IRC Logs

*******************************************************************
[03:02] Rotty ([email protected]) left irc: Remote closed the connection
[03:22] sublett ([email protected]) left irc: "I like food, food is good!"
[03:42] Company ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[03:44] Company ([email protected]) joined #gstreamer.
[03:45] mxpxpod ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[03:55] walters ([email protected]) joined #gstreamer.
[04:11] somex1 ([email protected]) joined #gstreamer.
[04:24] hadley ([email protected]) joined #gstreamer.
[04:30] jimmy_dean ([email protected]) joined #gstreamer.
[04:31] <sri> y0 gstreamer babies, whassup?
[04:32] <Misirlou> alright, you fuckin derelicts, where the FUCK is the documentation, huH? HUH?!
[04:32] <jimmy_dean> lol
[04:33] <Misirlou> (sri told everyone to say it)
[04:34] Action: sri laughs.
[04:34] <sri> uh oh.
[04:34] <sri>  /kick Misirlou
[04:34] <sri> :)
[04:35] <Company> gstreamer docs are in http://freedesktop.org/cgi-bin/viewcvs.cgi/gstreamer/ - they're written in what other people call C
[04:36] <Misirlou> no no
[04:36] <Misirlou> sorry to cause a fuss
[04:36] <Company> :p
[04:36] <Misirlou> I was just trying to get a select few to laugh
[04:36] <Misirlou> no offense was meant, honest
[04:36] <Company> i know
[04:36] <sri> I laughed thats all that counts.
[04:36] <Company> a real newbie would have gotten the "no time to write, lazy, ..." answer that you all know from when you wwere new ;)
[04:37] <sri> I already said company was my favorite guy. :)
[04:37] <Misirlou> :)
[04:37] <jimmy_dean> ahh, good times
[04:42] harshy ([email protected]) left irc: "I quit for now"
[04:46] <sri> this place is boring for an multimedia framework
[04:46] <sri> where's the damn disco ball?
[04:47] <sri> where's the excitement man?
[04:47] Nick change: Jaramir -> Jara[zZ]
[04:47] Action: jimmy_dean hangs up a shiny new disco ball, pops in a CD
[04:49] <jimmy_dean> Company: do you have any experience loading a LADSPA plugin with GStreamer or know who might have that experience that might be able to assist me?
[04:52] <Company> nope
[04:53] <Company> and if there's someone playing with ladspa plugins it would be wingo, but he's rarely in here
[04:53] <jimmy_dean> Company: alright, maybe taaz would...he's the one who suggested it to me in the first place
[04:53] <jimmy_dean> Company: the author of the LADSPA equalizer told me I should either just use his plugin or make a more efficient, easier design bandpass based equalizer
[05:07] <taaz> i haven't tried ladspa plugins for over a year
[05:10] <jimmy_dean> taaz: in general though might it be the same now?  was some kind of data conversion needed?
[05:27] <walters> any gst hackers awake who could approve a patch?
[05:30] <taaz> jimmy_dean: no idea
[05:30] <taaz> walters: i'd look but what do i know about gst anymore ;)
[05:31] Miko5881 ([email protected]) left irc: Remote closed the connection
[05:31] <walters> heh
[05:32] <walters> taaz: i was pretty surprised how little i knew (and still don't know, really), after ds committed the caps changes
[05:32] <walters> and everything stopped working, and i tried to fix it myself :)
[05:32] aldug ([email protected]) joined #gstreamer.
[05:32] <taaz> all voodoo i say
[05:33] <taaz> from afar it does appear that gst is heading the right direction though
[05:33] <taaz> mainly due to a number of talented hackers working on it now
[05:34] <taaz> after omega and wtay vanished i feared it would die off
[05:34] <walters> well i'll post this patch anyways: http://web.verbum.org/files/gst-error.patch
[05:35] Action: taaz <- no clue about new error system
[05:38] <Company> walters: you should ask thomasvs to be sure, but I think it's ok
[05:38] <Company> and yes, we need a third core hacker - if only one that has the knowledge and provides ideas
[05:39] <sri> Company: what about the one that thomasvs is hiring?
[05:39] <sri> this cream of tequila thing has a wierd taste.
[05:39] <Company> i have no clue if that is a talented hacker...
[05:40] <Company> and he's hired, so he'll not have his heart in it and focus on getting payed
[05:40] <walters> Company: what do i return if i encounter an error in my _get function?
[05:40] <walters> Company: _EVENT_INTERRUPT?
[05:40] <Company> walters: NULL
[05:40] <walters> Company: hm, ok.
[05:40] <taaz> who's hired for what?
[05:40] Action: walters is hacking a bit more on iradiosrc
[05:41] <Company> walters: interrupt events are only inserted by schedulers to break loops in loopbased elements
[05:41] <walters> Company: ok.
[05:41] <sri> taaz: hired for gstraemer
[05:42] <walters> Company: i just noticed that's what gst_pad_pull does
[05:42] <sri> Company: unless he's a gstreamer hacker already?
[05:42] <Company> sri: yeah
[05:42] <taaz> sri: who?
[05:42] <Company> sri: but who'd that be?
[05:43] <sri> Company: dunno..I can't imagine thomasvs hiring someone off the street to work on gstreamer.
[05:43] <sri> Company: who isn't as at least committed as he was to gstreamer.
[05:44] <Company> sri: i don't know of anyone of the current hackers he could hire
[05:44] <Company> sri: and i don't think wtay or omega want to do that job either
[05:44] <Company> and then you'Re left with off street
[05:44] <sri> Company: okay..well we'll see.
[05:46] <Company> you sound like someone who knows something that he doesn't wanna tell...
[05:51] mathrick|sleep ([email protected]) left irc: Remote closed the connection
[05:51] ds-work_ ([email protected]) joined #gstreamer.
[05:51] alley_cat ([email protected]) left irc: "May the Source be with you!"
[05:51] <sri> me?
[05:51] <sri> not at all.
[05:52] <sri> and I'm not being coy either.  just what I read on thomas's blog.  I hardly know thomas, he certainly wouldn't be telling me anything> :)
[05:54] <ds-work_> Company: could you commit gettext.patch in gst-autoplug?
[05:55] Action: ds-work_ wonders why anyone would bother keeping it a secret
[05:55] Nick change: ds-work_ -> ds-gromit
[05:56] <sri> ds-work: indeed.
[05:59] <walters> hmm.
[05:59] <walters> how can i format a ssize_t?
[05:59] <walters> i guess i could just cast it to an int and use %d
[06:00] <walters> pretty doubtful it'll be bigger than an int anyways
[06:03] <Company> ds-work: done
[06:04] <Company> gnight
[06:07] ChrisHJW ([email protected]) joined #gstreamer.
[06:21] ChriHJW_log ([email protected]) joined #gstreamer.
[06:25] harshy ([email protected]) joined #gstreamer.
[06:40] ChrisHJW_log ([email protected]) left irc: Connection timed out
[06:40] ChrisHJW ([email protected]) left irc: Connection timed out
[06:53] <bluejay|sleep> walters: %zd
[06:53] Nick change: bluejay|sleep -> bluejay
[06:54] <walters> hm
[06:54] <walters> `z' was introduced in ISO C99.  
[06:54] <bluejay> Anyone know why gst-plugins make is failing in gst-libs/gst/riff?
[06:54] <bluejay> walters: yeah. I don't think there's an earlier portable way.
[07:10] hadley ([email protected]) left irc: "Leaving"
[07:24] kmaraas ([email protected]) left irc: Read error: 110 (Connection timed out)
[08:15] ChrisHJW ([email protected]) joined #gstreamer.
[08:35] Nick change: bluejay -> bluejay|sleep
[09:21] somex1 ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[09:42] Shoragan ([email protected]) joined #gstreamer.
[09:43] kmaraas ([email protected]) joined #gstreamer.
[10:04] kmaraas ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:07] kmaraas ([email protected]) joined #gstreamer.
[10:23] sri ([email protected]) left irc: "Client exiting"
[10:40] zahir ([email protected]) joined #gstreamer.
[10:40] <zahir> hi all
[10:46] zahir ([email protected]) left irc: "byez"
[10:55] markov ([email protected]) joined #gstreamer.
[11:15] apoc ([email protected]) joined #gstreamer.
[11:49] hallibaby ([email protected]) joined #gstreamer.
[12:05] Shoragan ([email protected]) left irc: "Leaving"
[12:17] dolphy ([email protected]) joined #gstreamer.
[12:17] markov ([email protected]) left irc: "bbl"
[12:18] <dolphy> morning
[12:21] kmaraas ([email protected]) left irc: Read error: 110 (Connection timed out)
[12:24] Action: dolphy is trying to debug opt
[12:24] <apoc> morning dolphy
[12:27] walters ([email protected]) left irc: Remote closed the connection
[12:39] <dolphy> anybody knows a little about opt scheduling ?
[12:40] iain ([email protected]) joined #gstreamer.
[12:45] <dolphy> iain ! :)
[12:45] <dolphy> iain: i was hoping to see you
[12:48] <iain> ou oh
[12:48] <iain> what've I broken now?
[12:49] The_Company ([email protected]) joined #gstreamer.
[12:50] <dolphy> iain: i don't think that's you :)
[12:50] <dolphy> iain: http://getcontrol.sourceforge.net/files/gstplay_pipeline.png
[12:50] <dolphy> iain: look at that gstplay pipeline
[12:50] <dolphy> iain: switch is loop based, identity is chain based, visualization element is chain based
[12:51] <dolphy> iain: when playing that pipeline no scheduler is able to handle that situation
[12:51] <dolphy> iain: and opt tells :
[12:51] <iain> wow, thats a nice pipeline
[12:51] <dolphy> in this case, the group was running and we wanted to swtich to it,
[12:51] <dolphy>        * this is not allowed in the optimal scheduler (yet)
[12:53] <iain> I wouldn't know what its needing :(
[12:53] <dolphy> iain: hmm
[12:53] <dolphy> iain: basicgthread is not able to schedule that neither anyway
[12:54] <iain> so, opt has detected a deadlock...
[12:54] Shoragan ([email protected]) joined #gstreamer.
[12:54] <dolphy> iain: yeah i don't know if i get the picture clearly
[12:55] <dolphy> iain: but seems opt doesn't want to switch to a group that is already running
[12:55] <iain> oh, it might be because both the two chain based things are in the same "group" as each other
[12:55] kmaraas ([email protected]) joined #gstreamer.
[12:55] <dolphy> iain: so switch is not able to pull from 2 elements being in the same group
[12:55] <iain> because its being called when the loop element is trying to get data from its sinks
[12:55] <dolphy> iain: exactly
[12:55] <iain> dolphy: possibly, yes
[12:56] <dolphy> iain: how does opt decides what element is in what group
[12:56] <dolphy> iain: because for example 2 videotestsrc connected to switch in a bin works just nicely
[12:56] mathrick ([email protected]) joined #gstreamer.
[12:56] <mathrick> re
[12:57] <dolphy> mathrick: hi
[12:58] <iain> dolphy: docs/random/wtay/scheduling-ideas explains how the groups are organised
[12:58] <iain> dolphy: maybe if you put visualisation_bin as a thread?
[12:58] <iain> vis_bin sorry
[12:58] <dolphy> iain: but then i would have to add a queue.. 
[12:59] <iain> hmm, true
[12:59] <dolphy> get-based plugins are put in the same group as a peer loop based element
[12:59] <dolphy> so 2 videotestsrc and switch are in the same group 
[13:00] <dolphy> and switch will pull from 2 elements in the same group
[13:02] <iain> I guess?
[13:04] <dolphy> maybe setting identity to be loop based wil split the group
[13:04] <dolphy> iain: right ?
[13:04] <iain> it might
[13:07] <mathrick> dolphy: managed to get that projector working?
[13:09] Company ([email protected]) left irc: Read error: 110 (Connection timed out)
[13:09] <dolphy> mathrick: not to the wanted resolution :)
[13:09] <dolphy> mathrick: maybe it simply can't
[13:09] <dolphy> mathrick: so it's running 1024x768
[13:09] <dolphy> iain: nope that doesn't solve the problem
[13:10] <iain> "opt is a sea of mysteries"
[13:11] <mathrick> dolphy: you tried doing it @ some crappy refresh rate?
[13:11] sublett ([email protected]) joined #gstreamer.
[13:12] <mathrick> dolphy: it may be that it "works" in 1280, but with eye-poking other params
[13:14] <dolphy> mathrick: well yeah but i tried a lot
[13:16] <dolphy> mathrick: with no success
[13:16] Action: mathrick wonders what is the rule behind button placement on GNOME Taskbar
[13:17] <mathrick> dolphy: hmm, maybe DDC was right, it can't do it
[13:17] <mathrick> dolphy: but then docs lie
[13:17] <dolphy> yup
[13:20] Rotty ([email protected]) joined #gstreamer.
[13:21] <dolphy> iain: well making the vis_bin a thread makes it work
[13:21] <dolphy> iain: :-/
[13:23] <iain> dolphy: but you don't like having a queue?
[13:28] hallibaby ([email protected]) left irc: "Bye bye"
[13:33] hallibaby ([email protected]) joined #gstreamer.
[13:33] alley_cat ([email protected]) joined #gstreamer.
[13:35] <dolphy> iain: well the less threads there are the better i feel
[13:37] LeRoutier ([email protected]) joined #gstreamer.
[13:38] <iain> dolphy: true
[13:38] <LeRoutier> hello
[13:40] jdahlin ([email protected]) joined #gstreamer.
[13:41] <mathrick> hi LeRoutier
[13:41] <mathrick> yo jdahlin, hallibaby, alley_cat
[13:42] Action: alley_cat *yawns*
[13:43] <LeRoutier> found a dupe to close : http://bugzilla.gnome.org/show_bug.cgi?id=132328
[13:44] <LeRoutier> and this one was fixed yesterday : http://bugzilla.gnome.org/show_bug.cgi?id=131932
[13:57] <dolphy> requested pads are never released
[14:04] <LeRoutier> hum, was wrong for second bug. fixed in avidemux, not asfdemux
[14:08] <dolphy> i just marked 132328 as duplicate
[14:08] <dolphy> the other one does not seem fixed
[14:09] <LeRoutier> yep. it's the same compilation problem but not in the same file
[14:13] apoc ([email protected]) left irc: "Leaving"
[14:18] <LeRoutier> dolphy, who's the maintainer of gst-launch-gui ?
[14:21] TheCompany ([email protected]) joined #gstreamer.
[14:22] <LeRoutier> hello Company
[14:25] Action: mathrick notices lots of Companies around #gstreamer
[14:26] <mathrick> does that mean gst is becoming industry standard? :)
[14:26] <LeRoutier> yep, saw he was already there but too late 
[14:27] <mathrick> LeRoutier: probably his other self won't be around too long, I expect it to timeout soon
[14:28] <LeRoutier>  
[14:28] <LeRoutier> avi no more crashes gst-player for me. but only have sound, funky colors on screen (but no video) and tons of similar error messages on the console
[14:29] <LeRoutier> WARNING **: Index entry XXX has invalid stream nr YYY
[14:29] <mathrick> LeRoutier: what video sink do you have set in gconf?
[14:29] <LeRoutier> xvimagesink
[14:30] <LeRoutier> updating my bug report as symptoms have changed
[14:30] <dolphy> LeRoutier: i think ronald is currently working on avi
[14:32] <LeRoutier> ok
[14:32] The_Company ([email protected]) left irc: Read error: 60 (Operation timed out)
[14:33] <LeRoutier> bug updated : http://bugzilla.gnome.org/show_bug.cgi?id=131050
[14:44] <LeRoutier> dolphy, in gst-player, i have : 
[14:44] <LeRoutier> CRITICAL **: file gstfilesrc.c: line 710 (gst_filesrc_check_filesize): assertion `GST_FLAG_IS_SET (src ,GST_FILESRC_OPEN)' failed
[14:45] <LeRoutier> each time i press play. to play a wav file
[14:47] <LeRoutier> just closed : http://bugzilla.gnome.org/show_bug.cgi?id=131331   as playing this wav file now work in gst-player
[14:57] <LeRoutier> dolphy: seems .asf files (wma & wmv) duration is badly calculated for gst-player slider bar. usually, it says it ends around 8 or 10s even for audios/videos of several minutes. 
[14:59] <kmaraas> is gstreamer CVS broken at the moment?
[14:59] <kmaraas> I get lots of "Could not find desired version x.xx of <file>"
[15:00] <LeRoutier> hum, trying it
[15:00] <kmaraas> ok, just one it seems
[15:01] <LeRoutier> during checkout i suppose ? as i only did an update and it worked well
[15:01] <kmaraas> I get a lot of changes every time I do cvs update
[15:02] <LeRoutier> remove your local repo and re-do a checkout. sometimes local repo get corrupted
[15:02] <kmaraas> hmm, maybe my checkout was just way broken
[15:02] <kmaraas> yeah
[15:09] markov ([email protected]) joined #gstreamer.
[15:13] <mathrick> hi markov
[15:14] Nick change: markov -> afk
[15:20] <LeRoutier> bbl
[15:20] LeRoutier ([email protected]) left irc: "Leaving"
[15:34] Uraeus ([email protected]) joined #gstreamer.
[15:37] sublett ([email protected]) left irc: "I like food, food is good!"
[15:37] <Uraeus> ello
[15:38] <mathrick> yo Uraeus
[15:44] Nick change: TheCompany -> Company
[15:59] <Company> question for the C gurus (do we have any here?):
[16:00] <Company> if i don't use __attribute__ ((packed)), are 64bit machines free to expand guint32 struct fields to 64 bits?
[16:14] Nick change: afk -> re
[16:16] Nick change: re -> markey
[16:17] <mathrick> Company: yes. AFAIK the only guanrantee ISO C does is that: struct { int i; } x; x.i = 10; int i = (int)x; is correct and i == 10. Otherwise compilers are free to do any padding they wish. And FWIW, I don't think that ((packed)) is _guaranteed_ to work, it's just way of kindly asking compiler to do it.
[16:19] _axel ([email protected]) joined #gstreamer.
[16:22] <_axel> i know this sounds ridiculous, but im getting a total hardware crash when running a gstreamer hello world app
[16:22] <_axel> seems related to the soundcard but i just find it all extremely odd
[16:24] bitshifter ([email protected]) joined #gstreamer.
[16:26] foser ([email protected]) joined #gstreamer.
[16:39] bitshifter ([email protected]) left irc: Remote closed the connection
[16:41] <mathrick> Company: nice one about 64-bit cleanliness: http://docs.sun.com/db/doc/806-0477/6j9r2e2b2?a=view
[16:50] bitshifter ([email protected]) joined #gstreamer.
[16:50] sublett ([email protected]) joined #gstreamer.
[16:57] <bitshifter> Hi everyone. I'm having problems with gstreamer 0.6.4 on debian sid and alsa. It seems that not even 'gst-launch fakesrc num_buffers=3 ! fakesink' results in any output similar to that in the FAQ (to be precise: no output between 'Running pipeline' and 'execution ended'). Any ideas?
[17:00] The_Company ([email protected]) joined #gstreamer.
[17:02] <dolphy> hmm this visualization thing in gstplay is driving me completely crazy
[17:06] <mathrick> The_Company: got my brilliant answers?
[17:07] Nick change: mathrick -> mathrick|food
[17:11] alley_cat ([email protected]) left irc: "May the Source be with you!"
[17:13] alley_cat ([email protected]) joined #gstreamer.
[17:15] <bitshifter> anyone?
[17:23] Nick change: mathrick|food -> mathrick
[17:24] Company ([email protected]) left irc: Read error: 110 (Connection timed out)
[17:27] <bitshifter> k, so the FAQ docs at http://www.gstreamer.net/docs/cvs/faq/html/chapter-using.html are wrong then?
[17:28] <dolphy> bitshifter: ?
[17:28] <mathrick> bitshifter: looks like it's output of gst-launch -v, i'll check that
[17:28] <bitshifter> mathrick: oh yes, you're right. sorry about that
[17:29] <mathrick> bitshifter: np, there's type there indeed
[17:29] <mathrick> s/type/typo/
[17:29] <bitshifter> I thought something was really really wrong with my setup
[17:29] <mathrick> bitshifter: nope, it just wouldn't make sense to flood users with so much info :)
[17:30] <bitshifter> so any known problems with gst 0.6.4 on debian sid and alsasink?
[17:30] <mathrick> bitshifter: yep
[17:30] <bitshifter> I get this:  ALSA lib pcm_hw.c:370:(snd_pcm_hw_sw_params) SNDRV_PCM_IOCTL_SW_PARAMS failed: Invalid argument
[17:30] <bitshifter> all other apps work fine
[17:30] <mathrick> bitshifter: alsasink is generally considered broken in 0.6
[17:31] <mathrick> bitshifter: although you can try using ` alsasink device="hw:0" '
[17:31] <bitshifter> mathrick: oh, okay. I found a ML msg to that respect, but it was about 0.6.0 and I figured it got fixed by now 
[17:31] <bitshifter> mathrick: I'll try that, thanks
[17:32] <bitshifter> nope, same thing
[17:32] <bitshifter> oh well
[17:33] Action: bitshifter goes looking for NAS sink that he saw somewhere some time ago
[17:33] <mathrick> bitshifter: nassink is 0.7 work
[17:34] <mathrick> bitshifter: if you're trying to start with gst, and don't have specific version constraints, don't bother with 0.6, it's obsolete anyway
[17:35] <_axel> but 0.6 code works in 0.7?
[17:35] Uraeus ([email protected]) left irc: "Client exiting"
[17:37] <bitshifter> mathrick: k, thanks for all the help, I'll have a look at 0.7 then
[17:37] <mathrick> _axel: not necessarily, it is possible to write code that will work in both, but many things have changed, and most advanced code needs to be ported. It's not very big of porting, but still
[17:37] <_axel> i see
[17:38] <_axel> what about a simple pipeline converting .ogg to wav? i just had a struggle writing that this morning... will it work in 0.7?
[17:38] <mathrick> _axel: should, although I'd need to see it
[17:39] <mathrick> _axel: most changes are in programmatic access to gstreamer, although gst-launch syntax has changed (read: got unbroken) a bit too
[17:40] <dolphy> iain: do you think that putting the visualization elements directly in the main bin would help schedulers ?
[17:46] Shoragan ([email protected]) left irc: Read error: 60 (Operation timed out)
[17:49] <dolphy> The_Company: ping ?
[17:51] Shoragan ([email protected]) joined #gstreamer.
[17:55] jimmy_dean ([email protected]) left irc: "Carrier Signal Lost"
[18:00] Nick change: The_Company -> Company
[18:00] <Company> dolphy: pong
[18:01] <mathrick> Company: you seem to have some net problems today
[18:02] <Company> mathrick: yeah, both me and my brother were running file sharing tools, each one using 60% of the available upstream bandwidth ;)
[18:02] <mathrick> Company: hehe :)
[18:02] <mathrick> Company: use HTB, dude
[18:02] <Company> HTB?
[18:02] <mathrick> Company: iproute2 is really nice
[18:03] <mathrick> Company: yep, one of queueing algos in iproute2, works like charm
[18:04] <Company> haven't found a good tutorial about it so i gave up, because all I did didn'T work
[18:04] _axel ([email protected]) left irc: "Leaving"
[18:04] <mathrick> Company: there is very good one, Advanced Traffic Shaping in Linux
[18:04] <Company> problem is that it's of no use unless I also get good pings for other apps
[18:04] <mathrick> s/Shaping/& Howto/
[18:04] bitshifter ([email protected]) left #gstreamer ("(null)").
[18:05] <Company> and that means the DSL modem queue may not be filled
[18:05] <mathrick> Company: that's what HTB is useful for
[18:06] <sublett> Company: http://lartc.org/
[18:06] <mathrick> Company: you set high prio for control packets, and they just get processed before bandwith hogs, but don't hinder them in any way
[18:06] <mathrick> s/with/width/
[18:07] <Company> that still doesn't give me low ping for quake
[18:07] <mathrick> Company: no prob in setting prio for quake packets too :)
[18:07] <mathrick> Company: it just requires writing rule for that
[18:07] <Company> that still doesn't give me low ping because the modem queue is full
[18:07] <sublett> Company: look at the "Wonder Shaper" script at the lartc page...it will give you low ping and all other goodness
[18:08] <mathrick> Company: argh, read my words: it *priority* queue with HTB
[18:08] <mathrick> s/it/it's/
[18:08] <dolphy> Company: autoplugging of mp3 with id3tags really does weird stuff
[18:08] <mathrick> Company: really, take a look at that page sublett gave you
[18:09] <Company> i'm doing atm
[18:11] <dolphy> Company: looks like spider is always trying to connect from x-gst-tags 
[18:11] <dolphy> Company: so it plugs a lot of id3tag elements
[18:11] <mathrick> dolphy: ** (gst-player:9412): WARNING **: xvimagesink was unable to obtain the X11 context.
[18:11] <dolphy> and apparently after doing that it tells "reached EOS"
[18:11] <mathrick> dolphy: when starting gst-player
[18:11] <dolphy> mathrick: you probably have another application using Xv
[18:12] <dolphy> running
[18:12] <mathrick> dolphy: nope
[18:12] <Company> dolphy: hm
[18:12] <dolphy> mathrick: gst-launch-0.7 videotestsrc ! xvimagesink
[18:13] <Company> dolphy: i think i fixed that once...
[18:15] <mathrick> dolphy: hmm, mathrick@megumi:~/Dev/gstreamer/gst-player$ gst-launch-0.7 -v videotestsrc ! xvimagesink
[18:15] <mathrick> RUNNING pipeline ...
[18:15] <mathrick> ERROR: pipeline doesn't want to play.
[18:15] <mathrick> but I'm sure I don't have Xv running
[18:15] <dolphy> Company: well i have latest cvs
[18:16] <dolphy> mathrick: hmmm
[18:16] <dolphy> mathrick: add --gst-debug-level=5 2>/tmp/trace.log and send me that  trace
[18:16] <mathrick> dolphy: as you wish :)
[18:19] <mathrick> dolphy: are you OK with dcc?
[18:19] <dolphy> should work
[18:19] <dolphy> doesn't :)
[18:19] <mathrick> Company: btw, did you get my C guru question answer?
[18:19] <dolphy> julien at moutte dot net
[18:20] <mathrick> dolphy: k
[18:20] <dolphy> and add a xvinfo
[18:20] <dolphy> please
[18:20] <mathrick> dolphy: k
[18:20] <dolphy> and bzip2 it :)
[18:20] <Company> mathrick: yeah
[18:20] <mathrick> dolphy: argh, I'm stupid
[18:20] Action: mathrick has disabled nvidia binary drivers
[18:21] <mathrick> dolphy: tried 2.6 recently ;)
[18:21] Action: dolphy kicks mathrick
[18:23] <dolphy> Company: http://getcontrol.sourceforge.net/files/gstplay_pipeline.png with that pipeline i get scheduling errors with both opt and basicgthread
[18:23] <dolphy> Company: the issue seems to be that switch is loop based and pulls from identity colorspace and visualization colorspace
[18:24] <dolphy> Company: any idea how to have a working situation
[18:24] <dolphy> ?
[18:25] <mathrick> dolphy: anyway, I got lots of that ** (gst-player:9489): WARNING **: Index entry 14158 has invalid stream nr 1311 for avis, and nice SEGV for mpegs
[18:25] <dolphy> mathrick: that's not from gst-player
[18:25] <dolphy> mathrick: i get the same
[18:26] <dolphy> mathrick: probably ronald made some fixes in avidemux
[18:26] <dolphy> mathrick: and something s wrong with indexes
[18:27] <mathrick> dolphy: hmm, right, gst-launch gets them too
[18:27] <mathrick> dolphy: but mpegs segfaulted even before that avi breakage
[18:27] <mathrick> dolphy: or, when they did not, playback was really crappy
[18:27] <dolphy> mathrick: yeah mpeg is broken since a long time :)
[18:27] <Company> dolphy: no idea
[18:28] <mathrick> dolphy: in gst-player only? or in gst as a whole?
[18:28] <dolphy> mathrick: as a whole
[18:28] <mathrick> dolphy: ah, k
[18:29] <dolphy> Company: hmm that's a big issue for me
[18:29] <dolphy> Company: i m completely blocked now
[18:30] <mathrick> dolphy: hum, I didn't even know it's so broken now ;)
[18:31] <Company> mpeg works here, but what do i know ;)
[18:32] <Company> dolphy: if it doesn't work with the schedulers, making more threads will help there ;)
[18:32] <Company> but really, i have no clue about schedulers
[18:32] hallibaby ([email protected]) left irc: Remote closed the connection
[18:33] <dolphy> Company: i thought you masterized basicgthreads at least
[18:33] <dolphy> Company: i tried making the vis bin a thread
[18:33] <dolphy> Company: with 2 queues for in/out/put
[18:33] <dolphy> Company: that completely break negotiation
[18:34] jaiserca ([email protected]) left irc: "Terminando cliente"
[18:34] <dolphy> Company: for whatever reason ffmpeg_mpeg4 refuses link because video window is 16x16
[18:38] Nick change: Jara[zZ] -> Jaramir
[18:39] <Company> dolphy: for starters, switch has to go into the video queue
[18:40] <Company> s/queue/thread/
[18:40] <Company> dolphy: because you'll end up with a circle otherwise and that's always problematic
[18:41] <dolphy> Company: so i should have 2 queues in front of switch and put that in video_thread?
[18:41] <Company> that probably won't work either, because vis fills the queue with 16x16 images before the video decoder figures out the real size
[18:42] <dolphy> Company: hmm
[18:42] Action: Jaramir is away: The data on your hard drive is out of balance.
[18:43] <dolphy> Company: the only thing i want is to have vis connected while i am looking at the identity to know if video signal is coming out of spider
[18:43] <dolphy> Company: and then disconnect vis if identity emits handoff
[18:45] <dolphy> Company: anyway i already tried that design
[18:46] <dolphy> Company: and having the queue being connected to switch with nothing on its sink pad was not working
[18:50] <Company> mathrick: the wondershaper sucks
[18:50] <Company> mathrick: s/sucks/isn't good enough/
[18:52] <dolphy> Company: what about that one http://getcontrol.sourceforge.net/files/gstplay_pipeline.png
[18:53] <Company> dolphy: that should work - at least if every element used in there works correctly
[18:53] <Company> or i haven't found the error yet ;)
[18:54] <dolphy> Company: ok i ll try to implement it
[18:54] <dolphy> Company: that gstplay pipeline is a mind breaker
[18:54] <dolphy> :-/
[18:55] Action: Company knows why he wants to have a way to make it more complex step by step ;)
[18:56] hallibaby ([email protected]) joined #gstreamer.
[19:00] hadley ([email protected]) joined #gstreamer.
[19:02] <Company> I _should_ have a ping of 100, but I get 150
[19:02] <Company> booo
[19:03] <Company> and I need 50
[19:08] alley_cat ([email protected]) left irc: "May the Source be with you!"
[19:22] <mathrick> Company: it probably needs some tweaking. I'm not fluent in using HTB myself, only friend of mine uses it and it (it == HTB) works very well, but I don't know if he uses specifically wondershaper or some other ruleset
[19:22] <mathrick> Company: do you have assymetric DSL?
[19:22] <mathrick> s/assymetric/asymmetric/
[19:24] <mathrick> Company: first and most important thing is to correctly determine your maximum bandwidth in both direction, and never ever allow it to saturate if you want to keep low ping
[19:28] <mathrick> btw, does anyone know why NickServ complains that in *!*[email protected] username is too long?
[19:35] Action: dolphy is making some progress :)
[19:35] <dolphy> gst_element_release_request_pad is a nice friend :)
[19:39] <dolphy> hmm gst_pad_pull is blocking
[19:39] <dolphy> that's annoying
[19:40] <dolphy> Company: how should switch know that it can't pull from a pad that is not throwing data ?
[20:06] The_Company ([email protected]) joined #gstreamer.
[20:07] <mathrick> The_Company: wb, how's going with HTB?
[20:08] <dolphy> The_Company: have you seen my question about pads ?
[20:10] jimmy_dean ([email protected]) joined #gstreamer.
[20:10] <mathrick> lo jimmy_dean
[20:10] Action: mathrick takes a nap
[20:11] <jimmy_dean> mathrick: yo mathrick
[20:11] Nick change: mathrick -> mathrick|nap
[20:11] <jimmy_dean> mathrick|nap: I feel the same way...lazy day
[20:26] Zeenix ([email protected]) joined #gstreamer.
[20:26] <Zeenix> hi
[20:29] Company ([email protected]) left irc: Read error: 110 (Connection timed out)
[20:39] somex1 ([email protected]) joined #gstreamer.
[20:44] Zeenix ([email protected]) left irc: "Leaving"
[20:54] <The_Company> wee, network testing leads to disconnects ;)
[20:55] <The_Company> mathrick|nap: the main problem with the whole thing is that i really can'T make the mtu much larger than 200 and then I have a lot of overhead
[21:00] Nick change: The_Company -> Company
[21:03] walters ([email protected]) joined #gstreamer.
[21:14] sublett ([email protected]) left irc: "I like food, food is good!"
[21:38] ChrisHJW ([email protected]) left irc: 
[21:41] <dolphy> hmm
[21:41] <dolphy> seems i lost my connection
[21:41] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[21:42] dolphy ([email protected]) joined #gstreamer.
[21:42] <dolphy> Is it possible to pull from a pad in non blocking mode ?
[21:42] Nick change: mathrick|nap -> mathrick
[21:42] wheels ([email protected]) left irc: Read error: 113 (No route to host)
[21:43] <mathrick> Company: I don't really understand what problems with MTU you have
[21:43] <mathrick> Company: 200 is awfully small, DSL likes having much larger one, like 1420
[21:44] <mathrick> Company: but anyway, what does it have to do with HTB
[21:44] <mathrick> ?
[21:44] Ermshta ([email protected]) joined #gstreamer.
[21:45] <Ermshta> does totem or gst-player are supposed to work with gstreamer0.7.3?
[21:46] Nick change: Jaramir -> Jara[zZ]
[21:48] <mathrick> Ermshta: AFAICT no, GstPlay used by these two is under heavy work now, and 0.7.3 is very old these days
[21:48] <Ermshta> hmmm ok
[21:49] <Ermshta> so if I "move" all the gst*0.7*.pc [pkgconfig] file to another folder and compile totem.. it should detect gstreamer 0.6.4 and be able to compile with it nah?
[21:53] alley_cat ([email protected]) joined #gstreamer.
[21:54] <mathrick> Ermshta: there's another way, you should set PKG_CONFIG_PATH to dir containing .pkg files
[21:55] <Company> mathrick: the problem is that if you have 128kbps DSL and an average package is 1492 bytes, you need ~100 ms to send that packet
[21:56] <mathrick> Company: hmm, OK, so what is the problem when you set lower MTU ?
[21:56] <Company> mathrick: and that means your ping is ~50ms greater than normal, which is not acceptable for shooters
[21:56] <Company> lower mtu means more overhead
[21:56] <mathrick> Company: k, and something in between, like 800?
[21:58] <mathrick> Company: and, that means even with no other bandwidth users, and no HTB, you get pings too high for your needs, right?
[21:59] wheels ([email protected]) joined #gstreamer.
[21:59] <mathrick> lo wheels
[22:00] <Company> mathrick: no
[22:00] <markey> yoyo skott
[22:00] <Company> mathrick: because if there's no other bandwidth user the chance that there's already a packet in transit is 0
[22:01] <Company> but if you saturate the bandwidth, the chance is 1
[22:01] <mathrick> Company: but you still get 100ms for transmition, right?
[22:02] <mathrick> Company: and purpose of HTB in this case is to get such behavior that for shooter packets there is no other user
[22:03] <Company> mathrick: the only thing that HTB can do is select the next packet
[22:03] <Company> mathrick: it does that pretty good
[22:03] <mathrick> Company: yep, and it bases its choice on priorities you assigne them
[22:03] <Company> mathrick: correct
[22:03] <mathrick> Company: so what you should do is to set highest prio for that packets you want processed first
[22:03] <Company> mathrick: but if there's already a packet in transit, that packet has to be finished first
[22:04] <Company> mathrick: and that's the problem
[22:04] <mathrick> Company: hmm, ic now
[22:05] <mathrick> Company: so you should indeed lower your MTU
[22:05] <mathrick> Company: did values ~800 do the trick?
[22:05] apoc ([email protected]) joined #gstreamer.
[22:05] <mathrick> yo apoc
[22:05] <apoc> hey mathrick
[22:06] <Company> mathrick: MTU/16 is the average added ping
[22:06] <Company> mathrick: so 800 means +50ms
[22:06] zeenix ([email protected]) joined #gstreamer.
[22:07] <Company> mathrick: and i don'T want more than +20ms
[22:07] <mathrick> hiya zeenix
[22:07] <mathrick> Company: hrmpf, that's tough
[22:07] <Company> mathrick: so i'm forced to either go very low with the MTU or to stop downloading
[22:07] <zeenix> hi
[22:07] <mathrick> Company: I don't see a way to remedy that
[22:08] <mathrick> Company: other than lowering max bandwidth of your downloads
[22:08] <zeenix> anyone having nvidia tnt2?
[22:08] <mathrick> zeenix: me
[22:09] <mathrick> Company: this should decrease avg penalty incurred by packet already in transit to acceptable value
[22:09] <zeenix> mathrick: which distro?
[22:09] <mathrick> zeenix: Debian Sid
[22:09] <zeenix> mathrick: Any problem(s) ?
[22:11] <mathrick> zeenix: hmm, like? Only problem I have w/ 4496 is that they aren't what I'd call rock stable
[22:11] <zeenix> mathrick: 4496?
[22:11] <mathrick> zeenix: latest version I know of
[22:12] <mathrick> zeenix: not that new tho
[22:12] <mathrick> zeenix: earlier revs were rock solid, but had serious issues after several hours of working
[22:12] <zeenix> mathrick, latest version of what? card? driver? distro?
[22:13] <mathrick> zeenix: driver revision
[22:13] Ermshta ([email protected]) left irc: Read error: 110 (Connection timed out)
[22:13] <zeenix> mathrick: i think i should try the old version, the new version does'nt seem to work
[22:14] <mathrick> zeenix: not at all?
[22:15] <zeenix> mathrick: you can say that
[22:18] <mathrick> zeenix: previously I used 3123, and they worked OK, apart from that they would start to be choppy after a while - everything would freeze each 5-10s for a little while, very annoying
[22:24] sub_pop ([email protected]) joined #gstreamer.
[22:39] <mathrick> zeenix: ahh, now I realized that I never actually mentioned I'm using NVidia binary drivers :)
[22:45] <apoc> dilinger: ping ?
[22:47] Action: mathrick notices there is new NVidia drivers revision out
[22:50] hadley ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[23:10] somex1 ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[23:22] hadley ([email protected]) joined #gstreamer.
[23:41] ChrisHJW ([email protected]) joined #gstreamer.
[23:44] hallibaby_ ([email protected]) joined #gstreamer.
[23:52] ChrisHJW ([email protected]) left irc: Read error: 60 (Operation timed out)
[23:57] alley_cat ([email protected]) left irc: "May the Source be with you!"
[00:00] --- Sun Jan 25 2004
[00:08] apoc ([email protected]) left irc: "Leaving"
[00:09] hallibaby_ ([email protected]) left irc: "Bye bye"
[00:17] <mathrick> jimmy_dean: ping?
[00:18] war[sheep]` ([email protected]) joined #gstreamer.
[00:18] war[sheep]` ([email protected]) left irc: Remote closed the connection
[00:22] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[00:23] BBB ([email protected]) joined #gstreamer.
[00:24] <mathrick> lo BBB
[00:24] <BBB> howdy
[00:24] <BBB> Company: da?
[00:25] <mathrick> BBB: funky domain you have :)
[00:27] <BBB> :)
[00:28] Nick change: harshy -> harshyOUT
[00:29] <mathrick> BBB: what exactly does that mean?
[00:32] <BBB> doctor's place
[00:34] markey ([email protected]) left irc: "g'nite"
[00:53] Shoragan ([email protected]) left irc: "Leaving"
[00:58] <dilinger> apoc: pong
[00:58] <dilinger> apoc: i just got an ultrasparc; i'll be able to test out MA again on a big endian system :)
[00:59] <mathrick> dilinger: apoc is no longer here :)
[00:59] <dilinger> heh.  right, i knew that. ;)
[00:59] <dilinger> oh well
[01:03] <mathrick> BBB: SDL implements XOverlay?
[01:04] <BBB> yes
[01:05] <mathrick> BBB: i thought that as SDL runs on various display systems, it would also have its own overlay method
[01:05] <mathrick> BBB: so even SDL on DirectFB is still XOverlay?
[01:06] alley_cat ([email protected]) joined #gstreamer.
[01:07] <BBB> mathrick: we don't have a fb iface yet
[01:07] <BBB> and yes, if we had a fb iface, it'd support that too if runnning on that
[01:07] <mathrick> BBB: hmm, strange
[01:07] sub_pop ([email protected]) left irc: "Client exiting"
[01:09] <mathrick> BBB: so SDL isn't display system dependent, but sdlsink is / will be?
[01:10] <BBB> no
[01:10] <BBB> we support xoverlay if it runs on X, we support fboverlay if it runs on dfb
[01:11] <BBB> but we need a way to connect iut to the app
[01:11] <Company> BBB: pong (my answers aren't i18n'd yet)
[01:12] Action: BBB kicks mp2 decoding
[01:12] <BBB> eh, for typefinding
[01:12] <BBB> your 'probability' calculation in mpeg system stream typefinding, is that based on anything real?
[01:13] <BBB> it causes issues here
[01:14] <Company> no it's not
[01:14] <BBB> in streams with video inside a system stream (without header at start), I randomly get video or system mpeg detected
[01:14] <BBB> ok, so I can just change it to max?
[01:15] <Company> you're just supposed to return 100% for "absolutely sure" and below for less ssure :)
[01:15] <BBB> (or max-1, don't care)
[01:15] <Company> dunno, fix it so it works :)
[01:15] alley_cat ([email protected]) left irc: "May the Source be with you!"
[01:16] <BBB> k
[01:23] alley_cat ([email protected]) joined #gstreamer.
[01:26] foser ([email protected]) left irc: "[ I want to believe ]"
[01:26] <BBB> hm
[01:26] <BBB> MPEG-with-spider gives audio but no video
[01:27] <BBB> I thought it was the other way around?
[01:28] <BBB> hm, oh, wait, I was using mpegdemux .audio_00 ! ...
[01:28] <BBB> doh
[01:35] Action: mathrick goes to sleep
[01:35] <mathrick> nite
[01:35] Nick change: mathrick -> mathrick|sleep
[01:53] BBB ([email protected]) left irc: "Client exiting"
[01:54] Miko5881 ([email protected]) joined #gstreamer.
[01:55] <walters> hmm
[01:55] <walters> how do i load an element by mimetype in 0.7?
[01:57] <Company> element by mimetype?
[01:58] zeenix ([email protected]) left irc: "Leaving"
[01:59] <walters> Company: http://gstreamer.net/docs/current/manual/html/ch13s03.html
[01:59] <walters> Company: i don't want to hardcode "mad" in rb for iradio
[01:59] <walters> Company: instead i want to grab an element that can decode audio/mpeg.
[02:01] <Company> you know that gstreamer and gnome-vfs or apache mimetypes don't match, right?
[02:02] <walters> yes
[02:03] <Company> anyway, if you have a mimetype, you can just create a caps from it and then look for an element that can connect to those caps
[02:03] <Company> gst_caps_new_simple (mimetype, NULL) should work
[02:04] <walters> i guess it would be better to get autoplugging to work for iradio
[02:05] <Company> it doesn't work if you put typefind in front?
[02:07] <walters> /pipeline0/typefindelement0.sink: caps = application/octet-stream
[02:07] <Company> hum
[02:08] <Company> can you give me the stream you test with?
[02:08] <walters> here's one test
[02:08] <walters> /build/gstreamer-0.7/bin/gst-launch-0.7 -v gnomevfssrc iradio-mode=1 location=http://broadband.afterhoursdjs.org:8008 ! typefind ! fakesink
[02:10] <Company> that's anm mp3 stream?
[02:11] <walters> yeah
[02:13] <Company> typefinding is a thing that is supposed to work 100%
[02:14] iain ([email protected]) left irc: "leaving"
[02:14] <Company> it's actually one of the best designed parts of gstreamer and certainly one i'm extremely proud of :)
[02:15] <walters> it does seem pretty nice
[02:15] <walters> man, gstreamer is very cool.
[02:16] <walters> i like just being able to replace osssink with filesink to save this stream to a file
[02:16] <Company> yeah, and after that you use filesrc ! audio/x-raw-int,... ! osssink to play it back :)
[02:23] <walters> this is kinda odd
[02:23] <walters> if i stream it directly
[02:23] <walters> the type is detected as application/octet-stream
[02:23] <walters> but if i play it from filesrc, it's correctly detected as audio/mpeg
[02:24] <Company> yeah, it's probably a buffer size issue or so
[02:24] <walters> hm
[02:26] <walters> i have gnomevfssrc hacked to return a huge buffer on the first get though
[02:26] <walters> 8000 bytes anyways
[02:27] <Company> that shouldn't matter anymore
[02:27] <Company> (though it probably still has a bug somewhere as we can see)
[02:27] hallibaby ([email protected]) left irc: "Bye bye"
[02:28] <walters> if we can get rb to not hardcode mad anymore i will be sooooooo happy
[02:28] <walters> hardcoding mad just feels incredibly dirty
[02:29] <Company> gst-launch-0.7 -v gnomevfssrc location=http://broadband.afterhoursdjs.org:8008 ! typefind ! fakesink returns audio/mpeg here
[02:29] <walters> hmmm
[02:29] <Company> but with iradio-mode=1 nothing happens
[02:29] <walters> is that reliably reproducible?
[02:29] <walters> ahhhh
[02:29] <walters> interestin
[02:29] <walters> very interesting
[02:30] <walters> i'll feel especially dumb if this is all my fault
[02:30] <Company> :p
[02:30] <Company> you don't need to make the first buffer bigger for typefinding btw
[02:30] <walters> ok
[02:30] <walters> i'll fix that now
[02:30] <Company> so gnomevfssrc ! typefind should always detect the correct type
[02:31] <walters> dude
[02:31] <walters> somehow
[02:31] <walters> it was that code
[02:31] <Company> if someone make gnomevfs implement the uri interface and someone fixes gst-typefind, it should even be possible to run gst-typefind-0.7 http://...
[02:32] <walters> word
[02:32] Action: walters updates rb to use spider
[02:32] <Company> so that means typefinding is still bugfree :)
[02:33] <walters> i wonder what the bug was
[02:33] Action: Company continues being proud of that subsystem :)
[02:34] <walters> hmmm
[02:35] <walters> no audio.
[02:37] Action: Company just realized that "liebe" is a better sounding word than "love"
[02:38] <walters> walters@nexus> shell/rhythmbox --dry-run --gst-debug=spider:5,spideridentity:5
[02:38] <walters> DEBUG          spider(13167) gstspider.c(215):gst_spider_request_new_pad: successuflly created requested pad autoplugger:src_0
[02:38] <walters> DEBUG          spider(13181) gstspider.c(594):gst_spider_plug_from_srcpad: trying to plug from sink_ident:src to src_0
[02:38] <walters> DEBUG          spider(13181) gstspider.c(619):gst_spider_plug_from_srcpad: no chance to plug from sink_ident to src_0
[02:38] <walters> hmm
[02:39] <Company> that's correct
[02:39] <Company> thouzgh it should probably say "no chance to plug directly"
[02:39] <Company> unless i'm missing something...
[02:40] <Company> hm shit
[02:40] <Company> typefind on that stream doesn't always work :/
[02:44] <walters> wow
[02:44] <walters> i get a *lot* of deep_notify signals
[02:44] <walters> looks like 15-20 per second
[02:44] <Company> where from?
[02:45] <walters> not sure
[02:48] <walters> hm, this is weird
[02:49] <walters> ohhh
[02:49] <walters> i'm stupid.
[02:49] <walters> very stupid :)
[02:50] <Company> yes, i've isolated a bug
[02:51] <Company> ok, now I'm gonna fix it :)
[02:51] <walters> cool
[02:53] Misirlou ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[02:53] Misirlou ([email protected]) joined #gstreamer.
[02:57] <walters> word
[02:57] <walters> it mostly works!
[02:57] <walters> ** (rhythmbox:27164): WARNING **: couldn't set the right number of channels, enjoy the speed difference
[02:57] <walters> heh
[02:58] <walters> * committed [email protected]/rhythmbox--mainline--0.7--patch-177
[02:58] <walters> word
[02:58] <walters> we no longer hardcode mad for iradio
[02:59] <walters> now i just need to find a vorbis station to test...


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