IRC Logs for #crux Tuesday, 2010-04-06

*** wawowe has quit IRC00:27
*** Zaba has quit IRC01:10
pitillogood morning01:11
*** Rotwang has joined #crux01:13
*** Zaba has joined #crux01:19
*** lasso|qt has joined #crux01:58
*** tadzik has joined #crux02:42
tadzikhowdy, kde4 repo is evolving03:14
*** cjg has joined #crux03:20
*** m4rku5 has quit IRC03:27
*** Rotwang has quit IRC03:34
*** Rotwang has joined #crux03:50
*** tadzik has quit IRC04:07
*** cjg has quit IRC04:43
*** talonz has joined #crux04:57
*** jdolan has joined #crux06:03
*** ChanServ sets mode: +o jdolan06:03
*** cjg has joined #crux06:10
*** jdolan has quit IRC06:26
*** talonz has quit IRC07:13
*** jdolan has joined #crux08:00
*** ChanServ sets mode: +o jdolan08:00
*** cjg has quit IRC08:09
*** cjg has joined #crux08:09
thrice`wow, alan is doing a great job :P  looks awesome08:12
*** jdolan has quit IRC08:15
*** jdolan has joined #crux08:17
*** ChanServ sets mode: +o jdolan08:17
*** Rotwang has quit IRC08:18
*** jdolan has quit IRC08:24
*** tadzik has joined #crux08:27
thrice`mm, new mesa didn't take long08:29
*** cjg has quit IRC08:54
teK_yay, windows does have a pkgmgr..09:05
teK_installing the telnet client is as easy as C:\> pkgmgr /iu:"TelnetClient" :p09:06
luxhcase sensitive?09:06
teK_at least it seems so.09:07
cruxbot[opt.git/2.6]: scummvm: 1.0.0 -> 1.1.009:13
*** jdolan has joined #crux09:39
*** ChanServ sets mode: +o jdolan09:39
*** cjg has joined #crux10:21
*** _mavrick61 has quit IRC10:41
*** rauz has quit IRC10:41
*** joacim has quit IRC10:41
*** _mavrick61 has joined #crux10:42
*** rauz has joined #crux10:42
*** joacim has joined #crux10:42
*** BettO has quit IRC10:55
*** BettO has joined #crux10:56
*** jdolan_ has joined #crux10:57
*** jdolan has quit IRC10:59
*** lasso|qt has quit IRC11:01
*** Rotwang has joined #crux11:19
tilmanwhy did we reject the idea of using xz instead of gz for our _packages_?11:47
tilman(ie *.pkg.tar.gz)11:47
thrice`I still think it makes sense for the iso11:47
thrice`there was a recent ML post about it, iirc11:50
tilmankind of fucking annoying that the mesademos tarball is now 16 mb11:53
thrice`they are worthless anyway, I vote to kill it :>11:55
tilmannot worthless. we need glxinfo and glxgears11:55
frinnstit would take a loooooong time to build packages on my eeepc11:59
frinnsteven more than it already does, with xz11:59
tadzikit's not build, it's install. Installing is nothing compared to building anyway11:59
tadzikach, compressing12:00
frinnstwell, compressing is part of "building", sort of :)12:02
frinnstits fine on fast systems, id not mind it on my desktop12:02
thrice`a variable in pkgmk.conf would probably be harmless12:02
tilmanSubject: [ANNOUNCE] xf86-video-ati 6.13.012:04
tilmancool shit. kms release.12:04
thrice`the firmware requirement goes away in .34 ?12:05
frinnstiirc, all firmware currently shipped will get thrown out.. or so i heard12:06
frinnstmight be complete bollox though :>12:07
tilmanwhat's with the damn nick changes?12:07
frinnstits just to piss you off12:07
frinnstswedish conspiracy12:08
tilmanfrinnst: yeah, i see a pattern12:08
RotwangI had to ping frinnst to know that12:08
frinnstgood to know12:21
*** cjg has quit IRC12:21
tilmandoes anyone know whether mesa is required to install glew.h, glxew.h and wglew.h? ;)12:41
frinnstits in the mesa3d package.. or what do you mean?12:42
tilmando programs actually use these headers?12:42
thrice`gentoo has:  "rm -f "${D}"/usr/include/GL/{glew,glxew,wglew}.h \12:42
frinnstoh, no idea12:43
thrice`and they do /usr/include/GL/glut.h too12:43
frinnstwas that the reason for the 7.8.1 release? one of those headers?12:43
tilmanthe glew headers came from the MesaDemos tarball, too12:43
tilmanso dropping that tarball means also dropping glew headers12:43
cruxbot[xorg.git/2.6]: xorg-libpixman: updated to
cruxbot[xorg.git/2.6]: mesa3d: updated to
cruxbot[opt.git/2.6]: libdrm: updated to
tilmanboom boom12:46
tilmanthrice`: thanks for the heads up re. 7.8.112:46
tilmanfrinnst: have a diff for xorg-server 1.8.0? i think i need some additional configure switches... on my laptop it tried to access ${prefix}/something :D12:47
frinnstyeah, it looks for fonts in ${(prefix}, i noticed that oo12:48
tilmangreat :|12:48
frinnstb0rked makefiles?12:49
tilmanxkb/Makefile:COMPILEDDEFAULTFONTPATH = ${prefix}/share/fonts/X11/misc/,${prefix}/share/fonts/X11/TTF/,${prefix}/share/fonts/X11/OTF/,${prefix}/share/fonts/X11/Type1/,${prefix}/share/fonts/X11/100dpi/,${prefix}/share/fonts/X11/75dpi/12:49
tilmanthat variable is the one i guess12:50
thrice`it's b0rked for sysconfdir too12:53
tilmanthrice`: how did you notice?12:53
frinnstthrice`: really? or is it just the xorg.conf.d dir youre refering to?12:54
frinnsti think that was done on purpose12:54
thrice`tilman, it goes to /usr/etc :>12:54
thrice`frinnst, yes, but it can't be overridden, which is wrong12:54
tilmanah, so easily fixed with --sysconfdir=/etc12:54
frinnsttilman: no :/12:54
thrice`no :)12:54
frinnstyou need to patch the makefiles12:54
thrice` /usr/etc regardless12:54
frinnstplaces the xorg.conf.d in /usr/lib/X11/12:55
tilmanmy gut says /etc/X11/xorg.conf.d is better :D12:55
frinnstand then you can have custom xorg.conf.d in /etc/X11 if you want12:55
thrice`tilman, me too :>12:55
cruxbot[contrib.git/2.6]: syslinux: updated to 3.8612:55
thrice`you can sed $(sysconfdir)#/etc' config/Makefile.*   if you're inclined, or just "mv" :P12:56
frinnstyeah but i thought the xorg.conf.d that the ports installs would be populated with loads of configs12:56
tilmanmakes sense...12:56
tilmandriver packages can install to /usr/lib, and users may override their shit in /etc12:56
frinnstand then allow for custom sysadmin configs in /etc12:56
frinnstjust like udev does, with /lib/udev/rules.d etc12:56
thrice`i *think* for true-hotplugging (ala hal), udev needs gudev12:57
tilmanthat would mean x server now requires glib12:57
tilmanaha, it uses glib at least for some unit tests12:58
tilmanfrinnst: see the "xorg-server 1.8.0" thread on xorg@lists.freedesktop.org13:01
tilmanfourth mail, by stefan dirsch13:01
tilmanand keithp's reply13:01
tilmandebian will use $(libdir), too, apparently (julien cristau's reply)13:03
tilmanlooks like i'll be waiting for 1.8.113:04
frinnstyeah, probably for the best :)13:04
tilmantrying KMS now13:05
tilmanshould probably go shopping for food first ;)13:05
frinnstfor some retarded reason DVI-0, DVI-1 gets switched with KMS13:16
frinnstatleast on my card13:16
*** jdolan__ has joined #crux13:34
*** jdolan_ has quit IRC13:35
*** cjg has joined #crux13:42
*** jue has joined #crux13:47
*** ChanServ sets mode: +o jue13:47
thrice` seems to be acting strange here.  could be my work network, however13:58
thrice`   :D:D14:00
*** lasso has joined #crux14:01
juejust read the log and saw the short discussion about xz-compressed packages14:07
jueTBH I still don't see the point to use the very _slow_ xz to spare some disk space14:08
thrice`decompression speed is about the same as gzip, from what I can tell14:09
jueanother weak point of xz is its huge memory usage14:09
tadzikI think it's a good idea to make this user's choice14:09
juebut again, what's the point?14:10
tadzikISO I think14:10
thrice`an iso that is 30% smaller or so, I guess is the only point14:10
juewell, at least the i586 should have gz packages, because such systems tend to have low memory14:11
juesure, that's not important14:11
thrice`I personally don't see a point in it for creating packages, since I usually prune my /usr/ports/packages regularly :)14:12
jueyeah, that's my thoughts as well14:13
juebut, I agree, ISO size matters14:13
teK_it'd be _very_ cool to include *more* IDE/S-ATA/NIC drivers on the CRUX ISO14:14
jueteK_: sure, why not14:15
thrice`I did a very small test when the 2.6 RC's were out, of just gunzip on {core,opt,xorg}/*.pkg.tar.gz, and then used xz to archieve, and it seemed to save about 30% or so14:15
teK_jue: 'd be great14:15
jueteK_: as always, patches against iso.git welcome ;)14:17
thrice`I did my 2.6 install from a ubuntu CD iirc :(  I don't think I had blank CD's at the time, and the 64-bit iso did not support usb or so14:18
*** BettO has quit IRC14:19
jaegerteK_: missing any particular ones that are important?14:20
teK_I'll have to ask a friend, he used some lenovo laptop14:21
teK_the .ko files come at really low costs, so we could make some customers happy[tm]14:22
*** cjg has quit IRC14:23
*** BettO has joined #crux14:23
*** cjg has joined #crux14:23
thrice`I'm excited for 2.7.  hopefully gcc releases something new soon14:25
tadzikwill there be attributes in 2.7? :)14:25
tadzikor this INSTALL/REJECT feature for pkgadd.conf14:27
juetadzik: INSTALL works with current pkgadd14:29
frinnsti think they released a rc today14:29
*** sinistre has joined #crux14:32
frinnstjue: do you have a repo for it yet?14:33
juefrinnst: sorry, not yet14:34
*** BettO has quit IRC14:37
*** jdolan__ has quit IRC14:37
thrice`hopefully the "ignore-new-files.patch" for pkgadd too :)14:37
juethere's one open point wrt gcc, graphite loop optimizations, tilman will look into that14:38
juethrice`: I've put it into TODO2714:38
thrice`that is beyond my understanding :(14:38
thrice`does it speed up compiles, or create smaller binaries, or?14:38
teK_loop optimization14:39
thrice`sorry, I'm not a CS guy by nature :(14:40
*** jdolan has joined #crux14:41
*** ChanServ sets mode: +o jdolan14:41
teK_a primitive example would be: for(i=0;i<4;i++) do_smth(i);  == do_smth(1); do_smth(2); do_smth(3); do_smth(4);14:42
teK_but I'm no expert in this, either. Someone at my uni works on graphite, though14:42
thrice`so it optimizes the order which it does things?14:42
tadzikor rather 0..314:42
tilmanoh boy14:43
tilmankms on r420-dual-dvi = no workie14:44
tilmanand my backup kernel image doesn't help either since it's the modules that are bad ;)14:44
teK_radeon.modeset = 0?14:45
tilmanaaah, good point. forgot about that14:47
tilmanwhat key do i press to drop to the lilo prompt?14:47
thrice`mm, shift?14:48
teK_both will work14:48
teK_i.e. one of them is the right one ;)14:48
tilmanblindly logging in and running sudo /sbin/reboot ftw14:49
sinistrehey, I thought you gyus had stopped jamming all keys on the keyboard when the c64 went into the basement ;)14:50
tilmancrap, doesn't seem to work this time14:50
cruxbot[opt.git/2.6]: libpcap: update to 1.1.114:50
cruxbot[opt.git/2.6]: tcpdump: update to 4.1.114:50
thrice`livecd time :>14:51
thrice`tilman, do you need the firmware or so?14:51
tilmanthe only cd i found is broken14:51
sinistrebootfloppy! :D14:51
tilmanthrice`: i did install radeon-ucode, but that's for r600 and greater afaik. i have an r42014:52
thrice`oh :(14:52
tilmani have /lib/firmware/radeon/R420_cp.bin14:52
tilmani guess that one came from the kernel14:52
tilmandon't see any errors in /var/log/messages14:55
tilmanmaybe i didn't enable the framebuffer console14:56
frinnstcommon error14:56
thrice`kill your other vga= stuff?14:56
thrice`mm, I guess it shoudl still work with that14:57
*** BettO has joined #crux14:57
frinnstyou need fbcon14:57
cruxbot[opt.git/2.6]: fcron: update to 3.0.614:57
*** lennart has joined #crux15:02
*** lennart has joined #crux15:02
tilmani actually had fbcon. but only as a module15:06
*** cjg has quit IRC15:08
tilmannow it's broken in a different way15:14
tilmanactually the system just came up15:15
tilmanin graphics mode15:15
tilmanseems like the initialization of the card took ... LONG15:15
*** tadzik has quit IRC15:17
tilmanmaybe some timeout15:17
sinistreaah, yeah.15:17
sinistrethat's with kms?15:17
sinistreit takes like forever here, when I've tried it. Like 3 minutes or so.15:17
frinnstits instant here15:18
sinistreIt's a known issue iirc.15:18
frinnstwhat card?15:18
tilmandmesg says: failed to load firmware15:18
lennartsinistre: did you install all the needed firmware?15:19
sinistrelennart: I didn't see anything mentioned in the logs as missing.15:20
lennartsinistre: ok, my hd4200 needed some and took a long time only when they were missing15:21
tilmandmesg|grep "firmware: requesting" would show it15:21
sinistreok. well, I can't check now, since I reverted because of other issues. Just thougth it could be worth adding that someone else had seen the problem before.15:22
tilmansinistre: you did install the 'radeon-ucode' port, right? ;)15:23
sinistreactually, no, but I don't think that's related ;)15:23
frinnsthd3200 should need extra firmware15:24
tilmanaccording to syslog, there's no 30s delay btw15:25
frinnstso you are obviously mistaken :D15:26
* sinistre shrugs15:27
*** lasso_ has joined #crux15:28
*** ulughbegh has joined #crux15:29
tilmanhelp from says "work is underway to provide support for R6xx, r7xx and newer hardware (radeon from hd2xxx to hd4xxx)15:29
thrice`try .34 :>15:29
tilmanwhat about the config options related to firmware15:31
*** lasso has quit IRC15:31
tilmanthe 2nd oen is STANDALONE15:32
thrice`no clue, my video card doesn't require silly blobs of crap15:32
*** Zaba has quit IRC15:32
teK_nouveau proved to be stable for 2D usage for me15:33
*** Zaba has joined #crux15:36
*** lennart has quit IRC15:37
*** ulughbegh has quit IRC15:38
tilmanwhy would the firmware loading fail?15:39
tilmando i need to tell the kernel to look in /lib/firmware?15:40
teK_didn't have to15:41
tilmanwhat's the correct value for UEVENT_HELPER_PATH these days?15:41
thrice`your radeon stuff is static, or as a module?15:42
tilmann/m, udev readme ftw15:44
thrice`I've never changed my path, and /lib/firmware is definitely used (for my wlan card)15:46
tilmani remember i had trouble with my wlan chip's firmware on the laptop, too15:49
*** lasso_ has quit IRC15:49
tilmannext step is to blacklist radeon frame buffer and radeon modules. We need to do this to avoid firmware initialisation problem, if it is not accessable before rrot FS is mounted. This should be added to /etc/modprobe.d/blacklist.conf:"15:58
sinistrethat's interesting. Thanks.15:59
tilmanso it's back to module land15:59
*** jue has quit IRC16:01
*** jue has joined #crux16:02
*** ChanServ sets mode: +o jue16:02
tilmanbbl, hopefully16:03
sinistrehals und beinbruch. :p16:03
tilmanyup, that fixed the delay problem16:04
tilmanbut now the boot process isn't as pretty16:04
tilmansince most of the kernel messages end up on the vga console ;)16:04
sinistreisn't that what "portland" is about?16:05
sinistreor was it "plymouth"...?16:05
* sinistre gets confused by all the funny fedora stuff16:05
*** lasso_ has joined #crux16:09
*** lasso_ has quit IRC16:11
*** Rotwang has quit IRC16:13
*** jue has quit IRC16:14
*** Zaba has quit IRC16:14
*** tri1 has joined #crux16:21
*** Zaba has joined #crux16:22
tilmanthat plymouth stuff requires initrd :x16:23
*** jdolan has quit IRC16:23
thrice`I get about 5 lines before DRM changes it to sexiness16:28
*** Zaba has quit IRC16:28
frinnsti have quiet on, so i think its 0 :)16:29
*** Zaba has joined #crux16:38
thrice`yes, but you probably don't get tux displayed then? :>16:54
tri1when will the next crux version be released ?17:35
*** szympro has joined #crux17:35
frinnstnot until gcc 4.5 gets released probably17:38
frinnstjust a guess though17:38
*** tri1 has left #crux17:42
*** szympro has quit IRC17:47
*** sinistre has quit IRC18:05
teK_tilman: either ctrl OR (not the exclusive one) shift will bring LILO to halt18:36
*** nthwyatt has quit IRC19:27
*** nthwyatt has joined #crux19:35
*** jdolan has joined #crux20:46
*** ChanServ sets mode: +o jdolan20:46
*** _mavrick61 has quit IRC21:48
*** Dudde has quit IRC21:48
*** _mavrick61 has joined #crux21:49
*** Dudde has joined #crux21:50
*** jdolan has quit IRC22:00
*** nthwyatt has quit IRC22:40

Generated by 2.11.0 by Marius Gedminas - find it at!