IRC Logs

*******************************************************************
[03:08] aldug ([email protected]) left irc: Connection timed out
[03:23] md` ([email protected]) joined #gstreamer.
[03:41] taaz ([email protected]) joined #gstreamer.
[03:41] taaz ([email protected]) left irc: Remote closed the connection
[03:42] taaz ([email protected]) joined #gstreamer.
[03:43] taaz ([email protected]) left irc: Remote closed the connection
[03:53] alley_cat ([email protected]) left irc: "May the Source be with you!"
[04:19] taaz ([email protected]) joined #gstreamer.
[04:52] aldug ([email protected]) joined #gstreamer.
[05:05] Rotty ([email protected]) left irc: Remote closed the connection
[05:33] jcsston` ([email protected]) joined #gstreamer.
[05:41] jcsston ([email protected]) left irc: Read error: 60 (Operation timed out)
[05:46] Nick change: jcsston` -> jcsston
[05:47] <jcsston> Could GStreamer be considered COM-like?
[06:29] ChriHJW_log ([email protected]) joined #gstreamer.
[06:34] Action: bitshifter wished that --[enable|disable]-plugin-builddir warning of configure would be printed at the beginnig and not at the end
[06:35] ChrisHJW_log ([email protected]) left irc: Connection timed out
[06:50] mathrick|sleep ([email protected]) left irc: Remote closed the connection
[06:55] <bitshifter> *** No rule to make target `gstplay.c', needed by `libgstplay_0.7_la-gstplay.lo'.  Stop.
[06:55] <bitshifter>   (current anon cvs)
[07:16] ChrisHJW ([email protected]) joined #gstreamer.
[07:59] Nick change: harshyOut -> harshy
[08:11] lilo ([email protected]) got netsplit.
[08:11] KyleB ([email protected]) got netsplit.
[08:11] khalek ([email protected]) got netsplit.
[08:11] khalek ([email protected]) returned to #gstreamer.
[08:11] lilo ([email protected]) returned to #gstreamer.
[08:11] KyleB ([email protected]) returned to #gstreamer.
[08:13] taaz ([email protected]) got netsplit.
[08:18] aeyakovenko ([email protected]) joined #gstreamer.
[08:19] taaz ([email protected]) got lost in the net-split.
[08:20] <aeyakovenko> Is it possible to request multiple src pads from any element as it was a tee?  If not, I think this would be a great feature.
[08:25] Misirlou ([email protected]) left irc: Read error: 110 (Connection timed out)
[08:35] Misirlou ([email protected]) joined #gstreamer.
[08:35] taaz ([email protected]) joined #gstreamer.
[08:45] markey ([email protected]) joined #gstreamer.
[08:50] markey ([email protected]) left irc: "later"
[09:24] aeyakovenko ([email protected]) left #gstreamer.
[09:44] jcsston ([email protected]) left irc: Read error: 60 (Operation timed out)
[09:45] Company ([email protected]) joined #gstreamer.
[09:45] <Company> good morning angels
[09:46] <bitshifter> morning Charl^H^H^H^Hompany
[09:47] <bitshifter> say, is   filesrc location=id3v1_tagged.mp3 ! typefind ! spider ! application/x-gst-tags ! fakesink  already supposed to extract the ID3v1 tag?
[09:47] <bitshifter> doh, nm
[09:48] Action: bitshifter just remembered that he asked the same thing 2 days ago
[09:48] <Company> and yes it's supposed to :)
[09:49] <bitshifter> I think I filed a bug report because it doesn't (you said spider somehow messes it up or so)
[09:49] <Company> it shouldn't
[09:49] <Company> the only thing that might not work is id3v2 tags at the end
[09:59] aeyakovenko ([email protected]) joined #gstreamer.
[10:08] <lilo_booter> hmm - got what might seem like the dumbest question in the world :-) - but what the heck... i've been giving some thought to a command line oriented multitrack video editor, and i have arrived at a set of switches to do everything that i would want to do (user freindly it ain't, consistent it is :-)... question is, would someone care to review a doc and point out stuff i've missed?
[10:09] <Company> depends on how good the doc is :)
[10:10] <lilo_booter> heh - being the author of that, i'm not really the one to ask on quality issues ;-)
[10:12] <lilo_booter> i'll message the url to anyone who's interested - it might be construed as off topic here given that the approach doesn't involive gstreamer (or at least, it doesn't yet...)
[10:15] <Company> well, i have no clue about video editing, so it's probably useless to read it, but i'll do it anyway if you'd like me to :)
[10:15] <lilo_booter> sure - thanks :-)
[10:21] markey ([email protected]) joined #gstreamer.
[10:27] apoc ([email protected]) joined #gstreamer.
[10:27] <apoc> hi
[10:35] <thomasvs> lilo_booter: I want to give it a go too
[10:35] <lilo_booter> cheers thomasvs :-)
[10:37] mathrick ([email protected]) joined #gstreamer.
[10:37] <Company> thomasvs: the oggdemux patch solved a problem for you?
[10:38] markey ([email protected]) left irc: "bbl"
[10:39] Action: Company will keep using wondershaper
[10:39] <Company> it doesn't work for filesharing and ego shooter though
[10:39] <Company> but it works for file sharing and ssh :)
[10:42] wheels ([email protected]) left irc: "this->istanbul()"
[10:52] <thomasvs> Company: yep - it flushes the queue on playback
[10:53] <thomasvs> Company: still have the same issue though, of after seeking getting an impossible position
[10:53] <thomasvs> Company: only know it happens immediately :)
[10:53] <thomasvs> Company: fwiw, with esdsink it seems to work
[10:53] <thomasvs> Company: maybe osssink still references an absolute clock somewhere ?
[10:53] <Company> thomasvs: hm, will investigate - but the flushing upstream is wrong
[10:53] <Company> thomasvs: oggdemux should flush downstream manually
[10:54] <Company> thomasvs: (when the seek event told it to flush)
[10:55] <thomasvs> Company: what do you mean ? it should only flush if the seek coming from upstream is also a flush seek ?
[10:56] <Company> thomasvs: yeah
[10:56] <Company> thomasvs: and it should flush downstream in the discont
[10:57] <Company> thomasvs: i'll fix that and make you test when i'm done
[10:58] <Company> thomasvs: the flush flag for seeking was included because people sometimes want immediate seeks and sometimes want "seek only after all data is processed" - wtay needed the second way for gnonlin
[11:15] <thomasvs> what would be a good use example of the second ? (so that I understand the problem better)
[11:19] <Company> if you have a pipeline like filesrc ! tee ! sink1 tee0. ! { queue ! sink2 }
[11:20] <Company> and wanted to seek back when sink1 reaches EOS
[11:20] <Company> and want to have all data reach sink2
[11:21] <Company> if you'd flush anyway, the queue might be flushed before sink2 has recieved all data up to the EOS
[11:22] <thomasvs> hm, that seems like it would be an application-specific choice, no ?
[11:22] <thomasvs> ie, for example the queue would be flushing or nonflushing
[11:22] <thomasvs> where for playback you probably prefer flushing
[11:24] <Company> well, you might want to seek on sink1 anyway
[11:24] <Company> with a flush
[11:29] Action: Company doesn't get this
[11:30] <Company> why does g_return_if_fail (GST_IS_BUFFER (buffer)); change the value of buffer?
[11:30] sxpert ([email protected]) joined #gstreamer.
[11:37] <Company> gdb...
[11:37] ensonic ([email protected]) joined #gstreamer.
[11:38] <ensonic> re
[11:39] Uraeus ([email protected]) joined #gstreamer.
[11:39] <apoc> hi Uraeus
[11:39] <Uraeus> morning apoc
[11:45] <Company> i still wonder why theora looks so jerky
[11:45] <Uraeus> maybe cause it isn't finished yet?
[11:51] <khalek> well vp3 isn't exactly new technology
[11:52] <bitshifter> gst-register-0.7: relocation error: /usr/local/lib/gstreamer-0.7/libgstintfloat.so: undefined symbol: gst_float2_2_int_get_type
[11:52] <bitshifter> any idea what that might be?
[11:53] <Uraeus> Company: afaik so where you the last person to tuch the intfloat pluins, any idea on bitshifter's issue?
[11:56] <Company> bitshifter: recompile the plugin and be sure to have gst-plugins/gst/intfloat/float22int.c
[11:58] <Company> thomasvs: what's the difference between GST_TAG_FLAG_ENCODED and GST_TAG_FLAG_DECODED ?
[11:59] BBB ([email protected]) joined #gstreamer.
[12:00] <mathrick> lo BBB
[12:00] <BBB> thomasvs: I'd like to request one change to doc building
[12:00] <BBB> I require an internet connection to do 'make' in docs/pwg/
[12:00] <BBB> that's not good for us old-fashioned poor little 56k developers
[12:00] <BBB> lo mathrick
[12:01] <Uraeus> BBB: http://bugzilla.gnome.org/show_bug.cgi?id=129342 <- patch for gnome sound recorder
[12:02] <BBB> I know, and it's Ted's task to apply pending patches
[12:02] <BBB> he's maintainer
[12:03] <BBB> (and he rejected the patch because it doesn't apply...)
[12:04] <BBB> jade:<URL>http://www.oasis-open.org/docbook/xml/4.2/dbpoolx.mod:3529:0:E: could not resolve host "www.oasis-open.org" (try again later)
[12:04] <BBB> *curse*
[12:07] Action: mathrick curses @ topposters
[12:08] <Company> BBB: that's a wrong xml catalog most likely :)
[12:09] <mathrick> BBB: parser should have some --nonet option, although it's probably called from makefile, right?
[12:12] alley_cat ([email protected]) joined #gstreamer.
[12:12] <BBB> Company: it used to work just fine
[12:12] <BBB> mathrick: ues
[12:12] <mathrick> BBB: huh?
[12:13] <Company> "united earth species" probably
[12:13] <BBB> yes
[12:13] <BBB> ;)
[12:15] <mathrick> BBB: so one hacky way is to add that option manually to makefile. The proper way is probably to kick thomasvs ;)
[12:16] <mathrick> BBB: I've had similar pbs with xsltproc, adding -nonet solved them
[12:16] <ensonic> BBB: have a look at /etc/xml/catalog
[12:16] <BBB> I'll kick thomas
[12:16] <BBB> I can't find the option myself
[12:16] <mathrick> BBB: btw, Debian?
[12:16] <ensonic> i use docbook without net connection
[12:17] <ensonic> BBB: xsltproc has alternatively the --nonet option (requires patching the makefile thougH)
[12:17] <mathrick> ensonic: read above :)
[12:18] <ensonic> mathrick: thats why i recommned fixing the catalog ;-)
[12:18] <BBB> how do I fix the catelog?
[12:18] <mathrick> BBB: it depends on you using Debian
[12:21] <BBB> no
[12:22] Action: BBB downloads and adds random characters in catalog file
[12:22] <mathrick> BBB: then I don't know ;)
[12:22] <ensonic> BBB: you can use the xmlcatalog command
[12:22] <mathrick> BBB: on Debian it takes running some smartass script for that
[12:23] <mathrick> BBB: that's because Debian traditionally ships with broken catalog, so people have written scripts for that ;)
[12:23] <BBB> ensonic: hm, how?
[12:23] <mathrick> BBB: #xmlcatalog ?
[12:23] <ensonic> BBB: mmompl
[12:24] <mathrick> BBB: IIRC, that's what that script does, after setting lots of magic env vars :)
[12:24] <mathrick> BBB: wanna me DCC it?
[12:25] <ensonic> BBB: example : xmlcatalog --noout --add "system" "http://gitk.sourceforge.net/giml.dtd" "file:///filepath/giml.dtd" /etc/xml/catalog;
[12:25] <ensonic> BBB: replace with docbook urls and repeat with public idetifier
[12:25] <BBB> xmlcatalog --add rewriteURI http://www.oasis-open.org/docbook/xml/4.2/dbpoolx.mod file:///home/rbultje/projects/dbpoolx.mod /etc/xml/catalog ?
[12:26] <ensonic> BBB no rewiteURI has a different syntax
[12:27] Action: BBB does system
[12:27] <BBB> can't I automate all this?
[12:27] <ensonic> public publicId="-//OASIS//DTD DocBook XML V4.1.2//EN" "file:///.../docbookx.dtd"
[12:27] <BBB> it's still accessing internet
[12:27] <ensonic> BBB: you have to do this only once ;-)
[12:28] <alley_cat> but for all dtds you want to use without fetching them from the net
[12:28] <BBB> exactly
[12:28] <BBB> and that sucks
[12:28] Action: alley_cat got quite a long script for that in his docbook-dtd package
[12:30] <BBB> I'll stop working on docs, this doesn't work
[12:30] <BBB> it still thinks it wants to fetch them from internet and still claims that I'm not connected to the internet
[12:30] <mathrick> BBB: is your DCC disabled?
[12:31] <BBB> no
[12:31] <BBB> normally it works
[12:31] pb_ ([email protected]) joined #gstreamer.
[12:32] <mathrick> BBB: apparently not from my box
[12:32] <BBB> :p
[12:32] <BBB> I'll fix this later
[12:32] <mathrick> BBB: anyway, I have script that should be pretty easy to fix for your distro
[12:32] <BBB> not now, not here
[12:32] Action: BBB has a party in a few hours
[12:33] <BBB> I'll fetch it tomorrow
[12:33] Action: BBB kicks stupid documentation
[12:33] <BBB> what idiot ever got the idea that people have internet access?
[12:34] <mathrick> BBB: SGML catalog design commitee ;P
[12:34] <BBB> that's worse than microsoft's idea of hosting all your personal data online on their servers, only available through encryprtion keys that they seel to you for 1$/minute
[12:34] <BBB> they're on crack
[12:34] <BBB> anywya
[12:35] <BBB> I'm off
[12:35] <BBB> bye!
[12:35] <Company> the xml tools should save everything locally they fetched online
[12:35] BBB ([email protected]) left irc: "Client exiting"
[12:35] <Company> and next time use the local copy
[12:35] <mathrick> Company: that would mean they won't use updated DTDs if any
[12:35] <mathrick> Company: that's why you need to build catalog manually
[12:37] <Company> DTDs aren't updated
[12:37] <Company> or catalogs would be bad
[12:37] <ensonic> good packages update the catalog for the user when doing 'make install'
[12:39] <aeyakovenko> is it currently possible to ask for multiple pads from any element, as if it was a tee?  if its not, i think it would be a good feature
[12:40] <Company> no, it's not a good idea
[12:40] <aeyakovenko> how come?
[12:40] <Company> if you need that, use a tee
[12:40] <Company> because it makes handling pads much more complex
[12:40] <Company> you'd have to duplicate tee's code in every element
[12:41] <aeyakovenko> how about a way to create a element with a tee attached, like a bin, that you can ask for src pads from
[12:41] <Company> why is it so complicated to attach the tee manually?
[12:42] <aeyakovenko> its not, but i ended up wrapping that in a class anyways, just thought it might be a handy feature for someone else
[12:43] <Company> dunno, it would be quite complex to code
[12:43] <aeyakovenko> really?  just a wrapper around an element and a tee, the element is linked to the tee, when you ask for a pad, you ask the tee for it
[12:43] <Company> as you have to make sure you automatically remove tees when request pads get removed, you need to handle requesting pads and all that
[12:44] <Company> and there's obviouly elements with loads of srcpads
[12:44] <aeyakovenko> ok
[12:47] <aeyakovenko> i am prolly gonna get to writing an rtp element this weekend, i had some questions to ask you.  Jrtplib supports multicasting and stuff like that, so would i want to have a list of destination properties?
[12:48] <aeyakovenko> it makes more sense to me to have each destination as a sink, but they should somehow share the actual rtpstack resource
[12:49] <mathrick> aeyakovenko: are you basing it on existing rtp plugin(s)?
[12:49] <Company> no, that should be only one element
[12:50] <aeyakovenko> mathrick: to be honest i havent looked at it, and i heard that its not really complete, ive used jrtplib before, and its an accelent library
[12:50] <aeyakovenko> mathrick: but its in c++, which is really not that big of a deal
[12:51] <mathrick> aeyakovenko: ah, k, if you have something that'll accelerate your work, then it's ok
[12:51] <mathrick> aeyakovenko: and if it isn't too wicked C++, it shouldn't be much of hassle
[12:51] <aeyakovenko> mathrick: if you google for it it should come up, it has a really intuitive interface, bit i want to make it work with gstreamer
[12:53] <aeyakovenko> Company: so would i have something similar to the location property?
[12:53] <Company> aeyakovenko: yeah, you could invent a way to spcify multiple IPs, like using comma or something
[12:53] <Company> 1.2.3.4:600,4.3.2.1:245
[12:54] <aeyakovenko> Company: how would you be able to add and remove them on the fly?
[12:54] <Uraeus> aeyakovenko: you are using the live.com library?
[12:54] <aeyakovenko> Company: I guess a addip/removeip
[12:54] <Company> aeyakovenko: handling lists is a thing properties can't really do
[12:55] <Company> aeyakovenko: having add/remove properties seems bad
[12:55] <aeyakovenko> Uraeus: here is a link http://lumumba.luc.ac.be/jori/jrtplib/jrtplib.html  this may not be the best rpt lib out there, but ive used this one on other projects and i like it
[12:55] <Company> aeyakovenko: i'd say use g_object_get, add/remove from the string you got and then g_objecxt_set it again
[12:55] <aeyakovenko> Company: jrptlib stores the destinations internaly, i am just wondering how to make its interface accessible through gstreamer
[12:56] <bitshifter> Company: int2float problem is gone (had to do rm -f /usr/local/lib/gstreamer-0.7 /usr/local/lib/libgst*  and recompile gst + plugins from scratch though)
[12:56] <Uraeus> aeyakovenko: http://www.live.com/liveMedia/ <- ok cause this is the one used by other projects like mplayer and I think xine, so it has a good mindshare
[12:57] apoc ([email protected]) left irc: "Leaving"
[12:58] <mathrick> aeyakovenko: it's just a matter of translating it to / from a list of strings
[12:59] <mathrick> aeyakovenko: just use GQuarks, and you're mostly done
[12:59] <aeyakovenko> Uraeus: cool, ill check it out.  do you know if its a just a bear bones rtp stack, or does it do codec stuff as well?
[13:00] <Uraeus> aeyakovenko: as far as I know it is just bare bones
[13:00] <aeyakovenko> mathrick: what about adding and removing destinations on the same element?
[13:01] <Uraeus> aeyakovenko: and I hope you end up using it, since it does more protocols than jrtplib which means we can add plugins for more protocols later without adding yet another dependency, and since mplayer and vlc use it then it also sees more bugfixing
[13:01] <aeyakovenko> Uraeus: ok great.  ill check it out then.  i am sure its all the same, the rtp is not very complicated
[13:01] apoc ([email protected]) joined #gstreamer.
[13:01] <Uraeus> wb apoc
[13:02] <mathrick> aeyakovenko: hmm? I don't think I get you, all you need to do is store list of strings for destinations, and when new one is set, compare it to old one. GQuarks make this efficient and easy
[13:03] <Company> you don't want to use GQuarks for IPs
[13:03] <mathrick> Company: they're pure IPs? Then even better
[13:04] <apoc> Uraeus: does gst-player work ?
[13:04] <aeyakovenko> i prolly need to learn more about gstreamer first
[13:04] <Uraeus> apoc: there are still some sync issues, but I just played an avi in it
[13:04] <mathrick> aeyakovenko: http://developer.gnome.org/doc/API/2.0/glib/glib-Quarks.html if you're interested
[13:04] <apoc> Uraeus: did you test some mov/svq3 or wma ?
[13:05] sublett ([email protected]) joined #gstreamer.
[13:05] <thomasvs> Uraeus: what issues still ?
[13:05] <Uraeus> thomasvs: well using the Matrix video the audio and video is not in sync for me
[13:05] <Uraeus> apoc: I tried a mov and gst-player crashed
[13:06] <Uraeus> apoc: want me to send you the mov file?
[13:06] <apoc> Uraeus: yes please
[13:06] <apoc> Uraeus: svq3 ?
[13:06] <aeyakovenko> my question is more on how to tell the elment that i want to add a destination or that i want to remove one
[13:07] <Uraeus> apoc: I have a wmv file which seems to play in gst-player, but only sound, the screen is black
[13:07] <thomasvs> Uraeus: move the window, see if that helps
[13:08] <aeyakovenko> would it be soemthing like  g_object_set (G_OBJECT (element), "destination", "add","blah.com", NULL);
[13:08] <aeyakovenko> but thats really ugly, and doesnt really fit with how everything else works
[13:09] <Company> aeyakovenko: always set all destinations you want to have
[13:09] <Company> aeyakovenko: and then the element figures out itself if some where added or removed
[13:09] <Company> (yes, thats suboptimal)
[13:09] <aeyakovenko> Company: well that works
[13:09] <Company> (but it's the way glib works)
[13:11] <Uraeus> thomasvs: no, moving the window doesn't make any difference
[13:11] <aeyakovenko> Company: what about a rtpencoder with a  src and a sink? you can then just link it to a udpsink
[13:11] <Uraeus> thomasvs: but my CVS is from yesterday so I might be missing some fixes
[13:12] <apoc> Uraeus: what is the video codec of your wmv ?
[13:12] <aeyakovenko> Company: that would take some effort though
[13:12] Action: sxpert is away: call for lunch...
[13:12] <apoc> Uraeus: you are sending the mov file at @yahoo.fr or @libertysurf.fr ?
[13:13] <mathrick> aeyakovenko: there was discussion about such details somewhere back in September, take a look at -devel archives
[13:13] <aeyakovenko> ok
[13:13] <Uraeus> apoc: actually it is 12MB so I ended up uploading it to our media area on SF
[13:13] <Uraeus> apoc: I give you the URL when its done
[13:14] <Uraeus> apoc: how do I see which codec the wmv uses?
[13:14] <apoc> Uraeus: with mplayer for example
[13:14] <Company> aeyakovenko: this is ok if an rtpencoder does not need to know about IPs and only sends data, but doesn't receive any
[13:14] <Uraeus> apoc: Selected video codec: [ffwmv1] vfm:ffmpeg (FFmpeg M$ WMV1/WMV7)
[13:15] <apoc> Uraeus: ok so gst should play it fine :(
[13:15] <aeyakovenko> Company: that would be a problem, rtppackets carry ip info i believe
[13:15] hallibaby ([email protected]) left irc: "Bye bye"
[13:15] <apoc> Uraeus: does it work with gst-launch ?
[13:16] <Uraeus> apoc: when I run gst-player on the command line with the video it says: ERROR        GST_PADS( 7064) gstpad.c(2260):gst_pad_set_explicit_caps: <qtdemux0> failed to negotiate (try_set_caps returned REFUSED)
[13:16] <Uraeus> Segmentation fault
[13:16] <Uraeus> sorry, that was the quicktime
[13:17] <aeyakovenko> Company: yea, i think it uses it to generate the CSRC field
[13:17] <Uraeus> apoc: no, doesn't play with gst-launch either 
[13:18] <Uraeus> apoc: should any .mov trailer from quicktime.com work? any you know work I can test?
[13:19] <apoc> Uraeus: ftp://ftp.mplayerhq.hu/MPlayer/samples/V-codecs/SVQ3/Vertical400kbit.sorenson3.mov
[13:21] <Company> aeyakovenko: in that case it sounds like a stupid idea to seperate the sink and the rtp encoder
[13:22] <Company> aeyakovenko: especially because people might just want to connect rtpenc to something different then
[13:23] <Company> hm, either theora is sloooow or i don't understand it
[13:24] <aeyakovenko> Company: yea, I think your suggestion is the best way to go
[13:24] <aeyakovenko> Company: i was just thinking in terms of a large multicast server or something, if you had lots of connected clients connect/disconnect pretty frequently
[13:25] <Company> you could maybe invent some caps like application/x-ip and use ips during caps nego, but i'm not sure that would make sense
[13:25] <Company> aeyakovenko: that would be handled by the element automatically, wouldn't it?
[13:26] <Uraeus> apoc: http://www.gstreamer.net/media/large/ -> the lord of the rings parody trailer (12MB)
[13:26] <ensonic> does gst-player-0.6 work with gstreamer-0.7.3?
[13:26] <Company> no, you need gst-player CVS
[13:27] <ensonic> Company: thx, is a tar.gz release planned for gst-player-0.7.X ?
[13:27] <aeyakovenko> Company: yes, but, both the user and the element would have to rebuild the list every time
[13:27] <Company> ensonic: dunno, you'd have to ask dolphy about that
[13:27] <Uraeus> apoc: ok, gst-player gives me the same error with the file you gave me the URL to as I get with the file I just gave you the URL to
[13:27] <Company> aeyakovenko: not really, you only build the list on demand by the application
[13:28] <Company> aeyakovenko: since the application normally doesn't care as the element handles it all itself, it's pretty rare that you need to build the list
[13:29] <aeyakovenko> Company: i think you would need to rebuild it anytime you want to add or remove a destination, rtp doesn't handle any session initiation
[13:29] sxpert_ ([email protected]) joined #gstreamer.
[13:30] <Company> aeyakovenko: dunno
[13:31] <mathrick> aeyakovenko: why would you need any session initiation?
[13:32] <aeyakovenko> mathrick: not in gst, in general just to use rtp you need to use another protocol to initiate a session
[13:34] <mathrick> aeyakovenko: yes, but you can store session state (initiated or not) in your element, and just initiate it for newly added dests
[13:34] <mathrick> aeyakovenko: or am I missing something?
[13:34] <ensonic> on gstreamer/modules it says that gst-editor-0.6.0 is available, but I can only find 0.5.0, which does not compile (gst-element-ui.c)
[13:34] <mathrick> ensonic: cvs -z3 co gst-player <-- works for me :)
[13:34] <ensonic> irks ;-)
[13:35] <aeyakovenko> mathrick: well no, you are right. what i am trying to figure out is how to modify the sesison state, change the current set of destinations, Company suggested that i just always set the destination to some list, and the element figures out who is removed and who is added.  the problem with that is if you have a large list that changes frequently
[13:36] dragongirl ([email protected]) joined #gstreamer.
[13:36] <mathrick> aeyakovenko: frequently? as long as you don't do it 1000 times per second, it shouldn't be a problem :)
[13:37] <aeyakovenko> mathrick: hehe, well i promise i wont
[13:37] <mathrick> aeyakovenko: traversing list of 100 items doing hash lookup isn't that much of deal i think :)
[13:38] <aeyakovenko> mathrick: its also ugly, because the user and the element have to rebuild the list
[13:38] <mathrick> aeyakovenko: and recipients list probably shouldn't be larger than that
[13:38] <mathrick> aeyakovenko: no
[13:39] <mathrick> aeyakovenko: at least, not element
[13:39] <mathrick> aeyakovenko: list returned for client will probably have to be built each time, that's true
[13:39] <aeyakovenko> mathrick: how would it choose which destinations to remove, it has to compare the new state to the saved
[13:40] <mathrick> aeyakovenko: yes, but if you store it efficiently (ie hashtable), doing lookup for 100 items nad changing status of 10 probably is cheap. It's building list for client out of hashmap that could be expensive
[13:41] <Company> it's funny that people think about optimization before there's a speed problem
[13:41] <mathrick> aeyakovenko: and when client gets the list, modifing it is also easy
[13:41] <aeyakovenko> Company: hehe
[13:42] <mathrick> aeyakovenko: so as long as you don't get above 100 connections changing at 1000/s rate, it's fine :)
[13:43] <Uraeus> Company: all pipelines using movie.avi/mpg ! spider ! ximagesink fails (while using xvimagesink works)
[13:43] <aeyakovenko> mathrick: i could also create a new sink for every destination, just waste memory
[13:43] <mathrick> aeyakovenko: you loose multicast then
[13:43] <mathrick> s/loose/lose/
[13:43] <aeyakovenko> yea
[13:44] dragongirl ([email protected]) left #gstreamer ("Leaving").
[13:44] <mathrick> aeyakovenko: write elem first, and when you get 100 clients simultanously, come back and cry here ;P
[13:45] <Company> Uraeus: that's because you need ffcolorspace in between
[13:45] <aeyakovenko> hehe
[13:45] <mathrick> Company: shouldn't spider connect it automatically?
[13:46] <mathrick> s/connect/insert/
[13:46] <aeyakovenko> Uraeus: by the way, that live.com lib is about 5 times larger then jrtplib 70k to 350k, and it doenst look like you can strip out just the rtp stuff
[13:46] <Uraeus> aeyakovenko: true, but as I said long term we want plugins for all those protocols, so in total the dependency is smaller (from a gstreamer viewpoint)
[13:47] sxpert ([email protected]) left irc: Read error: 110 (Connection timed out)
[13:47] <Company> mathrick: maybe, but mit never did this, so i'm not going to break all sorts of apps by doing this ;)
[13:48] <Uraeus> Company: mit?
[13:48] <mathrick> Company: heh, and why didn't it do that? no rank assigned to colorspace?
[13:48] <mathrick> Uraeus: s/mit/it/ ;)
[13:48] <mathrick> Uraeus: m is close to spacebar :)
[13:49] <Uraeus> ah :)
[13:49] <Company> if someone decides to write a rtp plugin, he is free to chose the rtp lib he wants
[13:49] Action: mathrick has memorizes qwerty layout long time ago ;)
[13:49] Action: Uraeus wasn't aware that we didn't autoplug colorspace plugins
[13:49] <mathrick> hmm, s is close to d apparently ;P
[13:49] <Company> he can even use an rtp lib that is 9MB if he likes
[13:49] <Uraeus> Company: agreed, just tossing an opinion into the ring
[13:49] <Company> if someone doesn't like it he's free to write another one - we'll then take the best one :)
[13:50] <Uraeus> Company: could be a she that doesn't like it, don't be sexist ;)
[13:50] <Company> mathrick: yes, there wasn't a rank assigned - maybe there are other reasons, dunno
[13:50] <Company> Uraeus: you sound like only woman complain :p
[13:50] <aeyakovenko> time to sleep, later
[13:51] aeyakovenko ([email protected]) left #gstreamer.
[13:51] <Company> mathrick: spider autoplugs float2int though (dunno about audioconvert)
[13:51] <Company> mathrick: and i intend to change it for the new autoplugger
[13:51] <mathrick> Company: hmm, why?
[13:52] <mathrick> Company: ie, why change that?
[13:52] <Company> because you want to autoplug
[13:52] <Company> you don't want to plug parts yourself
[13:52] <Uraeus> Company: are you and ds working on the same autoplugger or are you doing one each?
[13:52] <mathrick> <Company> mathrick: spider autoplugs float2int though (dunno about audioconvert)
[13:52] <mathrick> <Company> mathrick: and i intend to change it for the new autoplugger
[13:52] <Company> Uraeus: the same
[13:52] <mathrick> Company: so what exactly do you want to change?
[13:53] <Company> mathrick: i want to plug converters automatically (any sort)
[13:53] <mathrick> Company: ahh, your imprecise way of speaking ;P
[13:53] <mathrick> s/speaking/writing/ actually ;)
[13:54] <Company> :p
[13:54] <mathrick> Company: will the new autoplugger only use ranked elems?
[13:54] <Company> mathrick: dunno, probably configurable or so
[13:56] <mathrick> Company: b/c if only ranked were to be used, then you could either break spider behavior by assigning it, or not use them
[13:57] <Company> no, i couldn't break it, i could expose its breakage ;)
[13:57] <mathrick> Company: break == "change so that much of current users start to barf" ;)
[13:58] <Company> whatever :p
[14:00] <mathrick> Company: although not plugging in colorspace is kinda illogical, so they should be grateful for you fixing it ;)
[14:13] <mathrick> make[4]: Entering directory `/home/mathrick/Dev/gstreamer/gstreamer/docs/manual'
[14:13] <mathrick> cd build && xmllint -noout -valid manual.xml
[14:13] <mathrick> error : Operation in progress
[14:13] <mathrick> manual.xml:56: I/O error : failed to load external entity "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd"
[14:13] <mathrick> ]>
[14:13] <mathrick>   ^
[14:13] <mathrick> manual.xml:58: element book: validity error : No declaration for attribute id of element book
[14:13] <mathrick> <book id="index">
[14:13] <mathrick>                 ^
[14:13] <mathrick> thomasvs: same as BBB's error
[14:24] sublime ([email protected]) joined #gstreamer.
[14:24] <sublime> hello
[14:25] <bitshifter> mathrick: are you on debian stable by any chance?
[14:26] subleet ([email protected]) joined #gstreamer.
[14:26] <mathrick> bitshifter: nope, Sid here
[14:27] <mathrick> hi sublime
[14:32] <sublime> what's up mathrick
[14:33] <mathrick> sublime: everything's fine here :)
[14:33] <bitshifter> mathrick: k, just wondering because debian bug #211617 (fixed since docbook-util 0.6.13-1) used to cause more or less the same probs for me
[14:35] <mathrick> bitshifter: it may be my catalog's still b0rked, ie. I ran that script, but it's not the one mentioned in jamesh post (it was inaccessible), rather something similar I googled out
[14:35] <thomasvs> [gst-cvs] [thomas@otto gst-player]$ gst-launch filesrc location=/home/thomas/gst/media/good/south.mp3 ! mad ! esdsink
[14:35] <thomasvs> RUNNING pipeline ...
[14:35] <thomasvs> ERROR        GST_PADS(14236) gstpad.c(2260):gst_pad_set_explicit_caps: <mad0> failed to negotiate (try_set_caps returned REFUSED)
[14:35] <thomasvs> ERROR: from element /pipeline0/mad0: Internal GStreamer error: pad problem.  File a bug.
[14:35] <thomasvs> Additional debug info:
[14:35] <thomasvs> gstpad.c(2260):gst_pad_set_explicit_caps:
[14:35] <thomasvs> failed to negotiate (try_set_caps returned REFUSED)
[14:35] <thomasvs> anyone have an idea why <mad0> would be dropped from the debug info for some reason ?
[14:36] <mathrick> thomasvs: any comments on docs build breakage?
[14:36] Action: thomasvs reads backlog
[14:37] <thomasvs> mathrick: looks like your catalog isn't set up properly ?
[14:37] <mathrick> thomasvs: well, I did my best (ie. ran the script). BBB reported the same, and he's not on debian
[14:37] <mathrick> thomasvs: I'll try running it once more
[14:39] <mathrick> urk, it has some problems :\
[14:39] Action: thomasvs can't figure out why <> stuff gets dropped from the debug
[14:39] BBB ([email protected]) joined #gstreamer.
[14:39] <mathrick> wb BBB
[14:40] <BBB> thomasvs: here yet?
[14:40] sublett ([email protected]) left irc: Read error: 113 (No route to host)
[14:41] <Uraeus> BBB: he is here :)
[14:42] <mathrick> /home/mathrick/Downloads/build-xml-catalog-for-debian.sh: line 1: /usr/share/sgml/docbook/stylesheet/xsl/nwalsh/VERSION: No such file or directory
[14:42] <mathrick> bitshifter: what pkgs do you have installed?
[14:42] <BBB> is he not-afk?
[14:42] <mathrick> BBB: just talked with him a second ago :)
[14:42] <ensonic> about the gstreamer manual, wouldn't it be good to include dates when writing '... this component has yet to be written' (I am reading "30.4.3 timelined parameters" at the moment)
[14:43] <mathrick> bitshifter: I have it (it == VERSION) in docbook-xsl, but it apparently wants some sgml- package
[14:43] <bitshifter> ah, one sec
[14:44] <mathrick> ensonic: that's the plan, all newly written docs should be version stamped
[14:44] <mathrick> BBB: you doing that, right?
[14:44] <mathrick> BBB: <thomasvs> mathrick: looks like your catalog isn't set up properly ?
[14:45] <mathrick> BBB: so you should probably set up yours too
[14:46] <BBB> mathrick: I don't know a thing about those
[14:46] <BBB> I'm just writing text
[14:59] BBB ([email protected]) left irc: "Client exiting"
[15:08] Uraeus ([email protected]) left irc: "Client exiting"
[15:17] sublime ([email protected]) left irc: Read error: 110 (Connection timed out)
[15:17] <thomasvs> BBB keeps leaving too soon
[15:19] <mathrick> thomasvs: he has good reason, a ticking one ;)
[15:21] markey ([email protected]) joined #gstreamer.
[15:24] <mathrick> lo markey
[15:26] <markey> re
[15:27] ds-gromit ([email protected]) got netsplit.
[15:28] <Company> is it normal that i have an incredibly nice green picture in gst-player?
[15:28] Shoragan ([email protected]) joined #gstreamer.
[15:28] ds-gromit ([email protected]) returned to #gstreamer.
[15:28] <thomasvs> Company: if it's quicktime, then yes
[15:28] <Company> it's divx avi
[15:29] <Company> and the picture is 16x16
[15:30] <thomasvs> no, that sounds like a nego bug then
[15:30] Action: Company rebuilds plugins...
[15:30] <thomasvs> http://thomas.apestaart.org/download/screenshots/gst-player-error-debug.png
[15:30] <thomasvs> opinions ?
[15:31] <thomasvs> cool, works with translation too
[15:31] <Company> if the text is selectable, it's fine
[15:31] <thomasvs> it is
[15:31] <thomasvs> although
[15:31] <thomasvs> I think it would look better if it was one dialog, and when you click Debug it would add some text to the original dialog
[15:32] <ensonic> in the dev-docs on can read about GstDParam, GstDParamSmooth and Timelined DParans, where the later is still TODO. I've just had a look at the source (0.7.3 libs/gst/control) and there isn't even a c-source for timelined params. Is there already a concept about it somewhere?
[15:32] <mathrick> thomasvs: Terminal(14)? Isn't it a bit too much for a single desktop? ;)
[15:32] <thomasvs> mathrick: no, that's how I work :)
[15:32] <thomasvs> mathrick: usually four terminals on one screen, and several desktops for different projects
[15:33] <mathrick> thomasvs: me too, but I use more than one desktop for terminals
[15:33] <mathrick> thomasvs: digging through 14 of them doesn't sound like fun :p
[15:34] <thomasvs> mathrick: well, I never have to dig
[15:34] <mathrick> thomasvs: howso?
[15:34] <thomasvs> I just go to the right workspace, and there are up to four terminals visisble there
[15:34] <mathrick> thomasvs: hmm, so you list all windows in taskbar, not from single desktop only?
[15:34] <thomasvs> Does anyone care if I terminat each error with period ?
[15:35] <thomasvs> mathrick: yeah, I don't understand why other people can live with only current desktop windows
[15:35] <mathrick> thomasvs: AFIAK HIG does ;)
[15:35] <mathrick> thomasvs: it's the only true way :)
[15:35] <thomasvs> mathrick: I switch a lot more by clicking on an app than going to the workspace
[15:35] <thomasvs> mathrick: what is the only true way ?
[15:35] <mathrick> thomasvs: single desktop only
[15:36] <mathrick> thomasvs: if I wanted full list, I'd have used top-right corner thingy for that
[15:44] Rotty ([email protected]) joined #gstreamer.
[15:46] robUx4 ([email protected]) joined #gstreamer.
[15:46] <robUx4> lo
[15:46] <mathrick> lo robUx4
[15:46] <robUx4> I'm almost done with the compilation of GStreamer under Windows
[15:47] <robUx4> I'm just facing a *lot* of problems with the debug log
[15:47] <markey> guys question: can GST handle shoutcast/icecast streams, or does it rely on GNOMEVFS for that?
[15:47] <robUx4> it seems that there is no real way to get totally rid of it as it is now
[15:47] <robUx4> using GST_DISABLE_GST_DEBUG is just not enough
[15:48] <robUx4> some stuff like GST_CAT_PARAMS gets undefined in this case
[15:48] <markey> mathrick: do you know it?
[15:48] <markey> the streams thingie
[15:53] <mathrick> markey: AFAIK gnomevfs
[15:54] <markey> you are not sure?
[15:54] <markey> who could I ask
[15:55] <mathrick> markey: hmm, though question. Best would be probably to run ./configure | tee plugins.log in gst-plugins, and check plugins list :)
[15:55] <mathrick> s/though/tough/
[15:55] <mathrick> markey: or grep sources
[15:56] <markey> thx
[15:58] <mathrick> markey: and as always, questions of this kind are welcome on -devel list :)
[15:58] <markey> is there some kind of #gst-devel?
[16:03] <mathrick> markey: [email protected] 
[16:03] <markey> ok
[16:06] <ensonic> can anyone make a statement about my question from 15:32 (timelined dparams)?
[16:14] markey ([email protected]) left irc: "bbl"
[16:16] Shoragan ([email protected]) left irc: "Leaving"
[16:21] <ensonic> success i've now build the gst-plugins-0.7.3 on solaris as well (and posted 3 more patches)
[16:22] <mathrick> ensonic: if you're tweaking build system, we'd appreciate if you do this with CVS HEAD
[16:22] <mathrick> ensonic: this way much less chance to conflict with or duplicate things already done
[16:22] <ensonic> I have run gst-register afterward and that generates the registry in my home dir (even for root), the manual says that the registry is in /etc/gstreamer/reg.xml. using gst-register --gst-registry=/etc/... dos not help
[16:23] <mathrick> ensonic: and 0.7.3 is pretty obsolete now anyway
[16:23] <ensonic> mathrick: is the release very old?
[16:23] <ensonic> okay
[16:23] <ensonic> mathrick: I thought I download the tar.gz first and see if gst is the right thing we need for the app we are going to write ;-)
[16:25] <mathrick> ensonic: you're trying to make that buzz thing, right?
[16:25] <ensonic> mathrick: we want to
[16:26] <ensonic> mathrick: atleast to get the ball rolling, as we know many buzz-plugin developers who would gladly port their machines onto a living project ...
[16:26] <mathrick> ensonic: AFAIK DParams are stable, but most of possible transition types aren't written yet (ie. they'll be written when someone needs them badly enough ;)
[16:27] <ensonic> mathrick: thats going to be use (smae applies for midi ;-) ) - anyway I currently checkout cvs and rebuild :-(
[16:28] <ensonic> use=us 
[16:28] <mathrick> ensonic: best post it on -devel, with detailed area of interested, requirements and problems, and have a look at our MIDI docs, in random/, MIDI is under construction now
[16:29] <ensonic> mathrick: we are far from that, it has time
[16:29] <mathrick> ensonic: some preliminary synth support was posted on -devel recently, you may want to take a look
[16:29] <ensonic> errm is the CVS already up on freedesktop.org, the URL is broken
[16:29] <mathrick> ensonic: AFAIR it uses DParams
[16:29] <mathrick> ensonic: yes, it's on fdo
[16:30] <ensonic> what is the CVSROOT
[16:30] <mathrick> :pserver:[email protected]:/cvs/gstreamer
[16:32] <mathrick> ensonic: are you subscribed to -devel? if not, better subscribe
[16:32] <ensonic> mathrick: just done, sf.net mailing list had trouble this week and so I had problems subscribing
[16:33] <mathrick> ensonic: ah, may be, it sucked hard recently :\
[16:33] <mathrick> ensonic: archives ae available on sf and gmane
[16:34] <mathrick> s/ae/are/
[16:34] Nick change: mathrick -> mathrick|afk
[16:34] <ensonic> gmane ?
[16:34] <ensonic> gmane.org - found it
[16:35] <Company> yeah, gst-player is an incredible 16x16 green picture for me
[16:35] <Company> i'm not going to see if video is in sync that way ;)
[16:41] <ensonic> mathrick|afk: i am probably really be off using cvs, the version I build in the last two days just core-dumps, with: sincesrc ! esdsink
[16:49] <thomasvs> Company: whatever video you use ?
[16:53] <Company> thomasvs: my standard test AVI, divx and mp3
[16:54] <Company> alternatively mpeg, doesn't work either
[16:54] <Company> gst-launch without that scaling and colorbalance stuff works fine
[16:55] Marsupilami23 ([email protected]) joined #gstreamer.
[16:55] jcsston ([email protected]) joined #gstreamer.
[17:00] <ensonic> is there a shortcut function to get e.g. all source-only plugins from the registry (i am currently reading the API docs and I am at GstRegistryPool)
[17:01] <thomasvs> grr
[17:01] <thomasvs> of course when you really WANT an error from the core to test error handling you can't make it error
[17:03] jcsston` ([email protected]) joined #gstreamer.
[17:05] <Company> thomasvs: set your audiosink to identity ! audio/x-raw-float ! alsasink and try again ;)
[17:05] <Company> alternatively set it to ximagesink
[17:05] jcsston ([email protected]) left irc: Nick collision from services.
[17:05] Nick change: jcsston` -> jcsston
[17:33] <ensonic> is there a guide from how to build the cvs version, I am a bit confuse about the --disable-plugin-builddir option
[17:34] <ensonic> I have checked out gstreamer, gst-plugins and gst-editor and I am going to [autogen.sh ... , make all install[ them one after the other
[17:35] <ensonic> is that not the way to do it?
[17:35] <Company> ./autogen.sh --disable-plugin-builddir && make && make install
[17:36] <Company> that works for me
[17:36] <ensonic> shit so I have to run autogen again (that will take another 30 minutes) grmpf
[17:36] <ensonic> why is that on by default then?
[17:37] <Company> because autogen does developer builds
[17:37] <Company> and those aren't installed
[17:37] <Company> you can just run configure, too
[17:37] <ensonic> it *is* a developer build
[17:37] <ensonic> what sense does the system make without any plugins
[17:38] <Company> well, you can either run installed or uninstalled
[17:39] <ensonic> atleast this could be put into the README
[17:40] <Company> file a bug about it :)
[17:40] Action: Company always runs installed anyway
[17:41] <ensonic> but when you run it without installing where are the plugins from??, so seems that I am going the safe way (the other seems to be not documented)
[17:44] <thomasvs> ensonic: I run uninstalled almost all the time
[17:44] <thomasvs> http://www.freedesktop.org/~gstreamer/data/doc/gstreamer/head/faq/html/chapter-developing.html#developing-uninstalled-gstreamer
[17:50] Uraeus ([email protected]) joined #gstreamer.
[17:51] Uraeus ([email protected]) left irc: Remote closed the connection
[17:53] ct_ ([email protected]) joined #gstreamer.
[17:54] Nick change: khalek -> kh_zZz
[17:54] Action: Company wonders why seeking in oggs worked
[17:54] <ensonic> thomasvs: oh, thanks - anyway now I leave the autogen as it is and go for the installed scheme
[17:54] <Company> i fixed 5 bugs already and it's still not right ;)
[17:56] <ensonic> oay I lied, autogen just took 20 minutes :-;
[17:57] Uraeus ([email protected]) joined #gstreamer.
[17:59] <Uraeus> ello
[18:01] ds-gromit ([email protected]) got netsplit.
[18:02] ds-gromit ([email protected]) returned to #gstreamer.
[18:03] <Uraeus> ds-gromit: around?
[18:05] apoc ([email protected]) left irc: "Leaving"
[18:12] apoc ([email protected]) joined #gstreamer.
[18:15] sirux ([email protected]) joined #gstreamer.
[18:15] sirux ([email protected]) left irc: Client Quit
[18:18] <Company> thomasvs: could you try seeking in vorbis again?
[18:18] <Company> thomasvs: after updating gst-plugins of course
[18:21] hallibaby ([email protected]) joined #gstreamer.
[18:21] hallibaby ([email protected]) left irc: Client Quit
[18:26] <thomasvs> damn, I can't trigger a single error from the player except for mad not negoing with esd :(
[18:26] <thomasvs> Company: what's the easiest way to improve the seeking algorithm in oggdemux ?
[18:26] <thomasvs> Company: updating and recompiling right now ...
[18:28] <Company> thomasvs: no clue, but it's hard
[18:29] <Company> thomasvs: and unless you have a really clever idea where you'Re sure it definitely works in every case i would keep it as is - it's good enough
[18:31] <thomasvs> Company: well, for one thing it could maintain a cache for example, using the indexing stuff
[18:31] <thomasvs> Company: b/c it still has some lag on seeking
[18:31] <thomasvs> Company: how did vorbisfile do it ?
[18:32] <Company> are you sure that's not because of the queues in the player?
[18:33] <Company> and it's pretty much instantly here
[18:33] jcsston ([email protected]) left irc: Read error: 60 (Operation timed out)
[18:34] herzi ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[18:35] robUx4 ([email protected]) left irc: "Baby wants to bleep"
[18:35] <Company> but yeah, you could make it use an index if you want to spend the time
[18:37] herzi ([email protected]) joined #gstreamer.
[18:42] <ensonic> cvs gstreamer still compiles ;-), but i'll have to leave, cu on monday
[18:45] <ensonic> ciao
[18:45] ensonic ([email protected]) left #gstreamer.
[18:56] bitshifter ([email protected]) left irc: Remote closed the connection
[18:57] bitshifter ([email protected]) joined #gstreamer.
[18:59] Nick change: mathrick|afk -> mathrick
[19:04] <mathrick> hmm, is ffmpeg to be used solely on keep-your-own-tree basis? Or is it intended to be shared lib?
[19:06] <Company> the ffmpeg people suck at releasing
[19:07] <mathrick> Company: so intention is to be shared lib, but practice is to maintain private trees?
[19:08] <Company> i dunno what their intention is
[19:09] markey ([email protected]) joined #gstreamer.
[19:11] <mathrick> wb markey
[19:11] <markey> re
[19:15] <mathrick> Company: ok, let me rephrase. If I ever had idea of depending on "libffmpeg", does that mean I'm idiot to begin with, or just that I lived under the rock for last 3 years? ;)
[19:17] <Company> mathrick: you'd be an idiot if you wanted to depend on "libffmpeg" and you'd be an idiot if you didn't use ffmpeg somehow
[19:17] <mathrick> Company: ok, so libffmpeg never existed?
[19:17] <Company> mathrick: there are releases
[19:18] <Company> mathrick: every other year or so
[19:18] Action: mathrick draws his lart
[19:19] <mathrick> Company: argh, so do they release it in form buildable as shared lib, or not?
[19:20] <thomasvs> mathrick: whlie ffmpeg do releases, they don't want it build as a shared lib
[19:21] <Company> isn't that thing buildable shared?
[19:21] <mathrick> thomasvs: thanks, I get impression Company is playing with me ;)
[19:22] <Company> i really thought that thin was buildable shared and we just didn't use it because it was to infrequent :/
[19:22] <Company> you learn a new thing every day....
[19:24] <thomasvs> it's builddable shared, they just don't want anyone to depend on it as buildable
[19:24] <thomasvs> it's not fpic though
[19:24] <mathrick> thomasvs: do you know why they want it this way?
[19:25] <thomasvs> mathrick: because they want to break API/ABI whenever they feel like it
[19:25] <mathrick> thomasvs: and if it isn't fpic, it's not really "shared"
[19:25] <thomasvs> and because they suck :)
[19:25] <mathrick> thomasvs: ah, that's a reason ;)
[19:31] kmaraas ([email protected]) joined #gstreamer.
[19:34] jcsston ([email protected]) joined #gstreamer.
[19:36] <thomasvs> gst-launch videotestsrc ! agingTV ! xvimagesink
[19:36] <thomasvs> doesn't work anymore, can someone reproduce this ?
[19:37] <Uraeus> [cschalle@localhost cschalle]$ gst-launch-0.7 videotestsrc ! agingTV ! xvimagesink
[19:37] <Uraeus> RUNNING pipeline ...
[19:37] <Uraeus> ERROR: pipeline doesn't want to play.
[19:39] <Company> agingtv outputs rgb, xvimagesink requires yuv
[19:40] <Uraeus> yup, adding ffcolorspace makes it work
[19:41] Uraeus ([email protected]) left irc: "Client exiting"
[19:44] <thomasvs> Uraeus: so for you it works ?
[19:44] <thomasvs> crap
[19:44] <thomasvs> Company: seems to work now with both esdsink and osssink
[19:44] <thomasvs> Company: what did you fix ?
[19:51] <Company> thomasvs: discont and flush event handling
[19:52] <thomasvs> [gst-cvs] [thomas@otto gst-plugins]$ gst-launch videotestsrc ! colorspace ! agingTV ! colorspace ! xvimagesink
[19:52] <thomasvs> WARNING: erroneous pipeline: could not link agingtv0 to colorspace1
[19:52] <thomasvs>          Trying to run anyway.
[19:56] <Company> ffcolorspace
[19:56] <Company> or fix colorspace ;)
[19:57] <thomasvs> heh :9
[19:57] <thomasvs> Company: if an app wants to know when a seek has completed - how should it do that ?
[19:57] Action: mathrick notes that "I wish media support on Linux were so good as on windoze" is becoming "I wish I'd never have to go back to windoze" :)
[19:59] <Company> thomasvs: a seek is not completed...
[20:02] <Company> thomasvs: that convept just does not exist
[20:05] <thomasvs> Company: right - but if we would want to have that, then what would we do ?
[20:05] <thomasvs> Company: here's an example
[20:05] <thomasvs> you have a player with a slider
[20:05] <thomasvs> you want to block updating of the slider until seek is ocmpleted
[20:06] <Company> don't update the slider for half a second
[20:07] <thomasvs> ugh, that's bogus :)
[20:07] <mathrick> Company: hmm, too naive
[20:07] jdahlin ([email protected]) joined #gstreamer.
[20:07] <Company> you cannot expect a streaming framework to do such a thing
[20:07] ct_ ([email protected]) left irc: Read error: 110 (Connection timed out)
[20:08] <mathrick> Company: seeking in between keyframes on slow CPU can take 5s
[20:08] <thomasvs> Company: why can't you expect a framework to do that ?
[20:08] <jdahlin> Company, why can't just a signal be emitted (or something similar) when the seek is finished and the data is read from a new position?
[20:08] <Company> who should emit that signal?
[20:08] <thomasvs> Company: all it would need is for some way to detect that element where you inject the seek has seen a successful seek
[20:08] <jdahlin> it would be very useful for an application to block the update of the position of the gui control until such an event is emitted
[20:09] <Company> maybe the element does never see that
[20:09] <Company> you'd say the seek is done when the seek event was successfully sent upstream and a discont arrived at the element where you inserted the seek
[20:09] <mathrick> Company: does sink get DISCONT on seek?
[20:10] <Company> that would work _unless_ there is no discont emitted or a discont was emitted earlier for other reasons
[20:12] <mathrick> Company: when could be discont emitted before seek, and not get emitted afterwards?
[20:13] <Company> before if the file is "broken" and you jump over some data
[20:13] <Company> and not at all if the seek was not far enough so the same data should be displayed
[20:13] <Company> and probably other reasons
[20:14] <mathrick> Company: if the same data is displayed, there is no reason to advertise seek, IMO
[20:14] <mathrick> Company: I'm rather interested in that other reasons
[20:15] <mathrick> s/reason to advertise/need to advertise/
[20:15] <Company> lemme think that through
[20:16] <Company> (all of this doesn't work anyway, because there is no real way to peek inside the stream)
[20:16] <mathrick> thomasvs, jdahlin: hmm, isn't querying whatever elem you usually get time from enough?
[20:16] <mathrick> thomasvs, jdahlin: ie, if time is updated, your gui control will get updated too
[20:17] <thomasvs> mathrick: here's the thing
[20:17] <jdahlin> mathrick, no, you'll still get some things reported before it seeked
[20:17] <thomasvs> mathrick: a) you drag the slider to a new pos
[20:17] <thomasvs> b) this tells gst to do a seek
[20:17] <thomasvs> c) this seek takes time to process, and in between you still get a callback for the current time, which is still at the previous time, so it jumps back
[20:17] <thomasvs> d) as the seek is processed, the slider jumps back forward because the seek is done and the callback gets triggered again
[20:18] <mathrick> thomasvs, jdahlin: there is little practical difference between "time advanced by 1s" and "time advanced by (-)100s"
[20:18] <Company> the easiest way is to block "jumping back"
[20:18] <mathrick> thomasvs: hmm, lemme think
[20:18] <thomasvs> mathrick: all other frameworks solve this because they block the update callback until the seek is done
[20:18] <Company> for a specified amount of time
[20:19] <Company> btw: every element is free to initiate seeks itself, so it might just seek back on a discont (for whatever stupi reason)
[20:19] <mathrick> thomasvs: hmm, I don't get it. Seek callback should be called only on user triggered slider move, right?
[20:19] <mathrick> thomasvs: otherwise, you'd end up seeking every second b/c slider is updated once per second
[20:20] <Company> slider is updated more often
[20:20] <mathrick> Company: technical detail
[20:20] <Company> 1024 pixels/4 minutes
[20:20] <mathrick> Company: "specified time" isn't solution
[20:20] <Company> i think it is a good enough solution
[20:21] <mathrick> Company: at least, not better than char filename[30];
[20:21] <Company> you don't even know if the pipeline will actually seek
[20:21] <mathrick> thomasvs: I still don't see your concern
[20:21] <Company> no, it's not that much better than filename[30], but it's not possible to do better
[20:22] <Company> imagine you watch a partially downloaded file wher you have minutes 1-10 and 50-60 and you seek to 30 minutes
[20:22] <Company> what should happen?
[20:23] <mathrick> Company: hmm, it should probably stall on undefined (== "last known") time
[20:23] <thomasvs> mathrick: the slider gets updated through an idler, say, every 1/10th of a sec
[20:24] <mathrick> Company: until someone seeks to valid time
[20:24] <thomasvs> mathrick: when you drag it to a different position, the idler is still active
[20:24] <thomasvs> mathrick: so the idler will put it back to it's original position because it queries the position at a time where the seek isn't completed yet
[20:24] <mathrick> thomasvs: does that emit ::value-changed?
[20:24] <thomasvs> mathrick: does what emit that ?
[20:24] <mathrick> thomasvs: it doesn't matter
[20:24] <Company> just disable the slider for $time
[20:25] <mathrick> thomasvs: does idler updating trigger GtkScale::value-changed?
[20:25] <thomasvs> mathrick: idler updating just sets the position, period
[20:25] <thomasvs> Company: that's not a good solution - that's just wet finger guessing
[20:25] <mathrick> thomasvs: hmm, ok, so why is idler setting bogus position for, say, 3s concern to you?
[20:25] <thomasvs> mathrick: run the player, you'll see it soon enough :)
[20:26] <thomasvs> mathrick: because the visual effect is
[20:26] <thomasvs> a) slider is at point A
[20:26] <Company> thomasvs: yes - but it works
[20:26] <thomasvs> b) user slides it to B)
[20:26] <thomasvs> c) app throws it back to A
[20:26] <thomasvs> d) app throws it back to B
[20:26] <mathrick> thomasvs: ahh, I understand now
[20:26] <thomasvs> mathrick: ie, instead of just staying at the pos you set it to, it jumps back and forth
[20:26] <Company> thomasvs: at least it works better than just assuming stuff about the pipeline
[20:26] <thomasvs> which, to a user, says, "boy these guys don't know how to do it, they suck"
[20:27] <thomasvs> Company: so instead of assuming, why don't we find a right way to do it ?
[20:27] <thomasvs> a way that says, "hey, you asked me to seek, I can tell you know that the seek is done" ?
[20:27] <mathrick> Company: is it possible for app to connect to DISCONT?
[20:27] <thomasvs> Company: just like gtk does similar things for widgets ?
[20:27] <Company> thomasvs: if you find one go ahead :)
[20:27] <Company> thomasvs: gtk works synchronously
[20:30] <Company> i can even imagine seeks that take 10+ seconds if you stream a seekable stream over some lame web connection
[20:30] <Company> s/web/net/
[20:32] <Company> the actual problem is that you use the time display for seeking
[20:37] <mathrick> Company: hmm, is adding "seek-complete" event that much of hassle? It could be emitted whenever seek is completed, or just in cases where there is no discont emitted
[20:37] <mathrick> Company: this way element that never suppresses discont wouldn't have to bother implementing it
[20:47] Zeenix ([email protected]) joined #gstreamer.
[20:48] <Zeenix> hello
[20:49] <mathrick> lo Zeenix
[20:51] kmaraas ([email protected]) left irc: Read error: 110 (Connection timed out)
[20:57] apoc ([email protected]) left irc: Connection timed out
[20:57] apoc ([email protected]) joined #gstreamer.
[21:07] Shoragan ([email protected]) joined #gstreamer.
[21:20] sublime ([email protected]) joined #gstreamer.
[21:47] jdahlin ([email protected]) left irc: "Leaving"
[21:57] Company ([email protected]) left irc: Remote closed the connection
[21:59] apoc_ ([email protected]) joined #gstreamer.
[22:07] apoc ([email protected]) left irc: Read error: 60 (Operation timed out)
[22:07] sublime ([email protected]) left irc: Read error: 110 (Connection timed out)
[22:13] trow ([email protected]) joined #gstreamer.
[22:27] ct_ ([email protected]) joined #gstreamer.
[22:43] subleet ([email protected]) left irc: "I like food, food is good!"
[22:44] ct_ ([email protected]) left irc: Client Quit
[22:51] hyriand ([email protected]) left irc: Remote closed the connection
[22:53] hyriand ([email protected]) joined #gstreamer.
[23:17] hallibaby ([email protected]) joined #gstreamer.
[23:17] <hallibaby> Offtopic: i think i need a regexp guru. is here anybody who can help?
[23:36] jcsston` ([email protected]) joined #gstreamer.
[23:38] <mathrick> hallibaby: how much of guru do you need?
[23:40] Action: ds is a regexp guru
[23:40] jcsston ([email protected]) left irc: Nick collision from services.
[23:40] Nick change: jcsston` -> jcsston
[23:41] Action: mathrick is pretty fluent
[23:41] <thomasvs> so, the player throws this:
[23:41] <thomasvs> http://thomas.apestaart.org/download/screenshots/gst-player-error-1.png
[23:42] <thomasvs> and when you click on information
[23:42] <thomasvs> http://thomas.apestaart.org/download/screenshots/gst-player-error-2.png
[23:42] <thomasvs> any ideas to improve this ?
[23:42] <thomasvs> (btw, I'm a bit disappointed that I get such a bad error message out of something that basically is "esd cannot handle this samplerate" - I wish I could improve on this somehow)
[23:43] <ds> it's often hard to figure out who is at fault in negotiation problems
[23:44] <bitshifter> thomasvs: if the error message details are always that short, then why not show it all the time by default? no point having an extra button there for that ...
[23:44] <bitshifter> thomasvs: you could also see how it looks if you put the details into a frame
[23:45] <mathrick> bitshifter: there's no guarantee on message length, besides, it shouldn't scare users with cryptic messages by default
[23:46] <thomasvs> bitshifter: the debug message is untranslated and userunfriendly, only for debugging
[23:46] <thomasvs> bitshifter: I tried doing the frame thing though
[23:46] <thomasvs> ds: yeah, agreed, but we really should make it easier than this though.
[23:46] <thomasvs> ds: maybe the debug should contain caps on both sides or something, to give a hint
[23:46] Action: mathrick is constanatly amazed by how much repeated letters it is possible to use in single word ;)
[23:47] <ds> thomasvs: the way we do it, it's often not possible to tell which two elements _actually_ can't negotiate
[23:47] <ds> thomasvs: because the error may show up somewhere else
[23:47] <ds> for example
[23:48] <ds> in ... mad ! audioscale ! audioconvert ! esdsink
[23:48] <ds> if (hypothetically) mad only supported float audio, you could get negotiation errors in any of the links
[23:48] <bitshifter> thomasvs: what about putting the message into a GtkTextView, possibly with a 'copy to clipboard' button or so?
[23:49] <ds> you can almost always select text in dialog boxes
[23:49] <ds> no need to pretend it's special
[23:49] <mathrick> isn't GtkMessageBox label selectable by default?
[23:50] <mathrick> s/Box/Dialog/
[23:52] <mathrick> am I wrong, or it isn't possible to create GtkMessageDialog from Glade?
[23:54] Action: sxpert_ is back (gone 10:24:13)
[23:55] Nick change: sxpert_ -> sxpert
[23:59] sublett ([email protected]) joined #gstreamer.
[00:00] --- Sun Feb  1 2004
[00:23] <Zeenix> eek i get this error on compling gst: 
[00:23] <Zeenix> /usr/bin/msgfmt: nl.po: warning: Charset missing in header.
[00:23] <Zeenix>                                  Message conversion to user's charset will not work.
[00:23] <Zeenix> /usr/bin/msgfmt: nl.po: headerfield `PO-Revision-Date' missing in header
[00:23] <Zeenix> /usr/bin/msgfmt: nl.po: headerfield `Last-Translator' missing in header
[00:23] <Zeenix> /usr/bin/msgfmt: nl.po: headerfield `Language-Team' missing in header
[00:23] <Zeenix> /usr/bin/msgfmt: nl.po: headerfield `MIME-Version' missing in header
[00:23] <Zeenix> /usr/bin/msgfmt: nl.po: headerfield `Content-Type' missing in header
[00:23] <Zeenix> /usr/bin/msgfmt: nl.po: headerfield `Content-Transfer-Encoding' missing in header
[00:23] <Zeenix> nl.po:9:2: parse error
[00:23] <Zeenix> nl.po:9: keyword "nl" unknown
[00:24] <Zeenix> nl.po:425:2: parse error
[00:24] <Zeenix> nl.po:425: keyword "nl" unknown
[00:24] <Zeenix> nl.po:436:2: parse error
[00:24] <mathrick> Zeenix: looks like borked checkout
[00:25] markey ([email protected]) left irc: "nite"
[00:26] <Zeenix> mathrick: what does borked-out means?
[00:27] <mathrick> Zeenix: borked cvs checkout, ie. your cvs failed somewhere in the process
[00:27] herzi ([email protected]) left irc: Read error: 110 (Connection timed out)
[00:30] <Zeenix> mathrick: no errors but what does these lines (during the update) means:
[00:30] <Zeenix> C po/fr.po
[00:31] <Zeenix> C po/nl.po
[00:34] <mathrick> Zeenix: it means conflicts while updating
[00:34] <mathrick> Zeenix: rm these files and redo
[00:35] <mathrick> Zeenix: of course you use -dP options while updating?
[00:36] <Zeenix> mathrick: i tried -dP & it started to speak marsian language with me :)
[00:36] <Zeenix> mathrick: i was getting a strange looking endless XML 
[00:37] trow ([email protected]) left irc: "Leaving"
[00:37] <mathrick> Zeenix: heh, it means you have some conflicting changes piled up. Unless you've modified these files, just rm all conflicting ones and retry
[00:45] <Zeenix> mathrick: i think i got it..
[00:51] pb_ ([email protected]) left irc: "Client exiting"
[00:51] <hallibaby> @mathrick: sorry for delay. was a bit busy with this *peep* regexp?
[00:51] <mathrick> hallibaby: what do you want to do?
[00:52] <hallibaby> @mathrick: is there any possibility to difference [xxx]bla and [xxx]blub with end-of-line and any other whitespace?
[00:52] <mathrick> hallibaby: huh?
[00:52] <mathrick> hallibaby: could you clear up a bit?
[00:52] <hallibaby> [xxx](.*)\s does not work for replacement :(
[00:53] <hallibaby> @mathrick: i would like to colour [xxx]bla and any other words prefixed by [xxx]
[00:54] <hallibaby> mathrick: i thought the regexp: [xxx](.*)\s should find any prefixed [xxx] sentence.
[00:54] <mathrick> hallibaby: ok, so you want to match any word prefixed by [xxx]
[00:54] <hallibaby> but :(
[00:54] <hallibaby> yes
[00:55] <hallibaby> followed by any whitespace or newline (\n)
[00:55] taf2 ([email protected]) joined #gstreamer.
[00:55] <mathrick> hallibaby: \[xxx\][^$]*[[:blank:]]
[00:55] <mathrick> hallibaby: try this one
[00:55] <hallibaby> check it out 
[00:57] <hallibaby> mathrick: perheps i shoould mention it in php
[00:57] <hallibaby> ?
[00:57] <mathrick> hallibaby: one moment
[00:57] <mathrick> \[xxx\][^\n]*
[00:58] <mathrick> hallibaby: it will match whole sentence up to, but not including, newline
[00:58] <hallibaby> try again :)
[00:59] <mathrick> hallibaby: where sentence == everything from [xxx] on
[00:59] <mathrick> hallibaby: works here :)
[00:59] <mathrick> hallibaby: in Emacs :)
[00:59] <mathrick> hallibaby: probably \n confuses it
[00:59] <mathrick> hallibaby: it's usually tricky to match \n in various regexp implementations
[01:00] <mathrick> hallibaby: or, try doubling every \
[01:00] <hallibaby> it finds the prefix but not the following :(
[01:00] <hallibaby> this result i know :(
[01:01] <hallibaby> i try to replace found exps by \\1 \\2 and so on
[01:01] <mathrick> hallibaby: it's strange
[01:01] <hallibaby> regexp is hard
[01:02] <mathrick> hallibaby: no it's not
[01:02] <hallibaby> maybe for you ;)
[01:02] <mathrick> hallibaby: it's just quirky implementations that make using it hard
[01:02] <hallibaby> but shouldn't \s match any whitspace following?
[01:03] <mathrick> hallibaby: i don't know \s, it's some extension :)
[01:03] <mathrick> hallibaby: posix says [[:blank:]] is what you should use
[01:03] <hallibaby> and isn't told that \\0 \\1 and \\2 and so on matches the found part
[01:03] <mathrick> hallibaby: yes, they do
[01:03] <mathrick> hallibaby: and what about that?
[01:04] <hallibaby> then i begin to be confused?
[01:04] <mathrick> hallibaby: so you want to match _sentence_ or _word_ prefixed with [xxx]?
[01:04] <hallibaby> i try to find prefixed words like [xxx] or to be precise [GMC]
[01:04] <hallibaby> yes
[01:04] <hallibaby> and i want to replace them with colour in html via php
[01:05] <hallibaby> prep_replace
[01:05] <mathrick> hallibaby: ok, should "[xxx]foo bar" match up to foo or up to bar?
[01:05] <hallibaby> "[xxx]foo bar" should result in "[xxx]foo" and then "bar"
[01:06] <hallibaby> but it could happen to be only "[xxx]foo" or "[xxx]foo(\ \.\n\t and so on)". tricky!
[01:07] <mathrick> argh, I don't get you
[01:07] <mathrick> hallibaby: is bar in matches or not?
[01:07] <hallibaby> php thinks also ;)
[01:07] <hallibaby> bar is not in match
[01:08] <mathrick> hallibaby: ok, so it's different from what you told earlier :)
[01:08] <hallibaby> sorry, my english explanation may confuse :(
[01:09] Nick change: kh_zZz -> khalek
[01:09] <mathrick> \[xxx\][[:graph:]]*[[:blank:]]*
[01:10] <mathrick> hallibaby: try this one
[01:10] <hallibaby> i will :)
[01:11] <mathrick> hallibaby: so far it only matches, if you want to split it in parts, it'll be a bit different
[01:11] ChrisHJW ([email protected]) left irc: Read error: 60 (Operation timed out)
[01:11] <hallibaby> [:graph:] isn't recognize as POSIX class name?!?!? is there any abbreviation for this?
[01:11] <hallibaby> php tells me :(
[01:12] <mathrick> hallibaby: no, graph is the only one I know about
[01:14] <mathrick> hallibaby: you can try [[:alnum:][:punct:]] instead of [[:graph:]]
[01:14] <hallibaby> is more tricky as i thought. but you you know some good docs about regexp? some url or so? 
[01:14] <mathrick> hallibaby: or, maybe \w will work
[01:16] <mathrick> hallibaby: info gawk is particularly good and verbose
[01:17] <mathrick> hallibaby: it depends on you having gawk docs installed however
[01:17] <mathrick> hallibaby: also it discusses in detail slash escaping rules which may also apply to php
[01:18] <hallibaby> think i should read more ;) thanks
[01:19] <hallibaby> but php seems to recognize [[:graph:]] if written as [:graph:]!?!?
[01:19] <hallibaby> no more posix violation explationations
[01:20] <mathrick> hallibaby: hmm, it's serious posix violation then
[01:20] <Misirlou> PHP sucks. :)
[01:21] <mathrick> hallibaby: try matching [[:graph:]] to "asdasdasd"
[01:21] <mathrick> hallibaby: does it apply?
[01:21] <hallibaby> i like php. but i come to regexp????? everytime i have these problems. in shell scripts it seems easier to me. maybe i'm stupid :(
[01:22] <mathrick> hallibaby: some implementations are really quirky
[01:22] <mathrick> hallibaby: did the above apply?
[01:23] <hallibaby> does it work too good :)==) every word is changed!
[01:23] <hallibaby> ahhhh... trying something
[01:24] <hallibaby> :)=======))) yes
[01:24] <hallibaby> yes
[01:24] Last message repeated 1 time(s).
[01:24] <hallibaby> thanks
[01:24] <mathrick> hallibaby: np
[01:24] <hallibaby> :)==)
[01:25] <hallibaby> i'm happy
[01:25] <mathrick> hallibaby: regexp is really easy after you get it completely for the first time :)
[01:25] <mathrick> hallibaby: from then on, it's only matter of few trials and having good regex builder handy ;)
[01:26] <hallibaby> i used it in shell scripts. wasn't really hard! you're right. but in php i have a better look at it.
[01:26] <hallibaby> there some options smUi. need to read about them
[01:27] <mathrick> hallibaby: I suggest you use some builder for that, like one in Emacs, or standalone, there are plenty available
[01:27] <hallibaby> "\[xxx[[:graph:]]*[:blank:]*}smUi" was the prob. but deleting smUi and the worked as you told
[01:27] <mathrick> hallibaby: good builder can really speed up your regex work :)
[01:27] <hallibaby> builder?
[01:27] <hallibaby> i use only brain
[01:28] <hallibaby> that was the problem here
[01:28] <hallibaby> :(
[01:28] <mathrick> hallibaby: yep, regex construction assistant tool
[01:28] <hallibaby> but i should use shell more for testing for example
[01:29] <hallibaby> sh** my english is horrible. delete last sentence
[01:30] <mathrick> hallibaby: builders are better suited for that, they provide on the fly visual feedback, so you can spot errors almost instantly
[01:31] <hallibaby> any suggestion?
[01:31] herzi ([email protected]) joined #gstreamer.
[01:32] <mathrick> hallibaby: I use Emacs elisp package for that, and as whole my work is done with emacs, I have no need for separate app
[01:33] <mathrick> hallibaby: if you're not emacs user however, there's a flock of them, just look on freshmeat, or your distro pkgs repository
[01:34] <hallibaby> i will have a look at emacs or some emacs-like apps :)=) but thanks for all even if offtopic ;)
[01:35] <hallibaby> now it's time to go to bed ;)
[01:35] <mathrick> hallibaby: true :)
[01:35] Action: mathrick goes to bed
[01:35] <mathrick> nite
[01:35] <hallibaby> even so
[01:35] <hallibaby> nite
[01:35] Action: hallibaby is away: I'm busy
[01:35] Nick change: mathrick -> mathrick|sleep
[01:46] KyleB ([email protected]) left irc: Read error: 110 (Connection timed out)
[02:03] taf2 ([email protected]) left irc: Read error: 110 (Connection timed out)
[02:09] wereHamster ([email protected]) joined #gstreamer.
[02:15] <wereHamster> If I have a pipeline with a filesrc followed by a demuxer, the demuxer has to have the ability to go forward and backward in the stream.. right? Imagine a file which has audio/video not interleaved, So the demuxer first reads a video frame (at the beginning of the file) and then the corresponding audio frame (somewhere at the end of the file).
[02:18] Nick change: harshy -> harshyOut
[02:22] <wereHamster> I should perhaps read the source...
[02:22] wereHamster ([email protected]) left irc: 
[02:24] Zeenix ([email protected]) left irc: "Leaving"
[02:44] edmack ([email protected]) joined #gstreamer.
[02:46] <edmack> Hi, I'm using Fedora, have installed libmad and id3lib, and am having difficulties getting an mp3 decoder to appear in gst-inspect-gui
[02:46] <edmack> [edmack@linpc1 edmack]$ rpm -qa | grep gst | grep mad
[02:46] <edmack> gstreamer-mad-0.6.3-0.fdr.1.rh90
[02:51] <edmack> hmmm Plugin /usr/lib/gstreamer-0.6/libgstmad.so failed to load


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