IRC Logs

*******************************************************************
[03:00] md` ([email protected]) joined #gstreamer.
[03:02] <Company> ds-work: do we support multiple connections between elements?
[03:02] <Company> ds-work: like src ! oneton ! ntoone ! sink  onton0. ! ntoone0.
[03:08] <ds-work> Company: yes.  wingo will cause you harm if you break that
[03:08] <ds-work> Company: I have no personal experience that it actually works, however
[03:08] <Company> ds-work: it works if you push on oneton in the same order you pull from ntoone
[03:09] <Company> which basically means you could just use one pad and define yourself a caps that provides that...
[03:09] <ds-work> Company: well, technically, ntoone should be pulling from pads equally
[03:09] <Company> equally != in predefined order
[03:10] <ds-work> no, as in using gst_pad_select(), if it works/worked
[03:10] <Company> yeah, that would be ok
[03:10] <Company> (in fact, it'd be easy to add support for that in entry, but since noone uses it...)
[03:11] Action: mathrick goes to sleep
[03:11] <ds-work> nobody uses it because it's broken :)
[03:11] <ds-work> maybe
[03:11] <Company> gst_pad_select sucks because it claims you must use gst_pad_pull afterwards but doesn't actually do that for you
[03:11] Nick change: mathrick -> mathrick|sleep
[03:11] <ds-work> I think wingo uses it
[03:12] <Company> neither core nor plugins use it
[03:12] <Company> apart from aggregatr
[03:12] <ds-work> $(all elements) != gstreamer + gst-plugins + gst-ffmpeg
[03:12] <Company> i know
[03:13] <Company> i'm pretty sure wingo doesn't use gst_pad_select btw
[03:14] <ds-work> perhaps
[03:14] <Company> wingo uses opt and opt doesn't implement it
[03:16] spyder__ ([email protected]) joined #gstreamer.
[03:16] Rotty ([email protected]) left irc: "Bye"
[03:17] Action: Company just decided entry is not gonna support it
[03:17] <ds-work> uh...
[03:17] <Company> (unless someone else does it of course)
[03:17] <ds-work> it actually _would_ be useful if the schedulers supported it
[03:17] <Company> not pad_select
[03:17] <ds-work> but it certainly isn't a priority
[03:18] ChriHJW ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[03:18] ChrisHJW ([email protected]) joined #gstreamer.
[03:18] <Company> the current way multiple connections between elements are handled
[03:18] <Company> we need to change the API for gst_pad_select
[03:19] <ds-work> gtk-doc can kiss my ass
[03:19] <ds-work> gst_pad_select_2()
[03:19] <Company> to GstData *gst_pad_select (GList *pads, GStPad **used_pad);
[03:19] <ds-work> sure
[03:20] <Company> otherwise people will not call gst_pad_pull
[03:21] spyder482 ([email protected]) left irc: Read error: 113 (No route to host)
[03:21] <ds-work> how can they use gst_pad_select() currently and not use gst_pad_pull()?
[03:22] <Company> gst_pad_select returns the pad that has a buffer
[03:23] <Company> the scheduler caches the buffer and provides it when the element (hopefully) calls gst_pad_pull after that
[03:24] <ds-work> ah, ok
[03:24] <ds-work> so it's just that you need to add a stupid workaround in the scheduler because of an API misfeature
[03:24] <ds-work> I say add gst_pad_select_2() instead
[03:26] <Company> doesn't solve a thing
[03:27] <ds-work> why?
[03:27] <ds-work> which problem?
[03:27] <Company> unless we add a new function to GstSchedulerClass's vtable
[03:28] <ds-work> we can do that, as we have available padding
[03:28] thaytoo ([email protected]) joined #gstreamer.
[03:28] <Company> hm
[03:28] <Company> i don't need it for entry i think
[03:29] <Company> it's just that we need to make sure to change this in 0.9
[03:30] <Company> ds-work: can you note that in you 0.9 changes file?
[03:31] bitshifter ([email protected]) left irc: "bye"
[03:32] <ds-work> added
[03:32] <ds-work> I called it gst_pad_pull_many()
[03:32] <ds-work> since that makes more sense, imo
[03:32] <Company> fine with me
[03:33] <Company> we can implement that already, just use gst_pad_select and then gst_pad_pull ;)
[03:33] <Company> gst_pad_pull_select makes most sense to me
[03:34] sub_pop ([email protected]) left irc: "Client exiting"
[03:35] <ds-work> select() is misnamed.  let's not continue the tragedy :)
[03:35] bitshifter ([email protected]) joined #gstreamer.
[03:35] <Company> "many" sounds like getting many buffers
[03:35] <ds-work> besides, it doesn't do exactly what select does
[03:35] <ds-work> well, it does now
[03:36] <Company> yeah
[03:36] md` ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[03:38] sub_pop ([email protected]) joined #gstreamer.
[03:41] arana ([email protected]) left irc: Read error: 110 (Connection timed out)
[03:45] spyder482 ([email protected]) joined #gstreamer.
[03:54] spyder__ ([email protected]) left irc: Read error: 110 (Connection timed out)
[04:00] sub_pop ([email protected]) left irc: "Client exiting"
[04:00] Nick change: mxpxpod -> mx|gone
[04:05] sub_pop ([email protected]) joined #gstreamer.
[04:06] <Company> grah
[04:06] <Company> AC_CHECK_LIB doesn't work with C++
[04:06] <Company> at least not if you use namespaces
[04:19] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[04:28] <wheels> Company: re
[04:29] <Company> wheels :)
[04:30] <Company> wheels: how do i write an autoconf check for libkio?
[04:30] <Company> (without needing the admin dir of KDE cvs ;)
[04:32] <wheels> Company: Give me a sec -- I'll just read the relevant portion of the admin dir.  :-)
[04:32] <wheels> Looks like I picked a bad night to be away for a bit.  You guys could have used some lecturing better.  ;-)
[04:32] <wheels> s/better/earlier/
[04:33] <Company> hehe
[04:33] Action: wheels is still half asleep.  For some reason I've already woken up after going to bed at 23:00...
[04:33] <Company> we could repeat this while i wait for g++ to compile :p
[04:33] <wheels> Company: Well, you won't find me defending g++.  :-)
[04:34] <wheels> Company: There's a reason that I still use GCC 2.95.  ;-)
[04:35] <Company> wheels: anyway, i'm gonna commit kiosrc soon - i'll leave integrating it properly with KDE up to you ;)
[04:36] <wheels> Company: Ok, so does that mean "please nitpick"?  :-)
[04:36] <Company> wheels: you do have commit access, right?
[04:36] <wheels> Company: yeah
[04:36] <Company> wheels: i'll just commit it and leave it up to you :p
[04:36] <wheels> Company: I think I've done all of 3 commits, but...  :-)
[04:37] <wheels> Company: Hmm, I guess this is going into post-0.8.x at this point though, eh?
[04:37] <Company> wheels: no, it'll be part of 0.8.1
[04:37] <wheels> Company: Ah, ok -- I wasn't really sure what the policy was for new plugins in the stable branch...
[04:38] <Company> wheels: we'll keep HEAD the devel branch for a while
[04:38] <wheels> Hmm, ok -- for some reason I thought it was branched at 0.8.0 -- not just tagged...
[04:39] <Company> nope
[04:39] <Company> we're trying to see if this method works better than last time
[04:39] <Company> ... where we had to force all devels to backport their fixes
[04:39] <wheels> Company: do you have a "CCMAIL:..." equivalent with xdg cvs?
[04:40] <Company> wheels: nope - people are reading the commits list though
[04:40] <wheels> Company: it would have been nice to CCMAIL [email protected]...
[04:41] <wheels> I may just find and forward the CVS mail.
[04:41] Action: wheels would have trouble living without CCMAIL at this point.  It's truly an important part of my life.  :-)
[04:42] Action: wheels does a preperatory rebuild to work out all of the 2.95 issues for the night.  :-)
[04:42] <Company> yeah, it sounds like a clever concept
[04:42] <Company> you could file an enhancement bug - it shouldn't be that hard to adapt the mail script
[04:43] <wheels> Company: Hmm, does GNOME not have such?
[04:43] Action: wheels uses it all the time for "CCMAIL:[email protected]" and so on...
[04:44] <wheels> Dammit.  Gst wants a newer bison now...
[04:44] <Company> nope, don't think so
[04:44] <Company> it does?
[04:44] <wheels> configure: error: Your bison version is too old, v1.35 or later is required.
[04:44] <wheels>   configure failed
[04:45] Action: wheels doesn't rebuild Gst more than once every couple of weeks lately since it seems like every time that I do I have to upgrade something or another.  ;-)
[04:45] <Company> yeah, there were some issues with 1.28
[04:45] <Company> uhm, bison 1.28 is at least a decade old, no?
[04:45] <Company> what do you run? slackware 2.0?
[04:47] <wheels> Company: SuSE 8.0  :-)
[04:47] <wheels> Company: Which is really only about 2 years old...
[04:47] <ds-work> bison-1.35 is relatively recent, iirc
[04:48] <wheels> But it's ok -- I just ripped of the SUSE 9.0 package and rebuilt it for 8.0...
[04:48] <wheels> s/of/off/
[04:49] <Company> 25.03.02
[04:49] <bitshifter> http://bugzilla.gnome.org/show_bug.cgi?id=127838 could be closed then, I suppose
[04:50] <wheels> Company: My system is pretty hacked up -- I've rebuilt most of the major stuff in it -- but I keep it because (a) it would take me a long time to duplicate my setup if I moved to a later distro, (b) I'm using older gcc and binutils which are much faster than current versions and (c) it's a decent testbed for a distro that's not really that old...
[04:51] <Company> bitshifter: thanks
[04:53] <Company> wheels: update gst-plugins and fix it :)
[04:53] <wheels> Company: will shortly  :-)
[04:55] Action: Company wonders if he should sttart an argument with wheels about why kde sucks now ;)
[04:55] <wheels> Company: Sure, we can call it "Reflections on the divine"  :-)
[04:57] <Company> we could call it "why you don't make stuff global variables", too
[04:57] <Company> but that sounds much too technical
[04:58] <wheels> Oh, and I still forgot to find the appropriate check...
[04:58] <Company> feel free to update it
[04:59] <Company> the only cool thing now is that the author of kiosrc is [email protected] ;)
[05:01] <wheels> Yeah, you can tell from the code.  :-)
[05:02] <wheels> (It mostly uses C programming conventions instead of C++.)
[05:03] <Company> i tried to keep C++ in kioreceiver.cpp and use C in kiosrc.cpp
[05:04] <wheels> Company: I guess you probably need an extern "C" { ... } block in the header as well...
[05:06] <Company> wheels: G_BEGIN_DECLS does that
[05:06] <wheels> ah
[05:07] Nick change: mx|gone -> mxpxpod
[05:08] <wheels> Company: One nitpick of course is that it's "true" and "false" in C++...
[05:10] <Company> wheels: I _tried_ to use it, but the habit, the habit...
[05:11] <wheels> Company: Sure -- fortunately when I write C the compiler tends to catch such things.  :-)
[05:11] <Company> yeah, the C++ compiler catches a lot of other things... ;/
[05:17] <Company> i suck
[05:17] <Company> cvs ci without specifying the file...
[05:18] <wheels> heh
[05:23] dholmes ([email protected]) left irc: "Leaving"
[05:42] <wheels> Company: Hmm, are all of the checks for every plugin really done in just one configure.ac?
[05:44] <Company> wheels: yup
[05:44] <wheels> eek  ;-)
[05:44] <Company> wheels: fortunately they're alphabetical ;)
[05:45] <wheels> Company: Ok, looks like first order will of business will be rewriting the check...
[05:46] arana ([email protected]) joined #gstreamer.
[05:46] spyder482 ([email protected]) left irc: Read error: 110 (Connection timed out)
[05:47] <Company> wheels: how about using pkg-config?
[05:47] <wheels> Company: KDE, you mean?
[05:47] <Company> wheels: yeah
[05:48] <wheels> Company: I would suspect that we will in KDE 4.  It was just such an annoyance when people first started using it that nobody was in a hurry...
[05:49] <Company> it's really nice right now
[05:49] <Company> our configure checks all mutate to 1-liners :)
[05:51] <wheels> Company: Yeah, well, the nasty thing is to do these checks properly (again, sans KDE admin dir) it'll not be small.  :-)
[05:51] <wheels> Company: (Just getting the Qt check right takes a bit.)
[05:52] jcsston ([email protected]) joined #gstreamer.
[05:54] <Company> wheels: isn't that a qt bug? all libs gst uses are able to use pkg-config one-liners
[05:54] <wheels> Company: Well, if it were reported as a bug I would mark it as WISHLIST.  ;-)
[05:55] <Company> yeah, ok - but at least it'd be a bug then
[05:55] <Company> grrr - i hate that make dist updates files
[05:55] <Company> same goes for gtk-doc
[05:56] <Company> the only person that may update checked in files is ME and it's MANUALLY
[05:56] <Company> stupid tools
[05:59] <wheels> Company: Should I put the Qt check in with the KDE check or give it its own spot in the file?
[06:03] <Company> wheels: it's fine with me if you give it its own check
[06:03] <Company> wheels: in fact it's probably better
[06:03] <Company> wheels: i don't mind additinal m4 files either
[06:04] <wheels> Heh.  AOL forked GTK.  :-)
[06:05] <wheels> Company: We'll see how large it ends up being.  If it's really bit it's probably best to put it in a separate file for clarity.
[06:05] <wheels> Company: But I actually like putting configure checks in with the projects using them and just having the build system stuff them all together...
[06:09] Action: wheels notes that there's one thing that C and C++ programmers everywhere agree on:  m4 sucks worse.
[06:09] <Company> AOL forked GTK?
[06:10] <wheels> Company: Well, I actually just looked at the source and it was just minor changes in one tool, but yeah -- they're distributing it with their new AIM apparently.
[06:14] <Company> as long as they use free software, they can even invent a new widget set
[06:15] <wheels> Well, for the most part they don't.  :-)
[06:16] <Company> hm, using gthread cothreads sucks
[06:16] Surphaze ([email protected]) joined #gstreamer.
[06:16] <Company> ds-work: http://www.xmailserver.org/libpcl.html - does that code look good to you?
[06:17] <Surphaze> do I have to recreate the pipeline in order to change the location?
[06:18] <Company> Surphaze: short answer: no - longer answer: depends on the plugins you use - if you use spider, then yes
[06:19] <Surphaze> ah yes I am using spider
[06:19] <Surphaze> I was wondering why it wasn't working
[06:19] <Surphaze> will unlinking them then relinking them be enough?
[06:21] <Company> you definitely need to recreate the spider element
[06:21] ChrisHJW ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[06:21] ChrisHJW ([email protected]) joined #gstreamer.
[06:21] <Surphaze> ok thanks
[06:30] Surphaze ([email protected]) left irc: Read error: 60 (Operation timed out)
[06:31] <wheels> Company: Is there already a variable around with the X include path?
[06:31] Action: wheels didn't notice it when just doing a quick glance...
[06:32] <Company> wheels: AC_PATH_XTRA defines X_CFLAGS
[06:32] mxpxpod ([email protected]) left irc: Connection timed out
[06:32] <Company> wheels: see autoconf docs :)
[06:33] <wheels> Well, I'm actually fairly comfortable with autoconf at this point.  This however doesn't mean that I don't hate it.  ;-)
[06:36] <Company> i can live with it, too
[06:39] <wheels> Ok, next question -- is there already a check somewhere for the x86_64 "lib64" thingie?  i.e. /usr/lib64?
[06:40] <Company> no idea
[06:40] <Company> and i'm going to bed now, too
[06:41] <wheels> Alright.  I'll break gst-plugins while you're sleeping and then leave.  :-)
[06:51] mxpxpod ([email protected]) joined #gstreamer.
[07:55] ChrisHJW_log ([email protected]) joined #gstreamer.
[07:58] ChrisHJW ([email protected]) left irc: 
[08:16] grub_booter ([email protected]) left irc: Remote closed the connection
[08:34] <ds> why does bin, thread, pipeline, and queue have GST_RANK_PRIMARY?
[08:35] lilo_booter ([email protected]) joined #gstreamer.
[08:43] sub_pop ([email protected]) left irc: "Client exiting"
[08:50] alley_cat ([email protected]) got netsplit.
[08:50] alley_cat ([email protected]) returned to #gstreamer.
[08:51] kmaraas ([email protected]) left irc: "Leaving"
[09:00] sub_pop ([email protected]) joined #gstreamer.
[09:05] arana ([email protected]) left irc: "Leaving"
[09:08] foser ([email protected]) joined #gstreamer.
[09:18] Nick change: mathrick|sleep -> mathrick
[09:18] <mathrick> morning
[09:19] <ds> gettin' that way
[09:19] <thaytoo> crazy ppls - it's 17:20! ;)
[09:21] Action: mathrick is amused by another of lcamtuf's findings
[09:22] <mathrick> ds: where did you found that AOL/Gtk story ?
[09:22] <mathrick> hmm, no, it was wheels
[09:23] <foser> is the addicam guy from the ML also on irc ?
[09:24] <ds> mathrick: huh?
[09:25] <mathrick> ds: misread, I guess using colored nick in XChat is bad for me ;)
[09:26] <mathrick> ds-work and wheels share the same green color, so I end up bothering random ppl
[09:26] <ds> Company: PCL was partially the basis of gstreamer's cothreads implementation
[09:27] <ds> Company: and if we want something external, we'd use GNU Pth
[09:54] thaytoo ([email protected]) left irc: "heading off"
[09:59] apoc_ ([email protected]) joined #gstreamer.
[10:15] apocz ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:23] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[10:30] jdahlin ([email protected]) joined #gstreamer.
[10:37] thomasvs ([email protected]) joined #gstreamer.
[10:49] markey ([email protected]) joined #gstreamer.
[10:51] <thomasvs> jdahlin: http://build.fluendo.com:8080/FC1-plug/events/227/log -> conflicts
[10:53] dolphy ([email protected]) joined #gstreamer.
[10:54] <jdahlin> thomasvs: eek
[11:08] markey ([email protected]) left irc: "leaving"
[11:12] mathrick ([email protected]) left irc: ""Shutdown""
[11:15] sublett ([email protected]) joined #gstreamer.
[11:28] mathrick ([email protected]) joined #gstreamer.
[11:29] Nick change: mathrick -> mathrick|Uni
[11:41] pb_ (~pb@2002:3e03:42cc:1:20b:6aff:fe19:f290) joined #gstreamer.
[11:48] Shoragan ([email protected]) joined #gstreamer.
[11:54] LeRoutier ([email protected]) joined #gstreamer.
[11:54] <LeRoutier> hello
[11:55] <LeRoutier> got a small question for rhythmbox people : which one to use against gst 0.8 CVS / gtk 2.4 / gnome 2.6 ? 0.6.10 or 0.7.1 ?
[12:14] ChrisHJW ([email protected]) joined #gstreamer.
[12:17] <Rotund> 0.6.10 would be more stable
[12:17] <Rotund> probably a better choice
[12:18] <LeRoutier> yes for stability, but in term of features, beside iPod support that i don't need, what 0.7.1 has that 0.6.10 don't ?
[12:21] <LeRoutier> i suppose the more integrated to gnome 2.6 is rhythmbox 0.7.1, 0.6.10 beeing the last stable for gnome 2.4
[12:23] <LeRoutier> bbl, lunch
[12:23] Nick change: LeRoutier -> LeRaway
[12:28] Rotty ([email protected]) joined #gstreamer.
[12:38] kmaraas ([email protected]) joined #gstreamer.
[12:42] wheels ([email protected]) left irc: "using sirc version 2.211+KSIRC/1.3.9"
[12:51] markey ([email protected]) joined #gstreamer.
[12:54] LeRaway ([email protected]) left irc: "Leaving"
[13:22] jonathang ([email protected]) joined #gstreamer.
[14:17] thaytan ([email protected]) joined #gstreamer.
[14:18] jonathang_ ([email protected]) joined #gstreamer.
[14:20] The_Company ([email protected]) joined #gstreamer.
[14:20] jonathang ([email protected]) left irc: Nick collision from services.
[14:21] Nick change: jonathang_ -> jonathang
[14:38] Company ([email protected]) left irc: Read error: 110 (Connection timed out)
[14:41] mathrick|Uni ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[14:50] bilboed_skool ([email protected]) joined #gstreamer.
[14:50] bilboed_school ([email protected]) left irc: Connection reset by peer
[14:53] ensonic ([email protected]) joined #gstreamer.
[14:56] Nick change: The_Company -> Company
[14:57] <Company> good morning angels
[15:00] <Misirlou> hi bernie mac
[15:02] thaytan ([email protected]) got netsplit.
[15:02] ensonic ([email protected]) got netsplit.
[15:02] alley_cat ([email protected]) got netsplit.
[15:03] ensonic ([email protected]) returned to #gstreamer.
[15:03] thaytan ([email protected]) returned to #gstreamer.
[15:03] alley_cat ([email protected]) returned to #gstreamer.
[15:15] ChrisHJW_log ([email protected]) left irc: 
[15:17] ChrisHJW_log ([email protected]) joined #gstreamer.
[15:19] markey ([email protected]) left irc: "leaving"
[15:25] pb_ (~pb@2002:3e03:42cc:1:20b:6aff:fe19:f290) left irc: "bbl"
[15:29] ChriHJW ([email protected]) joined #gstreamer.
[15:31] ChrisHJW ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[15:34] kmaraas ([email protected]) left irc: Client Quit
[15:47] md` ([email protected]) joined #gstreamer.
[16:03] <Company> gthread based cothreads suck
[16:03] <Company> because there's certainly one gthread that the kernel scheduler gives lower priority
[16:04] <Company> and that makes your audio jerky and bam, bad experience
[16:08] wheels ([email protected]) joined #gstreamer.
[16:19] <Company> f*ck
[16:19] <Company> more debugging needed...
[16:19] <Company> some element is signalling too many EOS
[16:35] ChrisHJW ([email protected]) joined #gstreamer.
[16:35] ChriHJW ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[16:52] trow ([email protected]) joined #gstreamer.
[17:09] Action: wheels grunts -- oh, how I love CVS over SSH.  ;-)
[17:09] <wheels> (Esepecially when I want to do a co on another machine.  ;-) )
[17:09] kwm ([email protected]) joined #gstreamer.
[17:09] <Company> what's the problem with ssh then?
[17:11] <wheels> Company: rather key-based cvs-over-ssh...
[17:11] <Company> copy your private key or use anoncvs
[17:12] <wheels> Company : Well, but each machine can only have one private key, right?
[17:12] Action: wheels is doing a gst checkout at work...
[17:13] <wheels> (I just did an anonymous one, but that's more annoying since I'm planning on changing stuff...)
[17:13] <Company> no idea
[17:13] mxpxpod ([email protected]) left irc: "To quit or not to quit.. wait... was that a question?"
[17:14] <Company> i _think_ you can have multiple keys, but i'm not sure
[17:14] <Company> you certainly can run ssh-add more than once
[17:14] <Company> ...now that i think about it
[17:15] <Company> anyway, you could still just replace the file for the checkout ;)
[17:15] <wheels> Company : yeah, that'd be fun.  ;-)
[17:18] <wheels> This is all combined with the fun that I'm having to tunnel through my home machine too since I only have KDE CVS officially at work...
[17:18] <Company> heh
[17:18] <Company> well, you can say you're working on KDE with kiosrc
[17:19] <wheels> Company : Well, no -- it's a port forwarding issue.  There's an official port forward to the kde cvs server for me...
[17:19] <wheels> Company : (Since we're behind a proxy at work -- I end up tunneling all kinds of stuff over a hacked up SSH connection to my home machine.)
[17:21] <Company> cuuute
[17:21] <Company> where do you work?
[17:21] <wheels> Company : SAP LinuxLab
[17:21] <Company> ah
[17:32] walters ([email protected]) joined #gstreamer.
[17:32] ensonic ([email protected]) left irc: Read error: 60 (Operation timed out)
[17:52] ensonic ([email protected]) joined #gstreamer.
[17:57] kmaraas ([email protected]) joined #gstreamer.
[17:58] <ChrisHJW> test
[17:59] bilboed_skool ([email protected]) left irc: Remote closed the connection
[17:59] bilboed_school ([email protected]) joined #gstreamer.
[18:02] BBB ([email protected]) joined #gstreamer.
[18:02] Action: BBB waves
[18:05] iain ([email protected]) joined #gstreamer.
[18:07] <dolphy> hi
[18:07] <iain> hi
[18:13] <Company> man, someone tell this "comp?any" guy to comment his patches better
[18:18] <Company> grr
[18:18] <iain> Company: did you get my entry bug?
[18:18] <Company> could cvs please indent my code, too instead of just indenting the stuff i commit and leaving me with diffs
[18:18] <Company> iain: yeah, and i actually think it's a bug in what you do
[18:19] <iain> oh?
[18:19] <Company> you connect oneton twice to marlins sink
[18:19] <Company> and hope that the order in that oneton pushes and the order in that marlins sink pulls are the same
[18:20] <iain> hmmm...
[18:21] <iain> surely it doesn't matter?
[18:22] <Company> it's true that this might work (and indeed it does in your case) but the moment someone changes the order in that oneton pushes, your app dies a horrible deadlock
[18:22] thomasvs ([email protected]) left irc: "Client exiting"
[18:22] <Company> entry detects that when deciding what to schedule next and throws the error
[18:23] <iain> ah...
[18:23] <iain> so what should I do instead?
[18:24] <Company> using a different thread for the sink and 2 queues in between would definitely work
[18:24] <Company> another thing that would work if it were implemented would be to use gst_pad_select
[18:28] thomasvs ([email protected]) joined #gstreamer.
[18:33] <Company> thomasvs: any plans on when we do 0.8.1 ?
[18:34] ensonic ([email protected]) left irc: Read error: 110 (Connection timed out)
[18:34] <thomasvs> Company: for sure after bugs are properly milestoned so I have some idea of what went on :)
[18:35] <Company> thomasvs: because after 0.8.1 i was planning on doing the changes needed for rb's release-device-on-pause
[18:35] <Company> which might break some misbehaving plugins
[18:36] ChriHJW ([email protected]) joined #gstreamer.
[18:36] ChrisHJW ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[18:36] <iain> Company: oneton ! queue ! {sink} or oneton ! {queue ! sink}?
[18:36] <thomasvs> Company: well, I guess a) we need to check what has changed and if we should release, and b) you should mail what you plan on changing and what it will break so we can help getting it fixed inbetween releases
[18:37] <Company> iain: doesn't matter
[18:38] <Company> thomasvs: will do b) when i'm sure the testcase for that (the entry scheduler) works good enough
[18:39] <Company> thomasvs, walters: fwiw, i have device releasing on paused properly working locally (with mp3 and ogg at least)
[18:39] <walters> Company: rock
[18:40] lupusBE ([email protected]) joined #gstreamer.
[18:40] <iain> Company: ok
[18:44] markey ([email protected]) joined #gstreamer.
[18:50] bilboed_skool ([email protected]) joined #gstreamer.
[18:50] bilboed_school ([email protected]) left irc: Connection reset by peer
[18:53] apwbd ([email protected]) joined #gstreamer.
[18:57] teuf ([email protected]) joined #gstreamer.
[18:59] <Company> markey: seen kiosrc yet? :)
[18:59] <markey> I'm waiting for it to materialize :)
[18:59] <markey> we must have faith
[19:00] <Company> markey: it's in gst-plugins cvs
[19:00] <markey> wo did it?
[19:00] <Company> <--
[19:00] <markey> !!!
[19:00] <markey> does it work?
[19:00] <Company> it works with gst-launch
[19:01] <markey> hmm
[19:01] <markey> means?
[19:01] <Company> i have no kde app to test it with
[19:01] <markey> anyway, that's pretty cool
[19:01] <markey> good news
[19:01] <markey> did you use kioreceiver.cpp?
[19:01] <Company> yes
[19:01] <markey> I can test it later on
[19:01] jcsston ([email protected]) joined #gstreamer.
[19:02] <markey> went smoothely?
[19:02] <Company> i needed to figure out some stuff, but after that it worked pretty well
[19:02] <markey> coolness :)
[19:03] <markey> can it play shoutcast streams yet?
[19:03] <markey> kiosrc ! mad
[19:03] ChriHJW ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[19:03] <Company> lemme check
[19:03] <Company> it's still pretty rough though
[19:04] <Company> error handling and stuff like that is missing
[19:04] <markey> I'd love to help but this gst-register problem is driving me insane
[19:04] ChrisHJW ([email protected]) joined #gstreamer.
[19:04] <markey> that app makes me mad
[19:04] <Company> you'll see if this works
[19:04] <Company> i already made wheels look at it ;)
[19:10] <BBB> where's tjansen actually?
[19:11] Action: BBB notes how funny it is that all sort of random gnome and kde developers are working together on common goals while the crowd continues the ongoing fud war "kde sucks" "but not nearly as bad as your gnome!"
[19:11] <BBB> funny
[19:13] <iain> submit a story to dot.kde.org and /.
[19:13] <iain> mentioning it
[19:13] <markey> tjansen is away
[19:13] <markey> he'll prolly be back later on
[19:14] <iain> my theory is that we simply don't publicise the achievments we make enough
[19:14] <markey> too much marketing blub sucks, too
[19:14] <markey> OTOH
[19:14] <markey> :)
[19:14] <iain> so the /. shit can only see "C is dead", "The only QT app in SuSE is YaST"
[19:14] <markey> yep :)
[19:15] <iain> markey: its not marketing blurb if it's simply "We've done this"
[19:15] <markey> let's better get this working right first
[19:15] <iain> markey: marketing blurb is like that KDE "review" that savannah did on dot.kde.org a few weeks back
[19:15] <markey> I liked it ;)
[19:15] <iain> as a PR piece it was great
[19:15] <markey> savanna is cute
[19:15] <iain> as a review it sucked ass
[19:16] <BBB> so you're saying I should write a review on ongoing kde/gnome cooperation and submit it to slashdot?
[19:16] <BBB> dude
[19:16] <BBB> not a chance
[19:16] <BBB> people don't wanna read good news
[19:16] <iain> it should have been called "Reasons why KDE is great (and one thats really about linux, but I'll call it a KDE great to make it better)"
[19:17] <iain> and I've never seen savanna, so...
[19:17] <markey> nobody has seen her
[19:17] <markey> but we're all convinced she's kute
[19:17] <iain> oh right
[19:18] <iain> in my experience...its not a good assumption to make :)
[19:20] sub_pop ([email protected]) left irc: "Client exiting"
[19:20] <iain> BBB: get uraesus to do it
[19:22] <Company> iain: notice the k in kute - those guys probably think KDE is kute, too which explains a lot then
[19:24] <iain> oh, thats true
[19:27] wheels_ ([email protected]) joined #gstreamer.
[19:28] wheels ([email protected]) left irc: Nick collision from services.
[19:28] Nick change: wheels_ -> wheels
[19:31] Nick change: apoc_ -> apoc
[19:31] Uraeus ([email protected]) joined #gstreamer.
[19:31] <Uraeus> g'day
[19:32] <apoc> hi Uraeus
[19:32] <Uraeus> hello apoc
[19:38] <Company> ds-work: ping
[19:40] sub_pop ([email protected]) joined #gstreamer.
[19:45] ensonic ([email protected]) joined #gstreamer.
[19:54] <Company> wow
[19:55] <Company> it took only an hour to test if kiosrc plays http streams
[19:55] <Company> because rate probing broke my soundcard
[19:56] <Company> anyway: markey: kiosrc doesn't play digitally imported - it gets some admin page instead of the actual stream
[19:56] <Company> markey: maybe because of forwards or i don't know what
[19:56] <markey> might be 
[19:56] <markey> DI is a bit weird
[19:57] <markey> Company: playback is stable?
[19:57] pb_ (~pb@2002:5160:4550:1:240:95ff:fe30:aa84) joined #gstreamer.
[19:57] <Company> i wouldn't call it "stable" - i'd say it mostly works
[19:57] <markey> performance ok, too?
[19:57] <markey> amazing :)
[19:58] <wheels> Company : s/mostly works/mostly builds/  ;-)
[19:58] <markey> tell me about plugin build trouble..
[19:59] <Company> pfff
[19:59] Action: wheels fixed the thing for detecting the Qt dir yesterday, but there's still some type issue that I had a bit of problems with last night...
[19:59] Action: Company whispers pkg-config
[20:00] <wheels> kiosrc.cpp:84: conversion from `int' to `enum GTypeFlags'
[20:00] sjoerd ([email protected]) joined #gstreamer.
[20:00] <teuf> <slashdotmode>pkgconfig is evil gnome technology!!!</slashdotmode>
[20:00] <markey> indeed :)
[20:00] <Company> wheels: eeks, didn't i commit that yet?
[20:01] <wheels> teuf : Well, the problem is that pkgconfig really sucks unless everything is using pkgconfig.  :-)
[20:01] <markey> ack
[20:01] <teuf> wheels: yeah, so things not using pkgconfig suck ;)
[20:01] <Company> wheels: unless current kde which just sucks no matter what?
[20:02] <wheels> teuf : So I think the KDE strategy -- since it mostly sits on top of other components -- has been to wait until stuff that's lower down in the food-chain switch...
[20:02] <ensonic> teuf, pkgconfig is badly maintained :-( I had submitted a bug+patch nearly one year ago, which fixes a fclose() they have forgotten, the patch has quite many vote, but gets not commited
[20:03] <Company> wheels: KDE should have done that with the programming language, too - you would still be using the right language then
[20:03] <ensonic> wheels, everything but KDE already uses it (okay everything which still is alive)
[20:03] Action: wheels is really confused.  I'm trying to update glib and for some reason it's not building the shared libs...
[20:03] <Company> wheels: fix for that bug committed to core
[20:04] <wheels> ensonic : But it didn't when KDE 3 was just coming around.
[20:04] <wheels> ensonic : I suspect that KDE will use it in KDE 4.
[20:04] <wheels> Company : Assembler?  :-)
[20:04] <Company> wheels: no, just macro voodoo (or does that count as assembler?)
[20:05] <wheels> Company : Oh, sorry -- I didn't realize that there's much of a difference between assembler and C.  :-)
[20:05] <wheels> But I did forget about the macros.  ;-)
[20:07] jdahlin ([email protected]) left irc: "Leaving"
[20:07] <Company> hrm
[20:07] <Company> what size does KIO provide for streams?
[20:07] thomasvs ([email protected]) left irc: Read error: 113 (No route to host)
[20:08] <wheels> "size"?
[20:08] <Company> length of file
[20:10] <wheels> just a sec
[20:13] <wheels> Company : KIO::Job::totalSize(...)
[20:14] <wheels> Company : (that's a signal)
[20:14] <Company> wheels: yeah, and what does that tell me for an mp3 stream?
[20:14] <Company> wheels: -1? 0? MAXINT?
[20:15] <wheels> Company : Well, the signal is just likely never triggered.
[20:16] <Company> that sounds like a reasonable idea
[20:16] <Company> i'll debug this after lunch
[20:16] jcsston` ([email protected]) joined #gstreamer.
[20:16] <wheels> Hmm, I sleep weird hours, but at least I use standard time of the day to refer to meals.  :-)
[20:16] <wheels> 20:15 is an odd time for "lunch"  ;-)
[20:19] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[20:22] alley_cat ([email protected]) left irc: Read error: 60 (Operation timed out)
[20:22] apoc ([email protected]) left irc: "Leaving"
[20:24] lupusBE ([email protected]) left irc: Remote closed the connection
[20:26] <wheels> does gst support musepack?
[20:26] Action: wheels always answers the guys asking about it in JuK "No, JuK the backends don't support it; don't bother me..."  :-)
[20:29] <Uraeus> what is musepack?
[20:30] <wheels> Uraeus : It's some format that's used a lot by Kazzaa weenies.  ;-)
[20:31] <iain> really? what for? I've never seen it
[20:32] <Rotund> Isn't it another lossless format?
[20:32] <wheels> iain http://www.musepack.net/
[20:33] <iain> oh, I don't get lossless stuff from kazaa
[20:33] <iain> wouldn't know how to search for it
[20:33] <Rotund> wait. it's lossy
[20:36] <Rotund> It says it's open source, but I don't see a link to the source!
[20:37] alley_cat ([email protected]) joined #gstreamer.
[20:39] <wheels> Rotund : where does it say it's OSS?
[20:39] <Rotund> News: 25-02-04 MPC is OpenSource and Development goes on
[20:39] thomasvs ([email protected]) joined #gstreamer.
[20:40] dolphy ([email protected]) left irc: "Network down, IP Packets delivered via UPS"
[20:41] <Rotund> found it.
[20:41] <wheels> Rotund : ah -- http://www.personal.uni-jena.de/~pfk/MPP/index2.html#decoder-source
[20:41] <Rotund> yeah.  took long enough
[20:41] <ensonic> ciao
[20:41] ensonic ([email protected]) left #gstreamer ("Leaving").
[20:46] Action: wheels is tempted to change his responces from "not supported" to "I don't care"  ;-)
[20:47] <Rotund> I dunno.  I'd like to get a lot of decoders in gstreamer, but I won't have time to work on it for a while
[20:48] spyder482 ([email protected]) joined #gstreamer.
[20:49] sublett ([email protected]) left irc: "Why are all the really good songs short?"
[20:51] mxpxpod ([email protected]) joined #gstreamer.
[20:51] <Company> create a bug for musepack
[20:51] <Company> provide the links to the source
[20:51] <Company> make our format tracker bug depend on it
[20:51] <Company> oh: provide testfiles
[20:55] jonathang ([email protected]) left irc: "Leaving"
[21:04] <wheels> Hmm, apparently just the source for the decoder is available -- not the encoder.
[21:05] <Company> good menough
[21:05] <Company> we don't want to encode to proprietary formats anyway
[21:06] LeRoutier ([email protected]) joined #gstreamer.
[21:06] <LeRoutier> H3ell oh
[21:10] Uraeus ([email protected]) left irc: "Client exiting"
[21:17] sublett ([email protected]) joined #gstreamer.
[21:21] mathrick ([email protected]) joined #gstreamer.
[21:22] <mathrick> yo
[21:39] <ds-work> @
[21:46] <Company> ds-work: i solved the problem i pinged you about earlier
[21:46] <Company> ds-work: do you have any plans about 0.8.1?
[21:47] jcsston` ([email protected]) left irc: Read error: 110 (Connection timed out)
[21:47] BBB ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[21:47] benow ([email protected]) joined #gstreamer.
[21:47] <benow> heyas.
[21:48] Uraeus ([email protected]) joined #gstreamer.
[21:48] <benow> heya Uraeus
[21:48] <Uraeus> hi benow
[21:48] <Uraeus> benow: how goes work on the java bindings?
[21:49] <mathrick> hi Uraeus, benow
[21:51] <alley_cat> any idea when there will be a totem release working with gst 0.8?
[21:51] <Uraeus> alley_cat: I think the latest totem release have most/all gstreamer related fixes in
[21:51] <ds-work> Company: not really
[21:52] <ds-work> Company: documentation polishing, mainly
[21:52] <alley_cat> totem 0.99.9 doesn't build with 0.8, it still checks for gstreamer-0.7
[21:54] <mathrick> 0.99.9 is release, isn't it?
[21:55] <mathrick> I you need cvs totem for it to be considered "recent"
[21:55] <mathrick> s/I/I think/
[21:57] apwbd ([email protected]) left irc: Read error: 110 (Connection timed out)
[21:58] Action: wheels just started a debate with some kde folks about pkg-config in KDE 4.  :-)
[21:59] <mathrick> what does kde use atm?
[21:59] <wheels> mathrick : a mixture of stuff
[21:59] <mathrick> like for example?
[21:59] <wheels> mathrick : the issue is that it's a lot more of a PITA to maintain, but it works better than pkg-config usually does.  ;-)
[22:00] <mathrick> dunno, pkg-config mostly works here, never had any pbs
[22:00] Action: ds-work never has a problem with pkg-config
[22:00] <mathrick> and it's sooo easy to add one liner for check
[22:00] <wheels> ds-work : that's only because you're working in a GNOME environment.
[22:00] <ds-work> it's so reliable, I can't imagine using anything else
[22:00] <LeRoutier> and far better than ugly m4 macros
[22:01] <mathrick> and you need to explain it to ppl less, because pkg-config is used everywhere
[22:01] <wheels> mathrick : it's easy to write checks -- but getting the search path right can be a PITA...
[22:01] <alley_cat> getting cvs totem is a bit hard right now with cvs.gnome.org being down
[22:01] <mathrick> wheels: maybe, but once you do that, it's copy/paste setup at most
[22:01] <wheels> mathrick : No, it's really not -- GNOME has pushed it into the stuff that it depends on (and I think that's good), but it really isn't "everywhere" yet...
[22:01] <mathrick> wheels: it works consistently
[22:02] <mathrick> wheels: everywhere for values of everywhere "on my box" :)
[22:02] <ds-work> wheels: true, X and KDE don't use it :)
[22:02] <Company> X does
[22:02] <Company> at least fdo X does
[22:02] <ds-work> Company: that doesn't really count
[22:03] <mathrick> wheels: almost all of libs I interact with use pkg-config
[22:03] <mathrick> wheels: so it's dead easy to compile against another one
[22:03] <wheels> mathrick : Right.  Read my line above.  :-)
[22:03] <Company> mathrick: you're writing assembler code for your own processor ;)
[22:03] <mathrick> Company: yep, and so what? :)
[22:04] <mathrick> Company: I built it, I can asm it, right? :)
[22:04] <Company> mathrick: you'Re not using a lot of libs there ;p
[22:04] <wheels> mathrick : Many of the libs that KDE uses still don't.  Plus GNOME tends to install a copy of every friggin' external lib into the GNOME prefix, so it's not hard to find the .pc files.  ;-)
[22:04] <mathrick> Company: well, no, but that isn't the only env I develop against :)
[22:05] <mathrick> wheels: too bad for KDE :P. And .pc being easy to find is advantage, IMO :)
[22:05] <Company> gnome tends to install a copy of every lib?
[22:06] <Company> is that a suse illness? ;)
[22:06] <ds-work> jhbuild silliness
[22:07] <wheels> mathrick : Actually it's bitten me a couple of times (and is why I don't use Gst with JuK at home) because if GNOME can't find a .pc file it wants a new copy.  When I then link JuK against GStreamer it picks up the stuff in my GNOME install dir and causes some really weird lib interaction stuff...
[22:07] <wheels> Company : GARGNOME actually -- it at least installs a new copy of anything that it can't find a .pc file of.
[22:07] <mathrick> wheels: that's pretty obvious
[22:07] <mathrick> no .pc == not installed
[22:07] <Company> i don't use please-break-my-env-with-cvs-crap stuff
[22:08] <wheels> mathrick : Right, that's a big problem for migration.  ;-)
[22:08] <Company> i either break it manually so i know what i broke or i use packages
[22:08] <mathrick> wheels: why?
[22:08] <wheels> mathrick : Or if you're working on a system where that assumption is not true...
[22:08] <mathrick> wheels: that means your pkg-config is broken == not installed ;P
[22:09] <wheels> Company : Hmm?  GARGNOME doesn't build CVS.  I mean I keep a GNOME build around, but I'm not hard-core enough to run CVS.  ;-)
[22:09] <mathrick> easy, consistent, predictable
[22:09] <mathrick> wheels: yep it does, at least that's the reason of its existence ;)
[22:10] <wheels> mathrick : Well, let's say that I'm using libfreetype -- my installed version doesn't have an associated .pc file, so GARGNOME pulls down a new one.  This is annoying.  ;-)
[22:10] <Company> garnome is compiling from source for dummies
[22:10] <mathrick> wheels: but by all means correct
[22:10] <wheels> mathrick : As compared to a non-PC check which would stand a good change of finding my current one.
[22:10] <Company> so there's a reason why it's braindead
[22:10] <wheels> Company : Or in my case people that don't really give a shit.  ;-)
[22:11] <Company> kinda like gentoo for weirdoes ;)
[22:11] <mathrick> wheels: and have zillion places it can break in
[22:11] <wheels> Company : Yeah, something like that.  ;-)
[22:11] <mathrick> wheels: pkg-config is really easy - no .pc == not installed
[22:11] <wheels> mathrick : Right, that's what people hate.  ;-)
[22:11] <mathrick> wheels: it's simple enough for 5-years old to remember ;P
[22:12] <Company> wheels: why do they hate it? do they prefer "no .pc, let's start using AC_TRY_LINK and AC_CHECK_LIB"?
[22:12] <mathrick> wheels: and everyone just *loves* spaghetti m4 code ;)
[22:12] <wheels> mathrick : Because when trying to switch that means what people see is "I have this; my old system was able to find it and work with it and now with pkg-config it can't"
[22:12] <ds-work> if a lib is really installed, then why isn't the .pc file installed?
[22:12] <wheels> Company : Users, not people writing the checks.  It's definately easier on developers.
[22:13] <Company> wheels: ah - gstreamer uses the "try .pc file first, if not, use the spaghetti m4 code" at first
[22:13] <wheels> ds-work : Let's say that the lib didn't start using .pc files until v99.5 but your program works fine with v98.2.
[22:13] <Company> wheels: later on we remove the spaghetti code when we'Re sure everyone has the pkgconfig stuff
[22:13] <mathrick> "Please upgrade to x.y.z which is correctly flagged for pkg-config, or in case you have it already, check your .pc file is in correct place" <- simplest support you can get ;)
[22:14] <wheels> mathrick : Right, again -- kind of easy for developers, but sucks for users compared to before...
[22:14] <ds-work> wheels: it's a transient problem
[22:14] <Company> mathrick: unless x.y.z needs libv a.b.c and libw d.e.f which you don't have and those require...
[22:14] <mathrick> wheels: that's life, progress has its price...
[22:14] <mathrick> after all, it's kinda MORA
[22:15] <mathrick> (Migrate Once, Run Afterwards ;)
[22:15] <wheels> mathrick : Well, that essentially what I'm arguing.
[22:15] <wheels> mathrick : To be clear -- I'm in the "pro" crowd -- but I'm just trying to give the other side too...
[22:17] <mathrick> wheels: I know the arguments of other side, but it's rather pointless really. You cannot always migrate painlessly, otherwise it wouldn't be called migration
[22:18] <mathrick> it's a choice between 1) Migrate 2) Profit, or 1) Bear the pain 2) goto 1
[22:21] <LeRoutier> my CS teachers always told me that if they saw a single goto, they'll cut my hands :( well, i disgress
[22:21] <benow> Uraeus: not bad, works for hello world, ready to check in for that.  i'm having problems with bitrotten shout2send, however.
[22:21] <mathrick> yeah, and in 1st semester CP class, we weren't allowed to use exit() in our Pascal (yuck!) programs
[22:22] <mathrick> but that doesn't mean exit (or goto) is _never_ allowed
[22:23] <LeRoutier> oh my god, an horrible crasher bug in gstreamer : "gstreamer/docs/random/events:  - The appplication" . hum, well, not really, just a typo in the doc
[22:36] jdahlin ([email protected]) joined #gstreamer.
[22:44] jonathang ([email protected]) joined #gstreamer.
[22:47] jcsston ([email protected]) joined #gstreamer.
[22:49] jdahlin ([email protected]) left irc: "Leaving"
[22:51] spyder482 ([email protected]) left irc: Read error: 113 (No route to host)
[23:01] teuf ([email protected]) left irc: "Client exiting"
[23:19] jcsston ([email protected]) left irc: Read error: 110 (Connection timed out)
[23:31] Uraeus ([email protected]) left irc: "Client exiting"
[23:31] md` ([email protected]) left irc: 
[23:31] BBB ([email protected]) joined #gstreamer.
[23:31] <BBB> thomasvs: I'll handle addicam's email, ok? I'm working on that:)
[23:40] md` ([email protected]) joined #gstreamer.
[23:41] kwm ([email protected]) left irc: "Leaving"
[23:44] <wheels> Company : ah, with your core changes kiosrc builds.  :-)
[23:45] <thomasvs> wheels: so your freetype installed by your distro doesn't have a pc file ?
[23:45] <thomasvs> wheels: what distro are you using ?
[23:46] <wheels> thomasvs : an old SuSE
[23:47] <thomasvs> wheels: I don't understand.  why is the .pc file not being installed ?
[23:47] <LeRoutier> .pc file only installs if pkg-config is present on some packages
[23:47] <thomasvs> LeRoutier: it is ? why ? that's broken
[23:47] <thomasvs> LeRoutier: and I don't think I've ever seen a package do that btw.
[23:47] <Company> wheels uses suse 8.0
[23:48] <LeRoutier> thomasvs, let me check i did not say crap
[23:48] <Company> that's back in the days when bison was only at version 1.28 and pkgconfig wasn't invented yet
[23:48] <Company> it was that time when people compiled their code by hand
[23:48] <thomasvs> wheels: now that novell is standardising on qt, might as well upgrade your old suse/novell box :)
[23:48] <wheels> Company : So your memory extends all the way back to 2002?  Impressive.  ;-)
[23:49] <Company> wheels: no, it extends back to yesterday where we had that discussion ;)
[23:49] <wheels> thomasvs : Well, I actually stick with it because upgrading would mean a slower system for the stuff I do...
[23:49] <Company> wheels: 2002 was a year where gstreamer didn't even play back avi i guess ;)
[23:50] <thomasvs> Company: actually, it did :)
[23:50] <Company> thomasvs: that heavily depended on the avi and how you played it back :p
[23:50] <thomasvs> Company: true, but it still does
[23:50] Action: thomasvs had a few avi's that didn't thumbnail today
[23:51] <markey> wheels: got it working?
[23:51] <wheels> markey : Well, building...
[23:51] Action: markey holding breath
[23:52] <wheels> Actually -- Company -- do you have your test case around?
[23:53] <Company> wheels: test case?
[23:53] <wheels> Company : (And heck -- I've even been messing with GStreamer since ~2002.  ;-) )
[23:53] <wheels> Company : for kiosrc
[23:54] <Company> you mean like gst-launch kiosrc location=http://gstreamer.net ! fdsink ?
[23:54] <wheels> Company : Oh, that works?
[23:54] <Company> wheels: it should
[23:54] Action: wheels tries...
[23:55] <wheels> gst-register-0.8: kcmdlineargs.cpp:684: static int * KCmdLineArgs::qt_argc(): Assertion `argc >= (args->count()+1)' failed.
[23:55] <wheels> Abort
[23:55] <markey> heh, fun with gst-register
[23:55] <markey> I friggin hate it
[23:55] <Company> that's because i didn't know what to put there
[23:55] <Company> worked for me ;)
[23:56] <ds-work> who's gonna write mozilladec?
[23:56] <wheels> Company : Yeah, but it's at least speaking my language now.  I'll see what I can do.  :-)
[23:56] <wheels> Back in a minute -- switching clients...
[23:56] wheels ([email protected]) left irc: "[BX] Gary Coleman uses BitchX. Whatchoo talkin bout foo?"
[23:56] <Company> ds-work: should mozilladec render into buffers or to X?
[23:57] wheels ([email protected]) joined #gstreamer.
[23:57] <Company> ds-work: src ! mozilladec ! warpTV ! xvimagesink would surely be interesting ;)
[23:57] <mathrick> what is mozilladec? elem for decoding mozilla? :)
[23:57] <Company> it's a stupid idea i guess
[23:58] Action: Company remembers he wanted to write a libgame element
[23:58] <wheels> Company: I think for The Mozilla Way (tm) it would of course have to be incorporated into Mozialla -- ideally by copying and forking all of GStreamer and putting it in their repository and then porting it to XUL.  ;-)
[23:59] Action: ds-work wonders if someone will write a gtk frontend for the khtml engine
[00:00] --- Wed Mar 31 2004
[00:00] <Company> they'd need a backend first
[00:00] <Company> or that frontend depends on openoffice in the end ;)
[00:00] <LeRoutier> is khtml something like the libgtkhtml 1, 2.x, 3.X and gecko ? (but kde people were smart enought to only use 1)
[00:00] thaytan ([email protected]) left irc: Read error: 60 (Operation timed out)
[00:01] <jonathang> it's the html renderer used by konqueror and safari
[00:01] thomasvs ([email protected]) left irc: "Client exiting"
[00:01] <wheels> Company: Well, there's already a layer for porting KHTML to other toolkits.
[00:02] <wheels> Company: (Apple wrote it for Safari)
[00:02] <ds-work> isn't KHTML GPL?
[00:02] <wheels> LeRoutier: It's a full rendering engine, but the codebase is much smaller than Gecko.
[00:02] <wheels> ds-work: LGPL -- like all of kdelibs.
[00:04] <ds-work> oh yeah.  but it depends on Qt
[00:04] <wheels> ds-work: Well, sort of depends on Qt -- since it's obviously possible to port it to other things...
[00:04] <jonathang> ds-work, you wouldn't imagine Apple using Qt? :)
[00:05] <ds-work> er, actually I should have said, "as compiled on my system, it depends (at least indirectly) on Qt"
[00:06] <ds-work> jonathang: no, Apple is a business.  They will use whatever is expedient
[00:07] <jonathang> given the state of Qt on MacOS X, I don't think they'd use it for their web browser...
[00:07] <wheels> jonathang: hmm?
[00:07] <ds-work> jonathang: at most, I imagine that KHTML only uses the glib-equivalent code in Qt, which I imageine works perfectly well on OS X
[00:07] <ds-work> btw, Qt works fine on OS X, too
[00:08] Action: wheels wouldn't have expected them to use it for their web browser, but I'm not aware of any particular problems with Qt on OS X.
[00:08] Company ([email protected]) left irc: Read error: 54 (Connection reset by peer)
[00:08] Company ([email protected]) joined #gstreamer.
[00:08] <wheels> ds-work: Well, it handles the widgets in the rendering window too...
[00:08] <jonathang> no, there are no problems. it's just that it's not... well integrated (compared to cocoa or carbon)
[00:09] <wheels> jonathang: Well, but that doesn't matter much for a rendering engine.  :-)
[00:09] <mathrick> Safari uses KHTML ported to Cocoa, not Qt one
[00:09] <jonathang> ds-work, i think that khtml renders the buttons and such with the graphical toolkit (Qt in our systems, cocoa in osx)
[00:10] <wheels> mathrick: Right.  I said that.  :-)
[00:10] <mathrick> wheels: I know, but they didn't hear it obviously :>
[00:10] <jonathang> the only thing i wanted to prove is that, like wheels and mathrick said, khtml is now independent of the toolkit
[00:10] Action: Company hates his hard disk
[00:12] abo_ ([email protected]) joined #gstreamer.
[00:14] ChrisHJW ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[00:14] ChrisHJW ([email protected]) joined #gstreamer.
[00:15] thomasvs ([email protected]) joined #gstreamer.
[00:16] <jonathang> and Apple also ported the API to objective C
[00:17] <thomasvs> yay, ssh-add and ssh-agent are working
[00:17] Action: thomasvs feels a bed coming on
[00:17] <thomasvs> night guys
[00:17] <ds-work> arg
[00:17] <wheels> jonathang: Actually "Objective C++" was what they called it.
[00:17] <wheels> jonathang: (Basically C++ with a base object.)
[00:17] <jonathang> and then the "apple" khtml were ported to GNUstep :)
[00:18] <jonathang> wheels, yep, that's a way of calling C++ classes and methods from objective C
[00:18] <Company> w00t
[00:18] Last message repeated 1 time(s).
[00:18] <wheels> Company: j00 r0k d00d!!
[00:18] <Company> german TV RTL just had something about the RIAA equivalent in germany
[00:18] <jonathang> but it's rather limited because i think you can't subclass C++ classes in objective C
[00:18] <Company> and they said _bad_ stuff
[00:19] <wheels> Company: The manager will be fired tomorrow.  ;-)
[00:19] <wheels> (programming manager)
[00:19] <Company> yeah
[00:19] <wheels> Of course since this is Germany he'll of course still be around for  6 months.  :-)
[00:20] <Company> whatever
[00:20] <Company> as long as they say bad stuff it's ok
[00:21] <jonathang> night everybody
[00:21] jonathang ([email protected]) left irc: "Leaving"
[00:21] BBB ([email protected]) left irc: "Download Gaim: http://gaim.sourceforge.net/"
[00:22] thomasvs ([email protected]) left irc: Read error: 104 (Connection reset by peer)
[00:24] thomasvs ([email protected]) joined #gstreamer.
[00:28] ChrisHJW ([email protected]) left irc: Remote closed the connection
[00:29] ChrisHJW ([email protected]) joined #gstreamer.
[00:33] <LeRoutier> good night
[00:33] LeRoutier ([email protected]) left irc: "Leaving"
[00:40] spyder482 ([email protected]) joined #gstreamer.
[00:49] steve_b_ ([email protected]) joined #gstreamer.
[00:52] markey ([email protected]) left irc: "g'nite slashdotters"
[00:52] <wheels> Company: Oh, another  C++-ism -- assigning a C-string to "char *" is wrong since a C-string is const in C++.
[00:53] <Company> wheels: huh? where?
[00:53] <wheels> Company: suck_it  :-)
[00:54] <wheels> Company: it doesn't complain here since you guys don't use -pedantic though
[00:54] md` ([email protected]) left irc: Read error: 60 (Operation timed out)
[00:54] <Company> wheels: that was the "don't segfault you stupid KDE thingy, do something reaction - i only wanted it to run ;)
[00:54] <wheels> Company: Hmm, it still doesn't run here.  :-)
[00:55] <Company> it worked for me
[00:55] md` ([email protected]) joined #gstreamer.
[00:55] <wheels> Company: I guess you don't have KDE compiled with debugging?
[00:55] <wheels> Company: (i.e. the asserts are compiled out)
[00:55] <Company> i use default debian unstable packages
[00:55] <wheels> Company: Ah, yep -- that explains it.
[00:55] walters ([email protected]) left irc: "out"
[00:56] <Company> i don't want to create a KApplication anyway, but your stupid API forces me to :p
[00:57] <wheels> You could actually probably get away with a KInstance.  I may mess with that later.
[00:58] <Company> the important thing to get right is kiosrc.cpp:process_events
[00:59] <wheels> Company: Well, maybe not.
[00:59] <Company> ?
[00:59] <wheels> Company: there are other ways to initiate an event loop
[00:59] <wheels> Company: But like I said -- I'll mess with that later.  It may be that a kapp pointer is needed for some other things.
[01:00] <Company> i could file a bug "kio shouldn't require X"
[01:01] <wheels> Company: You mean link to it?
[01:01] <Company> yeah
[01:01] <Company> it shouldn't require a KApplication either
[01:01] <wheels> Won't happen any time soon...
[01:01] <wheels> Since the kio lib includes related dialog boxes and such...
[01:01] <Company> that's what I expected
[01:01] Action: ds-work points at the quote in Rupert
[01:01] <Company> right
[01:02] <iain> whats the quote?
[01:02] <ds-work>  "there's a style of programming that takes object-oriented design, and intermingles all the classes so that you can't extract one without depending on OpenOffice" --ds-work, 29-mar-04
[01:02] <iain> nice
[01:03] <iain> and there was me being told about how nice the kde libraries were
[01:03] <iain> I shouldn't listen to oGo
[01:03] <Company> they look very nice
[01:03] <Company> until you start doing something unusual
[01:04] <Company> like trying to use them for non-kde stuff
[01:04] <ds-work> like linking with GNOME :)
[01:04] <Company> if xine were written that way, kde would still not have a media player ;)
[01:04] <ds-work> my housemate is watching the LoTR trilogy today
[01:04] <wheels> Company: neither would GNOME.  ;-)
[01:06] <Company> wheels: i'm pretty sure you can run gnome nicely from a different thread - you can at least use gnome-vfs without gtk
[01:07] <iain> yeah, we did the gui/gui-less split nicely in most places
[01:07] <wheels> Company: Right but gnome-cvs == kioslaves != libkio
[01:07] <iain> even marlin is split like that :)
[01:07] <Company> where does kio use its cache?
[01:07] <wheels> Company: hmm?
[01:07] <Company> wheels: gnomevfs == libkio, gnome-vfs modules == kioslaves
[01:07] <Company> s/use/put/
[01:07] steve_b ([email protected]) left irc: Read error: 110 (Connection timed out)
[01:08] KoRnouille ([email protected]) joined #gstreamer.
[01:08] <Company> i want to force it to redownload the file
[01:08] <wheels> Company: Well, the truth is somewhere between those two.  :-)
[01:08] <wheels> Company: libkio does significantly more than gnome-vfs does.
[01:10] <Company> well, i wouldn'T count error dialogs and progress dialogs "significant", but it does more
[01:10] kmaraas ([email protected]) left irc: "Leaving"
[01:11] <Company> you could say gnome-vfs does more, too - gnome-vfs provides a synchronous API...
[01:14] <Company> the only problem i see is that you cannot cut out the parts of kde that you think are cool and use them - you more or less have to use either all of it or nothing
[01:14] sublett ([email protected]) left irc: "Why are all the really good songs short?"
[01:14] <wheels> Company: Well, that's by design, sort of.
[01:15] <mathrick> anyone played with GDA / Mergeant ?
[01:16] <wheels> Company: The goal of kdelibs is producing something for the KDE desktop.  If something happens to be useful elsewhere, great, but it's not meant to be a set of generic components that just happen to work well together.
[01:17] <Company> wheels: kde might have missed the lesson about "small is beautiful"
[01:17] <iain> Company: you've not noticed the size of kde tarballs :)
[01:18] <wheels> iain: ...as compared to the number of GNOME ones?  ;-)
[01:18] <Company> wheels: you can see this throughout gnome that they seperated packages by functionality and provide clean (more or less) interfaces
[01:18] <iain> wheels: I'd rather have a lot of small packages, so I don't need to update so much
[01:19] <wheels> Company: Well, it's just a cultural difference.  GNOME is more geared towards producing indepenant components and then tying them all together on the desktop.  KDE is designed with the single goal of being a desktop framework.
[01:19] <Company> wheels: well, it kinda ends up the same
[01:19] <iain> gnome being aimed at general reuse in everywhere possible, kde being aimed at domination
[01:20] <iain> muhahahha
[01:21] <Company> note for next /.: KDE is like XFree86, gnome is like GNU
[01:22] AlexCONRAD ([email protected]) left irc: Read error: 110 (Connection timed out)
[01:22] <wheels> Company: Well, the GNOME approach is a more traditional UNIX approach to development.
[01:22] <Company> yeah
[01:22] <Company> go unix! ;)
[01:22] <wheels> Company: But I think KDE does lend itself more to integration of components because of its structure.  There are advantages and disadvantages to both.
[01:23] <Company> yeah, gnome sucks very hard at vertical integration
[01:23] kmaraas ([email protected]) joined #gstreamer.
[01:24] <wheels> Company: And it's carried on into development styles too -- I think we talked about "generalists vs. specialists" once...
[01:24] <iain> hey, evolution 1.5 is integrated into the gnome panel clock and it kicks ass :)
[01:25] <Company> yeah, that's what i'm talking about
[01:26] Action: wheels enjoys living in a world without ACLs.  :-)
[01:27] <Company> dd if=/dev/hda7 of=/dev/hda7 skip=7468124 seek=7468124 count=5000 bs=4096 conv=noerror
[01:27] <Company> i love my hd
[01:28] <mathrick> bah, GDA sucks
[01:28] Action: mathrick is back into using emacs for everything :)
[01:28] Action: wheels high-fives mathrick :-)
[01:28] <Company> hehe
[01:28] <Company> gnome developers use vim, kde developers use emacs ;)
[01:29] <mxpxpod> emacs...
[01:29] <wheels> Well, not everything.  Like I don't check my mail with it.  :-)
[01:29] <iain> vim sucks
[01:29] <mxpxpod> bah
[01:29] <mxpxpod> vim may suck, but gvim rules :)
[01:29] <Misirlou> FIGHT FIGHT FIGHT FIGHT
[01:29] <iain> I use emacs...but I want to use gedit
[01:29] <wheels> Company: Well, most of the older developers use XEmacs -- most of the new guys use vim or kate.  (Where "new" == "been around for less than 3-ish years")
[01:29] <iain> but I can't do away with emac's sexy indentation ability
[01:29] <mathrick> iain: why you want gedit? you have emacs
[01:30] <iain> mathrick: because emacs does so much stuff that annoys me, but I don't have the ability to fix
[01:30] <Company> wheels: i started 2000 but made a 1 year pause - so i guess i'm the new kind of guys ;)
[01:30] Action: iain is old-skool
[01:31] <mathrick> iain: hmm, like what?
[01:31] <wheels> Company: Well, I started Unix development in 1996, but didn't become active in KDE until 2001 (but was doing Qt stuff for a year or two before that).
[01:31] CompIBook ([email protected]) joined #gstreamer.
[01:31] <mathrick> emacs does precisely what you say it to
[01:32] <CompIBook> no, it does what you put into the correct brace
[01:32] <iain> mathrick: well, it complains about missing font locks, doesn't seem to want to put my c-source files into c mode...
[01:32] <CompIBook> any HD vendor suggestion? :/
[01:32] <iain> button 4 and 5 are undefined
[01:32] <mathrick> iain: hmm, seems like broken install
[01:32] <Misirlou> Western Digital
[01:33] <mathrick> Maxtor
[01:33] <iain> mathrick: yes, it annoys me and I can't fix it
[01:33] <mathrick> but check it to be silent
[01:33] <mathrick> iain: distro?
[01:33] <iain> mathrick: my own
[01:33] <mathrick> iain: ahh, so that's your own problem then ;)
[01:33] <Misirlou> LFS kekeke
[01:33] <wheels> mathrick: CompIBook: Actually I've found that the cheap line of pretty much every brand sucks, and the expensive ones tend to be ok.  :-)
[01:34] <iain> mathrick: its a redhat with mostly compiled things
[01:34] <mathrick> iain: but seriously, default emacs distibution is pretty much complete
[01:34] <iain> Misirlou: no, not LFS
[01:34] <mathrick> and works out-of-box, too
[01:34] <Misirlou> yeah
[01:34] <Misirlou> I install from CVS
[01:34] <iain> mathrick: I do like emacs, but there's generally things that annoy me
[01:34] <CompIBook> wheels: then give me something that's expensive and works :)
[01:34] <iain> mathrick: and I'd just like to use gedit to remove all the crap that emacs has that I never use
[01:35] <CompIBook> wheels: i have 2 40 gig HDs that are half-dead so i have enough space to put trash on now ;)
[01:35] <mathrick> iain: you change them, or seacrh wiki, or write your own
[01:35] Company ([email protected]) left irc: Nick collision from services.
[01:35] <iain> mathrick: I don't want to have to do that
[01:35] <mathrick> iain: well, wiki is easy
[01:35] <iain> mathrick: I just want something that works
[01:35] <iain> mathrick: without me having to do things
[01:35] <mathrick> iain: and is your best friend with emacs
[01:35] <Misirlou> CompIBook: WD2500JB
[01:36] <mathrick> iain: well, then emasc really isn't for you. TASATLNCE, after all :)
[01:36] <ds-work> argh!
[01:36] <mathrick> s/L/A/
[01:36] <Misirlou> TASATLNCE?
[01:36] <wheels> CompIBook: I've been pretty happy with my WDC WD800JB which was my most recent addition.
[01:36] Action: ds-work is surrounded by young 'uns
[01:37] <mathrick> There Ain't Such A Thing As Not Customized Emacs
[01:37] Shoragan ([email protected]) left irc: "Leaving"
[01:37] <iain> mathrick: I've been using emacs since 1995
[01:37] <ds-work> er
[01:37] <mathrick> what is 'uns ?
[01:37] <iain> mathrick: and its the best thing available, but I want something else
[01:37] <CompIBook> guns without g i spose
[01:37] <Misirlou> mathrick: contraction of "young ones"
[01:37] <ds-work> in any english dialect where "ain't" is considered proper, that would be "there ain't no such thing..."
[01:38] <mathrick> iain: well, I'm surprised you don't have it change into super coding machine of your own yet
[01:38] <mathrick> Misirlou: ahh, ok
[01:38] <iain> mathrick: because I never gave two shits about emacs-lisp
[01:38] <CompIBook> ds-work: when did you start coding free software?
[01:38] <Misirlou> iain: honestly, I don't think anyone other than RMS does
[01:38] <ds-work> 1992
[01:38] <iain> mathrick: or the ability to make my text editor into a hovercraft
[01:38] <mathrick> ds-work: mebbe, I'm copying it from 'net :)
[01:38] <CompIBook> 1992?
[01:39] <CompIBook> we didn't even have net access back then
[01:39] Action: mathrick does
[01:39] <CompIBook> i got dialup in 1999
[01:39] <Misirlou> I was four years old then
[01:39] <mathrick> ie, I give a shit
[01:39] <CompIBook> and it was hellishly expensive
[01:39] <iain> mathrick: each to their own
[01:39] <mathrick> wow, I was 3 then
[01:39] <mathrick> hmm, no
[01:39] <mathrick> wrong
[01:39] <mathrick> my brother was :)
[01:39] <mathrick> I was 9
[01:39] <iain> I was 14
[01:40] <CompIBook> i was 13 and coded basic on my atari
[01:40] <ds-work> although I did some free software stuff in the 80's, too
[01:40] sub_pop ([email protected]) left irc: "Client exiting"
[01:40] <mathrick> I was learning pascal back then
[01:40] <CompIBook> and played games
[01:40] Action: CompIBook was top ten german starcraft player for a reason ;)
[01:40] <Misirlou> hah
[01:40] <ds-work> I didn't really get into stuff until about 95
[01:41] <iain> yeah, 95 was when I got to use the internet for the first time
[01:41] <Misirlou> ds-work: What kernel did you use in the '80s?
[01:41] sub_pop ([email protected]) joined #gstreamer.
[01:41] Action: ds-work strains to remember the early 90's
[01:41] Action: wheels had FTP access via compuserve in ~1993.  :-)
[01:41] <iain> on a 10 day computer course at oxford uni, where they started to teach us stuff....and then we found xpilot
[01:41] <wheels> And some gopher.  :-)
[01:42] <CompIBook> first thing i did in the real world was send in a patch to make gst-player build against gtk2
[01:42] <ds-work> Misirlou: Atari TOS, MiNT, OS/9000, OS/9, SysV, HP/UX
[01:42] <ds-work> DOS, OS/2
[01:42] <wheels> argl -- segmentation faults with gst-launch...
[01:43] <Misirlou> k, I think you are more qualified than all of us
[01:43] <ds-work> Pr1me... (don't ask)
[01:43] <CompIBook> wheels: that's most likely inside a plugin ;)
[01:43] <wheels> Hmm, in popt actually.
[01:43] <ds-work> Misirlou: no, it just means I win old geezer contests until some _real_ old geezer appears
[01:44] <CompIBook> we could start attracting developers in #freedesktop
[01:44] <ds-work> knowing the differences in commands between OS/9 and unix is completely useless to me now, even if I could remember them
[01:44] <Misirlou> ah
[01:46] Action: mathrick envies ds-work a bit
[01:46] <iain> ds-work: forgetting useless stuff is no big deal :)
[01:47] Action: CompIBook likes having someone like ds-work around - that guy has at least fallen into some traps
[01:47] <iain> although I'm amazed I can still remember a 13 digit phone number perfectly 7 years since I last needed to dial it
[01:47] <ds-work> CompIBook: that doesn't mean I won't fall into them again.  I have very little long-term memory
[01:48] <CompIBook> ds-work: that's why i only do stuff when many people agree
[01:48] <CompIBook> ds-work: but youngsters tend to do the youngster errors
[01:49] <CompIBook> i'd like to have someone in here that has hacked on the thing before xanim
[01:50] <Misirlou> you relative youngsters make geezer errors too
[01:50] Action: ds-work remembers the days when creating an MPEG movie involved a list of .Y, .U and .V files in a directory
[01:52] <Misirlou> I just can't remember what they are
[01:52] <ds-work> "back in my day, we calculated motion vectors by hand... and we LIKED it"
[01:53] <CompIBook> yeah, and there's a reason why i don't even know what a motion vector it
[01:55] ChrisHJW ([email protected]) left irc: Remote closed the connection
[01:55] pb_ (~pb@2002:5160:4550:1:240:95ff:fe30:aa84) left irc: "Client exiting"
[01:56] ChrisHJW ([email protected]) joined #gstreamer.
[01:56] <CompIBook> uuuuuh shit
[01:56] <Misirlou> what?
[01:57] <CompIBook> my hard disk
[01:57] iain ([email protected]) left irc: "leaving"
[01:57] <CompIBook> Cannot open root device
[01:58] <mathrick> uh-oh, sounds bad
[01:58] ChrisHJW ([email protected]) left irc: Remote closed the connection
[01:59] ChrisHJW ([email protected]) joined #gstreamer.
[02:00] <CompIBook> reiserfsck --rebuild-tree crashed ;)
[02:00] <CompIBook> due to bad sectors
[02:05] <ds-work> chuck the disk
[02:06] sri ([[email protected]) left irc: "RRRrrrrrrrr...pfft."
[02:07] sri ([[email protected]) joined #gstreamer.
[02:07] Action: mathrick needs some sleep
[02:07] <mathrick> nita
[02:07] <mathrick> s/a/e/
[02:08] <Misirlou> gute nacht
[02:08] Nick change: mathrick -> mathrick|sleep
[02:09] abo_ ([email protected]) left irc: Read error: 113 (No route to host)
[02:11] thomasvs ([email protected]) left irc: Read error: 110 (Connection timed out)
[02:12] <CompIBook> ds-work: i'd like to get it into working condition so i can get hold of my data first
[02:12] <CompIBook> not that it's important, but it'd be nice to keep my LFS system as it is
[02:13] Rotty ([email protected]) left irc: Remote closed the connection
[02:15] <wheels> Ok, per my pkg-config rant earlier today:
[02:15] <wheels> checking for mad >= 0.15 id3tag >= 0.15... Package mad was not found in the pkg-config search path.
[02:15] <wheels> But the libmad tarballs don't have a .pc file.
[02:16] <CompIBook> i'm pretty sure the 0.15b ones have it
[02:16] <wheels> Nope.
[02:17] <CompIBook> sure?
[02:17] <CompIBook> i have them here on my debian box
[02:17] <wheels> philemon: /home/scott/downloads/src/libmad-0.15.1b> find | grep pc
[02:17] <wheels> ./depcomp
[02:18] Action: wheels was trying to get mad working again in preparation for some kiosrc testing.  ;-)
[02:19] <wheels> But also the default check doesn't work now either.  ;-)
[02:19] <wheels> s/default/fallback/
[02:20] <CompIBook> you don't have it installed in /usr, right?
[02:20] <wheels>  /usr/local
[02:20] <CompIBook> hrm
[02:21] foser ([email protected]) left irc: "[ I want to believe ]"
[02:22] <CompIBook> it checks for id3tag and mad at once
[02:22] <wheels> CompIBook: And I have both.
[02:22] <wheels> CompIBook: I just installed the latest version of each (and read the configure check)...
[02:23] <CompIBook> if he didn't change stuff from 0.15.0 to m0.15.1 that broke it...
[02:23] <wheels> Hmm, I take that back.  It worked this time...
[02:23] <CompIBook> youn can't blame it on mad anyway
[02:23] <CompIBook> s/mad/pkgconfig/
[02:23] <CompIBook> since in that case we still use the fallback
[02:24] <wheels> CompIBook: Well, I'm not really blaming pkg-config -- just saying that it's a frustration for people when they're looking for pc files that don't exist.  :-)
[02:25] <wheels> CompIBook: (On the last one I had several shells going and apparently started my gst-plugins configure before the make install in libid3tag, and then just noticed when the configure finished and it didn't work.)
[02:25] <CompIBook> i was under the impression mad includes a pkgconfig file now
[02:25] <CompIBook> i would have looked if i still had / on that box...
[02:26] <wheels> philemon: /home/scott/downloads/src/libmad-0.15.1b> grep -ri '\.pc' *
[02:26] <wheels> CHANGES:  * Added `samplerate' and `channels' fields to synth.pcm struct to allow
[02:27] <wheels> So not even a mention of such (in case it was supposed to be written during make install or something.)
[02:27] <CompIBook> nope, you're right
[02:28] <CompIBook> mad is buggy and ships without .pc file
[02:28] <CompIBook> all distros add one though
[02:28] Action: CompIBook just checked the debian diff file ;)
[02:29] <wheels> Where "all distros" == "Debian and maybe Redhat"  ;-)
[02:30] <CompIBook> definitely fedora
[02:31] <CompIBook> anyway, you're qt check is already longer than all xiph checks combined
[02:32] <wheels> CompIBook: Right.  It's ugly and a pain to write, but it works.  :-)
[02:33] <CompIBook> same for the xiph check
[02:34] <CompIBook> without the ugly and the pain part
[02:34] <wheels> CompIBook: Well, but Xiph is packaged much more concistantly.
[02:35] <CompIBook> gtk? glib? libgnome?
[02:36] <CompIBook> wargh
[02:36] <CompIBook> where's /dev/hda7 on devfs?
[02:36] <wheels> CompIBook: Actually the funny thing is that Qt does have a .pc file, but you'd still need all of that stuff to find out where it is.  :-)
[02:39] jcsston ([email protected]) joined #gstreamer.
[02:39] <CompIBook> nope, that's the job of the user to install it correctly
[02:39] <CompIBook> just as it's the job of the user to make ldd find his libs
[02:40] <wheels> CompIBook: Right, and that's why people hate pkg-config.  ;-)
[02:40] <CompIBook> wheels: that's probably why they hate libs, too :/
[02:41] <CompIBook> wheels: qt won't be found either if i install it in /my/random/path
[02:42] <CompIBook> or /opt/desktops/kde/qt3
[02:42] <wheels> CompIBook: it will if you have QTDIR set
[02:42] <wheels> CompIBook: Which is standard in the KDE / Qt world.
[02:42] <CompIBook> wheels: so is PKG_CONFIG_PATH
[02:42] <wheels> CompIBook: Yeah, I made this same argument earlier today.  :-)
[02:43] <CompIBook> i'm a bit annoyed by KDE doing everything different by now...
[02:44] <wheels> CompIBook: Different from GNOME?  :-)
[02:44] <CompIBook> and mozilla
[02:45] <CompIBook> and basically every other thing but XFree
[02:45] <wheels> CompIBook: Not really.
[02:45] <wheels> CompIBook: But everyone feels that way when they first come into contact with another large project.
[02:45] <wheels> CompIBook: I was pretty much endlessly frustrated with building Gst for a very long time.  ;-)
[02:46] <wheels> (Now I can deal with it, but still bitch.)
[02:47] <CompIBook> gst-plugins is a very bad example though
[02:47] <wheels> I would guess that (unpatched) well under half of the libs that come with a standard distro install (let's say without GNOME or KDE) come with .pc files.
[02:49] <CompIBook> yeah
[02:50] <CompIBook> the base system doesn't use pc files to start
[02:50] <CompIBook> but it's gaining ground
[02:50] <CompIBook> especially in the media world
[02:50] <wheels> Yeah, well, and like I said -- I was pushing for KDE to use it in KDE 4.
[02:50] <CompIBook> go wheels!
[02:51] <CompIBook> :)
[02:51] <wheels> CompIBook: did you miss this earlier?
[02:51] <CompIBook> nope, i just missed to say go wheels! back then
[02:51] <wheels> :-)
[02:51] <CompIBook> is there anyone objecting to this?
[02:51] <CompIBook> apart from the usual flamers?
[02:51] <wheels> Yeah, I'd say most of the core-devels.  ;-)
[02:52] <CompIBook> why?
[02:52] <wheels> Mostly because just about all of them have been burned by it at some point or another...
[02:52] <Misirlou> core devils
[02:52] Action: Misirlou ducks
[02:53] ChrisHJW ([email protected]) left irc: Remote closed the connection
[02:53] <wheels> CompIBook: It still may happen.  We'll see, but this was the first time that I had brought it up...
[02:53] jcsston ([email protected]) left irc: Read error: 60 (Operation timed out)
[02:53] ChrisHJW ([email protected]) joined #gstreamer.
[02:54] Action: CompIBook should make the ximian guys kick the suse guys ;)
[02:54] <wheels> CompIBook: Yeah, 15 againt 500, eh?  ;-)
[02:56] <CompIBook> wheels: ximian has dave camp, so they're almost guaranteed to win


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click