T2 IRC Log: 2011-04-21

This is the log as captured by an IRC bot in the channel. The statements are those of the individual people and might not neccessarily reflect the policy and legal rules as set forth by the T2 SDE Project.

« prev | next »

--- Log opened Thu Apr 21 00:00:06 2011
07:56 -!- vanner_ [~vanner_@nat2.rnd.stcnet.ru] has joined #t2
08:14 -!- mjungwirth9 [~mjungwirt@213.174.252.112] has quit [Ping timeout: 276 seconds]
08:26 < rxr> moin
08:54 -!- mjungwirth9 [~mjungwirt@213.174.252.112] has joined #t2
09:35 -!- koan [~koan@unaffiliated/koan] has quit [Ping timeout: 260 seconds]
09:35 -!- koan [~koan@174.136.97.66] has joined #t2
09:35 -!- koan [~koan@174.136.97.66] has quit [Changing host]
09:35 -!- koan [~koan@unaffiliated/koan] has joined #t2
10:27 -!- Iguleder [~root@unaffiliated/iguleder] has joined #t2
10:27 < Iguleder> hello guys, I have a question
10:27 < Iguleder> I have a fix for the glibc inlining problem that prevents it from building with -Os, how can I commit my changes?
10:31 < rxr> post a link here, or even better mail it to the mailing list, t2@t2-project.org
10:31 < Iguleder> can i post a pastebin link with a diff?
10:32 < rxr> sure
10:32 < Iguleder> alright, will do :)
10:33 < Iguleder> i guess someone will merge it, am i right?
10:34 < rxr> if it looks plausible
10:35 < Iguleder> i'd like to join t2 development, I'm a distro developer and I work on t2 atm, i feel bad when i tweak it and don't commit my changes
10:40 < rxr> sounds great
10:50 < CIA-79> rene * r39004 /trunk/package/base/sgutils/sgutils.desc: * changed sgutils build priority (100.000 -> 110.000)
11:23 < Iguleder> finally, here it is -> http://pastebin.com/xCbET0g1
11:23 < Iguleder> tested in 3-4 T2 builds
11:26 < rxr> all with -Os, or also one with e.g. -O2 ?
11:27 < Iguleder> should work too
11:27 < Iguleder> the problem is the same
11:28 < Iguleder> -O2 works, that's what the original glibc.conf changes too
11:30 < rxr> maybe O2 breaks with the inline change?
11:30 < rxr> I test build it just in case :-)
11:31 < Iguleder> i'm running another -Os build atm
11:32 < Iguleder> btw, the credit for this patch goes to a Puppy Linux guy who calls himself technosaurus
11:33 < Iguleder> i just created a fresh patch against 2.13 and did the T2 side of this fix
13:50 < vanner_> i have error at compile kernel26: Couldn't open directory /tmp/tmp.nirOdv9cfm/lib/modules/2.6.38.2-dist: No such file or directory. directory /tmp/tmp.nirOdv9cfm/lib/modules/ is exists but 2.6.38.2-dist was not created
14:14 -!- Iguleder [~root@unaffiliated/iguleder] has quit [Quit: Leaving]
15:23 -!- mjungwirth10 [~mjungwirt@213.174.252.112] has joined #t2
15:24 -!- mjungwirth9 [~mjungwirt@213.174.252.112] has quit [Ping timeout: 276 seconds]
15:30 < CIA-79> rene * r39005 /trunk/package/base/glibc/ (Os.patch glibc.conf):
15:30 < CIA-79> technosaurus @PuppyLinux, via IRC:
15:30 < CIA-79> * fixed glibc to build with -Os, optimized for size
15:30 < CIA-79> rene * r39006 /trunk/package/network/zabbix/zabbix.desc: * updated zabbix (1.8.4 -> 1.8.5)
15:30 < CIA-79> rene * r39007 /trunk/package/network/squid/squid.desc: * updated squid (3.1.12 -> 3.1.12.1)
15:33 -!- mjungwirth10 [~mjungwirt@213.174.252.112] has quit [Ping timeout: 276 seconds]
15:37 < CIA-79> rene * r39008 /trunk/package/x86/syslinux/syslinux.desc: * updated syslinux (4.03 -> 4.04)
15:39 -!- mjungwirth10 [~mjungwirt@213.174.252.112] has joined #t2
15:51 < vanner_> rxr: can you help with error at compilation kernel?
15:54 < rxr> ouhm, what's up?
15:55 < vanner_> i have error at compile kernel26: Couldn't open directory /tmp/tmp.nirOdv9cfm/lib/modules/2.6.38.2-dist: No such file or directory. directory /tmp/tmp.nirOdv9cfm/lib/modules/ is exists but 2.6.38.2-dist was not created
15:55 < rxr> hm, which target, what config, which cpu architecture?
15:57 < vanner_> target my, based on embedded-minimal, architecture arm, config my, but i'm not sure that everything is done correctly :(
15:57 -!- ofauchon [~ofauchon_@LCaen-156-56-13-165.w193-251.abo.wanadoo.fr] has joined #t2
15:57 < ofauchon> hi
15:59 < vanner_> what set default config for my target? i copy config to target/my_target/kernel.conf
16:00 < rxr> vanner_: guess the kernel did not build and then you get additional errors while trying to generate an initrd or so nontheless?
16:00 < rxr> have you taken a closer look into the error log?
16:00 < rxr> hi ofauchon
16:00 < vanner_> kernel was built, error at module install
16:03 < vanner_> http://paste.pocoo.org/show/375839/
16:06 < rxr> Warning: you may need to install module-init-tools
16:06 < rxr> ???
16:08 < vanner_> i have 'X module-init-tools' in pkgsel
16:26 -!- mjungwirth10 [~mjungwirt@213.174.252.112] has quit [Ping timeout: 276 seconds]
16:31 < rxr> puh, was it built?
16:32 < rxr> I generally recommend to build a t2 included target for the start, and only try adapting it when it fully built without error and some t2 knowledge was gained
16:33 -!- mjungwirth10 [~mjungwirt@213.174.252.112] has joined #t2
16:42 -!- mqueiros [~mqueiros@83.132.255.195] has joined #t2
16:49 < vanner_> rxr: i tried build embedded-minimal yesterday, but i have error in uclibc. i create own target with new version uclibc
16:59 < ofauchon> vanner, I had problems with embedded-minimal too ...
17:01 -!- vanner_ [~vanner_@nat2.rnd.stcnet.ru] has quit [Quit: KVIrc 4.1.1 Equilibrium http://www.kvirc.net/]
17:01 < ofauchon> vanner, can you build your own target ?
17:04 < CIA-79> rene * r39009 /trunk/package/archiver/lzma/lzma.desc: * changed lzma build priority (102.050 -> 103.000)
17:08 -!- T-One [~T-One@194.183.141.242] has quit [Quit: Verlassend]
17:10 < rxr> vanner: well, but just copying and updating some things will not magicallly cure it
17:10 < rxr> vanner: and this way others, and I loose the abbility to double-check and fix
17:10 < rxr> e.g. I currently have a glibc arm cross build running for the QA of it, ...
17:10 < rxr> if there is something wrong with uclibc it should be fixed in trunk for all
17:13 < ofauchon> rxr what about night build to check build system sanity ?
17:13 < ofauchon> automated night builds
17:13 < rxr> even on my 8core builds take to long for completey nightly builds
17:13 < rxr> though on that older 8 core poor Intel VT performance is partially to blame for that, sigh!
17:13 < rxr> but yes, I already put a lot of effort to test build stuff
17:14 < rxr> thing is there are just so many architecture, libc and other fancy combinations
17:14 < ofauchon> you're right
17:14 < rxr> without a Google-, or Facebook-like we can not possibly test them all
17:14 < rxr> and on top all those open source bits are just so quicklz moving, yesterday stuff built, and tomorrow some update breaks whatever, often pretty unrelated bit and such
17:15 < rxr> this is also why we update the linux headers so seldomly, though I recently had to for KVM on ppc64, ...
17:15 -!- mqueiros [~mqueiros@83.132.255.195] has quit [Quit: Ex-Chat]
17:16 < rxr> best thing really is people using example targets and reporting stuff back, instead of hiding it off-side in some wiki, blog, or even worse: private tree
17:18 < rxr> but latest AMD test boxes I put together have way better performance than this 8core Xeon
17:18 < ofauchon> rxr... can you confirm me that if i want to delete eg : "/etc/fgetty/1" , I have to create a /target/xxx/rootfs/etc/fgetty/000-1 file with '#!rm -fr 1' ?
17:18 < rxr> that is 6core AMD Phenom II outperforms the 8core Intel Xeon by quite a bit, ...
17:19 < rxr> so I got some more cpu power at my fingertips, but I also need it for other things :-)
17:19 < ofauchon> rxr I would install a software RAID-5 with dedicated SATA controllers, and SSD disks .
17:19 < rxr> ouh, you could probably also just name it 1 and skip the manual cmdline 1
17:20 < rxr> note that the rootfs overlay only works for some packaging formats, e.g. live-iso
17:20 < rxr> not for classic installer packages
17:20 < ofauchon> outch ...
17:20 < ofauchon> that don't work for splashfs ?
17:21 < rxr> what kind of image output do you use?
17:22 < ofauchon> Initial RAM disk (initrd) / RAM FS (initramfs)
17:22 < rxr> btw. can't wait for the AMD Bulldozer cores, hope they will be jet more fun to deal with :-)!
17:23 < rxr> hm, maybe rootfs overlay is not taken into account for it, yet, but that would be like a one-liner to also apply the rootfs overlay for it :-)
17:23 < rxr> the rest works with the initrd output, just not the overlay_
17:23 < rxr> ??
17:24 < rxr> just looked it up
17:25 < rxr> it already is supposed to be taken into account for the initrd output :-)
17:25 < ofauchon> I'm not really skilled with t2 ... I found the rootfs override method.
17:25 < ofauchon> ah ok
17:25 < rxr> echo "Overlaying root file-system with target defined files ..."
17:25 < rxr> target/share/initrd/build.sh
17:25 < rxr> so it should work, do you see a line like this at the end of the build?
17:26 < rxr> note that the overlay is taken from your target's target/$name/rootfs
17:26 < ofauchon> I'm @ work , I can't connect home ...
17:26 < rxr> ah, ok
17:26 < rxr> so you got the infos here for later to take a look
17:26 < ofauchon> I'll check that tonite
17:27 < ofauchon> yeah.. I'll test that
17:31 < rxr> what kind of target are you working on?
17:33 < ofauchon> embedded 486 geode ... with compact flash and serial
17:33 < ofauchon> http://www.pcengines.ch/wrap.htm
17:33 < rxr> ah, that one
17:34 < rxr> but 486 is an understatement, wasn't that geode LX?
17:34 < ofauchon> I use it as a 6lowpan to ipv[46] gateway
17:34 < rxr> it"s more like a i585+
17:35 < ofauchon> You may be right. I'm not sure
17:35 < rxr> have some Geode boards here, ... GX2, LX and the OLPC IIRC
17:36 < ofauchon> I can read : AMD Geode SC1100 CPU (fast. 486 core)
17:36 < rxr> maybe that is another than the LX
17:36 < rxr> maybe that was the Elan series or so
17:37 < rxr> maybe you can paste it's /proc/cpuinfo when you get back to it here into the chat
17:37 < ofauchon> my board is an outdated one I use for test purpose. I may move to '433 (LX700) or 500 MHz (LX800) AMD Geode LX CPU'
17:37 < ofauchon> http://www.pcengines.ch/alix.htm
17:38 < rxr> ic
17:38 < rxr> still interesting cpuinfo from the older, did not yet got in touch with one of these
17:38 < rxr> interested in what it is telling
17:39 < ofauchon> next time I boot T2 on it I'll tell you
17:39 < rxr> welcome
17:39 < ofauchon> I just need to change fgetty to use serial tty so I can log in it
17:39 < ofauchon> -> rootfs override :-)
17:40 < rxr> will have friends over for easter, will not be here that often for those days
17:40 < rxr> but I'll have the scrollback and log in any case
17:41 < ofauchon> great
18:17 < CIA-79> rene * r39010 /trunk/package/graphic/libjpeg-turbo/ (. libjpeg-turbo.cache libjpeg-turbo.conf libjpeg-turbo.desc): * added libjpeg-turbo (1.1.0) - The Independent JPEG Group's JPEG software with SIMD enhancements
18:19 < rxr> http://www.t2-project.org/packages/libjpeg-turbo.html
18:50 -!- ofauchon [~ofauchon_@LCaen-156-56-13-165.w193-251.abo.wanadoo.fr] has quit [Ping timeout: 246 seconds]
18:54 -!- Grumpywolfe [~Grumpywol@adsl-99-88-46-146.dsl.stl2mo.sbcglobal.net] has joined #t2
18:56 < Grumpywolfe> rxr I have been doing some looking into the xorg-server issue and want to know why they chose to disable-extras and disable-introspection and could this in any way be why it will not build
20:14 -!- Iguleder [~root@unaffiliated/iguleder] has joined #t2
20:15 < Iguleder> rxr: thanks
20:18 < Iguleder> well, I guess it's quiet now, so in case you see this later: lzo needs --enable-shared ;)
20:18 -!- Iguleder [~root@unaffiliated/iguleder] has left #t2 ["Leaving"]
20:27 -!- ofauchon [~ofauchon_@85-171-129-34.rev.numericable.fr] has joined #t2
21:45 < ofauchon> rxr. my target was not using "copy_and_parse_from_source" but "copy_from_source", this is the reason "#!rm -fr xxxx" where not executed.
--- Log closed Fri Apr 22 00:00:06 2011