IRC Logs

*******************************************************************
[03:01] Marsupilami23 ([email protected]) joined #gstreamer.
[03:01] walters ([email protected]) left irc: "out"
[03:06] sublett ([email protected]) left irc: "I like food, food is good!"
[03:09] sub_pop ([email protected]) left irc: "Client exiting"
[03:12] walters ([email protected]) joined #gstreamer.
[03:16] thaytan ([email protected]) left irc: Read error: 110 (Connection timed out)
[03:25] Shoragan ([email protected]) left irc: "Leaving"
[03:29] <mathrick> nite
[03:29] Nick change: mathrick -> mathrick|sleep
[03:32] thaytan ([email protected]) joined #gstreamer.
[03:32] <Company> hum
[03:33] <Company> if (someuint < -someint) {
[03:33] <Company> is there a nicer way to write that?
[03:33] <Company> if (someuint < -someint && someint < 0) {
[03:34] <Company> that's what i meant
[03:35] Misirlou ([email protected]) joined #gstreamer.
[03:35] <ds-work> I'd write it the other way
[03:36] <taaz> some people would use abs() too
[03:39] <ds-work> we should turn on -Wsign-compare
[03:39] <Company> absolutely
[03:40] <taaz> depends, is it good news or bad? ;)
[03:40] <ds-work> generally indicates bugs
[03:40] <Company> we should turn on ever warning that doesn't issue too much false warnings (like -Wall in forte)
[03:41] <Misirlou> -Wall -Werror -pedantic-errors
[03:42] <Company> what's pedantic?
[03:43] <ds-work> pedantry should be avoided
[03:43] <Misirlou> hehe
[03:44] <Company> no, it shouldn't
[03:44] <Company> (i looked that word up now...)
[03:44] <Company> we even fixed function prototypes so we don't need to cast them...
[03:49] <Company> uh shit
[03:49] <Company> element's base time could go below 0
[03:50] <herzi> is it possible to install gstreamer 0.7 next to 0.6 or will they conflict?
[03:51] <ds-work> that's not so good
[03:51] <ds-work> herzi: no conflicts at all
[03:51] <Company> ds-work: happens when you seek forwards
[04:01] <Company> ds-work: GST_PADDING is 4?
[04:02] <ds-work> yeah
[04:02] <Company> GstClock just lost all its padding on 32bit machines...
[04:03] <ds-work> be careful you didn't replace void *'s with ints
[04:04] <Company>  guint8         padding[sizeof(gpointer) * 4 - sizeof (GstClockTime) * 2];
[04:04] <Company> that should work, no?
[04:04] <ds-work> heh
[04:04] mxpxpod ([email protected]) joined #gstreamer.
[04:05] <ds-work> 0-length arrays aren't portable, iirc
[04:05] <Company> is there a reason to reset clocks?
[04:05] <ds-work> when you go to READY?
[04:06] <Company> yeah
[04:06] <Company> currently the clock is supposed to tell you the time in the pipeline it manages
[04:07] <Company> in my model the clock just tells you the time
[04:07] <Company> the element tells you its own time
[04:09] <Company> so if you wanna know the time inside a pipeline, you can't just ask the clock, you need to ask the element you'Re interested in
[04:10] <Company> hum, a nice effect of that is that srcs and sinks can be at different positions in the stream, so you have buffering
[04:10] <ds-work> that makes sense, because some elements may be "ahead"
[04:10] <Company> it only breaks the current "clock" model
[04:10] <ds-work> osssink will always be slightly ahead of xvideosink
[04:10] <Company> why?
[04:11] <ds-work> at least, the source pad will
[04:11] <ds-work> er, "sink" pad
[04:11] <Company> it will have consumed more data, yeah
[04:11] <ds-work> yeah
[04:11] <Company> but osssink won't output it
[04:13] <Company> hm, my stuff deprecates quite a lot in gstclock.h
[04:13] <Company> because of changing clocks from relative to absolute time
[04:14] <Company> it doesn'T make sense to activate clocks anymore
[04:14] <Company> clocks don't handle disconts anymore
[04:15] <ds-work> this is starting to sound like too big of a change
[04:15] <Company> i'll keep resetting in
[04:15] <Company> (the code)
[04:15] <Company> but deprecate it
[04:16] <Company> activation and discont handling won't work anymore though
[04:17] <Company> because elements rely on clocks being active
[04:17] <Company> hm
[04:17] <Company> only PLAYING element rely on that...
[04:20] <Company> activation and resetting is broken if you have 2 pipelines
[04:20] <Company> so this is a bugfix, cool :)
[04:25] herzi_lap ([email protected]) joined #gstreamer.
[04:25] <herzi_lap> configure: WARNING: Sissy ! By asking to not build the tests known to fail, you hereby waive your right to customer support.  If you do not agree with this EULA, please press Ctrl-C before the next line is printed.  By allowing the next line to be printed, you expressly acknowledge your acceptance of this EULA.
[04:25] <herzi_lap> nice, go on
[04:27] <Company> typical thomasvs :)
[04:29] jdahlin ([email protected]) left irc: Read error: 110 (Connection timed out)
[04:29] jdahlin ([email protected]) joined #gstreamer.
[04:31] <walters> (rhythmbox:4963): GStreamer-CRITICAL **: Could not find a compatible pad on element id3tag34 to link to src_0:sink
[04:31] <walters>  
[04:31] <walters> hmm
[04:32] <Company> that is not critical
[04:32] <Company> and should only be a GST_DEBUG max
[04:33] <Company> ds-work: i need a good default
[04:33] <ds-work> Company: 0
[04:34] <Company> ds-work: clocks allow "clock events" (like state changes) to stay in sync
[04:34] <Company> ds-work: so if multiple events happen shortly after each other, they use the same timestamp
[04:34] <Company> ds-work: how long is "shortly" ?
[04:34] <ds-work> 200 ms
[04:35] <Company> hm
[04:35] <Company> isn't that a bit long?
[04:35] <ds-work> approximate length of human perception of time differences
[04:35] <ds-work> it's also 2 timeslices
[04:35] <Company> i'll use 100 for now
[04:35] <Company> just to be sure
[04:36] <Company> when do humans notice A/V sync issues?
[04:36] <ds-work> iirc, around 100 ms
[04:39] <Company> can you use sizeof in #if statements?
[04:40] <ds-work> no
[04:41] <Company> is there a macro for pointer sizes?
[04:42] <ds-work> yes, but it's better to use a union, imo
[04:43] <Company> we'll fix that tomorrow anyway
[04:43] <Company> so i'll just forget it
[04:43] <Company> or whenever thomas commits his stuff
[04:44] <ds-work> Company: at the least, add the fields now
[04:47] <Company> ds-work: i'm at it
[04:49] <Company> heh, i like gabucino
[04:49] <Company> at least he's honest :)
[04:49] Marsupilami23 ([email protected]) left irc: Remote closed the connection
[04:54] Marsupilami23 ([email protected]) joined #gstreamer.
[04:55] <thaytan> what's the best way to fix a valid pointer type punning?
[04:55] <thaytan> use a temporary var?
[04:55] <ds-work> fix the punning
[04:55] <thaytan> or is there some syntax that tells the compiler 'this pun is ok'
[04:56] <ds-work> use a union
[04:56] <thaytan> in this case, it's passing a gpointer* to retrieve user_data
[04:56] herzi_lap ([email protected]) left irc: "Leaving"
[04:56] <ds-work> then pass a gpointer *
[04:57] <Company> gpointer temp = retrieval; retrieve (&temp);
[04:57] <thaytan> yeah, so use a temporary and cast it
[05:00] <ds-work> it's all rather funny, since mplayer has never been very clean with its licensing
[05:03] <Company> yeah
[05:04] <Company> but mplayer certainly is a nice thing to put into your dvd player
[05:04] <Company> little cpu, it's a big binary anyway, ...
[05:04] <Company> maybe KISS even included the win32 binary codecs ;)
[05:04] <ds-work> my housemate has a KISS DVD player
[05:04] <ds-work> it's rather nice, actually
[05:04] Marsupilami23 ([email protected]) left irc: "I don't wanna grow up, I'm a Toys R Us Kid. There's a million toys at Toys R Us that I can play with!"
[05:05] <ds-work> it plays almost anything
[05:05] <ds-work> however, it doesn't seek well at all
[05:05] <ds-work> and occasionally, the codecs go all flooey
[05:06] <thaytan> hey, that sounds like mplayer to me
[05:06] <Company> yeah, i just thought that, too
[05:06] <Company> does it print out loads of stuff on screen before starting the movei? ;)
[05:06] <ds-work> however, I think mplayer's mpeg4 code looks better than KISS's
[05:06] <ds-work> maybe it's just old ffmpeg :)
[05:08] Marsupilami23 ([email protected]) joined #gstreamer.
[05:11] Marsupilami23 ([email protected]) left irc: Client Quit
[05:11] <ds-work> maybe we should email KISS and tell them to start using GStreamer :)
[05:13] <Company> they'd need to write a player
[05:13] <Company> or add seeking to gst-launch
[05:13] Action: Company imagines gst-launch with lirc support
[05:14] <ds-work> just use ! lircnav ! xvimagesink
[05:14] <ds-work> that's not a bad idea for an element, actually
[05:14] <Company> that's what i just thought, too
[05:14] <Company> and i do have a remote...
[05:15] <Company> /* dedicated to KISS technologies */
[05:21] <ds-work> thaytan was doing stuff with dvdnav to handle rc-type events
[05:22] Action: Company deletes code
[05:23] <Company> it's so much easier to just do gst_element_wait than requesting some id first
[05:23] <ds-work> presumably there was a reason for it
[05:24] <Company> maybe
[05:25] <Company> wtay had many ideas
[05:25] <Company> btw: we still need to make sure all elements send proper disconts at the start of a stream
[05:26] <Company> s/start of stream/negotiation/
[05:26] <ds-work> ha
[05:26] <Company> alsasink relies on that
[05:29] <Company> even the API docs for clocks are wring
[05:29] <Company> "the time of a clock is always increasing"
[05:29] <Company> dude, what does seeking back do?
[05:31] <Company> the code looks more and more like my current tagging code: unfinished
[05:31] Nick change: harshy -> harshyMovie
[05:33] Action: ds-work wonders how to play a movie backwards
[05:33] <Company> that's hard
[05:34] <Company> why does matroska need a clock?!
[05:34] <Company> and why does it (de)activate it?
[05:34] <ds-work> it would demonstrate success of a seeking implementation
[05:34] <Company> try it
[05:35] <Company> get_length, and always seek to length-timestamp and take that picture
[05:35] <Company> but most demuxers only seek to keyframes
[05:35] <ds-work> yeah, you'd have to cache video between keyframes, and play each frame in reverse order
[05:36] <Company> what the hell?
[05:37] <Company> mpegparse/demux does clocking...
[05:37] <ds-work> mpegparse makes sense
[05:37] <ds-work> since it was supposed to be used by tcpsink
[05:37] <ds-work> tcpsink should be doing the clocking, though
[05:39] <Company> tcpsink doesn't know the data format though
[05:39] <Company> but if someone does clocking, it should be tcpsrc
[05:40] <Company> a sink should just accept as fast as possible
[05:40] <ds-work> tcpsink sync=true should send out packets just as fast as the timestamps say to
[05:40] <Company> ah
[05:40] <ds-work> mpegparse should provide those timestamps
[05:40] <Company> i got src and sink wrong again
[05:41] <ds-work> I tend to get really screwed up with srcpad/sinkpad and src/dest
[05:42] <Company> same here
[05:42] <Company> tcpsink does clocking btw
[05:45] Marsupilami23 ([email protected]) joined #gstreamer.
[05:56] <Company> it's always funny when you look over element code and realize noone had a clue what the code does :)
[05:59] <ds-work> heh, like negotiation?
[05:59] <Company> in this case gst_element_clock_wait
[05:59] <Company> and the interesting jitter parameter
[06:01] <Company> and it's funny that dolphy really wrote 2 elements that are 99% equal
[06:05] Misirlou ([email protected]) left irc: "asdf"
[06:05] hadley ([email protected]) joined #gstreamer.
[06:08] <Company> i'm an asshole
[06:08] <Company> always moaning about others
[06:08] <Company> i wonder if that's good or bad
[06:08] <ds-work> it tends to piss people off
[06:09] <Company> yes, but it gets stuff done
[06:09] <ds-work> dolphy wrote xvimagesink and ximagesink separately so as to completely avoid the xvideosink mess
[06:10] <ds-work> no, it doesn't get stuff done
[06:10] <Company> they should have the same base class
[06:10] <Company> GstVideoSink most probably
[06:10] <ds-work> complaining that people are bad coders helps nothing except making yourself feel better
[06:11] <ds-work> I think he intends to merge them eventually
[06:11] <ds-work> we talked about it in Spain
[06:11] <Company> i dunno
[06:12] <Company> (i know that complaining about others being bad coders is bad, that wasn't the question)
[06:12] <Company> the question was if being an asshole helps API decisions sometimes
[06:13] <Company> especially in the simpler cases
[06:13] <ds-work> better would be to show people their poor code, and encourage them to write better code
[06:14] <ds-work> most people want to improve themselves; being told "you suck" doesn't help in that direction
[06:14] <ds-work> people will just wander off to different projects
[06:15] <Company> yeah, that wasn't the question either
[06:16] <Company> i know that i'm more direct than the average person
[06:17] <ds-work> you also have demonstrated that you have less experience than, say, thomasvs or dolphy
[06:18] <Company> less experience?
[06:25] <ds-work> in terms of project management
[06:25] <ds-work> and, uh, interpersonal skills
[06:30] <Company> that's not (just) experience but unwillingness to communicate that way
[06:30] <Company> i decided some time ago that i prefer to say things the way i think them
[06:31] <Company> and at the same time decided i'm not going to do management of persons
[06:31] <ds-work> and eventually, you will realize that that doesn't help the project, or even your own goals
[06:32] <ds-work> anyway, nobody finds you annoying (I think)
[06:32] <Company> this hasn't happened since i decided that, but i'm not that old yet...
[06:33] <Company> i always prefer to be the one doing technical stuff and leaving the communication stuff to someone else
[06:34] <ds-work> well, if it was that obvious, you would have noticed already :)  You're not dumb
[06:36] <Company> i have perfect A/V sync
[06:38] <Company> and it's ABI and API compatible :)
[06:39] Action: ds-work moos
[06:39] <Company> it only changes the meaning of something...
[06:40] <Company> to be exact: the meaning of gst_clock_get_time
[06:52] Action: ds-work goes home
[07:03] Nick change: trow -> trow_away
[07:17] thaytan ([email protected]) got netsplit.
[07:17] kmaraas ([email protected]) got netsplit.
[07:17] lilo ([email protected]) got netsplit.
[07:17] jdahlin ([email protected]) got netsplit.
[07:17] walters ([email protected]) got netsplit.
[07:17] hyriand ([email protected]) got netsplit.
[07:17] ds-work ([email protected]) got netsplit.
[07:17] hadley ([email protected]) got netsplit.
[07:17] aldug ([email protected]) got netsplit.
[07:18] hadley ([email protected]) returned to #gstreamer.
[07:18] jdahlin ([email protected]) returned to #gstreamer.
[07:18] walters ([email protected]) returned to #gstreamer.
[07:18] aldug ([email protected]) returned to #gstreamer.
[07:18] hyriand ([email protected]) returned to #gstreamer.
[07:18] ds-work ([email protected]) returned to #gstreamer.
[07:18] thaytan ([email protected]) returned to #gstreamer.
[07:18] kmaraas ([email protected]) returned to #gstreamer.
[07:18] lilo ([email protected]) returned to #gstreamer.
[07:21] <Company> 8 hour hackathon - clocking fixed, Company happy
[07:21] <Company> but it's 7 am
[07:21] <Company> gnight
[07:21] Company ([email protected]) left irc: Remote closed the connection
[07:22] hadley ([email protected]) left #gstreamer ("Leaving").
[07:27] lilo ([email protected]) got netsplit.
[07:27] lilo ([email protected]) returned to #gstreamer.
[07:28] lilo ([email protected]) got netsplit.
[07:28] lilo ([email protected]) returned to #gstreamer.
[07:28] lilo ([email protected]) got netsplit.
[07:28] <dilinger> man
[07:28] lilo ([email protected]) returned to #gstreamer.
[07:28] <dilinger> i wish i could still do that
[07:28] <dilinger> start hacking at midnight, finish up when it gets light out
[07:28] <dilinger> damned day job..
[07:32] Nick change: mathrick|sleep -> mathrick|uni
[07:33] Action: mathrick|uni wonders if Company hasn't Uni going or something
[07:43] Nick change: harshyMovie -> harshy
[07:46] matrixise ([email protected]) joined #gstreamer.
[07:55] Marsupilami23 ([email protected]) left irc: "I don't wanna grow up, I'm a Toys R Us Kid. There's a million toys at Toys R Us that I can play with!"
[07:58] thaytan ([email protected]) left irc: Read error: 110 (Connection timed out)
[07:59] Marsupilami23 ([email protected]) joined #gstreamer.
[07:59] Marsupilami23 ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[08:17] mathrick|uni ([email protected]) left irc: Read error: 110 (Connection timed out)
[08:35] <walters> ds: here?
[08:37] matrixise ([email protected]) left irc: "Leaving"
[08:43] <ds> somewhat
[08:48] <thomasvz> ds: see anything wrong with my list setter function ?
[08:49] <thomasvz> ds: or should I tranplant it to core ?
[08:49] Nick change: thomasvz -> thomasvs
[08:53] scureboy ([email protected]) joined #gstreamer.
[08:53] <scureboy> hello..
[08:53] <scureboy> any pointer for the fastest mmx/sse of quantize for intra/inter?
[08:54] <ds> thomasvs: er, what?
[08:58] <thomasvs> ds: in my patch to audio.c there was a (start of) a gst_structure list set function, which was missing from the core
[08:58] <thomasvs> ds: is this useful for the core or not ?
[09:00] <ds> possibly
[09:00] <ds> gst_structure_set_field_list() or something
[09:00] <thomasvs> heh, yeah, that's why I'm asking if yo uwould like to look over it
[09:01] <thomasvs> so if you agree with the way I do it, then I can move it to core and finish it
[09:01] <ds> I didn't want to add to much code with switches on GTypes, though
[09:02] <ds> I'd like it to work in a more extensible way
[09:02] <thomasvs> well, forcing to do that on a higher level is worse imo
[09:02] <thomasvs> I didn't quite feel like having to do that in every _getcaps of every audio element
[09:03] <ds> I had to do it exactly once in the entire conversion of gst-plugins
[09:03] <ds> in some mpeg plugin
[09:03] <ds> that is, construct a list
[09:03] <thomasvs> with your remove trick you mean ?
[09:03] <thomasvs> I'm guessing the more extensible trick would be to use macros to handle the switching and typecasting ?
[09:03] <ds> no
[09:03] somex1 ([email protected]) left #gstreamer.
[09:04] <ds> something like glib's type collection stuff
[09:04] <ds> er, value colloection
[09:04] <ds> values register how they interact with va_list
[09:05] <ds> it's rather crappy, but I never researched it further
[09:06] <thomasvs> hm, so every time you could have used a list you used _remove/_intersect instead ?
[09:07] <ds> well, I think _remove and _intersect is better
[09:07] <ds> there were only a few cases where I needed that, too
[09:07] <ds> we don't use lists very often
[09:07] <thomasvs> and is there some way to set a structure field based on a string representation ?
[09:08] <ds> no, but there should be
[09:09] <ds> you could deserialize a GstValueList value
[09:10] <ds> by the way, we should get a static allocation of GTypes from the glib people
[09:10] <thomasvs> is there a way to do that ?
[09:11] <ds> yes, some project has already done that
[09:11] <ds> not something for 0.8, though
[09:11] <ds> static gboolean
[09:11] <ds> gst_value_from_string (GValue *value, const char *s)
[09:14] kmaraas ([email protected]) left irc: "Leaving"
[09:14] <ds> only mpeg2enc and gstid3types use gst_value_list construction stuff
[09:15] scureboy ([email protected]) left #gstreamer.
[09:18] scureboy ([email protected]) joined #gstreamer.
[09:23] kmaraas ([email protected]) joined #gstreamer.
[09:24] swentel ([email protected]) joined #gstreamer.
[09:30] <walters> ds: so negotiation fails completely now with ogg metadata
[09:30] <ds> what pipeline?
[09:31] <walters>  /build/gstreamer-0.7/bin/gst-launch-0.7 -t gnomevfssrc location=/audio-video/music/Evanescence/Fallen/Tourniquet.ogg ! typefind ! spider ! application/x-gst-tags ! fakesink
[09:34] <ds> walters: it's a spider problem (actually, probably a rank problem)
[09:35] <walters> ds: i know we've discussed this before, but i don't see where the filtercaps come into play 
[09:35] <walters> ds: i see them getting set when i set a breakpoint on gst_parse_perform_link
[09:36] <walters> ds: but later, when gst_spider_plug_from_srcpad is called, it's trying to plug ANY to ANY
[09:36] <ds> walters: that's because it's pluggin oggdemux
[09:36] <ds> which is wrong
[09:36] <walters> hmm.
[09:37] <ds> it's supposed to be oggdemux ! vorbistag, but apparently that doesn't happen
[09:38] <ds> but using that directly doesn't work either
[09:40] <ds> talk to Company
[09:42] <walters> actually
[09:42] <walters> oggdemux ! vorbistag works here
[09:44] <walters> except for (process:18230): GStreamer-CRITICAL **: Could not find a compatible pad on element oggdemux0 to link to vorbistag0:sink
[09:44] <ds> sorry
[09:44] <ds> should be oggdemux .src ! vorbistag
[09:44] <ds> src is a sometimes pad
[09:44] <ds> so you have to specify it
[09:44] <walters> indeed, that appears to do nothing at all
[09:50] <ds> http://www.schleef.org/~ds/patch-oggdemux
[09:50] <ds> that pulls tags from vorbis files here
[09:50] <walters> that's kind of a hack, no?
[09:50] <ds> don't care
[09:51] <ds> it's technically not fixable with the current spier
[09:51] <ds> spider
[09:51] <walters> hmm.
[09:52] <ds> also, someone needs to write a vorbisdec asap
[09:55] thomasvs ([email protected]) left irc: Read error: 110 (Connection timed out)
[09:57] mathrick ([email protected]) joined #gstreamer.
[09:57] <mathrick> re
[10:10] kmaraas ([email protected]) left irc: "Leaving"
[10:11] <walters> walters@nexus> /build/gstreamer-0.7/bin/gst-launch-0.7 --gst-debug=GST_AUTOPLUG=5:spider=5 -t gnomevfssrc location=/audio-video/music/Evanescence/Fallen/Tourniquet.ogg ! typefind ! spider ! application/x-gst-tags ! fakesink
[10:11] <walters> DEBUG          spider(19169) gstspider.c(215):gst_spider_request_new_pad: successuflly created requested pad spider0:src_0
[10:11] <walters> RUNNING pipeline
[10:11] <walters> DEBUG          spider(19169) gstspider.c(594):gst_spider_plug_from_srcpad: trying to plug from sink_ident:src to src_0
[10:11] <walters> DEBUG          spider(19169) gstspider.c(622):gst_spider_plug_from_srcpad: found a link that needs 1 elements
[10:11] <walters> DEBUG          spider(19169) gstspider.c(635):gst_spider_plug_from_srcpad: 1 elements must be inserted to establish the link
[10:11] <walters> DEBUG          spider(19169) gstspider.c(501):gst_spider_create_and_plug: Adding element vorbisfile0 of type vorbisfile and syncing state with autoplugger
[10:12] <walters> how is that possible?
[10:12] <walters> i mean, the link between spider and fakesink is filtered by application/x-gst-tags
[10:12] <walters> and vorbisfile can't do that
[10:16] <walters> oh well, gotta sleep.
[10:16] walters ([email protected]) left irc: "out"
[10:16] KoRnouille ([email protected]) joined #gstreamer.
[10:27] thomasvs ([email protected]) joined #gstreamer.
[10:31] KoRnouille ([email protected]) left irc: "Leaving"
[10:41] wheels ([email protected]) joined #gstreamer.
[10:43] LeRoutier ([email protected]) joined #gstreamer.
[10:43] <LeRoutier> hello
[11:04] <LeRoutier> is someone here using v4l2 in gstreamer ?
[11:05] dolphy ([email protected]) joined #gstreamer.
[11:05] <dolphy> morning
[11:05] <LeRoutier> hi dolphy
[11:06] <mathrick> y0 d01phy ;)
[11:07] <LeRoutier> seems v4l2 plugin looks for videodev2.h  but i only have videodev.h with my v4l2 driver
[11:12] sack ([email protected]) left irc: "Client Exiting"
[11:12] <thomasvs> dolphy: what joystick type do you have at home ?
[11:12] <thomasvs> the wireless one ?
[11:14] <dolphy> thomasvs: logitech wingman
[11:19] <scureboy> guys,
[11:19] <scureboy> any pointer for tutorial about mpeg quantization ?
[11:22] <mathrick> scureboy: http://www.google.com/search?q=mpeg%20quantization%20tutorial
[11:23] <scureboy> hehe
[11:23] <scureboy> ok tq
[11:23] <LeRoutier> argh, gentoo rsync mirrors are all down.
[11:31] <LeRoutier> anyone have kernel 2.6.x here ?
[11:35] <LeRoutier> can you do this please and paste me the result in PV or on the chan : ll /usr/include/linux/videodev*.h
[11:38] sxpert_work ([email protected]) left irc: "Leaving"
[11:44] <LeRoutier> as no one seems awaken, i created this one : http://bugzilla.gnome.org/show_bug.cgi?id=131320
[11:46] <danb> LeRoutier: i'm awake... hold up
[11:46] <danb> -rw-r--r--    1 1046     1046        14390 2003-12-17 21:00 /usr/include/linux/videodev.h
[11:46] <danb> -rw-r--r--    1 1046     1046        26327 2003-12-17 20:59 /usr/include/linux/videodev2.h
[11:46] <danb> those are for 2.6.0
[11:46] <LeRoutier> ok, so bttv author fucked this up
[11:47] <dolphy> LeRoutier: you probably applied a wrong patch
[11:47] <LeRoutier> as when v4l2 is present in 2.4.x, there is only header n°1 
[11:48] <LeRoutier> dolphy : this one : http://bytesex.org/patches/2.4.24-1/
[11:48] <LeRoutier> official bttv patch against kernel 2.4.24
[11:49] <LeRoutier> and my videodev.h has some places where it talks about v4l2 specific members in structures
[11:49] sxpert_work ([email protected]) joined #gstreamer.
[11:50] <LeRoutier> argh. got a videodev2.h in my /usr/src/linux-2.4.24 tree but it did not install
[11:50] <thomasvs> bah
[11:50] <thomasvs> we need a GStreamer team/Foundation for copyright stuff
[11:54] <sxpert_work> morning people
[11:56] Shoragan ([email protected]) joined #gstreamer.
[12:05] <LeRoutier> thomasvs, did the last commit you did fixed the es.po problem ?
[12:09] <LeRoutier> configure.ac:ALL_LINGUAS="de es no nl" => only de.po exists
[12:11] <mathrick> LeRoutier: Log message:removing empty translation files
[12:11] <mathrick> Removed files:    
[12:11] <mathrick>     po              : es.po nl.po no.po
[12:12] <LeRoutier> well, so the ALL_LINGUAS should be changed in order for gstreamer to compile
[12:13] <thomasvs> LeRoutier: I'm working on it, hang on a little bit
[12:13] foser ([email protected]) joined #gstreamer.
[12:13] <LeRoutier> thomasvs, k, no hurry
[12:13] <LeRoutier> hello foser
[12:14] <foser> oi
[12:14] iain ([email protected]) joined #gstreamer.
[12:16] Action: thomasvs tries to figure out how he can get one of those tagging strings to show up
[12:19] <thomasvs> GStreamer options:
[12:19] <thomasvs>   --gst-version                     Druk de GStreamer versie af
[12:19] <thomasvs>   --gst-fatal-warnings              Maak alle waarschuwingen fataal
[12:19] <thomasvs> whee
[12:20] gheet ([email protected]) joined #gstreamer.
[12:20] Action: mathrick sees maany a's aall aaround ;)
[12:21] <iain> waat iz dat?
[12:21] <iain> I loved flying KLM to south africa
[12:21] <LeRoutier> i wonder why i did avoid this language that much when i was at school.
[12:21] <iain> to see how many words in dutch and english were so similar
[12:22] <iain> except for the dutch looked like it was ebonics
[12:22] <thomasvs> iain: well, south african is like dutch from the 16th century
[12:22] <iain> :)
[12:23] <iain> aaaafrickkkaaans
[12:24] scureboy ([email protected]) left #gstreamer.
[12:25] <mathrick> hmm, what lib is needed for AAC plugin? faad / faac?
[12:26] <LeRoutier> faad for decoding/ faac for encoding
[12:26] <LeRoutier> or a52dec for decoding too
[12:26] <mathrick> k, and is AAC == MP4?
[12:29] <LeRoutier> really don't know. from what emerge search faad faac says, yes
[12:30] wheels ([email protected]) left irc: "using sirc version 2.211+KSIRC/1.3.9"
[12:30] <LeRoutier> i'm not sure for a52dec. might be AC3
[12:30] <mathrick> LeRoutier: looks like A/52 != AAC, at least from plugins desc
[12:32] Nick change: LeRoutier -> LeRaway
[12:34] Miko5881 ([email protected]) left irc: "ChatZilla 0.8.31 [Mozilla rv:1.4/20030928]"
[12:42] sublett ([email protected]) joined #gstreamer.
[12:44] thaytan ([email protected]) joined #gstreamer.
[12:50] <iain> yey scott defended my point on "the dot"
[12:55] <thomasvs> ds: ping ?
[12:58] <iain> oh planet gstreamer could take feeds from the GStreamer applications too
[12:58] <iain> not just the developers
[12:58] <dolphy> yeah
[12:58] <dolphy> iain: are you really working on that ?
[12:58] <iain> no
[12:59] <iain> just had a thought
[12:59] <iain> but itd be cool
[12:59] <thomasvs> iain: go ahead and do it
[13:01] <iain> once again, I've got no time :)
[13:02] <thomasvs> "different priorities"
[13:02] <thomasvs> sounds beter
[13:02] <dolphy> well so let's forget about that idea then :)
[13:02] <dolphy> nobody can work on that now :)
[13:04] Nick change: LeRaway -> LeRoutier
[13:07] <iain> yeah...my list of things to do: Exams, tidy room, marlin, file selector, take photos of things down at harbour, reinstall 2 computers, move DSL over to something better
[13:07] <iain> play THUG
[13:07] <iain> so I've got lots of other priorities :)
[13:09] jdahlin_ ([email protected]) joined #gstreamer.
[13:12] jdahlin ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[13:15] Safari_Al ([email protected]) joined #gstreamer.
[13:17] BBB ([email protected]) joined #gstreamer.
[13:20] <iain> kdedevelopers.net blogs hardly ever get updated, and when they do, they are soooo terminally dull
[13:20] Action: Safari_Al looks at that url.
[13:21] <Safari_Al> It seems the KDE lads are selling steak knives.
[13:21] <Safari_Al> Also 6 second abs.
[13:21] <iain> .org sorry
[13:21] <Safari_Al> maybe this site is more interesting :)
[13:21] <iain> it sounds it
[13:21] <Safari_Al> it has the lowest prices online!
[13:22] <Safari_Al> iain, how does your marlin compare to something like sweep?
[13:22] <iain> my abs works in milliseconds
[13:22] <Safari_Al> wow.
[13:22] <iain> 6 seconds for abs (-2) seems really slow
[13:22] <Safari_Al> well, it is KDE
[13:22] <iain> ah yeah, c++ linking issues slow it down a bit
[13:23] <Safari_Al> that's the unfortunate reality.
[13:23] <iain> umm, marlin is gtk2 so looks nicer, and it loads lots more formats (thanks to gstreamer)
[13:24] <iain> it can load large files cos it uses disk space rather than memory to store it...
[13:24] <Safari_Al> I ask because I went to an audio day at linux.conf.au yesterday, they mentioned sweep (ugly looking, btw) but no one mentioned marlin
[13:24] <iain> but basically I think Marlin will end up being a better application
[13:24] <Safari_Al> cool!
[13:24] <iain> but sweep is the more advanced at the moment
[13:24] <Safari_Al> Also, your roadmap is by far the better.
[13:24] <iain> yeah, I'm (unofficially) endorsed by a norwegian pop singer :)
[13:24] <iain> for various meanings of endorsed
[13:25] <dolphy> BBB: yo
[13:25] <Safari_Al> that's what really makes a good sample editor in these times
[13:25] <dolphy> BBB: we really need something global in interfaces
[13:25] <dolphy> BBB: a kind of flag specifying if they are hardware or software
[13:25] <mathrick> hi BBB
[13:26] <iain> Safari_Al: well, her songs probably need lots of sample editting so its only far
[13:26] <LeRoutier> hum, can no more play wav files with gst-player
[13:26] <iain> Safari_Al: remove those little off notes, loop the drums
[13:26] <mathrick> BBB: update on matroska playback - it plays now, although still spews lotsa criticals
[13:26] <dolphy> BBB: i implemented colorbalance in xvimagesink and i will need to make the same in gst/videofilter/gstvideobalance.c
[13:26] <dolphy> BBB: but i will need to know when i search for elements implementing interface which one is hardware/software to make a choice
[13:27] <Safari_Al> iain, well, they say that behind every success is 1% inspiration, 99% perspiration^H sample editing
[13:27] <BBB> dolphy: add it to the class/interface
[13:27] <BBB> dolphy: instead of the instance
[13:28] <BBB> that's fine
[13:28] <BBB> I do that in profiles, too
[13:28] <BBB> the profile data is stored in the class/interface
[13:28] <BBB> (it isn't instance-specific)
[13:29] <iain> Safari_Al: well, hopefully marlin can be at least some of that editing, eventually...
[13:29] <dolphy> you mean putting the flag in the GstColorBalanceClass structure
[13:29] <BBB> yes
[13:29] <Safari_Al> excellent.
[13:30] <thomasvs> BBB: can you update core, cd po, make install, then export LANG=nl_NL and see if you have some translations done ?
[13:30] Action: mathrick notices using queue now takes enormous amount of mem
[13:32] <dolphy> BBB: we should do that for mixer too
[13:32] <iain> yey...marlin loads audio from mpegs again :D
[13:32] <iain> beat that sweep :)
[13:33] <Safari_Al> iain, can it load from mods?
[13:33] <iain> should be able to if GStreamer can load mods
[13:33] Action: iain doesn't know
[13:34] Action: iain goes looking for a mod to test
[13:34] <Safari_Al> gst supposedly can do it, but I've never had success...
[13:35] <iain>  GStreamer-CRITICAL **: file gstpad.c: line 1306 (gst_pad_try_set_caps): assertion `!GST_FLAG_IS_SET (pad, GST_PAD_NEGOTIATING)' failed
[13:35] <thomasvs> yeah, mods work
[13:35] <iain> hmm, that happens
[13:35] <iain> and then it just sits there
[13:35] <thomasvs> might be broken due to caps nego though
[13:36] Action: iain tries a divx just for a laugh
[13:37] <BBB> dolphy: indeed
[13:37] <BBB> dolphy: feel free to add it...
[13:37] <iain> "Loading /home/iain/Media/Video/Movies/Orange County.avi: (10%) Estimated time remaining 2 minutes 9 seconds"
[13:37] <iain> kick ass
[13:38] <BBB> avi works here...
[13:38] <iain> here too
[13:38] <BBB> :)
[13:38] <LeRoutier> hum, lucky boys
[13:38] <BBB> I'll commit MPEG seeking fixes in a second
[13:38] <BBB> or, rather, add to bugzilla
[13:38] <BBB> it needs some review by experts
[13:38] <iain> hmm, I think we're getting out of the trough that GStreamer was in for a while
[13:38] <BBB> thomasvs: what's autopoint?
[13:38] <BBB> iain: we'd better ;)
[13:39] <iain> BBB: just in time really
[13:39] <BBB> I know
[13:39] <iain> mmmmmmm jeff buckly covering van morrison
[13:39] <BBB> and we're far from there
[13:40] <iain> we're getting there though, slowly
[13:40] <iain> which is good
[13:40] <iain> when marlin starts working again that means things are getting better
[13:41] <BBB> thomasvs: and where do I get 0.12?
[13:41] <iain> ** (lt-marlin:29794): CRITICAL **: file bytestream.c: line 256 (gst_bytestream_peek): assertion `len > 0' failed
[13:41] <iain> hmm, a few of those appeared at the end of decoding the avi
[13:42] <iain> but it doesn't seem to have harmed it too much
[13:42] <BBB> I might have to remove the assertions and make them GST_WARNING() instead
[13:42] <BBB> I'm quite picky on the standards
[13:42] <BBB> and files aren't
[13:42] <BBB> matroska has the same
[13:43] <BBB> but if it plays, I'm happy
[13:43] Action: BBB will fix MPEG now, and will then work on some bugs
[13:43] <dolphy> BBB: we should maybe have a global enum for interface type no? 
[13:43] <BBB> interface type?
[13:43] <BBB> :?
[13:43] <iain> hmmm, it loads, but Marlin seems to think that Orange County is 8hours long
[13:43] <dolphy> BBB: GST_INTERFACE_HARDWARE
[13:43] <BBB> how do you mean?
[13:43] <BBB> no...
[13:43] <BBB> just per interface where you want it
[13:43] <BBB> it doesn't make sense for most interfaces anyway
[13:43] <dolphy> BBB: ok so enum { } GstColorBalanceType
[13:43] <iain> oh cos it thinks the sample rate is 4000hz
[13:44] <dolphy> BBB: hard/soft
[13:44] <BBB> it makes sense - by accident - for colorbalance, tuner and mixer
[13:44] <iain> have to investigate that sometime
[13:44] <dolphy> ok
[13:44] <thomasvs> BBB: it's in gettext
[13:44] <thomasvs> BBB: you don't have gettext ?
[13:44] <BBB> where do I get that?
[13:44] <BBB> yes
[13:44] <BBB> but RH9 has 0.11
[13:44] <BBB> not 0.12
[13:44] <thomasvs> aargh
[13:44] <thomasvs> stupid rh
[13:44] <BBB> no, stupid you
[13:44] <thomasvs> when are you going to upgrade ? :)
[13:44] <BBB> ;)
[13:44] <BBB> well, do you have packages please?
[13:44] <BBB> ;)
[13:44] <thomasvs> get the fedora core 1 rpm
[13:45] <BBB> ok, on fedora.redhat.com?
[13:45] <thomasvs> should be fine for something like gettext
[13:45] <thomasvs> download.fedora.us
[13:45] <thomasvs> or ayo.freshrpms.net
[13:45] <thomasvs> your pick
[13:45] <iain> awww it doesn't load m4a's yet :/
[13:45] <Safari_Al> what in the name of glod are they?
[13:45] <BBB> iain: I know, I have to fix autoplugging for that
[13:45] <BBB> I have a lot of work left :/
[13:48] <iain> oh, ok
[13:49] <iain> Safari_Al: ipods file format
[13:49] <BBB> djeeh, fedora is slow
[13:49] <iain> ipod/itunes
[13:49] <BBB> (MPEG-4 AAC)
[13:50] <BBB> has anyone ever played with PROT_WRITE in filesrc?
[13:51] <BBB> I have issues with seeking if I enable PROT_WRITE
[13:58] thaytan ([email protected]) left irc: Read error: 60 (Operation timed out)
[14:02] <thomasvs> what does that do ?
[14:02] <BBB> autopoint: File mkinstalldirs has been locally modified.
[14:02] <BBB> autopoint: *** Some files have been locally modified. Not overwriting them because --force has not been specified.
[14:02] <BBB> autopoint: *** Stop.
[14:02] <BBB> PROT_WRITE gives write access to the buffers
[14:03] <BBB> I'm not sure if that's a good idea, but ohwell...
[14:03] Action: BBB uses another hack now
[14:03] <BBB> ...
[14:03] <BBB> O-)
[14:03] <BBB> it's local, so not for in CVS
[14:03] <BBB> anyway... what do I do about the above?
[14:04] <dolphy> is it possible that an element is called in its chain method with a null data ?
[14:04] <BBB> no
[14:04] <dolphy> weird
[14:04] <LeRoutier> BBB: seems it happens
[14:04] <BBB> hm...
[14:04] <BBB> that's a bug then
[14:04] <dolphy> when playing a wav with gst-plugins/examples/gstplay/player
[14:04] <dolphy> xvimagesink receives some data buffers
[14:05] <LeRoutier> same with gst-player
[14:05] <dolphy> being null
[14:05] <LeRoutier> perhaps not related but before the crash, i got : GStreamer-CRITICAL **: Could not find a compatible pad on element wavparse0 to link to src_1:sink
[14:06] <BBB> it could be a bug
[14:06] <BBB> in any case, report in bugzilla
[14:06] <BBB> I hope to find some time to debug it today
[14:06] <LeRoutier> ok
[14:06] <dolphy> hmm that's not possible
[14:06] <dolphy> g_return_if_fail (data != NULL);
[14:07] <dolphy> i m checking that anyway in xvimagesink
[14:07] <dolphy> "never trust gdb"
[14:11] <thomasvs> BBB: mkinstalldirs locally modified ?
[14:11] <thomasvs> how the hell did you do that :)
[14:12] <LeRoutier> http://bugzilla.gnome.org/show_bug.cgi?id=131331
[14:16] <BBB> thomasvs: no... but I got that error
[14:16] <BBB> thomasvs: maybe because my gettext version is incompatible to my autoconf/automake version?
[14:16] <BBB> LeRoutier: checking
[14:17] <BBB> oh, PS, thomasvs:  -> +//#define _(String) gettext (String)never use '//' style comments
[14:17] <LeRoutier> oh, i've got this crash only under gstplay/player and gst-player, not when building a pipeline in gst-launch (even with spider in the pipeline)
[14:18] <BBB> I'l check those later
[14:18] <BBB> first trying to get pipelines bugfree
[14:18] <BBB> gst-player will come... tomorrow, not today
[14:18] <BBB> </evil>
[14:18] <dolphy> gstplay is a pipeline :)
[14:19] <dolphy> nothing less nothing more :)
[14:19] <BBB> true, true
[14:19] <BBB> bah
[14:19] <BBB> so is gst-record ;)
[14:19] <BBB> still not bugfree
[14:19] <thomasvs> BBB: where's that from ?
[14:19] <thomasvs> that's a mistake commit
[14:19] <LeRoutier> well, the evil part is that playing a wav file, it tries to initialize a video widget
[14:20] <dolphy> i guess that's spider
[14:20] <dolphy> maybe having a gst-launch pipeline
[14:20] Safari_Al ([email protected]) left irc: "Leaving"
[14:20] <BBB> thomasvs: your commit a few minutes ago
[14:20] <dolphy> with spider being connected to both a video sink and audio sink would trigger the same crash
[14:20] <thomasvs> BBB: can you explain to me how gnome-sound-recorder works ?
[14:20] <thomasvs> BBB: yeah, but what file ?
[14:20] <thomasvs> BBB: I was reading through it, and it seems it first records to wav then converts ?
[14:20] <BBB> --- gst-i18n-app.h      13 Jan 2004 11:29:59 -0000      1.1+++ gst-i18n-app.h      13 Jan 2004 12:46:47 -0000      1.2@@ -30,6 +30,7 @@ #include "gettext.h" /* included with gettext distribution and copied */  /* we want to use shorthand _() for translating and N_() for marking */+//#define _(String) gettext (String) #define _(String) gettext (String) #define N_(String) gettext_noop (String)
[14:20] <BBB> thomasvs: yes, exactly
[14:20] <BBB> that's how the old one did it, too
[14:21] <BBB> recording and saving is separated
[14:21] <BBB> bit ugly, but ohwell
[14:21] <thomasvs> BBB: do you think that's a good idea ? because I want to change it
[14:21] <BBB> it's not mine
[14:21] <thomasvs> right
[14:21] <BBB> I don't care much
[14:21] <thomasvs> I want to directly record to the file
[14:21] <BBB> gst-record does that
[14:21] <BBB> don't spend too much time on it
[14:21] <BBB> I'll replace it with gst-record in gnome-3 or so
[14:21] Company ([email protected]) joined #gstreamer.
[14:33] <mathrick> Company: don't you have any uni or anything that'd stop you from staying until 7:30 in the morning? ;)
[14:34] <Company> mathrick: yes, i have uni - but i just start later then...
[14:35] <mathrick> Company: hmm, you have much freedom in choosing your schedule then
[14:36] <mathrick> Company: how much is 'later'? :)
[14:37] <BBB> uni is fun
[14:37] <Company> mathrick: 2 hours today
[14:37] <BBB> I'm a student and I work fulltime at a media company
[14:37] <BBB> that's odd
[14:37] <mathrick> BBB: yes, that's odd. How do you manage to get through your finals?
[14:38] Action: mathrick is having hard time every semester end
[14:40] <thomasvs> [gst-cvs] [thomas@otto po]$ rm -f nl.gmo && /usr/bin/msgfmt -c --statistics -o nl.gmo nl.po
[14:40] <thomasvs> 99 translated messages.
[14:40] <thomasvs> whee
[14:40] Action: Company just ignores semester ends and studies and studies and studies...
[14:41] <thomasvs> Company: does mad currently report tags for bitrate and so on ?
[14:41] <Company> thomasvs: no, but it should
[14:42] <Company> thomasvs: that requires some detection code for vbr though
[14:42] <thomasvs> Company: I'm guessing it's a simple copy from the id3tag code, using the old streaminfo signaller ?
[14:43] <Company> there is no streaminfo signaller anymore...
[14:43] <Company> and id3tag doesn't detect bitrates, because bitrates require looking at audio data
[14:43] <thomasvs> no, it's ifdeffed out, but I mean, I can just use that old code to check how to implement it
[14:43] <Company> yeah
[14:43] Action: thomasvs now tries to i18n-ize grammar.y
[14:43] <Company> most likely
[14:44] <Company> shouldn't be that hard
[14:44] <Company> (I hope)
[14:46] <Company> thomasvs: how are you going to change gst_element_error ?
[14:47] <thomasvs> +static void                    gst_element_error_func_2        (GstElement* element, GstElement *source, GError *error, gchar *debug);
[14:47] <thomasvs> and
[14:47] <thomasvs> +void
[14:47] <thomasvs> +gst_element_error_2
[14:47] <thomasvs> +(GstElement *element, GQuark domain, gint code, gchar *message, gchar *debug)
[14:48] <thomasvs> and then a macro that I still need to write, but should be simple
[14:48] matrixise ([email protected]) joined #gstreamer.
[14:50] <Company> hm
[14:50] <Company> why not (GstElement *element, GQuark domain, gint code, gchar *format_message, gchar *format_debug, ...)
[14:51] <Company> and make both format strings take arguments in the same order?
[14:51] <thomasvs> not sure I follow
[14:51] <thomasvs> you mean, first specify the two formats, then specify all args for format 1, then for format 2 ?
[14:51] <thomasvs> btw, is it useful at all to translate the parsing strings ?
[14:51] <Company> no
[14:52] <Company> use the same args for format 1 and format 2
[14:52] <Company> translate parsing strings?
[14:53] <thomasvs> yeah, the parsing code in grammar.y - stuff like "could not link element ... to ..."
[14:53] <thomasvs> Company: why would I want to use the same args for both formats ? they're different
[14:54] <Company> thomasvs: most likely the translated message is a subset of the debug message
[14:54] <thomasvs> Company: no, they are completely separate
[14:54] <Company> thomasvs: then you can format the debug message to suit the order of the translated message and add some arguments
[14:54] <thomasvs> the debug message is in addition to the error message
[14:54] <Company> thomasvs: give me an example where they are
[14:55] <thomasvs>     gst_element_error_2 (GST_ELEMENT (src),
[14:55] <thomasvs> +                         GST_RESOURCE_ERROR,
[14:55] <thomasvs> +                         GST_RESOURCE_ERROR_NOT_FOUND,
[14:55] <thomasvs> +                        g_strdup_printf ("File \"%s\" does not exist",
[14:55] <thomasvs> +                                          src->filename),
[14:55] <thomasvs> +                         NULL);
[14:56] <thomasvs> hm, hang on, that has NULL so that doesn't help
[14:56] <thomasvs> hm, it's even wrong, so that needs fixing :)
[14:56] <thomasvs>    gst_element_error_2 (GST_ELEMENT (src),
[14:56] <thomasvs> +                           GST_RESOURCE_ERROR,
[14:56] <thomasvs> +                           GST_RESOURCE_ERROR_OPEN_READ,
[14:56] <thomasvs> +                           g_strdup_printf ("Could not open file \"%s\" for reading", src->filename),
[14:56] <thomasvs> +                           g_strdup_printf ("error message: %s", strerror (errno)));
[14:56] <thomasvs> this one is better
[14:56] <mathrick> thomasvs: why is NULL wrong?
[14:57] <thomasvs> Company: (on a side note, should I or shouldn't I translate parsing errors ?
[14:57] <thomasvs> mathrick: it's not wrong, but Company asked for an example with both message and debug
[14:57] <mathrick> ah, k
[14:57] <Company> gst_element_error_2(GST_ELEMENT (src),GST_RESOURCE_ERROR, GST_RESOURCE_ERROR_OPEN_READ,_("Could not open file \"%s\" for reading"), "file %s could not be opened for reading: %s", src->filename, strerror(errno))
[14:58] <Company> thomasvs: i guess you should translate them - they're shown to the user in at least gst-launch
[14:58] <mathrick> Company: one word: i18n
[14:58] <Company> thomasvs: and probably should be in gst-gconf, too
[14:58] <thomasvs> Company: so I thought of that too, but I'm not sure it is better
[14:58] <Company> mathrick: hu?
[14:59] <Company> thomasvs: it get's rid of that stupid strdup_printf, that's a huge advantage
[14:59] <mathrick> Company: remember, that what is just superset of sentence in english, may require totally different structure in other langs
[14:59] <thomasvs> Company: the macro would be GST_ELEMENT_ERROR (src, GST_RESOURCE, OPEN_READ, (_("Could not open ..."), filename)), ("file could not be opened: %s", error));
[15:00] <thomasvs> Company: hm, I guess that would still make it possible to have it your way too
[15:00] <thomasvs> Company: it is confusing though
[15:00] <mathrick> Company: so forcing them to take the same params is risky
[15:00] teuf ([email protected]) joined #gstreamer.
[15:00] <thomasvs> Company: yeah, the params need to be separate
[15:00] Nick change: jdahlin_ -> jdahlin
[15:00] <mathrick> Company: and, what is the gain from doing it your way?
[15:01] <thomasvs> Company: when parsing something like ("format1", "format2", "arg1a", "arg1b", "arg2a") ...
[15:01] <Company> mathrick: the API is clean
[15:01] <thomasvs> ... can it figure out where arg2a starts ?
[15:01] <Company> thomasvs: no
[15:01] <mathrick> Company: i don't think so
[15:01] <Company> mathrick: cleaner than using (thingy1), (thingy2)
[15:01] <thomasvs> Company: so your trick only works when the args in the debug are a superset of the args in the message ?
[15:01] <Company> thomasvs: yes
[15:01] sublett ([email protected]) left irc: "I like food, food is good!"
[15:02] <Company> thomasvs: i don't see why that shouldn't be enforced anyway btw
[15:02] <mathrick> Company: for me, if you pass msg1 and msg2, specifying it as (msg1), (msg2) is much more straightforward
[15:02] <thomasvs> Company: I gave you an example where they were different, and you "fixed" it by adding an argument that's redundant :)
[15:03] <Company> thomasvs: i don't consider it redundant
[15:03] <thomasvs> Company: I'm sure there are lots of cases where there's no point at all in repeating arguments in the debug message, since they're already in the error
[15:03] <Company> thomasvs: that forces you to always output the error besides the debug message though
[15:03] <thomasvs> Company: on top, I don't see a problem at all with the strdup, since it's for error messages, so it doesn't happen often
[15:03] <thomasvs> Company: why would you ever show the debug message without hsowing the error ?
[15:04] <thomasvs> Company: the writeup we discussed never talked about wanting to do it
[15:04] <Company> thomasvs: GST_ERROR_OBJECT (error, debugmessage) for example
[15:04] <Company> thomasvs: or users filing bug reports and only pasting the "extended" message
[15:05] <thomasvs> Company: well, we can't account for EVERYTHING a user does to be perfect :) the intent is clean and clear, so, that's just as valid as saying "what if a user files a bug report and doesn't attach the debug message at all"
[15:06] <Company> thomasvs: well, the debug message is quite useless without the _untranslated_ error message, too
[15:06] <thomasvs> We'll figure it out soon enough - but for now I'm sticking with what I think is correct based on the previous discussions all of us had.
[15:07] <thomasvs> Company: I understand why you would want to work around the g_strdup_printf
[15:07] <thomasvs> Company: but I have less of a problem with it, and I'm not sure we need to invent reasons to validate working around the g_strdup_printf just because we don't like it :)
[15:07] <thomasvs> BBB: hoe vertaal ik "bin" ? ton ?
[15:08] <Company> i think it's good to enforce the debug message being a superset of the error message and not an addition, too
[15:09] <mathrick> Company: IMO they are complementary, not related hierarchically
[15:09] <thomasvs> Since the error message can be from the enum when specified as NULL, I don't really see that work
[15:09] <thomasvs> But given the discussions on email, it's clear to me they need to be decoupled
[15:10] <thomasvs> Company: anyway, feel free to do a proposal based on the patch I'm going to do, I can always be convinced otherwise
[15:10] <Company> thomasvs: ok, will discuss this on the list then
[15:11] <Company> mathrick: the problem is that i as a developer have to use the debug message only, because i most likely don't understand the error message
[15:11] <Company> mathrick: "udlanik czar ampramoch" would tell me nothing, even if it just meant "could not allocate memory"
[15:12] <mathrick> Company: hmm, this is a problem indeed, but this also narrows possible sources of debug info
[15:12] <BBB> thomasvs: 'container'?
[15:12] <BBB> 'opslag'
[15:12] <BBB> oid
[15:12] <mathrick> Company: b/c it can't be chosen freely, but instead has to be similar to error msg
[15:13] <Company> mathrick: that's most likely the case anyway, because both messages relate to the same error...
[15:14] <Company> mathrick: but if you can think of a case where some argument should be part of the error message and should not be part of the debug message, please tell me :)
[15:14] Action: mathrick thinks hard ;)
[15:16] <Company> mathrick: btw, I was advocating the approach thomasvs is currently taking and even implemented it, but after thinking more about it, I came to the conclusion that I like more what I advocate now
[15:16] <thomasvs> Company: one possibility might be somehow forcing the nontranslated error message as an append to the debug one
[15:16] <thomasvs> which shouldn't be too hard actually.
[15:16] <thomasvs> since I have to merge in source location too anyway
[15:17] <Company> if you do that
[15:18] <thomasvs> [gst-cvs] [thomas@otto po]$ ../tools/gst-launch  fakesrc ! fakes
[15:18] <thomasvs> WAARSCHUWING: foutieve pijplijn: geen element "fakes"
[15:18] <thomasvs>               Probeer toch uit te voeren.
[15:18] <thomasvs> BEZIG met pijplijn ...
[15:18] <thomasvs> FOUT: pijplijn wil niet spelen.
[15:18] <thomasvs> [gst-cvs] [thomas@otto po]$
[15:18] <Company> you could just do real_debug_format = stdup_printf ("%s - %s", message_format, debug_format) and then append the args
[15:18] Action: thomasvs really doesn't like translated apps anyway
[15:19] <BBB> pijplijn?
[15:19] <Company> "pijplijn wil niet spelen." sounds a lot like an angry kid - "Charlie doesn't wanna play!"
[15:19] <BBB> 'kan' niet spelen, I think
[15:20] <jdahlin> thomasvs: translated apps rocks
[15:20] <Company> no they don't - google doesn't support searching for error messages in the original format yet
[15:21] <Company> when google has the po files and translates error messages back, then it's a bit closer to rocking
[15:22] <thomasvs> jdahlin: the words sound stupid in the native language
[15:22] Nick change: trow_away -> trow
[15:22] <Company> and it's ugly when you have "speelen", "pause" and "stop"
[15:22] <thomasvs> jdahlin: it is helpful though for learning spanish, in a way :)
[15:23] <jdahlin> thomasvs: just think about the ones who can't understand english and that you're helping them to use a computer
[15:23] <mathrick> but it's undeniable, that *many* users won't ever touch untraslated app
[15:23] ChrisHJW ([email protected]) joined #gstreamer.
[15:23] <mathrick> s/tras/trans/
[15:23] <thomasvs> jdahlin: yeah, I know
[15:23] <BBB> jdahlin: well... the thing is that for us, developers, it's a bit odd...
[15:23] <BBB> but ohwell, if it makes others happy
[15:23] <thomasvs> jdahlin: I'm saying, it sucks when you *do* know english
[15:24] <jdahlin> BBB, I know, most developers like in english
[15:24] <thomasvs> jdahlin: I'm not advocating removing it
[15:24] <BBB> "pijplijn wil niet spelen"
[15:24] <BBB> :X
[15:24] <thomasvs> BBB: yeah, sounds stupid no ?
[15:24] <Company> it's ok to do translations, but they're just a help for non-native speakers, not something that "rocks" :)
[15:24] <jdahlin> thomasvs: No, I prefer translated apps every single day
[15:24] <BBB> "pijplijn kan niet spelen"
[15:24] <BBB> sounds better
[15:24] scureboy ([email protected]) joined #gstreamer.
[15:24] <jdahlin> thomasvs: it's more "professional" and bla bla
[15:24] <BBB> but "pijplijn" is still stupid
[15:24] <jdahlin> agreed.
[15:24] <thomasvs> BBB: I don't care atm really, I just want to get the framework in place
[15:25] <thomasvs> BBB: ok, so I'll test on rh9 now for you
[15:25] <BBB> hey, I'll talk dutch to you whenever I come to Spain, evne if it's only so you don't forget it ;)
[15:25] <Company> BBB: what would you call it in dutch?
[15:25] scureboy ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[15:25] <BBB> dunno
[15:26] <Company> BBB: i think it's the correct term - though i dunno why you translate it, it's pipeline in german, too :)
[15:26] <thomasvs> BBB: hehe :)
[15:26] <thomasvs> Company: really ? pipeline ?
[15:26] <Company> "peipleine" would be ugly...
[15:26] <thomasvs> and bin ?
[15:26] <Company> yeah
[15:26] <thomasvs> wait, I don't follow
[15:26] <thomasvs> the right translation is peipleine or pipeline ?
[15:27] <Company> "pipeline"
[15:27] <BBB> bin is 'container' or so, imo
[15:27] <BBB> and why would you translate these terms?
[15:27] <BBB> in dutch, they really don't make any more sense than in english
[15:27] <Company> leo says "Behälter", which sounds ok to me
[15:28] <Company> (leo is dict.leo.org)
[15:28] <thomasvs> BBB: that's the question really. what sort of things do you translate and what sort of things don't you
[15:28] <thomasvs> BBB: ok, so 0.11.4 has autopoint too, so I updated autogen.sh, please try again
[15:29] Action: thomasvs takes a pee break
[15:29] <Company> interesting question is: how to translate "pad"? :)
[15:33] <BBB> thomasvs: I already installed 0.12
[15:36] <Company> can we finally up our libtool requirement to 1.5 btw? (or did we already do this?
[15:36] <BBB> yes please
[15:37] gaijin025 ([email protected]) joined #gstreamer.
[15:38] gaijin025 ([email protected]) left #gstreamer.
[15:38] scureboy ([email protected]) joined #gstreamer.
[15:38] gaijin025 ([email protected]) joined #gstreamer.
[15:43] <gaijin025> hi there
[15:43] <gaijin025> is it true that ligos mpeg encoder can go as fast as average 8ms per frame??
[15:43] <thomasvs> no idea
[15:43] <thomasvs> what's ligos ?
[15:44] <gaijin025> well, a commercial mpeg software encoder.. i think the name was ligos..
[15:44] <gaijin025> http://www.ligos.com
[15:48] matrixise ([email protected]) left irc: "Leaving"
[15:52] gaijin025 ([email protected]) left #gstreamer.
[15:53] ChrisHJW ([email protected]) left irc: Connection timed out
[16:00] trow ([email protected]) left irc: "Leaving"
[16:02] <iain> thomasvs: + running autopoint ...
[16:02] <iain> cvs [checkout aborted]: no such tag gettext-0_12
[16:02] <iain> /usr/bin/autopoint: line 259: fatal_error: command not found
[16:02] <iain> where do I get autopoint?
[16:02] <thomasvs> iain: I fixed that
[16:02] <thomasvs> iain: needs an update to configure.ac
[16:02] <thomasvs> update and try again
[16:02] <thomasvs> iain: you already have autopoint btw, but it does some funky cvs handling
[16:06] <thomasvs> iain: crap, sorry, commit broke
[16:06] <thomasvs> iain: give me a minute
[16:07] <iain> + running autopoint ...
[16:07] <iain> autopoint: File mkinstalldirs has been locally modified.
[16:07] <iain> autopoint: *** Some files have been locally modified. Not overwriting them because --force has not been specified.
[16:07] <iain> autopoint: *** Stop.
[16:07] <iain>  
[16:07] <iain> autopoint failed
[16:07] <thomasvs> yeah I know, fix is coming for that too
[16:07] <iain> oh no, its not the same crap thing as gettextize does is it?
[16:08] <iain> can we not use glib-gettext stuff?
[16:09] <thomasvs> it's not the same crap, don't worry
[16:09] <thomasvs> in fact, it's 100% pure gettext
[16:09] <thomasvs> doesn't use gettextize
[16:09] <iain> we're not going to end with intl/ dirs are we?
[16:10] <iain> and was gettextize not 100% pure gettext as well?
[16:10] <thomasvs> iain: SETTLE DOWN
[16:10] <thomasvs> iain: TAKE SOME PILLS
[16:10] <thomasvs> iain: WAIT A MINUTE
[16:10] <iain> I'm not worried, I'm just trying to understand what its doing :)
[16:11] <thomasvs> if you want to understand what it's doing, check the as-gettext tarball I mailed to the list
[16:11] <thomasvs> iain: autopoint is the autogen-friendly tool in gettext
[16:11] <iain> ok
[16:11] <thomasvs> gettextize is the wizard that touches everything
[16:11] <thomasvs> autopoint is the nice well-behaving counterpart
[16:11] <iain> I see
[16:11] <thomasvs> iain: I have as little love for gettextize and glib-gettextize as you, be sure
[16:11] <Company> good cop / bad cop
[16:12] <thomasvs> iain: ok, now the fixes are in
[16:12] <thomasvs> iain: please let me know if after updating it has a problem
[16:13] <iain> ok
[16:13] scureboy ([email protected]) left #gstreamer.
[16:13] <Company> i think is should write a speedclock and make pipelines work with that
[16:13] <Company> at least that would allow for some bufixing
[16:16] Company ([email protected]) left irc: Remote closed the connection
[16:18] <thomasvs> ugh - GST_ELEMENT_ERROR is used in an enum, that sucks
[16:19] matrixise ([email protected]) joined #gstreamer.
[16:21] <dolphy> lol
[16:21] <dolphy> i m really not the good person to translate gstreamer to french with my USA keyboard :)
[16:23] <teuf> will this gstreamer i18n stuff add many new strings to translate ?
[16:24] <thomasvs> teuf: not many, but quite some yeah
[16:25] <thomasvs> teuf: why ?
[16:25] <teuf> thomasvs: I'm wondering how much translators will hate you ;)
[16:25] <thomasvs> teuf: hehe :)
[16:25] <thomasvs> teuf: we need to find them first anyway
[16:25] <dolphy> yeah nobody will get angry atm :)
[16:26] <teuf> find them ? just mail gnome-i18n about it
[16:26] <thomasvs> teuf: well, IIRC gnome doesn't want to do projects not hosted by gnome anymore
[16:26] <dolphy> gstreamer is not on gnome cvs 
[16:27] <teuf> oh yeah, I forgot that ;)
[16:30] <thomasvs> anyone have a guide/tips on how to use ## in #define's ?
[16:32] <iain> thomasvs: okay, it works
[16:32] <BBB> thomasvs: for what?
[16:33] <BBB> ## stringifies the given parameter
[16:35] <thomasvs> BBB: #define GST_ELEMENT_ERROR_2(el, domain, code, message, debug) gst_element_error(GST_ELEMENT(el), GST_ ## domain ## _ERROR, GST_ ## domain ## _ERROR_ ## code, message, debug)
[16:35] <thomasvs> BBB: is the second argument legal ?
[16:36] <iain> thomasvs: it takes what is passed into domain
[16:36] <iain> and makes GST_WHATEVER_ERROR
[16:36] <thomasvs> hm, so something is wrong then, because it doesn't work
[16:36] alley_cat ([email protected]) left irc: Read error: 60 (Operation timed out)
[16:36] Action: thomasvs writes simple C test
[16:41] mxpxpod ([email protected]) left irc: "Guns kill people as much as spoons made Rosie O'Donnell fat.."
[16:45] <thomasvs> damn, I'm a dope
[16:46] <BBB> thomasvs: yes, it's legal
[16:46] <BBB> it concatenates the strings together to a new string representation
[16:46] <BBB> somehwat icky, but useful
[16:47] Nick change: mathrick -> mathrick|nap
[16:47] alley_cat ([email protected]) joined #gstreamer.
[16:48] <thomasvs> BBB: yeah, my mistake was to be using element_error instead of my own element_error_2
[16:49] <BBB> ok :)
[17:01] ChrisHJW_log ([email protected]) joined #gstreamer.
[17:01] <LeRoutier> bbl
[17:01] LeRoutier ([email protected]) left irc: "Leaving"
[17:06] <thomasvs> is there a way to see preprocessor expansion ?
[17:08] <teuf> thomasvs: gcc -E ?
[17:09] <thomasvs> cpp works too
[17:09] ChrisHJW ([email protected]) joined #gstreamer.
[17:09] <thomasvs> YES
[17:09] Action: thomasvs nailed the macro
[17:10] Action: thomasvs is afraid of the compilers this won't work on
[17:10] <BBB> oh god why do I hate MPEG so much
[17:10] <BBB> rather
[17:10] <BBB> why do I hate this stupid WMP so much
[17:12] <thomasvs> BBB: is this example fine for you:
[17:12] <thomasvs>       GST_ELEMENT_ERROR_2 (src, RESOURCE, OPEN_READ,
[17:12] <thomasvs>                            (_("Could not open file \"%s\" for reading"), src->filename),
[17:12] <thomasvs>                            ("error message: %s", strerror (errno)));
[17:12] <BBB> I was thinking lately... why don't we just stick to one error message?
[17:12] <BBB> and if we want a second one, let's just use GST_WARNING() right below it
[17:13] <BBB> just an idea
[17:13] <taaz> btw, have you all every used the diff -p option?  it's pretty nice.  we should add that to a suggested way to submit patches
[17:13] <taaz> s/every/ever/
[17:13] <BBB> what's that?
[17:14] <taaz> guesses the c function
[17:14] <thomasvs> BBB: what do you mean, one error message ?
[17:14] <BBB> well, you're now defining two, right?
[17:15] <BBB> I'd say, let's just use one, translated to UI and untranslated to log
[17:15] matrixise ([email protected]) left irc: "Leaving"
[17:15] <thomasvs> BBB: no, one error message, translated, and additional debug info, so you guys can fix them
[17:15] <BBB> and if we need more info in the log, add a GST_WARNING()
[17:15] <thomasvs> huh ?
[17:15] <BBB> so the macro then contains one:
[17:15] <thomasvs> how are we going to get the debug info then ?
[17:15] <BBB> from the same function
[17:15] <BBB> look:
[17:16] <BBB> GST_ELEMENT_ERROR (bla, ("blabla %s bla", "bla"));
[17:16] <BBB> ->
[17:17] <BBB> #define GST_ELEMENT_ERROR ( gst_element_error (.., do something with _(second argument)); GST_ERROR (do somethign with second argument);)
[17:17] <BBB> then we have the same message translated and untranslated
[17:17] <thomasvs> I am completely lost as to what you mean, but as far as I can follow, this won't bring the untranslated error to the app
[17:17] <BBB> what I'm affraid of is that having to define two messages for each error will not help much
[17:17] <thomasvs> BBB: moreover, this doesn't allow us to give debug info
[17:17] <thomasvs> BBB: you don't have to, that's the point
[17:18] <thomasvs> BBB: you can get by fine with specifying NULL, NULL
[17:18] <thomasvs> and it will pull the "message" out of the code database, and the "debug" it will generate with an empty string, but with source code location info
[17:20] <BBB> but can't you do that with one string?
[17:20] <BBB> it's macro's
[17:21] <BBB> wht I want is something that will use the given value in the macro twice, once for translated error display to the app and once for debug purposes with line number et all
[17:21] <BBB> but just with one argument, not two
[17:22] <BBB> because I fear that nobody will really make a difference between the two, which means we'll basically have a small piece of bloat in our API
[17:23] <BBB> I'll try to write down... hold on
[17:23] <thomasvs> BBB: send me a mail, I'm sure you're discussing stuff that we didn't want to do anyway
[17:23] <thomasvs> BBB: you'll see in the patch
[17:28] <BBB> see privwindow ;)
[17:29] <taaz> try -p for the patch ;)
[17:29] ChrisHJW ([email protected]) left irc: Remote closed the connection
[17:30] ChrisHJW ([email protected]) joined #gstreamer.
[17:31] ChrisHJW ([email protected]) left irc: Remote closed the connection
[17:31] ChrisHJW ([email protected]) joined #gstreamer.
[17:39] Uraeus ([email protected]) joined #gstreamer.
[17:40] Nick change: Jara[zZ] -> Jaramir
[17:41] trow ([email protected]) joined #gstreamer.
[17:48] <Uraeus> hi trow
[17:48] <trow> hi Uraeus
[17:52] BBB ([email protected]) left irc: "Client exiting"
[17:53] <trow> gst_play_seek_to_time doesn't seem to work in HEAD... at least not when totem tries to use it.
[17:53] <trow> Is that something which is known to be broken?
[17:53] smoke ([email protected]) joined #gstreamer.
[17:54] <thomasvs> dolphy: trow is asking you something :)
[17:54] <thomasvs> trow: what is your audiosink set to ?
[17:54] Action: thomasvs puts his money on esound
[17:54] <trow> osssink
[17:54] <thomasvs> damn :)
[17:55] <trow> And xvimagesink for video.  The gst_element_seek calls both return TRUE, but nothing seems to happen.
[17:55] <dolphy> hmm
[17:55] <dolphy> trow i ll take a look at that
[17:55] <dolphy> tonight
[17:55] <trow> dolphy: thanks
[17:56] <dolphy> trow: have you seen the colorbalance added in totem  ?
[17:56] <trow> I saw your ChangeLog message, but I haven't tried it yet.
[17:56] Action: trow tries it...
[17:57] <trow> Hmm.  totem segfaulted while I was sliding the contrast slider.
[17:58] <thomasvs> trow: ah, please file a backtrace for that
[17:58] <thomasvs> trow: we need to fix that then
[17:59] <trow> X is crapping out with a BadValue.
[17:59] ChrisHJW ([email protected]) left irc: 
[18:00] Action: Jaramir is away: Backbone Scoliosis
[18:00] Action: Jaramir is back (gone 00:00:02)
[18:00] <thomasvs> trow: add an xvinfo output to the bug report
[18:00] <thomasvs> trow: maybe your xv doesn't support colorbalance
[18:01] <trow> thomasvs: Well, I can change the contrast somewhat, but it crashes if I move the slider past a certain point.
[18:01] <trow> So something is probably making a bad assumption about the range of a parameter.
[18:02] <dolphy> BadValue ?
[18:02] <dolphy> hmm
[18:02] <dolphy> yeah put your xvinfo
[18:06] <trow> Interesting.  The ColorBalanceChannel is returning min_value of -1000 and a max_value of 1000.  But it gets a BadValue for values < 0.
[18:08] Nick change: Jaramir -> Jara[zZ]
[18:11] <trow> Ok, I think I found the problem.
[18:11] <trow> In xvimagesink.c, the min_value and max_values are hardwired as -1000 and 1000.
[18:11] <trow> However, my xvinfo reports the min and max values for XV_CONTRAST as 0 and 8191.
[18:12] <trow> Of course, keep in mind that I don't really understand any of this. :)
[18:12] swentel ([email protected]) left irc: 
[18:13] somex1 ([email protected]) joined #gstreamer.
[18:15] <thomasvs> trow: hm, sounds like the cause of the bug to me :)
[18:16] <trow> The XvGetPortAttribute claims that XV_BRIGHTNESS and XV_CONTRAST are in the range [-1000,1000], so this wasn't a totally unreasonable thing for someone to do.
[18:16] Nick change: harshy -> harshyWork
[18:18] <thomasvs> trow: hm, sounds like a driver bug then ?
[18:18] <thomasvs> trow: what driver do you use ?
[18:19] <thomasvs> trow: for me, it reports them all correctly
[18:20] <trow> I've got an nvidia card, using the nv driver.
[18:20] <trow> There must be a way to query the attributes to get the min and max values.
[18:21] Action: trow wonders where to find the xvinfo source code
[18:21] <trow> OK, XvQueryPortAttributes.
[18:22] <trow> It always makes me nervous when I need to call a function that starts with a capital X...
[18:23] mx|school ([email protected]) joined #gstreamer.
[18:23] <mx|school> dolphy: have you gotten a chance to check out the bug report I keyed you in on last night? (this is mxpxpod)
[18:31] Company ([email protected]) joined #gstreamer.
[18:31] <dolphy> mx|school: nope
[18:31] <dolphy> mx|school: i m at work :)
[18:31] <mx|school> dolphy: ok, that's cool
[18:31] <mx|school> just checking up
[18:31] <mx|school> :)
[18:32] <mx|school> dolphy: do you have any notions as to what it could be?
[18:32] <dolphy> trow: ok then it's normal
[18:32] <dolphy> trow:  XV_CONTRAST as 0 and 8191.
[18:32] <dolphy> trow:  i ll fix that
[18:32] <dolphy> trow: nope
[18:32] <dolphy> err
[18:32] <dolphy> mx|school: nope
[18:33] <mx|school> k
[18:33] <trow> dolphy: I've fixed it in xvimagesink.
[18:33] <trow> dolphy: Who owns xvimagesink.  Should I send you the patch?
[18:33] <dolphy> trow: me
[18:33] ds-work ([email protected]) left irc: Read error: 110 (Connection timed out)
[18:34] <trow> dolphy: OK, I'll send you the patch for review in just a minute.
[18:35] <thomasvs> trow: the problem with that is this
[18:35] <dolphy> trow: great thx
[18:35] <thomasvs> trow: gst-register can be run without X active
[18:35] <thomasvs> trow: so we can't query for the property
[18:35] Nick change: mathrick|nap -> mathrick
[18:36] <dolphy> thomasvs: that's not a problem
[18:36] <dolphy> thomasvs: gst-register is not setting the elements to READY right ?
[18:36] <thomasvs> why not ? it seems the nvidia have different min and max, and you can only query it through XVQueryPortAttributes, no ?
[18:36] <thomasvs> oh, right
[18:37] <thomasvs> yep, I was talking about other things
[18:37] <thomasvs> so the interface properties are runtime ?
[18:38] <dolphy> absolutely
[18:38] <dolphy> so basically we just need to query the xv adapter to get the min/max
[18:38] <dolphy> i set them to -1000 + 1000
[18:38] <dolphy> because that's what the X doc was telling
[18:38] <dolphy> seems the doc wasn't telling the truth :)
[18:39] <trow> Never trust a man page. :)
[18:39] <dolphy> trow: you sent your patch ?
[18:40] <dolphy> got it
[18:40] <trow> dolphy: Yeah, I just mailed it a second ago.
[18:41] <dolphy> trow: any reason why you are not using glib strcmp  ?
[18:42] <trow> dolphy: Does glib provide a straight strcmp implementation these days?
[18:42] <trow> I thought it only provided string functions that weren't standard, like g_strcasecmp.
[18:43] Nick change: mx|school -> mxpxpod
[18:44] <trow> Yeah, there is no g_strcmp.
[18:44] <dolphy> trow: i usually use g_ascii_strcasecmp
[18:44] somex1 ([email protected]) left #gstreamer.
[18:44] matrixise ([email protected]) joined #gstreamer.
[18:44] <trow> dolphy: That works too.  I'll change it.
[18:45] teuf ([email protected]) left irc: "Client exiting"
[18:45] <trow> dolphy: Is there anything else you'd like me to change, or is it OK to commit?
[18:46] <thomasvs> dolphy,trow: sounds more like an nvidia driver bug to me
[18:47] somex1 ([email protected]) joined #gstreamer.
[18:47] <trow> thomasvs: Maybe it is a bug, but xv does give us a mechanism for querying attributes to find their allowed ranges, so we might as well use it...
[18:47] <thomasvs> trow: yeah
[18:47] <dolphy> yeah it s better to query
[18:48] <dolphy> trow: i ll commit that
[18:48] <dolphy> trow: in a few secs
[18:48] <trow> dolphy: OK, great.
[18:53] mxpxpod ([email protected]) left irc: "bbiab"
[18:57] <dolphy> trow: it's in. be sure to rebuild the interface lib in gst-plugins/gst-libs/gst too  as i added something to colorbalance changing the class structure size
[18:59] <trow> dolphy: OK, thanks.
[19:00] <dolphy> trow: thank you :)
[19:00] <dolphy> trow:  i ve a local version of gst-totem here with 4 sliders
[19:00] <dolphy> trow: for the complete color balance
[19:00] <dolphy> trow: i sent the patch to hadess
[19:01] <dolphy> trow: he needs to discuss some stuff with seth before we can go and merge it
[19:02] <trow> dolphy: It would also be nice if there was a "restore sensible defaults" button, since totem will remember whatever values you set.
[19:05] <dolphy> yeah
[19:05] <dolphy> good idea
[19:06] mxpxpod ([email protected]) joined #gstreamer.
[19:06] <trow> Maybe the GstColorBalanceChannel should store the initial value to use as a default.
[19:06] <dolphy> would you file a bug report for that button idea
[19:07] <dolphy> k going back home
[19:07] <dolphy> bbl
[19:07] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[19:14] <Company> weeeheee
[19:14] kmaraas ([email protected]) joined #gstreamer.
[19:15] <Company> we only have an oss clock and an alsa clock, right?
[19:15] <thomasvs> I think so, yes
[19:15] <thomasvs> Company: I'm trying to solve the issues by creating an E_() and a D_() macro
[19:16] <thomasvs> Company: blocked on something now though, a preprocessor error I don't follow
[19:16] <iain> oh bugger dolphy's gone :/
[19:16] <Company> preprocessor errors are best solved running the source through cpp :)
[19:16] <thomasvs> I know
[19:16] <thomasvs> but I don't see why it's complaining :)
[19:18] <thomasvs> oh, hang on
[19:18] <thomasvs>   printf ("D_(NULL): %s\n", (((void *)0) ? g_strdup_printf ("%s"((void *)0), "string") : g_strdup_printf ("FIRST WAS NULL")));
[19:19] <thomasvs> that doesn't work well, concatening those two
[19:19] <thomasvs> n/m, I'll stick with the current approach for now
[19:29] pb_ ([email protected]) joined #gstreamer.
[19:36] mxpxpod ([email protected]) left #gstreamer ("Client exiting").
[19:53] Nick change: thomasvs -> thomasvs-home
[20:13] thomasvs-home ([email protected]) left irc: No route to host
[20:15] ds-work ([email protected]) joined #gstreamer.
[20:15] Nick change: trow -> trow_lunch
[20:16] <ds-work> whazzap?
[20:17] Action: ds-work 's computer died overnight, so there's no backlog to read
[20:18] <Uraeus> ds-work: you where chosen as project leader and also chosen to fix all bugzilla entries during next 12 hours :)
[20:20] <ds-work> Uraeus: is that my initiation ritual?
[20:20] <mathrick> Uraeus: you filthy liar! it was 48, not 12
[20:20] <ds-work> I can fix all the bugs in the next 12 hours, but not in the most useful way
[20:20] <Uraeus> ds-work: yes, if you manage that you are officially a member of GSta-pi
[20:20] <ds-work> FIXED, NOTABUG
[20:22] <Uraeus> ds-work: btw, did you like Company's sync patch? I guess your approval would be nice before commiting?
[20:22] <ds-work> #1234546 merged with #134567 "GStreamer is the suck"
[20:22] Action: ds-work fears regressions
[20:22] <ds-work> and I want to see _every_ plugin fixed first (bastard I am)
[20:23] <Uraeus> before the patch?
[20:23] <ds-work> before applying, yes
[20:24] <taaz> is it possible to write a testsuite for sync?
[20:24] <Uraeus> ok, please reply to the mail then. I didn't realize it cause regressions
[20:24] <ds-work> a certain developer has a habit of moving on before finishing projects
[20:24] <Uraeus> :)
[20:24] AFK ([email protected]) joined #gstreamer.
[20:24] <Company> ds-work: what plugins need fixage?
[20:24] Nick change: AFK -> Markey-
[20:24] <Markey-> hi
[20:24] <Uraeus> hi AFK
[20:25] <Company> ds-work: alsaclock and ossclock work, all elements that sync work, too
[20:25] <Markey-> how could I get access to the currently playing audio samples?
[20:25] <Markey-> from my app
[20:25] <Uraeus> Markey-: I notice that you and wheels are seldom in here at the same time, is this a case of a split personality or are you really two persons ? :)
[20:25] <Markey-> hehehe
[20:25] <Markey-> shhhh.. big time secret
[20:25] <Markey-> well, I'd like to calculate an FFT
[20:26] <ds-work> * ds-work fears regressions
[20:26] <ds-work> not
[20:26] <ds-work> * ds-work sees regressions
[20:27] <Uraeus> Company: if I apply your patches to my local tree, will rb and sound juicer keep on working?
[20:27] <ds-work> Markey-: write an element
[20:28] <Markey-> that really necessary?
[20:28] <Markey-> hacking a plugin?
[20:28] <Company> Uraeus: they should
[20:28] <Uraeus> ok, I test then :)
[20:28] <Markey-> I just need to access an array of the samples regularly
[20:28] <Uraeus> didn't thomasvs make something for that? 
[20:31] <ds-work> you can always use an identity and connect to the handoff signal
[20:32] <Company> handling data is the job of plugins
[20:32] <Company> it's the only reason something like gstreamer exists...
[20:32] Action: taaz notes these common questions indicate the need for an examples section in the docs
[20:33] Action: Company notes someone needs to write that
[20:33] Action: Uraeus notes that taaz will update the docs today :)
[20:33] Action: taaz hides
[20:33] <taaz> Markey-: what are you going to do with the fft output?
[20:33] <Markey-> taaz: visualization
[20:34] <Markey-> we already use it with our aRts engine
[20:34] <Company> visualization is done by plugins...
[20:34] <taaz> that's the sort of thing that would work well as a plugin
[20:34] <Uraeus> http://www.osnews.com/story.php?news_id=5648 <- its live!!
[20:34] <Markey-> no, we do it our way
[20:34] <taaz> take audio on one side and spit out video on the other
[20:35] <Markey-> right, but we still need access to the sample data :)
[20:35] <Uraeus> Markey-: looks that thomasvs level plugin, it has a demo app in the source three
[20:35] <ds-work> Markey-: if you do it the gstreamer way, with ximagesink, then it will actually be synchronized with the sound output
[20:35] <ds-work> independent of the output device
[20:36] <Markey-> FFT routine is in place, I just need to feed it :)
[20:36] #gstreamer: mode change '+o Uraeus' by ChanServ!ChanServ@services.
[20:36] Topic changed on #gstreamer by [email protected]: GStreamer article on OSNEWS -> http://www.osnews.com/story.php?news_id=5648
[20:36] #gstreamer: mode change '-o Uraeus' by [email protected]
[20:38] <Company> ds-work: ossclock and alsaclock work, esdsink is as good as always, xvimagesink and ximagesink synchronize to avi and quicktime perfectly - is there anything else that needs doing to please you?
[20:39] <Uraeus> Company: mpeg :)
[20:39] <taaz> Uraeus: nice article
[20:39] <ds-work> Company: I want to see that every element is converted.  I don't know what other elements might need
[20:39] <Uraeus> taaz: thanks :)
[20:40] <Company> Uraeus: works like a charm
[20:41] <Company> ds-work: every element is converted
[20:42] <ds-work> Company: that's what I wanted to hear :)
[20:43] <ds-work> Company: so when I compile gst-plugins with the patch, it won't fail because of deprecated functions?
[20:46] <Company> ds-work: apart from the elements in ext, that I forgot again (and am fixing now), no
[20:52] thomasvs ([email protected]) joined #gstreamer.
[20:53] <taaz> why keep the old clocking api if it's broken and the deprecated functions are basically void anyway?
[20:53] <Company> because we're API stable
[20:54] <Company> and the clock_entry stuff still works
[20:54] <Company> I just want to get rid of it in 0.9 so noone should start using it now...
[20:54] <Company> in fact, the core still uses it internally
[20:55] <taaz> maybe add some debug/warning stuff in those functions that now do nothing
[20:56] <Company> sound like a good idea
[20:57] <thomasvs> use GST_DISABLE_DEPRECATED
[20:57] <taaz> yeah, but who knows about that? ;)
[20:57] <thomasvs> that's not an excuse :)
[20:57] <taaz> that reserved padding fu looks scary.  why *4 vs *GST_PADDING?
[20:59] <Company> taaz: yeah, *GST_PADDING is better
[20:59] <Company> taaz: we'll reset padding anyway when thomasvs' stuff goes in
[20:59] <taaz> the patch has "//".  the comment police will get angry
[21:00] <Uraeus> what?? who put those in?? who are so evil :)
[21:00] <ds-work> we don't support any compilers (for other reasons) that don't support //
[21:00] <Uraeus> such // slipups will enrage KDE hackers against us :)
[21:01] <Uraeus> ds-work: I thought the Irix compiler didn't support // (and people have successfully used that)
[21:01] <ds-work> does it support variadic macros?
[21:01] <Uraeus> no idea what it supports or not
[21:02] <ds-work> although we're pretty close to not needing variadic macros now
[21:02] <Uraeus> anyway I think we should hold on to only using /**/ comments since the code is 100% so currently
[21:03] <ds-work> I prefer /* */ for comments, but // for disabling a few lines of code
[21:03] <ds-work> and #if unused for large blocks of code
[21:05] <Company> I use // for comments that are temporarily, so I know wherre to look at later
[21:06] Action: taaz liking this -p diff option
[21:08] <thomasvs> Company: me too
[21:08] <thomasvs> ds-work: can you redo your german translation ? :)
[21:10] Action: Company thinks he'd be better qualified for that :p
[21:10] <ds-work> indeed
[21:10] matrixise ([email protected]) left irc: "Leaving"
[21:12] <thomasvs> Company: heh, well go ahead :)
[21:15] matrixise ([email protected]) joined #gstreamer.
[21:16] walters ([email protected]) joined #gstreamer.
[21:17] <iain> Company: if you want you can finish up marlin's too
[21:17] <iain> de: 245 translated messages, 20 fuzzy translations, 9 untranslated messages.
[21:17] <iain> thomasvs: and you can do dutch. :)
[21:17] <iain> nl: 263 translated messages, 7 fuzzy translations, 4 untranslated messages.
[21:18] <Uraeus> hi walters
[21:18] <thomasvs> hm :)
[21:19] <walters> yo
[21:20] Action: Uraeus sees cool comments start appearing on osnews :)
[21:20] gheet ([email protected]) left #gstreamer.
[21:20] meck ([email protected]) joined #gstreamer.
[21:21] <iain> just wait till the "KDE WILL NEVER USE GSTREAMER" people get on board
[21:21] Action: Company finds more annoying bufs wrt clocking
[21:21] <Company> bugs, too
[21:21] <Company> you know what's interesting?
[21:22] <Uraeus> Company: hope they are only annoying, not problematic :)
[21:22] <Company> i didn't know half of the stuff in the article
[21:22] <Uraeus> Company: like what?
[21:22] <Company> like the handhelds.org thingy
[21:22] <Uraeus> ah :)
[21:22] <thomasvs> Company: it's pretty cool, that guy showed it to us at guadec
[21:23] <Uraeus> guy == pb_
[21:23] <thomasvs> phil ? he's here ?
[21:23] <thomasvs> oh, hi phil :)
[21:23] Action: Company must be missing something a lot of times
[21:23] <thomasvs> guess I should keep up with nicks
[21:23] <pb_> thomasvs: heh.  hello :-)
[21:24] Zeenix ([email protected]) joined #gstreamer.
[21:25] <Zeenix> hi
[21:25] <ds-work> could someone write vorbisdec?
[21:25] <Uraeus> hi Zeenix
[21:25] <Uraeus> I thought it was vorbis demux we where missing
[21:25] <Uraeus> not decode
[21:25] <iain> no, we have oggdemux
[21:25] <iain> we need vorbisdec I think
[21:26] <ds-work> Company: is there anything that can be done about spider autoplugging correctly for both ogg tag extraction and ogg decoding?
[21:26] <Uraeus> I was thinking about oggmux :)
[21:27] <Company> ds-work: that should already work
[21:27] <ds-work> Company: it doesn't
[21:28] <ds-work> ./it.ogg ! spider ! application/x-gst-tags ! fakesink
[21:28] <ds-work> ./it.ogg ! spider ! audio/x-raw-int ! fakesink
[21:28] <ds-work> I can get one or the other working at a time, but not both
[21:30] Action: walters wishes gdb knew about macros
[21:31] <taaz> walters: it does a bit doesn't it?
[21:31] <walters> does it?
[21:31] <taaz> http://sources.redhat.com/gdb/current/onlinedocs/gdb_10.html
[21:32] dr88dr88 ([email protected]) joined #gstreamer.
[21:32] <Uraeus> cool :) ->
[21:32] <Uraeus>  Great article
[21:32] <Uraeus>  By Marshall (IP: 202.7.32.---) - Posted on 2004-01-13 20:25:47
[21:32] <Uraeus> This is great article.
[21:32] <Uraeus> OS news has been producing some great articles lately and this one is right up there. Great work.
[21:32] <walters> hmm
[21:32] Action: Uraeus does a ego trip
[21:32] <Company> 1 of 10 tests did not behave as expected (1 unexpected passes)
[21:32] <walters> i wonder what i need to do then to get it to know about gst macros
[21:32] <taaz> Uraeus: don't we have c++ bindings?  one comment about that...
[21:32] Action: Company hates the threading tests
[21:33] <Uraeus> taaz: well we have the KDE binding, the gtkmm style one is kinda in hiatus atm afaik
[21:33] <Uraeus> then again I don't know how up-to-date the kde binding is either, could be that it is just as outdated atm as the gtkmm style one
[21:35] Nick change: trow_lunch -> trow
[21:36] <Uraeus> trow: had a nice lunch?
[21:36] <Uraeus> trow: btw, you know that I sent a mail to rml?
[21:38] <trow> Uraeus: What did you mail rml about?
[21:39] Action: ds-work wants a cheese slicing laser
[21:40] <Uraeus> trow: retrace interupt counter
[21:42] <trow> I have no idea what a retrace interrupt counter is.  I guess that is why I'm not a kernel hacker. :)
[21:42] <Uraeus> trow: well its something we need in the kernel to be able to get near perfect audio& video sync
[21:43] <Uraeus> trow: I can forward you the mails if you want?
[21:43] <taaz> go on #livid and ask vektor about it ;)
[21:46] <walters> hmm.
[21:46] Action: mathrick finishes reading
[21:47] <ds-work> we don't need "near perfect" sync
[21:47] <Uraeus> mathrick: hope you liked it :)
[21:47] <ds-work> we just need "unperceptible"
[21:48] <mathrick> Uraeus: nice article, although I'd put more emphasis on apps & core being almost totally independent of plugins. Something that i hate the most in mplayer for ex
[21:48] <taaz> syncing with retrace interrupt will stop problems with tearing in the output.
[21:48] <walters> how do i go from a pad to its element?
[21:48] <walters> gst_pad_get_parent appears not to do it...(or maybe i'm misunderstanding something)
[21:49] <ds-work> that is how to do it
[21:49] <mathrick> that is, I hate mplayer *not* being independent, and forcing you to recompile everytime you'd like to add some format
[21:49] <walters> ds-work: oh, i see the bug in my code.
[21:49] <mathrick> Uraeus: but article's really nice, and great wrt. pc issues ;)
[21:50] <Zeenix> have anyone ever used gstreamer on a 386 system?
[21:51] <mathrick> Zeenix: you mean, 386 like in i386 == x86, or 386 == 80386 chip?
[21:52] <Zeenix> mathrick: 80386
[21:52] Action: Uraeus just got mail from eugenia that the article got approved for slashdot :)
[21:53] <Zeenix> Uraeus: cool, which article?
[21:53] <Uraeus> Zeenix: see topic :)
[21:53] <mathrick> is gstreamer.net ready to be slashdotted? ;)
[21:53] <Uraeus> mathrick: it is still on SF, so it should be good :)
[21:54] Action: mathrick fears people may get discouraged by gstreamer.net being heavily outdated
[21:54] <walters> p (char*)g_type_name ((((GTypeInstance*) (gst_pad_get_real_parent (srcpad)))->g_class)->g_type)
[21:54] <walters> fun
[21:54] <Uraeus> mathrick: ok update it quickly :)
[21:54] <walters> $18 = 0x8174ba0 "GstSpiderIdentity"
[21:56] <ds-work> ah, _that's_ why you want macros :)
[21:56] Rayban ([email protected]) joined #gstreamer.
[21:56] <walters> ds-work: you could say that :)
[21:56] <Rayban> great article on osnews, guys :)
[21:57] <Uraeus> thanks Rayban
[21:57] <ds-work> walters: Company wrote an object dumping function for that purpose, IIRC
[21:57] <walters> ds-work: what's it called?
[21:57] <Zeenix> Uraeus: good job, you'll be paid double this month :)
[21:57] <Rayban> quick question ...  is there any support for MPEG transport streams within gstreamer?
[21:58] <Uraeus> Zeenix: will I be paid in a libcaca plugin that works 100% ? :)
[21:58] <walters> Uraeus: so you're normally paid in libcaca plugins that work 50%? :)
[21:58] <Uraeus> walters: yeah, it is hell I tell you :)
[21:59] Action: Zeenix throws something at Uraeus for calling his plugin as hell..
[21:59] Eugenia ([email protected]) joined #gstreamer.
[22:00] <Zeenix> hi Eugenia
[22:00] <Eugenia> hey guys, the article on OSNews will be on Slashdot later today :)
[22:00] <mathrick> hi Eugenia
[22:00] <Eugenia> hi all
[22:00] Action: mathrick still has chills after reading libcaca source code
[22:00] <mathrick> Uraeus: IMO it's wonder it works at all ;)
[22:01] <Eugenia> no wonder with such name ("caca" in greek means "shit" :)
[22:01] <Uraeus> hi Eugenia, thanks for publishing the article :)
[22:01] <Eugenia> np
[22:01] <Rayban> good story
[22:02] <Eugenia> btw, if you want to check out the traffic after slashdot hits, check here: http://bilbo.counted.com/2/42699/1/?date=0
[22:02] <Eugenia> (there are pageviews per hour)
[22:02] <Zeenix> Eugenia: grab sam on #videolan for such a name 
[22:02] <Eugenia> heh
[22:02] <mathrick> Zeenix: he hangs on videolan?
[22:02] <Zeenix> mathrick: yup
[22:03] <Eugenia> I have emailed Sam in the past quite a lot... last night I instaled VLC 0.7 on FreeBSD and all I see is a black screen (all other stuff works, but no matter what video output used, it is all black :P)
[22:03] <mathrick> great, I have to ask him some inconvenient q's ;)
[22:03] <Eugenia> ok, got to run, need to put a washing machine going on and undo the christmas tree :P
[22:03] <Zeenix> mathrick: np but dont tell him that I told you of his hiding place :)
[22:03] <Eugenia> see ya all
[22:03] pb_ ([email protected]) left irc: Remote closed the connection
[22:04] <mathrick> cya Eugenia
[22:04] Eugenia ([email protected]) left irc: Client Quit
[22:04] <mathrick> Company: am I right gsttags is implemented completely on plugins side, w/o any special core support?
[22:06] <ds-work> no, GstElement is responsible for moving tags downstream
[22:06] <Rayban> hrm...  looks like mpeg transport streams are unsupported in gstreamer.  perhaps it might be an easy task to implement
[22:07] pb_ ([email protected]) joined #gstreamer.
[22:08] <Uraeus> walters: btw, you really need to get the mp3 tag editing in RB going, I am tired of listening to Annie Lenix :)
[22:08] <walters> Uraeus: i know.
[22:09] <mathrick> ds-work: i.e no special support, right?
[22:09] <walters> Uraeus: again, it's pretty much all there in rb
[22:09] <walters> Uraeus: since like december 24
[22:10] <Uraeus> walters: the other day you said tagging wasn't enabled in RB for mp3 yet, only Ogg
[22:15] Action: Uraeus curses the a52dec Makefile for making troubles
[22:16] Nick change: Jara[zZ] -> Jaramir
[22:17] <Zeenix> strange if the masks i get are in BE & they should convert on GUINT32_TO_LE not on GUINT32_TO_BE
[22:17] <Zeenix> s/&/then
[22:20] <mathrick> Zeenix: what macro do you use? I think it should be GUINT32_FROM_BE, as libcaca seems to expect host endianness
[22:21] <walters> Uraeus: * commited [email protected]/rhythmbox--mainline--0.7--patch-137
[22:21] <walters> Uraeus: it segfaults for me in gst now
[22:21] <Zeenix> mathrick: you are correct but still, it's not compatible
[22:22] <mathrick> Zeenix: compatible w/ what?
[22:22] <Uraeus> walters: ok, hopefully Company will update his RB copy then and debug why things crash
[22:22] <walters> Uraeus: i'll take a look too.
[22:23] <Zeenix> mathrick: after conversion i get it as ABGR & libcaca wants it to be BGRA
[22:24] <Uraeus> hmm, to bad BBB isn't here, Eugenia posted a question herself to osnews about webcams and gstreamer
[22:24] <mathrick> Zeenix: and what mask do you pass to it?
[22:24] <mathrick> Zeenix: ARGB?
[22:24] <Company> mathrick: you're wrong - gsttag* in the core does tagging
[22:25] <Company> and yes, Uraeus, Company is probably going in direction Rhythmbox with his coding stuff next. Though he's going to solve the boring "is this an audio file?" issue first
[22:25] <mathrick> Company: ok, so it requires additional support from core?
[22:25] <Uraeus> Company: ok sorry :) /me sensed some irritation :)
[22:25] Action: Jaramir is away: You need to upgrade your VESA local bus to a MasterCard local bus.
[22:26] Action: mathrick ROTFLs
[22:26] <Company> mathrick: yes, very little though
[22:26] <mathrick> Company: is that support tag specific? or more generic and could be used for something else?
[22:27] <Company> mathrick: that's "tag specific", though tags are very generic in itself
[22:27] <Company> why d'you ask?
[22:27] <mathrick> Company: writing elaborated comment ;)
[22:27] <Zeenix> mathrick: well, it works when i specify the masks as ABGR (sorry my previous statement was wrong)
[22:27] dolphy ([email protected]) joined #gstreamer.
[22:27] <Uraeus> hi dolphy
[22:28] <dolphy> yop
[22:28] <mathrick> y0 d01phy
[22:31] <Company> yay, gnu has an operating system now!
[22:31] <Rayban> what about hurd?  :)
[22:31] <Company> who'd have thought that the first complete gnu system would be embedded? ;)
[22:32] <mathrick> Company: hmm? what is that?
[22:32] <Company> mathrick: ecos (see slashdot)
[22:34] <Uraeus> ds-work: is the vnc plugin of yours working?
[22:39] <ds-work> Uraeus: yes
[22:39] <Uraeus> ds-work: can it be used to record vnc sessions? for educational videos etc?
[22:40] <mathrick> dolphy: was doing anything necessary to add dvdnav support to gst-player? or did it Just Work after thaytan's hacking?
[22:41] <Company> ds-work: clocking works, the core distchecks, the plugins are currently running - is there anything else I should do?
[22:43] dr88dr88 ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[22:44] <walters> Company: /build/gstreamer-0.7/bin/gst-launch-0.7 -t gnomevfssrc location=/audio-video/music/Evanescence/Fallen/Tourniquet.ogg ! typefind ! spider ! application/x-gst-tags ! fakesink
[22:46] <ds-work> Company: fix walters's bug?  :)
[22:46] <ds-work> Uraeus: yes
[22:47] <Uraeus> ds-work: ok, I will try and see if I can find a new developer to create a desktop session recorder GUI (unless gst-rec can do it just as easy)
[22:47] <Company> ds-work: broken getcaps function
[22:48] <ds-work> Uraeus: gst-rec is fine, as long as it displays the stream while recording
[22:48] <ds-work> Company: eh? where?
[22:48] <walters> ok.
[22:48] <walters> (gdb) bt
[22:48] <walters> #0  rb_debug_stop_in_debugger () at rb-debug.c:141
[22:48] <walters> #1  0x08073457 in log_handler (domain=0x40088840 "GStreamer", level=G_LOG_LEVEL_CRITICAL,
[22:48] <walters>     message=0x40bc0294 "file gstpad.c: line 1275 (gst_pad_renegotiate): assertion `GST_PAD_LINK_SINK (pad)' failed", data=0x0) at rb-debug.c:159
[22:48] <walters> #2  0x4febed69 in g_logv () from /usr/lib/libglib-2.0.so.0
[22:48] <walters> #3  0x4febef74 in g_log () from /usr/lib/libglib-2.0.so.0
[22:48] <walters> #4  0x40058bc3 in gst_pad_renegotiate (pad=0x8450688) at gstpad.c:1275
[22:48] <walters> #5  0x40f7e4af in gst_id3_tag_do_caps_nego (tag=0x8470010, buffer=0x812f600) at gstid3tag.c:766
[22:48] <walters> #6  0x40f7eef9 in gst_id3_tag_chain (pad=0x84508bc, data=0x812f600) at gstid3tag.c:926
[22:49] <Company> ds-work: spider identity, i'm fixing it atm
[22:49] <dolphy> ds-work: may i implement colorbalance interface in gst/videofilter/gstvideobalance.c ?
[22:49] <iain> dolphy: I got a crash with gst-totem. Do you want to see the bt by email or do I file it against Totem?
[22:49] <ds-work> dolphy: didn't you already?
[22:49] <Company> ds-work: works now, i'll commit it with the clocking stuff
[22:49] <Company> :p
[22:50] <dolphy> ds-work: nope
[22:50] <dolphy> ds-work: i was wondering today if this element is still working :)
[22:50] <dolphy> ds-work: never used it
[22:51] <dolphy> iain: pv msg
[22:51] <iain> dolphy: okay...
[22:51] <ds-work> colorbalance is hard to break, as it is a subclass of videofilter
[22:51] <dolphy> ds-work: does it have a mode where data simply pass through ?
[22:52] <dolphy> ds-work: and what is the ideal pipeline to get it to work with ximagesink ?
[22:52] <dolphy> ds-work: ffcolorspace ! videobalance ! ffcolorspace ! ximagesink ?
[22:52] <ds-work> dolphy: videotestsrc ! videobalance ! ffcolorspace ! ximagesing
[22:53] <ds-work> videobalance only works with I420, iirc
[22:53] <mathrick> dolphy: what was amount of code changed in gst-player to get basic dvdnav? 0?
[22:53] <dolphy> yeah but i need it in gstplay pipeline
[22:53] <ds-work> dolphy: by default, videobalance is passthru
[22:53] <Company> ds does evil I420 plugins, I always do RGBx
[22:53] <dolphy> mathrick: not that much 
[22:54] yf12s ([email protected]) joined #gstreamer.
[22:54] <dolphy> mathrick: the hard work was in the plugins
[22:54] <Company> our plugins are hard to interface ;)
[22:54] <mathrick> dolphy: did it work out of the box after thatytan's work?
[22:54] <ds-work> Company: actually, I do it in the most natural colorspace.  which is usually I420
[22:54] <Company> yeah, same here
[22:54] <Company> cairo is naturally RGB :)
[22:54] <ds-work> exactly
[22:54] <dolphy> mathrick: well it worked almost out of the box after the work of many people :)
[22:55] <dolphy> mathrick: getting navigation in gstreamer has been a long road :)
[22:55] <Company> ds-work: I figured audioconvert would be a nice test plugin for liboil
[22:55] <mathrick> dolphy: i know that, but that's for marketing ;)
[22:55] <taaz> new EffecTV out
[22:55] <Company> ds-work: because it has a nice way to specify optimization functions
[22:55] <ds-work> Company: yeah, the conv_ stuff is somewhat tested
[22:56] <mathrick> dolphy: did you change anything to get nav at all after thaytan did it with gst-launch?
[22:56] <ds-work> Company: my goal is to conver libaudioresample
[22:56] <mathrick> dolphy: i'm not talking about fancy support, but the most basic one
[22:56] <dolphy> mathrick: i had to send navigation events to the element supporting the NAVIGATION interface
[22:56] <ds-work> Company: actually, liboil is just ripped out of libaudioresample
[22:56] <dolphy> mathrick: x[v]imagesink in that case
[22:56] <Uraeus> ok, disting is broken currently, nothing in ext seems to be disted. If any of you changed something which could have caused it please look at it. I have yet to understand why it happens
[22:57] <ds-work> btw, who is going to write a lircnav element?
[22:57] <mathrick> dolphy: so there was some change in player's code necessary?
[22:57] <dolphy> yup
[22:57] <dolphy> few but yes
[22:57] <dolphy> and that s normal
[22:57] <dolphy> the video widget of gst-player and totem catches events
[22:57] <dolphy> to see mouse motion
[22:58] <dolphy> so they need to resend those events in the pipeline
[22:58] <mathrick> dolphy: ic, and rough number of LOCs necessary?
[22:59] <dolphy> mathrick: 40
[22:59] <mathrick> dolphy: thx :)
[22:59] <dolphy> what the hell are you writing :)
[23:00] <dolphy> ds-work: what lircnav would do ?
[23:00] <Company> send nav events from remotes
[23:01] <mathrick> dolphy: heh, comment on OSNews :)
[23:03] mxpxpod ([email protected]) joined #gstreamer.
[23:04] <Company> someone should employ Eugenia for bug finding
[23:04] <Company> "i need a webcam on FreeBSD that has no drivers. Can I use gst?"
[23:04] <mxpxpod> Company: no, she'd be like, "there's a gigantic bug!! it's called gnome!"
[23:05] <Company> sure you can use gst, mylady, but you'd need to write a driver there, too
[23:05] <Company> d'oh
[23:06] <Company> thomasvs: do you run make distcheck on your test computers? and if so, could you install alsalib?
[23:06] BBB ([email protected]) joined #gstreamer.
[23:06] <Uraeus> hi BBB !
[23:06] <BBB> hi
[23:06] <Uraeus> BBB: my article is now online at osnews :)
[23:06] <mathrick> hi BBB
[23:07] <Zeenix> guys, I am discusing the masks issue with sam on videolan & we seem to be getting into nowhere but confusion. Let me paste some last lines from him:
[23:07] <Zeenix> <sam> you seem to be positive that the original BE masks were 0xff0000, 0xff00, 0xff
[23:07] <Zeenix> <sam> which is ARGB
[23:07] <Zeenix> <sam> but from what I see, your data looks more like RGBA
[23:07] <Zeenix> <sam> so the original BE masks should have been 0xff000000, 0xff0000, 0xff00
[23:07] Nick change: harshyWork -> harshy
[23:08] <Zeenix> so any clue?
[23:08] <BBB> Zeenix: I mailed you about that...
[23:08] <BBB> the masks are BE
[23:08] <BBB> and they're really correct ;)
[23:09] <BBB> Uraeus: guess what I'm reading right now
[23:09] <mathrick> BBB: matroska issues evolved into new, interesting form, I think i will wait w/ bzilla until problems stabilise a bit ;)
[23:09] <Zeenix> BBB: yes you did 
[23:10] <Zeenix> BBB: but the problem is still there
[23:10] <BBB> mathrick: I'll try to debug it asap... I'm just not very much available right now
[23:10] <mathrick> BBB: currently, they change w/ every cvs up ;)
[23:10] <BBB> maybe tomorrow
[23:10] <BBB> mathrick: nice bug report ;) :p
[23:10] <mathrick> BBB: I don't have too much time either
[23:10] <Markey-> hey BBB :)
[23:10] <BBB> Uraeus: nice article! :)
[23:10] <BBB> hi Markey- :)
[23:10] <mathrick> BBB: hence this sparse info ;)
[23:11] <Markey-> BBB: do you see a way of accessing currently playing audio samples?
[23:11] <Markey-> I need to feed my FFT routine
[23:11] <Markey-> with audio data
[23:11] <BBB> And then if things go as planned I guess we see a GStreamer-using video player and maybe even a video recorder bundled with GNOME 2.8 <- Uraeus, that is confidential :p
[23:11] <BBB> Markey-: identity, and connect to the handoff signal
[23:11] <BBB> Markey-: that's what most people use
[23:11] <BBB> it's basically a hack, but a very well working one
[23:12] <Markey-> let me look at that signal
[23:12] <Uraeus> BBB: does gst-rec work with ds-work's new vnc plugin? it would be cool if people could use it for recording education videos etc.
[23:12] <mathrick> BBB: IMO it's not really more of hack than letting data out of pipeline at all is ;)
[23:12] <BBB> I didn't test, but I guess it does
[23:13] <BBB> if it uses 'Source/Video' as category, it will be autodetected
[23:13] <Markey-> BBB: do you possibly know some existing code I could look into?
[23:13] <BBB> for using it?
[23:13] <BBB> mathrick: true :p
[23:13] <Markey-> yep
[23:14] <BBB> Markey-: identity = gst_element_factory_make ("identity", "my_element"); g_signal_connect (identity, "handoff", G_CALLBACK (cb_handoff), NULL);
[23:14] <BBB> and cb_handoff looks like:
[23:14] <BBB> static void cb_handoff (GstElement *identity, GstData *data, gpointer data)
[23:14] <Markey-> thanks a lot!
[23:14] <BBB> check GST_IS_BUFFER (data) { ... do your processing here ... }
[23:14] <Markey-> you rock
[23:15] <BBB> no prob :)
[23:15] <Markey-> I'll try it
[23:16] Misirlou ([email protected]) joined #gstreamer.
[23:17] <BBB> Uraeus: very nice article!
[23:17] <Uraeus> BBB: thanks :)
[23:19] <dolphy> hmm videobalance has strange values
[23:20] Action: Misirlou reads
[23:21] jdahlin_ ([email protected]) joined #gstreamer.
[23:21] jdahlin ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[23:21] <dolphy> ds-work: [0,2] for contrast and saturation [-1,1] for brightness and hue
[23:21] <dolphy> ds-work: moreover the colorbalance interface passes a gint 
[23:21] <dolphy> ds-work: so it kind of reduce the settings possibilities :)
[23:21] <dolphy> ds-work: compared to Xv which use a range from [-1000,1000]
[23:21] <Company> export G_MININT to G_MAXINT and calculate based on that
[23:22] <Misirlou> Cool, I didn't know we had a Tremor plug-in.
[23:22] thaytan ([email protected]) joined #gstreamer.
[23:22] <BBB> dolphy: I just sent a reply to your CVS commit for the type in colorbalance
[23:22] <mathrick> yo thaytan
[23:22] <BBB> dolphy: please add a default value in base_init() in colorbalance.c
[23:22] <iain> Uraeus: "That said I don't think many in the GNOME and KDE community takes MAS serious anymore after their abysmal track record for the last years."
[23:22] <iain> Uraeus: nice :)
[23:23] <Uraeus> hi thaytan
[23:23] <dolphy> BBB: apart from that is it ok ?
[23:23] <Misirlou> Uraeus: Really glad to see Vorbis and Theora mentioned!
[23:23] <dolphy> BBB: we should probably add a macro GST_COLOR_BALANCE_TYPE ()
[23:24] steve_b ([email protected]) joined #gstreamer.
[23:24] <dolphy> steve_b: hi
[23:24] <steve_b> hey
[23:24] <Uraeus> hi steve_b
[23:24] <steve_b> i'm at linuxconf :)
[23:24] <Uraeus> steve_b: seen the article?
[23:24] <dolphy> lucky you
[23:24] <steve_b> loading now
[23:25] Action: Jaramir is back (gone 00:59:53)
[23:25] <steve_b> i sat next to linus at the speakers dinner last night
[23:25] <trow> dolphy: So I was trying to figure out why seeking doesn't work in a GstPlay, and I couldn't figure out where the seek event ends up.  Since xvimagesink doesn't implement send_event, the seek event goes to the pad.  And it isn't clear to me where it ever ends up...
[23:26] <BBB> dolphy: all fine, please go ahead :)
[23:26] <dolphy> trow: it s normal event handling of GstElement
[23:26] <BBB> iain: where does he say that? ;)
[23:26] <dolphy> trow:  it goes upstream in the pipeline
[23:27] <thaytan> Uraeus: good morning :)
[23:27] <thaytan> gotta go - official conference opening is on
[23:27] <trow> dolphy: I didn't see where the pad ever handed it back to the element.  And I set a breakpoint on gst_element_send_event which never got tripped after the initial call on the xvimagesink.
[23:28] <Uraeus> thaytan: make sure to tell people to check out osnews for gstreamer article :)
[23:28] <Company> trow: events are handled by pads only
[23:28] <mathrick> yay, finally done writing ;) That surely took long for such a brief text
[23:28] <walters> it's a bug to call gst_pad_renegotiate on a not-negotiated pad, right?
[23:28] <trow> Company: Oh, OK.  I guess I'd assumed that they just propagated along the pad but would end up back on an element.
[23:28] <mathrick> http://anime.swat.pl/~mathrick/gst.txt <-- could anyone take a look and correct terrible mistakes I've done for sure? ;)
[23:29] <Company> which reminds me
[23:29] <iain> BBB: one of the comments
[23:29] <Company> ds-work: is there a way for elements to force renegotiation?
[23:29] <walters> Company: the only way i see is via try_set_caps
[23:30] <dolphy> trow: the event goes from one pad to another
[23:30] <walters> Company: i am looking at gstid3tag.c:766
[23:30] <dolphy> trow: until it reaches an element which implements an event handler method
[23:31] Nick change: Jaramir -> Jara[zZ]
[23:31] steve_b ([email protected]) left irc: "keynote time"
[23:31] <Company> walters: yeah, that's where i found that, too
[23:32] <trow> dolphy: So in GstPlay's pipeline, which element should be handling the seek events?
[23:33] <Company> everyone
[23:34] <Company> depends on the format
[23:34] <thaytan> Uraeus: ok :)
[23:34] <thaytan> turns out we have net access from the main hall ;)
[23:35] <Company> in a simple mp3 pipeline, the mp3 decoder would translate the seek to a bytebased seek and pass it on to the filesrc which would then change the offset it reads from
[23:35] <Company> in a video app it's a bit mo0re complicated
[23:36] <iain> mathrick: you haven't mentioned marlin either... 
[23:36] <Company> most likely the demuxer translates it to bytebased and passes it on
[23:36] <iain> mathrick: no-one mentions marlin
[23:36] <Company> marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin marlin
[23:36] <Zeenix> some please answer one simple stupid question: the data is kept in the NE & ONLY the masks are in BE?
[23:36] <Company> i doo :)
[23:36] <mathrick> mushroom mushroom ;)
[23:36] <iain> Company: thanks :)
[23:37] <Uraeus> iain: post a marlin comment to osnews :)
[23:37] <Company> Zeenix: the data is kept in BE, the msaks are host endian
[23:37] <mathrick> iain: :) ok, how should i describe it? what ass kicking things does it do that show gst's ultimate power? :)
[23:38] <Company> Zeenix: so you do red_part = pixel && GINT_TO_BE (red_mask)
[23:38] <Company> Zeenix: red_part = pixel & GINT_TO_BE (red_mask) obviously
[23:38] <iain> mathrick: it works occasionally
[23:38] Nick change: Jara[zZ] -> Jaramir
[23:38] <Company> the roadmap
[23:38] <iain> mathrick: and its usually the first application to stop working on any changes made to gst
[23:38] <Company> marlin is all about the roadmap
[23:39] <mathrick> iain: i'm afraid you'd have to describe it yourself then ;)
[23:39] <iain> mathrick: nah, you don't need to mention it :)
[23:39] <Company> iain: opt still has a bug with event handling btw
[23:39] <iain> Company: is it a one line fix?
[23:40] <Company> iain: i'm 100% it is
[23:40] <Company> ... possible, that it is
[23:40] <mathrick> iain: btw, I don't really have a clue about audio editing, so marlin isn't my everyday used app ;)
[23:40] <iain> when you're translating a po file and you don't need to translate a string how do you tell intltool that?
[23:41] <iain> mathrick: funny thing is
[23:41] <iain> mathrick: neither do I
[23:41] <mathrick> iain: heh :)
[23:41] <iain> mathrick: as I've said before I'm just making it up as I go along
[23:41] Action: BBB will look at media playback tomorrow...
[23:41] <BBB> I might actually fix .m4a if someone uploads some sample files to gst.net/media/
[23:42] <iain> I have one, but I don't know if its legal
[23:42] <Zeenix> BBB: but did'nt you tell me that masks are in BE?
[23:42] <mathrick> BBB: do .mkv with mp4 audio stream count in? This is the nice one I have so much fun with ;)
[23:43] <BBB> Zeenix: yes
[23:43] <BBB> mathrick: please upload :)
[23:43] <BBB> but I'd like an actual .m4a, too
[23:43] <mathrick> BBB: is it public access area>
[23:43] <BBB> ok, thnx
[23:43] <mathrick> ?
[23:43] <Zeenix> BBB: Company: so are you both telling me the opposite?
[23:43] <BBB> oh, no, ehm
[23:43] <BBB> bugzilla?
[23:44] <iain> BBB: do you want my m4a of dubious legality?
[23:44] <BBB> iain: sure... can you dcc it tomorrow?
[23:44] <BBB> Zeenix: no, company is saying the same
[23:44] <mathrick> BBB: maybe http w/ nice connection, and you will upload it yourself?
[23:45] <BBB> sure
[23:45] <Zeenix> <Company> Zeenix: the data is kept in BE, the msaks are host endian
[23:45] <BBB> mathrick: if I have it locally, it's fine
[23:45] <BBB> Zeenix: that's the same ;)
[23:45] <mathrick> BBB: on sec
[23:45] <iain> BBB: I can upload it somewhere other than gst.net if you'd prefer
[23:45] <BBB> Zeenix: point being that on LE, you need to byteconvert one of the two to use it
[23:45] <BBB> iain: it's all fine with em
[23:45] <BBB> iain: as long as I have it ;)
[23:45] <iain> okay
[23:46] <iain> BBB: http://62.188.60.41/~iain/01_Civili.m4a
[23:47] <BBB> please remind me tomorrow
[23:47] <BBB> it'd take me an hour here
[23:47] <iain> ok
[23:47] <Uraeus> Markey-: maybe you could post my gstreamer article to the dot?
[23:47] <BBB> ty ;)
[23:47] <Markey-> Uraeus: not really, I don't post there
[23:47] <Markey-> Uraeus: maybe wheels?
[23:48] <Uraeus> Markey-: he is not online, you are :)
[23:48] <Zeenix> BBB: so the data is in ABGR format?
[23:48] <iain> I could do it
[23:48] <iain> they love me on the dot :)
[23:49] <Markey-> Uraeus: no really I don't know the dot guys personally
[23:49] <Markey-> but I can ask someone
[23:49] <Markey-> wait
[23:49] <BBB> Zeenix: depends how you look at it, but yes :)
[23:50] <BBB> Uraeus: we could do a small campaign on gnome.org for media file support... "if your file doesn't play with gst-player, please add it to bugzilla so we can fix it"
[23:50] <BBB> Uraeus: not now, but next week or in two weeks or so
[23:50] <BBB> Uraeus: is that ok?
[23:50] <mathrick> BBB: http://anime.swat.pl/~mathrick/[ax]arc_the_lad_-_episode_10_[BCF58409].mkv
[23:52] <BBB> mathrick: can I get that link tomorrow too please?
[23:52] Action: BBB kicks his stupid analog 56k modem
[23:52] <Company> who was the guy complaining about flac taking a long startup on PPC?
[23:52] lmjohns3 ([email protected]) joined #gstreamer.
[23:52] <mathrick> BBB: sure, you'll get at 7:41 in daily IRC logs i presume?
[23:53] <Uraeus> BBB: sure :)
[23:53] <dolphy> Company: mxpxpod:
[23:53] <Uraeus> hi lmjohns3
[23:53] <lmjohns3> hi hi !
[23:53] <lmjohns3> been quite a while ...
[23:53] <Uraeus> lmjohns3: long time no see :)
[23:53] <Company> mxpxpod: you had the flac ppc problem?
[23:53] <Company> leif :)
[23:53] <Company> hi
[23:53] <lmjohns3> Uraeus: i liked your article on osnews
[23:53] <mxpxpod> Company: yes
[23:53] <Uraeus> lmjohns3: thanks :)
[23:53] <Company> mxpxpod: it works fine on my Debian unstable iBook with current HEAD
[23:53] <mxpxpod> hmm
[23:53] <ds-work> Company: use gst_pad_negotiate() and make sure your getcaps function works correctly
[23:53] <lmjohns3> Uraeus: glad you're here. i'm working on adding midi elements !
[23:53] <Markey-> Uraeus: everyone seems to be more or less asleep.. KDE guys go to bed early ;)
[23:54] <Uraeus> lmjohns3: cool!!! 
[23:54] <Markey-> Uraeus: maybe just submit the article via mail
[23:54] <ds-work> Company: or maybe it's renegotiate()
[23:54] <Zeenix> BBB: so my (converted) masks should be: red: 0x000000ff, green: 0x0000ff00, blue: 0x00ff0000 ?
[23:54] <Company> ds-work: and that ensures that the old link is removed?
[23:54] <lmjohns3> Uraeus: i looked through the code from that fellow in taiwan, it looks ok but needs some cleaning up.
[23:54] <BBB> hi leif
[23:54] <lmjohns3> BBB: hi
[23:54] <ds-work> Company: it's supposed to
[23:54] <lmjohns3> BBB: so i'm also adding a sequencer interface along with the midi stuff
[23:54] <dolphy> mxpxpod: works fine here too
[23:54] <BBB> Zeenix: for 32bpp, they should be GUINT32_FROM_BE()... for 24bpp, they should be GUINT32_FROM_BE()>>8
[23:54] <lmjohns3> BBB, Uraeus: wanted to bounce some ideas off you if you have time ...
[23:54] lypanov ([email protected]) joined #gstreamer.
[23:54] <BBB> yay! \o/
[23:55] <dolphy> mxpxpod: just checked
[23:55] <lypanov> dum di dum
[23:55] <BBB> lmjohns3: oh, then I suck
[23:55] <Uraeus> hi lypanov
[23:55] <lypanov> hey Uraeus 
[23:55] <BBB> lmjohns3: because I need sleep
[23:55] <BBB> lmjohns3: is tomorrow morning ok?
[23:55] <lmjohns3> BBB: doh !
[23:55] <BBB> lmjohns3: or email...
[23:55] <BBB> ok, ok, I'll stay online another 15 minutes
[23:55] Action: Uraeus declares new record in IRC channel participation :)
[23:55] <lmjohns3> BBB: i'll talk with Uraeus and maybe send an email ?
[23:55] <lmjohns3> rock !
[23:55] <mxpxpod> dolphy: hmmm
[23:56] <Uraeus> lypanov: you are here to tell me you are putting my article onto the dot? :)
[23:56] Action: BBB starts his clock at '15 minutes encounting'
[23:56] <dolphy> mxpxpod: i tried south.flac
[23:56] Action: ds-work ops and kicks some people
[23:56] <lypanov> hehe
[23:56] <mxpxpod> dolphy: where's that at?
[23:56] <mathrick> BBB: would you like me to throw that link at you tomorrow, or you'll get it from logs?
[23:56] <dolphy> mxpxpod: which is available on web site
[23:56] <lmjohns3> ok, so quickly, i looked at this fellow's code (the midi elements he wrote) and there seems to be code for a midiparse element (converts file information to midi channel information) ...
[23:56] <taaz> Uraeus: get wtay and omega here too ;)
[23:56] <lypanov> Uraeus: just reading it ;-)
[23:56] <BBB> mathrick: throw it at me so I don't forget
[23:56] <dolphy> mxpxpod: http://gstreamer.net/media/incoming
[23:56] <lmjohns3> ... and also code for a midisink element
[23:56] <mathrick> BBB: ok, if i don't forget, too ;)
[23:57] <lmjohns3> yikes---people everywhere ! :)
[23:57] <Company> boohoo, my gedit doesn't work, evil gtk HEAD
[23:57] <BBB> lmjohns3: that's good.... is midisink a 'soundcard midi output' or a 'midi file output'?
[23:57] <mxpxpod> dolphy: what version of glib are you using?
[23:57] <Zeenix> BBB: oh, ok >> 8 was the key
[23:57] <lmjohns3> BBB: it's a soundcard output. i think we could do a midienc element that would be suitable for file output.
[23:58] <BBB> Zeenix: you're converting a 24bits number, GUINT32_FROM_BE() does 32, thus the >>8
[23:58] <BBB> ok...
[23:58] <BBB> lmjohns3: but that absolutely rocks
[23:58] <BBB> I mean, it means we have quite some required code available
[23:58] <dolphy> mxpxpod: 2.2.3
[23:58] <BBB> we just need to tweak it right
[23:58] <BBB> right?
[23:58] wheels ([email protected]) joined #gstreamer.
[23:59] <mxpxpod> dolphy: hmm, it works fine here too...
[23:59] <lmjohns3> my ponderings started coming up with the sequencer interface ... should we have an interface, or a superclass ?
[23:59] <Uraeus> hi wheels!
[23:59] <lmjohns3> BBB: yes, if you were talking to me, most of the code is already written.
[23:59] <Zeenix> BBB: thanks. BTW you should always expect such stupidities from me :)
[23:59] <mxpxpod> dolphy: is that a flac or an ogg/flac?
[23:59] <wheels> hi Uraeus 
[23:59] <Zeenix> Uraeus: sam was'nt expecting libcaca to become so much famous
[00:00] --- Wed Jan 14 2004
[00:00] <lypanov> hey wheels 
[00:00] <lypanov> you feel like writing dot article?
[00:00] <Uraeus> hmm, with both Markey- and wheels here now I guess my theory about one person with split personality has to go :)
[00:00] <mathrick> yay, Xchat is lit up like Christmas tree ;)
[00:00] <Misirlou> mathrick
[00:00] Action: lypanov would like wheels to write something, the users like scott :)
[00:00] Action: wheels tells Uraeus about the possibility of using two clients at once.  ;-)
[00:01] <mathrick> Misirlou: yes? :)
[00:01] <lypanov> hehe
[00:01] Action: Misirlou tries to light up mathrick's X-Chat :)
[00:01] <mathrick> hehe
[00:01] <lypanov> Uraeus: nice article btw
[00:01] <BBB> lmjohns3: simple... superclasses are for sharing code between elements... interfaces are for sharing a common control method for several elements with an application
[00:01] <thaytan> Uraeus: very nice
[00:01] <thaytan> I just finished reading it
[00:01] <lmjohns3> Uraeus, BBB: how bout i make additions/changes in the gstreamer_and_midi file, then you all can read through it ?
[00:02] <Uraeus> lmjohns3: sound good, to many people here now to follow any discussion :)
[00:02] <BBB> wheels: Uraeus wants you to do something ;)
[00:02] <lmjohns3> BBB: right, but i think we could use both in the sequencer case. there are several possible elements that could implement a sequencer, but those elements would all end up needing to share a lot of common code.
[00:02] <wheels> BBB : Yeah, I know -- but I've told him several times that I don't feel the same way about him.  ;-)
[00:02] <BBB> lmjohns3: that's perfect... if you add stuff to CVS, I'll read it, I pretty much read every commit
[00:02] <lmjohns3> BBB: i suppose that would point to the superclass method over an interface ...
[00:03] <lmjohns3> BBB: ok, will do.
[00:03] <BBB> lmjohns3: you can do both
[00:03] <dolphy> mxpxpod: hmm i don't know :)
[00:03] <BBB> lmjohns3: videosink/xoverlay is an example of that
[00:03] <dolphy> mxpxpod: give us a file that doesn't work then
[00:03] <lmjohns3> BBB: right ... might do it that way :)
[00:03] <mxpxpod> dolphy: also, what compression level is it?
[00:03] <mxpxpod> dolphy: ok
[00:03] <BBB> videosink is shared code between video outputs... xoverlay is the application interface to that
[00:03] Action: ds-work wanders off to #gstreamer-devel
[00:03] <lmjohns3> BBB: oh, cool, i'll check that out for inspiration.
[00:03] <BBB> so we have both
[00:03] <mxpxpod> dolphy: where can I upload to?
[00:03] <BBB> ds-work: does that exist?
[00:04] <mathrick> BBB: try /j #gstreamer-devel and you'll see ;)
[00:04] <BBB> hm, scary
[00:04] <BBB> 3 people
[00:04] <BBB> that's like worse than when gstreamer started
[00:05] <Uraeus> heh :)
[00:05] Action: wheels warns the crowd that the -devel trick never works.  It just takes a few days and then it's similarly off-topic.  :-)
[00:05] <lypanov> yup
[00:05] <taaz> not really... i was the 4th person to hang out here ;)
[00:05] <BBB> wheels: oh, and Uraeus wanted to tell you that, but also something else ;)
[00:05] <mathrick> BBB: were you the 3rd one to join in? right after wtay?
[00:05] <Uraeus> wheels: we want you to post my osnews article to the dot :)
[00:06] <taaz> used to just be wtay, omega, and sometimes richardb
[00:06] <lypanov> Uraeus: if you write a summary i'll do it ;-)
[00:06] <BBB> mathrick: no, fifth or so
[00:06] <Uraeus> ok, time for me to sleep, larry is not happy when I am tired at work :)
[00:06] <BBB> chiefhighwater
[00:06] <lypanov> Uraeus: sleep sucks :)
[00:06] <BBB> I think I came along with dennis
[00:06] <BBB> dennis actually kicked me into this
[00:07] <lmjohns3> does anyone have a spare 12" powerbook they want to give me ?
[00:07] <BBB> nite Uraeus
[00:07] <mathrick> BBB: who's dennis? :)
[00:07] <wheels> Uraeus : Yeah, give me a bit -- I'm still in my "just got home from work" stupor...
[00:07] <BBB> mathrick: some funny guy ;)
[00:07] <mathrick> night Uraeus
[00:07] <Uraeus> lypanov: Just write about Juk and amaroK and point to the article :)
[00:07] <Uraeus> ok, night everyone :)
[00:07] Uraeus ([email protected]) left irc: "Client exiting"
[00:07] <lmjohns3> night Uraeus 
[00:07] <wheels> dammit -- he's not very good for discussions.  ;-)
[00:07] <lypanov> hehe :)
[00:08] <BBB> lypanov: ohmy, you're from KDE too?
[00:08] <lypanov> BBB: no. i'm an evil dwarf
[00:08] <lypanov> ask scott he'll agree
[00:08] <BBB> wheels?
[00:08] <lypanov> yeah
[00:08] <wheels> BBB : We're taking over.  (The evil dwarves that is.)
[00:08] <lypanov> hehe :)
[00:08] <BBB> dwarf?
[00:08] <BBB> uhm
[00:08] Action: BBB thinks
[00:08] <BBB> are those scary?
[00:08] <BBB> I mean
[00:08] <lypanov> well, i am at least
[00:08] <BBB> I can just squash them with my feet if they annoy me
[00:09] <BBB> squeeze, even
[00:09] <lypanov> and scott is pretty fearsome when fighting admittedly
[00:09] <lypanov> so, he could be compared to a dwarf i guess
[00:09] <BBB> does scott fight when he drinks?
[00:09] <BBB> if so, is he a brit?
[00:09] <BBB> if so: auch
[00:09] <wheels> BBB : We'll have you bitching about C in no time.  :-)
[00:09] <lypanov> hehe
[00:09] <lypanov> wheels: lol
[00:09] <lypanov> yeah. c sucks
[00:09] <lypanov> but then. i hate C++ too ;-)
[00:09] <BBB> that sounds interesting
[00:09] Action: wheels waits for it...
[00:09] <iain> wheels: thanks for "defending" my point on the dot
[00:10] <BBB> a KDE dwarf saying c++ sucks
[00:10] <BBB> :p
[00:10] <mathrick> c is fine, c++ not :)
[00:10] <lypanov> mathrick: they both suck
[00:10] <BBB> wheels: /me bitching about C? impossible
[00:10] <BBB> :p
[00:10] <BBB> lypanov: where you one of the persons mentioned in the KDE article in Trouw one-and-a-half week ago?
[00:11] <mathrick> lypanov: nope, C has it's usage, but c++ is really horrendous for anything more complicated due to unfortunate choice of multiple inheritance instead of ifaces
[00:11] <mxpxpod> dolphy: I'm uploading to my church's server right now... gimme a sec and I'll have it for you
[00:12] <wheels> iain : Ah, which one was that?
[00:12] matrixise ([email protected]) left irc: "Leaving"
[00:12] <iain> wheels: the one where I pointed out why eazel didn't waste 50M$ on Nautilus
[00:12] <iain> wheels: and pissed people of by saying "fuck"
[00:13] <iain> s/of/off/
[00:13] <mathrick> lypanov: C combines all the power and flexibility of assembly with all the elegance and maintainability of assembly ;) But really, it's nice because it's so simple
[00:13] <iain> I love how people's priorities are all wrong
[00:13] <iain> its perfectly legit to slander people, call them liars and cheats
[00:13] <iain> and yet you say "fuck" and suddenly you're the evil one
[00:14] <wheels> iain : Ah, yeah.  Well, yeah -- I mean there may be a valid debate between GTK and Qt, but Eazel having bad management has little to do with that.  ;-)
[00:14] steveb_ ([email protected]) joined #gstreamer.
[00:14] <iain> wheels: yeah. exactly :) Pity some people didn't get that point at all
[00:14] matrixise ([email protected]) joined #gstreamer.
[00:15] <ds-work> mathrick: obviously written by someone with no expericene in assembly
[00:15] <iain> (notably derek kite (who I had some respect for) and some guy david (who I've seen being clueless before))
[00:15] <steveb_> thaytan, do you have wireless?
[00:16] <mathrick> ds-work: just you wait until I have my assembly classes ;)
[00:16] Action: BBB really goes to bed now
[00:16] <BBB> byebye
[00:16] BBB ([email protected]) left irc: "Client exiting"
[00:16] <wheels> iain : It's generally well intended, but that's typical from the "fan clubs" of both desktops (which is mostly what the dot is anyway)...
[00:17] <lypanov> re
[00:17] Action: sxpert re-enables ivorbis building/installing
[00:17] <mathrick> ds-work: but it's not mine either, I just like that quote, not entirely agree with it ;)
[00:17] <lypanov> um.. missed BBB
[00:17] <lypanov> mathrick: ummmmmmmmmmmmmmmmmm
[00:17] <lypanov> mathrick: thats my response :)
[00:18] <lypanov> mathrick: but, really. c is more complex
[00:18] <mathrick> lypanov: I think i didn't get it :)
[00:18] <lypanov> hehe :)
[00:18] <mathrick> lypanov: assembly is obviously more complex
[00:19] <lypanov> i meant c
[00:19] <iain> wheels: maybe
[00:19] <lypanov> is more complex than c++
[00:19] <lypanov> at a usage level
[00:19] <mathrick> lypanov: no way
[00:19] <lypanov> "usage level"
[00:19] <lypanov> you just learn c + glib + crap
[00:19] <mathrick> lypanov: yep, "usage level"
[00:19] <lypanov> its not a simple subset or anything
[00:20] <lypanov> you end up having to learn just as much
[00:20] <lypanov> as me with qt
[00:20] <lypanov> if not more
[00:20] <mathrick> lypanov: but then, it's usable, contrary to C++ multi inheritance
[00:21] <lypanov> balls
[00:21] <lypanov> but anyway
[00:21] <lypanov> i hate c++
[00:21] <lypanov> i just hate c more thats all
[00:21] <lypanov> :)
[00:21] <mathrick> lypanov: believe me, up until recently i thought C++ was nice lang :)
[00:21] <lypanov> i learnt c++
[00:21] <lypanov> then c
[00:21] <lypanov> then c++
[00:21] <lypanov> then c
[00:21] <lypanov> repeat
[00:21] <lypanov> ad infintinum
[00:21] <lypanov> they both suck
[00:22] <mathrick> lypanov: but then, I started to use it seriously, and came to hate C++ :)
[00:22] <mathrick> lypanov: for nice programming, there is python
[00:22] <lypanov> python sucks
[00:22] <lypanov> ruby ruby ruby
[00:22] <mathrick> lypanov: for quick, portable, or low-level, there is C
[00:22] <lypanov> quick
[00:22] <wheels> Wow, that took longer than I expected. lypanov's getting more reserved.  ;-)
[00:22] <lypanov> hehe
[00:22] <mathrick> s/quick/fast/
[00:22] <sxpert> mathrick: I did a project 100% in c++, and came to hate it 2 months in the project
[00:23] <lypanov> hey sxpert 
[00:23] <mathrick> sxpert: same here
[00:23] <lypanov> everyone at fosdem again?
[00:23] <lypanov> sxpert: stl and all that crap i guess
[00:23] <lypanov> yeah. i agree
[00:23] <lypanov> i prefer std c than that shit
[00:23] <lypanov> but. i use qt
[00:23] <lypanov> not pure C++
[00:24] <mathrick> lypanov: afaik, ruby's just pyhton w/o meaningful indentation
[00:24] <mathrick> lypanov: STL is actually one of the nicer things in C++, it's really impressive
[00:24] <wheels> sxpert : mathrick : C++ is definitely not a language to go at for 2 months and make a call on it.  Heck, I think you need an extra 6 months just to get out of C habits.  :-)
[00:24] <mathrick> lypanov: but then, it requires you to use C++ ;P
[00:24] Action: sxpert reboots carpc to see if update of alsa and gstreamer worked
[00:25] <dolphy> i  m off to bed
[00:25] <mathrick> dolphy: night
[00:25] <sxpert> wheels: well, you haven't seen that project I was in... complete windowing library for a dos app :D
[00:26] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[00:26] <lypanov> mathrick: python is ugly shit
[00:26] <lypanov> i mean
[00:26] <lypanov> self.__repr__(self, param1)
[00:26] <lypanov> what is that shit?
[00:26] <wheels> sxpert : That doesn't change the above.  :-)
[00:26] <lypanov> boost is the nicest thing in c++
[00:26] <lypanov> stl sucks
[00:26] <mathrick> wheels: I don't claim being C++ lawyer, but I definitely dislike fighting C++ for two days, and failing to compile *interfaces only*
[00:27] <mathrick> lypanov: boost?
[00:27] <mxpxpod> Company: can you check and see if this flac works for you?
[00:27] <mathrick> lypanov: one of that fancy libs out there>
[00:27] <sxpert> of course, STL is a tangled mess of pure evil
[00:27] <lypanov> mathrick: yup
[00:27] <mathrick> s/>/?/
[00:27] Eugenia ([email protected]) joined #gstreamer.
[00:27] <lypanov> sxpert: yup
[00:27] <Eugenia> story is now on slashdot
[00:27] steveb_ ([email protected]) left irc: Remote closed the connection
[00:27] <lypanov> hey Eugenia 
[00:27] <Eugenia> hiya
[00:27] <wheels> mathrick : Well, I came to C++ from a background of several years of C programming.  I really didn't consider myself a C++ guy for at least a year...
[00:27] <lypanov> Eugenia: was it you that several years back used to like allos?
[00:28] <Eugenia> allos?
[00:28] <lypanov> no then i guess :)
[00:28] <lypanov> cachekernel / freedows
[00:28] <lypanov> must have been another eugenia how very confusing :)
[00:28] <Eugenia> ah, I know of freedows but I didn't really specifically "liked" it :)
[00:28] <lypanov> hehe :)
[00:28] <lypanov> i remember chatting to a eugenia ages back on irc
[00:29] <lypanov> and always wondered afterwards if it was actually you :)
[00:29] <Eugenia> hmm, possibly not
[00:29] <lypanov> :)
[00:29] <Eugenia> when was that?
[00:29] <lypanov> um.. 99 or so i guess ;)
[00:29] <Eugenia> hmm
[00:29] <lypanov> it was funny
[00:29] <lypanov> rik van riel was also on the project
[00:30] <lypanov> i wonder what happened to ramon...
[00:30] <Company> mxpxpod: what flac?
[00:30] Action: mathrick is back to work on his *C++* project
[00:30] Eugenia ([email protected]) left irc: Client Quit
[00:30] <lypanov> mathrick: good luck. i feel the pain :)
[00:30] Nick change: mathrick -> mathrick|code
[00:31] <mathrick|code> lypanov: thx :)
[00:31] <lypanov> :)
[00:32] <Company> iain's modpoints are so used now
[00:32] mxcl ([email protected]) joined #gstreamer.
[00:34] <lypanov> hey Company btw
[00:34] <Markey-> hey mxcl 
[00:34] <mxcl> Markey-: sup?
[00:34] <lypanov> mxcl -> amarok devel right?
[00:34] <mxcl> lypanov: that's me
[00:34] <lypanov> ah
[00:34] <lypanov> yoyo mxcl then :)
[00:34] <mxcl> heh, greets
[00:35] <sxpert> Gstreamer hits slashdot (or the other way around ?)
[00:36] Action: Misirlou moderates sxpert as -1 confusing
[00:36] Action: sxpert shaves off standard floppy support from carpc kernel... 3 seconds saved, at least
[00:37] <sxpert> Misirlou: well, the fact that a gstreamer story hits the main slashdot page will surely slashdot the gstreamer pages
[00:37] <sxpert> hence the sentence above ;D
[00:37] <Misirlou> hmm
[00:38] <Misirlou> I never thought about it that way.
[00:38] Action: Misirlou moderates sxpert as +2 riddling
[00:38] <Company> gst is hosted on sf
[00:38] <sxpert> hehehe
[00:38] <Company> so it's -1 owned
[00:38] <sxpert> lol
[00:39] <lypanov> hehe
[00:40] <iain> well, at least the /. effect will drain money from themselves
[00:40] <sxpert> lol
[00:41] <iain> if we'd have been using the freedesktop pages then I might have felt bad
[00:41] <Misirlou> ha ha
[00:43] Action: ds-work changes the video.h framerate to [1.0,MAX]
[00:43] <sxpert> ok, tremor should be now compiled in when available 
[00:44] <Zeenix> ds-work: how do i specify multiple formats of RGB in my pad_template
[00:44] <ds-work> Zeenix: use ';'
[00:45] <ds-work> Zeenix: for example, "video/x-raw-rgb; video/x-raw-yuv"
[00:46] <Zeenix> ds-work: should i be using strings or the #defines in gst-libs/video.h ?
[00:47] <ds-work> Zeenix: the defines _are_ strings
[00:50] Markey- ([email protected]) left irc: "g'nite"
[00:50] <Zeenix> ds-work: I know but i could'nt find a way to concat them.. e.g look at this code to see how stupid i am:  GST_STATIC_CAPS (GST_VIDEO_CAPS_RGB) ";"  GST_STATIC_CAPS (GST_VIDEO_CAPS_RGBx)
[00:52] <ds-work> Zeenix: you need to concatinate _strings_, not caps
[00:52] <Zeenix> ds-work: yeah debugged my stupidity
[00:52] <Zeenix> :)
[01:02] Action: mxpxpod wants to point out that Company is the man
[01:06] yf12s ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[01:22] thread ([email protected]) joined #gstreamer.
[01:22] Action: thread just finished reading the osnews article
[01:23] <thomasvs> lmjohns3: dude, a fine time you pick to get back
[01:23] <thread> ...way nifty stuff. I didn't even know you guys existed :)
[01:23] <ds-work> we don't, actually
[01:23] <thread> heh
[01:23] <ds-work> this channel is really just a bunch of bots
[01:23] <lypanov> hehe
[01:23] <ds-work> the real developers are off working
[01:24] <Misirlou> We just sit around on IRC and eat glass in real life.
[01:24] <mxpxpod> you guys eat glass?
[01:24] Action: lypanov prefers bricks
[01:24] <thread> <-- paint chips
[01:24] <ds-work> only safety glass for me
[01:24] <Misirlou> mxpxpod: You don't?
[01:24] <mxpxpod> I've been eating nails this whole time hoping that would make me a gst devel
[01:24] <mxpxpod> :(
[01:25] <Misirlou> Sorry.
[01:25] <Misirlou> Maybe we could make an exceptio.
[01:25] <Misirlou> n
[01:25] <thread> it's just too bad I use KDE over gnome
[01:25] deadchip ([email protected]) joined #gstreamer.
[01:25] <deadchip> is there a callback with a GstThread for notification when it's done playing?
[01:25] Action: mathrick|code notes gst is *not* GNOME specific
[01:25] <thomasvs> ds-work: pokey
[01:25] <mathrick|code> hi deadchip :)
[01:25] Action: thomasvs hopes ds-work bots back
[01:25] <thread> mathrick|code: right.. I gathered that
[01:25] <deadchip> hey mathrick|code :))
[01:26] <ds-work> I AM NOT ZIPPY THE PINHEAD!  GET IT STRAIGHT!
[01:26] <ds-work> um, et
[01:26] <ds-work> er
[01:26] <deadchip> is the "shutdown" signal it?
[01:26] <iain> ds-work: currency GBP 456.34 NOK
[01:26] <Company> deadchip: ste change?
[01:26] <thomasvs> deadchip: the state-change signal should work
[01:26] <thomasvs> ds-work: that sounds about right
[01:26] Action: Company can't write state
[01:26] <lmjohns3> thomasvs: hey there
[01:26] <deadchip> ok thanks
[01:27] Action: deadchip checks
[01:27] <thomasvs> Company: the state police is after you ?
[01:27] <deadchip> hmm i can't find it in the docs
[01:28] <thomasvs> deadchip: gst-inspect-0.7 thread
[01:28] foser ([email protected]) left irc: "[ I want to believe ]"
[01:29] <mxpxpod> Company: awesome! it works!
[01:29] <deadchip> bah
[01:29] <deadchip> for one i was reading wrong docs
[01:29] <deadchip> and then dist-upgrade reverted to 0.6
[01:29] <mxpxpod> Company: but now I'm getting an error using esdsink
[01:29] <ds-work> iain: .no and .uk should start using euros
[01:30] <sxpert> ds-work: they will, when the USD will be sufficiently broken :D
[01:30] <iain> ds-work: .no isn't a member of the EU
[01:30] Action: mathrick|code is seriously annoyed by docs randomly (dis)appearing in Debian's Devhelp
[01:30] <sxpert> iain: yet :D
[01:30] <mxpxpod> RUNNING pipeline
[01:30] <mxpxpod> ERROR: /pipeline0/esdsink0: not negotiated
[01:30] <mxpxpod> ERROR       scheduler(17282) gstoptimalscheduler.c(2083):gst_opt_scheduler_iterate: [GstOptScheduler@0x10020510] in error state
[01:30] <mxpxpod> execution ended after 3 iterations (sum 4220000 ns, average 1406666 ns, min 94000 ns, max 2804000 ns)
[01:32] <deadchip> mathrick|code: devhelp is a little weird yes
[01:33] <Company> mxpxpod: start esd first?
[01:33] wheels ([email protected]) left irc: Read error: 60 (Operation timed out)
[01:33] <iain> well, its good to see that there was one comment of worth on the /. gstreamer article
[01:33] <Company> there was?
[01:33] <Company> did you post something?
[01:34] <sxpert> lol
[01:34] <deadchip> it's stolen from osnews again though
[01:35] <iain> http://developers.slashdot.org/comments.pl?sid=92710&cid=7967870
[01:35] <iain> that one
[01:36] thaytan ([email protected]) left irc: Remote closed the connection
[01:38] hadley ([email protected]) joined #gstreamer.
[01:39] Zeenix ([email protected]) left irc: "The whole problem with the world is that fools and fanatics are always so certain of themselves, but wiser people are so full"
[01:39] <deadchip> which argument is the state? it doesn't seem to pass a GstElementState
[01:40] thaytan ([email protected]) joined #gstreamer.
[01:40] <deadchip> ok i guess i can query the thread..
[01:40] wheels ([email protected]) joined #gstreamer.
[01:41] <thomasvs> deadchip:   g_signal_emit (G_OBJECT (element), gst_element_signals[STATE_CHANGE],
[01:41] <thomasvs>                  0, old_state, GST_STATE (element));
[01:41] <thomasvs> so it looks like it emits both old and new GstElementState
[01:41] <Company> iain: yeah, but we solved that one (both in the actual code and with GstPlay)
[01:42] <deadchip> thomasvs: thanks
[01:42] <Company> iain: though we didn't solve it yet on the plugin front, but we're working on it
[01:42] <Company> videofilter, audiofilter and other superclasses
[01:43] <iain> no, thats what he's complaining about
[01:43] <iain> we've abstracted everything away
[01:43] Nick change: thomasvs -> thomasvz
[01:43] <iain> he claims you need to access the intimate details of plugins to be able to build anything useful
[01:44] <Company> which is not true
[01:44] <iain> well, I know that and you know that, but he doesn't seem to
[01:45] <mxpxpod> Company: esd is running
[01:45] <deadchip> does a stopped GstThread jump to GST_STATE_READY ?
[01:45] <Company> nope, state changes must be done explicitly
[01:46] <deadchip> ok
[01:46] <Company> only if it's done playing on EOS it jumps to PAUSED
[01:46] <deadchip> that was my question, thanks :)
[01:47] <lmjohns3> i'm off, posting my midi stuff to the devel list. cheers !
[01:47] lmjohns3 ([email protected]) left irc: "leaving"
[01:52] <Company> people in videos speak when they open their mouths again
[01:53] <mxpxpod> Company: if I just have rhythmbox 0.7 installed, will it suffix the executables with 0.7?
[01:54] <deadchip> did autoplugging change from 0.6 to 0.7?
[01:54] Nick change: Jaramir -> Jara[zZ]
[01:54] <Company> mxpxpod: rhythmbox will not suffix its executables, you need to use a different prefix there
[01:54] <Company> deadchip: what do you mean "change"?
[01:54] <mxpxpod> Company: I mean gstreamer... sorry
[01:55] <deadchip> Company: it seems there is no type GstAutoplug anymore
[01:55] <Company> mxpxpod: yes, gstreamer suffixes all its executables
[01:55] <mxpxpod> Company: ok
[01:55] <Company> deadchip: yes, that was removed - autoplugging is done by elements only
[01:55] <mxpxpod> Company: btw, I just closed the bug
[01:55] <deadchip> ok
[01:55] Nick change: jdahlin_ -> jdahlin
[01:55] Action: Company is actually going to work on autopluggers now
[01:56] <Company> i think
[01:56] <mxpxpod> heh
[01:56] <deadchip> are there any updated docs as on how to use it?
[01:56] <Company> but it doesn't mean i'm done tomorrow ;)
[01:56] <Company> deadchip: use the spider element and it autoplugs for you
[01:56] <deadchip> mm maybe my question is not appropriate
[01:56] <deadchip> Company: ok
[01:56] <Company> most of the time it autoplugs at least :)
[01:57] <deadchip> :)
[02:02] je4d ([email protected]) joined #gstreamer.
[02:06] pb_ ([email protected]) left irc: "Client exiting"
[02:06] Misirlou ([email protected]) left irc: "asdf"
[02:07] lypanov ([email protected]) left irc: "leaving"
[02:10] csimpson ([email protected]) joined #gstreamer.
[02:11] smoke ([email protected]) left irc: 
[02:13] trow ([email protected]) left irc: "Leaving"
[02:17] <csimpson> exit
[02:17] <csimpson> grr...sorry, wrong window
[02:17] csimpson ([email protected]) left #gstreamer ("I'm out.").
[02:21] Action: wheels is surprised by the amount that the discussion on /. is about Gst + KDE.
[02:21] mxcl ([email protected]) left irc: "leaving"
[02:23] <Company> FOSS discussions on /. always end up with KDE
[02:23] <Company> or MS
[02:23] <deadchip> is typefind borqued in 0.7?
[02:24] <deadchip> sorry in case my question's redundant
[02:25] <wheels> Company: KDE people say that they always end up being about GNOME.  :-)
[02:25] <Company> wheels: that's the same statement
[02:26] <ds-work> deadchip: no
[02:26] <Company> ds-work: help!
[02:26] <Company> ds-work: how can a queue contain buffers while not being negotiated?
[02:26] <deadchip> ok prob pebkac on my side
[02:27] <Company> #0  g_log (log_domain=0x40096a00 "GStreamer", log_level=G_LOG_LEVEL_CRITICAL, format=0x4009a800 "file %s: line %d (%s): assertion `%s' failed") at gmessages.c:540
[02:27] <Company> #1  0x40043e2c in gst_caps_copy (caps=0x0) at gstcaps.c:165
[02:27] <Company> #2  0x4006f2fa in gst_queue_getcaps (pad=0x81ac298) at gstqueue.c:326
[02:27] <ds-work> Company: huh?  I fixed that
[02:27] meck ([email protected]) left irc: "Leaving"
[02:27] <Company> gst-launch-0.7 /mnt/windows-alt/Temp/Pixar\ -\ Knick\ Knack.mpg ! mpegdemux ! mpeg2dec ! { queue ! xvimagesink } mpegdemux0. ! mad ! { queue ! alsasink device=hw:1 }
[02:27] <Company> (simulates spider, that had the same error
[02:27] Action: ds-work hrmmms...
[02:28] <ds-work> it's probably about time to check that pads are negotiated before moving data
[02:29] iain ([email protected]) left irc: "takk"
[02:30] <mxpxpod> hmm...
[02:30] <mxpxpod> I think the optimalscheduler is messed up in cvs
[02:30] <Company> ds-work: could that thingy in the queue be an event?
[02:30] <ds-work> Company: dunno
[02:30] <ds-work> Company: how do I get queue to always store buffers?
[02:31] <ds-work> i want to force it to hold buffers, to check that it won't renegotiate
[02:31] <Company> ds-work: you mean "don't give out last buffer" ?
[02:31] <ds-work> yeah
[02:31] <Company> ds-work: no idea
[02:32] <ds-work> min-treshold-buffers doesn't seem to work
[02:33] <Company> ds-work: events may pass before negotiation
[02:33] <Company> ds-work: and the queue indeed contains an event
[02:33] <Company> ds-work: i'll fix it
[02:36] Action: ds-work turns off the annoying -Woptions that he turned on yesterday
[02:36] Mithrandir_ ([email protected]) joined #gstreamer.
[02:38] Miko5881 ([email protected]) joined #gstreamer.
[02:40] <Company> what does that do?
[02:41] <ds-work> there were 3 or 4
[02:41] <ds-work> that were just annoying
[02:41] <ds-work> -Wall -Wstrict-prototypes -Wwrite-strings \
[02:41] <ds-work> +       -Wsign-compare -Wcast-align
[02:43] Action: ds-work does a s/treshold/threshold/g on gstqueue.c
[02:48] <ds-work> Company: do you think we should call g_object_notify() when queue changes the number of buffers in the queue?
[02:49] <ds-work> current-level-buffers property
[02:49] <taaz> doesn't that happen alot?
[02:49] <ds-work> every time a buffer passes :)
[02:49] <ds-work> twice
[02:49] <ds-work> from different threads :)
[02:49] mxpxpod ([email protected]) left irc: "Guns kill people as much as spoons made Rosie O'Donnell fat.."
[02:50] <taaz> how about this... write a performance test first ;)
[02:50] <taaz> or just use lat.c and check before/after stats
[02:50] <taaz> hrm.. i guess that code doesn't use queues though
[02:51] <Company> ds-work: how about "remove the properties" - especially because i'm strictly against signals from different threads?
[02:51] <ds-work> I thought you had that fixed :)
[02:52] <Company> not the deadlocks
[02:53] Action: Company finds a bug in gstplay
[02:53] Action: ds-work notices that 'videotestsrc sync=false ! { queue ! ximagesink }' is evil
[02:54] <ds-work>  /pipeline0/thread0/queue0: current-level-buffers = 250
[02:54] <wheels> Hmm, first update since the move to Freedesktop: cvs server: cannot find module `common' - ignored
[02:55] Miko5881 ([email protected]) left irc: Remote closed the connection
[02:55] <deadchip> i'm getting this error using spider:
[02:55] <deadchip> ERROR       scheduler( 9387) gstoptimalscheduler.c(2080):gst_opt_scheduler_iterate: [GstOptScheduler@0x8068448] in error state
[02:55] Miko5881 ([email protected]) joined #gstreamer.
[02:56] <Company> ds-work: yeah, i haven't changed the default max thresholds yet
[02:56] <deadchip> what could be wrong there?
[02:57] <Company> ds-work: i haven't found a satisfying value for those that doesn't make video stutter
[02:57] <ds-work> it should probably be time-based
[02:59] <ds-work> if you have both audio and video queues, they probably compete, too


-------------------------------------------------------
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html