy ([email protected]) joined #gstreamer.
[22:46] Uraeus ([email protected]) left irc: Read error: 110 (Connection timed out)
[22:46] Action: mathrick is dead tired
[22:46] <mathrick> nite people
[22:47] Nick change: mathrick -> mathrick|sleep
[22:47] somex1 ([email protected]) left #gstreamer.
[22:50] jcsston ([email protected]) got lost in the net-split.
[22:51] Nick change: KA_afk -> KA
[22:52] <ds-work> Company: not on unnegotiation, no
[22:54] <Company> i did it explicitly on PAUSED=>READY
[22:56] <robUx4> bye everyone
[22:56] robUx4 ([email protected]) left irc: "Baby wants to bleep"
[23:06] jcsston ([email protected]) joined #gstreamer.
[23:08] sublett ([email protected]) left irc: "I like food, food is good!"
[23:08] <jcsston> Does GStreamer require variable argument marcos? MSVC doesn't support them, AFAIK. Could it instead, use use marcos like GST_PAD_FORMATS_FUNCTION1 (for one arg) GST_PAD_FORMATS_FUNCTION2 (for 2) and so on?
[23:09] <Company> ok, anyone complaining about gstreamer mem usage, please run again - the worst offenders should be gone now
[23:09] <Company> MSVC doesn'T support variable arguments?
[23:10] <Company> how does printf work on windows?
[23:10] <dolphy> BBB, Company: while you are both there
[23:10] <dolphy> BBB, Company: are we all ok to add gst_x_overlay_redraw or _expose (GstXOverlay *overlay);
[23:11] <ds-work> macros, not functions
[23:11] <jcsston> <Company> MSVC doesn'T support variable arguments? <-- it supports functions with variable arguments, but not #define / marcos
[23:11] <ds-work> does it support inline functions?
[23:11] <dolphy> BBB, Company: that will just put the current image on the drawable
[23:12] <jcsston> ds-work: yes..., MSVC can do function inlining
[23:12] <BBB> it should do that automatically
[23:12] <BBB> (well, that's my idealistic idea)
[23:12] <BBB> is that possible?
[23:12] <BBB> by messing with the X event queue or so?
[23:13] <ds-work> jcsston: make sure that glib got the right type of vararg macros
[23:13] <ds-work> jcsston: G_HAVE_ISO_VARARGS or G_HAVE_GNUC_VARARGS
[23:13] <dolphy> BBB: not when PAUSED
[23:13] <ds-work> if it's vaguely c99-ish, it will have the former
[23:13] <dolphy> BBB: and that's what sux actually
[23:13] <dolphy> BBB: except if we create an event thread but i m sure we don't want to do that
[23:14] <jcsston> neither is defined
[23:14] <BBB> nah
[23:15] <Company> let me say that again: msvc doesn't do varargs macros?
[23:15] Action: Company votes running gnu cpp then
[23:15] <BBB> dolphy: well, go ahead if you think it's needed... I can't really speak, I'm off development so...
[23:15] <dolphy> Company: ?
[23:16] <dolphy> _expose or _redraw ?
[23:16] <Company> dolphy: whatever you like
[23:16] <BBB> expose
[23:16] <BBB> I guess
[23:17] <BBB> redraw is so java'ish
[23:17] <dolphy> ok
[23:17] foser ([email protected]) left irc: "[ I want to believe ]"
[23:17] <Company> what does "expose" mean?
[23:17] Action: ds-work snickers at libmba
[23:17] Action: Company looks it up
[23:17] <ds-work> http://www.ioplex.com/~miallen/libmba/
[23:17] <jcsston> Company: yes, both GST_PAD_QUERY_TYPE_FUNCTION(functionname, ...) and GST_PAD_QUERY_TYPE_FUNCTION(functionname, a...) cause synax errors
[23:17] <BBB> show off to the outer world
[23:18] <Company> yeah, whatever you like :)
[23:19] Action: ds-work grumbles
[23:19] <ds-work> GST_QUERY_TYPE_FUNCTION() is another stupid static function factory
[23:20] <Company> yeah
[23:20] <BBB> I kind of like them
[23:20] <BBB> though they're a bit hackish
[23:20] foser ([email protected]) joined #gstreamer.
[23:20] <Company> and GST_BOILERPLATE, too
[23:20] <Company> though GST_BOILERPLATE is not varargs
[23:20] foser ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[23:21] <Company> hm, who uses libmba?
[23:21] <BBB> GST_BOILERPLATE is weird, I have to get used to that
[23:21] <ds-work> it came up in the search for 'msvc variadic macros'
[23:21] Action: ds-work votes for GstStaticQueryType
[23:22] <Company> BBB: when you got used to it, you understand gnome code more easily ;)
[23:22] Action: dolphy wonders how to put the current frame in the drawable without copying it 
[23:22] <dolphy> :-/
[23:22] <jcsston> Would a patch with GST_PAD_FORMATS_FUNCTION1 (for one arg) GST_PAD_FORMATS_FUNCTION2 (for 2) be accepted? or would that be to widespread of a change?
[23:22] <BBB> Company: oh, I'm perfectly fine with gnome code... it's not that it;'s hard to understand, it's really that it's "new" so I have to learn to use it instead of writing all that glib code manually
[23:23] <BBB> er
[23:23] <BBB> c/p'ing ;)
[23:23] <Company> yeah
[23:23] <Company> i like it now :)
[23:23] <ds-work> jcsston: probably not
[23:23] <Company> though i usually don't like function generating macros...
[23:23] <ds-work> jcsston: it would be better to find a good way to fix it
[23:25] <jcsston> ds-work: I know the 1, 2, 3 method is bit hackish and not very clean :/ But it's the only working way I've found (I ran into this problem with some debugging macros before)
[23:26] <Company> you can make the debugging stuff inline functions
[23:26] Action: ds-work agrees with Company
[23:26] <ds-work> the function factory macros are, of course, harder
[23:26] <ds-work> which is why they're evil
[23:27] <Company> and GST_ELEMENT_ERROR, too ;)
[23:31] ploum ([email protected]) left irc: "Leaving"
[23:31] <dolphy> ds-work: not that gst_x_overlay_expose will not make a gst-launch pipeline refresh :-/
[23:31] <dolphy> s/not/note
[23:32] Action: ds-work moos a moo of understanding
[23:32] <ds-work> I need to resist the urge to fix GST_QUERY_TYPE_FUNCTION()
[23:33] <Company> deprecate it if you want
[23:35] <ds-work> are we removing deprecated code as part of the freeze?
[23:37] <Company> no
[23:37] walters ([email protected]) left irc: "out"
[23:37] <Company> i don't think
[23:37] <Company> we could
[23:38] <BBB> we can fix it internally
[23:38] <BBB> but deprecation stuff should be allowed by default
[23:39] <ds-work> we've sort of treated GST_DISABLE_DEPRECATED as "fix your crap _now_" recently
[23:40] <ds-work> btw, we suck at freezing
[23:40] <ds-work> just wanted to let everyone know
[23:40] sjoerd ([email protected]) left irc: Remote closed the connection
[23:40] <BBB> I didn't change anything recently
[23:40] <Company> well, we run with -DGST_DISABLE_DEPRECATED, so we need to fix our stuff _now_ :)
[23:41] <Company> deprecating stuff without notice is ok from my point of view
[23:42] <Company> anybody running releases with deprecated active is in serious need of slapping anyway...
[23:46] Action: BBB is about to totally give up on releasing
[23:46] <BBB> I've compiled libquicktime five times now
[23:46] <BBB> and each time it failed
[23:46] <BBB> that software is supposed to *FIX* quicktime4linunx issues
[23:46] <BBB> but just fails worse
[23:46] <ds-work> what do we need libquicktime for?
[23:47] <BBB> not we
[23:47] <BBB> something else
[23:47] <ds-work> ah
[23:48] <BBB> apoc said qtdemux doesn't seek... does it?
[23:48] <ds-work> no
[23:48] Action: ds-work lazy
[23:48] Action: BBB kicks ds-work
[23:49] Action: BBB puts on TODO list
[23:49] <BBB> does qt have indexes or so?
[23:49] <ds-work> yes
[23:49] <ds-work> seeking should be trivial
[23:50] <BBB> k
[23:50] <Company> that's what they always say at first...
[23:50] <Company> (though it's certainly easier than with mpeg or ogg)
[23:52] <dolphy> wheee
[23:52] <dolphy> expose works fine in totem :)
[23:53] <BBB> ?
[23:53] <BBB> Company: mpeg is easy
[23:53] <BBB> Company: mpeg is very easy
[23:54] <Company> is it?
[23:54] Action: Company has no clue about mpeg
[23:54] <BBB> sure is
[23:54] <BBB> especially CBR mpeg
[23:54] <Company> it's the only common container format i've never looked at
[23:54] <Company> haha, CBR :/
[23:55] <BBB> seek to bytes=time*muxrate in filesrc
[23:55] <BBB> and muxrate is in the MPEG header
[23:55] <BBB> there's a patch in bugzilla which enables proper seeking in VBR MPEG
[23:55] <BBB> but I didn't clean it up yet
[23:55] Action: ds-work comments that seeking will be _insane_ in swfdec
[23:56] pb_ ([email protected]) joined #gstreamer.
[23:56] <Company> do it like ogg does it
[23:56] <Company> seek to beginning and fast forward ;)
[23:56] <BBB> noooooo
[23:56] <BBB> please don't do that
[23:56] <BBB> at least build a file index table on startup
[23:57] <Company> i do that
[23:57] <Company> you are free to optimize it when you think it's too slow
[23:57] <dolphy> hmm strange that doesn't work for Xv :-/
[23:58] <BBB> Company: oh, I thought you re-seeked for every new seek
[23:59] <Company> yeah, i re-seek for every seek
[23:59] <BBB> oh conme on
[23:59] <BBB> you know what I mean :p
[23:59] Nick change: KA -> KA_afk
[23:59] <ds-work> dolphy: Xv, you need to reput the image
[23:59] <ds-work> dolphy: so it knows the new location
[00:00] --- Wed Feb  4 2004
[00:00] <dolphy> ds-work: that's what i m doing
[00:00] <dolphy> ds-work: for both xv and xshm
[00:00] <dolphy> ds-work: i m putting the image again
[00:00] <dolphy> ds-work: but with xv it works "sometimes:
[00:00] <Company> BBB: everytime a seek happens, oggdemux seeks to the start of the stream if it's a seek backwards and then discards buffers until the seek position is reached
[00:00] <dolphy> s/:/"
[00:00] <dolphy> which is weird
[00:01] <BBB> oi
[00:01] markey ([email protected]) left irc: "nite"
[00:04] <dolphy> i ll work on that some more tomorrow
[00:05] <dolphy> x[v]imagesink with expose method are in cvs
[00:05] <dolphy> totem with gstreamer backend implements it
[00:06] <dolphy> night dudes
[00:06] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[00:10] yippi ([email protected]) left irc: "Client exiting"
[00:12] sjoerd ([email protected]) joined #gstreamer.
[00:16] yippi ([email protected]) joined #gstreamer.
[00:19] pb_ ([email protected]) left irc: "Client exiting"
[00:20] pb_ ([email protected]) joined #gstreamer.
[00:25] <Company> mp3 typefinding and gst_init are memleak free
[00:25] <Company> at least according to valgrind
[00:26] <BBB> did anyone look at mpeg video stream typefinding?
[00:26] <BBB> there's a read error in it according to valgrind
[00:39] Company ([email protected]) left irc: Remote closed the connection
[00:41] Company ([email protected]) joined #gstreamer.
[00:41] walters ([email protected]) joined #gstreamer.
[00:43] sublett ([email protected]) joined #gstreamer.
[00:44] <thomasvs> so, shall I release tomorrow morning ? :)
[00:45] ChriHJW ([email protected]) left irc: Read error: 110 (Connection timed out)
[00:45] <walters> thomasvs: is there a blocker?
[00:45] <thomasvs> walters: well, mpeg video playback doesn't work correctly, but no one is working on it
[00:45] <BBB> thomasvs: please do
[00:46] BBB ([email protected]) left irc: "Bezig met afsluiten"
[00:46] <thomasvs> walters: also, pause is broken, as is playing multiple files from the player
[00:46] <walters> thomasvs: i can make rb not set the sink to NULL on pause, that should work around that issue
[00:46] <thomasvs> walters: but I think it's better to release anyway as long as people understand it's still a devel and we're spending the next cycle of bugfixing
[00:46] <walters> yeah
[00:47] <thomasvs> well, if no one complains, I'll release tomorrow
[00:47] Nick change: thomasvs -> thomasvz
[00:48] yippi ([email protected]) left irc: "Client exiting"
[00:48] <walters> cool
[01:01] thaytoo ([email protected]) joined #gstreamer.
[01:07] water_ ([email protected]) joined #gstreamer.
[01:07] Nick change: water_ -> water
[01:07] water ([email protected]) left #gstreamer ("The water has parted.").
[01:14] Shoragan ([email protected]) left irc: "Leaving"
[01:22] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[01:27] C ([email protected]) joined #gstreamer.
[01:28] thomasvz ([email protected]) left irc: Read error: 60 (Operation timed out)
[01:28] <C> what is the proper way to "pausing" a Gst pipeline when it's playing in a different thread than the one that will be pausing it?
[01:29] <ds-work> gst_element_set_state()
[01:30] <C> hmm, i tried that and i get ::: (process:28866): GStreamer-CRITICAL **: file gstelement.c: line 2357 (gst_element_set_state): assertion `GST_IS_ELEMENT (element)' failed
[01:30] <ds-work> are you calling it with an element?
[01:30] <C> gst_element_set_state (pipeline, GST_STATE_NULL);
[01:31] <ds-work> it may be a bug
[01:31] <ds-work> does it happen with any pipeline?
[01:32] <C> i only have 1 pipeline in my code... i click "play" and it plays my song.  i press "stop" and i get the error, song keeps playing
[01:33] <C> i have the same line used to stop the pipeline when the song ends, and it seems to work (song stops when it should)
[01:33] <ds-work> what is your pipeline?
[01:33] trow ([email protected]) left irc: "Leaving"
[01:35] <C> http://rafb.net/paste/results/N2441038.html
[01:37] pb_ ([email protected]) left irc: "Client exiting"
[01:38] thomasvz ([email protected]) joined #gstreamer.
[01:39] <C> is there a way to make a pipeline play in a detached thread by using GST methods?
[01:43] <ds-work> create a GstThread instead of a GstBin
[01:44] <C> oh thank you so much!  my threading isn't so good...
[01:51] eventide_ ([email protected]) joined #gstreamer.
[01:54] eventide ([email protected]) left irc: Nick collision from services.
[01:54] Nick change: eventide_ -> eventide
[01:54] sjoerd ([email protected]) got netsplit.
[01:54] sublett ([email protected]) got netsplit.
[01:55] sjoerd ([email protected]) returned to #gstreamer.
[01:56] sublett ([email protected]) returned to #gstreamer.
[01:57] jobi (~jobi@3ffe:b80:3:2fbc:0:0:0:2) got netsplit.
[01:58] jobi (~jobi@3ffe:b80:3:2fbc:0:0:0:2) returned to #gstreamer.
[01:58] kmaraas ([email protected]) got netsplit.
[01:58] AlexCONRAD ([email protected]) got netsplit.
[01:58] kmaraas ([email protected]) returned to #gstreamer.
[01:58] AlexCONRAD ([email protected]) returned to #gstreamer.
[01:59] <taaz> anyone up to latest debian and having doc build failures?
[01:59] <taaz> cd build && xmllint -noout -valid faq.xml
[01:59] <taaz> faq.xml:20: I/O error : failed to load external entity "file:///home/tobi/bulk/xml/schemas/docbook/4_2/dtd/docbookx.dtd"
[01:59] <taaz> i have no idea who tobi is ;)
[02:00] <ds-work> not here
[02:00] <taaz> i can't even find that tobi string anywhere on my system
[02:01] <taaz> anyone have a clue what things like that mean?
[02:01] <ds-work> probably a registry problem
[02:02] <taaz> yeah, but i can't find where
[02:03] Rotty ([email protected]) left irc: Remote closed the connection
[02:04] <C> http://rafb.net/paste/results/S2945678.html   i'm now using gst_thread, and when i call the STOP function it doesn't seem to be able to 'control' the pipeline in another thread (i get same error i got before when press stop)
[02:05] Action: taaz curses...  not in {/etc,/usr/share}/{xml,sgml}
[02:09] <C> never mind, sorry i fixed it
[02:09] C ([email protected]) left #gstreamer.
[02:09] Action: taaz curses at himself... old catalog file from 2002(?!) being used with total random garbage in it.
[02:10] Action: taaz didn't notice until xmlcatalogs started working in debian today ;)
[02:10] <taaz> walters: how soon till docbook hacks get to leave cdbs? ;)
[02:10] <walters> taaz: soono...
[02:10] <walters> soon
[02:13] <taaz> anyone know proper xinclude syntax to pull in a docbook section with a certain id?
[02:13] <taaz> my old hack in gst-python docs seems to fail now
[02:13] <taaz> need to do something with xpointer
[02:17] <taaz> ah, got it...
[02:17] sublett ([email protected]) left irc: "I like food, food is good!"
[02:18] <taaz> in case anyone cares ;)  change <xi:include href="AUTHORS.xml#authors" xmlns:xi="http://www.w3.org/2001/XInclude"/> to <xi:include href="AUTHORS.xml" xpointer="xpointer(//section[@id='authors'])" parse="xml" xmlns:xi="http://www.w3.org/2003/XInclude"/>
[02:19] <taaz> xml is so clear and concise
[02:32] Company ([email protected]) left irc: "Client exiting"
[02:34] Company ([email protected]) joined #gstreamer.
[02:51] C ([email protected]) joined #gstreamer.
[02:51] <C> how can i make a pipeline tell me when its done without a 'while' loop?
[02:53] <ds-work> connect to the eos signal
[02:55] <C> ok


-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn