IRC Logs

*******************************************************************
[03:00] <danb> glib, no; gstreamer, yes ;)
[03:00] <danb> do the glib docs hold these answers?
[03:00] <Company> yeah
[03:01] <danb> hrm
[03:01] <Company> the docs about main loops/context
[03:03] iain ([email protected]) left irc: "Tomorrow is just the same as the day before"
[03:04] <Company> gstreamer's examples/thread/thread.c does that, btw
[03:07] <Company> Linux 2.6 rules - gdb dlopening madness, compile and a grep through /usr/lib at once
[03:07] <Company> no sound hickup, everything is smooth
[03:07] <danb> Company: this one? http://gstreamer.net/docs/current/manual/html/ch22.html
[03:09] <danb> err, no, that's probably helloworld2.c
[03:09] Action: danb extracts source
[03:09] Action: Company shoots all sorts of people
[03:11] Action: danb looks at thread.c
[03:12] <danb> Company: i see no explicit use of idle handlers in thread.c. am i missing something?
[03:14] <Company> no you're not
[03:14] <Company> i should've looked better at that example...
[03:15] <Company> sorry
[03:15] <danb> np
[03:15] <danb> that's the example i'm following (helloworld2.c, actually, but it does the same), and getting problems i can't explain
[03:16] <danb> looks like deadlock or something, but it's a little complicated so i figured i'd learn a little more about what's really going on instead of blindly flipping all the switches available to me
[03:16] <Company> heh
[03:16] <danb> so i'll dig around for some useful glib docs and might be back in a while
[03:16] <Company> yeah, we're seriously lacking docs...
[03:18] <jowenn> cu
[03:18] jowenn ([email protected]) left irc: Remote closed the connection
[03:29] <danb> Company: can you point me to any examples of uses of g_idle_add?
[03:30] <Company> Rhythmbox :o
[03:31] <danb> hehe
[03:31] Action: danb fails to ingest (probably) tens of thousands of lines of C code
[03:35] <Company> it's only one file
[03:35] <Company> monkey-media/player-gst.c or something
[03:35] <Company> but it's still enough code to look at ;)
[03:36] <danb> jesus
[03:36] <danb> lines?
[03:36] <danb> wait
[03:36] <Company> that one file isn't particularly long
[03:36] <danb> oh, player-gst
[03:36] <danb> oh, within rb
[03:37] <danb> must be monkey-media/monkey-media-player-gst.c
[03:38] <Company> yeah, that one
[03:38] <danb> hrm... 857 lines
[03:38] <danb> i'll take a look
[03:38] <danb> :P
[04:09] <danb> wow... glib blows away all kinds of type-safety
[04:14] <dilinger> that's what happens when you try to do crazy OO stuff in C
[04:14] <dilinger> but i'll take that over crappy c++ template implementations any day :)
[04:15] <danb> and i'll take a good c++ impl over any of that ;)
[04:16] <danb> i'd rather increase c++ demand so that the crappy c++ impls get pushed out of use, honestly
[04:16] <Company> and i'll just continue using descriptive function names over left shifting my data onto the screen
[04:16] <mathrick> danb: *bzz* wrong. GObject *is* typesafe
[04:17] <danb> mathrick: explain?
[04:19] <mathrick> danb: all G_ cast macros are runtime checked for being or correct type
[04:20] <mathrick> s/or/of/
[04:20] <dilinger> well, run-time type safety isn't quite comparable to compile-time type safety
[04:21] <dilinger> it's a lot nicer to have the compiler catch typing issues than some a user who manages to do something in an order that hasn't been tested by developers
[04:22] <danb> yeah, compile checks are nice
[04:22] <danb> at least the types are checked at all
[04:22] <Company> only if you want to
[04:24] <mathrick> nite
[04:26] Nick change: mathrick -> mathrick|sleep
[04:38] Bast ([email protected]) joined #gstreamer.
[04:39] Bast ([email protected]) left irc: Client Quit
[04:44] harshy|lap ([email protected]) joined #gstreamer.
[04:45] harshy|lap ([email protected]) left irc: Client Quit
[04:57] <LeRoutier> good night
[04:57] LeRoutier ([email protected]) left irc: "Leaving"
[04:59] khalek ([email protected]) joined #gstreamer.
[05:47] Nick change: khalek -> kh_out
[05:52] <danb> Company: i've read up some on glib. can you now tell me more about this idle handler technique? i'm not sure that an idle handler is what i want...
[05:53] <danb> and why it's preferred to having the callback kill the thread that originated the signal
[05:58] Action: danb makes some popcorn
[06:10] Action: danb takes a gander at monkey-media-player-gst.c
[06:13] <danb> hrm
[06:13] <danb> maybe it is what i want
[06:21] Misirlou ([email protected]) left irc: "asdf"
[06:21] <jimmy_dean> So is using a LADSPA plugin from an application through GStreamer any different than a normal GStreamer plugin?
[06:27] thaytoo ([email protected]) left irc: "leaving"
[06:34] Action: danb tests...
[06:34] Action: danb fails
[06:34] <danb> fuck
[06:34] ChriHJW_log ([email protected]) left irc: Read error: 110 (Connection timed out)
[06:35] <steve_b> Company: still here?
[06:47] hallibaby_ ([email protected]) joined #gstreamer.
[06:48] harshy|lap ([email protected]) joined #gstreamer.
[06:49] harshy ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[06:49] <steve_b> jimmy_dean: using ladspa should be like other gstreamer functions
[06:49] <steve_b> jimmy_dean: why?
[06:49] <jimmy_dean> steve_b: k, cause I'm trying to use one in rhythmbox
[06:49] <jimmy_dean> steve_b: the equalizer plugin
[06:49] <steve_b> that won't work ;)
[06:50] <jimmy_dean> why not?
[06:50] Nick change: harshy|lap -> harshy
[06:51] <steve_b> all ladspa plugins use mono float audio at their format. you will need a pipeline which deinterleaves the audio, connects the 2 seperate streams to 2 ladspa plugins, then converts back to an interleaved int format
[06:52] <jimmy_dean> steve_b: hmm, so do you do that within gstreamer or then within rhythmbox?
[06:54] <steve_b>            -ladspafoo-
[06:54] <steve_b>           /           \
[06:54] <steve_b> int2float-             float2int-
[06:54] <steve_b>           \           /
[06:54] <steve_b>            -ladspafoo-
[06:54] <steve_b> or somesuch
[06:54] hallibaby ([email protected]) left irc: Read error: 60 (Operation timed out)
[06:54] <jimmy_dean> steve_b: two separate ladspa equalizer plugins will use up a heck of a lot of CPU is my guess...it's a full fourier transform-based equalizer
[06:55] <steve_b> i suggest you write a gstreamer plugin which does eq on interleaved int audio - that would be much appreciated
[06:56] <jimmy_dean> steve_b: well I was writing one till taaz told me about the ladspa one...and actually it uses the same FFTW library that I was using...the same DFT call
[06:56] <jimmy_dean> steve_b: so I can just use the rest of it's algorithm and then I'm close to being done...just have to connect some element properties for external apps to communicate with the plugin
[06:56] <steve_b> well try the ladspa approach if you want. you won't know if its fast enough until you try it.
[06:57] <jimmy_dean> steve_b: well doing all of the conversions sounds almost as hard
[06:57] <jimmy_dean> plus I don't know how to do that...this is my first experience programming with GStreamer
[06:58] <jimmy_dean> but who wants to use an equalizer that sucks up CPU %
[06:59] <steve_b> once you have something working it can be redesigned to not suck
[07:01] <jimmy_dean> steve_b: very true...that is my main goal to...to get something that at least is changing the magitudes of the different bands
[07:05] <jimmy_dean> steve_b: so how does one properly talk to a GStreamer plugin from an external application?
[07:05] <jimmy_dean> steve_b: are there any docs on doing this, or a good example I can look at?
[07:06] <steve_b> either through its object properties which can be clumsy, or if you can write an interface (like an eq interface!)
[07:06] <steve_b> we have a mixer interface which you could base it on
[07:06] <jimmy_dean> steve_b: ok, mind pointing me to any docs and/or source code?  I have Gstreamer CVS checked out
[07:07] <jimmy_dean> steve_b: an interface sounds generic like you could have multiple equalizer plugins and the higher level app not really caring which one it uses then...is that the idea?
[07:07] <steve_b> exactly
[07:07] <steve_b> gst-plugins/gst-libs/gst/mixer
[07:07] <steve_b> volume and osssink implement it
[07:08] <jimmy_dean> steve_b: ok, so then rhythmbox would have an example on how to access that interface since it uses the volume plugin I believe
[07:08] Nick change: kh_out -> khalek
[07:09] <steve_b> dunno yet. its kinda new. try libgstplay
[07:11] <jimmy_dean> steve_b: ok, thanks a lot
[07:24] thaytan ([email protected]) joined #gstreamer.
[07:26] <danb> is gst_element_set_state thread-safe?
[07:28] ChrisHJW ([email protected]) joined #gstreamer.
[07:29] <danb> ChrisHJW: do you know if gst_element_set_state is thread safe?
[07:34] <thaytan> danb: yes, it is supposed to be
[07:38] hadley ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[07:42] <danb> thaytan: good, thanks
[07:42] <danb> that gives me another reasonable option to explore
[07:44] Action: danb sleeps
[07:49] hallibaby_ ([email protected]) left irc: "Bye bye"
[07:54] Shoragan ([email protected]) joined #gstreamer.
[07:55] Uraeus ([email protected]) joined #gstreamer.
[07:55] <Uraeus> good morning everyone
[07:58] <thaytan> Hi Uraeus
[08:00] <Uraeus> thaytan: got the rest of the DVD stuff working now?
[08:00] ChrisHJW ([email protected]) left irc: Read error: 110 (Connection timed out)
[08:02] <thaytan> 'rest of' is a very broad term
[08:02] <Uraeus> well the stuff you where working on then :)
[08:02] <thaytan> and the answer is no ;)
[08:02] <thaytan> I hope to have a patch on the list tonight
[08:10] <Uraeus> ugh, seems like my advogato entry about Everaldo even made it onto some kde-mailing lists, not sure I planed for it to be that published
[08:12] <Uraeus> hmm, guess with planetgnome the days where I can use my blog for heartsighs is over :)
[08:13] <thaytan> yeah, it must be hard to blog when you know people are actually going to read it ;)
[08:20] ChrisHJW_log ([email protected]) joined #gstreamer.
[08:20] <Uraeus> well I expect people to read it, but not to react to it :) a blog is to me sort of a braindump, unlike a mail which is making a statement :)
[08:34] harshy ([email protected]) left irc: "I quit for now"
[08:38] <Uraeus> ping thomasvs
[08:44] kmaraas ([email protected]) left irc: "Leaving"
[08:44] pgoel ([email protected]) joined #gstreamer.
[09:15] rk ([email protected]) joined #gstreamer.
[09:16] <rk> good morning.
[09:16] <Uraeus> morning rk
[09:16] <Uraeus> rk: talked anymore with our french friends?
[09:17] <rk> Uraeus: No. I didn't take his coordinates and he didn't visit this channel for the last few days. :(
[09:20] <rk> Uraeus: I see you were involved in the Gnonlin project, don't you?
[09:22] <Uraeus> rk: only as webmaster
[09:23] <Uraeus> rk: he is away this week, but he should be back next week toghether with the 10 other french nonlin hackers
[09:25] <rk> Uraeus: Pity. I have few questions regarding gnonlin.
[09:33] <Uraeus> rk: well all I know is that they plan on start by updating gnonlin to latest API's etc, since it is a somewhat bittrotten atm
[09:34] <rk> Uraeus: I've done it on the weekend. 
[09:34] <rk> Uraeus: And that's why I need to consult with somebody who developed it.
[09:39] BBB ([email protected]) joined #gstreamer.
[09:44] hallibaby ([email protected]) joined #gstreamer.
[09:50] <BBB> mornin'
[09:51] <Uraeus> rk: [email protected] <- he is original developer
[09:51] <Uraeus> morning BBB
[09:51] <Uraeus> ok, gotta go to work
[09:51] Uraeus ([email protected]) left irc: "Client exiting"
[09:52] <BBB> and erik walthinsen is original developer too
[09:52] somex1 ([email protected]) left irc: Remote closed the connection
[09:52] <rk> BBB: Thanks.
[09:57] <BBB> (what do you need it for?)
[09:57] <BBB> (if I may ask? ;) )
[09:58] <rk> BBB: :) I adopted gnonling to the current gstreamer API and have some questions regarding gnonlin design and the code
[10:01] <BBB> ah :)
[10:01] <BBB> then you'd have to ask wim
[10:01] <BBB> erik didn't develop gnonlin, afaik
[10:01] <BBB> I know bits about the edsign, btw, so if you have simple questions you can ask me
[10:02] <rk> BBB: There is a 'design' file in the source tree that describes overall design and tree hierarchy
[10:02] <rk> And it's a bit out of dat
[10:02] <BBB> yes
[10:03] <rk> s/dat/date
[10:03] <BBB> although the general point of it should still apply
[10:03] <BBB> right?
[10:03] <rk> Yes.
[10:03] <BBB> you meant the part that describes what objects gnonline should have, what each should do (wrap one object, wrap one 'scene', timeline) etc
[10:04] <BBB> gnonline-element -> gnonlin-pipeline -> gnonlin-timeline, IIRC
[10:04] <rk> There is a non described GnlVlayer class. What was the point of creating it?
[10:05] <BBB> do you have a link to the SF webCVS?
[10:05] Action: BBB doesn't know that class
[10:05] <rk> Yes, I'm in the webcvs now
[10:06] <BBB> for me... ;)
[10:06] thomasvs ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:07] <rk> design is here:
[10:07] <rk> http://cvs.sourceforge.net/viewcvs.py/*checkout*/gnonlin/gnonlin/docs/random/design?content-type=text%2Fplain&rev=1.5
[10:07] <rk> all the sources are here:
[10:07] <rk> http://cvs.sourceforge.net/viewcvs.py/gnonlin/gnonlin/gnl/
[10:08] hallibaby ([email protected]) left #gstreamer ("Client exiting").
[10:09] <rk> According to design GnlGroup is derived from the GnlComposition but actually they're both derived from the GnlVlayer class.
[10:10] <rk> That in turn derived extends GnlLayer
[10:11] <BBB> it seems a layer is a collection of pipelines
[10:11] <BBB> so a vlayer is - probably - a collection of parallel-running pipelines
[10:11] <BBB> (whereas a hlayer would be a collection of sequentially-running pipelines)
[10:11] <BBB> (but halayer doesn't seem to exist)
[10:12] <BBB> I guess composition is a hlayer, sort of
[10:14] <BBB> hm, according to the design, composition is a vlayer...
[10:14] <BBB> erm
[10:14] <BBB> ...
[10:14] Action: BBB thinks he doesn't know ;)
[10:15] <rk> BBB: :) OK. Thank you.
[10:15] Action: rk continues reading the code
[10:23] Misirlou ([email protected]) joined #gstreamer.
[10:28] thomasvs ([email protected]) joined #gstreamer.
[10:29] <BBB> thomasvs: if I add AC_CANONICAL_TARGET/HOST, how do I retrieve the target platform?
[10:30] <thomasvs> BBB: I've never known either
[10:30] <thomasvs> BBB: but I never needed it either
[10:30] <BBB> hm...
[10:30] <thomasvs> BBB: can you check if v4l2 stuff compiles for you and let me know what doesn't ?
[10:30] <thomasvs> I don't have the headers
[10:30] <BBB> will do in a sec
[10:30] <BBB> am first fixing up ffmpeg a bit more
[10:30] Action: BBB updates CVS already
[10:32] Nick change: bluejay|sleep -> bluejay
[10:38] dolphy ([email protected]) joined #gstreamer.
[10:40] <BBB> damn, gedit broke
[10:40] <BBB> gstpad.c:2208: warning: null format string
[10:40] <BBB> ?
[10:42] <thomasvs> hm, strange
[10:42] <thomasvs> I wasn't getting that for regular c, but for cplusplus files
[10:43] <thomasvs> am looking into why now
[10:43] <thomasvs> what gcc version ?
[10:44] <BBB> 3.2
[10:46] <BBB> I think the issue is that you make it result in g_strdup_printf NULL instead of (NULL), no?
[10:48] <BBB> or we'll simply have to define all error messages already
[10:48] <BBB> I can't build right now
[10:48] <BBB> ..
[10:48] <dolphy> morning
[10:48] <dolphy> Company: you looked at audioconvert ?
[10:48] <thomasvs> BBB: yeah, but for c it doesn't have a problem
[10:49] <thomasvs> dolphy: what's wrong with it ? I fixed up some things in it last week
[10:50] <dolphy> thomasvs: filesrc ! spider ! audioconvert ! volume ! osssink doesn't work at my computer at home
[10:50] <dolphy> thomasvs: with matrix.avi file it makes the noise of a BMW M3 on a racing circuit
[10:53] <thomasvs> heh
[10:53] <thomasvs> what does it say with the debug from audioconvert ?
[10:55] <dolphy> haven't looked that deep
[10:55] <dolphy> i m reworking the gstplay pipeline 
[10:55] <dolphy> and i put audioconvert in it
[10:55] <dolphy> according to Company advice
[10:55] <dolphy> it make the video play double speed and break the sound
[11:01] <BBB> video follows the sound clock
[11:01] <BBB> so obviously, if sound breaks, so does video
[11:02] <dolphy> yeah
[11:06] Action: thomasvs really wonders about how many announcements coaster can make
[11:07] <BBB> does coaster exist for real now?
[11:07] Company ([email protected]) left irc: Remote closed the connection
[11:09] <BBB> thomasvs: (NULL) instead of NULL doesn't fix it for m :(
[11:09] <BBB> I suppose g_strdup_printf(NULL) is invalid
[11:17] <thomasvs> BBB: hm, that's strange
[11:17] <thomasvs> how come it works fine here ?
[11:17] <thomasvs> dolphy: can you try ?
[11:17] <BBB> which glibc?
[11:18] <BBB> I think G_GNUC_PRINTF(x,y) is a glibc define
[11:18] <BBB> (as in, glib forwards it to glibc)
[11:20] pb_ ([email protected]) joined #gstreamer.
[11:21] <thomasvs> 2.3.2
[11:21] <thomasvs> hm, that sucks.
[11:22] <thomasvs> BBB: g_strdup (NULL) works for sure according to the manual
[11:22] <thomasvs> g_strdup_printf (NULL), I'm not sure
[11:23] harshy ([email protected]) joined #gstreamer.
[11:24] <BBB> g_strdup() works, yes
[11:24] <BBB> but printf requires a first argument, I think
[11:24] <BBB> sounds normal to me...
[11:24] <BBB> since the first argument is a format identifier
[11:25] Action: BBB doesn't know ISO C that well
[11:25] <thomasvs> NULL is a format, the question is if it's valid
[11:26] <thomasvs> ok, so maybe I should try with "" instead
[11:26] harshy ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[11:26] <thaytan> printf NULL will kill most libc
[11:26] <thaytan> it'll segfault on solaris, for example
[11:27] <thomasvs> hm, ok then
[11:27] Action: thomasvs resorts to changing NULL to ("")
[11:29] Action: BBB feels sorry for thomasvs
[11:30] Nick change: bluejay -> bluejay|sleep
[11:30] <thomasvs> BBB: don't worry, I'm happy I got this in
[11:30] <thomasvs> BBB: just tying up loose ends.
[11:32] Action: BBB wonders where bitrate is stored in MPEG video packet headers
[11:55] Misirlou ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[11:58] <BBB> ugh, it isn't there
[11:58] Action: BBB boinks head
[12:09] mtearle ([email protected]) joined #gstreamer.
[12:16] wheels ([email protected]) left irc: "work"
[12:16] KoRnouille ([email protected]) joined #gstreamer.
[12:21] hallibaby ([email protected]) joined #gstreamer.
[12:21] hallibaby ([email protected]) left irc: Client Quit
[12:35] Company ([email protected]) joined #gstreamer.
[12:43] <thomasvs> BBB: can you update again and build again ?
[12:43] <Company> BBB: what's the ffmpeg list?
[12:44] Action: Company needs to flame them for exporting symbols like "mp3_init"
[12:50] <thomasvs> Company: hehe, yeah, good luck
[12:50] <thomasvs> Company: I already told them once, but they didn't really care much
[12:55] <BBB> working on it...
[12:55] <BBB> Company: I'm working on that slowly...
[12:55] <BBB> Company: I think that if we supply patches, they'll accept them (sort of)
[12:55] <BBB> we gotta push them a bit
[12:56] <BBB> I've hd the wierdest things that they didn't wanna fix, but they'd apply patches instantly ifn I provide them myself :)
[12:56] <BBB> thomasvs: updating CVS now
[12:57] <Company> problem is that current musicbrainz and ffmpeg clash
[12:57] <Company> so if you do a complete build of current gst-plugins you get a crash on gst-register
[12:58] <Company> the musicbrainz guys told me they've fixed it for the upcoming 2.0.3, so it won't be a problem this time...
[13:00] <Company> but i'm gonna up the version requirement of musicbrainz to > 2.0.2 anyway to get around people asking about segfaults in ffmpeg
[13:04] <BBB> yup
[13:04] <BBB> thomasvs: seems to work fine now
[13:40] Gorilla_ ([email protected]) joined #gstreamer.
[13:49] sublett ([email protected]) joined #gstreamer.
[13:49] <Gorilla_> is anyone around??
[13:49] <BBB> well... I see 38 people online
[13:49] <BBB> so I guess so
[13:50] <Company> i"m nopt around, im anedge
[13:51] <Gorilla_> :-P... bascially I am trying to find an application that will take a video stream and save it to an mpeg file.. gstreamer seems to fit okay but I must be misunderstanding something..
[13:51] Action: BBB nods @ mpeg
[13:53] <BBB> mpeg and asf have a slight issue... you need to know the bitrate... and we don't really do any of that...
[13:54] mark ([email protected]) joined #gstreamer.
[13:55] Nick change: mark -> mark_p
[13:59] Nick change: rk -> rk|away
[14:05] iain ([email protected]) joined #gstreamer.
[14:05] <BBB>  msgid "Could not determine type of stream"-msgstr ""+msgstr "Kon het type van de stroom niet herkennen"
[14:05] <BBB> 'stroom'?
[14:05] Company ([email protected]) left irc: Remote closed the connection
[14:05] <BBB> is dat niet datatype?
[14:06] <mark_p> trying to compile latest released gstreamer and am getting this error on the mpeg2 decoder plugin:  gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I../../gst-libs -D_LARGEFILE_SOURCE -D_FILE_OF FSET_BITS=64 -pthread -I/usr/include/gstreamer-0.7 -I/usr/include/glib-2.0 -I/usr/li b/glib-2.0/include -I/usr/include/libxml2 -DGST_DISABLE_DEPRECATED -Wall -I/usr/incl ude/mpeg2dec -O2 -g -pipe -march=i386 -mcpu=i686 -MT libgstmpeg2dec_la-gstmpeg2dec.l o -MD -MP -M
[14:06] <mark_p> F .deps/libgstmpeg2dec_la-gstmpeg2dec.Tpo -c gstmpeg2dec.c  -fPIC -DPIC -o .libs/libgstmpeg2dec_la-gstmpeg2dec.o
[14:06] <mark_p> gstmpeg2dec.c: In function `gst_mpeg2dec_chain':
[14:06] <mark_p> gstmpeg2dec.c:498: warning: implicit declaration of function `mpeg2_pts'
[14:06] <mark_p> gstmpeg2dec.c:622: error: `PIC_FLAG_PTS' undeclared (first use in this function)
[14:06] <mark_p> gstmpeg2dec.c:622: error: (Each undeclared identifier is reported only once
[14:06] <mark_p> gstmpeg2dec.c:622: error: for each function it appears in.)
[14:06] <mark_p> gstmpeg2dec.c:623: error: structure has no member named `pts'
[14:06] <mark_p> make[3]: *** [libgstmpeg2dec_la-gstmpeg2dec.lo] Error 1
[14:06] <mark_p> anyone know whats happening?
[14:06] <BBB> you've got mpeg2dec-.0.4.0
[14:06] <BBB> there's a patch in bugzilla, latest CVS has it applied too
[14:06] <BBB> it didn't make it into 0.7.3...
[14:07] <mark_p> BBB: thanks!
[14:07] <BBB> np
[14:19] <khalek> the link to the gst-launch documentation on the mainpage is missing the html dir in its url
[14:21] <thomasvs> khalek: ?
[14:23] <khalek> under the group of links titled "Extensive Development Tools" the url should be http://www.gstreamer.net/docs/current/manual/html/ch26s02.html not http://www.gstreamer.net/docs/current/manual/ch26s02.html
[14:23] <khalek> later being a 404
[14:26] <thomasvs> ah, ok
[14:26] Action: thomasvs needs to finish the site sometime very soon
[14:39] Nick change: rk|away -> rk
[14:45] ChrisHJW ([email protected]) joined #gstreamer.
[14:46] <mark_p> BBB: does gstreamer have support for subtitles in matroska files?
[14:48] <BBB> no
[14:48] <BBB> we read them
[14:49] <BBB> but we don't process them
[14:49] <BBB> (we don't have that muhc plugin stuff really)
[14:49] <BBB> er
[14:49] <BBB> subtitle plugin stuff
[14:49] <mark_p> does gstreamer support subtitles for dvds?
[14:50] <thomasvs> mark_p: thaytan is adding that this week
[14:50] <mark_p> thomasvs: cool!
[14:50] <mark_p> so it will be in 0.8
[14:52] Action: BBB has no clue what will be in 0.8
[14:52] <BBB> I hope we'll be able to get all playback formats back working before that
[14:52] <mark_p> Excellent. gstreamer looks like a really nice piece of software :)
[14:57] mark_p ([email protected]) left irc: "Leaving"
[14:59] <thaytan> hey, someone said my name
[15:00] <thomasvs> yeah
[15:00] <thomasvs> and someone wants you to do something
[15:03] mtearle ([email protected]) left irc: "Client exiting"
[15:12] dennis_ ([email protected]) got netsplit.
[15:12] Jara[zZ] ([email protected]) got netsplit.
[15:12] dennis_ ([email protected]) returned to #gstreamer.
[15:12] Jara[zZ] ([email protected]) returned to #gstreamer.
[15:12] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by calvino.freenode.net
[15:14] KoRnouille ([email protected]) got netsplit.
[15:14] herzi ([email protected]) got netsplit.
[15:14] mathrick|sleep ([email protected]) got netsplit.
[15:14] danb ([email protected]) got netsplit.
[15:14] KoRnouille ([email protected]) returned to #gstreamer.
[15:14] herzi ([email protected]) returned to #gstreamer.
[15:14] mathrick|sleep ([email protected]) returned to #gstreamer.
[15:14] danb ([email protected]) returned to #gstreamer.
[15:14] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by weber.freenode.net
[15:17] thomasvs ([email protected]) got netsplit.
[15:17] Shoragan ([email protected]) got netsplit.
[15:17] thomasvs ([email protected]) returned to #gstreamer.
[15:17] Shoragan ([email protected]) returned to #gstreamer.
[15:17] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by kornbluth.freenode.net
[15:18] sjoerd ([email protected]) got netsplit.
[15:18] ChrisHJW ([email protected]) got netsplit.
[15:18] pb_ ([email protected]) got netsplit.
[15:18] Gorilla_ ([email protected]) got netsplit.
[15:18] sxpert ([email protected]) got netsplit.
[15:18] desrt ([email protected]) got netsplit.
[15:18] bluejay|sleep ([email protected]) got netsplit.
[15:18] taaz ([email protected]) got netsplit.
[15:18] dilinger ([email protected]) got netsplit.
[15:18] ChrisHJW ([email protected]) returned to #gstreamer.
[15:18] Gorilla_ ([email protected]) returned to #gstreamer.
[15:18] pb_ ([email protected]) returned to #gstreamer.
[15:18] sxpert ([email protected]) returned to #gstreamer.
[15:18] bluejay|sleep ([email protected]) returned to #gstreamer.
[15:18] desrt ([email protected]) returned to #gstreamer.
[15:18] sjoerd ([email protected]) returned to #gstreamer.
[15:18] dilinger ([email protected]) returned to #gstreamer.
[15:18] taaz ([email protected]) returned to #gstreamer.
[15:18] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by brunner.freenode.net
[15:23] sjoerd ([email protected]) got netsplit.
[15:23] desrt ([email protected]) got netsplit.
[15:23] bluejay|sleep ([email protected]) got netsplit.
[15:23] pb_ ([email protected]) got netsplit.
[15:23] dilinger ([email protected]) got netsplit.
[15:23] sxpert ([email protected]) got netsplit.
[15:23] taaz ([email protected]) got netsplit.
[15:23] Gorilla_ ([email protected]) got netsplit.
[15:23] ChrisHJW ([email protected]) got netsplit.
[15:23] Shoragan ([email protected]) got netsplit.
[15:23] thomasvs ([email protected]) got netsplit.
[15:23] KoRnouille ([email protected]) got netsplit.
[15:23] mathrick|sleep ([email protected]) got netsplit.
[15:23] herzi ([email protected]) got netsplit.
[15:23] danb ([email protected]) got netsplit.
[15:23] Jara[zZ] ([email protected]) got netsplit.
[15:23] dennis_ ([email protected]) got netsplit.
[15:23] sublett ([email protected]) got netsplit.
[15:23] sxpert_work ([email protected]) got netsplit.
[15:23] BBB ([email protected]) got netsplit.
[15:24] taaz ([email protected]) returned to #gstreamer.
[15:24] dilinger ([email protected]) returned to #gstreamer.
[15:24] sjoerd ([email protected]) returned to #gstreamer.
[15:24] desrt ([email protected]) returned to #gstreamer.
[15:24] bluejay|sleep ([email protected]) returned to #gstreamer.
[15:24] sxpert ([email protected]) returned to #gstreamer.
[15:24] pb_ ([email protected]) returned to #gstreamer.
[15:24] Gorilla_ ([email protected]) returned to #gstreamer.
[15:24] ChrisHJW ([email protected]) returned to #gstreamer.
[15:24] Shoragan ([email protected]) returned to #gstreamer.
[15:24] thomasvs ([email protected]) returned to #gstreamer.
[15:24] sxpert_work ([email protected]) returned to #gstreamer.
[15:24] BBB ([email protected]) returned to #gstreamer.
[15:24] sublett ([email protected]) returned to #gstreamer.
[15:24] dennis_ ([email protected]) returned to #gstreamer.
[15:24] Jara[zZ] ([email protected]) returned to #gstreamer.
[15:24] KoRnouille ([email protected]) returned to #gstreamer.
[15:24] herzi ([email protected]) returned to #gstreamer.
[15:24] mathrick|sleep ([email protected]) returned to #gstreamer.
[15:24] danb ([email protected]) returned to #gstreamer.
[15:24] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by weber.freenode.net
[15:26] LeRoutier ([email protected]) joined #gstreamer.
[15:26] <LeRoutier> hello
[15:27] <LeRoutier> thomasvs, v4lX is still not compiling, even after your latest fix
[15:28] gheet ([email protected]) joined #gstreamer.
[15:28] <thomasvs> LeRoutier: yes, but it's not my fault
[15:28] <thomasvs> well
[15:29] <thomasvs> depends on how you look at it, but ... :)
[15:31] <LeRoutier> hum, from the topic, official site is now at fdo but latest news (OSNews article) is only on gstreamer.net
[15:31] <LeRoutier> or is fdo official "beta" site ?
[15:32] #gstreamer: mode change '+o thomasvs' by ChanServ!ChanServ@services.
[15:32] Topic changed on #gstreamer by [email protected]: New website under construction - http://www.freedesktop.org/~gstreamer/ | GStreamer article - http://www.osnews.com/story.php?news_id=5648
[15:32] #gstreamer: mode change '-o thomasvs' by [email protected]
[15:32] <BBB> I will, hold on
[15:33] <BBB> thomas, how do I give errors now?
[15:34] <BBB>     gst_element_error (scene, CORE, NEGO, NULL,
[15:34] <BBB>                        "We're not set up properly");
[15:34] <BBB> is that ok?
[15:35] <thomasvs> almost
[15:35] <thomasvs> NEGO -> NEGOTIATION
[15:35] <BBB> ah, ok
[15:35] <thomasvs> wrap the debug in () as well
[15:35] <BBB> oh, yeah, righ
[15:35] <thomasvs> and, I wouldn't give that as a debug message :)
[15:36] <thomasvs> if it's the error where nego failed before and thus some parameters aren't set ...
[15:36] <thomasvs> I'd make it a general one like
[15:36] <BBB> as if I care... it's an unfinished plugin
[15:36] <BBB> error handling is really the last thing I care about ;)
[15:36] <thomasvs>        ("format wasn't negotiated before chain function"));
[15:36] <thomasvs>                                                                                 
[15:36] <BBB> I first need it to work
[15:36] <BBB> ;)
[15:36] <thomasvs> heh :)
[15:36] <thomasvs> well, you asked :)
[15:36] <thomasvs> want me to explain why NULL, NULL is valid for message and debug ?
[15:37] <BBB> I suppose it uses the default message then?
[15:37] <BBB> i.e. the one for CORE_ERROR_NEGOTIATION?
[15:38] <dolphy> thomasvs: so how should we move on that HAVE_V4L2 define ?
[15:38] <BBB> I took a look at videodev.h for v4l2.  Can someone explain me why akernel header defines HAVE_V4L2 unconditionally, and how we can get thisfixed upstream ?
[15:38] <BBB> eh?
[15:38] <BBB> it doesn't
[15:38] <BBB> if you have v4l2, it's defined
[15:38] <BBB> else, it isn't
[15:38] <dolphy> BBB: but anyway that conflicts with config.h
[15:38] <BBB> no
[15:39] <BBB> beucase that one uses HAVE_VIDEO4LINUX2
[15:39] <dolphy> ../../config.h:320:1: "HAVE_V4L2" redefined
[15:39] <dolphy> In file included from gstv4lelement.h:40,
[15:39] <dolphy>                  from v4l_calls.h:23,
[15:39] <dolphy>                  from gstv4lelement.c:30:
[15:39] <dolphy> /usr/include/linux/videodev.h:7:1: this is the location of the previous definition
[15:39] <BBB> dxr3 fails to compile here, too
[15:39] <BBB> ???
[15:39] <BBB> I've never ever had that error before
[15:39] <BBB> and I've been messing around with v4l2 for a LONG time
[15:39] <dolphy> neither did i
[15:39] <dolphy> first time today :)
[15:40] <dolphy> #ifndef __LINUX_VIDEODEV_H
[15:40] <dolphy> #define __LINUX_VIDEODEV_H
[15:40] <dolphy>                                                                                 
[15:40] <dolphy> #include <linux/types.h>
[15:40] <dolphy> #include <linux/version.h>
[15:40] <dolphy>                                                                                 
[15:40] <dolphy> #define HAVE_V4L2 1
[15:40] <dolphy> #include <linux/videodev2.h>
[15:40] <dolphy>                                                                                 
[15:40] <dolphy> #ifdef __KERNEL__
[15:40] <dolphy>                                                                                 
[15:41] <dolphy> #include <linux/device.h>
[15:41] <dolphy> #include <linux/poll.h>
[15:41] <dolphy> #include <linux/mm.h>
[15:41] <dolphy> that's from /usr/include/linux/videodev.h
[15:41] <BBB> ehm, our symbol used to be called HAV_EVIDEO4LINUX
[15:41] <BBB> etc.
[15:41] <BBB> HAVE_V4L is wrong
[15:41] <thomasvs> BBB: a system header should neve ever have something like  #define HAVE_V4L2 1
[15:41] <thomasvs> so that's a bug anyway
[15:41] <BBB> eh?
[15:41] <BBB> sure it's not
[15:41] <BBB> it's defined to 1 on systems that have videodev2.h
[15:41] <BBB> it's part of the 2.4.x v4l2 patch
[15:41] <thomasvs> that's namespace pollution
[15:41] <BBB> by defualt, it's 0
[15:41] <dolphy> HAVE_V4L2="yes"
[15:41] <dolphy> that's in configure.ac
[15:42] <BBB> we should use VIDEO4LINUX{,2}
[15:42] <BBB> not V4L{,2}
[15:42] <dolphy> well we won't make the kernel guys change :)
[15:42] <dolphy> so let's change ours
[15:42] <thomasvs> BBB: videodev.h literally says #define HAVE_V4L2 1
[15:42] <BBB> thomasvs: then you have the v4l2 patchj applied :)
[15:42] <thomasvs> BBB: no other kernel headers define a HAVE_ like this
[15:43] <dolphy> thomasvs: yeah and we should have HAVE_VIDEO4LINUX2 = "yes"
[15:43] <thomasvs> BBB: I don't have anything at all, this is dolphy's header
[15:43] <thomasvs> WEEHEE
[15:43] Action: thomasvs has the first error translated from a gst-plugins plugin
[15:43] <dolphy> i have debian ppc header :)
[15:44] <dolphy> the strange thing is that my header hasn't change
[15:44] <dolphy> since at least several weeks
[15:44] <dolphy> so something changed somewhere
[15:46] <dolphy> LeRoutier: what error did you get ?
[15:47] <BBB> dolphy: I dunno, I've never seen this before
[15:47] <BBB> ywya, go ahead change it please
[15:47] Action: BBB is doing other stuff
[15:47] Company ([email protected]) joined #gstreamer.
[15:48] <thomasvs> the only thing is different is this
[15:48] <thomasvs> the header gets included twice
[15:48] <thomasvs> which I'm about to fix
[15:48] <thomasvs> but I'm still stumped as to why the header does a define on HAVE_V4L2
[15:48] ct_ ([email protected]) joined #gstreamer.
[15:49] <thomasvs> I'm assuming the mixer stuff having translation marked strings also need to be translated ?
[15:51] <BBB> yes
[15:51] <BBB> that was marked so it could be used later on
[15:51] <BBB> ;)
[15:52] <thomasvs> ok
[15:52] Action: thomasvs adds that too then
[15:54] Nick change: mathrick|sleep -> mathrick
[15:55] <BBB> woohoo! weehee!
[15:55] <BBB> I've just transcoded my first file using only ffmpeg elements in gstreamer
[15:55] <mathrick> mornin'
[15:55] <BBB> (who's gonna kill me first? :+)
[15:55] <mathrick> well, sort of mornin ;)
[15:55] Action: BBB is playing with ffmux_*/ffdemux_*
[15:56] <mathrick> BBB: you're evil ;)
[15:56] <BBB> you only noticed now?
[15:56] Action: mathrick reads backlog
[16:01] <LeRoutier>  
[16:01] <LeRoutier> isn't it wrong ?
[16:01] <LeRoutier> #ifdef HAVE_CONFIG_H
[16:01] <LeRoutier> #  include "config.h"
[16:01] <LeRoutier> #endif
[16:01] KoRnouille ([email protected]) left irc: "Leaving"
[16:01] <LeRoutier> i suppose this include defines HAVE_CONFIG_H
[16:01] <LeRoutier> so we should have #ifndef   no ?
[16:02] <BBB> no
[16:02] <BBB> every command line says "-DHAVE_CONFIG_H"
[16:02] <thomasvs> so uhm
[16:03] <thomasvs> BBB: Phone-in and Phone-out, what are they ?
[16:03] <thomasvs> (Ie, I think I know, but I want it confirmed or denied)
[16:04] <BBB> I have no clue
[16:04] <BBB> see the OSS specs
[16:05] <thomasvs> to me it seems they mean telephone in and headphone out
[16:05] <BBB> no
[16:05] <BBB> there's a separate headphone, no?
[16:05] <thomasvs> don't think so
[16:05] <BBB> #define SOUND_MIXER_PHONEIN     20      /* Phone input */
[16:05] <BBB> #define SOUND_MIXER_PHONEOUT    21      /* Phone output */
[16:06] <BBB> ... so helpful
[16:07] <thomasvs> yeah :)
[16:07] <thomasvs> ah well
[16:07] <thomasvs> guess people will file a translation bug anyway
[16:07] <thomasvs> -ENOCARE
[16:07] <BBB> make it short please
[16:07] <BBB> labels in the mixer shouldn't be too lon
[16:07] <BBB> so Tel.-in and Tel.-uit in dutch
[16:07] <BBB> or sop
[16:07] <BBB> -p
[16:08] <mathrick> BBB: reminds me "Details" button in windoze sys monitor. "Allocated virtual pages - shows allocated virtual pages" ;)
[16:09] <BBB> yes, same thing
[16:09] <BBB> annoying
[16:09] <BBB> ...
[16:09] <BBB> ;)
[16:09] <mathrick> brb
[16:09] Nick change: mathrick -> mathrick|afk
[16:15] <thomasvs> BBB: how can I test the ossmixer i18n ?
[16:15] ChriHJW ([email protected]) joined #gstreamer.
[16:17] <BBB> thomasvs: in gst-mixer
[16:19] <thomasvs> in gnome-media ?
[16:19] <thomasvs> it doesn't have custom gnome txl's ?
[16:23] <dolphy> the V4L thing does not seem to be a modification in configure.ac
[16:28] <LeRoutier> nope, the problem is that gst-i18n.h re-includes config.h which defines HAVE_V4L2
[16:29] <dolphy> i see
[16:30] teuf ([email protected]) joined #gstreamer.
[16:30] <teuf> hi
[16:30] <thomasvs> hm, it is annoying that config.h doesn't have include-once
[16:33] <BBB> thomasvs: no, gst-mixer uses the mixer-provided strings
[16:33] <BBB> public headers should never include config.h?
[16:33] Action: BBB kicks thomasvs
[16:34] Uraeus ([email protected]) joined #gstreamer.
[16:34] <thomasvs> BBB: it's not a public header
[16:34] ChrisHJW ([email protected]) left irc: Read error: 110 (Connection timed out)
[16:35] <BBB> of course it is
[16:35] <BBB> and even if it isn't
[16:35] <BBB> (it is: all plugins include it, no?)
[16:35] <Uraeus> yodi yo yo
[16:35] alley_cat ([email protected]) joined #gstreamer.
[16:35] <thomasvs> some plugins include it, and it's not a public header
[16:35] <thomasvs> it's private, it's not installed
[16:35] <Company> gst_private.h includes config.h, too i think
[16:35] <Uraeus> hi alley_cat
[16:36] <thomasvs> yeah, it does
[16:36] <alley_cat> hi
[16:36] <LeRoutier> i think all plug-ins already include config.h.  so no need to reinclude it in i18n specific header
[16:36] <Uraeus> alley_cat: did you manage to get anywhere with that NAS plugin I sent you?
[16:37] <alley_cat> Uraeus: didn't try yet, i think i'll have some time for it this evening
[16:37] <Uraeus> alley_cat: ok, I be around :)
[16:39] <LeRoutier> BBB: still having fun with ffmpeg shell script hell or is the nightmare over ? (about your blog)
[16:44] <BBB> still having nightmares
[16:44] <BBB> but it's almost done
[16:44] <BBB> :)
[16:47] <BBB> Company: shall I add that we're willing to provide patches?
[16:48] <Company> BBB: sure
[16:48] pippin ([email protected]) joined #gstreamer.
[16:48] <Company> BBB: s//ff_/ shouldn't be too hard ;)
[16:49] <BBB> I know
[16:49] <BBB> but these guys really do nothing except if you spoonfeed them patches
[16:51] Evaso ([email protected]) joined #gstreamer.
[16:51] <Evaso> hi everybody
[16:51] <Uraeus> ho Evaso
[16:52] <Evaso> gstreamer guys
[16:52] <Evaso> do u have read this manifesto: http://linuxaudio.org/en/press/linux_audio_press_release_1.txt ?
[16:53] <Uraeus> Evaso: I mailed them on sunday asking if we could join :)
[16:53] <Evaso> you have posted on mailing list?
[16:53] <Uraeus> Evaso: no I mailed the guy listed on top of the contact page
[16:54] <Uraeus> Evaso: Daniel James
[16:54] <Evaso> http://linuxaudio.org/en/press/linux_audio_press_release_1.txt
[16:54] <Evaso> ops sorry
[16:55] <Evaso> to me good initiative seems one
[16:55] <Uraeus> Evaso: sure, that is why we want to join :)
[16:55] <LeRoutier> is it something more than a press release for now ?
[16:56] <Evaso> leroutier
[16:56] <Evaso> take a look to the members:
[16:56] <Uraeus> LeRoutier: not really
[16:56] Action: thomasvs isn't too fond of having replaygain tags added to core
[16:57] <Uraeus> LeRoutier: my idea is that it doesn't hurt us to join so...
[16:57] <Evaso> angula alsa ardour audacity JACK rosegarden etc. tc.
[16:57] <LeRoutier> Evaso, having numerous/known/rich members is one thing. having specs or code is another
[16:58] <Company> thomasvs: is there a way to debug shell scripts? autogen on solaris gives me errors
[16:58] <LeRoutier> for now, it seems to have a goal and people that want it 
[16:58] <Evaso> yes but do u want to enter it when the things are alredy done?
[16:58] <Company> thomasvs: what's the problem with replaygain?
[16:59] <thomasvs> Company: set -x at the top might help
[17:00] <thomasvs> Company: well, tags in general I'm sort of ambivalent about them being in the core and not as a helper lib ...
[17:00] <thomasvs> ... but having replaygain tracks just means anyone can add whatever tag he wants
[17:00] <thomasvs> ... which I'm about to do with my custom tag system that I use in Dave/Dina (because, replaygain tags are conceptually wrong, but that's another discussion :))
[17:01] <thomasvs> What I mean is, a) I dislike having media-specific subsystems in gstreamer and b) the fact that the only way to add tags is to put them in the core
[17:01] steve_b ([email protected]) left irc: Read error: 110 (Connection timed out)
[17:02] <LeRoutier>  
[17:02] <Company> thomasvs: tags require the core for 2 reasons where a fix is now known but didn't make 0.8
[17:03] <Company> thomasvs: and tags are not media specific
[17:03] <LeRoutier> thomas, now that config.h is no more included, it's gst/avi  that does not compile :)
[17:03] <Company> thomasvs: well, the tags itself are media specific, but not the tagging subsystem
[17:03] <thomasvs> Company: media-specific in the sense that replaygain tracks have no relevance for anything beyond audio/video
[17:03] <thomasvs> Company: yeah, that's what I mean.
[17:03] <thomasvs> Company: I don't mind having a tagging system in core
[17:04] <thomasvs> Company: it's the fact that the only way to add media-specific tags is to do it in the core
[17:04] <Company> thomasvs: I consider the whole tagging system experimental anyway until we know where we go with it (whcih requires apps using it first)
[17:04] <thomasvs> Company: sure, I agree - I just like it being experimental outside of the actual core :)
[17:05] <thomasvs> and am willing to help on moving it somewhere, because I feel bad about wanting to add my quantaudio tags to the core
[17:05] <thomasvs> LeRoutier: you have stale source, update first
[17:05] <Company> thomasvs: you don't need to
[17:05] <Company> thomasvs: apps and plugins are free to add their own tags
[17:05] <teuf> btw, is there a generic id3 tagger ? lame tagging doesn't seem to support many tags
[17:05] <thomasvs> Company: ah - so why is the replaygain in the core then ?
[17:06] <Company> thomasvs: 1) because it's simpler to add support from other elements
[17:06] <LeRoutier> thomasvs, well, it's gstreamer that does not compile, not gst-plugins. and my CVS tree is up to date
[17:06] <Company> thomasvs: 2) because everything is added there at the moment
[17:07] <thomasvs> LeRoutier: you pasted a build failure in gst-plugins/gst/avi
[17:07] <thomasvs> Company: ok, right - but that just means that today I should add my tags there too ?
[17:07] <LeRoutier> hum, let me check if my brain is so damaged
[17:07] Evaso ([email protected]) left #gstreamer ("Client exiting").
[17:07] pb_ ([email protected]) left irc: "Client exiting"
[17:07] <thomasvs> which is why I would like to help change it so it makes more sense.
[17:08] <thomasvs> Company: because afaict the replay gain tags would only ever be useful for audio plugins anyway, so it wouldn't hurt to have them in a gst-plugins lib
[17:10] <Company> thomasvs: if you want to get the tags out of the core, go ahead and move them out - i just don't think it's worth it right now when we don't even know what tags people want
[17:11] <LeRoutier> thomasvs, got a brain damage, gstreamer compiles well, you're right
[17:11] <Company> thomasvs: and if you have tags that are likely supported by more than 1 plugin and useful in more than 1 app, put em in the core for now
[17:11] <BBB> I'd rather only see 'recommended' tags in core
[17:11] Action: BBB thinks thomas has a poitn
[17:15] <dolphy> is that normal that make in gstreamer core always recompile everything even if you touch a very small thing
[17:16] <LeRoutier> thomasvs, i just verified. my gst-plugins CVS tree is clean and up to date and gst/avi does not compile because i18n stuff isn't defined
[17:18] BOFH_ ([email protected]) joined #gstreamer.
[17:18] <BOFH_> Afternoon - anyone alive care to fix a small bug in the latest CVS GST-Plugins?
[17:19] <BOFH_> (or I can finish the bug report if that is easier)
[17:19] <Company> if the bug really is small, sure 
[17:20] <Company> hm
[17:20] <Company> that shell doesn't support test -e
[17:21] <BOFH_> make that more than one bug...
[17:21] <BOFH_> all "cannot build" bugs.
[17:21] <BBB> BOFH_: it's usually easier to use bugzilla
[17:21] <BBB> we're pretty active on bugzilla
[17:21] <BOFH_> for some reason, the makefile is not including the X11 libs for some plugins.
[17:21] <BOFH_> so far, the overlay plugin, and xwindowlistener
[17:21] <BBB> oh, I just fixed that for v4l
[17:21] <BBB> please submit a bugzilla report
[17:22] <BBB> I'll fix it tonight
[17:22] <BBB> ok?
[17:22] <BBB> it's fairly easy
[17:22] <BOFH_> aight..  lemme finish patching manually, and finish the build so I know if any more are gonna cause the same prob.
[17:22] <BOFH_> aye - I noticed..  I am not a coder by any stretch, and I managed to fix ;)
[17:25] <BBB> :)
[17:26] <Company> my uni is running glib 2.0.3
[17:26] <Company> o m g
[17:26] <BBB> so?
[17:26] <Company> gstreamer HEAD doesn't build
[17:30] <thomasvs> dolphy: depends on what you touch
[17:30] <thomasvs> [gst-cvs] [thomas@otto avi]$ cvs stat gstavi.c
[17:30] <thomasvs> ===================================================================
[17:30] <thomasvs> File: gstavi.c          Status: Up-to-date
[17:30] <thomasvs>  
[17:30] <thomasvs>    Working revision:    1.2
[17:30] <thomasvs> LeRoutier: what does it say for you ?
[17:30] BBB ([email protected]) left irc: "Client exiting"
[17:32] pb_ ([email protected]) joined #gstreamer.
[17:34] <LeRoutier> looking at it
[17:34] <LeRoutier> ===================================================================
[17:34] <LeRoutier> File: gstavi.c          Status: Up-to-date
[17:34] <LeRoutier>  
[17:34] <LeRoutier>    Working revision:    1.2
[17:34] <LeRoutier>    Repository revision: 1.2     /cvs/gstreamer/gst-plugins/gst/avi/gstavi.c,v
[17:34] <LeRoutier>    Sticky Tag:          HEAD (revision: 1.2)
[17:34] <LeRoutier>    Sticky Date:         (none)
[17:34] <LeRoutier>    Sticky Options:      (none)
[17:35] <thomasvs> ok, grep i18n gstavi.c ?
[17:35] <LeRoutier> so, we have the same version
[17:35] <LeRoutier> grep i18n gstavi.c
[17:35] <LeRoutier> #include "gst/gst-i18n-plugin.h"
[17:36] <thomasvs> you were telling me the source didn't include the header
[17:36] <thomasvs> grep shows it does
[17:36] <LeRoutier> no, i just said that i18n stuff was not defined
[17:36] <LeRoutier> LC_ALL and setlocale
[17:37] <thomasvs> hm
[17:37] <thomasvs> why am I not getting that ?
[17:37] <thomasvs> is anyone else getting the same thing ?
[17:37] <thomasvs> LeRoutier: what system are you on ?
[17:37] <LeRoutier> gentoo kernel 2.4 + bttv/v4l2 official patches
[17:37] Nick change: mathrick|afk -> mathrick
[17:37] <LeRoutier> gnome 2.4 + gstreamer/gst-plugins/gst-player CVS HEAD
[17:39] sub_pop ([email protected]) left irc: "Client exiting"
[17:39] Action: thomasvs wonders on where locale.h would best be included to fix it
[17:40] <thomasvs> I just plump for the easy way out I guess
[17:42] <BOFH_> Looks like there is only one plugin / lib that fails to build - the only other one I had was V4L - and that has apparently been fixed since I initially checked ouy CVS.
[17:43] rk ([email protected]) left irc: "bounging home"
[17:43] <BOFH_> wait... scratch that..
[17:43] <BOFH_> gst-plugins I don;t have CVS of.
[17:51] Rotty ([email protected]) joined #gstreamer.
[17:52] <LeRoutier> thomas, be happy, the error went away
[17:52] <LeRoutier>  
[17:52] <LeRoutier> and now, we have the next error, not i18n related but related to error functions which changed lately
[17:54] <jimmy_dean> Can anyone suggest a way of using gst_launch-0.7 to play an mp3 file that doesn't click and has normal timing?  Right now I have this but you can barely tell it's a song: gst-launch-0.7 filesrc location=my.mp3 ! mad ! audioconvert ! equalizer ! osssink
[17:54] <LeRoutier> v4l2_calls.c:55:40: macro "gst_element_error" requires 5 arguments, but only 2 given
[17:54] <jimmy_dean> equalizer is my plugin but it's acting as a passthrough right now
[17:56] <pippin> how is the plans of gstreamer wrt higher bitdepths for images(frames within video),. is 8 bit/channel considered good enough?, or has the system been designed to allow an easy transition to 16bit, gdouble as possible sample models?
[17:56] <thomasvs> pippin: yeah, it could be anything
[17:56] <thomasvs> pippin: just a matter of making sure the plugins understand the format
[17:57] <thomasvs> jimmy_dean: how about dropping audioconvert and equalizer ?
[17:57] <jimmy_dean> thomasvs: can't drop equalizer...that's my plugin I'm making and testing
[17:57] <jimmy_dean> thomasvs: Company gave me that line a few nights ago so I've just been using it to run my plugin
[17:58] <jimmy_dean> thomasvs: hmm, without audioconvert it's not music at all, only clicks
[17:58] <pippin> thomasvs: are such formats defined?,. or would it mean a rewrite of the used plug-ins if you needed higher bitdepths?
[17:59] Nick change: LeRoutier -> LeRaway
[18:01] <jimmy_dean> thomasvs: oh, seems to be if I take audioconvert out then it works, but it clicks a lot if I g_print() out a lot of stuff...is this expected?
[18:02] <teuf> jimmy_dean: yep
[18:02] <jimmy_dean> teuf: so printing out to the screen is really bogging it down in other words
[18:03] <BOFH_> uncool..   in libgstmedia-info.a - undefined reference to gst_caps_get_structure and gst_caps_free etc.
[18:03] <BOFH_> (compiling nautilus-media)
[18:04] <jimmy_dean> Does anyone here know anything about equalizers or has designed one in the past?
[18:04] <jimmy_dean> in software obviously :)
[18:06] <BOFH_> no anoncvs access to gstreamer?
[18:06] <Uraeus> BOFH_: there is, freedesktop offers it
[18:06] <BOFH_> the links there point to SF.net
[18:06] <Uraeus> BOFH_: http://www.freedesktop.org/Main/GettingInvolved
[18:07] <BOFH_> t/y
[18:07] <Uraeus> BOFH_: ugh, ok I update the pages there / thomas we need to move on getting the website off SF :)
[18:10] <thomasvs> jimmy_dean: of course - g_printing heavily impacts realtime processes
[18:11] <thomasvs> jimmy_dean: since it is blocking on terminal throughput
[18:11] harshy ([email protected]) joined #gstreamer.
[18:11] <thomasvs> pippin: it would probably not need a rewrite, more an adaptation
[18:11] <jimmy_dean> thomasvs: ok, well it makes sense...just was wondering if it was that sensitive
[18:11] <thomasvs> pippin: it depends on the plugins in question though
[18:11] Nick change: harshy -> harshyWORK
[18:12] <thomasvs> LeRaway: ok, next set of patches is in for v4l2
[18:12] <jimmy_dean> thomasvs: I've got a GUI written for Rhythmbox and it's all tied in with signals, etc...all I have to do now is take the algorithm from the LADSPA equalizer and adapt it to put in my equalizer plugin for gstreamer
[18:13] <pippin> thomasvs: well there doesn't seem to be a lot of image processing plugins yet anyway,. I'm just trying to think ahead about the possiblity of interaction between gstreamer and gegl
[18:13] <BOFH_> Uraeus, prolly wanna use /cvs/gstreamer for login and co commands - just to ppl don;t get overzealous and justr grab gstreamer - and end up with the whole site etc.
[18:13] <BOFH_> why am I telling you this? silly me)
[18:13] <thomasvs> pippin: there are some video effects if  that's what you mean though
[18:13] <thomasvs> pippin: and I'm sure lots of people want more added :)
[18:14] <thomasvs> pippin: the base class for videofilters is there
[18:15] <BOFH_> there - got gstreamer and gst-plugins from CVS...
[18:15] <BOFH_> now I can see if the same bugs that were in the release are still there before I submit a bugreport.
[18:15] <BOFH_> (prolly not as BBB fixed the v4l bug)
[18:15] <Uraeus> BOFH_: yeah, I am stealing the SF CVS page and updating it to work for freedesktop :)
[18:16] <BOFH_> heh - that works.
[18:17] <pippin> thomasvs: what I'm trying to figure out is in which ways it would make the most sense to integrate GEGL ( http://www.gegl.org/ ) and gstreamer,. I see them as very likely playmates in the future
[18:20] dolphy_ ([email protected]) joined #gstreamer.
[18:21] dolphy ([email protected]) left irc: Read error: 113 (No route to host)
[18:21] <thomasvs> pippin: to me it looks like gstreamer would have a gegl plugin, so that all video plugins could conceivably be pushed down to the gegl level
[18:21] <thomasvs> pippin: bear in mind though that video effects and image effects can have very different requirements
[18:23] <jimmy_dean> Is there a way for me to obtain the sampling rate for a given input in my plugin's chain function?
[18:23] <pippin> thomasvs: I don't think for instance the v4l, ffmpeg etc kind of plug-ins belong at the gegl level,. 
[18:24] <pippin> thomasvs: even though it's nice to have them there,. after some thinking,. one might end up with both a gstreamer wrapper in gegl, and a gegl wrapper in gstreamer,. the latter would probably be the most useful,. but sometimes just using gstreamer plug-ins for the input could make sense
[18:26] <BOFH_> wow that is so odd...   I installed gstreamer...  then tried to install gnome-media, and it complained about not having gst.h - that happened twice.
[18:27] <BOFH_> then I installed GST Plugins, and all of the sudden, gst.h was there.
[18:30] <Uraeus> BOFH_: ok, CVS info on the freedesktop website is fixed i the /developer section. Hope you like it :)
[18:30] teuf ([email protected]) left irc: "Client exiting"
[18:33] <BOFH_> Uraeus, prolly wanna include the "login" command there...
[18:33] <Uraeus> BOFH_: just removed it, don't think it is needed
[18:33] <BOFH_> hmm..  indeed for anyone who knows CVS, it's not...
[18:34] <jimmy_dean> hmm, teuf told me that perhaps the sampling rate can be obtained through caps but he wasn't sure how that works...can someone help me with that?
[18:34] <BOFH_> they should be able to figure out that part themselves...
[18:34] <BOFH_> (jut going on what most other packages do - they always include the login command)
[18:34] <thomasvs> BOFH_: no, "login" isn't needed
[18:34] <Uraeus> BOFH_: what I mean is that I tried the command listed for anon-cvs and it works even if you don't do the login first
[18:34] <thomasvs> BOFH_: you can check out iwthout doing login first
[18:34] <thomasvs> pippin: with video filters I specifically mean transformations, not sources or sinks
[18:35] <BOFH_> ahh..  ok then.  
[18:35] <BOFH_> that's a new one on me.
[18:35] <Uraeus> BOFH_: you have spent to much time on SF :)
[18:35] <thomasvs> I actually never ever saw a project where it was necessary
[18:35] <thomasvs> I think the only thing it does is make it set your currently used default CVSROOT
[18:35] Action: BOFH_ commits that to brain::usefulTrivia
[18:37] <BOFH_> wow..   gstplay sample takes forever to get through doing the autogen.sh script.
[18:41] <dolphy_> ??
[18:41] mathrick ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[18:41] Nick change: dolphy_ -> dolphy
[18:41] <Uraeus> BOFH_: gst-plugins you mean?
[18:41] <BOFH_> yes
[18:42] mathrick ([email protected]) joined #gstreamer.
[18:44] <BOFH_> well, here goes building gst-plugins again - hopefully the bugs have been fixed ;
[18:44] <BOFH_> ;)
[18:44] <Uraeus> BOFH_: btw, you shouldn't use the root account, especially when IRC'ing :)
[18:45] <BOFH_> I know...
[18:45] <BOFH_> my user account can;t do much ATM without alot of hacking around.
[18:45] <thomasvs> BOFH_: make another user account
[18:47] <BOFH_> meh.  unless my being on the irc nets as root risks someone else's box, to hell with em if they wanna hack my system ;)  Nothin here worth gettin anyhow.
[18:47] <BOFH_> LOL
[18:47] <pippin> thomasvs: I'll probably start thinking about wrapping gst for my own project as well,. perhaps just a input element wrapper,.. some of the code is quite boring to write ;P http://schweden.mine.nu/bauxite/screenshots/illpip.png
[18:48] <thomasvs> pippin: looks interesting, but I don't understand what the app does ?
[18:48] <thomasvs> what does it do ?
[18:49] <Company> muhaha, the old "don't use root" discussion :o
[18:49] <pippin> allows you to construct graphics processing pipelines,. intended as a testing ground for gegl
[18:51] <Company> thomasvs: 
[18:51] <Company> "gstpad.c", line 2208: syntax error before or at: 0
[18:51] <Company> "gstpad.c", line 2213: cannot recover from previous errors
[18:51] <Company> cc: acomp failed for gstpad.c
[18:51] <Company> I think using NULL in gst_element_error is not doable
[18:51] <Company> as g_strdup_printf NULL; isn'T valid C
[18:52] <Company> (that was with forte btw)
[18:52] <thomasvs> Company: right - I was thinking of either moving to (NULL) or ("")
[18:52] <thomasvs> Company: but I would need to test what I have now
[18:53] <thomasvs> Company: does it work if you change the calls tehre to (NULL) with forte ?
[18:54] sub_pop ([email protected]) joined #gstreamer.
[18:55] <Company> thomasvs: it compiles...
[18:55] <alley_cat> Uraeus: got the nassink integrated into the build system, now just need to port it to current gstreamer :)
[18:56] <Uraeus> alley_cat: cool :)
[18:56] <Uraeus> alley_cat: looking forward to seeing how the configure test ended up looking, cause I could never get it to find the .h file it was supposed to :)
[18:57] <alley_cat> yeah, that took me some time to figure out
[18:57] <alley_cat> but it's actually quite easy
[18:58] <thomasvs> Company: ok, will work on getting NULL -> (NULL) then
[18:58] <Company> thomasvs: but it segfaults
[18:58] <thomasvs> Company: ok, will work on getting NULL -> ("") then
[19:00] <Company> and we need the warning added to convfigure that barks for p_strdup_printf NULL;
[19:01] Rotty ([email protected]) left irc: Remote closed the connection
[19:02] <thomasvs> you lost me on that last one
[19:02] <thomasvs> Company: hm, is there a way that (NULL) would not segfault through something we'd do in the gst_element_error_printf ?
[19:03] <thomasvs> ie, what specifically is making it segfault ?
[19:03] <thomasvs> the va list stuff ?
[19:03] <Company> probably the fact that format strings in libc may not be NULL
[19:03] hadley ([email protected]) joined #gstreamer.
[19:03] kmaraas ([email protected]) joined #gstreamer.
[19:03] <thomasvs> no, because the _printf wrapper checks it, so it doesn't call a printf if it's NULL
[19:04] <Company> g_strdup_printf message;
[19:05] <Company> that sounds a lot like it calls a printf function before checking ;)
[19:06] <thomasvs> you sure you're not looking at the oldish version ?
[19:06] <Company> and that one doesn't bark _only_ because NULL is defined as  ((void) 0) instead of (void) 0
[19:07] <thomasvs>   gst_element_error_printf message, \
[19:07] <thomasvs>   gst_element_error_printf debug, \
[19:09] <BOFH_> there - bigreport filed.
[19:11] <Company> ah
[19:11] <Company> that was a recent addition :)
[19:12] <Company> could you make that function __gst_element_error_printf
[19:12] <Company> since it's not supposed to be used outside of gst_element_error
[19:12] <thomasvs> yeah, I could
[19:12] Action: Uraeus breaks dolphy's skies
[19:13] <thomasvs> Company: I wanted to figure out too if there are compilers that don't define NULL as ((void) 0), because I quite liked that accident
[19:13] Nick change: LeRaway -> LeRoutier
[19:13] <LeRoutier> thomasvs, you forgot one more : v4l2-overlay_calls.c:64:50: macro "gst_element_error" requires 5 arguments, but only 3 given
[19:16] <dolphy> Uraeus: i have snowboard :)
[19:18] <Company> thomasvs: we should try to write portable code, not depend on "accidents"
[19:18] gheet ([email protected]) left #gstreamer.
[19:20] <thomasvs> Company: yeah, I know
[19:21] <thomasvs> Company: I'm just not sure how other compilers do it
[19:21] <Uraeus> dolphy: you should coordinate your weekends better; what about a combined ski/wedding weekend? weddings at ski resorts as so popular now :)
[19:21] <thomasvs> LeRoutier: check again, I updated now
[19:22] somex1 ([email protected]) joined #gstreamer.
[19:22] Rotty ([email protected]) joined #gstreamer.
[19:23] <jimmy_dean> mathrick: ping?
[19:23] <mathrick> jimmy_dean: pong
[19:24] <jimmy_dean> mathrick: hey, are you aware of a way to get the sampling rate for the sound data coming into a plugin's chain function?
[19:24] <jimmy_dean> mathrick: I really need to either calculate this or obtain it from something
[19:24] <mathrick> jimmy_dean: you can get it from caps, but I can't tell you how to do it unfortunately :\
[19:25] <jimmy_dean> mathrick: do you know where I can find this out, i.e. documentation or someone else who might know?
[19:25] <mathrick> jimmy_dean: ask Company
[19:25] <jimmy_dean> mathrick: ok
[19:25] <jimmy_dean> Company: are you aware of a way to get the sampling rate for the sound data coming into a plugin's chain function?
[19:25] <Company> jimmy_dean: you defined this with your caps
[19:26] <Company> jimmy_dean: and unless you haven't changed it, it's hardcoded to 44100
[19:26] <jimmy_dean> Company: ahh right
[19:26] <jimmy_dean> Company: ok, I didn't make that connection...is GStreamer actively re-sampling the data then to that rate even if it's recorded at a different rate?
[19:27] <Company> jimmy_dean: either that or it refuses to connect
[19:27] <jimmy_dean> Company: ok, cool...is this a true statement then?
[19:28] <jimmy_dean> Company: num_samples = GST_BUFFER_SIZE(buf)/sizeof(double);
[19:28] <Company> jimmy_dean: you can also accept different caps in your plugin, that requires getting to know about linkfunctions and getcaps functions, the caps system and negotiation ;)
[19:28] <jimmy_dean> Company: yeah, I'll do that after I get my plugin working for the hardcoded case which is proving to be quite a challenge but hopefully worth it in the end
[19:28] <Company> jimmy_dean: i don't think it is
[19:29] <jimmy_dean> Company: steve_b suggested I make an equalizer interface too so more than one backend plugin could provide equalization
[19:29] <Company> jimmy_dean: you use 16bit
[19:29] <Company> double isn't 16bit ;)
[19:29] <jimmy_dean> Company: oh...int is?
[19:30] <Company> yeah, an equalizer interface is another thing that should be done - but later
[19:30] <LeRoutier> thomas : one more v4l2 compile error. i paste in PV
[19:30] <jimmy_dean> Company: so should I put sizeof(int) then?
[19:31] <Company> jimmy_dean: 16bit is always 2 bytes, so you can use a literal 2 there
[19:31] <Company> jimmy_dean: or sizeof(gint16) if you wish
[19:32] <jimmy_dean> Company: oh I gotcha...ok...do you know of anyone who might want to help me tackle this in the long-run?  It's taking up a ton of time and might go faster with someone who's knowledgable in GStreamer
[19:33] <jimmy_dean> Company: not that I mind but I'd like to see this get done in a reasonble amount of time and done with high quality integration into GStreamer
[19:33] <Company> jimmy_dean: dunno
[19:33] <mathrick> <Company> jimmy_dean: 16bit is always 2 bytes, so you can use a literal 2 there <-- by making this statement, you explicitly drop support for Honeywell machines ;)
[19:34] <jimmy_dean> mathrick: Honeywell machines...that's familiar but I can't quite place where I've heard that term from
[19:34] <Company> jimmy_dean: considering noone did it yet, the chances are pretty slim anyone will do it for you...
[19:34] <LeRoutier> Honeywell, sounds like an old sixties mainframe for me
[19:35] <mathrick> jimmy_dean: these were large mainframes, back from PDP days, they had 9 bits byte, and 36 bits word :)
[19:35] <jimmy_dean> Company: well I'm not saying do it for me, more like help me do parts that only an experienced GStreamer coder would know how to do well
[19:35] <jimmy_dean> mathrick: ahh, right!
[19:35] <jimmy_dean> mathrick: just read that in my OS Design book
[19:36] <alley_cat> yay!
[19:36] Action: alley_cat just listens to a mp3 stream output through nassink :)
[19:36] <jimmy_dean> Company: like I have the knowledge as well as anyone to do the equalizer algorithm, but doing an interface and having the plugin deal with int and float data inputs is something I don't know about
[19:37] <Uraeus> alley_cat: cool :)
[19:37] <jimmy_dean> alley_cat: is that nas for network audio server?
[19:38] <iain> jimmy_dean: just make it deal with either int or float
[19:38] <alley_cat> jimmy_dean: yes
[19:38] <iain> (I would suggest int, unless its easier to do float)
[19:38] <alley_cat> Uraeus: still some warnings i need to get rid of so it compiles with -Wall, but otherwise works so far :)
[19:38] <jimmy_dean> alley_cat: sweet, so that would allow a gstreamer app to send it's sound to a remote X session?
[19:38] <jimmy_dean> iain: well i'm doing int right now to start with
[19:39] <jimmy_dean> iain: converting to double for the fourier transform
[19:39] <iain> jimmy_dean: thats all you need to do
[19:39] <iain> jimmy_dean: if someone needs to pass float data in then they can use float2int ! equalizer
[19:39] <jimmy_dean> iain: do i need to cast back to int to output the data from my chain function?
[19:39] <iain> if you're outputting int then yes, probably
[19:40] <jimmy_dean> iain: well I'd like to avoid conversions like that if possible and also to be well-rounded as to support high-end audio people
[19:40] <Company> if you need doubles anyway, just use doubles and let a converter plugin do the rest
[19:40] <alley_cat> jimmy_dean: output to remote systems should work with esdsink already, with nassink it should work, but i can't try, only have one box
[19:41] <jimmy_dean> Company: ok, so you're saying it's fine for me to output as a double?
[19:41] <Company> no plugin should ever do conversion itself
[19:41] <jimmy_dean> ok
[19:41] <Company> unless of course it's a conversion plugin
[19:41] <jimmy_dean> Company: makes sense
[19:41] <Company> (i dunno if we have a double conversion plugin though)
[19:42] <mathrick> hmm, will float2int ! int2float output the same data as was inputted? or is it lossy conversion?
[19:43] <Company> you can't convert 32bit => 16bit => 32bit in a lossless way
[19:43] <jimmy_dean> Company: I'd like to separate this plugin later to be a time domain to frequency domain convertor and then just a equalizer plugin
[19:43] <Company> you'd just store it in 16bits in that case ;)
[19:44] <mathrick> Company: that what i thought :)
[19:44] <mathrick> so, it'd be better to take float as the default format for eq, right?
[19:45] <Company> mathrick: yeah, especially because there's quite some float audio decoders nowadays (like vorbis)
[19:45] walters ([email protected]) joined #gstreamer.
[19:45] <mathrick> yo walters
[19:45] <Uraeus> yo walters
[19:45] <walters> hey hey
[19:46] <Uraeus> walters: is the removal of the slider from rb a design decision or a gst bug decision?
[19:47] <walters> Uraeus: it happens because gst doesn't give any length.
[19:48] BOFH_ ([email protected]) left irc: "Leaving"
[19:51] Nick change: Jara[zZ] -> Jaramir
[19:53] <thomasvs> jimmy_dean: check the audio helper lib
[19:53] <thomasvs> gst-libs/gst/audio
[19:53] <thomasvs> jimmy_dean: if it doesn't work, let me know and I'll fix it
[19:54] <jimmy_dean> thomasvs: ok...there are docs in there describing things like caps, etc?
[19:54] <thomasvs> jimmy_dean: it's supposed to provide helper functions to get common stuff you want
[19:55] <thomasvs> jimmy_dean: if it misses stuff, we can add it there if you let me know what you need
[19:55] <jimmy_dean> thomasvs: ahh, I see...ok
[19:55] <jimmy_dean> thomasvs: that'd be a huge help, I'll let you know
[20:01] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[20:11] <alley_cat> is there some coding guide for gstreamer?
[20:11] <alley_cat> this nas plugin has very strange indentation, i'd like to clean it up a bit
[20:11] <Uraeus> alley_cat: I think we basically follow the gnome code guidelines for indentation etc.
[20:16] thomasvs ([email protected]) left irc: Read error: 60 (Operation timed out)
[20:20] Misirlou ([email protected]) joined #gstreamer.
[20:20] walters ([email protected]) left irc: "brb"
[20:28] walters ([email protected]) joined #gstreamer.
[20:29] ElRoutier ([email protected]) joined #gstreamer.
[20:29] <ElRoutier> re
[20:29] Gorilla_ ([email protected]) left #gstreamer ("Leaving").
[20:30] steve_b ([email protected]) joined #gstreamer.
[20:30] <Uraeus> hi steve_b
[20:34] <steve_b> hi
[20:34] <Uraeus> steve_b: congrats on a great talk at the conference :)
[20:37] <ds> mmmm
[20:37] <ds> I have to get a CD from work
[20:37] <ds> should I drive or bike?
[20:38] <Uraeus> ds: jogg
[20:40] <steve_b> Uraeus: thanks
[20:40] hadley ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[20:41] <steve_b> off to work now :(
[20:41] <ds> Uraeus: that would be best, but I just took a shower
[20:42] <Uraeus> ds: well bike then :)
[20:42] <ds> (apparently, to preclude myself from running)
[20:42] Action: Jaramir is away: Collapsed Backbone
[20:42] <alley_cat> what could be the reason for gst-launch hanging when i put a audioscale element before the nassink?
[20:43] <alley_cat> and what's the use of the fixate function?
[20:47] LeRoutier ([email protected]) left irc: Read error: 110 (Connection timed out)
[20:47] Nick change: ElRoutier -> LeRoutier
[20:47] Nick change: LeRoutier -> LeRaway
[20:51] apoc ([email protected]) joined #gstreamer.
[20:51] <apoc> yo
[20:52] <Uraeus> hey apoc
[20:52] <jimmy_dean> How should I keep track of internal plugin data that I can access in my plugin functions?
[20:53] mark_p ([email protected]) joined #gstreamer.
[20:56] <jimmy_dean> Specifically I mean I'd like to initialize my equalizer when the plugin inits and store some data from calculations and then access it later in my plugin's chain function
[21:03] <mathrick> jimmy_dean: use your element for that
[21:03] <jimmy_dean> mathrick: you're going to have to be more specific...I'm new to all of this
[21:04] dolphy ([email protected]) joined #gstreamer.
[21:04] Nick change: LeRaway -> LeRoutier
[21:04] <apoc> Salut dolphy :)
[21:04] <jimmy_dean> mathrick: do you mean like this...this is from the adder plugin:
[21:04] <jimmy_dean> gstelement_class->request_new_pad = gst_adder_request_new_pad;
[21:04] <jimmy_dean>   gstelement_class->change_state    = gst_adder_change_state;
[21:05] Action: apoc wonders why wma is broken now ...
[21:05] <mathrick> jimmy_dean: it's normal GObject thing, rather than gst specific. Just do typedef struct { GstElem *parent; int my_field; } GstEq;, and then from chain func you can get it by:
[21:06] <mathrick> jimmy_dean: GstEq *this = GST_EQ(GST_PAD_PARENT(pad));
[21:06] <mathrick> jimmy_dean: now, whenever you want to access my_field, just do this->my_field
[21:08] somex1 ([email protected]) got netsplit.
[21:08] sjoerd ([email protected]) got netsplit.
[21:08] desrt ([email protected]) got netsplit.
[21:08] bluejay|sleep ([email protected]) got netsplit.
[21:08] dilinger ([email protected]) got netsplit.
[21:08] sxpert ([email protected]) got netsplit.
[21:08] taaz ([email protected]) got netsplit.
[21:08] walters ([email protected]) got netsplit.
[21:08] alley_cat ([email protected]) got netsplit.
[21:08] Shoragan ([email protected]) got netsplit.
[21:08] Jaramir ([email protected]) got netsplit.
[21:08] dennis_ ([email protected]) got netsplit.
[21:08] sublett ([email protected]) got netsplit.
[21:08] sxpert_work ([email protected]) got netsplit.
[21:08] harshyWORK ([email protected]) got netsplit.
[21:08] herzi ([email protected]) got netsplit.
[21:08] danb ([email protected]) got netsplit.
[21:08] lilo ([email protected]) got netsplit.
[21:08] dolphy ([email protected]) got netsplit.
[21:08] Rotty ([email protected]) got netsplit.
[21:08] jimmy_dean ([email protected]) got netsplit.
[21:09] dolphy ([email protected]) returned to #gstreamer.
[21:09] walters ([email protected]) returned to #gstreamer.
[21:09] Rotty ([email protected]) returned to #gstreamer.
[21:09] somex1 ([email protected]) returned to #gstreamer.
[21:09] harshyWORK ([email protected]) returned to #gstreamer.
[21:09] alley_cat ([email protected]) returned to #gstreamer.
[21:09] danb ([email protected]) returned to #gstreamer.
[21:09] herzi ([email protected]) returned to #gstreamer.
[21:09] Jaramir ([email protected]) returned to #gstreamer.
[21:09] dennis_ ([email protected]) returned to #gstreamer.
[21:09] sublett ([email protected]) returned to #gstreamer.
[21:09] sxpert_work ([email protected]) returned to #gstreamer.
[21:09] Shoragan ([email protected]) returned to #gstreamer.
[21:09] sxpert ([email protected]) returned to #gstreamer.
[21:09] bluejay|sleep ([email protected]) returned to #gstreamer.
[21:09] desrt ([email protected]) returned to #gstreamer.
[21:09] sjoerd ([email protected]) returned to #gstreamer.
[21:09] dilinger ([email protected]) returned to #gstreamer.
[21:09] taaz ([email protected]) returned to #gstreamer.
[21:09] jimmy_dean ([email protected]) returned to #gstreamer.
[21:09] lilo ([email protected]) returned to #gstreamer.
[21:09] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by tenn.freenode.net
[21:10] kmaraas ([email protected]) left irc: "Leaving"
[21:10] somex1 ([email protected]) got netsplit.
[21:10] sjoerd ([email protected]) got netsplit.
[21:10] desrt ([email protected]) got netsplit.
[21:10] bluejay|sleep ([email protected]) got netsplit.
[21:10] taaz ([email protected]) got netsplit.
[21:10] dilinger ([email protected]) got netsplit.
[21:10] sxpert ([email protected]) got netsplit.
[21:10] Shoragan ([email protected]) got netsplit.
[21:10] walters ([email protected]) got netsplit.
[21:10] alley_cat ([email protected]) got netsplit.
[21:10] dennis_ ([email protected]) got netsplit.
[21:10] Jaramir ([email protected]) got netsplit.
[21:10] sublett ([email protected]) got netsplit.
[21:10] sxpert_work ([email protected]) got netsplit.
[21:10] herzi ([email protected]) got netsplit.
[21:10] danb ([email protected]) got netsplit.
[21:10] harshyWORK ([email protected]) got netsplit.
[21:10] lilo ([email protected]) got netsplit.
[21:10] dolphy ([email protected]) got netsplit.
[21:10] jimmy_dean ([email protected]) got netsplit.
[21:10] Rotty ([email protected]) got netsplit.
[21:10] walters ([email protected]) returned to #gstreamer.
[21:10] somex1 ([email protected]) returned to #gstreamer.
[21:10] harshyWORK ([email protected]) returned to #gstreamer.
[21:10] alley_cat ([email protected]) returned to #gstreamer.
[21:10] taaz ([email protected]) returned to #gstreamer.
[21:10] dilinger ([email protected]) returned to #gstreamer.
[21:10] sjoerd ([email protected]) returned to #gstreamer.
[21:10] desrt ([email protected]) returned to #gstreamer.
[21:10] bluejay|sleep ([email protected]) returned to #gstreamer.
[21:10] sxpert ([email protected]) returned to #gstreamer.
[21:10] Shoragan ([email protected]) returned to #gstreamer.
[21:10] sxpert_work ([email protected]) returned to #gstreamer.
[21:10] sublett ([email protected]) returned to #gstreamer.
[21:10] dennis_ ([email protected]) returned to #gstreamer.
[21:10] Jaramir ([email protected]) returned to #gstreamer.
[21:10] herzi ([email protected]) returned to #gstreamer.
[21:10] danb ([email protected]) returned to #gstreamer.
[21:10] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by weber.freenode.net
[21:10] dolphy ([email protected]) returned to #gstreamer.
[21:10] Rotty ([email protected]) returned to #gstreamer.
[21:10] jimmy_dean ([email protected]) returned to #gstreamer.
[21:10] lilo ([email protected]) returned to #gstreamer.
[21:10] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by tenn.freenode.net
[21:11] <mathrick> wb, filthy splitters ;)
[21:11] <jimmy_dean> there were only 4 of us left here :)
[21:14] <mathrick> jimmy_dean: did you get my answer?
[21:14] <jimmy_dean> mathrick: nope
[21:14] <mathrick> <mathrick> jimmy_dean: it's normal GObject thing, rather than gst specific. Just do typedef struct { GstElem *parent; int my_field; } GstEq;, and then from chain func you can get it by:
[21:14] <mathrick> * apoc wonders why wma is broken now ...
[21:14] <mathrick> <mathrick> jimmy_dean: GstEq *this = GST_EQ(GST_PAD_PARENT(pad));
[21:14] <mathrick> <mathrick> jimmy_dean: now, whenever you want to access my_field, just do this->my_field
[21:15] lilo ([email protected]) got netsplit.
[21:15] lilo ([email protected]) returned to #gstreamer.
[21:15] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by kurland.freenode.net
[21:15] lilo ([email protected]) got netsplit.
[21:15] lilo ([email protected]) returned to #gstreamer.
[21:15] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by kurland.freenode.net
[21:15] <mathrick> hmm, apoc wasn't essential for my answer, but whatever ;)
[21:15] <jimmy_dean> mathrick: ok, the existing functions seem to pass around a GstEqualizer structure...can I do it that way too?
[21:16] <mathrick> jimmy_dean: that's even better
[21:16] <jimmy_dean> mathrick: ok, I'll be consistent then :)
[21:16] <mathrick> jimmy_dean: basically, that GstEqualizer is your 'this', except that it is explicit
[21:16] <jimmy_dean> mathrick: thanks...and I see that I define the new members in my header file
[21:17] kmaraas ([email protected]) joined #gstreamer.
[21:17] <jimmy_dean> mathrick: so there is only one instance of my plugin created then during a lifetime of playing one mp3 file for instance, right?  And my element will survive throughout that entire lifetime?
[21:19] <mathrick> jimmy_dean: not exactly, your elements are created and destroyed when needed, which often isn't equal to playback of one file.
[21:19] lilo ([email protected]) got netsplit.
[21:19] lilo ([email protected]) returned to #gstreamer.
[21:19] #gstreamer: mode change '+b *!*omg@*.client2.attbi.com' by kurland.freenode.net
[21:19] <jimmy_dean> mathrick: so how/when can I assume that my data that I store is accurate?
[21:20] <mathrick> jimmy_dean: you should listen for EOS event, I think 
[21:20] <jimmy_dean> mathrick: got an example for me?
[21:20] <mathrick> jimmy_dean: but better ask Company / or ds-work
[21:20] <mathrick> jimmy_dean: i don't think so, I'm not that familiar with codebase
[21:21] <jimmy_dean> mathrick: ok, thanks
[21:22] <jimmy_dean> mathrick: this is all somewhat confusing but I'm learning
[21:23] <mathrick> jimmy_dean: I'm still very new to gst too :) I just have a little more knowledge of GObject
[21:23] <jimmy_dean> mathrick: ok, that's cool
[21:27] <alley_cat> anyone got a hint why this pipeline: "gst-launch-0.7 filesrc location=Kamelot-Lunar_Sanctum.mp3 ! mad ! audioscale ! nassink" doesn't play, the same without audioscale does?
[21:27] <mathrick> alley_cat: did you try with -v ?
[21:28] <mathrick> alley_cat: and --gst-debug-level=5 ?
[21:29] <mathrick> jimmy_dean: btw, I already used that Panther theme to get nicer optionmenus in Glossy ;)
[21:30] <jimmy_dean> mathrick: excellent...glad you like it :)
[21:30] <mathrick> jimmy_dean: now, I only have to rip sliders out of Panther, and my aqua theme is almost done :)
[21:30] <jimmy_dean> mathrick: hehe, you should try and make it like the real aqua sliders...they have a triangle shaped slider
[21:31] <mathrick> jimmy_dean: hmm? do you have any screenie of that?
[21:31] <alley_cat> ok, if i interpret the output of gst-launch -v correctly the nassink doesn't like changing caps
[21:31] <jimmy_dean> mathrick: just search google images...I'm sure you can find a picture there
[21:32] <mathrick> jimmy_dean: that was to be expected, "aqua slider" yield aquaparks pictures mostly ;)
[21:32] <jimmy_dean> mathrick: lol, try "screenshot of mac os x"
[21:33] <jimmy_dean> screenshots
[21:33] <jimmy_dean> err
[21:33] <mathrick> jimmy_dean: http://images.google.com/imgres?imgurl=developer.apple.com/techpubs/macosx/Cocoa/TasksAndConcepts/ProgrammingTopics/Slider/Tasks/art/SliderNoTicksHor.gif&imgrefurl=http://developer.apple.com/techpubs/macosx/Cocoa/TasksAndConcepts/ProgrammingTopics/Slider/Tasks/UsingSliderTickMarks.html&h=28&w=86&prev=/images%3Fq%3Dmacosx%2Bslider%26svnum%3D10%26hl%3Dpl%26lr%3D%26ie%3DUTF-8%26inlang%3Dpl%26sa%3DG
[21:33] <mathrick> ughh, that's long one ;)
[21:33] <alley_cat> too bad my esd is broken, otherwise i could try the same with esdsink ...
[21:34] <mathrick> jimmy_dean: it doesn't look triangle
[21:34] <jimmy_dean> mathrick: yeah, like that
[21:34] <jimmy_dean> there is one that is triangle
[21:34] <jimmy_dean> it's basically a triangle that points to the right
[21:35] Action: ds-work writes two CDs, in case one is broken
[21:35] <jimmy_dean> just like that round ball shape you see in the pic
[21:35] <mathrick> hmm, others found are triangle
[21:46] Nick change: bluejay|sleep -> bluejay
[21:49] <mathrick> hum, it looks like gtk sliders have no tickmarks, or I'm unable to find them
[21:49] <jimmy_dean> Anyone who's semi-familiar with LADSPA...how can the LADSPA equalizer plugin just assume that there are 1024 samples of incoming data?
[21:49] <mathrick> s/them/property for then/
[21:50] Uraeus ([email protected]) left irc: Remote closed the connection
[21:51] mark_p ([email protected]) left irc: Read error: 110 (Connection timed out)
[21:53] <jimmy_dean> the LADSPA equalizer plugin sets a constant called FFT_LENGTH to be 1024 and that's the number of data elements that it sends into the discrete fourier transform function...how can it always assume that it has 1024 data elements?
[21:55] <mathrick> jimmy_dean: are you talking about LADSPA plugin itself or gst plugin using LADSPA one?
[21:55] <jimmy_dean> mathrick: the LADSPA plugin itself...I'm creating a gst plugin based off of it's algorithemn
[21:55] <jimmy_dean> algorithm
[21:56] <mathrick> jimmy_dean: maybe it caches remaining ones somewhere, or iterates over samples in 1024 packets?
[21:56] <jimmy_dean> mathrick: because I do this right now which is the same thing as FFT_LENGTH except I calculate it: num_samples = GST_BUFFER_SIZE(buf)/sizeof(gint16);
[21:57] thaytan ([email protected]) left irc: "See y'all"
[21:59] <mathrick> jimmy_dean: but setting num_samples to real value doesn't break anything I assume? So just ignore it and move along ;)
[21:59] <jimmy_dean> mathrick: the reason why I ask is because I'm trying to setup some of the calculations that only need to be done once and that's when the plugin initializes for the first time...but I don't have access to buf then
[21:59] <jimmy_dean> mathrick: yeah, but the question is what real number for right now?
[21:59] <mathrick> jimmy_dean: what exactly do you need to do?
[22:01] <mathrick> jimmy_dean: ie. do you need audio data for that one time calculations?
[22:01] <jimmy_dean> mathrick: basically I setup a bunch of data arrays that have to do with grouping the bands and preparing the FFTW library before it actually takes the DFT in my chain function
[22:01] <jimmy_dean> mathrick: well I need either a constant number or access to the audio data buffer in my init function
[22:02] <jimmy_dean> mathrick: I call my init_eq function from this function: static void
[22:02] <jimmy_dean> gst_equalizer_init (GstEqualizer *equalizer)
[22:03] <mathrick> jimmy_dean: you only need number of samples?
[22:03] <jimmy_dean> mathrick: right
[22:05] <mathrick> jimmy_dean: then just postpone it. add some 'set' gboolean field to elem, and when chain func is called for the first time, do what you need. Later, when you update to do proper caps nego, you could avoid that lengthy preparations in actual data processing by retrieving buffersize from caps and doing it then
[22:06] <jimmy_dean> mathrick: alright, it'll be a very inefficient equalizer until that point :)
[22:06] <mathrick> jimmy_dean: at least, i think it should be doable at nego time
[22:06] <mathrick> jimmy_dean: why? you do that only once
[22:06] <mathrick> jimmy_dean: so only first buffer is expensive
[22:07] <jimmy_dean> mathrick: wait, were you saying to setup a boolean to make sure I only do that init stuff once in my chain function?
[22:07] <mathrick> jimmy_dean: yep
[22:07] <jimmy_dean> ahh ok, yeah...good idea
[22:07] <jimmy_dean> I misread you the first time
[22:07] <mathrick> jimmy_dean: and flush it on eos / renego
[22:07] <jimmy_dean> not sure what that is
[22:08] Action: Jaramir is away: Recursive traversal of loopback mount points
[22:08] <mathrick> jimmy_dean: EOS == End Of Stream, renego == caps renegotiation, it happens under various circumstances, but you can add it later
[22:09] <mathrick> jimmy_dean: don't bother with these now, you don't do nego at all now
[22:09] apoc ([email protected]) left irc: "Leaving"
[22:10] <jimmy_dean> mathrick: ok, I'm gonna need a lot of help polishing this thing up later :)
[22:10] <jimmy_dean> mathrick: in terms of plugins, it's a hack right now
[22:10] <mathrick> jimmy_dean: do you install event handler for your element?
[22:10] <jimmy_dean> mathrick: I have no idea :)
[22:10] <jimmy_dean> mathrick: depends on if the template did that or not
[22:10] <mathrick> jimmy_dean: it's already better than FFT_LENGTH ;)
[22:11] <jimmy_dean> mathrick: yeah, steve_b told me though that LADSPA always assumes mono, interleaved float data...so that's probably why everything can be a constant
[22:11] <mathrick> jimmy_dean: gst-inspect-0.7 equalizer | grep AWARE
[22:11] <jimmy_dean> mathrick: nothing back, so no
[22:12] <mathrick> jimmy_dean: you will need to add it later then, if you have data structures that need to be initialized on new datastream
[22:12] <jimmy_dean> mathrick: ok
[22:13] <mathrick> jimmy_dean: but for now, it's relatively safe to ignore that, at most RB will crash when changing tracks ;)
[22:13] <jimmy_dean> mathrick: gotcha...well I'll be happy to get it to equalize anything first
[22:13] <jimmy_dean> mathrick: the rest is cake because people know how to do the rest
[22:14] Misirlou ([email protected]) left irc: Read error: 60 (Operation timed out)
[22:14] <mathrick> jimmy_dean: and you'll be better than XMMS, which AFAIK by default equalizes mp3's only :)
[22:15] <jimmy_dean> mathrick: only does mpegs, not just by default...it can't equalizer anything else
[22:15] Shoragan ([email protected]) left irc: "Leaving"
[22:16] <mathrick> jimmy_dean: 'by default' == without EQ XMMS plugin
[22:17] <jimmy_dean> ahh, yes :)
[22:17] <jimmy_dean> mathrick: the days of xmms are numbered :)
[22:17] <mathrick> jimmy_dean: but that's stranger, means that XMMS architecture must be incredibly limited
[22:17] <mathrick> s/stranger/strange/
[22:17] <jimmy_dean> mathrick: well just the eq, it's a very simple one
[22:18] <alley_cat> xmms architecture is pretty limited
[22:18] <mathrick> jimmy_dean: but one'd assume it equalizes audio data, not some specific file type
[22:19] <alley_cat> try supporting a container that can contain mp3/vorbis/ac3/aac audio in xmms, you'll have to decode everything yourself
[22:20] <jimmy_dean> mathrick: well it doesn't do a DFT and it uses filters to select each band...so it can't select bands if it doesn't take the dft, so it can only equalizer mpegs since they are stored in the frequency domain (or so they told me)
[22:23] apoc ([email protected]) joined #gstreamer.
[22:24] <mathrick> jimmy_dean: what exactly does DFT stand for?
[22:25] <jimmy_dean> mathrick: discrete fourier transform
[22:25] <jimmy_dean> mathrick: discrete instead of continuous time
[22:25] <mathrick> jimmy_dean: so it's other way to name time to freq translation?
[22:26] <jimmy_dean> mathrick: yeah, except it's on an array of data instead of an analog, continuous time signal...aka digital versus analog
[22:28] <mathrick> jimmy_dean: and OTOH FFT breaks continuous signal into series of components that are still continuous?
[22:29] <jimmy_dean> mathrick: well no, by definition each frequency component is just that, a component
[22:29] <jimmy_dean> mathrick: it's the difference of input basically
[22:32] <mathrick> jimmy_dean: hmm, could you explain more?
[22:33] <mathrick> jimmy_dean: by component, I mean one sine function, with given frequency, amplitude, and phase shift
[22:34] <jimmy_dean> mathrick: yeah, except on period of a sine function will give you one resulting spike
[22:34] <jimmy_dean> mathrick: the spike actually has a special name called a delta function
[22:34] <mathrick> jimmy_dean: Dirac's delta?
[22:35] <ds> mathrick: no
[22:35] <dolphy> ds: yop
[22:35] <dolphy> ds: got  5 mins to try a pipeline ?
[22:35] <jimmy_dean> mathrick: there yah go, not familiar with Dirac
[22:36] <ds> dolphy: got no up-to-date system
[22:36] <ds> dolphy: but I can try it later
[22:36] <dolphy> ds: should make the same issue anyway
[22:36] <mathrick> jimmy_dean: don't know if it's 'official' name, just called it that way on my basic EE course
[22:36] <dolphy> ds: videotestsrc ! ffcolorspace ! videoscale ! ximagesink
[22:36] <ds> oh yeah
[22:36] <ds> what about it?
[22:36] <dolphy> ds: the window is very small
[22:36] <jimmy_dean> mathrick: ahh ok, well then probably
[22:36] <ds> it works fine :)
[22:36] <ds> dolphy: so?
[22:36] <dolphy> ds: if i remove ffcolorspace it has a normal size
[22:37] <dolphy> ds: you think that s normal ?
[22:37] <ds> dolphy: yes, it's normal
[22:37] markov ([email protected]) joined #gstreamer.
[22:37] <mathrick> jimmy_dean: it's something what you get for ex by discharging capacitor w/o resistance at all, right?
[22:37] <mathrick> hi markov
[22:37] <markov> re
[22:38] Action: ds wonders if markov is a bot :)
[22:38] <dolphy> ds: well i don't understand why but anyway try that one now
[22:38] <dolphy> ds: videotestsrc ! switch ! ffcolorspace ! videoscale ! ximagesink
[22:38] <jimmy_dean> mathrick: yes, you would get that
[22:38] <dolphy> ds: video output is borked
[22:38] <jimmy_dean> mathrick: otherwise called an impulse
[22:38] <ds> dolphy: if you care about size, you'd use filtered caps
[22:39] <dolphy> ds: from all the tests i have done with switch
[22:39] <dolphy> ds: it seems that having 2 elements proxying caps connected together makes weird stuff in caps nego
[22:39] <jimmy_dean> mathrick: in mechanical engineering you can get an impulse by hitting something with a hammer...it's an instantaneous noise that is ideally not supposed to have any width
[22:39] <dolphy> ds: ffcolorspace and switch being the naughty ones i have tested most
[22:40] <ds> dolphy: it will a bit.  but only in relation to fixating
[22:41] <dolphy> ds: well as soon as i put switch in my gstplay pipeline the video is green and pink
[22:41] <mathrick> jimmy_dean: so, how is it related to FFT exactly?
[22:41] <dolphy> ds: and quite uniform
[22:42] <jimmy_dean> mathrick: well I was just making a comparison...it's like that except not exactly
[22:42] <jimmy_dean> mathrick: an FFT gives you a vertical bar but it does have a certain amplitude
[22:43] <mathrick> jimmy_dean: what about freq and phase shift? I think FFT is supposed to give you these too?
[22:43] <dolphy> ds: i m really trying to understand what's wrong with switch but reading the debug output 10 times didn't help
[22:43] <dolphy> ds: i see the proxying happening
[22:43] <dolphy> ds: from switch to colorspace and from colorspace to ximagesink
[22:43] <jimmy_dean> mathrick: yes, frequency is on the x-axis...meaning each vertical bar is located at a particular frequency
[22:43] <dolphy> ds: but in the end the intersection is yuv
[22:44] <mathrick> ok, anyone familiar with installing printer for CUPS on Debian?
[22:44] <Company> dolphy: proxy the link, too
[22:44] <dolphy> ds: which is completely weird as ximagesink gives rgb
[22:44] <dolphy> Company: not sure i understand what you mean?
[22:44] <Company> dolphy: you only proxy getcaps, which means you can have different caps negotiated on src and sink side
[22:44] <jimmy_dean> mathrick: and the phase shift is the arctan(complex/real) which comes from each vertical bar
[22:44] <ds> dolphy: it sounds like it's still early in the negotiation
[22:45] <Company> dolphy: you need to proxy the link function, too
[22:45] <mathrick> jimmy_dean: ahh, so these bars are on complex plane actually? good to know :)
[22:46] <jimmy_dean> mathrick: yes, definitely :)
[22:46] <jimmy_dean> mathrick: the magnitude of each vertical bar is the real part
[22:47] Action: mathrick tries to remember how exactly was frequency expressed on complex plane
[22:47] <dolphy> Company: are there helper function to do that ?
[22:47] <dolphy> Company: gst_pad_proxy_link ?
[22:47] <Company> dolphy: look at how idenitity does it
[22:47] <Company> dolphy: or tee
[22:47] <Company> dolphy: yeah, i think gst_pad_proxy_link is the name there
[22:48] <dolphy> gst_pad_proxy_pad_link
[22:48] <dolphy> Company: hmm i ll check that
[22:48] <dolphy> Company: thx
[22:48] Uraeus ([email protected]) joined #gstreamer.
[22:48] <dolphy> Company: need to proxy fixate too? 
[22:48] <Company> dolphy: nope
[22:50] <dolphy> Company: so basically switch should proxy link and getcaps on all sink and the single src
[22:50] <Company> dolphy: yeah
[22:50] <dolphy> Company: and try to set src caps on sink pads when requested ?
[22:51] <Company> yeah
[22:54] apoc ([email protected]) left irc: "Leaving"
[22:55] <alley_cat> hmm, is freedesktops anoncvs down? i only get timeouts
[22:58] somex1 ([email protected]) left #gstreamer.
[23:02] sub_pop ([email protected]) left irc: "Client exiting"
[23:03] ds ([email protected]) left irc: Read error: 60 (Operation timed out)
[23:04] zeenix ([email protected]) joined #gstreamer.
[23:04] <Uraeus> hi zeenix
[23:04] <zeenix> fedora sucks fedora sucks fedora sucks...
[23:04] <mathrick> hi zeenix
[23:04] <zeenix> hi BTW :)
[23:05] <dolphy> hmm core is currently broken
[23:05] <Uraeus> zeenix: no fedora rocks fedora rocks fedora rocks
[23:05] <dolphy> it does not build
[23:07] Action: alley_cat kicks freedesktop cvs into shape
[23:07] <alley_cat> can someone send me the current configure.ac and ext/Makefile.am of gst-plugins? i just want to make a patch for nassink
[23:11] <zeenix> Uraeus: redhat-config-packages works alright for you?
[23:11] <Uraeus> zeenix: YES
[23:13] <zeenix> Uraeus: here, it just give an error window (not showing an details of it) ASA it actually starts to install packages
[23:13] sub_pop ([email protected]) joined #gstreamer.
[23:14] <alley_cat> dolphy: thanks
[23:14] <dolphy> alley_cat: np
[23:17] kmaraas ([email protected]) left irc: "Leaving"
[23:21] <zeenix> what was the name of omega's current project?
[23:21] <zeenix> pbx-colo?
[23:23] <Company> pdxcolo i think
[23:23] <dolphy> anybody who can fix gstpad.c in the core ?
[23:23] <dolphy> thomasvs broke it it seems
[23:23] <Uraeus> dolphy: fire him :)
[23:26] <Uraeus> ok, gotta go to sleep; in a moment of weakness I said I be at work by 0800 tommorow; which turned out to be earler than the other people had planned to start, but then it was on the table and there was no return
[23:27] ds ([email protected]) joined #gstreamer.
[23:28] Uraeus ([email protected]) left irc: "Client exiting"
[23:29] wheels ([email protected]) joined #gstreamer.
[23:29] <mathrick> hi wheels
[23:29] <wheels> howdy
[23:30] Action: dolphy fixes gstelement.h to get core to build again
[23:32] <dolphy> committing
[23:35] <taaz> anyone here live in san diego?
[23:40] <dolphy> hmm riff is broken too
[23:41] <dolphy> oops
[23:43] <dolphy> forgot to cvs update -dP in gst-libs :)
[23:45] <LeRoutier> dolphy, don't worry, last time i checked, v4l2 did not compile
[23:45] <LeRoutier> because of i18n changes
[23:47] <dolphy> will fix that then
[23:47] <dolphy> when i see the error
[23:47] <dolphy> currently building
[23:48] kmaraas ([email protected]) joined #gstreamer.
[23:49] <dolphy> xvidenc..
[23:49] <dolphy> hmm
[23:53] Nick change: harshyWORK -> harshy
[23:53] BBB ([email protected]) joined #gstreamer.
[23:55] <LeRoutier> re BBB
[23:55] <mathrick> yo BBB
[23:56] <LeRoutier> dolphy, did not happen for you ? (the v4l2 compile error)
[23:56] <BBB> lo
[23:56] <BBB> should be fixed now
[23:56] <BBB> but anyway
[23:56] <BBB> I'll try tomorrow
[23:57] Action: BBB is finishing reports right now
[23:57] <dolphy> LeRoutier: nope got xvidenc before
[23:57] <dolphy> LeRoutier: fixing
[23:58] <dolphy> missing GST_LIBRARY_ERROR_ENCODE in gsterror.h
[23:58] <dolphy> ok built gst-plugins successfully
[00:00] --- Tue Jan 20 2004
[00:00] <BBB> oh my god
[00:00] <BBB> these mplayer people just did it
[00:01] <BBB> they flamed and intimidated a guy so badly that he offered to buy them a new server
[00:01] <Company> ?
[00:01] <BBB> ... no clue how they did that
[00:01] <BBB> his question: "what big a server do you guys need?"
[00:01] <BBB> I don't even need to see the answer, these people will just scream that ten quad xeons isn't enough for their ego-boosting product, bla bla bla
[00:02] Action: BBB deletes whole thread from inbox
[00:04] Nick change: bluejay -> bluejay|gone
[00:12] <dolphy> requested pads are disappearing when unlinked ?
[00:13] <BBB> I don't think we ever agreed on a final solution for that
[00:13] <BBB> ocmpany wrote some doc on it, it says who it's supposed to be
[00:14] <Company> request pads disappear when unlinked unless you have a reference to them
[00:14] <dolphy> so i can link unlink to switch a gazillion time
[00:15] <dolphy> there will be only one pad
[00:15] <Company> if it's implemented correctly, then yes :)
[00:17] <dolphy> removing an element from a bin unrefs it ?
[00:17] <dolphy> it s not really clear in gstbin.c code
[00:18] <Company> if you remove an element from a bin, the bin gives up its reference to the element
[00:18] <Company> so yes, it's unreffed
[00:18] <dolphy> gst_object_unparent does it
[00:18] <Company> yeah
[00:18] <dolphy> i didn't see that one at first sight
[00:19] <dolphy> ok
[00:19] <dolphy> Company: i found my issue with the video thread
[00:19] <dolphy> Company: i need to put the vis_queue, goom and colorspace in a bin that i add remove each time i want vis
[00:20] <dolphy> Company: otherwise  they will never go to play because no data coming from src
[00:20] <LeRoutier> dolphy, so, v4l2 compiled for you ?
[00:21] Misirlou ([email protected]) joined #gstreamer.
[00:26] <ds> where's thomasvs?
[00:26] Action: ds kicks thomasvs
[00:26] <Company> the big commits always shuffle stuff :)
[00:27] <ds> no, it's a directory with bad permissions in cvs
[00:27] Action: Company is happy he only uses cvs to connect to cvs
[00:28] <ds> well, check out '.', and you'll understand why, too
[00:31] markov ([email protected]) left irc: "nite"
[00:32] <dolphy> rahh
[00:33] <dolphy> when i have an audio only media spider is not sending data to video queue then the video thread stops after a while
[00:33] <dolphy> shit
[00:34] Action: dolphy reconsiders the gstplay pipeline
[00:36] <BBB> you have video-thread, vis-thread and video-output-thread?
[00:36] <dolphy> nope
[00:36] <dolphy> BBB: http://getcontrol.sourceforge.net/files/gstplay_pipeline.png
[00:37] <dolphy> BBB: if there s no video spider is not connecting the highest queue
[00:37] <BBB> *check*
[00:37] <dolphy> BBB:  and video thread fails to get to PLAYING
[00:38] <dolphy> well i guess that's still not the right pipeline
[00:39] <BBB> you make scheduling a hell, dude
[00:40] <dolphy> well if you have better proposal :)
[00:40] <BBB> make sure the video bits that are before the switch are not in the thread
[00:40] <BBB> then, it'll work
[00:40] <dolphy> putting them in the main thread ?
[00:43] <dolphy> well i m too tired to think now
[00:43] <dolphy> going to bed
[00:43] <dolphy> night
[00:43] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[00:46] sublett ([email protected]) left irc: "I like food, food is good!"
[00:49] pippin ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[00:59] bforbes ([email protected]) joined #gstreamer.
[01:12] The_Company ([email protected]) joined #gstreamer.
[01:13] Company ([email protected]) left irc: Nick collision from services.
[01:13] Nick change: The_Company -> Company
[01:16] thaytoo ([email protected]) joined #gstreamer.
[01:16] BBB ([email protected]) left irc: "Client exiting"
[01:21] Nick change: bforbes -> mxpxpod
[01:22] harshy ([email protected]) left irc: "I quit for now"
[01:30] jowenn ([email protected]) joined #gstreamer.
[01:30] <jowenn> hi
[01:32] <jowenn> Is it possible to get a change of filesrc's location property recognized without deleting and recreating the whole pipeline ?
[01:38] sub_pop ([email protected]) left irc: "Client exiting"
[01:39] harshy ([email protected]) joined #gstreamer.
[01:46] <jowenn> cu
[01:46] jowenn ([email protected]) left irc: "using sirc version 2.211+KSIRC/1.2.1"
[01:58] <LeRoutier> good night
[01:58] LeRoutier ([email protected]) left irc: "Leaving"
[02:00] harshy ([email protected]) left irc: "I quit for now"
[02:07] alley_cat ([email protected]) left irc: "May the Source be with you!"
[02:08] sub_pop ([email protected]) joined #gstreamer.
[02:12] mxpxpod ([email protected]) left #gstreamer ("Client exiting").
[02:12] zeenix ([email protected]) left irc: "Leaving"
[02:14] Misirlou ([email protected]) left irc: "asdf"
[02:14] <jimmy_dean> if the pad being passed into my chain function is NULL, shouldn't this line fail?  g_return_if_fail (GST_IS_PAD (pad));
[02:17] <Company> yes
[02:18] <Company> but if the pad is null, that's a bug in gstreamer :)
[02:18] <jimmy_dean> Company: ok, well what about these three lines from gdb when my plugin segfaults:
[02:18] <jimmy_dean> Program received signal SIGSEGV, Segmentation fault.
[02:18] <jimmy_dean> [Switching to Thread 16384 (LWP 7322)]
[02:18] <jimmy_dean> 0x401338a9 in gst_equalizer_chain (pad=0x0, in=0x8061694) at gstequalizer.c:461
[02:18] <jimmy_dean> 461           gains[i] = db_table[LIMIT(gain_idx, 0, 999)];
[02:18] <jimmy_dean> why is pad=0x0 ?
[02:19] <Company> never trust gdb after a segfault
[02:19] ct_ ([email protected]) left irc: "Client exiting"
[02:19] <Company> it's probably one of the arraysx that's out of bounds
[02:19] <jimmy_dean> Company: ok, how do you use it to debug your program then? :)
[02:19] <jimmy_dean> Company: yeah, I do a "p i" in gdb and it says 0
[02:20] <Company> and gain_idx ?
[02:20] <jimmy_dean> 740
[02:20] <Company> and db_table[740] ?
[02:20] <jimmy_dean> Cannot access memory at address 0x1720
[02:21] <Company> there you have it :)
[02:21] <jimmy_dean> yeah, this is interesting stuff...I'm having to do some guess work with the differences between LADSPA and GStreamer and what types of data comes in
[02:22] <jimmy_dean> Company: they assume 1024 data elements coming in from the sound buffer
[02:22] <jimmy_dean> Company: I cannot do that, so that makes for an interesting time translating all of these arrays and their upper limits
[02:23] <jimmy_dean> Company: is the author of the LADSPA plugins still around?
[02:23] <jimmy_dean> Company: maybe I should email that person/persons
[02:24] <Company> no idea
[02:24] <jimmy_dean> I'm really close to making this work...I just have to solidify a couple of concepts about certain variables and what they represent exactly and then it should work on a basic level
[02:25] Rotty ([email protected]) left irc: Remote closed the connection
[02:25] iain ([email protected]) left irc: "I'm the kinda guy who leaves the scene of a crime"
[02:26] wheels ([email protected]) left irc: Read error: 113 (No route to host)
[02:32] <mathrick> hmm, are the v4l2 issues fixed yet?
[02:33] <jimmy_dean> Company: hmm, pad has to be zero because everything that I print off in gdb having to do with the element data members is zero in my chain function and I know they are getting set up above in my init function
[02:35] <Company> yes, but that's only because the segfault screwed your data
[02:35] <Company> if you set a breakpoint before the swegfault, that value won't be NULL
[02:36] <jimmy_dean> Company: actually, how do you set a breakpoint...I never knew how to do it with gdb
[02:36] wheels ([email protected]) joined #gstreamer.
[02:37] <mathrick> jimmy_dean: bt <line or function>
[02:37] <mathrick> s/bt/b/
[02:38] <jimmy_dean> mathrick: you do that before you run a program?
[02:38] <jimmy_dean> mathrick: how's it know what source file you're talking about?
[02:38] <mathrick> jimmy_dean: basically, you can do that anytime, provided gdb is taking your input at the time
[02:39] <mathrick> jimmy_dean: if it's ambigous, use srcfile:<line or function>
[02:39] <mathrick> jimmy_dean: although, this sometimes has funny issues
[02:39] <jimmy_dean> mathrick: ok, thanks...I'll give it a try...how do you step through code line by line then?
[02:39] <mathrick> jimmy_dean: s or n
[02:40] <mathrick> jimmy_dean: i think quoting helps if it doesn't want to load function from given file
[02:40] <mathrick> jimmy_dean: s means step (into), and n is next (line, w/o stepping into function calls)
[02:41] <jimmy_dean> mathrick: ok, thanks a lot
[02:41] <mathrick> jimmy_dean: there's also u (until), which always advances in terms of source line, so is useful for getting out of loops and such
[02:41] <mathrick> jimmy_dean: all of these can take numerical params
[02:42] <mathrick> jimmy_dean: and, you can enable / disable breakpoints with ena / dis <number>
[02:42] <mathrick> jimmy_dean: and d stands for delete, to remove breakpoint
[02:43] <jimmy_dean> mathrick: what about working directory...if I'm launching gst_launch does it assume my current plugin directory that I'm in in the shell or gst_launches working directory?
[02:43] <mathrick> jimmy_dean: w is watch (hardware if possible), useful, but beware of pointer vs. variable issues, it's sometimes funky wrt that
[02:44] <mathrick> jimmy_dean: the dir you started gdb from, but don't know for sure
[02:44] <mathrick> jimmy_dean: you can always cd inside gdb
[02:45] <mathrick> ok people, time to sleep for me
[02:45] <mathrick> nite
[02:46] <jimmy_dean> this is not working :)
[02:46] <jimmy_dean> ok, night
[02:46] <mathrick> jimmy_dean: what isn't working?
[02:47] <jimmy_dean> mathrick: I did a cd to my src directory within gdb, still can't find my source file though when I do: b gstequalizer.c:429
[02:47] <jimmy_dean> when I know it exists and I can see it from gdb with tab completion
[02:47] <mathrick> jimmy_dean: try quoting it
[02:48] <jimmy_dean> still no
[02:48] <mathrick> jimmy_dean: like 'gstequalizer.c;:429, or 'gstequalizer.c:429'
[02:48] <mathrick> jimmy_dean: i told you, it has issues sometimes
[02:48] <jimmy_dean> yeah, I'd say
[02:48] <mathrick> jimmy_dean: :)
[02:48] <Company> it only works when the plugin is loaded
[02:48] <jimmy_dean> ahh, maybe that was it
[02:48] <Company> which doesn't work until the dlopen call
[02:48] <mathrick> jimmy_dean: ah, forgot to tell you that :)
[02:49] <jimmy_dean> yeah, it set it then
[02:49] <Company> gdb keeps it across restarts
[02:50] <Company> ... of your app
[02:50] <Company> <mathrick> hmm, are the v4l2 issues fixed yet? <-- it's now
[02:50] <jimmy_dean> well it erased any knowledge of gstequalizer.c once I restart the app though and run from the beginning
[02:51] <Company> it keeps complaining it can'T set the breakpoint
[02:51] <Company> but when it loads the plugin it sets it and the complaining stops
[02:51] <mathrick> Company: hmm? you mean they are, right?
[02:51] <Company> they are fixed
[02:51] <Company> it was only one issue (missing include)
[02:51] <mathrick> Company: thx
[02:52] <mathrick> jimmy_dean: and use hw watchpoint generously, they are invaluable, but you must be extra careful about what is adress, and what actual variable you want to watch
[02:52] <mathrick> jimmy_dean: takes some time to practice it out
[02:52] <jimmy_dean> mathrick: yeah, time is something I don't have anymore for a few days :)
[02:53] <jimmy_dean> mathrick: this is taking a long time...all I want is an equalizer :)
[02:53] <mathrick> jimmy_dean: btw, usually on x86 you can st up to 8-10 hw watchpoints
[02:53] <mathrick> jimmy_dean: and you're getting knowledge instead ;P
[02:53] <jimmy_dean> yes, very true
[02:53] <jimmy_dean> I'll thank myself in the long run
[02:53] <mathrick> ok, now i'm really asleep :)
[02:53] <jimmy_dean> night
[02:54] Nick change: mathrick -> mathrick|sleep
[02:58] Nick change: Jaramir -> Jara[zZ]


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