IRC Logs for #crux Saturday, 2011-01-08

*** andarius has joined #crux02:11
andariussalutations and greetngs02:11
*** andarius has quit IRC02:55
*** mike_k has joined #crux03:11
*** aubic has joined #crux03:35
*** pitillo has quit IRC04:11
*** pitillo has joined #crux04:13
tilmansynchronizing your clock in a cronjob that's run @daily is a bad idea05:20
tilmanit might cause cron to run the @daily jobs twice in a row05:20
tilman:)05:20
mike_kchrony fits better, as it would try to adjust the clock speed. why not?05:24
tilmanhehe, that's what jue told me05:25
tilmani like the simplicity of (open)rdate05:26
mike_kanyone using vdpau for video playback offload?05:29
tilmanno. i'm using a multithreaded ffmpeg05:33
teK_tilman: simply in what way? ntpdate works like sudo ntpdate pool.ntp.org (I'm just wondering)05:35
tilmanntpdate is about as simple as rdate05:37
tilmanbut chrony isn't afaict05:38
teK_kk05:38
mike_ktilman: little googling gives no hints. is multithreaded ffmpeg tree used in current mplayer snapshots?05:38
tilmanmike_k: no :)05:38
tilmanmike_k: it's a little tricky ;>05:38
tilmanmike_k: in the mplayerhq.hu news section there was an article about ongoing work on mt-ffmpeg, with a link to a gitorious repo05:39
tilmanyou need to get that ffmpeg code05:39
tilmanthen get a recent mplayer checkout05:40
tilmanrm -r mplayer/ffmpeg05:40
mike_kI get it05:40
tilmanmv gitorious-checkout-of-ffmpeg mplayer/05:40
tilman*crossfingers that it builds*05:40
tilman;)05:40
mike_kthanks. it is an option. but dunno if it would produce less heat than singlethreaded method. I am particulary concerned about not turning case fans...05:41
mike_kvdpau kind of works, but there seems to be wrong library location in opt/nvidia port.05:42
tilmanah, i see05:42
pedjamike_k: check nvidia port in opt-x86_64, it has required symlinks, so mplayer can use vdpau.you will also need libvdpau, for which I have a port, if you want.06:04
mike_kpedja: required stuff seems to be in contrib06:05
mike_kthanks. have you informed jue?06:05
pedjareally?i rolled my own a while back, so I never checked the contrib.06:06
pedjavdpau is great, 3-5% cpu use, even on HD.06:07
mike_kreally. works for me06:07
mike_kpedja: opt-x86_64/nvidia links do not work for me. can you try just "ln -s /usr/lib/vdpau/libvdpau_nvidia.so.$version /usr/lib/libvdpau_nvidia.so"?06:15
mike_kthat single link makes it work in mplayer and xbmc. with another driver version however (current opt/nvidia)06:16
mike_khttp://www.nvnews.net/vbulletin/showthread.php?t=14452506:16
mike_ktheir support seems claiming that is correct place to put that library06:16
mike_koh, wait, I should contact the port maintainer )06:17
pedja"libvdpau_nvidia.so.1 needs to be in /usr/lib/vdpau/libvdpau_nvidia.so.1"06:18
pedjathat is what symlinks in my nvidia port do.06:19
mike_kdamn I've read it another way around. and somehow that does not work for me.06:21
pedjaadditional link to libvdpau_nvidia.so is there because earlier snapshots of mplayer required it, iirc06:21
pedjamike_k: http://paste.lisp.org/+2JDI06:24
pedjathat works for me :)06:24
mike_khttp://paste.lisp.org/display/118423 - like that?06:25
pedjawhere did you get libvdpau_trace.so.260.19.29?06:26
mike_kmaybe just a newer driver contains that06:26
mike_kthat is not essential06:27
pedjatrace* is a part of libvdpau, not nvidia.06:27
mike_kpkginfo dont lies06:29
mike_klibvdpau_trace.so.260.19.29 is part of nvidia06:30
pedjabut it should not be, that is my point.libtrace is provided by libvdpau.06:31
mike_kmaybe another trace for more low-level stuff06:32
mike_kok, it works only with /usr/lib/vdpau being in ld.so.conf.d06:39
pedjalibvdpau.so.1 => /usr/lib/libvdpau.so.1 (0x00007f277fe5b000)06:42
pedja(from ldd /usr/bin/mplayer)06:42
*** pitillo has quit IRC06:42
pedjathere is no reason to call libvdpau_nvidia directly, that is a job of libvdpau.so.06:43
*** pitillo has joined #crux06:43
mike_kthan I just don't get why it does not work without explicit ld.so.conf path in both mplayer and xbmc06:45
pedjaremove it from ld.so.conf, and run ldd on mplayer, that might give you something.06:47
pedjaif you have both nvidia and libvdpau installed, there is no reason for it not to work...06:48
mike_kldd lists the same lib as yours. but AFAIK, apps can do dlopen() or something similar with hardcoded (part of) path in a way not visible to ldd.06:52
mike_kI'll leave it as is, and we'll see on a driver upgrade in your tree if it's an issue06:53
tilmanmike_k: correct. ldd doesn't look trhough the binary for random dlopen calls or somesuch :)08:00
juepedja, mike_k: if there's something we can adjust in opt/nvidia to improve the port, please tell08:16
jueeven though I remember that vdpau will be removed from nvidia soon, so we have to create our own vdpau port08:17
mike_kjue: I am still not sure. but x86 opt/nvidia works with vdpau only with manual symlinking plus ld.so.conf path. I want someone to confirm/bust that08:18
pedjajue: I based nvidia port from x64 repo on port from opt.08:18
pedjaonly difference are libvdpau_nvidia symlinks.08:18
juelooking for the announcement for that ...08:18
mike_kjue: like nvidia official announcement?08:20
pedjaand libvdpau wrapper from freedesktop.org should be imported to opt,08:20
jueah, found it -> http://www.nvidia.com/object/linux-display-ia32-260.19.12-driver.html08:20
jueTBH, I'm not really interested in the whole nvidia shit ...08:23
pedjaheh08:24
*** Zaba has quit IRC08:28
pedjaafaict, I am the only one using nvidia port from x64, and it is in my own best interest not to fsck it up :)08:29
juepedja: what about the additional stuff wrt desktop/pixmap ?08:30
pedjawhat additional stuff are you refering to?08:30
juein the x86_64 port08:31
*** Zaba has joined #crux08:31
pedjaoh, desktop file and icon for nvidia-settings?08:32
jueyeah08:32
jueI'd prefer if opt/opt_x86_64 ports are more or less the same and don't have different features etc.08:36
pedjaI sometimes use nvidia-settings GUI to check up if conky is reporting proper temps.08:36
juedon't get me wrong, I've no problem to add that to the opt port08:37
pedjaI have sent frinnst port that I use, and since he does not use nvidia binary driver, iirc, he does not care :)08:37
pedjaj/k08:37
pedjaI never thought of that, tbh.08:38
pedjaitworksforme[tm] :)08:38
pedjanvidia-settings GUI menu integration and vdpau related stuff are just extensions to vanilla nvidia port.08:43
juemike_k: sorry, you mean that you have to add /usr/lib/vdpau to ld.so.conf?08:43
mike_kjue: yes, I had to. and symlinks like in x86-64 port.08:44
pedjamplayer should link to /usr/lib/libvdpau.so, which loads nvidia backend.08:45
teK_tell me if there's anything I can do08:45
teK_+ hi folks08:46
pedjait does here, so there is no need for adding /usr/lib/vdpau dir to ld.so.conf.d08:47
juehmm, strange08:47
pedjamike_k: run configure for mplayer with libvdpau dir removed from ld.so.conf, and check where mplayer looks for libvdpau in configure.log.08:49
mike_kpedja: good catch. I'll try08:49
thrice`omg, new mesa :o08:57
mike_kpedja: it just passes -lvdpau to cc, nothing specific. rebuild with ld.so.conf omitted changed nothing. maybe just this driver version issue.08:59
pedjano 'VDPAU...not found' message?09:00
*** mike_k_ has joined #crux09:01
pedja"Checking for VDPAU ... no",to be exact.09:02
Zaba'checking for VDPAY... no, not checking, nevermind'09:03
mike_k_pedja: "... yes"09:04
*** mike_k has quit IRC09:04
*** mike_k_ is now known as mike_k09:04
mike_kso, mplayer just wants /usr/lib/libvdpau{.la,.so}. later during playback with forced vo/codec list it fails _unless_ ld.so.conf is adjusted. that is my story (09:07
juewell, just looked at mplayer's configure.log, fails for me because of #include vdpau/vdpau_x11.h fails09:11
mike_khttp://pastie.org/pastes/1440197/text09:12
mike_kjue: I'd leave it until someone else encounter the problem. it's a minor issue for most of users.09:13
juewhere does the includes come from?09:13
mike_klibvdpau09:14
mike_k*from contrib09:14
jueok09:16
juemike_k: I'll follow your advise and do nothing at the moment :)09:21
mike_ksure09:21
*** jdolan has joined #crux09:24
*** ChanServ sets mode: +o jdolan09:24
*** lasso has joined #crux09:52
teK_//510:18
teK_sorry.10:18
teK_is http://forum.mods.de/bb/img/smilies/smiley-pillepalle.gif blinking or animated for you with firefox?10:19
jueblinking10:23
teK_jumanji aka webkit shows it as an animated non-blinking smiley10:25
teK_tbh I think is a firefox bug10:25
cruxbot[core.git/2.7]: iproute2: update to 2.6.3710:25
mike_kblinking10:31
juemike_k: IMO it should be sufficient to have *.so.1 symlinks, could you test that for me, please?10:41
mike_kjue: having http://pastie.org/pastes/1440397/text and running "rm libvdpau_nvidia.so" breaks it. si the *.so is essential for me10:45
jueok, thanks10:48
mike_knp10:50
mike_kit's my issue afterall10:50
aubicteK_: blinking animations is a cairo issue10:56
aubicworks fine for me with cairo 1.8.1010:56
aubictilman: cairo outdated10:57
tilmanchecking..10:57
tilmanffs, my list of ports to update is large :x10:58
thrice`lol10:58
cruxbot[xorg.git/2.7]: mesa3d: updated to 7.9.1.11:02
tilmantake this for now11:02
tilmani wonder why cairo only offers gzip'd tarballs11:08
tilmanlibpng 1.5.0 has been released11:25
tilman-> revdep buildfest (at least)11:25
mike_k(11:25
jue.oO11:30
tilmani'm going to hold off on it for now11:30
tilmanit's only been rleased two days ago11:30
juetilman: +111:31
juemaybe we could update xz instead ;)11:31
tilmanjue: sure, feel free to do it11:35
cruxbot[opt.git/2.7]: cairo: updated to 1.10.2.11:36
cruxbot[opt.git/2.7]: cryptsetup: updated to 1.2.0.11:36
cruxbot[opt.git/2.7]: freetype: updated to 2.4.4.11:36
cruxbot[opt.git/2.7]: git: updated to 1.7.3.5.11:36
cruxbot[opt.git/2.7]: shared-mime-info: updated to 0.90.11:36
cruxbot[opt.git/2.7]: oprofile: updated to 0.9.6.11:36
cruxbot[opt.git/2.7]: libogg: updated to 1.2.2.11:36
cruxbot[opt.git/2.7]: lftp: updated to 4.1.2.11:36
juetilman: you've seen the xz port in my repo?11:37
tilmanyou'd like me to test it i guess?11:38
juesure, if you like so, but I don't think it's really necessary11:39
juejust comment about the idea behind, it adds a symlink for liblzma from 5.0.0 to 0, which is needed for bsdtar until a rebuild of pkgutils11:41
jueadditionlay, I'd bump pkgutils version to 'force' a rebuild11:42
*** tri has joined #crux11:47
juetilman: ?11:48
tilmansounds like a plan :)12:00
jueso we go for it?12:01
pedjatilman: is mesa3d missing talloc as a dep?12:02
tilmani think you meant "...until a rebuild of _bsdtar_ " though, no?12:02
tilmanpedja: possible. hang on.12:02
juetilman: yeah, meant "rebuild of libarchive" of course12:03
juethe same for "version bump"12:04
tilmanpkgutils should get a version bump, too, thoug. so it makes use of the new xz12:07
jueyeah, right12:09
pedja"No package 'talloc' found" at configure.12:10
tilmanpedja: yah, sec12:10
cruxbot[xorg.git/2.7]: mesa3d: depends on talloc.12:11
tilmansorry12:11
pedjabusy day, eh?12:11
tilmanjue: just tested an update like that (xz, libarchive, pkgutils). works nicely12:11
tilmanpedja: it's not as bad as the hundreds of xorg updates yesterday ;D12:13
juegood, let's go for it, I'll do it12:13
tilmanthank you12:14
cruxbot[xorg.git/2.7]: mesa3d: also depends on libxml2-python.12:14
pedjais there an equivalent to 'prt-get depinst', but for updating ports?12:17
pedjaone that will install deps, and then update the port?12:18
cruxbot[core.git/2.7]: xz: update to 5.0.012:20
cruxbot[core.git/2.7]: libarchive: forced rebuild for liblzma 5.0.012:20
cruxbot[core.git/2.7]: pkgutils: forced rebuild for liblzma 5.0.012:20
*** mike_k has quit IRC12:27
tilmanpedja: i don't think so12:29
pedjait would be rather complicated to implement, and that situation is not that common, methinks.12:32
pedjaI use yapo almost as much as prt-get directly, so it is not much of a big deal.12:33
pedjabtw, after updating mesa3d, should I rerun 'gl-config use xorg ; use nvidia'?I always forget ;)12:34
tilmanprobably12:35
tilmanand you should have run "gl-config use xorg" _before_ updating it iirc12:35
pedjaargh12:37
*** jdolan has quit IRC12:37
pedjaI'll build it now, and update it after.12:38
pedjathanks12:38
trihi12:50
trii have often problems with downloads from sf.net12:50
pedjatilman: http://paste.lisp.org/+2JDP12:51
trihave you ever thought about mirroring the sf.net (or maybe all) src-packages from the port-db to your servers?12:52
juetri: that's out of question, we don't have the time/capacity/bandwidth to do that12:56
juewith "we" I mean the official project, if someone else volunteers to do that, fine12:58
juebut most of your troubles might be caused by old/wrong sourceforge URL's12:59
pedjawould you like to donate a server?13:00
pedjatri: check this out http://crux.nu/Wiki/FaqPortsSystem#ntoc813:00
pedjathat might help13:00
pedjafrom the wth-were-they-thinking department13:03
pedjahttp://bit.ly/h0fG9713:03
*** RyoS has quit IRC13:07
*** RyoS has joined #crux13:08
jueI've added a modified nvidia port to my private repo for testing, it removes the conflicts with libvdpau and installs everything into lib, symlinks for the libraries added too13:15
juethat should fix mike_k's problem13:15
tripedja: i know this faq entry but that don't fix the problem13:16
trii think most of the problems i have are old or wrong sourceforge urls13:16
trione solution is to update the port or the other solution is to find the src-package somewhere else and save it in the distfiles folder13:17
juepedja: would you mind ...?13:17
*** pedja has quit IRC13:25
*** pedja has joined #crux13:26
*** Zaba has quit IRC13:26
*** Zaba has joined #crux13:37
*** acrux has quit IRC14:00
*** acrux has joined #crux14:03
pedjajue: sorry, I was afk.what's up?14:04
*** acrux has quit IRC14:10
*** BRAINY has joined #crux14:11
BRAINYIs there a way to install CRUX without compiling a kernel?14:11
pedjause one from the iso?14:12
BRAINYand how would i do that?14:12
juepedja: http://crux.morpheus.net/irc/#13:15:2114:18
* pedja reads log before netsplit14:19
pedjaoh, you modified nvidia port?cool14:19
*** BRAINY has quit IRC14:19
pedjaI really think libvdpau_nvidia* libs should go to /usr/lib/vdpau/.14:23
pedjamike_k should test this.14:24
pedjaafk-Christmas dinner :)14:25
pedjagrr, dinner not ready yet...14:27
frinnstjue: xz: configure: WARNING: unrecognized options: --enable-dynamic14:30
cruxbot[core.git/2.7]: xz: removed invalid configure option15:01
juefrinnst: thanks15:01
juepedja: why?15:02
tilmanpedja: re. your mesa footprint mismatch; are you on x86_64?15:02
*** pitillo has quit IRC15:05
*** pitillo has joined #crux15:06
*** jdolan has joined #crux15:15
*** ChanServ sets mode: +o jdolan15:15
pedjatilman: yes15:17
pedjajue: just a moment, I am trying to dig up a post from nvidia dev wher I saw that.15:17
tilmanfrinnst: pedja reported a footprint mismatch for the new mesa port. seems to be x86_64 specific15:18
pedjasorry, tech support for family/friends :)15:18
*** acrux has joined #crux15:22
*** acrux has quit IRC15:31
*** acrux has joined #crux15:34
*** acrux|i586 has joined #crux15:39
*** acrux has quit IRC15:41
pedjajue: http://bit.ly/eDifzP and http://bit.ly/hQX19C16:02
pedja/usr/lib/vdpau is canonical/default path that libvdpau searches for libvdpau_nvidia.so.16:03
pedjadid I read that correctly?16:03
*** acrux|i586 has quit IRC16:12
*** acrux has joined #crux16:13
juepedja: I'd say yes, thanks for the links16:15
pedjanp16:15
frinnsthmm, i dont get any missmatches other than missing drivers16:23
frinnstoh, thats what i get too16:24
frinnstnevermind then :>16:24
frinnsthmm.. good way to handle it? reintroducing xorg-x86_64 would suck16:25
juepedja: just guessing, but the libvdpau.so that comes with nvidia might be the problem, overwriting that with the one from contrib/libvdpau should fix the problem16:26
pedjainstead of overwriting, why install it at all?it will be removed in future releases anyway.16:27
jueyeah, sure, the ports in my repo doesn't have it anymore, but opt/nvidia does16:28
jue*port16:28
*** acrux has quit IRC16:29
pedjaearly 2011, iirc, so removing it will make the port future proof :)16:29
pedjatilman: mesa3d configure: WARNING: unrecognized options: --with-demos16:35
pedjaand  DRI drivers: i915 i965 mach64 mga r128 r200 r300 r600 radeon savage tdfx unichrome swrast16:36
*** acrux has joined #crux16:36
pedjano i810 and sis listed.16:36
frinnst"seems to be x86_64 specific"16:37
*** tri has left #crux16:37
pedjaso, footprint mismatch should be ignored, then?16:37
frinnstyes16:37
pedjaok16:37
frinnstuntil i figure out how to handle it gracefully16:37
juepedja: and I'm just curious what is/was the reason for mike_k's problems16:37
frinnst(=never) :)16:38
*** acrux has quit IRC16:38
thrice` .footprint64 ?  ;)16:38
pedjajue: if he still has problems, he should remove nvidia and libvdpau, and check for stale symlinks/old libs.16:41
pedjait should not happen, but still16:41
*** acrux has joined #crux16:41
jueif my theory is correct he installed opt/nvidia after contrib/libvdpau16:43
pedjainstead of other way around16:44
pedjaI think you might be right.16:44
juebut anyway, more tomorrow ...16:44
juenight all16:44
pedjagood night16:44
thrice`later jue :>16:45
tilmanpedja: *sigh* thanks16:52
pedja:)16:52
cruxbot[xorg.git/2.7]: mesa3d: remove obsolete --with-demos configure switch.16:53
pedja# sis is missing because it has not been converted to use17:20
pedja# the new interface.  i810 are missing because there is no x86-6417:20
pedja# system where they could *ever* be used.17:20
pedja[from configs/linux-dri-x86-64]17:20
pedjaproblem solved :)17:20
pedja(sorry for the flood)17:22
*** acrux|i586 has joined #crux17:27
*** acrux has quit IRC17:29
jaegergrrr.... nvidia ION + HDMI is annoying me17:49
*** acrux|i586 has quit IRC17:56
*** acrux has joined #crux17:57
*** jdolan has quit IRC18:14
teK_Romster: wxpython 2.8.11.0 is available18:15
teK_Rotwang: wxGTK 2.8.11.0 is available18:33
teK_2.9.1/218:35
teK_oops18:35
*** lasso has quit IRC19:17
Romsterpedja, vdpau? i packaged libvdpau in contrib.19:38
pedjai know.19:38
Romsteri don't believe i'm missing anything important?19:42
pedjaafaict, nothing wrong with libvdpau port, if that is what you mean :)20:12
pedjanight all.20:13
*** aubic has quit IRC20:24
Romsteryeah i was just wondering.20:42
Romsterwhy can't cairo use bzip2 or xz damn 22MB gunzip.20:44
Romsteror is that gnuzip i forget now.20:44
Romsteryeah gnuzip...20:45
Romstersrc/gtk/_core_wrap.cpp:28693:49: error: 'const class wxShowEvent' has no member named 'IsShown'20:56
Romsterwonderful.20:56
Romsterstrange mesa3d failed to compile21:11
Romsterchecking for TALLOC... configure: error: Package requirements (talloc) were not met:21:11
thrice`it's a new dep21:11
Romsterdoh silly me for not checking that thanks.21:12
thrice`:>  np21:12
Romsterdo you notice that in say another 10 years there will be like a million dependencies.21:13
Romsteras everything gets cut into smaller packages.21:14
thrice`talloc is actually split from samba iirc?  quite strange indeed :-)21:14
jaegerAnyone use XBMC in crux?21:15
Romsterthat is rather strange.21:15
Romsterdon't even know what XBMC is.21:15
jaegerHTPC software, previously known as xbox media center, now it's xbmc21:16
jaegerruns on a lot of platforms now21:16
Romsterah xbox media center i do know of.21:16
jaegerI used to have it running on my zotac mag hd-nd01 but now I can't remember how I got 5.1 surround sound working21:17
Romsterah how could i forget i need newer wxgtk and i got no access to bump that so i really can't push this even if i bumped wxgtk on my system.21:19
*** Dudde has quit IRC21:34
*** mavrick61 has quit IRC21:34
*** mavrick61 has joined #crux21:35
*** Dudde has joined #crux21:35
*** jue_ has joined #crux21:47
*** ChanServ sets mode: +o jue_21:47
*** jue has quit IRC21:50
*** jue_ is now known as jue21:50

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!