IRC Logs

*******************************************************************
[03:11] ChrisHJW_log ([email protected]) left irc: Read error: 60 (Operation timed out)
[03:12] ChrisHJW_log ([email protected]) joined #gstreamer.
[03:13] Company ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[03:15] Company ([email protected]) joined #gstreamer.
[04:08] ChriHJW_log ([email protected]) joined #gstreamer.
[04:11] <taaz> blah... docs still failing for me.  docbook2ps pukes at me
[04:11] Action: taaz thought that was just for sgml docbook...
[04:13] ChrisHJW_log ([email protected]) left irc: Read error: 110 (Connection timed out)
[04:20] The_Company ([email protected]) joined #gstreamer.
[04:20] Company ([email protected]) left irc: Nick collision from services.
[04:21] Nick change: The_Company -> Company
[04:31] ChrisHJW_log ([email protected]) joined #gstreamer.
[04:41] ChriHJW_log ([email protected]) left irc: Read error: 110 (Connection timed out)
[04:50] <taaz> hmm... cleaned up build stuff to try again and it's been using 2.4GHz of cycles for waaay too long on pwg.ps :(
[04:50] jcsston ([email protected]) joined #gstreamer.
[05:01] charlie_ ([email protected]) joined #gstreamer.
[05:11] mxpxpod ([email protected]) joined #gstreamer.
[05:16] nhudson ([email protected]) joined #gstreamer.
[05:17] lilo_booter ([email protected]) left irc: Read error: 110 (Connection timed out)
[05:22] taf2 ([email protected]) left #gstreamer ("Leaving").
[05:23] mxpxpod ([email protected]) left #gstreamer ("Client exiting").
[06:01] charlie__ ([email protected]) joined #gstreamer.
[06:17] charlie_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[06:39] thaytoo ([email protected]) left irc: "leaving"
[07:01] charlie_ ([email protected]) joined #gstreamer.
[07:10] charlie__ ([email protected]) left irc: Read error: 60 (Operation timed out)
[07:36] Marsupilami23 ([email protected]) joined #gstreamer.
[07:40] Nick change: charlie_ -> lilo_booter
[07:56] ChrisHJW ([email protected]) joined #gstreamer.
[07:59] Marsupilami23 ([email protected]) left irc: "I don't wanna grow up, I'm a Toys R Us Kid. There's a million toys at Toys R Us that I can play with!"
[08:01] lilo_booter ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[08:01] lilo_booter ([email protected]) joined #gstreamer.
[08:10] walters ([email protected]) left irc: "z"
[08:13] aeyakovenko ([email protected]) joined #gstreamer.
[08:16] <aeyakovenko> during gst_elment_link does the gst_somelement_connect get called?
[08:19] thaytan ([email protected]) joined #gstreamer.
[08:20] <aeyakovenko> and does the GstCaps have the GstProps of the element that it belogs to?
[08:25] somex1 ([email protected]) joined #gstreamer.
[08:26] <aeyakovenko> oh ok, i think i get it, pads are defined by the element, so they can have whatever poperties the elment wants them to have
[08:26] <aeyakovenko> whats up with the website?
[08:28] Nick change: mathrick|sleep -> mathrick
[08:28] <mathrick> mornin
[08:44] jcsston ([email protected]) left irc: "VDubMOD watchout TCME is coming!"
[08:47] <aeyakovenko> morning?  were are you at
[08:49] <mathrick> aeyakovenko: poland
[08:49] <mathrick> aeyakovenko: and you?
[08:49] <aeyakovenko> mathrick: cool, i used to live near by, i am in sunny san diego, california, but i am from odessa, ukraine
[08:51] <aeyakovenko> hehe, i guess sort of near by
[08:51] <mathrick> aeyakovenko: ah, neighbourhood indeed :)
[08:53] <aeyakovenko> mathrick: i have an idea of how to implement this wrapper.  its not ideal to what rtp element should be, but it fits the library that i am using.  i am going to make three elements, an jrtplibmux, jrtplibsrc, and jrtplibsink, the mux will do all the rtp stuff, but linking the src and the sink elements to it will give you access to the jrtplib functions 
[09:04] lilo_booter ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[09:05] grub_booter ([email protected]) joined #gstreamer.
[09:13] <aeyakovenko> time to sleep
[09:13] aeyakovenko ([email protected]) left #gstreamer.
[09:16] kmaraas ([email protected]) left irc: "Leaving"
[09:20] Nick change: lilo -> levin
[09:20] Nick change: levin -> lilo
[09:21] Nick change: lilo -> levin
[09:21] Nick change: levin -> lilo
[09:28] ChrisHJW ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[09:31] markey ([email protected]) joined #gstreamer.
[09:32] somex1 ([email protected]) left #gstreamer.
[09:33] Nick change: KA_afk -> KA
[09:40] markey ([email protected]) left irc: "bbl"
[09:52] <KA> some time ago I googled for info about a equalization plugin. is someone working on it or have some ideas? I've found only a schema for eq+gain with ladspa and some discouraging thread on lists
[09:57] AlexCONRAD ([email protected]) left irc: "Leaving"
[09:58] KoRnouille ([email protected]) joined #gstreamer.
[10:02] charlie_ ([email protected]) joined #gstreamer.
[10:02] <mathrick> KA: yep, jimmy_dean is
[10:02] <mathrick> bbl
[10:02] Nick change: mathrick -> mathrick|Uni
[10:02] Nick change: charlie_ -> lilo_booter
[10:03] jdahlin ([email protected]) joined #gstreamer.
[10:04] thomasvz ([email protected]) left irc: Read error: 60 (Operation timed out)
[10:16] <KA> jimmy_dean: ping
[10:20] grub_booter ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:25] markey ([email protected]) joined #gstreamer.
[10:32] thomasvs ([email protected]) joined #gstreamer.
[10:38] markey ([email protected]) left irc: "yuck.. eating! can't have that"
[10:38] Kaetzchen ([email protected]) joined #gstreamer.
[10:42] alley_cat ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:45] Nick change: Kaetzchen -> alley_cat
[10:53] dolphy ([email protected]) joined #gstreamer.
[11:02] charlie_ ([email protected]) joined #gstreamer.
[11:11] Shoragan ([email protected]) joined #gstreamer.
[11:20] lilo_booter ([email protected]) left irc: Read error: 110 (Connection timed out)
[11:23] charlie__ ([email protected]) joined #gstreamer.
[11:30] Company ([email protected]) left irc: Remote closed the connection
[11:38] charlie_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[11:47] <thaytan> evening, all
[12:00] markey ([email protected]) joined #gstreamer.
[12:01] Nick change: markey -> markey|afk
[12:01] Nick change: charlie__ -> lilo_booter
[12:10] <jdahlin> Is it possible to do only demuxing with the vorbisfile plugin for 0.6?
[12:19] <thomasvs> jdahlin: nope
[12:25] ensonic ([email protected]) joined #gstreamer.
[12:31] <ensonic> can anyone please tell me whats the value of 'loaderdir' in gst-plugins/ext/gdk_pixbuf/Makefile, here is seems to be undefined, so 'make install' fails
[12:33] Company ([email protected]) joined #gstreamer.
[12:37] sublett ([email protected]) joined #gstreamer.
[12:38] alley_cat ([email protected]) left irc: "May the Source be with you!"
[12:52] <ensonic> or is the one who is doing 'gst-plugins/ext/gdk_pixbuf' here, current cvs failes to build
[12:54] <ensonic> it fails in mgstgdkanimation.c
[12:55] <bitshifter> I just built from (anonymous) cvs 10 minutes ago, and it built fine here
[12:56] <ensonic> it obiously misses some includes, as gcc say implicit declaraion of GST_CAPS_NEW, ...
[12:57] <ensonic> bitshifter: was the gst-plugins/ext/gdk_pixbuf/gstgdkanimation stuff build as well?
[12:57] <bitshifter> yep
[12:58] <bitshifter> wait
[12:58] <ensonic> bitshifter: and was it build withpout warning (gcc runs her per default of autogen.sh with -Werror)
[12:59] <Company> ensonic: that thingy isn't built by default
[12:59] <Company> ensonic: did you decide to explicitly build it?
[12:59] <ensonic> Company: I have not enabled it
[13:00] <ensonic> I run autogen.sh --prefix=... --disable-ffmpeg
[13:00] <bitshifter> ensonic: sorry. I have the file, but it's not built 
[13:00] <ensonic> then make install failes ask the var loaderdir is uncommented in Makefile, so I removed the comment and then it fail on compilation'
[13:02] <Company> ensonic: we have obviously not fixed that particular part of the code, but i think your Makefile fixes are suboptimal ;)
[13:02] <ensonic> Company: they are not mine, it's because install-loaderLTLIBARRIES is in Makefile
[13:04] <ensonic> i've now commented out the targets and see what happens, i do not yet understand from Makefile.am how they have been generated
[13:04] <ensonic> HAVE_GDK_LOADER must have become true here
[13:06] <Company> hum
[13:06] <ensonic> can you please tell what grep HAVE_GDK_LOADERS in Makefile (any Makefile) says
[13:07] <Company> it's false here
[13:08] <ensonic> means HAVE_GDK_LOADER_FALSE = # and HAVE_GDK_LOADER_FALSE =
[13:08] <ensonic> means HAVE_GDK_LOADER_FALSE = # and HAVE_GDK_LOADER_TRUE =
[13:09] <ensonic> here it is HAVE_GDK_LOADER_FALSE = and HAVE_GDK_LOADER_TRUE = #
[13:09] <ensonic> anyway by uncommenting the targets the make install run trough ...
[13:09] <Company> it is HAVE_GDK_LOADER_FALSE = and HAVE_GDK_LOADER_TRUE = # here
[13:10] <Company> which means it's false
[13:10] <ensonic> gst-register now says that I have 106 plugins with 246 features, lets see if they work ;-)
[13:14] <ensonic> how can I find out what possible audiosink I have?
[13:14] <Company> you either know it or you try every plugin you have ;)
[13:14] <Company> on solaris only esdsink works
[13:15] <ensonic> esdsink does not work - it says "can't open connection to esound server" although esdplay works
[13:15] <Company> is esd running?
[13:17] <Company> i know that noone wrote a solaris audiosink yet :)
[13:17] <ensonic> Company: yes, and when I run gst-launch sinesrc ! esdsink as user (not root as before, user is the owner of the esd proc) I get WARNIN: errorneous pipeline: could not link sinsrc0 to esdsink0
[13:18] <ensonic> when I run gst-launch sinesrc ! spider ! esdsink, it runs but is quiet
[13:20] <Company> try sinesrc ! audioconvert ! esdsink
[13:21] Nick change: markey|afk -> markey|f00d
[13:22] <ensonic> Company: segmentation fault
[13:23] <ensonic> Company: i'll be right back in aminute
[13:30] <ensonic> Company: i am back
[13:30] <Company> where does it segfault?
[13:31] Action: thomasvs is halfway through a wrapper binary for the versioned tools
[13:31] <ensonic> Company: the only diff in the caps of esdsink and sinesrc is that sinesrc has channels=1 and esdsink has channels=2
[13:32] <ensonic> Company: i'll run gdb on it
[13:32] <Company> i know, and the plugin for channel conversions is audioconvert :)
[13:34] <ensonic> segfault is at : gst_atomic_int_add() at gstatomic_impl.h:271
[13:34] <Company> eek
[13:34] <Company> from where is that called?
[13:35] <ensonic> gst_atomic_int_inc, gst_data_ref,gst_pad_event_default_dispatch, ....
[13:35] <dolphy> Company: seems the leek that trow was talking about are more related to flushing..
[13:35] <dolphy> s/leek/leak
[13:36] <Company> dolphy: possible, i didn't look at it :)
[13:36] <Company> ensonic: that sounds bad, especially because it's solaris dependant assembler code...
[13:37] <ensonic> Company: there is asm code in there ?
[13:37] <Company> ensonic: yeah, gstatomic_impl.h has atomic refcounting implementations in assembler - one of them in sparc
[13:38] <Company> ensonic: you could probably comment it out and use the default one with mutexes if you want a quick but dirty and slow fix
[13:38] <ensonic> Company: is the one who has done it still alive ?
[13:38] <Company> ensonic: probably, though i dunno who and where ;)
[13:39] <ensonic> Company: probably I am going to uncomment it as I just want to use this installation for testing
[13:39] <Company> should work but be slow
[13:39] Action: Company is off to lunch now though
[13:39] Company ([email protected]) left irc: Remote closed the connection
[13:40] mathrick|Uni ([email protected]) left irc: Remote closed the connection
[13:48] alley_cat ([email protected]) joined #gstreamer.
[13:50] Rotty ([email protected]) joined #gstreamer.
[14:03] Nick change: markey|f00d -> markey
[14:07] markey ([email protected]) left irc: "leaving"
[14:31] nhudson ([email protected]) left #gstreamer ("Leaving").
[15:06] <jimmy_dean> KA: pong
[15:07] <KA> jimmy_dean: mathrick said me you are working on an equalizer filter, aren't you?
[15:07] <KA> or thinking about it :)
[15:07] <jimmy_dean> KA: yes I am, though I haven't touched it for about a week and a half
[15:07] <jimmy_dean> KA: I tried porting the LADSPA algorithm to a GStreamer native plugin
[15:08] <jimmy_dean> KA: it compiles but it segfaults right now
[15:08] <KA> so it's from the schema on the web site
[15:08] <jimmy_dean> KA: yeah, the mbeq plugin
[15:08] <KA> ah, mbeq, well
[15:09] <KA> have you thinked on a bonobo interface too?
[15:09] <jimmy_dean> KA: the LADSPA author gave me some suggestions...he recommended I try and implement a bandpass-based equalizer which would be much simpler
[15:10] <jimmy_dean> KA: well not initially but I did create a test GUI for it for Rhythmbox
[15:10] <jimmy_dean> KA: longterm solution is probably a bonobo control, yes
[15:12] <jimmy_dean> KA: are you wanting to help out?
[15:12] <jimmy_dean> KA: I could use some help actually...I don't care how this thing gets done but I really want an equalizer in GStreamer
[15:17] <KA> I want it too
[15:18] <KA> yes, I can help
[15:18] <KA> my knowledge about gst and gnome is really low, I'm studing them since the last days
[15:19] <jimmy_dean> KA: hey, this was my first programming experience with both also
[15:19] <jimmy_dean> KA: I learned the most by studying the Rhythmbox code
[15:19] <KA> iea:)
[15:19] <KA> me too
[15:19] <KA> rb and gst-plugin tarball
[15:19] <KA> so, no slowdown problem for you, I suppose. well
[15:19] <jimmy_dean> KA: do you have any idea how an equalizer works, how to implement it?
[15:23] markey ([email protected]) joined #gstreamer.
[15:23] <KA> a bit, I'm an ecleptic programme ;), if you have doc about it let me read it :)
[15:25] <jimmy_dean> KA: no docs, sorry...I just have basic knowledge in theory since I'm an electrical engineering student
[15:25] <jimmy_dean> KA: I might be able to write a doc at least of the knowledge that I have
[15:25] <jimmy_dean> KA: but that wouldn't be something to risk the company on :)
[15:26] <KA> I'll try to ask some info to some ham-radio person
[15:26] <KA> probaly they've more bg :)
[15:26] <jimmy_dean> KA: you think so?  I never thought of that...why do you say that?
[15:29] <ensonic> jimmy_dean: KA: for an EQ use 1 low-shelf, <n> band-pass filters, 1 high shelf, google for a filter-cookbook (there are some post around with just the formulas for the filters)
[15:29] <KA> actually not merely for they ham-passion, anyway to have the radio license you'd have to know how bandwith filter know, among all the rest. it's only an idea
[15:30] <KA> how bandwith filter works...
[15:30] <jimmy_dean> ensonic: are those GStreamer plugins right now or are you just suggesting in general?
[15:31] <ensonic> jimmy_dean: i've just started with gstreamer a few days ago as well, anyway you are better of with looking up the formulars and code them into one plugin instead of chaining 10 separate filter plugins ;-) (dunno if there are any filter in gstreamer already)
[15:32] <KA> jimmy_dean: is mbeq in cvs?
[15:32] <ensonic> jimmy_dean: filters are easy to implement (just a loop with multiplications and sum), hard part is to calculate filter coeficients (therefore the cookbook)
[15:32] <jimmy_dean> ensonic: ahh gotcha...well I might search on filter cookbook but I already googled for equalizer designs without anything
[15:32] <jimmy_dean> ensonic: yeah, I know how it goes...this is all really basic circuit theory
[15:33] <ensonic> jimmy_dean: use google groups and or have a look at www.musicdsp.com (org?)
[15:33] <jimmy_dean> haven't seen that site, no
[15:33] <jimmy_dean> .org works
[15:34] <jimmy_dean> ensonic: that's a big help actually, thanks
[15:36] <KA> true
[15:36] <bitshifter> maybe the equalizer code in xmms is worth looking at as well
[15:38] <jimmy_dean> bitshifter: nope, it's not...it is an extremely poor and limited design that only works for mpeg data...the author told me himself
[15:41] <bitshifter> ah, I didn't know that
[15:42] <jimmy_dean> bitshifter: yeah, I found that out...they're not the nicest people in that chatroom :)
[15:45] C ([email protected]) left irc: Remote closed the connection
[15:45] <bitshifter> jimmy_dean: I know one of the bmp guys (xmms fork for gtk2) has done a new equalizer, but probably it has similar problems. Might not hurt to talk to him though
[15:46] <jimmy_dean> bitshifter: is that what was called beep?
[15:46] <bitshifter> yes
[15:46] <jimmy_dean> bitshifter: do they have an IRC room?
[15:46] <alley_cat> the beep equalizer works for all kinds of audio data, at least that's what's on the website
[15:46] <bitshifter> jimmy_dean: #bmp on irc.chatjunkies.org
[15:46] <alley_cat> only the vorbis plugin has been ported to it though
[15:47] <bitshifter> jimmy_dean: I don't think the new equalizer is in the code base yet, I think it's in a separate CVS branch
[15:47] <jimmy_dean> bitshifter: ok
[15:47] <jimmy_dean> alley_cat: yeah, that's the problem with their design...it's not a write once work with any data-type type of thing
[15:48] <ensonic> bitshifter: eq's in a mp3 player tend to reuse the alreay spluitted frequency bands, an eq for gstreamer would have to do the splitting itself
[15:48] <alley_cat> the new one is
[15:48] <alley_cat> but it uses a different plugin system, hence the porting need
[15:48] <jimmy_dean> alley_cat: ahh ok
[15:48] <alley_cat> the old xmms plugin system doesn't allow it
[15:48] <jimmy_dean> alley_cat: why doesn't the author just use GStreamer
[15:48] <alley_cat> no idea :)
[15:49] <alley_cat> that's what i will do when i finally get around to rewrite cplay
[15:49] <jimmy_dean> alley_cat: sounds like a bunch of duplicate effort that will probably be greyed out in the end by GStreamer based stuff
[15:49] <alley_cat> well, so far no one has written a gstreamer based player in the style of xmms
[15:50] <alley_cat> only those library style things
[15:50] <jimmy_dean> alley_cat: I know, I don't like that style anymore...unless they made it as nice as the new Winamp 5
[15:50] <jimmy_dean> alley_cat: old winamps and xmms are tacky now
[15:50] <alley_cat> all i want is a player with a playlist i can drop files/directories to from rox
[15:51] <alley_cat> libraries don't work if half of your stuff is wrongly tagged or not tagged at all
[15:51] <jimmy_dean> alley_cat: yeah, you're right
[15:51] <bitshifter> jimmy_dean: btw, I think 'larne' is the person to talk to in #bmp 
[15:51] <jimmy_dean> bitshifter: ok, thanks!
[15:52] <jimmy_dean> bitshifter: I'll ask later this evening...I've got to get ready to head down to school
[15:53] <jimmy_dean> KA: we'll talk more about this later...I'm excited to get some help on this....this will finally happen :)
[16:06] <KA> sure, use kalfa at bononia.it if you prefer (have doc, mail to put me in CC or to bounce :). I'll take a look into your code, is it in CVS? I cannot see it
[16:06] <KA> I've to go back to home
[16:10] Nick change: KA -> KA_afk
[16:10] <KA_afk> bye
[16:11] <jimmy_dean> KA_afk: no, it's not in CVS...it really doesn't do anything yet to be in CVS
[16:14] <hyriand> alley_cat: I'm making an xmms-like player with gstreamer backend
[16:15] <hyriand> alley_cat: well, not really XMMS-like, but a 'pretty skinned gui' player
[16:15] <alley_cat> hmm, skins
[16:15] <alley_cat> that's the part about xmms i don't like :P
[16:16] <hyriand> *sigh*
[16:16] <hyriand> :)
[16:16] <alley_cat> i added drag and drop support the the ruby-gstreamer simple-audio-player example, that was close to what i want
[16:16] <hyriand> anyway, it's more than just a gui.. it's a gstreamer-backend daemon / gui separated media player
[16:17] <alley_cat> but now i found cplay which basically does what i need, but has way too many bugs to not replace it some time
[16:17] <alley_cat> using what kind of communication? d-bus or bonobo?
[16:17] <jimmy_dean> use dbus
[16:17] <hyriand> atm, a custom crafted unix-socket protocol
[16:18] <jimmy_dean> dbus is a great standard
[16:18] <alley_cat> d-bus lacks activation support, that's one nice feature of bonobo
[16:18] <jimmy_dean> yeah, it's very new still though
[16:19] <jimmy_dean> but very useable
[16:19] <hyriand> and still pretty immature
[16:19] <hyriand> and the API hasn't been fixated yet
[16:19] <jimmy_dean> yeah, but that's not that big of a deal...the basics are pretty solid by now
[16:20] <jimmy_dean> the benefits of using it and the possibility of communicating with for example Project Utopia by Robert Love is the big benefit
[16:20] <hyriand> true, but writing an app that has dependancies like "gstreamer-cvs > 20040201 and D-BUS-cvs > 20040108" is kind of not cool
[16:21] <jimmy_dean> hyriand: yeah, you're probably right there
[16:21] <jimmy_dean> hyriand: d-bus will be officially supported in GNOME 2.6 though so I would assume it's going to be somewhat API stable by then
[16:21] <hyriand> which is why I've stalled development for now
[16:23] <hyriand> but it sort of works already, I use it myself all the time
[16:23] <jimmy_dean> very cool
[16:24] Company ([email protected]) joined #gstreamer.
[16:26] <hyriand> don't you just love the way keypresses get repeated under heavy load? or am I the only one suffering from that?
[16:26] <jimmy_dean> hyriand: I don't get that :)
[16:26] Action: hyriand kicks his kernel
[16:27] <jimmy_dean> hyriand: do you have DMA turned on on your hard disk?
[16:27] <hyriand> yes
[16:27] <jimmy_dean> hyriand: no idea then :)
[16:28] sublett ([email protected]) left irc: "I like food, food is good!"
[16:32] <hyriand> it'll probably come down to killing linus or andrew for those atkbd changes :)
[16:33] trow ([email protected]) joined #gstreamer.
[16:42] Nick change: markey -> markey|afk
[16:51] markey|afk ([email protected]) left irc: "bbl"
[16:52] jcsston ([email protected]) joined #gstreamer.
[17:19] Nick change: KA_afk -> KA
[17:22] sublett ([email protected]) joined #gstreamer.
[17:22] <ensonic> arg, on gstreamer (cvs HEAD), make install I not get : creating gstreamer-scan, GLib-ERROR **: The thread system is not yet initialized
[17:22] <ensonic> in gstreamer/docs/gst
[17:22] <ensonic> what the heck is this now
[17:24] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[17:24] <Company> run with --disable-docs-build
[17:25] <Company> the problem is that the doc build only runs g_type_init and not gst_init
[17:28] <ensonic> Company: I've started to rebuild 6 hours ago, I can't reautoscan again, I only want to try the stuff and it was compiling yestercday
[17:28] <ensonic> so is it new stuff, can I fix it locally
[17:29] <Company> no idea
[17:29] <Company> it's probably because of the default atomic refcounting
[17:30] <Company> which uses mutexes and requires an initialised threading...
[17:30] <ensonic> should i file a bug report for this?
[17:30] <Company> not a bad idea
[17:33] <thomasvs> ensonic: edit the Makefile on dir higher up and remove the docs dir from the subdirs
[17:34] <ensonic> thomasvs: done exactly that after biting a little chunk out of my table ;-)
[17:34] <Company> thomasvs: the -scan.c file is generated automatically, right?
[17:36] <thomasvs> Company: yep
[17:36] <thomasvs> Company: do you have some time to test out wrapper code I wrote for the tools so that they can run unversioned ?
[17:36] <thomasvs> Company: I'm wondering whether or not they're ok and if I should include them in the release
[17:38] <ensonic> with default mutexes instead of the sparc asm ones I now get a 'bus error' and not a 'segmentation fault' - is that better ;-)
[17:40] <ensonic> backtrace is : gst_structure_free / _gst_event_free / gst_data_unref / ... this time
[17:42] kmaraas ([email protected]) joined #gstreamer.
[17:42] Action: Company assumes solaris just plain sucks :p
[17:43] ain ([email protected]) joined #gstreamer.
[17:43] <ain> I'm back :D
[17:44] <Company> Weeeian is back :)
[17:45] <ain> the central heating repair man managed to blow out my cable modem
[17:45] <ain> and its taken me a week to get a new one
[17:45] <Company> thomasvs: i don't think i'm the right person to test those - i don't run 0.6
[17:45] <ain> but it came with a free wireless card, so I've got WLan in the house too now :)
[17:46] <Company> wlan in house rocks
[17:46] <ensonic> Company: I am now back with def mutexes
[17:46] <ensonic> Company: but with default mutexes instead of the sparc asm ones I now get a 'bus error' and not a 'segmentation fault' - is that better ;-)
[17:46] <ensonic> Company: backtrace is : gst_structure_free / _gst_event_free / gst_data_unref / ... this time
[17:46] <Company> probably not...
[17:47] <Company> what pipeline do you run?
[17:47] <ensonic> Company: sinesrc ! audioconvert ! esdsink
[17:47] <ain> Company: so how much stuff in GStreamer has been changed ni the last week and how long am I going to have to spend to get marlin working with it :p
[17:48] <Company> ain: we have Changelogs now :p
[17:48] <Company> ain: it should still work, but i don't know your voodoo-fu, so that doesn't say much
[17:53] <ensonic> There is an illegal memory access: GstStructure->fields points to 0xffffffff
[17:54] <Company> the question is: why does it try to free the structure
[17:54] <Company> hm
[17:54] yippi ([email protected]) joined #gstreamer.
[17:54] <Company> there should be break;'s between the cases
[17:55] <ensonic> Company: maybee because I run gst-inspect with -v to see something
[17:56] <Company> ensonic: in _gst_event_free, try putting break; between the case's
[17:58] <ensonic> between both?
[17:58] <Company> yeah
[17:58] <ensonic> forget about it ;-)
[17:59] <ensonic> Company: compiling again, may take a few minutes ;-)
[18:03] <ensonic> Company: it is making beeeeeep now, YEAH
[18:03] <ensonic> Company: can you please comiit the break's to CVS so that I don't have to file another bugzilla entry
[18:04] Shoragan ([email protected]) left irc: "Leaving"
[18:07] <ensonic> my next test was filesrc with mp3 ! mad ! esdsink and all that came our was noise, madplay in contrast played fine
[18:07] <Company> ensonic: i already have them here locally, so they'll be in on next commit
[18:08] <Company> are you on 32bit or 64bit sparc?
[18:12] <ensonic> Company: Ultra60 (Ultrasparc) the CPU is capable o 64, my Solaris (2.6) is 32 bit due (no time to update)
[18:13] somex1 ([email protected]) joined #gstreamer.
[18:15] <Company> i was wondering if it's a int size issue somewhere
[18:15] <Company> because it can't be endianness, as ppc works fine
[18:17] <KA> jimmy_dean: well, where I can download the source? can you send me a tarball?
[18:17] <ensonic> Company: that shoundn't be that case, I do a lot of development on thid box (lots of c stuff, glib, gtk, gnome, xml,...)
[18:18] <Company> ensonic: the problem might be raw data handling somewhere inside gstreamer
[18:21] Uraeus ([email protected]) joined #gstreamer.
[18:22] <Uraeus> hello
[18:22] markey ([email protected]) joined #gstreamer.
[18:22] <ensonic> Company: I now tried : filesrc location=xxx.wav ! audiofile ! esdsink and I can hear a bit of sound burried under noise
[18:23] <Company> ensonic: try wavparse
[18:24] <Company> i need a faster machine
[18:24] <Company> i can'T valgrind mp3s in realtime
[18:28] <ensonic> Company: when I do : filesource ! wavparse ! audioconvert ! esdsink - it sounds good
[18:29] <Company> well, at least one format that plays...
[18:30] <Company> does ffdec_mp3 instead of mad work by chance?
[18:30] <ensonic> Company: I needed to disable ffmpeg
[18:31] <Company> oh
[18:31] <Company> that's bad
[18:31] <ensonic> I have the ogg stuff, i'll that in a minute
[18:35] <ensonic> how can I find out how the ogg-decoder is called
[18:36] <Uraeus> thomasvs: seems docs are built now no matter what if I have the needed libs installed, seems to be no way to disable it (we used to have to enable it)
[18:37] <trow> What is the easiest way to use gstreamer to inspect the id3 tags on an mp3.  Is there a pipeline that will do it?
[18:38] sub_pop ([email protected]) joined #gstreamer.
[18:40] jcsston ([email protected]) joined #gstreamer.
[18:41] jcsston` ([email protected]) joined #gstreamer.
[18:41] <Company> trow: gst-launch -t filesrc ! spider ! osssink should do that
[18:41] jcsston ([email protected]) left irc: Nick collision from services.
[18:41] Nick change: jcsston` -> jcsston
[18:41] jcsston` ([email protected]) joined #gstreamer.
[18:41] jcsston` ([email protected]) left irc: Client Quit
[18:41] jcsston ([email protected]) left irc: Client Quit
[18:42] jcsston ([email protected]) joined #gstreamer.
[18:43] <trow> Company: thanks
[18:45] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[18:52] <thomasvs> Uraeus: could be, yeah
[18:54] <thomasvs> ok, I'll just commit the wrapper binary
[18:54] <thomasvs> I want it tested anyway
[18:54] <ensonic> how can I query the plugin registry to e.g. show me all plugins that e.g. generate audio/*
[18:55] <Company> not at all (unless you write an app to do that
[18:55] <Company> )
[18:56] <ensonic> Company: I am probaly doing tat as an exercise
[18:56] <ensonic> Company: by any chance do you know how the oggdecoder is called ?
[18:56] <Company> oggdemux
[18:56] <Company> or vorbisfile
[18:57] <Company> oggdemux ! vorbisdec ! float2intnew is the new replacement
[18:58] <ensonic> hmm, that does just one iteration and ends
[19:00] <Company> which one?
[19:00] <ensonic> oggdemux ! vorbisdec ! float2intnew
[19:01] <ensonic> when I use just vorbisfile then I get errors, stream is of different type
[19:02] <Company> then use an ogg file ;)
[19:05] <ensonic> Company: it is one
[19:05] <Company> are you sure?
[19:06] <ensonic> Company: the ogg-files have id3tags, is that a problem?
[19:06] <Company> yes, it is
[19:06] <Company> you need to put an id3tag in front in that case
[19:07] <Company> ((people putting id3tags in front of oggs should be forced to peel the bits of by hand btw)
[19:08] <ensonic> Company: it wasn't me ;-), anyway I have only id3types not id3tag and using id3types it runs but silent
[19:09] <Company> peel em off by hand :p
[19:09] <Company> or install libid3tag and compile the id3tag plugin
[19:09] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[19:13] somex1 ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[19:13] <ensonic> Company: do you knwo from where I can get the sources (only find rpms)
[19:15] <bitshifter> ensonic: google for underbit and mad audio
[19:15] <ds-work> /usr/bin/install: no se puede efectuar `stat' sobre `./html/index.sgml': No existe el fichero o el directorio
[19:15] <ensonic> erm already found it, mad is installed, but the lib not yet
[19:17] <ensonic> ... installing it
[19:18] Action: ds-work looks at thomasvs
[19:23] <ensonic> if the registry can be queried by features, how does an app provide a list of e.g. audiosinks to the user for choosing one
[19:24] <hyriand> my guess would be: they can sink audio/raw and don't provide a source
[19:24] Zeenix ([email protected]) joined #gstreamer.
[19:24] <Zeenix> hi
[19:24] <Company> my guess would be they take all elements from the class sink/audio
[19:25] <ensonic> Company: can you name a method in the api to get me on the right track for reading?
[19:25] <thomasvs> ds-work: why are you getting that in spanish ?
[19:25] <hyriand> my guess would be Company knows better than me :)
[19:26] <ds-work> thomasvs: LANG=es_US.UTF-8, probably
[19:26] <thomasvs> ds-work: could you sanity-check my gst-run code I just added ?
[19:26] <thomasvs> ds-work: I'm about to make a prerelease
[19:26] <thomasvs> ds-work: I'm assuming you want me to fix the doc build, if you do then some context would be nice
[19:27] <ds-work> ah, yes
[19:27] <ds-work> make install in docs/gst
[19:28] <thomasvs> ds-work: ok, could you send me make output after running make clean there so I can check it ?
[19:28] <ds-work> gst-launch --help doesn't tell me how to get 0.6
[19:28] <thomasvs> ds-work: I'm probably going to do a 0.6 service release that also installs versioned
[19:29] <thomasvs> ds-work: I didn't want to add a bunch of hacks just to try and figure out if you have some oldstyle 0.6 unversioned binaries
[19:29] <ensonic> thomasvs: do know how one could find all plugin-nbames for e.g. class sink/audio
[19:29] <thomasvs> ds-work: esp, since it would still just work
[19:29] <thomasvs> ensonic: check gst-player source, it has code to do that to get all of the visualisation effects
[19:30] <ensonic> thomasvs: thx
[19:30] <thomasvs> ds-work: feel free to tell me if the way I put it together is a good idea or not
[19:30] <ds-work> I have mostly versioned 0.6 binaries
[19:31] <thomasvs> you do ?
[19:31] <ds-work> Debian
[19:32] <thomasvs> hang on, you sure you rebuilt gst-launch from the new gst-run.c ?
[19:32] <ds-work> ok, gst-register
[19:32] <thomasvs> if you do --help, it should show you a few added flags at the top
[19:32] <thomasvs> one of which is --gst-list-mm
[19:32] <thomasvs> to list the m/m's for this tool
[19:32] <thomasvs> another is --gst-mm
[19:32] <thomasvs> to force a version for that tool
[19:32] gaisca ([email protected]) joined #gstreamer.
[19:32] <thomasvs> by default, it will use the most recent one it can find
[19:32] <ds-work> --gst-mm?
[19:33] <trow> There is a problem w/ tags, apropos bug 132519.
[19:33] <ds-work>  /usr/local/bin/gst-register --gst-list-mm
[19:33] <ds-work> should probably quit after listing the versions
[19:33] <trow> Rhythmbox uses the pipeline gnomevfssrc ! typefind ! spider ! application/x-gst-tags ! fakesink to discover tags
[19:33] <ensonic> many thanks to all who helped me so far!, see you tomorrow
[19:33] <trow> Which doesn't work on ID3v1 tags at the end of mp3 files.
[19:34] ensonic ([email protected]) left #gstreamer.
[19:34] <trow> In 132519, Ben Otte comments that spider is "very suboptimal" for detecting stuff at the end of a file.
[19:34] <thomasvs> trow: ben = Company
[19:34] <trow> Is this a spider bug, or should rb be using a different pipeline for that?
[19:34] <ds-work> --gst-mm doesn't seem like a good name for the option
[19:34] <thomasvs> ds-work: I'm open for suggestions, not sure yet either
[19:35] <thomasvs> ds-work: though it's mostly a packaging help
[19:35] <thomasvs> ds-work: something that's preferably not too long
[19:36] <ds-work> seems to work here
[19:36] <trow> thomasvs: Ah, thanks.
[19:36] <thomasvs> trow: we've had an unresolved discussion about whether or not tag elements should be allowed to seek at the end for id3 for example
[19:37] <thomasvs> trow: I'm not really sure why we're unable to resolve that discussion though :)
[19:38] <trow> thomasvs: What is the controversy?
[19:38] <thomasvs> ds-work: do you remember why it was bad for an element to seek to the end ?
[19:38] <thomasvs> trow: I think it was for nonseekable streams, eg internet radio
[19:38] <thomasvs> trow: but I wasn't convinced that was an issue, really
[19:38] <thomasvs> trow: feel free to bring it up on the ML for discussion though
[19:38] <trow> thomasvs: But what is the issue for files, which are certainly seekable.
[19:38] Action: trow thinks about his 30gb of id3v1-tagged mp3s.
[19:38] <ds-work> thomasvs: it's not bad.  you just need to make sure to not pull at the end, so you don't get the EOS
[19:39] <thomasvs> ds-work: so, there's no real argument against letting id3tag seek to the end to where an id3 tag might be ?
[19:39] <thomasvs> ds-work: how do you avoid pulling at the end ? :)
[19:40] <ds-work> seek to end-100 bytes, and pull once
[19:41] <ds-work> the first will be a 100 byte buffer
[19:41] <ds-work> the second pull will be EOS
[19:42] <thomasvs> trow: sounds like something you want to take a stab at ?
[19:44] <trow> thomasvs: I'm willing, but I'd need some hand-holding.
[19:45] <thomasvs> trow: let me check the source for a bit
[19:45] dilinger ([email protected]) left irc: Read error: 113 (No route to host)
[19:45] <trow> thomasvs: If you can just give me a quick sketch of what needs to be done, I'll give it a try.
[19:46] <trow> It is a pretty big showstopper for rhythmbox.
[19:46] <thomasvs> trow: well, first of all, I'm not completely sure where the problem is
[19:46] <thomasvs> trow: agreed, it needs fixing
[19:46] <thomasvs> trow: so, when using spider it's probably plugging mad instead of id3tag
[19:46] <thomasvs> hm
[19:46] <thomasvs> no
[19:46] <thomasvs> that's not correct
[19:47] <thomasvs> do you happen to see what spider plugs in that pipeline ?
[19:47] Action: thomasvs doesn't have mp3's to test on
[19:47] <thomasvs> or, if you could give me an mp3 with id3 at the end, I can reproduce and check
[19:48] <taaz> that gst-run code looks complex
[19:48] <taaz> i can't tell what it does just by looking at it ;)
[19:48] <thomasvs> taaz: it's commented ok no ?
[19:48] <thomasvs> taaz: quick tip - start in main () :)
[19:49] Action: thomasvs really hates how long gst-plugins takes to autogen
[19:49] Action: thomasvs really is going to do *something* about that
[19:49] <taaz> btw, taking out some of that sed stuff in the makefile for the man pages and such might break things if people actually try to use the program suffix configure option
[19:51] <ds-work> thomasvs: we could consolodate a lot of directories
[19:51] <ds-work> we have a habit of 1 plugin per directory, which is insane from an automake perspective
[19:52] <taaz> there's something wrong with "let's adapt to the tools" rather than "let's adapt the tools to us"
[19:52] <trow> thomasvs: When I do a gst-launch -t w/ the rb tag-discovery pipeline on an mp3 w/ the tag at the beginning, it says: found by element "id3tag0"
[19:53] <trow> taaz: Unfortunately, adapting the tools is non-trivial. :/
[19:57] <thomasvs> ds-work: I went halfway through implementing that
[19:57] <thomasvs> ds-work: but it is VERY boring work, and I gave up
[19:57] <thomasvs> ds-work: but it would increase speed a lot, since automake expansion is O(2) afaik
[19:57] <thomasvs> ds-work: and anyways, since I'm arguing for splitting up, it wouldn't make my point quite as well :)
[19:58] <thomasvs> trow: but with an mp3 with id3 on the end, what does it do ? just nothing ?
[19:58] Action: thomasvs deletes all fake gst-launch-0.5 commands from path
[19:59] <trow> thomasvs: that is right, it does nothing
[20:01] <thomasvs> trow: do you know how the debugging system works ?
[20:01] <thomasvs> trow: I can explain so you can help yourself figuring out some things
[20:05] <trow> thomasvs: No, I don't know how it works.  Anything you can tell me would be great.
[20:07] <thomasvs> trow: ok, it's not too hard, benjamin made it really good
[20:07] <thomasvs> so, any gst app can be run with --gst-debug
[20:08] <thomasvs> run gst-launch with --gst-debug-help for example
[20:08] <thomasvs> it will print out a whole bunch of categories
[20:08] <thomasvs> so, suppose you want to debug how spider works
[20:08] <thomasvs> you can run with --gst-debug=spider*:5
[20:08] <thomasvs> that will print ALL debug output it generates
[20:08] <thomasvs> :4 makes the system not print out LOG info, which is the sort of thing that gets printed on each iteration
[20:09] <thomasvs> I'm going to use that now to see what spider plugs in for that pipeline
[20:09] <thomasvs> or, I will once I get my stuff built
[20:10] <thomasvs> [gst-cvs] [thomas@otto gst-plugins]$ gst-launch gnomevfssrc location=../small.mp3 ! typefind ! spider ! application/x-gst-tags ! fakesink
[20:10] <thomasvs> RUNNING pipeline ...
[20:10] <thomasvs>  
[20:10] <thomasvs> (process:2846): GStreamer-CRITICAL **: file gstcaps.c: line 395 (gst_caps_get_structure): assertion `index < caps->structs->len' failed
[20:10] <thomasvs>  
[20:10] <thomasvs> (process:2846): GStreamer-CRITICAL **: file gststructure.c: line 267 (gst_structure_get_name): assertion `structure != NULL' failed
[20:10] <thomasvs> ds: there have been some other reports of some struct corruption as well
[20:10] <thomasvs> ds: any idea what oculd have happened ?
[20:10] <ds-work> thomasvs: yes
[20:11] <ds-work> thomasvs: this is all accessing structures after they've been freed
[20:11] <ds-work> there was an accidental 'return;
[20:11] <ds-work> in gst_structure_free()
[20:11] <taaz> thomasvs: sorry to be a pest about this but should i just commit that patch i sent you or post it to list or bugzilla or what?
[20:11] <ds-work> so these bugs were never seen before
[20:12] <ds-work> thomasvs: btw, I couldn't reproduce the doc build problem
[20:12] <thomasvs> taaz: no, you need to kick me repeatedly so I look at it
[20:12] <thomasvs> ds-work: so it's already fixed ? I just need to rebuild core ?
[20:12] pb_ ([email protected]) joined #gstreamer.
[20:12] <Company> typefind ! spider is supposed to detects ID3v1 tags at the end of files
[20:12] <taaz> thomasvs: ok.  i just feel like i'm being annoying by asking over and over ;)
[20:13] <Company> but looking at the debugging a bit it looks as if spider redoes typefinding and ignores typefind
[20:13] <thomasvs> taaz: I'm sure you'll get over that initial feeling :)
[20:14] <bitshifter> Company: I must confess I didn't really understand your follow-up to #132519 
[20:14] Action: thomasvs starts some make distchecks to run while going home
[20:14] <thomasvs> later
[20:14] <bitshifter> Company: because that's exactly what I did in the example pipeline
[20:14] <Company> bitshifter: i overread the "typefind" in the bugreport for some reason
[20:14] <taaz> ds-work: i can't build docs now with latest debian updates... actually, i'm not sure when i last had them build successfully.  docbook2ps fails on pwg.  (i thought docbook2ps was for sgml not xml anyway?)
[20:14] <Company> bitshifter: so ignore that comment :)
[20:14] jdahlin ([email protected]) left irc: "Klienten avslutas"
[20:16] <ds-work> thomasvs: what should we do about deprecated functions?
[20:17] <ds-work> when changing to -0.8?
[20:18] <Company> ds-work: any ideas on why filesrc ! typefind ! spider ! osssink doesn't negotiate typefind ! spider before starting to pass buffers? (do we even check that links are negotiated?)
[20:19] <ds-work> Company: um, no.  it should
[20:20] <Company> typefind does a lot of processing before starting to pass buffers, and it sets explicit caps before starting to pass buffers
[20:20] <Company> i'll have to investigate that...
[20:20] gaisca ([email protected]) left irc: Read error: 113 (No route to host)
[20:22] <Company> ds-work: how do i check if a pad has negotiated "ANY"? does gst_caps_is_any (gst_pad_get_negotiated_caps (pad)) work?
[20:23] <ds-work> it will negotiate application/octet-stream
[20:23] <ds-work> ANY is not fixed
[20:23] <Company> _is_any (get_allowed_caps (pad)) should work though
[20:24] <ds-work> yes
[20:25] <ain> gsta52dec.c:27:21: _stdint.h: No such file or directory
[20:25] <ain> In file included from gsta52dec.c:30:
[20:25] <ain> /usr/include/a52dec/a52.h:51: parse error before "mm_accel"
[20:25] <ain> /usr/include/a52dec/a52.h:53: parse error before '*' token
[20:25] <ain> /usr/include/a52dec/a52.h:55: parse error before "uint8_t"
[20:25] <ain> what version of a52dec do I need?
[20:25] <Company> ain: rerun configure
[20:26] <Company> ain: it should generate _stdint.h
[20:26] Action: Company wishes vektor had included the types he needs...
[20:26] <Company> ...or just used glib
[20:27] <ain> but that would have added dependancies!!!
[20:27] <ds-work> isn't that walken's lib?
[20:29] trow ([email protected]) left irc: "Leaving"
[20:30] mxpxpod ([email protected]) joined #gstreamer.
[20:34] thomasvs ([email protected]) left irc: Read error: 113 (No route to host)
[20:44] <ds-work> Company: what was wrong with set_explicit_caps() in qtdemux.c?
[20:52] <ain> dvdnavsrc.c:34:30: dvdnav/nav_print.h: No such file or directory
[20:52] <ain> and what version of dvdnav does one need?
[20:54] thomasvs ([email protected]) joined #gstreamer.
[20:57] <thomasvs> crap
[20:57] Action: thomasvs now also has doc build problems
[20:57] <thomasvs> what is it with these xml tools ? can't they just run correctly please ?
[20:59] Action: ds-work wonders why vim is randomly segfaulting
[21:07] somex1 ([email protected]) joined #gstreamer.
[21:10] <taaz> ain: that's a more recent livdvdnav change.  (ie, recent as of about a year ago).
[21:11] <taaz> thomasvs: are the docbook2foo tools supposed to work with xml?  or just sgml?
[21:11] <thomasvs> taaz: they work with both afaik
[21:11] <thomasvs> taaz: they build docs fine here
[21:11] <thomasvs> taaz: but I can't figure out how to make them not go on the net
[21:11] <thomasvs> taaz: I'm pretty sure I have everything I need locally, but it still breaks when I don't have net
[21:11] <thomasvs> taaz: exactly what problem are you seeing ?
[21:14] <taaz> docbook2ps either spins forever at 100% cpu or bails with errors i can't understand
[21:14] <taaz> on the pwg
[21:14] <taaz> hitting the net is a catalog issue most likely
[21:15] <thomasvs> taaz: the spin I can't do anything about
[21:15] <thomasvs> taaz: the errors I'd like to see
[21:15] <thomasvs> taaz: can you mail me your /etc/xml/catalog ?
[21:15] <thomasvs> my catalog has unversioned public id's for docbook, which baffles me
[21:15] <taaz> thomasvs: i could but it's not going to be useful ;)  debian just switched to a highly interlinked catalog system.  lots of files involved.
[21:19] <taaz> i think /etc/xml/catalog links to /etc/xml/docbook-xml.xml which links to /usr/share/blahblah/4.2/catalog which then has the mappings from URIs/pubids to actual dtd files
[21:23] <taaz> docbook2ps runs jw which runs jadetex which is off in neverneverland now.  running gdb on it not to useful due to lack of symbols and such
[21:26] mxpxpod ([email protected]) left irc: "Guns kill people as much as spoons made Rosie O'Donnell fat.."
[21:27] Action: taaz confused about all this
[21:35] <thomasvs> taaz: do we have alternatives ?
[21:38] <thomasvs> ==================================================
[21:38] <thomasvs> gstreamer-0.7.3.2.tar.gz is ready for distribution
[21:38] <thomasvs> ==================================================
[21:38] <taaz> in theory xmlto can do this stuff too... that fails for different reasons ;)
[21:39] <thomasvs> taaz: yeah, xmlto has been producing broken latex from the start on the machine
[21:39] <thomasvs> anyway, people like dv and twaugh told me docbook2 and docbook/xml was the wave of the future, so I followed
[21:39] <thomasvs> maybe I should release tarballs with built docs ? :)
[21:43] <thomasvs> taaz: so, how close to a gst-python release are we ?
[21:44] <taaz> update some version #s, maybe see if the docs make sense anymore, then do it
[21:44] <thomasvs> cool
[21:44] <taaz> and include a note that they won't build off uninstalled gst 0.7.x due to unapplied patches ;)
[21:45] <thomasvs> taaz: is the pkgconfig patch the only thing that breaks it ?
[21:45] <taaz> i have no idea if gst-python actually works though ;)
[21:45] <thomasvs> because then I can test tomorrow before releasing
[21:46] <taaz> the gstplay player.py example doesn't (or didn't yesterday) work too good
[21:47] <taaz> [33033.0.17] [33034.0.17] [33035.0.17] [33036.0.17] [33037.0.17] [33038.0.17]
[21:47] <taaz> [33039.0.17] [33040.0.17] [33041.0.17]
[21:47] <taaz> ! TeX capacity exceeded, sorry [pool size=340179].
[21:47] <taaz> docbook2pdf fails like that... prints out maybe 10000 lines of those #s first
[21:47] <taaz> 2ps prints out same thing if you sigkill it while its spinning
[21:47] <taaz> to which i say, blah
[21:48] <taaz> oh well.  i just won't build the pwg i guess.
[21:49] <ds-work> iirc, you need to increase some tex capacities
[21:49] <ds-work> that should be fixed, 'cuz I filed a bug
[21:50] <taaz> i've had to do that in the past.  i remember a twisty maze of incomprehensible vars with values that looked arbitrary.  tex is just weird.
[21:50] ds-gromit ([email protected]) left irc: Read error: 60 (Operation timed out)
[21:53] <taaz> manuals.mak has USERNAME ?= thomasvs... hmm
[21:54] <thomasvs> taaz: isn't that removed ?
[21:54] <thomasvs> taaz: it sure isn't necessary anymore
[21:54] Uraeus ([email protected]) left irc: "Client exiting"
[21:54] <thomasvs> man
[21:54] <thomasvs> I'm starting to think it's impossible to distcheck gst-plugins if gst is uninstalled
[21:56] <ds-work> it should be
[21:58] <thomasvs> possible, you mean ? :)
[21:58] <thomasvs> yeah, reading through some mails detailing some stuff
[21:58] <thomasvs> I guess I should make two test tarballs in order not to pull my hair out with gst-plugins
[21:58] ds ([email protected]) left irc: Read error: 110 (Connection timed out)
[21:59] <thomasvs> haha, that's funny
[21:59] <thomasvs> so I'm reading a havoc mail about gtk-all with the same problem, and a possible solution
[21:59] <thomasvs> second reply is from a guy named erik walthinsen
[22:03] <Zeenix> people! see this if you want to laugh: http://www.microsoft.com/mscorp/facts/default.asp
[22:06] jcsston ([email protected]) joined #gstreamer.
[22:08] Nick change: KA -> KA_afk
[22:14] walters ([email protected]) joined #gstreamer.
[22:17] <Company> hm
[22:17] <Company> would it be useful to have a G_BREAKPOINT in gst_element_error_full?
[22:18] <walters> thomasvs: what's the word?
[22:19] <thomasvs> walters: doing prerels right now
[22:19] <walters> thomasvs: cool
[22:19] <thomasvs> walters: unless people find silly stuff, I finalrelease tomorrow morning
[22:19] <walters> ok.
[22:19] <thomasvs> ds-work: that mem corruption in structs, that's fixed right ?
[22:19] <thomasvs> walters: you have time for doing a rhythmbox release tomorrow ?
[22:19] <walters> thomasvs: yep
[22:20] <walters> thomasvs: let me know if there's anything i can do to help with the gst release
[22:20] <thomasvs> walters: want to fix make distcheck on gst-plugins barfing on libtool running against uninstalled gstreamer ? :)
[22:20] <walters> ok, i'll take a look
[22:21] <thomasvs> walters: no, that was a joke
[22:21] <walters> ah
[22:21] <thomasvs> walters: spare yourself, seriously :)
[22:21] <walters> good :)
[22:21] <walters> libtool is...scary.
[22:21] <thomasvs> I'll make a test case so I can at least have some hope of debugging it
[22:21] <thomasvs> well, it allows us to run uninstalled, which is nice
[22:21] <thomasvs> walters: you could go over gst-run.c though to see if anything stupid is in there
[22:23] <thomasvs> ds-work: aren't I supposed to do something with the libversioning btw ?
[22:23] Action: thomasvs checks mail
[22:26] thaytan ([email protected]) left irc: Read error: 113 (No route to host)
[22:27] <ds-work> thomasvs: huh?
[22:27] <ds-work> Company: what was wrong with set_explicit_caps() in qtdemux.c?
[22:27] <thomasvs> ds-work: got your mail back, n/m
[22:28] <ds-work> thomasvs: no, I thnk there are still use-after-free bugs with structures
[22:28] <thomasvs> ds-work: oops.  any way we can catch them ?
[22:29] Action: thomasvs misses refcounting on caps/props :)
[22:29] <ds-work> they are caught, since structures are poisoned after free
[22:30] <Company> ds-work: autoplugging starts the moment you add the pad to the element
[22:31] <Company> ds-work: so you need to set the explicit caps before doing this to be sure spider gets it right
[22:31] <taaz> blah... all the 0.6.4 debian autobuilt archs fail due to some alsa api issue
[22:31] <thomasvs> Company: so svq3 should be autopluggable now ?
[22:32] <taaz> for those debian users out there, mpeg2dec 0.4.0b got in archives today
[22:32] <ds-work> Company: I added a test in gst-lint, and it catches about 5 more of those
[22:32] <Company> thomasvs: yes
[22:32] Action: ds-work fixes
[22:32] <thomasvs> Company: cool ! will try
[22:33] <taaz> ds-work: is it possible to add gst-lint check to make sure all class and object structs have padding?
[22:33] <thomasvs> ds-work: any chance we can get you to implement seeking in qtdemux ? :)
[22:33] <taaz> seems tricky to write...
[22:33] <ds-work> thomasvs: eventually, yes
[22:33] <Company> and any chance we can get him implement not seeking in qtdemux? ;)
[22:33] <ds-work> taaz: not all objects and structures need padding
[22:34] apoc_ ([email protected]) joined #gstreamer.
[22:35] Action: ds-work notes that ffmpegdemux, id3tag, and mpegdemux all have the explicit_caps problem
[22:36] <Company> id3tag?
[22:36] <ds-work> id3types.c
[22:36] <Company> that one can be deleted
[22:37] <Company> id3tag != id3types
[22:37] <ds-work> typo
[22:37] walters ([email protected]) left irc: "rebooting"
[22:42] apoc ([email protected]) left irc: Read error: 60 (Operation timed out)
[22:47] <Company> phew
[22:47] <Company> got iz
[22:49] <Company> thomasvs: gstclock.h has padding, too
[22:49] BBB ([email protected]) joined #gstreamer.
[22:51] <Company> how am i supposed to write tests for input or output elements?
[22:52] <Company> like osssink for example
[22:53] <Company> or alsasink where the tests where just disabled
[22:53] <BBB> you can't
[22:53] <BBB> it depends on devices which might not be available
[22:53] <BBB> you can at best write a sample app
[22:53] <BBB> sinesrc ! osssink
[22:53] <BBB> osssrc ! fakesink
[22:55] <Company> i want a test that gets run automatically not some sample app that noone looks at
[22:55] <BBB> that's what I said
[22:55] <BBB> <BBB> you can't
[22:55] <BBB> <BBB> it depends on devices which might not be available
[22:55] <Company> so there's no way to test sources or sinks?
[22:55] <Company> that sounds bad
[22:56] <BBB> I don't think there's a reliable way
[22:56] <ds-work> not in gstreamer/testsuite
[22:56] <ds-work> you can add it to gst-plugins/testsuite
[22:56] <thomasvs> Company: ugh, that looks weird :) no wonder I missed it.
[22:56] <thomasvs> Company: why is it like that ?
[22:56] <Company> thomasvs: to get guint64 correct
[22:57] <thomasvs> Company: shouldn't that just use some sort of union with the padding instead ?
[22:57] <Company> sizeof(guint64) / sizeof(gpointer) isn'?t const...
[22:57] <Company> ideally yes
[22:57] <Company> but since we were going to reset padding anyway...
[22:58] <Company> ds-work: it's in gst-plugins/testsuite - but it was just disabled by thomasvs so that's why i'm asking
[22:58] <ds-work> ah
[22:59] <ds-work> if it tests for existence of alsasink first, it shouldn't be a problem
[22:59] <ds-work> or whatever element
[22:59] <ds-work> we need a testsuite whose tests are allowed to fail
[23:00] <Company> thomasvs: why exactly did you disable alsasink tests?
[23:00] <ds-work> because alsasink testing is not appropriate for distcheck
[23:01] <thomasvs> right - they're automated tests, so they shouldn't interfer with the user's environment
[23:01] <thomasvs> make check is for making sure things work as expected, not for scaring the user with beeps :)
[23:01] <thomasvs> (esp. not for running tests that I don't have hardware for)
[23:02] <Company> it doesn't run those tests if you don't have alsa
[23:02] <thomasvs> it did here
[23:02] <thomasvs> I have alsa libs, but no alsa
[23:02] <Company> because you installed alsalib
[23:02] <thomasvs> right, but I don't have alsa
[23:02] <Company> yeah, that's a kinda broken setup...
[23:02] <thomasvs> anyways, it shouldn't try beeping as an automated test.
[23:02] <thomasvs> there's nothing broken about having libs installed
[23:03] <thomasvs> the automated test shouldn't depend on what the user is currently doing to fail or pass
[23:03] <thomasvs> it's for things that can be checked programatically
[23:03] <Company> so where do we put tests for those stuff?
[23:03] <thomasvs> something similar to the media testsuite
[23:03] <thomasvs> we should have a gst-test module where we put stuff like that
[23:03] <thomasvs> or just in tests/, as was done before
[23:03] <thomasvs> testsuite/ is for make check
[23:04] <Company> i'd like to have something i can just run with one command
[23:04] <Company> like "make check"
[23:04] <thomasvs> well, for example we can put a script in tests/ that runs them
[23:04] <thomasvs> and we can have make test from the Makefile if you want
[23:05] <Company> hm
[23:05] <thomasvs> I agree that we should have it, but just not as a check: target
[23:05] <Company> could we use makefile-foo to have a make check in tests/ ?
[23:05] <thomasvs> yeah
[23:05] <thomasvs> ehm, make test you mean right ?
[23:05] <Company> whatever
[23:05] <thomasvs> ====================================================
[23:06] <thomasvs> gst-plugins-0.7.3.2.tar.gz is ready for distribution
[23:06] <thomasvs> ====================================================
[23:06] Action: thomasvs is glad automake doesn't actually test if the code works to make that statement :)
[23:06] <Company> just something that is as similar as possible to make check
[23:06] <thomasvs> sure, file a bug so I don't forget, and I'll add it
[23:06] <thomasvs> (assign it to me so it doesn't get lost)
[23:10] Uraeus ([email protected]) joined #gstreamer.
[23:10] mathrick ([email protected]) joined #gstreamer.
[23:10] <Zeenix> Uraeus: wb
[23:10] <Uraeus> hi Zeenix
[23:10] <mathrick> re
[23:11] <Uraeus> Zeenix: I read the libcaca archive the other day, noticed that you have been quite active over there :)
[23:11] <Uraeus> ello mathrick
[23:11] <Zeenix> Uraeus: mailing-list archive?
[23:13] <thomasvs> Uraeus: exactly when are you arriving here friday ?
[23:14] <Uraeus> thomasvs: early in the morning, my plane leaves Oslo 07.30
[23:14] <Uraeus> thomasvs: let me find the exact date
[23:14] <Uraeus> Zeenix: yes, the archive
[23:14] <Uraeus> s/date/time/
[23:15] <thomasvs> Uraeus: and The Other ?
[23:15] <Uraeus> thomasvs: heh, ok not that Early :) due to transit in Copenhagen I will land at 12.10
[23:15] <Uraeus> thomasvs: no idea
[23:15] <BBB> Uraeus: are you rich? you're constantly going there :p
[23:15] <Zeenix> Uraeus: good that you reminded me of libcaca.. i needed to ask something...
[23:16] <Uraeus> BBB: I am not rich, I just choose not to focus on material things :)
[23:16] <taaz> Zeenix: make cacasink handle window resizes ;)
[23:17] <taaz> thomasvs: that clock padding might be able to revert to the gpointer style vs that guint8 hack
[23:17] <thomasvs> taaz: how ?
[23:17] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[23:19] foser ([email protected]) joined #gstreamer.
[23:19] <Zeenix> sam (libcaca maintainer) has proposed the following: he (or I) shall implement a special-mode in libcaca which would allow the app. to handle libcaca a buffer & libcaca shall draw it's graphics on that buffer. The implementation would be different for different drivers (X11, imlib etc) of course
[23:20] Action: mathrick wonders what is ASCII needed for in libcaca ;)
[23:20] <taaz> thomasvs: the comment out line below what you just changed?
[23:20] Action: Company fixes gstclock padding
[23:20] <taaz> commented
[23:20] <Zeenix> do anyone have any problem/objection/question on this proposition?
[23:21] <ds-work> Zeenix: it seems rather silly
[23:21] <ds-work> Zeenix: it might as well be a video filter then
[23:21] <thomasvs> Company: are you commiting that to CVS ?
[23:23] <Zeenix> ds-work: do you anything in mind that does'nt make libcaca X-specific?
[23:23] <Uraeus> mathrick: how can you have an ascii art renderer without ascii ? :)
[23:24] BBB ([email protected]) left irc: "Client exiting"
[23:24] <mathrick> Uraeus: simple, just use lotsa different graphic buffers, and eventually you'll obtain all-purpose graphics renderer you can as well drop ASCII from ;P
[23:25] <Company> thomasvs: yeah
[23:27] <thomasvs> Company: ok, let me know when it's in so I can prerel
[23:28] <Zeenix> ds-work: what is did was add two X-specific function in libcaca: caca_get_xid () and caca_set_parent_window ()
[23:28] <Zeenix> s/is/i
[23:28] <Zeenix> of course, the function did nothing when other drivers were being used
[23:29] <Zeenix> but sam did'nt like it as he want his lib to be totallay non-X-specific or even non-unix-specific
[23:33] <Company> thomasvs: done
[23:36] markey ([email protected]) left irc: "g'nite"
[23:36] Action: ds-work cringes at editing a 0.6-style plugin
[23:37] <Misirlou> heh
[23:37] <Company> that's good
[23:37] <Misirlou> You guys must have done a lot of work between 0.6 and 0.7. :) :)
[23:37] <Company> because it means we did good work for 0.7
[23:37] <thomasvs> any advance stabs at a release name for tomorrow ?
[23:38] <Company> taaz wanted to name it 0.8.3 or something
[23:38] Action: Zeenix still is unable to compile his app. from uninstalled gst
[23:38] <Company> other than that i'm not aware of release name wishes
[23:38] <ds-work> wooden eels?
[23:39] <Company> do we update the NEWS file now that we have a changelog?
[23:39] <NFusi0n> name tomorrows release....
[23:39] <NFusi0n> "FOR GREAT JUSTICE"
[23:40] <Misirlou> I like WOODEN EELS FOR GREAT JUSTICE.
[23:40] <Company> if we do it, we should name it "FOR GREAT JUSTICE!!!1122"
[23:40] <ds-work> "spider is dead.  Long live spider"
[23:41] <NFusi0n> "WOODEN EELS FOR GREAT JUSTICE!!one!!112!!1one!!!!"
[23:41] <Misirlou> I concur.
[23:41] <Misirlou> thomasvs: Make it so!
[23:41] <mathrick> yay "FOR GREAT JUSTICE" is cool
[23:42] Action: thomasvs wonders which movie he missed out on to understand
[23:42] <mathrick> and 1.0 will be "All your base are belong to us" obviously :)
[23:42] <Uraeus> thomasvs: call it  'in living Memory of Janet's boob that the americans are scared of' :)
[23:42] <mathrick> thomasvs: shame on you dude ;P
[23:42] <NFusi0n> mathrick: o of course
[23:42] <ds-work> 1.0 should be 'you have no chance to survive, make your time'
[23:42] <NFusi0n> 1.0 should be "MOVE ZIG"
[23:43] <NFusi0n> hahaha
[23:43] <Company> 1.0 should be "off on holidays"
[23:43] <Company> or something
[23:43] <Misirlou> nonono
[23:43] <mathrick> "Somebody set us up the bomb" ? ;)
[23:43] <Company> "first paid release"
[23:43] <Zeenix> Company: we are asking for a release name not you personal maters :)
[23:43] <Zeenix> s/you/your
[23:43] <mathrick> "We get signal" ? <-- pro audio release ;)
[23:43] <NFusi0n> mathrick: that should be for 1.0's bugfix release
[23:44] <Company> Zeenix: i'm not the paid one in here
[23:44] <mathrick> "Main screen turn on" <-- pro video one ;)
[23:44] <Company> in fact, i'm the only non-paid i tend to think ;)
[23:44] <Company> that thing is so old not even slashdot makes jokes about it
[23:44] <NFusi0n> that's what's so funny =D
[23:45] <mathrick> Company: that == all your base?
[23:45] <Company> "in soviet russia, gstreamer releasxes us" :(
[23:45] <mathrick> Company: it doesn't matter a single bit :)
[23:45] <NFusi0n> Company: hahahahaha
[23:45] <Misirlou> Company: !!!!!!!!!!!!!!! I like it
[23:45] <Misirlou> Or
[23:45] <Misirlou> "Step 1: GStreamer 1.0; Step 2: ???????????; Step 3: PROFIT!"
[23:46] <Company> that's not old enough
[23:46] <Company> and actually it's "Step 1: GStreamer; Step 2: lots of work; Step 3: PROFIT!"
[23:46] <Company> look at thomasvs :p
[23:47] <Company> you could certainly come up with a "very funny" sco related release name for a bugfix release, too
[23:48] <ds-work> "in soviet russia, sco sues you"
[23:48] <Zeenix> ds-work: ?
[23:49] <NFusi0n> in soviet russia, YOU SUE SCO!
[23:49] <Company> i like ds' idea
[23:49] Shoragan ([email protected]) joined #gstreamer.
[23:49] <Company> that makes people think
[23:49] <NFusi0n> (what if you get sued)
[23:49] <mathrick> nono, "in soviet russia, sco is right"
[23:49] Action: Zeenix needs to understand the idea before liking it :)
[23:50] Action: thomasvs is with Zeenix on this one
[23:50] <Company> (i don't like stupid slashdot jokes as release names at all)
[23:50] <ds-work> neither do I, really
[23:50] <mathrick> Company: that's why we should use "FOR THE GREAT JUSTICE" :)
[23:51] <Company> you could name it "already got 10 invitations to orkut"
[23:51] <NFusi0n> I think the all your base stuff is good
[23:51] <Company> that would be equally stupid
[23:51] <NFusi0n> it's not /. related..much...
[23:51] <mathrick> NFusi0n: no, it's 1.0 name :)
[23:51] <ds-work> however, it's about time to revive the stupid natalie portman/hot grits jokes
[23:51] <Misirlou> I don't know those.
[23:51] <taaz> did that gstreamer is teh suck thing start here?  i just noticed http://www.gnome.org/~markmc/blog//04022004 on planet gnome.
[23:51] <Company> that's require her to make a good movie again
[23:51] <NFusi0n> tje 
[23:52] <thomasvs> taaz: sort of
[23:52] <NFusi0n> the win32 release should be "down with dshow"
[23:52] <ds-work> 'X is the suck' did not come from gstreamer
[23:52] <thomasvs> taaz: it was a comment by oGALAXYo about gstreamer
[23:52] <thomasvs> from his armageddon posts I think
[23:52] <Company> man, you need to be more subtle
[23:52] <thomasvs> ds-work: the misspelled version is
[23:52] <mathrick> NFusi0n: nope, direct attacs are no good
[23:52] <Company> "down with dshow" is not subtle enough
[23:52] <taaz> ds-work: yeah i mean "is teh suck" specifically
[23:52] <ds-work> 'dshowsrc ! yr_mom ! dshowsink'
[23:52] <NFusi0n> hahaha
[23:53] <mathrick> ds-work: nice one :)
[23:53] <NFusi0n> dshowsrc ! yr_mom ! nullsink
[23:53] <Company> #ifdef __WIN32__
[23:53] <Company> gst_element_register ("dshowsrc, GST_TYPE_FAKESINK);
[23:54] <ds-work> '$(whatever) is the suck' certainly did not originate with gstreamer
[23:54] <mathrick> maybe 0.8 should be named "80/20" ?
[23:54] <Company> 0.8 isn't 80/20
[23:54] <mathrick> ds-work: s/the/teh/
[23:54] <Company> that would be too optimistic :p
[23:55] <ds-work> dang it, i misspelled the misspelling twice :)
[23:55] <mathrick> Company: 80% left ;P
[23:55] <Company> http://www.google.com/search?q=%22is%20teh%20suck%22
[23:55] <Company> either gstreamer really is teh suck or even that didn't originate her :p
[23:56] <Company> is cvs closed for release btw?
[23:56] <ds-work> http://episteme.arstechnica.com/6/ubb.x?a=tpc&s=50009562&f=39309975&m=2820942123
[23:56] <ds-work> search for 'teh'
[23:57] <taaz> Company: did you want to move set_state in gstelement.h up near other state funcs?
[23:57] <thomasvs> Company: sort of yeah
[23:58] <Company> taaz: it makes sense but isn't important enough to do now
[23:58] <Company> (especially because we'll do a header reorg when we open 0.9 anyway)
[23:58] <taaz> Company: you'll never remember...
[23:59] <mathrick> taaz: bugzilla? :)
[23:59] <Company> taaz: i have reorged headers before in 0.7...
[00:00] --- Thu Feb  5 2004
[00:04] <mathrick> thomasvs: what does "nijntje" mean? "nineteen" by any chance?
[00:04] hadley ([email protected]) joined #gstreamer.
[00:04] <thomasvs> mathrick: no, it's sort of like hello kitty but different and dutch
[00:04] <thomasvs> mathrick: kind of hard to explain :)
[00:05] <mathrick> thomasvs: heh :)
[00:05] <thomasvs> mathrick: ah, hang on, it's "miffy" in english
[00:05] <thomasvs> http://www.nijntje.nl/
[00:06] <mathrick> ahh, this one
[00:06] <mathrick> I've never came to idea it has its own name :)
[00:20] Uraeus ([email protected]) left irc: "Client exiting"
[00:28] eventide ([email protected]) left irc: "leaving"
[00:32] <thomasvs> taaz: so how do I "see" the problem with gst-python and uninstalled gst ?
[00:36] <taaz> point your pkg config path at the uninstalled dirs and run configure should probably do it
[00:37] <taaz> err... have to build too
[00:37] <taaz> it's going to grab the includedir var from the .pc files which currently points somewhere silly like the pkgconfig/ dir
[00:37] <taaz> when it really should point to where the include files are located
[00:38] <taaz> so when it trys to scan headers for the auto generated defs it won't find the proper ones
[00:38] <thomasvs> ok, cool
[00:38] <thomasvs> then I can verify that here and check your patch
[00:39] <taaz> i think the other comments, which i've now forgetten <g>, in that mail are relevant too.  just to make sure other things don't break along the way.  things still build for me here but i doubt i'm really testing it properly
[00:40] <taaz> like using -Ldir -lfoo vs dir/libfoo.la
[00:41] <thomasvs> taaz: do you know which of the two incantations is preferred ?
[00:46] <NFusi0n> Hey do any of you developers need 92mmfans? i'm down for a donation or two, i have tons of 92mm fans
[00:47] Zeenix ([email protected]) left irc: Read error: 110 (Connection timed out)
[00:48] <mathrick> k, time for me to sleep some before last exam :)
[00:48] <mathrick> nite guys
[00:48] <NFusi0n> 'nite :)
[00:48] Nick change: mathrick -> mathrick|sleep
[00:49] #gstreamer: mode change '+o thomasvs' by ChanServ!ChanServ@services.
[00:49] Topic changed on #gstreamer by !irc.freenode.net: New website under construction - http://www.freedesktop.org/~gstreamer/ | GStreamer article - http://www.osnews.com/story.php?news_id=5648 | test prereleases - http://freedesktop.org/~gstreamer/src/gstreamer/pre/gstreamer-0.7.3.2.tar.gz and http://freedesktop.org/~gstreamer/src/gst-plugins/pre/gst-plugins-0.7.3.2.tar.gz
[00:50] foser ([email protected]) left irc: "[ I want to believe ]"
[00:54] <taaz> thomasvs: i have no idea
[00:57] Action: Company has fixed 4 or 5 bugs since cvs is frosted
[00:57] <Company> somebody unfreeze it :)=
[00:58] <ds-work> if the prereleases are out, it sounds unfrozen
[00:58] <Company> thomasvs: may i commit again?
[01:01] dilinger ([email protected]) joined #gstreamer.
[01:01] <ds-work> ERROR        GST_PADS(29094) gstpad.c(2720):gst_real_pad_dispose: <'':serial_2059288011> still explicit caps audio/x-vorbis set
[01:02] <ds-work> that's on switching from one track to the next in rb
[01:02] <thomasvs> Company: sure, hope they're not too high-impact
[01:02] <ds-work> from a few-hours-old CVS
[01:02] <Company> ds-work: i added a warning there, because PAUSED=>READY should clear them
[01:03] <ds-work> it seg faults soon after
[01:04] <Company> it shouldn't
[01:04] Shoragan ([email protected]) left irc: "Leaving"
[01:05] <ds-work> #2  0x40822658 in gst_caps_replace (caps=0x8328ca0, newcaps=0x50b855)
[01:05] <ds-work>     at gstcaps.c:1013
[01:05] <ds-work> #3  0x4082d4c4 in gst_element_clear_pad_caps (element=0x0) at gstelement.c:2760
[01:05] <ds-work> #4  0x4082dc1a in gst_element_change_state (element=0x8326fd8)
[01:05] <ds-work>     at gstelement.c:2843
[01:05] <ds-work> anyway, ignore me until I update CVS
[01:05] <Company> k :)
[01:10] thaytoo ([email protected]) joined #gstreamer.
[01:14] sublett ([email protected]) left irc: "I like food, food is good!"
[01:17] Zeenix ([email protected]) joined #gstreamer.
[01:29] pb_ ([email protected]) left irc: "Client exiting"
[01:35] <Company> thomasvs: please think about adding "first try at theora decoding support" to the release notes for 0.7.4
[01:36] <thomasvs> Company: 'k
[01:36] Action: thomasvs needs to figure out where to put release notes now in the first place :)
[01:36] <Company> it's a nice political statewment ;)
[01:36] <Company> though it needs some detection love in configure and so on
[01:36] <taaz> why is that gst-run code needed?  it seems like something that's going to be hard to support/use
[01:37] <thomasvs> taaz: two reasons
[01:37] <thomasvs> a) docs simpler/simpler for user, since only unversioned tools are promoted
[01:37] <thomasvs> b) parallel installability of packages with unversioned tools is possible, since one package of unversioned tools can frontend for two packages of versioned tools
[01:38] <thomasvs> taaz: but I don't see how it's hard to support or use, it effectively "hides" the versioned binaries
[01:38] <thomasvs> I was even thinking of putting the real ones in libexec :)
[01:39] <taaz> it's a decent idea for single version installs.  but i see issues with multiple versions... at least from debian packaging standpoint
[01:39] <taaz> for insstance, now the packages have to be aware that only one can install the unversioned tools/docs/etc
[01:40] <taaz> which one is it?  newer one? older one?
[01:40] <thomasvs> taaz: it doesn't matter. they work for both.
[01:40] <thomasvs> taaz: that's the point
[01:40] <thomasvs> taaz: so, it's possible to have two lib packages installed, one for 0.7 and one for 0.8
[01:40] <Company> make a different package and depend on it in both versions
[01:40] <thomasvs> and only one package containing the frontends
[01:41] <thomasvs> as long as the lib packages just require a minimum version on the frontends package, this works fine
[01:41] <taaz> which source does that frontend package come from?
[01:41] <thomasvs> taaz: I brought this up repeatedly, and there is a bug for it, you're coming into the discussion kind of late :)
[01:41] <thomasvs> taaz: it doesn't matter what source it's from
[01:41] <Company> that'd up to the packager
[01:41] <thomasvs> taaz: it's preferrable that it's the most recent
[01:42] <thomasvs> taaz: but it doesn't really matter
[01:42] <taaz> thomasvs: sorry, i didn't even realize anyone was seriously considering doing this until i saw you mention it yesterday
[01:43] <thomasvs> taaz: I'm always serious about stuff I mention or file in bugzilla :)
[01:43] <thomasvs> taaz: anyways, feel free to log legitimate problems
[01:43] <taaz> i hear what you all are saying but i really don't know how to do this with debian packages in a sane way that is forwards/backwards compatible like you suggest it should be
[01:44] <thomasvs> taaz: but as a packager this was the last missing key I needed to make workable parallel-installable packages
[01:44] <thomasvs> taaz: what do you think doesn't work in the scheme ?
[01:44] <taaz> thomasvs: i think this is making parallel installable packages more difficult.
[01:44] <taaz> i had planned on making every package versioned
[01:45] <taaz> and assumed all the tools and docs and everything would be versioned.  so everything would be parallel installable
[01:45] <taaz> but unversioned tools are specifically not parallel installable since they have name conflicts
[01:45] <thomasvs> taaz: so just don't package them ?
[01:46] <taaz> i probably won't, and i'll get yelled at ;)
[01:46] <thomasvs> heh, your choice
[01:46] <thomasvs> I really don't see why you can't have a gstreamer-common deb that contains the most recent copy of the non-versioned scripts ...
[01:46] <thomasvs> and a libgstreamer06-tools and libgstreamer07-tools deb both containing different versioned tools
[01:46] <taaz> i probably can... but that makes some assumptions
[01:47] <taaz> such as if you depend on gstreamer-common x.y that the functionality will exist in X.Y+1
[01:47] <Misirlou> taaz: By the way, did you fix gstreamer-mpeg2dec?
[01:48] ain ([email protected]) left irc: "Punk as fuck"
[01:48] <taaz> Misirlou: yes, last week.  new libmpeg2 just got into archive today.
[01:49] <Misirlou> excellent
[01:52] <taaz> thomasvs: i'll see how it works out.  i'm not sure i'll install those wrappers though.  the more 'debian' way of doing it would be with the alternatives system so users could pick which version the aliases called.  doesn't have the fancy options though that gst-run does.
[01:53] <taaz> thomasvs: that code really looks like it should be a small(er) shell script rather than C ;)
[01:55] <thomasvs> taaz: I thought about making it a shell script, but then I couldn't use getopt
[01:56] <thomasvs> taaz: someone would complain it wouldn't run with sh, blah blah blah
[01:56] <thomasvs> anyway, I wrotes it, I gets to choose the language :)
[01:56] Nick change: thomasvs -> thomasvz
[01:56] <taaz> speaking of options...
[01:56] <thomasvz> taaz: but make sure to let me know exactly what you think is not correct
[01:56] <thomasvz> make sure you mail to the ML about it though
[01:57] <ds-work> -rwxr-xr-x    1 root     root         9092 2003-12-06 19:34 /usr/bin/getopt
[01:57] <taaz> i'm totally lost on how to handle gnome-side-popt and python-side-optparse at the same time
[01:58] <thomasvz> taaz: didn't I do that last time I worked on gst-python ?
[01:58] <thomasvz> I'm sure I was able to parse my options and gst options
[01:59] <taaz> hrm... maybe i'm confused then ;)
[01:59] <taaz> i had some issues the other day...
[02:00] <taaz> i think one is importing gst will do the popt parsing which will quit on --help, so you never get to see optparse --help options
[02:01] <taaz> optparse does the same thing but you can catch the quit signal and keep going
[02:01] <taaz> s/signal/exception/
[02:02] <taaz> anyway, it can wait.  just an annoyance at the moment
[02:16] jcsston ([email protected]) joined #gstreamer.
[02:18] 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 
[02:39] thomasvz ([email protected]) got netsplit.
[02:39] lilo_booter ([email protected]) got netsplit.
[02:40] grub_booter ([email protected]) joined #gstreamer.
[02:41] thomasvz ([email protected]) returned to #gstreamer.
[02:45] lilo_booter ([email protected]) got lost in the net-split.


-------------------------------------------------------
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