IRC Logs

*******************************************************************
[03:07] ds-work_ ([email protected]) joined #gstreamer.
[03:11] Nick change: ds-work_ -> ds-uhv
[03:18] <asdlfjkklsj> <LeRaway> asdlfjkklsj, it's a known problem, even with current CVS HEAD
[03:18] <asdlfjkklsj> well that sucks ;)
[03:19] <LeRaway> yep, true.
[03:19] <lmjohns3> ds-uhv: ping
[03:19] asdlfjkklsj ([email protected]) left #gstreamer (".").
[03:20] <ds-uhv> moo
[03:20] <lmjohns3> just wanted to check on this audiofiltertemplate ... the class->setup function gets called after caps are set, correct ?
[03:20] <ds-uhv> yes
[03:22] <lmjohns3> ok
[03:22] <lmjohns3> i also just added width = (int) 32 to standard float props in audio.h
[03:22] <lmjohns3> any objections to making the template use float audio by default ? :)
[03:26] <ds-uhv> yes
[03:27] <ds-uhv> because elements use audio.h expect very specific behavior
[03:27] <lmjohns3> ok, only half-serious anyway
[03:29] <ds-uhv> don't add a width to float caps
[03:29] <ds-uhv> caps definitions can't be changed without a migration plan
[03:29] ^iain^ ([email protected]) left irc: "leaving"
[03:30] <lmjohns3> ds-uhv: doh ... it's already part of GST_AUDIO_FLOAT_PAD_TEMPLATE_CAPS, but it's not in GST_AUDIO_STANDARD_FLOAT_<tab>
[03:30] <lmjohns3> s/STANDARD_FLOAT/FLOAT_STANDARD/
[03:31] <lmjohns3> it was in there before the caps change, i thought
[03:34] <ds-uhv> ah, it might be
[03:35] <ds-uhv> nothing handles it, though
[03:35] <ds-uhv> if you want to go through all the plugins and fix them, please do
[03:36] Rotty ([email protected]) left irc: Remote closed the connection
[03:45] hadley ([email protected]) joined #gstreamer.
[03:57] Nick change: Jaramir -> Jara[zZ]
[04:02] Nick change: bluejay -> bluejay|gone
[04:03] lmjohns3 ([email protected]) left irc: Read error: 110 (Connection timed out)
[04:56] LeRaway ([email protected]) left irc: "Leaving"
[05:10] <harshy> ok, whats the new trick to compile gstreamer cvs off freedesktop.org
[05:12] <ds-uhv> is it not checking out common?
[05:13] <harshy> no
[05:13] <harshy> common is already there
[05:13] <ds-uhv> what is wrong, then?
[05:13] <harshy> I just pulled down the entire gstreamer cvs stuff
[05:13] <harshy> it complains about common
[05:13] <harshy> even to know its there
[05:13] <harshy> then tries to checkout common and fails
[05:15] <ds-uhv> make a link from gstreamer/common to common
[05:15] <ds-uhv> the checkout code is broken
[05:16] <harshy> heh
[05:21] <ds-uhv> hmmm... it worked just fine here
[05:21] <ds-uhv> fresh checkout of gstreamer, ./autogen.sh, works
[05:22] <harshy> wierd
[05:22] <harshy> oh well
[05:22] <harshy> links work ok for now
[05:24] herzi_lap ([email protected]) joined #gstreamer.
[05:35] lilo ([email protected]) got netsplit.
[05:35] jimmy_dean|AFK ([email protected]) got netsplit.
[05:35] kmaraas ([email protected]) got netsplit.
[05:35] walters ([email protected]) got netsplit.
[05:35] ct_ ([email protected]) got netsplit.
[05:35] harshy ([email protected]) got netsplit.
[05:35] danb ([email protected]) got netsplit.
[05:35] sjoerd ([email protected]) got netsplit.
[05:35] md` ([email protected]) got netsplit.
[05:35] bluejay|gone ([email protected]) got netsplit.
[05:35] mathrick|sleep ([email protected]) got netsplit.
[05:35] ChrisHJW_log ([email protected]) got netsplit.
[05:35] ChrisHJW ([email protected]) got netsplit.
[05:35] desrt ([email protected]) got netsplit.
[05:35] taaz ([email protected]) got netsplit.
[05:35] dilinger ([email protected]) got netsplit.
[05:35] dennis_ ([email protected]) got netsplit.
[05:35] Jara[zZ] ([email protected]) got netsplit.
[05:35] TD[gone] ([email protected]) got netsplit.
[05:35] sxpert ([email protected]) got netsplit.
[05:35] sxpert_work ([email protected]) got netsplit.
[05:36] md` ([email protected]) returned to #gstreamer.
[05:36] bluejay|gone ([email protected]) returned to #gstreamer.
[05:36] walters ([email protected]) returned to #gstreamer.
[05:36] jimmy_dean|AFK ([email protected]) returned to #gstreamer.
[05:36] ct_ ([email protected]) returned to #gstreamer.
[05:36] sxpert ([email protected]) returned to #gstreamer.
[05:36] sxpert_work ([email protected]) returned to #gstreamer.
[05:36] TD[gone] ([email protected]) returned to #gstreamer.
[05:36] kmaraas ([email protected]) returned to #gstreamer.
[05:36] mathrick|sleep ([email protected]) returned to #gstreamer.
[05:36] ChrisHJW_log ([email protected]) returned to #gstreamer.
[05:36] harshy ([email protected]) returned to #gstreamer.
[05:36] lilo ([email protected]) returned to #gstreamer.
[05:36] ChrisHJW ([email protected]) returned to #gstreamer.
[05:36] danb ([email protected]) returned to #gstreamer.
[05:36] taaz ([email protected]) returned to #gstreamer.
[05:36] dilinger ([email protected]) returned to #gstreamer.
[05:36] sjoerd ([email protected]) returned to #gstreamer.
[05:36] desrt ([email protected]) returned to #gstreamer.
[05:36] Jara[zZ] ([email protected]) returned to #gstreamer.
[05:36] dennis_ ([email protected]) returned to #gstreamer.
[05:36] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by tenn.freenode.net
[05:49] Miko5881 ([email protected]) left irc: "ChatZilla 0.8.31 [Mozilla rv:1.4/20030928]"
[05:51] lilo ([email protected]) got netsplit.
[05:51] jimmy_dean|AFK ([email protected]) got netsplit.
[05:51] kmaraas ([email protected]) got netsplit.
[05:51] harshy ([email protected]) got netsplit.
[05:51] walters ([email protected]) got netsplit.
[05:51] danb ([email protected]) got netsplit.
[05:51] ct_ ([email protected]) got netsplit.
[05:51] sjoerd ([email protected]) got netsplit.
[05:51] desrt ([email protected]) got netsplit.
[05:51] mathrick|sleep ([email protected]) got netsplit.
[05:51] bluejay|gone ([email protected]) got netsplit.
[05:51] ChrisHJW ([email protected]) got netsplit.
[05:51] ChrisHJW_log ([email protected]) got netsplit.
[05:51] taaz ([email protected]) got netsplit.
[05:51] dilinger ([email protected]) got netsplit.
[05:51] md` ([email protected]) got netsplit.
[05:51] dennis_ ([email protected]) got netsplit.
[05:51] Jara[zZ] ([email protected]) got netsplit.
[05:51] TD[gone] ([email protected]) got netsplit.
[05:51] sxpert ([email protected]) got netsplit.
[05:51] sxpert_work ([email protected]) got netsplit.
[05:51] md` ([email protected]) returned to #gstreamer.
[05:51] bluejay|gone ([email protected]) returned to #gstreamer.
[05:51] walters ([email protected]) returned to #gstreamer.
[05:51] jimmy_dean|AFK ([email protected]) returned to #gstreamer.
[05:51] ct_ ([email protected]) returned to #gstreamer.
[05:51] sxpert ([email protected]) returned to #gstreamer.
[05:51] sxpert_work ([email protected]) returned to #gstreamer.
[05:51] TD[gone] ([email protected]) returned to #gstreamer.
[05:51] kmaraas ([email protected]) returned to #gstreamer.
[05:51] mathrick|sleep ([email protected]) returned to #gstreamer.
[05:51] ChrisHJW_log ([email protected]) returned to #gstreamer.
[05:51] harshy ([email protected]) returned to #gstreamer.
[05:51] lilo ([email protected]) returned to #gstreamer.
[05:51] ChrisHJW ([email protected]) returned to #gstreamer.
[05:51] danb ([email protected]) returned to #gstreamer.
[05:51] taaz ([email protected]) returned to #gstreamer.
[05:51] dilinger ([email protected]) returned to #gstreamer.
[05:51] sjoerd ([email protected]) returned to #gstreamer.
[05:51] desrt ([email protected]) returned to #gstreamer.
[05:51] Jara[zZ] ([email protected]) returned to #gstreamer.
[05:51] dennis_ ([email protected]) returned to #gstreamer.
[05:51] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by tenn.freenode.net
[05:52] lilo ([email protected]) got netsplit.
[05:52] jimmy_dean|AFK ([email protected]) got netsplit.
[05:52] kmaraas ([email protected]) got netsplit.
[05:52] harshy ([email protected]) got netsplit.
[05:52] walters ([email protected]) got netsplit.
[05:52] danb ([email protected]) got netsplit.
[05:52] ct_ ([email protected]) got netsplit.
[05:52] sjoerd ([email protected]) got netsplit.
[05:52] desrt ([email protected]) got netsplit.
[05:52] mathrick|sleep ([email protected]) got netsplit.
[05:52] bluejay|gone ([email protected]) got netsplit.
[05:52] ChrisHJW ([email protected]) got netsplit.
[05:52] ChrisHJW_log ([email protected]) got netsplit.
[05:52] taaz ([email protected]) got netsplit.
[05:52] dilinger ([email protected]) got netsplit.
[05:52] md` ([email protected]) got netsplit.
[05:52] dennis_ ([email protected]) got netsplit.
[05:52] Jara[zZ] ([email protected]) got netsplit.
[05:52] TD[gone] ([email protected]) got netsplit.
[05:52] sxpert ([email protected]) got netsplit.
[05:52] sxpert_work ([email protected]) got netsplit.
[05:52] md` ([email protected]) returned to #gstreamer.
[05:52] bluejay|gone ([email protected]) returned to #gstreamer.
[05:52] walters ([email protected]) returned to #gstreamer.
[05:52] jimmy_dean|AFK ([email protected]) returned to #gstreamer.
[05:52] ct_ ([email protected]) returned to #gstreamer.
[05:52] sxpert ([email protected]) returned to #gstreamer.
[05:52] sxpert_work ([email protected]) returned to #gstreamer.
[05:52] TD[gone] ([email protected]) returned to #gstreamer.
[05:52] kmaraas ([email protected]) returned to #gstreamer.
[05:52] mathrick|sleep ([email protected]) returned to #gstreamer.
[05:52] ChrisHJW_log ([email protected]) returned to #gstreamer.
[05:52] harshy ([email protected]) returned to #gstreamer.
[05:52] lilo ([email protected]) returned to #gstreamer.
[05:52] ChrisHJW ([email protected]) returned to #gstreamer.
[05:52] danb ([email protected]) returned to #gstreamer.
[05:52] taaz ([email protected]) returned to #gstreamer.
[05:52] dilinger ([email protected]) returned to #gstreamer.
[05:52] sjoerd ([email protected]) returned to #gstreamer.
[05:52] desrt ([email protected]) returned to #gstreamer.
[05:52] Jara[zZ] ([email protected]) returned to #gstreamer.
[05:52] dennis_ ([email protected]) returned to #gstreamer.
[05:52] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by tenn.freenode.net
[05:56] michael_-_ ([email protected]) joined #gstreamer.
[06:03] md` ([email protected]) left irc: 
[06:04] md` ([email protected]) joined #gstreamer.
[06:05] mathrick|sleep ([email protected]) left irc: Remote closed the connection
[06:13] Nick change: jimmy_dean|AFK -> jimmy_dean
[06:20] ChriHJW_log ([email protected]) joined #gstreamer.
[06:35] ChrisHJW_log ([email protected]) left irc: Read error: 110 (Connection timed out)
[06:44] hadley ([email protected]) left irc: "Leaving"
[07:41] thaytan ([email protected]) joined #gstreamer.
[08:03] sri ([email protected]) joined #gstreamer.
[08:03] <sri> gretings.
[08:04] <sri> erg greetings rather
[08:04] <sri> question on gst-plugins 
[08:04] <sri> waht is the devel library I need to get shoutcast2 plugin
[08:04] <sri> and second I seem to be running into a problem with gstmad.c
[08:05] <sri> it gives me this error:
[08:05] <sri> gstmad.c:241: error: `MAD_EMPHASIS_RESERVED' undeclared (first use in this function)
[08:05] <sri> debian package libmad is up to date..so maybe I need to get the latest libmad package?
[08:30] ct_ ([email protected]) left irc: "Client exiting"
[08:47] harshy ([email protected]) left irc: "I quit for now"
[08:47] harshy ([email protected]) joined #gstreamer.
[09:00] sub_pop ([email protected]) left irc: "Client exiting"
[09:33] mathrick ([email protected]) joined #gstreamer.
[09:33] <mathrick> mornin
[10:05] matrixise ([email protected]) joined #gstreamer.
[10:13] mathrick ([email protected]) left irc: Remote closed the connection
[10:15] mathrick ([email protected]) joined #gstreamer.
[10:26] matrixise ([email protected]) left #gstreamer ("Leaving").
[10:34] mathrick_ ([email protected]) joined #gstreamer.
[10:34] mathrick ([email protected]) left irc: "Client exiting"
[10:34] Nick change: mathrick_ -> mathrick
[10:56] kniteshade ([email protected]) joined #gstreamer.
[11:02] Uraeus ([email protected]) joined #gstreamer.
[11:07] <Uraeus> walters: around?
[11:08] <walters> yep
[11:08] <walters> not for long though... :)
[11:10] <walters> Uraeus: what's up?
[11:11] <Uraeus> walters: well I noticed over the last few days with fluctuating metadata support that if you load songs into RB without metadata then it seems there is no way except deleting playlist and loading metadata to get the metadata in if support for it is added in a newer version
[11:11] <walters> that's true
[11:11] <Uraeus> walters: my current CVS version seems to support loading Ogg metadata, but not mp3 metadata, is that correct?
[11:12] <walters> maybe once all this is stable and works, we can think about supporting upgrades in that way
[11:12] <Uraeus> ok
[11:12] <Uraeus> are you having trouble with the mp3 metadata support?
[11:12] <walters> mp3 works fine here
[11:13] <walters> although all my music is fairly pristine
[11:13] <walters> in terms of like not being filled with huge pictures, encapsulated in .wav, having broken charsets...
[11:14] Action: walters sighs
[11:14] <Uraeus> ok, I update CVS again :)
[11:14] <Uraeus> I just noticed you said in some CVS commit that you disabled mp3 support for now or something
[11:15] <walters> writing mp3s, yeah
[11:15] <walters> i need to sync up cvs again, one sec
[11:15] <Uraeus> walters: that didn't work as announced?
[11:15] <walters> no...
[11:15] Action: Uraeus had been hoping to use RB to clean up my tags
[11:15] <walters> i've tracked down the issue, but i'm not quite sure what the best way to fix it is
[11:15] <Uraeus> walters: have you told that to Company?
[11:15] <walters> so i'll talk to Company
[11:15] <walters> yeah
[11:16] <Uraeus> ah
[11:16] thomasvs ([email protected]) joined #gstreamer.
[11:18] <Uraeus> morning thomasvs
[11:18] <thomasvs> morning
[11:19] <Uraeus> thomasvs: did you get my SMS?
[11:19] walters ([email protected]) left irc: "z"
[11:20] Nick change: bluejay|gone -> bluejay|sleep
[11:20] <thomasvs> nope
[11:20] <thomasvs> about ?
[11:21] Nick change: harshy -> harshyZ
[11:29] <thaytan> yo yo
[11:30] <Uraeus> hi thaytan, back from the conference?
[11:30] <thaytan> yep, back to work in the morning
[11:30] <thaytan> it was a great week
[11:30] <Uraeus> thaytan: looked at some pictures, people seemed to be having a good time :)
[11:31] <thaytan> it is awesome every year :)
[11:31] <thaytan> I'm trying to finish off dvd subtitles and overlays
[11:31] <thaytan> a few event things and colour tables not working
[11:32] <Uraeus> cool, and I hope you got Malcomn to promise to commit his SMIL library when he got home from the conference :)
[11:32] <thaytan> I did hit him up about it
[11:32] <Uraeus> good :)
[11:32] <thaytan> he said he wasn't happy with it
[11:33] <thaytan> and I told him that if he put it up somewhere, we'd help him finish it
[11:33] <thaytan> I hadn't actually internalised the fact that he lives about 20 minutes from here, even though I know I knew it
[11:34] <Uraeus> thaytan: from the conference or from you?
[11:34] <thaytan> from me
[11:34] <Uraeus> ah good, then you can walk over at hit on his door whenever we need something done ;)
[11:35] <thaytan> heh, yeah
[11:35] <thaytan> we bought half our christmas presents in the shopping centre below his apartment
[11:40] Action: thomasvs syncs up his local copies again in a hope to commit
[11:42] <Uraeus> thomasvs: the patch you sent me missed a file
[11:42] <thomasvs> Uraeus: yeah, I know
[11:43] <thomasvs> sigh :)
[12:11] ds-uhv ([email protected]) left irc: Read error: 110 (Connection timed out)
[12:12] thomasvs ([email protected]) left irc: Read error: 113 (No route to host)
[12:16] herzi_lap ([email protected]) left irc: "Leaving"
[12:16] alley_cat ([email protected]) joined #gstreamer.
[12:29] herzi_lap ([email protected]) joined #gstreamer.
[12:29] michael_-_ ([email protected]) left irc: Read error: 60 (Operation timed out)
[12:38] kmaraas ([email protected]) left irc: "Leaving"
[12:41] dolphy ([email protected]) joined #gstreamer.
[12:41] <dolphy> morning :)
[12:42] <mathrick> hi dolphy
[12:50] pb_ ([email protected]) joined #gstreamer.
[13:26] Company ([email protected]) joined #gstreamer.
[13:29] kmaraas ([email protected]) joined #gstreamer.
[13:33] Action: Company nukes float audio
[13:34] <Company> morning btw
[13:35] <Uraeus> morning Company
[13:35] <mathrick> yo Company
[13:35] <Uraeus> mathrick: how goes the mms support?
[13:38] <mathrick> Uraeus: CS project deadline 23:59 GMT today, and after that I only have exams tomorrow I probably wouldn't be able to take anyway, so monday seems like fine day to sort out all that piled up things :)
[13:39] <mathrick> Uraeus: I honestly hope to be productive for gst from tomorrow on
[13:40] <Uraeus> ok :)
[13:41] <mathrick> Uraeus: and if you were asking about xine guys, well, most of them are okay, there are some of them hard to contact tho
[13:43] <Uraeus> mathrick: do you know who wrote the libmms stuff you started from?
[13:43] <Company> MajorMMS most likely ;)
[13:44] <mathrick> Uraeus: one of current xine devs
[13:44] <mathrick> Uraeus: ie. one of them is MajorMMS
[13:46] <Uraeus> ah :)
[13:46] <Uraeus> have you talked to him about LGPL licensing it?
[13:48] <mathrick> Uraeus: in case you haven't noticed, he is anonymous, so no chance to talk to him as "MajorMMS", aside from me not knowing his personality ;)
[13:48] <mathrick> Uraeus: but xine guys told me he *is* among ones i asked, so i guess it's okay
[13:49] <Uraeus> mathrick: ok, and all those you asked said yes?
[13:49] <mathrick> Uraeus: basically, yep, one of them required some more work tho ;)
[13:50] <mathrick> Uraeus: as he said himself, his contribution was negligible (3 lines), but it's nice to have single voted approval :)
[13:51] <taaz> srd
[13:51] <mathrick> Uraeus: he seemed to be confused a little about my intents, and range of LGPL too, but after all he said to go for it
[13:51] <taaz> oops
[13:52] <mathrick> Uraeus: btw, he's one of major contributors to libdvdnav
[13:52] <Uraeus> mathrick: which is LGPL I think?
[13:52] <Uraeus> hi taaz, a bit confused today? )
[13:52] <mathrick> Uraeus: *bzzz* wrong. GPL
[13:53] <Uraeus> mathrick: ah, well to be honest I don't think we 'ever' get the DVD stuff LGPL'ed anyway, all the current libs are GPL and some of them has told me clearly that they are under no circumstance willing to relicense
[13:54] <mathrick> Uraeus: that's why dvdnavsrc is marked GPL
[13:54] <mathrick> Uraeus: although plugins source is LGPL
[13:54] <mathrick> s/plugins/plugin's/
[13:54] <Uraeus> yup
[13:54] <taaz> same with mpeg2dec
[13:55] kniteshade ([email protected]) left irc: "I QUIT!"
[13:55] <Uraeus> well I still belive in adocating the LGPL to developers, since it is always easier to replace 3 libs than 5; and so on
[13:55] <mathrick> Uraeus: i was confused when BBB explained it to me, no wonder this guy wasn't too comfortable about gst being LGPL and using libdvdnav initially
[13:56] <Uraeus> mathrick: yeah, it is a confusing situation :)
[13:56] <mathrick> oh, btw, do anyone know of way to force gdb into loading given source file?
[13:56] <mathrick> it refuses to load source for dlopen()d symbols, which sucks :\
[13:57] <Company> huh?
[13:59] <mathrick> Company: it's CS related. It dlopen()s plugins for everything, and apparently it doesn't show me source line for them, which makes debugging crashes inside mesh loader impossible
[13:59] <mathrick> Company: of course, I did debug build of CS
[14:00] <Company> considering gstreamer dlopens all plugins and I always have the symbols there....
[14:00] <mathrick> Company: 'info sharedlibs' shows them as having symbols, but bt is pretty much useless anyway
[14:01] <Company> use g_module ;)
[14:02] <Company> or look at how they do the dlopen there at least
[14:02] <mathrick> Company: what could be the difference?
[14:03] <Company> no idea
[14:03] <Company> but there must be one if gst symbols are there and yours are not
[14:05] <thaytan> how can the plugin be LGPL if dvdnav is GPL ?
[14:06] <thaytan> surely dvdnav it counts as a derived work and hence requires GPL?
[14:06] <Company> yes
[14:06] <thaytan> let me try that sentence again
[14:06] <Company> but the code of the plugin is LGPL
[14:06] <mathrick> thaytan: it's binary license vs. source license
[14:06] <Company> so gstdvdnavsrc.c is LGPL and gstdvdnav.so is GPL
[14:07] <thaytan> I thought the GPL says that a derived work must be GPL
[14:07] <thaytan> and clearly dvdnavsrc is a derived work - it links to libdvdnav
[14:07] <Company> gstdvdnavsrc.c doesn't
[14:09] <mathrick> IMO it's fishy to treat linking as deriving
[14:09] <Company> it's fishy anyway
[14:09] <mathrick> this way, running on linux should be considered as deriving
[14:09] <Company> because the term "linking" isn't defined well enough
[14:10] <mathrick> yep, and there's where brilliant US law comes into play :\
[14:10] <Company> but the problem isn't linking anyway, the roblem is distributing
[14:10] <Company> and that's covered quite well
[14:11] <mathrick> Company: well, exact terms of distribution depend on whether you link, as this may make your binary GPL
[14:12] <Company> yes
[14:12] <Company> but that never was the problem
[14:13] <Company> the problem always was redistributing a GPL app without source
[14:13] <mathrick> Company: could you elaborate?
[14:13] <Company> see the linksys, KISS and whatever cases
[14:14] <Company> i don't know of any GPL case where the method of linking was the problem
[14:15] <thaytan> So, I have a navigation question
[14:16] <thaytan> dvdnavsrc receives a lookup table of colour values
[14:16] <thaytan> an array of 16 * 32bit ints
[14:16] <thaytan> and I need to ship it to mpeg2subt so it can use it for the subtitle overlays
[14:17] <thaytan> I'm placing it in a gst-structure as fields 'clut0' thru 'clut15', but this is somewhat suboptimal
[14:17] <mathrick> Company: and what about closed source app using dvdnacsrc? it's gpl, so makes the app violate it, right?
[14:17] <thaytan> it's a "get it working and reexamine later" issue, but any ideas are welcome
[14:17] <Company> mathrick: yes
[14:18] <Company> thaytan: the "$format needs setup data" problem is unsolved and i haven't found a better way to do it yet...
[14:18] <mathrick> Company: it's shit, by no means runtime linking should ever affect app legality
[14:19] <Company> mathrick: why not?
[14:19] <Company> mathrick: that'd mean you could take a GPL lib, modify the API to suit your needs and use it
[14:20] <thaytan> Company: the ability to attach an opaque type to an navigation event would help a bit here
[14:20] <mathrick> Company: huh?
[14:20] <thaytan> but this will work for now
[14:21] <Company> mathrick: take Linux kernel, add syscalls, make your closed source app use them, distribute
[14:22] <mathrick> Company: so what? all the source i modified is redistributed, and i really don't see 'calling binary object' as 'deriving'
[14:22] <Uraeus> well the plugin source is LGPL, the linked binary is GPL; I think it is pretty simple
[14:22] <Company> Uraeus: the plugin source does an #include <dvdnavsrc.h> - isn't that linking to GPL?
[14:23] <Uraeus> Company: the plugin source is just text, no more no less, it isn't until compilation you actually link
[14:23] <Company> mathrick: the modification might just be exporting 5 functions and noone knows why...
[14:23] sublett ([email protected]) joined #gstreamer.
[14:23] <Company> Uraeus: an .so file is just text, too, it doesn't contain GPL code
[14:23] <mathrick> Company: this way, I can come up with MOGL (My Own Great License), make a gst plugin that uses it, feed into gst using app, and voila, we have MOGL violation b/c apps didn't know MOGL is not to be touched
[14:24] <Company> Uraeus: both contain just a reference to GPLed stuff
[14:24] <Company> mathrick: i know, which is why gstreamer is designed to take plugin licenses into account
[14:24] <Company> mathrick: you'll not be able to open a plugin with license set to "MOGL" in current gst
[14:25] <Uraeus> Company: well the point is that when you run it you need to comply with the license of ALL the code, which means a) they all need to compatible in their licensing and b) the strictest license 'wins'
[14:26] <Company> Uraeus: yes, but that doesn't explain why the plugin's so needs to be GPL
[14:26] <mathrick> Company: mathrick: you'll not be able to open a plugin with license set to "MOGL" in current gst <-- why?
[14:26] <mathrick> Company: does that mean my app has to know every compatible license it could use?
[14:26] <Uraeus> Company: well I guess it is a question of at what point it stops being a separate entity and becomes part of a bigger unit
[14:27] <Company> mathrick: yes - you could use "other" or "undefined" (dunno exact name), bnuit that requires you to check if your license is compatible with what "other" menas
[14:27] <Company> Uraeus: exactly :)
[14:27] <Company> Uraeus: but it's not as clearcut as people make it
[14:28] <mathrick> Company: it's even more shit then
[14:29] <Uraeus> well in pratical terms it is, you know that you can take the file which includes the LGPL in its header and use as part of another LGPL piece of software and the whole is still LGPL, even if the file you took which had the LGPL header came packaged together with a bunch of GPL files
[14:29] <Company> mathrick: you're free to request adding support for MOGL to gstreamer :)
[14:30] <Uraeus> and the same is true for the opposite, you know that if you put 1 GPL file into a package with 10 LGPL files the whole will be considered GPL until someone replace that 1 GPL file with a LGPL one
[14:30] <Company> Uraeus: if I take GPL'ed .so plugin and link it to my own library, it still reports itself as GPL - that's a bit odd, isn't it? ;)
[14:31] <Uraeus> Company: then you need to update the factory information of that plugin :)
[14:32] <Company> Uraeus: see? for plugins you update afterwards, for source files you update before
[14:32] <Uraeus> Company: you would need to change the factory information in the source feil, there is no updating afterward
[14:32] <Uraeus> s/feil/file/
[14:33] <Company> Uraeus: yes, but why don't we set it to LGPL from the start? We dop this with source files, too
[14:34] <Company> Uraeus: both are a seperate LGPLed entity that "link" to GPL files, just in different ways
[14:34] <Uraeus> Company: the whole point of the factory information field is to tell application developers the 'result of plugin license+ library license' license of a plugin they might want to use
[14:34] <Uraeus> Company: if we changed it to be the license of the actaull plugin code it would be worthless
[14:36] <Company> so the plugin's .so file is LGPL or GPL?
[14:36] <Uraeus> I would say GPL
[14:36] <Uraeus> but I would also say that the question is rather academic
[14:37] <Company> why is a .so that links to GPL'd code GPL while a source file is not?
[14:37] <Company> the .so in itself only includes LGPL'd code
[14:38] Action: Company is just out to proof that it's not clearcut and confuse everyone
[14:38] Rotty ([email protected]) joined #gstreamer.
[14:39] <Uraeus> Company: it is clearcut enough, your arguments are highly theorectical, how often are you faced with a binary compatible LGPL library replacement for a GPL one?
[14:39] <Company> Uraeus: so source code that links stays LGPL, byte code that links is GPL
[14:40] <Company> ?
[14:40] <Company> Uraeus: what about interpreted Python?
[14:40] <Company> Uraeus: or intermediate languages?
[14:40] <thaytan> g'night, pplz
[14:40] <Company> gnight
[14:41] <mathrick> nite thaytan
[14:41] <mathrick> Company: of course it's fuzzy, and that's why I say it's shit
[14:41] <mathrick> <Company> why is a .so that links to GPL'd code GPL while a source file is not?
[14:41] <mathrick> Company: that's what I'm asking, why?
[14:42] <Company> noone knows :)
[14:43] foser ([email protected]) joined #gstreamer.
[14:44] <Uraeus> it is called legal precedent :)
[14:44] <Company> and then there's the readline story, too
[14:45] <Uraeus> and the no file 'become' GPL, it is just the whole that is to be treated as such
[14:45] <mathrick> Company: OTOH, closed app's vendor only distributes binary which is *not* linked to GPL'd plugin, so he doesn't violate GPL when somewhen in the future pipeline constructed contains GPL'd plugin
[14:46] <Company> mathrick: so there's an effective way to circumvent the GPL by just writing a plugin system and dlopening the code in question?
[14:46] <mathrick> GPL is only binary resulting in running that app, which isn't the one distributed by vendor :)
[14:46] <mathrick> Company: well, yes
[14:46] <mathrick> Company: if you just use that code, you're not deriving it
[14:46] <Company> depends
[14:47] <mathrick> Company: it's no different if you use it by system("my.gpl.app"), or dlopen("libmygpl.so"), after all it's yout app's user who initiates linking
[14:47] <Company> the GPL's intent is to say that deriving means using
[14:48] <mathrick> Company: and GPL grants that user right to use GPL'd code, that's for sure
[14:48] <Company> yes
[14:48] <Company> it's about distributing
[14:49] <mathrick> Company: vendor distributes app, we distribute gst. No problem here :)
[14:49] <mathrick> Company: and in case of appliances like KISS, it's also easy
[14:50] <Company> to comply with your view they just distribute the mplayer original
[14:50] <mathrick> Company: vendor just distributes appliance, with two binaries loaded. And it's user that requests linking by pressing "play"
[14:50] <Company> yeah
[14:50] <mathrick> so no GPL is violated IMHO
[14:50] <Company> which is not what the GPL is about
[14:51] <mathrick> GPL is about freedom, isn't it?
[14:51] <Company> yes, about developer's freedom
[14:51] <Uraeus> mathrick: depends on how you define freedom :)
[14:51] <Company> if I write teh suck media player and GPL it
[14:52] <mathrick> and I don't consider outlawing apps using gpl'd plugin any more fair than doing it by patents, or DMCA subpoening
[14:52] <Company> and KISS puts it into it's dvd player
[14:52] <Company> and i buy that player
[14:52] <Company> i want to be able to hack it
[14:52] <Company> so I need the full source of the firmware
[14:52] <mathrick> Company: if it uses *your* player, just do "cvs co" from your tree ;)
[14:53] <herzi_lap> .)
[14:53] <Company> mathrick: if they changed subtle parts of it that doesn't work...
[14:53] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[14:53] <mathrick> Company: it's vilolation only if they modified source of your player before compiling and uploading it
[14:53] <mathrick> Company: if they changed any of source, it's clearly violation
[14:54] <Company> mathrick: they didn't, they just wrote a binary patching app that patches the binary to include their code
[14:54] <mathrick> but linking is not "modifying source"
[14:54] <Company> mathrick: which is done by the user when he installs new firmware
[14:56] <mathrick> Company: here it's tricky, but court will probably say it's no go then, because patch has absolutely no sense w/o your exact app. So it's bit different IMO than just using library, or dlopen()ing plugin
[14:56] <Company> mathrick: that's why they say the library .so and the source .c file need to be GPL, too - because both don'T make sense with the GPLed lib
[14:57] <mathrick> Company: no, it's different
[14:57] <mathrick> Company: making ABI compatible lib to replace GPL'd one has sense
[14:57] <mathrick> Company: replacing patched app so that patch can be applied makes no sense
[14:58] <Company> mathrick: so it's about "making sense" ?
[14:58] <mathrick> Company: does your player app export API?
[14:58] <mathrick> Company: no, but lib does
[14:58] <mathrick> hmm, bbl
[14:58] Nick change: mathrick -> mathrick|food
[15:02] Rotty_ ([email protected]) joined #gstreamer.
[15:03] Rotty_ ([email protected]) left irc: Client Quit
[15:06] dolphy ([email protected]) joined #gstreamer.
[15:11] Nick change: mathrick|food -> mathrick
[15:11] <mathrick> b
[15:11] dolphy ([email protected]) left irc: Client Quit
[15:12] <mathrick> Company: the bottom line is, patch isn't "using" your player, but some player is "using" plugin
[15:13] <Company> so the correct way would be to write a binary that wraps the player, so that it "uses" it?
[15:14] <mathrick> Company: no, because player exports no API, it's not meant to be used except by system()
[15:14] <mathrick> Company: this is not the case with lib
[15:14] <Company> the lib isn't meant to be used by proprietary apss either
[15:16] <Company> s/proprietary apss/non-GPL apps/
[15:22] <ChrisHJW> http://lists.matroska.org/pipermail/matroska-general/2004-January/000675.html  <---- this is more than i can understand i fear
[15:23] <ChrisHJW> how can Real release Helix under the GPL, if its linking to closed code ?
[15:23] <mathrick> hmm, I don't think lib (or better yet plugin) could ever invalidate app, because it means *me* writing some GPL/MOGL'd lib requires *someone else* to take defensive steps against violating my license *after* he/she has written it. IMO it clearly goes against his/her freedom, as well as rule of presumed good will
[15:24] <herzi_lap> ChrisHJW, this is what debian puts into contrib (free software depending on proprietary sw)
[15:24] <ChrisHJW> herzi_lap ? is it possible, or not ? is the FSF clear on this ?
[15:24] <herzi_lap> it is still free software (but with aestetical errors)
[15:24] <ChrisHJW> lol
[15:24] <ChrisHJW> well
[15:25] <herzi_lap> isn't maplyer and the win32 binary stuff similar?
[15:25] <ChrisHJW> one could argue, as long as there is not even a single of their codecs released under GPL, its a clear violation
[15:25] <ChrisHJW> herzi_lap : yes, thats what i had to learn hard
[15:25] <Company> _if_ you write a GPL something and someone writes a plugin for it, that plugin _must_ be GPL
[15:26] <ChrisHJW> i had no idea, and dropped in #videolan to convince the guys to use the Real DLLs for RV9 playback .... they kicked my ass so hard, it hurt for hours
[15:27] <Company> if that something has another license than GPL it's ok though if you comply with that license
[15:27] <mathrick> Company: but if _my_ app uses LGPL framework, and _someone_ writes GPL plugin for it, does it make _me_ violate GPL?
[15:27] jowenn ([email protected]) joined #gstreamer.
[15:27] <Company> mathrick: no
[15:27] <jowenn> hi
[15:28] <Company> mathrick: it just makes you violate the GPL if you ship your app with the framework _and_ the offending plugin
[15:28] <jowenn> Is it possible to use gstreamer for mp3 playing from an non OS application ?
[15:29] <mathrick> Company: why? then I just ship two separate products - framework w/ plugins, and app using that framework
[15:29] <Company> mathrick: that's ok
[15:29] <Company> mathrick: but Redhat for example may not ship proprietary gst apps and GPL plugins in one distribution
[15:30] <mathrick> jowenn: heh, you just came when we're discussing that issues :)
[15:30] <mathrick> Company: why?
[15:32] <ChrisHJW> actually, thats how we could circumvent problems to load closed DLLs in GPLed players, make a L-GPL loader or wrapper for the closed stuff
[15:32] <ChrisHJW> but its not in the sense of the GPL, for sure
[15:32] <Company> mathrick: because you link the GPL'ed code to proprietary code _in_your_package
[15:33] <mathrick> Company: no, i ship 3 packages. framework, plugins, and app that uses framework and _potentially_ any of the plugins
[15:33] <Company> mathrick: in that case it would be ok
[15:34] <mathrick> Company: how does it make me violate GPL?
[15:34] <Company> mathrick: but Redhat ships one package - Redhat fedora, on one cd
[15:34] <Company> (or Redhat Enterprise or whatever)
[15:34] <mathrick> Company: no, it ships hella lots of packages, on one cd
[15:34] <Company> mathrick: the CD is one package
[15:35] <mathrick> Company: this way, no use of OpenSSH either
[15:35] <mathrick> Company: and X
[15:35] <Company> they're GPL compat, no?
[15:37] <Company> jowenn: GStreamer is LGP, so anything the LGPL allows is ok
[15:37] <Company> LGPL even
[15:39] <jowenn> from what I see the LGPL should allow proprietary apps, but aren't you using libmad, that on the other side is GPL, or am I mistaken ?
[15:40] <Company> yes
[15:40] <Company> but ffmpeg has an LGPL mp3 decoder
[15:43] <mathrick> Company: i think they're X licensed, at least it's not standard GPL, although may be compatible
[15:43] <jowenn> So if I use ffmpeg and feed the decoded file to gstreamer, I should not have a problem
[15:44] <Company> jowenn: right, if you don't explicitly request GPLed plugins, there shouldn'T be a problem
[15:45] <mathrick> jowenn: we're just discussing if it's okay for you to use mad plugin that uses GPL libmad. IMO it is okay
[15:45] <alley_cat> to be on the safe side just use the ffdec_mp3 plugin
[15:45] <alley_cat> it's faster anyway
[15:46] <Company> but it sounds like half the real bitrate ;)
[15:46] <mathrick> Company: as I see it, 'package' should be result of compilation of single source package, it'd be more or less equivalent to single .rpm
[15:46] <alley_cat> huh? it sounds exactly like mad here
[15:47] <jowenn> I don't see that plugin at http://www.gstreamer.net/status/?category=5
[15:48] <Company> alley_cat: for normal playback it's ok, but if you have more demanding sounds, you clearly hear it
[15:48] <alley_cat> i just used it for mp3 decoding in movie playback so far, for that it's ok
[15:48] <Company> mathrick: ask a judge about that, but I don't think it'll hold
[15:49] <Company> alley_cat: sure it's ok, but you can hear it
[15:49] <Company> jowenn: that list is totally outdated - we're lacking website maintainers....
[15:50] <mathrick> Company: this way, you can't ship GPL apps on the same cd as X ones, and as BSD ones, and MPL probably needs its own cd
[15:50] <Company> mathrick: if they link to each other you can't
[15:51] <mathrick> Company: ok, but that has nothing to do w/ being shipped in single package
[15:51] <Company> mathrick: i think it has
[15:52] dolphy ([email protected]) joined #gstreamer.
[15:52] <Company> mathrick: you can't simply use gstreamer to get around the "app must be GPL" requirement of mad
[15:53] jowenn ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[15:53] <mathrick> Company: i'm not getting around it, I think mad has no right to put these requirement in the first place, at least not with GPL
[15:53] <Company> mathrick: it has every right to do that
[15:53] <Company> mathrick: at least as long you convince judges otherwise
[15:54] Action: dolphy tried to get radeonfb to work with no success
[15:54] <mathrick> Company: but then, its not free
[15:55] <Company> mathrick: huh?
[15:57] jowenn ([email protected]) joined #gstreamer.
[15:57] <mathrick> Company: GPL is about preserving freedom of use, and freedom of source code. Locking someone out of use is not preserving freedom
[15:58] <Company> mathrick: it's a clear intent of the GPL to lock out every code that is not free
[15:59] <mathrick> Company: anyway, this whole discussion is a proof that GPL shouldn't be used for libs :)
[16:00] <Company> mathrick: it's proof that using GPL for libs is fishy
[16:00] <mathrick> Company: yes, that's what I meant
[16:02] <jowenn> Is it possible to crossfade songs with gst ?
[16:03] <Company> jowenn: it's not yet implemented
[16:03] <Company> jowenn: but it should not be hard to implement
[16:03] <mathrick> jowenn: 'possible' yes,. If someone implemented element for that - Markey- wanted to do that
[16:03] <mathrick> but I don't know how soon is he willing to do that
[16:06] hyriand ([email protected]) left irc: Read error: 110 (Connection timed out)
[16:06] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[16:07] <jowenn> I see
[16:09] herzi_lap ([email protected]) left irc: "Leaving"
[16:20] dolphy ([email protected]) joined #gstreamer.
[16:22] <dolphy> Company: a pipeline question
[16:23] <dolphy> Company: when should i put stuff in a thread exactly ?
[16:23] <dolphy> Company: for audio and video output i think they deserve their own thread
[16:24] <dolphy> Company: but for example src ! spider !  and then goom a tee, a switch
[16:24] <dolphy> Company: all of that can seat in the main thread or not ?
[16:25] <Company> dolphy: there's no reason for more than 3 threads in a normal video player
[16:25] <Company> dolphy: 1 thread for audio, 1 thread for video and 1 thread where it isn't yet clear if it's audio or video
[16:25] hyriand ([email protected]) joined #gstreamer.
[16:25] <Company> dolphy: the thread where it isn't clear should be as small as possible
[16:27] <dolphy> Company: hmmm
[16:28] <dolphy> Company: it's not that simple
[16:28] <dolphy> Company: basically just behind spider i need to connect a tee
[16:28] <dolphy> Company: for audio
[16:29] <dolphy> Company: on the first src pad of the tee i will have goom ! ffcolorspace and then it connects to switch
[16:29] <dolphy> Company: on the second pad i will have a connection to the audio thread
[16:29] <dolphy> Company: which contains queue ! osssink
[16:29] <dolphy> Company: on the video pad of spider i will have ffcolorspace and then switch
[16:29] <Company> dolphy: that sounds pretty clear to me:
[16:30] <dolphy> Company: and after switch the video thread
[16:30] <Company> dolphy: the tee is thread 3 (audio may still go to videosink or audiosink)
[16:30] <Company> dolphy: goom and ffcolorspacew are clearly video only
[16:33] <dolphy> Company: so how many threads ?
[16:33] <Company> dolphy: 3
[16:34] <Company> (unless that doesn't work for some reason)
[16:34] <dolphy> Company: thread 1 contains : src spider and ?
[16:34] <Company> dolphy: tee
[16:34] <dolphy> Company: and where goom goes ?
[16:36] <Company> dolphy: goom is video
[16:36] <dolphy> Company: well yeah but that would mean two entrances at the video thread
[16:37] <dolphy> Company: so 2 queues
[16:37] <Company> yeah
[16:37] <dolphy> Company: one for tee src1 and one for spider video src
[16:37] BBB ([email protected]) joined #gstreamer.
[16:38] <dolphy> Company: so video thread would have 2 queues, 3 colorspaces, goom, switch , videobalance, videoscale and ximagesink
[16:38] <Company> dolphy: yeah
[16:39] <Company> dolphy: if that leads to lags, you could make another thread for just ximagesink
[16:43] dolphy ([email protected]) left irc: Read error: 60 (Operation timed out)
[16:49] dolphy ([email protected]) joined #gstreamer.
[16:49] <dolphy> ggrrr adsl connection died
[16:55] <Uraeus> dolphy: btw, for me with gst-player or totem mpeg movies play fine for 4-5 seconds then they grind to a halt, while playing with gst-launch-ext works perfectly
[16:58] <dolphy> Uraeus: gst-launch-ext is not a good way to compare
[16:58] <dolphy> Uraeus: it does not use spider
[16:58] <dolphy> Uraeus: and builds very simple pipelines compared to libgstplay one
[16:58] <Uraeus> ah
[17:01] <Uraeus> dolphy: but we are sure the issues are caused by spider and not the scheduler?
[17:01] <dolphy> Uraeus: i don't think the scheduler has anything to see with that
[17:01] <dolphy> Uraeus: for example for me matrix.avi works in gst-totem
[17:01] <dolphy> Uraeus: but makes a segfault with gst-launch
[17:01] <BBB> hm
[17:01] Action: BBB will help fixing those in a week or so
[17:01] <BBB> first ffmpeg...
[17:01] <dolphy> Company: what's the last message you ve seen from me ?
[17:02] <Company> <dolphy> Uraeus: but makes a segfault with gst-launch
[17:02] <Company> :p
[17:02] <Company> <Company> dolphy: if that leads to lags, you could make another thread for just ximagesink
[17:02] <BBB> <dolphy> Company: so video thread would have 2 queues, 3 colorspaces, goom, switch , videobalance, videoscale and ximagesink
[17:02] <BBB> that
[17:02] <Company> that was the last message before your disconnect
[17:06] BBB ([email protected]) left irc: "Client exiting"
[17:07] <Company> vorbis sucks
[17:07] <Company> apparently noone ever tried to use anything but vorbisfile
[17:07] <Uraeus> so our vorbismux will be the first?
[17:08] <Company> there is no vorbismux, it's vorbisenc/dec
[17:08] <Company> but apart from that it looks like it
[17:12] <Uraeus> Company: what I meant was that when we implement a vorbismux we will be the first ever to do so, as vorbisfile afaik do not contain one right?
[17:13] <Company> there is no vorbismux
[17:13] <Company> and no, vorbisfile is a simple take-ogg-vorbis-file-output-sound thing anyway
[17:13] <Uraeus> Company: I thought you said we needed a vorbismux to get tag editing working?
[17:13] <Company> oggmux
[17:14] <Uraeus> ah
[17:14] <Uraeus> so our oggmux would be the frist ever then :)
[17:14] <Company> we need oggdemux and vorbisdec to replace vorbisfile
[17:14] <Company> and we need vorbisenc and oggmux for vorbis encoding
[17:14] <Company> and oggdemux, vorbistag and oggmux for tag changing
[17:15] <Uraeus> and currently we have none of those
[17:15] <Company> we have oggdemux and vorbistag
[17:15] <Company> and i'm trying to get a vorbisdec working
[17:15] <Uraeus> nice :)
[17:16] <dolphy> http://getcontrol.sourceforge.net/files/gstplay_pipeline.png
[17:16] <dolphy> Company, BBB can you take a look at this pipeline
[17:16] <dolphy> please
[17:17] <dolphy> Uraeus: do you know how i am supposed to add documentations to libgstplay such as this DIA diagram representing the pipeline ?
[17:18] <Uraeus> dolphy: seem we currently use XFIG, not sure how the 'source' for that is created
[17:18] <Company> dolphy: the volume lags far behind probably
[17:19] <Company> but other than that it looks quite ok
[17:19] <Company> (well, maybe apart from the fact that only I420 works without colorspace
[17:21] <Company> and you probably need quite some audioconvert plugins in the audio pipeline, too
[17:21] <dolphy> Company: audioconvert ?
[17:21] <dolphy> Company: what for ?
[17:21] <Company> at least after spider and in front of the visualizer
[17:21] <dolphy> Company: ah for mono output for example ?
[17:21] <Company> yeah
[17:22] <Company> or wav
[17:22] <Company> wav is little endian only
[17:23] <dolphy> i see
[17:23] <dolphy> audio convert works ?
[17:24] <Company> it should
[17:25] <dolphy> the reason why i put volume there
[17:25] <dolphy> is that i want volume to influence goom output
[17:31] <dolphy> http://getcontrol.sourceforge.net/files/gstplay_pipeline.png
[17:31] <dolphy> reload
[17:31] <jowenn> Is it possible to add an additional source to an adder, without interruppting the pipline
[17:33] <dolphy> Company: what about that I420 yuv thing ?
[17:33] <mathrick> dolphy: hmm, isn't it possible that audio sink would take different format than goom?
[17:34] <dolphy> mathrick: what do you mean ?
[17:35] <mathrick> dolphy: ie. shouldn't audioconvert be after tee?
[17:35] <dolphy> mathrick: on the way to goom ?
[17:36] <dolphy> mathrick: letting audio sink without any audio convert in between 
[17:36] <dolphy> ?
[17:36] <mathrick> dolphy: maybe on the way to audio sink too, depends on how tee handles nego
[17:37] <mathrick> hmm, volume is the one to make nego, right
[17:38] <mathrick> dolphy: bottom line is, i don't think we can assume that volume caps == goom caps == audio sink caps
[17:40] <dolphy> mathrick: goom and audio sink can probably interesect a common one
[17:40] <dolphy> mathrick: proxied by tee
[17:40] <mathrick> cute, galeon eats up some dirs when opening through file selector
[17:40] <dolphy> mathrick: and then audioconvert will transform what comes out of spider to that
[17:40] <mathrick> dolphy: probably? or are you really really sure?
[17:41] <dolphy> mathrick: if i was sure i would not loose time drawing the pipeline in dia to get thoughts :)
[17:41] <Company> if you'Re not sure, you need audioconvert
[17:42] <mathrick> dolphy: and remember there's volume besides goom and sink, too. It also needs to be able to intersect caps
[17:42] <dolphy> Company: what do you think about the new image ?
[17:43] <jowenn> is anonymous access to the sf repository disabled ?
[17:44] <dolphy> jowenn: sf repository is obsolete
[17:44] <dolphy> jowenn: use freedesktop
[17:44] <jowenn> Even the freedesktop page links to the sf one
[17:44] <Company> dolphy: I think it's ok
[17:44] <jowenn> what's the repository path ?
[17:44] <dolphy> jowenn: not sure it works anonymous
[17:45] <dolphy> jowenn: cvs.freedesktop.org:/home/cvs/gstreamer
[17:45] <Company> it does, but i don't know the path either
[17:46] Action: Company blames thomasvs for not updating the website
[17:46] <dolphy> jowenn: probably :pserver:[email protected]:/home/cvs/gstreamer
[17:48] <jowenn> found it, :pserver:[email protected]:/cvs/gstreamer works
[17:48] <Uraeus> I updated the developer page on the website to point to freedesktop
[17:49] <dolphy> Company: but that lets some audio elements in the main thread :)
[17:49] Rotty ([email protected]) left irc: Remote closed the connection
[17:51] <Company> dolphy: which one?
[17:51] <dolphy> Company: volume
[17:51] <dolphy> Company: audioconvert
[17:51] <Company> dolphy: they're not audio since they affect vdeo, too
[17:51] <dolphy> Company: ok
[17:52] <dolphy> Company:  i m implementing that pipeline then
[17:56] <mathrick> dolphy: 'pipeline' is no longer valid term for that beast ;)
[17:59] Nick change: harshyZ -> harshy
[18:00] <dolphy> mathrick: hehe :)
[18:01] <Company> i'd like to be able to switrch off all that additional "junk" somehow
[18:07] LeRoutier ([email protected]) joined #gstreamer.
[18:07] <LeRoutier> hello
[18:08] <mathrick> hi LeRoutier
[18:09] <Company> greetings from the other side of the rhine
[18:13] <dolphy> Company: ???
[18:13] <dolphy> Company: what additional junk are you talking about ? :)
[18:13] <Company> dolphy: videobalance, videoscale, goom, volume
[18:13] <dolphy> Company: videobalance is passthru
[18:14] <dolphy> Company: videoscale is passthru too
[18:14] <dolphy> Company: goom is not connected by default
[18:14] <dolphy> Company: the tee pad gets connected on request
[18:15] <Company> dolphy: videobalance forces conversions in ffcolorspace for a start
[18:15] Shoragan ([email protected]) joined #gstreamer.
[18:15] <Company> dolphy: switch is loopbased
[18:15] <dolphy> Company: and ?
[18:16] <Company> dolphy: volume forces conversions
[18:16] <Company> dolphy: and all of that must be taken into account when debugging
[18:17] <dolphy> Company: well i understand what you mean but libgstplay is intented to be a fullfeatured video/audio player library
[18:17] <Company> dolphy: debugging decoders, demuxers, clocking etc are a lot harder when there's 20 additional elements
[18:17] <dolphy> Company: so it has to be able to set colorbalance, do software video scaling and audio conversion, volume etc..
[18:18] <Company> dolphy: yeah, i'd just say it'd be nice if that could be switched off
[18:18] <dolphy> Company: any idea how ?
[18:19] <Company> dolphy: nope - i don't have that much clue about gstplay internals anyway
[18:19] <dolphy> Company: well that sounds like tough thing
[18:19] <Company> dolphy: i wasn't talking about switching off while playing anyway
[18:23] <dolphy> Company: well we'll see that later :-)
[18:24] <mathrick> dolphy: vis part could be switched off instantly, couldn't it?
[18:25] Marsupilami23 ([email protected]) joined #gstreamer.
[18:26] <dolphy> mathrick: yes
[18:26] <dolphy> mathrick: but switching off videobalance and videoscale out of the pipeline is tricky
[18:26] <dolphy> mathrick: you can make them passthru easily
[18:26] <dolphy> mathrick: but that's different
[18:27] <mathrick> dolphy: basically, it would require discovery of hw capabilities earlier, right?
[18:27] <Company> i'd like to have an easy way to just get src ! spider ! videeosink spider0. ! audiosink
[18:27] harshy ([email protected]) left irc: "I quit for now"
[18:27] <Company> just for debugging
[18:27] <jowenn> Sorry if I ask many silly question, but i've never used gstreamer before.  Is it possible from within a filter plugin to find out, how long an incoming stream is and at which position from beginning/end the current datablock is (in seconds) ?
[18:28] <Company> jowenn: you could use queries
[18:28] <Company> jowenn: gst_pad_query
[18:31] <jowenn> Company: Ah, thanks
[18:31] 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!"
[18:31] hallibaby ([email protected]) joined #gstreamer.
[18:31] <jowenn> Is it too coslty to call do GST_QUERY_POSITION for each iteration or is that buffered "locally"
[18:32] <Company> the position is transferred inside the buffer
[18:33] <Company> as a timestamp
[18:58] walters ([email protected]) joined #gstreamer.
[18:58] harshy ([email protected]) joined #gstreamer.
[19:13] <Uraeus> jowenn: what kind of application are you doing?
[19:15] herzi ([email protected]) joined #gstreamer.
[19:23] <jowenn> Uraeus: For work: some kind of touchscreen jukebox terminal
[19:23] <Uraeus> jowenn: ah, your company is one of those who supply systems to bars etc.?
[19:24] <jowenn> Uraeus: For fun, I'm trying to figure out how difficult development developmet with gstreamer is compared to artsd
[19:24] <Uraeus> jowenn: hope you think it is easier so far :)
[19:24] <jowenn> Uraeus: They want to enter that market
[19:26] <Uraeus> jowenn: your system is planned to be linux based then since you are looking at arts and gstreamer?
[19:27] <jowenn> yes
[19:27] <jowenn> It doesn't seem too difficult yet, just that I can't find a lot of documentation, except the API reference
[19:28] <Uraeus> jowenn: yeah, well we hope to get a lot of the docs updated over the next months as we prepare for our 0.8.0 release
[19:28] <Uraeus> s/updated/updated and extended/
[19:31] Action: jowenn is away: bbl
[19:31] sub_pop ([email protected]) joined #gstreamer.
[19:49] Action: dolphy regrets that gst-editor is not able to handle complex pipeline like gst-play 's one
[19:49] <Uraeus> dolphy: file a bug report and maybe thaytan will look into it
[19:56] Action: ds is happy to have a new normal mouse on his mac
[19:58] <dolphy> ds: http://getcontrol.sourceforge.net/files/gstplay_pipeline.png
[19:58] <dolphy> ds: do you think that's the good pipeline for libgstplay ?
[19:59] <Uraeus> ds: btw, unless someone else hasn't told you yet, you commited the autoplugger code to the wrong CVS repo (SF)
[20:02] <ds> Uraeus: I added it to fd.o as well
[20:04] <ds> dolphy: seems reasonable
[20:10] Nick change: LeRoutier -> LeRaway
[20:12] Action: ds netinstalls his new laptop
[20:16] water ([email protected]) joined #gstreamer.
[20:18] apoc ([email protected]) joined #gstreamer.
[20:18] <apoc> yo
[20:18] <Uraeus> hi apoc
[20:19] <Uraeus> apoc: did you commit those new plugins of yours to SF CVS instead of freedesktop?
[20:20] <apoc> Uraeus : no ... I'm about to commit the trm one to freedesktop
[20:20] jaiserca ([email protected]) left irc: "Terminando cliente"
[20:21] water ([email protected]) left #gstreamer ("The water has parted.").
[20:22] <Uraeus> apoc: ah, ok :)
[20:29] walters ([email protected]) left irc: "out"
[20:36] Action: jowenn is back
[20:37] teuf ([email protected]) joined #gstreamer.
[20:37] <mathrick> yo teuf
[20:37] <Uraeus> hi teuf
[20:38] <teuf> hi
[20:44] <dolphy> ds: videotestsrc ! videobalance ! ffcolorspace ! videoscale ! ximagesink
[20:44] <dolphy> ds: try that one
[20:48] Shoragan ([email protected]) left irc: "Leaving"
[20:57] <dolphy> Company: pre-full wait: 250 of 0-250 buffers in queue
[20:57] <dolphy> Company: that means the queue is full ?
[20:57] ds ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[21:00] <Company> dolphy: yep
[21:04] hadley ([email protected]) joined #gstreamer.
[21:13] ds ([email protected]) joined #gstreamer.
[21:18] Nick change: LeRaway -> LeRoutier
[21:20] sxpert ([email protected]) left irc: "Client exiting"
[21:21] sxpert ([email protected]) joined #gstreamer.
[21:27] LeRoutier ([email protected]) left irc: Remote closed the connection
[21:27] sxpert ([email protected]) left irc: "Client exiting"
[21:28] sxpert ([email protected]) joined #gstreamer.
[21:28] Zeenix ([email protected]) joined #gstreamer.
[21:29] <dolphy> Company: hmm weird
[21:29] <Zeenix> hi
[21:29] <dolphy> Company: with the new pipeline the queue gets full
[21:29] <dolphy> Zeenix: hi
[21:29] <dolphy> Company: and the playback stops after few seconds
[21:30] <apoc> hi Zeenix
[21:31] <Uraeus> Zeenix: new version of libcaca out now :)
[21:32] <Company> dolphy: stopping most likely happens when the audio queue is empty while the video queue is full
[21:32] <dolphy> Company: it s not really stopping
[21:32] <dolphy> Company: it s doing empty iterations
[21:32] <dolphy> Company: that s more exact :)
[21:33] <Company> --gst-scheduler=basicomega ....
[21:33] <dolphy> ah then it segfaults :)
[21:35] <dolphy> found the problem
[21:35] <dolphy> :)
[21:35] Action: dolphy slaps himself with a big stick
[21:36] <dolphy> Company: i got one element sitting in the thread not connected to anything
[21:36] <Company> that shouldn't segfault anyway, but whaatever
[21:37] <Zeenix> Uraeus: yeah, just openned my mailbox
[21:38] <dolphy> Company: strange the media plays double speed
[21:41] <dolphy> Company: in the pipeline i draw the visualization queue is not connected by default
[21:41] <dolphy> Company: does that mean that those elements won't state change ?
[21:43] <Company> dolphy: they should
[21:43] <dolphy> Company: well when i add them in the pipeline and connect them to switch
[21:43] <dolphy> Company: without connecting the queue 
[21:44] <dolphy> Company:  i get the full video queue and empty iterations
[21:44] <dolphy> Company: so i guess they aren't 
[21:45] thomasvs ([email protected]) joined #gstreamer.
[21:45] <dolphy> thomasvs: hey dude :)
[21:45] <apoc> hi thomasvs
[21:46] <thomasvs> evening people
[21:46] Action: thomasvs is about to commit
[21:47] <Zeenix> Uraeus: checked out the new features: support for window resizing ..
[21:47] <dolphy> trm plugin :)
[21:47] <thomasvs> trm plugin ?
[21:48] <dolphy> musicbrainz
[21:49] <thomasvs> yes I know :)
[21:49] <thomasvs> but is someone writing one ?
[21:52] <thaytan> something already hit cvs
[21:52] <thaytan> what is the memory layout of I420 ?
[21:53] <thomasvs> ouch
[21:54] <apoc> thomasvs, thaytan: yes it just compute the trmid from raw audio data
[21:54] <thomasvs> cvs update -dP takes OVER FIVE MINUTES here due to all my local changes
[21:54] <thomasvs> apoc: great stuff
[21:54] <thomasvs> in c I hope, not c++ ?
[21:54] <apoc> yes in c
[21:55] <thomasvs> cool stuff, will try it out soon
[21:56] <thaytan> someone tell me the memory layouts for I420 and YV12, and you can all have DVD subtitle/menu support RSN 
[21:57] <Company> thaytan: www.fourcc.org
[21:57] <thaytan> Company: I can't spot I420 there
[21:57] <mathrick> apoc: you're supposed to use mb server to compute trmid, right? does it mean you overcame that?
[21:57] <Company> thaytan: it's further down - planar formats
[21:58] <thaytan> aha!
[21:58] <thaytan> thanks :)
[21:58] <apoc> mathrick: the musicbrainz library contact the default mb_server. I should add a property for this ...
[21:59] <mathrick> apoc: so, it still uses mb server?
[21:59] <Zeenix> Uraeus: sam is making the rpms himself this time
[21:59] Action: mathrick hoped it wouldn't be necessary
[21:59] <mathrick> apoc: anyway, cool :)
[21:59] <apoc> mathrick: yes
[22:00] <thaytan> catch you all later
[22:00] thaytan ([email protected]) left irc: "See y'all"
[22:05] <harshy> anyway I can speed up ripping in s-j on a 2.6 kernel
[22:08] <jowenn> Where from do I get gstreamer-libs-0.7.pc. I
[22:08] <jowenn> 've compiled a cvs version of the gstreamer module , but I only get gstreamer-0.7.pc  and gstreamer-control-0.7.pc
[22:08] <Company> you need gst-plugins
[22:09] <jowenn> Ok, thanks I hope it is no problem to have gentoo's 0.6 version and the 0.7 installed in parallel
[22:10] <dolphy> jowenn: they can coexist
[22:18] <dolphy> ok so audioconvert is broken
[22:19] <dolphy> that's what is making the video play double the speed
[22:19] <dolphy> and the sound is a big thunder
[22:19] <dolphy> thomasvs: http://getcontrol.sourceforge.net/files/gstplay_pipeline.png
[22:19] <dolphy> thomasvs: i ve put the gstplay pipeline in a png graph
[22:26] steve_b ([email protected]) joined #gstreamer.
[22:27] somex1 ([email protected]) joined #gstreamer.
[22:29] BBB ([email protected]) joined #gstreamer.
[22:29] <BBB> hey thomasvs, how was the concert?
[22:30] <thomasvs> BBB: very good, both of them
[22:30] <thomasvs> ok, so I'll be breaking the build for half an hour while comitting all my stuff
[22:30] <thomasvs> hang on :)
[22:30] <BBB> eeeeeeeek
[22:30] <BBB> well
[22:31] <BBB> nevermind
[22:31] <BBB> I'm breaking stuff too
[22:31] <BBB> (and failing horribly)
[22:31] <thomasvs> dolphy: hm - too much colorspace elements
[22:31] <thomasvs> but looks nice :)
[22:32] <dolphy> thomasvs: can't do with less
[22:33] Action: BBB hates autotools
[22:33] <BBB> but! I almost have ffmpeg fully ported
[22:34] <thomasvs> BBB: does ffmpeg want it autotooled ?
[22:34] <dolphy> that's the good question to ask now that he finished :)
[22:34] <mathrick> BBB: do they really write it all *by hand* ?
[22:34] <thomasvs> mathrick: yes
[22:35] Action: mathrick ph34rs
[22:35] <BBB> mathrick: yes
[22:35] <BBB> and god... poor person that did that
[22:35] <BBB> that's ten times worse than what I'm doing
[22:35] <BBB> my script is like 250-300 lines now... their script is 1200 lines of pure evil
[22:36] <Zeenix> anyone know of are 'REAL' free hosting service?
[22:36] <Zeenix> s/are/a
[22:36] <BBB> gstreamer.net/media?
[22:36] <BBB> ;)
[22:39] <BBB> thomasvs: oh, I didn't answer your question
[22:39] <BBB> they do, but "sort of"
[22:39] <BBB> most of them seem to like automake/autoconf by now
[22:39] <BBB> but they still hate libtool
[22:40] <BBB> so I need to bring this carefully ;)
[22:41] <apoc> BBB : are you *autotooling* ffmpeg or our cvs copy of ffmpeg ?
[22:42] <thomasvs> BBB: make sure you use ds's autotooling of ffmpeg somehow too
[22:42] <BBB> that's different
[22:42] <BBB> that's simply one Makefile.am
[22:42] <BBB> I'm using the same list, but I can't really use more than that
[22:42] <BBB> it's no configure script or whatever, and that takes most of the time
[22:43] <BBB> then again, it's almost done
[22:43] <BBB> and should be simple to integrate in gst now
[22:44] <BBB> and I must say that quilt is really a perfect tool for the job :)
[22:44] Action: BBB is happily using it now
[22:46] <thomasvs> ok, commited
[22:46] <thomasvs> BBB: yeah, it's very god to use for this sort of thing
[22:47] <BBB> for many things.. I currently use it for my profiles code etc., too
[22:47] <BBB> and my scene detection plugin
[22:47] <BBB> all of which are unfinished
[22:48] <BBB> did anyone see Jorn's new music player?
[22:48] <BBB> apparently, he doesn't hate GStreamer that much after all ;)
[22:48] <dolphy> thomasvs: you know audioconvert ?
[22:56] <BBB> en ximian is alleen voor diegenen die er een CD van hebben
[22:56] <BBB> oops
[22:56] <BBB> gst_element_error (smpte, CORE, NEGOTIATION, NULL, NULL);
[22:56] <BBB> that
[22:56] <BBB> ;)
[22:56] <BBB> thomasvs: why NULL. NULL?
[22:56] <BBB> isn't that sort of...e mpty?
[23:05] hadley ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[23:09] <Company> GST_STREAM_ERROR_TYPE_NOT_FOUND, GST_STREAM_ERROR_WRONG_TYPE, GST_STREAM_ERROR_DECODE, GST_STREAM_ERROR_ENCODE, GST_STREAM_ERROR_DEMUX, GST_STREAM_ERROR_MUX, GST_STREAM_ERROR_FORMAT
[23:09] <Company> isn't that the same error?
[23:10] <BBB> GST_STREAM_ERROR_TYPE_NOT_FOUND -> I think that has to do with typefinding
[23:10] <BBB> wrong-type is a stream error ("input file is no vorbis file")
[23:10] <BBB> I guess decode/encode/.. can be 'process' errors
[23:11] <BBB> so you're sort of half-right ;)
[23:11] <Company> well, iot all comes down to "unexpected data"
[23:11] <BBB> type-not-found is more like 'unsupported'
[23:12] <BBB> wrong-type is more like 'stupid user error' or 'autoplugging is borked'
[23:12] <jowenn> is there a way to tell gstreamer where to find plugins / filters ?
[23:12] <Company> hey, autoplugging just plugs based on caps :p
[23:12] <BBB> jowenn: --gst-plugin-path=..
[23:12] <BBB> (at gst-register)
[23:13] <Company> I especially don't like the decode/encode/demux/mux differentiation
[23:13] <Company> that _might_ be useful for a media player, but not for a streaming framework
[23:13] <dolphy> well if we can have that level of detail that's quite good
[23:14] <dolphy> i really dislike when xine simply says "Stream error" all the time
[23:14] <Company> no
[23:14] <Company> it impl?ies stuff which isn'T there
[23:14] <jowenn> no environment variable ?
[23:14] <Company> implies, too
[23:14] <BBB> if you want to know the type of plugin, simple do gst_element_get_factory(elment)->klass
[23:14] <Company> yeah
[23:15] <BBB> or what was it? gst_element_factory_get_klass(gst_element_class_get_factory(GST_ELEMENT_GET_CLASS(element)))
[23:15] <BBB> or so
[23:15] <Company> what's float2int? encoder or decoder?
[23:16] <BBB> converter
[23:16] <BBB> filter
[23:16] <BBB> like colorspace
[23:16] <Company> GST_STREAM_ERROR_FILTER ?
[23:17] <Company> GST_STREAM_ERROR_CONVERTER ?
[23:17] <BBB> GST_STREAM_ERROR_PROCESS
[23:17] <BBB> think of some cute term
[23:17] <BBB> :)
[23:17] <Company> GST_STREAM_ERROR_FORMAT
[23:17] <BBB> possibly
[23:20] <Company> thomasvs: gst-libs/gst/gst-i18n-plugin.h is missing
[23:20] <jowenn> do I have to do something additionally to copying my library in the plugin directory to make eg gst-editor find my plugin ?
[23:21] <Company> run gst-register
[23:21] <Company> gst-register-0.7
[23:27] <Uraeus> g'night
[23:27] Uraeus ([email protected]) left irc: "Client exiting"
[23:27] Action: Company shoots thomasvs
[23:32] <steve_b> hi guys, i'm back from .au
[23:32] <ds> ugh
[23:32] <ds> my new laptop has a broken ethernet tranciever
[23:33] <ds> i have to return it
[23:33] <ds> which means I can't play with it in the meantime
[23:33] <dolphy> bad luck :-/
[23:34] <steve_b> my gst preso got voted in the top 4, so i had to do it again :/
[23:35] <BBB> that's good, no?
[23:35] <apoc> steve_b: is it available online ?
[23:36] <BBB> apoc: are you interested in getting the wmav2 thing back to work?
[23:36] <BBB> I have a ug in bugzilla that it's broken
[23:37] <steve_b> apoc: it should be soon
[23:37] <steve_b> BBB: yeah it was good really. we might be getting some more .au developers soon
[23:37] <apoc> steve_b: cool :)
[23:38] <BBB> now *that* is cool ;)
[23:38] <BBB> does that mean we'll have our next seminar in .au?
[23:38] <BBB> :)
[23:38] Action: BBB always wanted to go to sydney one day
[23:38] <apoc> BBB: yep
[23:38] <steve_b> heh
[23:38] <Company> we need a .au sponsor for that i guess
[23:39] <steve_b> there is a good crowd of media types from sydney
[23:41] <Company> thomasvs: wavparse is missing, too
[23:41] harshy ([email protected]) left irc: "I quit for now"
[23:42] Action: BBB sleeps
[23:43] <BBB> and wavparse still needs fixing
[23:43] <BBB> bye
[23:43] Action: BBB waves
[23:43] BBB ([email protected]) left irc: "Client exiting"
[23:44] <Company> thomasvs: v4l2 needs fixing, too
[23:47] Action: apoc needs his bed too
[23:47] <apoc> night
[23:47] <dolphy> night
[23:47] apoc ([email protected]) left irc: "Leaving"
[23:47] <dolphy> who is taking care of audioconvert ?
[23:50] <Company> give me a launch line that fails and i'll look at it
[23:53] teuf ([email protected]) left irc: Remote closed the connection
[23:54] <dolphy> well ... tomorrow :)
[23:54] <dolphy> night
[23:54] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[23:55] dolphy ([email protected]) joined #gstreamer.
[23:55] <dolphy> filesrc location="/home/dolphy/matrix.avi" ! spider ! audioconvert ! volume ! osssink
[23:55] <dolphy> Company: here you are :)
[23:55] <dolphy> Company: first try.. success :)
[23:56] Action: dolphy really goes to bed now :)
[23:56] dolphy ([email protected]) left irc: Client Quit
[23:56] wheels ([email protected]) joined #gstreamer.
[23:57] <wheels> Hmm, some very interesting GStreamer related stuff in here -- http://dot.kde.org/1074351596/
[23:58] <wheels> (It shows that GStreamer is building up mindshare within the KDE crowd -- though not speicifically developers -- on GStreamer in KDE.)
[00:00] --- Mon Jan 19 2004
[00:01] <Company> what's amarok?
[00:01] <Company> I need to steal this sonograph for a visualizer :o
[00:01] <wheels> Company: that's the one that Markey is working on
[00:02] <wheels> ...which he added GStreamer support to last week.
[00:02] Action: wheels remembers that he still needs to kick Markey about that since he broke the module by putting in a busted Gst check...
[00:02] <mathrick> wheels: know how to display comments on dot nested instead of thread view?
[00:03] <wheels> mathrick: Just turn up the number of comments that are visible -- top right.
[00:03] <mathrick> ahh, ok
[00:03] Action: wheels always leaves it at 1000.  :-)
[00:06] <Company> http://dot.kde.org/1074351596/1074375249/1074392392/1074404776/1074414978/1074417703/1074445825/
[00:06] <Company> woohoo :D
[00:08] <wheels> Company: Yeah, despite the jabs at Gnome that's actually a pretty good comment (and more or less the idea that I'm trying to encourage)
[00:11] <Company> yeah, if those (f)lamers don't have anything to say against gst, we've come a long way :)
[00:12] <wheels> Well, especially if you compare to the aRts comments in there.
[00:13] <wheels> And it's essentially the right attitude and what I've been saying all along -- whether it's GStreamer or whatever the choice should be made on technical reasons...
[00:14] <wheels> (Well, and some surrounding project issues -- it should be maintained, documented, have multiple people working on it, be reasonably open, etc.)
[00:14] <Company> and we get developers, developers, developers ... :)
[00:14] LeRoutier ([email protected]) joined #gstreamer.
[00:14] <LeRoutier> hello
[00:15] <LeRoutier> gst-plugins HEAD does not compile
[00:15] <Company> i know
[00:15] <Company> thomasvs error commit broke it
[00:15] <LeRoutier> k
[00:15] <Company> and he hasn't turned up yet to fix it
[00:15] <wheels> Company: Guess you saw that I've finally got a CVS account now.  I'll start doing ad hoc reverts of  your commits any day now.  :-)
[00:16] <Company> wheels: great - noone ever dared doing that before
[00:16] <Company> actually once that was done - thomasvs backed out my error changes in september and has taken 4 months to get them in shape ;)
[00:17] <jowenn> is there a plugin template that really works with current cvs ?
[00:17] ChrisHJW ([email protected]) left irc: Connection timed out
[00:17] <jowenn> the one in gst-template doesn't work for me
[00:18] <Company> no, they only work "sort of" unfortunately
[00:18] <jowenn> doesn't even compile
[00:19] kmaraas ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[00:19] <wheels> But there are lots of really small plugins...
[00:20] kmaraas ([email protected]) joined #gstreamer.
[00:21] <jowenn> hi wheels
[00:21] <wheels> howdy  :-)
[00:21] <thomasvs> jowenn: ah, send us a bug report for that then
[00:22] <wheels> jowenn: I think you're like KDE developer number 4 or 5 this week to turn up in here...
[00:22] <Company> it needs to be adapted for the caps stuff
[00:22] <Company> (the template)
[00:22] <jowenn> which would you recommend, I tried adder.c but I seem to miss something important since it segfaults during gst-register
[00:23] <Company> the alaw/mulaw plugins are very simple
[00:23] <Company> (i should have tested them beforehand)
[00:23] Action: Company hides
[00:24] <thomasvs> Company: what do you see wrong in wavparse and v4l2 ?
[00:24] <Company> thomasvs: unconverted gst_element_error
[00:25] <thomasvs> Company: are you sure ? my commit took a very long time, maybe you just didn't have it yet
[00:25] <thomasvs> it also broke the syncmail scripts completely
[00:25] <Company> i'll update once more
[00:26] <thomasvs> I'm running a test build as well, but it's taking quite some time t complete
[00:26] <Company> nope, some v4l files and wavparse seem to be missing
[00:26] <Company> gst-i18n-plugins.h is there now
[00:26] <thomasvs> yeah, that I forgot to commit
[00:26] <thomasvs> but the wavparse is fine locally.  hm.
[00:27] Action: thomasvs tries a fresh checkout
[00:27] <Company> maybe it didn't get committed somehow?
[00:27] <thomasvs> what working revision do you have ? mine's 1.51
[00:28] TD[gone] ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[00:28] Action: Company runs cvs up -dPA
[00:28] <thomasvs> http://freedesktop.org/cgi-bin/viewcvs.cgi/gst-plugins/gst/wavparse/gstwavparse.c?cvsroot=gstreamer
[00:29] <thomasvs> seems to be fine on cvs
[00:29] <Company> thomasvs: 1.47 - i reverted that some day to check something and didn't update it...
[00:29] <thomasvs> ok - so it's a local problem ?
[00:29] <thomasvs> same for v4l2 then ?
[00:29] <Company> http://freedesktop.org/cgi-bin/viewcvs.cgi/gstreamer/gst-plugins/sys/v4l2/v4l2_calls.c?rev=1.13&content-type=text/vnd.viewcvs-markup
[00:29] <Company> nope
[00:29] <Company> that one is missing
[00:31] <thomasvs> hm, that's strange
[00:31] <thomasvs> I distinctly remember doing them, but they're gone from my local branch
[00:31] Action: thomasvs checks patch set
[00:33] <thomasvs> hm, yeah, the commit broke there.
[00:33] thaytoo ([email protected]) joined #gstreamer.
[00:33] <thomasvs> I put in the patches now
[00:34] <thomasvs> seeing something else missing ?
[00:34] <thaytoo> yo, thomasvs 
[00:35] <mathrick> yo thaytoo
[00:35] <thomasvs> evening jan
[00:36] <thomasvs> I'll send a mail with some explanation tomorrow about the error stuff
[00:37] <jowenn> Perhaps you could give me a hint, it's not much code yet: http://www.jowenn.at/tmp/gstcf.tgz. I think it must be a problem of type registering
[00:38] <LeRoutier> argh, if i find the guy that 'packaged' libcaca for gentoo, i'll give him a shot
[00:39] <jowenn> it eventually crashes in gst_pad_set_chain_function, but I get some GST_IS_ELEMENT failed before
[00:41] <Company> thaytoo: did your tetribot put up a good fight?
[00:44] <thomasvs> well, if no one has further complaints about compilation, I guess I can head off to bed now :)
[00:45] <LeRoutier> it still does not compile
[00:47] <Company> thomasvs: v4l2 still doesn't compile
[00:55] <thaytoo> Company: I gave it up after 8 hours, half finished...
[00:55] <thaytoo> I decided that I could be doing something useful and continued on Gstreamer DVD support instead
[00:55] <thaytoo> the winner was a _clear_ winner anyway - about 3 times better than the next best
[00:57] iain ([email protected]) joined #gstreamer.
[00:58] <mathrick> hi iain
[00:59] <iain> hi there
[01:09] harshy ([email protected]) joined #gstreamer.
[01:26] <LeRoutier> hum, mjpegtools is such a mess on gentoo. it forces to install glib 1.2/gtk 1.2
[01:28] hadley ([email protected]) joined #gstreamer.
[01:40] water ([email protected]) joined #gstreamer.
[01:46] Zeenix ([email protected]) left irc: Read error: 110 (Connection timed out)
[01:56] sublett ([email protected]) left irc: "I like food, food is good!"
[01:59] pb_ ([email protected]) left irc: "Client exiting"
[02:09] water ([email protected]) left irc: "brb"
[02:10] water ([email protected]) joined #gstreamer.
[02:14] alley_cat ([email protected]) left irc: "May the Source be with you!"
[02:19] <Company> iain: is it me or is the float audio format really braindead?
[02:19] <Company> iain: the gst one?
[02:21] <iain> I don't know anything about it
[02:21] <iain> whats wrong with it?
[02:21] <Company> the fact that you use one pad per channel
[02:22] <iain> I didnt know that you did
[02:22] <Company> it's 1) kinda unsupported to connect multiple pads from different elements, 2) slow because of cothread switches and 3) a  hell of a lot code
[02:22] <Company> float2int does at least
[02:22] <jowenn> should gst-launch filesrc location=/home/jowenn/mp3/Thomas\ D./Thomas\ D.\ -\ Liebesbrief.mp3 ! ffdec_mp3 ! osssink work ?
[02:23] <iain> oh, i didn't realise that was a factor of the format
[02:23] <iain> I just thought that was "the way it was done"
[02:23] <jowenn> I mean theat ffdec_mp3 link
[02:23] <iain> (but its what I need to be done in marlin anyway)
[02:23] <iain> (so I didn't mind really)
[02:24] <Company> i did vorbisdec
[02:24] <Company> and vorbis outputs float audio by default
[02:25] <Company> jowenn: gst-launch-0.7 but yes, that works
[02:25] <Company> i had to write loads of code until it halfway worked
[02:26] <Company> that got me so annoyed I threw it all away and invented audio-x-raw-fl?at-2
[02:26] <iain> Company: maybe it is braindead...
[02:26] <Company> and audio/x-raw-float-2, too
[02:26] water ([email protected]) left #gstreamer ("The water has parted.").
[02:26] <jowenn> hmm. Here it terminates after two seconds without error message and without audio ouput
[02:27] <iain> Company: welll, if you want to change it go ahead :)
[02:28] <Company> iain: just wanted to know if there's a reason for using 1 pad per channel
[02:28] <iain> I didnt invent it...
[02:28] <iain> you'd have to ask wingo?
[02:29] <Company> will do
[02:29] <Company> i'll go with x-raw-float-2 for the time being
[02:29] <Company> x-raw-float and multipad doesn't work in autopluggers, so vorbis would be stereo...
[02:30] <Company> jowenn: you can only use chainbased elements if you only have one sinkpad
[02:30] <jowenn> that means ?
[02:30] <Company> jowenn: you need to use a loopfunction
[02:31] <jowenn> and that's causing the crashin in _init ?
[02:32] <Company> i didn't run it yet :)
[02:33] <Company> and i probably won't for a while because my plugins are broken here...
[02:33] <iain> oh no...now inkscape is having a preference for "How many pixels wide is the page shadow border"
[02:34] Action: iain had such high hopes for inkscape
[02:34] <Company> jowenn: s/gst_gst_crossfader_get_type/gst_crossfader_get_type/ in gstcrossfader.h
[02:36] foser ([email protected]) left irc: "[ I want to believe ]"
[02:37] <Company> unfortunately plugins can't complain about unresolved symbols...
[02:39] <Company> iain: float22int is 178 lines, float2int is 525 lines ;)
[02:40] <iain> nice
[02:40] <Company> the burden of working with request pads and all that
[02:40] m_wheels ([email protected]) joined #gstreamer.
[02:40] <Company> i'm currently packing the channels one after another, so it's not interleaved
[02:41] <Company> |<channel1-----><channel2----->|
[02:41] <Company> like that
[02:41] wheels ([email protected]) left irc: Nick collision from services.
[02:41] Nick change: m_wheels -> wheels
[02:41] <iain> would aggregator ! float22int work?
[02:42] <Company> nope
[02:42] <iain> oh
[02:42] <Company> because after aggregator it would still be 2 buffers
[02:43] Action: iain guesses he can hook MarlinSampleSrc to interleave the buffers
[02:43] <Company> i dunno which format would be best
[02:43] <Company> i should probably mail the list as i have no clue about "pro" audio
[02:44] <iain> me neither :)
[02:47] <jowenn> Company: Any idea, what could cause: assertion `GST_IS_ELEMENT (element)' failed in gst_crossfader_init ?
[02:48] <Company> jowenn: the fact that you run it as the plugin_init function probably :)
[02:49] <Company> jowenn: you need to put plugin_init in that GST_PLUGIN_DEFINE
[02:51] <Company> if you wanna do yourself a favour run with -Wall -Werror :)
[02:51] <danb> if i'm using a (gst) thread to iterate a pipeline, and the pipeline hits eos, which calls a callback, what is the state of the thread in the lifetime of the callback?
[02:51] <danb> speicifically, what if the callback tries to kill the thread (by setting its state to GST_STATE_NULL or GST_STATE_READY)?
[02:52] <danb> *specifically
[02:52] <jowenn> Company: That was the problem :)
[02:53] <Company> danb: that should work
[02:53] <Company> danb: the preferred method is to connect an idle handler that does that though
[02:56] <danb> connect an idle handler?
[02:56] <danb> rather, an idle handler?
[02:57] <Company> danb: (assuming you're using gtk/glib in your app: g_idle_add (handler, data)
[02:57] <danb> glib, yes; gtk, no
[02:58] <danb> Company: is there a more in-depth description of this technique? or do i need to squeeze it all out of you here and now? ;)
[02:59] <Company> danb: i was assuming you read the glib documentation :)


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