**** BEGIN LOGGING AT Tue Mar 29 02:59:57 2011 Mar 29 06:17:32 http://pastebin.com/4pQdchB0 has anyone been able to resolve this error? Mar 29 06:18:17 i'm trying to run gcc binary compiled with OE on armv5te device. Mar 29 06:31:35 03Antonio Ospite  07master * rbbbf99cde7 10openembedded.git/recipes/tzdata/tzdata_2011b.bb: Mar 29 06:31:35 tzdata: Updated to 2011d Mar 29 06:31:35 Signed-off-by: Antonio Ospite Mar 29 06:31:35 Acked-by: Mike Westerhof Mar 29 06:31:46 03Antonio Ospite  07master * r1c4b00fa6b 10openembedded.git/recipes/tzcode/tzcode-native_2011b.bb: Mar 29 06:31:46 tzcode-native: update to 2011d Mar 29 06:31:46 Signed-off-by: Antonio Ospite Mar 29 06:31:46 Acked-by: Mike Westerhof Mar 29 07:16:17 good morning Mar 29 07:16:40 varunb: looks like you have a problem with your path setting Mar 29 07:18:13 03Radoslav Kolev  07org.openembedded.dev * r61b5ea5fdf 10openembedded.git/recipes/libmodbus/ (libmodbus.inc libmodbus_2.0.3.bb libmodbus_git.bb): Mar 29 07:18:13 Added a recipe for libmodbus Mar 29 07:18:13 Signed-off-by: Radoslav Kolev Mar 29 07:18:13 Signed-off-by: Koen Kooi Mar 29 07:18:32 i tried finding cc1 but it's not present their in my rootfs Mar 29 07:22:07 03Koen Kooi  07org.openembedded.dev * ra95279947c 10openembedded.git/recipes/slim/slim_1.3.1.bb: Mar 29 07:22:08 slim: s:virtual/x11:virtual/libx11: Mar 29 07:22:08 Signed-off-by: Koen Kooi Mar 29 07:26:45 03Koen Kooi  07org.openembedded.dev * r7c2c60db9b 10openembedded.git/recipes/slim/slim_1.3.1.bb: Mar 29 07:26:45 slim: s:libjpeg:jpeg: Mar 29 07:26:45 Signed-off-by: Koen Kooi Mar 29 07:47:13 http://events.linuxfoundation.org/events/embedded-linux-conference/mixon Mar 29 08:38:58 hi florian Mar 29 08:39:29 good morning Mar 29 08:49:39 morning all Mar 29 08:49:47 hi bluelightning Mar 29 08:58:29 hi mickeyl Mar 29 09:10:03 morning Jay7 Mar 29 09:10:09 morning Mar 29 09:10:42 hm ka6sox still awake again? Mar 29 09:10:44 hi jay7 Mar 29 09:13:34 woglinde, an SA's work is NEVER done. Mar 29 09:14:47 * Jay7 confirm this :) Mar 29 09:14:58 SA is always working :) Mar 29 09:15:01 even when sleeping :) Mar 29 09:15:14 Jay7, too true Mar 29 09:15:30 tonight I am writing proceedure manual. Mar 29 09:16:09 * Jay7 have task about nfs + windows for this night.. Mar 29 09:16:52 well.. time to go Mar 29 09:16:54 * Jay7 afk Mar 29 09:32:34 03Philippe De Swert  07master * rbb63320bc4 10openembedded.git/ (2 files in 2 dirs): Mar 29 09:32:34 poppler 0.12.3: remove `DEFAULT_PREFERENCE = "-1"` Mar 29 09:32:34 Version 0.12.3 has been used in Ȧngström, minimal-{,uclibc} and SHR for over six months. Mar 29 09:32:34 Signed-off-by: Philippe De Swert Mar 29 09:32:34 Acked-by: Paul Menzel Mar 29 09:32:35 Acked-by: Martin Jansa Mar 29 10:33:36 phdeswer: O_o you everywhere! Mar 29 10:33:53 zumbi_, you too it seems :D Mar 29 10:34:11 just slacking :) Mar 29 10:34:44 Well here too. Since they cancelled the future I am just bugfixing. And as I made sure to have good code, there are no bugs... Mar 29 10:35:06 So I spent the last two days making my own OE distro for the pandaboard. Mar 29 10:35:59 phdeswer: uhm.. interesting.. I wanted to test wilink stuff on the panda Mar 29 10:36:46 phdeswer: do you have some quick notes on how to build your distro? Mar 29 10:37:37 zumbi_, I have some files. Could make you a git diff and send you the conf. Mar 29 10:37:45 However I still need to test it on the panda. Mar 29 10:38:28 But well I am happy already. My core image is 30Mb smaller than the Angstrom one. Mar 29 10:38:29 hi obi Mar 29 10:39:03 phdeswer: if you like to do that, I'll test it Mar 29 10:41:14 hi woglinde! Mar 29 10:41:56 zumbi_, ok. The build is quite long though... I am figuring out why the hell GTK+(-native) etc get built as I don't use them at all... And over 32Gb without rm_work... Mar 29 10:42:52 phdeswer: I guess it can run on the background Mar 29 10:43:29 zumbi_, ok. I'll send you the extra files and the small diff I have. Mar 29 10:45:00 phdeswer: thanks Mar 29 10:54:51 phdeswer, I'm confused why you need a new distro to make a smaller image Mar 29 10:55:28 Crofton, Just wanted to start from scratch to understand everything Mar 29 10:55:38 And I had trouble stripping out things from Angstrom. Mar 29 10:55:46 zumbi_, mail sent. Mar 29 10:56:27 it still sounds like most of your issues are not with the distro, but image construction Mar 29 10:59:16 phdeswer and sending your emails random users dont helps oe Mar 29 10:59:49 woglinde, what do you mean? Mar 29 11:00:14 Hello. does someone has a clue for this building issue? http://pastebin.com/raw.php?i=bRq7kdH4 I think I have to clean something but "bitbake -c clean libtool-cross Mar 29 11:00:14 " is not enought Mar 29 11:00:16 If you mean sending things to zumbi_ ... it is somebody I know => not random ;) Mar 29 11:01:32 GarthPS rm tmp and build all new Mar 29 11:02:16 woglinde: bouhh... sure ? what changed ? Mar 29 11:06:16 Crofton|work, Maybe. But well always good to learn new things. Got a better insight in things after this exercise. And then I do not have to deal with the per recipe workarounds for all the distros (like setting their own default_preferences etc) Mar 29 11:08:06 phdeswer, we just need to avoid distro proliferation :) Mar 29 11:08:59 Crofton|work, sure! But anyhow my goals are different so it might come in handy for me. Mar 29 11:09:42 Angstrom is general purpose. I wanted minimalistic for advanced users. So no automagic things like networkmanager, avahi, ... Mar 29 11:10:12 you should be able to crete an angstrom image that does not install them :) Mar 29 11:10:46 there are issues with build time config options, but they will be addressed as part of the oe-core effort Mar 29 11:12:13 hey! we are just testing a proof-of-concept, if something useful comes out of it, we'll surely let you know and even more, it let us get familiar with the build system. Mar 29 11:13:20 Crofton|work, sure. Just think a distro for my own use will meet my needs better. Allows me to learn more and I have to mess less with the tree. (No merging etc on each git pull) And there is no need to push it in OE anyway. Mar 29 11:16:09 And I have been catching up on the OE core stuff. Looking forward to help with that Mar 29 11:23:06 woglinde: because I have several build target in my tmp and for a target it built well so it would be a shame to throw all. Mar 29 11:57:59 phdeswer: I'm interessted of your work with a small image. Can I get some insights? Mar 29 11:59:10 silviof, what would you like to know? Mar 29 12:00:36 How to make an image smaller? Or what mine is about? Mar 29 12:01:09 o.O Mar 29 12:46:40 phdeswer: wrt you GTK+ things that you get, not sure if it drags in gtk, but insane.bbclass which is used in several distro's drags in lots of code as insane tests if .desktop files are proper Mar 29 12:52:38 eFfeM_work, hmm. I will try without that next time. Thanks for the tip. Mar 29 12:53:28 note that it disables all the sanity checks, actually I feel it would be better to somewhat parametrize the insane tests Mar 29 12:54:13 and if you really want small images make your own image recipe, and list exactly what packages you want to get in Mar 29 12:54:55 our images are < 8 MB, but then of course they are pretty dedicated Mar 29 12:55:11 eFfeM_work, Made own images also. Mar 29 12:55:39 eFfeM_work, was just learning how to do things. Mar 29 12:55:51 which is good :-) Mar 29 12:56:23 eFfeM_work, Thanks and I agree with you on the sanity checks. Mar 29 12:57:03 yw Mar 29 12:59:23 parameterizing insane just makes things more complex Mar 29 13:00:18 or maybe splitting Mar 29 13:00:55 so that you can decide what tests you want Mar 29 13:01:07 Crofton|work, Sure parameterizing would make it more complex. Splitting would be a good alternative. Mar 29 13:01:18 I am in favour of everything that makes we have to build less... Mar 29 13:01:58 most of us just use fast machines and budget time accordingly Mar 29 13:02:27 we have big issues to deal with :) Mar 29 13:03:01 Crofton|work, soon I will use 2-3 machines together to speed up building. But still I sometimes need to work on the move and my laptop is not that great for doing OE builds. Mar 29 13:03:18 laptops just plain suck for OE Mar 29 13:03:41 Anything I can find to reduce the amount of building will be reported for sure... Mar 29 13:42:01 phdeswer: NOTE: Tasks Summary: Attempted 6534 tasks of which 0 didn't need to be rerun and 0 failed. Mar 29 13:42:37 zumbi_, That was quick! Mar 29 13:42:50 Took over 8 hours on my laptop. Mar 29 13:42:58 phdeswer: 8 cores help Mar 29 13:43:19 and SSD disks too :) Mar 29 13:43:47 :D I think I really should put a quadcore in my game machine and start using that instead of my old dual Xeon builder... Mar 29 13:44:41 Anyway good to see my distro/image work seems to be good enough and reproducable :) Mar 29 13:47:07 phdeswer: well I need to find the image, load it into SD card and test Mar 29 13:50:58 gm! Mar 29 13:51:27 just tried a clean build of a fresh oe.dev pull and have a firefox build failure Mar 29 13:51:42 anyone seen this: http://build.sakoman.com/public/logs/task/41882.txt Mar 29 13:52:21 I get the same failure on both 32 and 64 bit ubuntu 10.04 Mar 29 13:52:46 clean build in both cases Mar 29 13:52:55 zumbi_, tmp/deploy/eglibc/images/ Mar 29 13:57:51 looking at config.status and comparing with an earlier, working build I see that HOST_LIBIDL_CFLAGS is being set properly in the working build and is null in the latest build Mar 29 14:02:14 phdeswer: thanks, found, but I cannot test it now, i am in the middle of finishing something else Mar 29 14:02:54 zumbi_, no worries. I have the same problem. Apart from being 180km from my Panda... Mar 29 15:58:35 Oh ya, if anyone is thinking of pulling the patch '..' in path stuff to 2011.03-maintenance they kinda need re-working in master (and oe-core) to use patchdir instead of kludging do_configure Mar 29 15:58:40 It's on my list, somewhere Mar 29 16:17:43 re Mar 29 16:30:56 sakoman: xpidl.h:49:18: error: glib.h: No such file or directory Mar 29 16:30:57 In file included from xpidl_util.c:4::xxpidl.h:49:18: error: glib.h: No such file or directory Mar 29 16:31:00 xpidl.h:52:24: error: libIDL/IDL.h: No such file or directory Mar 29 16:31:05 these are your errors Mar 29 16:33:20 sakoman: add mesa gconf to DEPENDS Mar 29 16:34:00 in firefox_3.6.8.bb Mar 29 16:34:07 that should/might solve it Mar 29 16:35:56 hi khem Mar 29 16:38:49 woglinde: hello Mar 29 16:50:42 khem: thanks, I will try that Mar 29 16:52:22 hi sakoman Mar 29 16:53:57 hey woglinde! Mar 29 17:09:25 khem: adding mesa gconf to DEPENDS did not work -- same failure Mar 29 17:13:36 hi stefan Mar 29 17:13:38 morning Mar 29 17:13:44 hi woglinde Mar 29 17:14:00 re ka6sox Mar 29 17:18:45 woglinde: I cherry-picked your latest changes. Sadly the ecj-bootstrap change broke cacao compilation (target, not host) Mar 29 17:18:54 woglinde: reverting it made it all work again. Mar 29 17:19:05 sakoman: hmmm Mar 29 17:19:24 woglinde: and I added the PARALLEL_MAKE="" in gnumal-native and the builds are *way* more stable now Mar 29 17:19:26 khem: does it build for you? Mar 29 17:19:49 top of oe.dev, clean build Mar 29 17:21:18 sakoman: no, I have not tried but I have fixed fennec lately Mar 29 17:21:29 sakoman: if thats an option you can use that for now Mar 29 17:21:37 I will see if I can reproduce this Mar 29 17:21:44 and if I am I will fix it Mar 29 17:23:26 khem: for now I have just disabled firefox in my image Mar 29 17:23:38 the desired header files exist Mar 29 17:24:20 it seems that configure for some reason now doesn't create HOST_LIBIDL_CFLAGS properly Mar 29 17:24:54 looking at a previous build on another machine, HOST_LIBIDL_CFLAGS is set properly Mar 29 17:25:03 in the latest build it is null Mar 29 17:25:24 no time today to dig into this, so I am just disbaling it for now Mar 29 17:27:14 stefan_schmidt whats broken? Mar 29 17:27:34 stefan_schmidt I got 10% aborts from cacao Mar 29 17:27:45 for gnumail Mar 29 17:28:00 hm I will test it myself Mar 29 17:28:42 stefan_schmidt, patch day soon? Mar 29 17:29:05 woglinde: http://pastebin.com/J92rxy06 Mar 29 17:29:20 woglinde: I cutted the other 900 errors messages above :) Mar 29 17:29:32 woglinde: it failed on the buildhost for every try Mar 29 17:29:55 woglinde: 10 out of 10 with it. Reverted it and its all good again Mar 29 17:30:19 stefan aehm consider to use cacao-hg on traget too Mar 29 17:30:57 ka6sox-home: I fear not. Exam on 11th and some work to do. After the exam I want to do a last patch day for the old patchwork and then kill it and only work on patches.oe.org Mar 29 17:31:35 woglinde: Thats new to me. Did you mention this before? Mar 29 17:31:50 stefan_schmidt, so you are referring to the one that khem setup as the "new" one? Mar 29 17:31:53 woglinde: I thought it was all about cacao-native hg versus jamvm Mar 29 17:32:15 ka6sox-home: the new one is patches.oe.org for me and the old patchwork.oe.org Mar 29 17:32:20 0.99.3 for target is old anyway Mar 29 17:32:41 and the best is Mar 29 17:32:46 no cacaoh needed anymore Mar 29 17:33:05 woglinde: I'm not against it. Just wondering how we get it stable with all this changes :D Mar 29 17:33:22 anyway I test jamvm git today and it gives deadlocks with ecj Mar 29 17:33:26 verynice Mar 29 17:34:07 woglinde: muahaha, I hope Tarent knows what a hell this openjdk stuff is :) Mar 29 17:34:27 * stefan_schmidt at leats knows why he wants Tarent to do the work and not me :) Mar 29 17:35:00 * woglinde wonders anyway why you need cacao at target at buglabs? Mar 29 17:35:22 woglinde: its get build when building openjdk-6 Mar 29 17:35:35 fix it Mar 29 17:35:42 no need for it Mar 29 17:36:18 woglinde: why is it in there at all then? Mar 29 17:36:34 I didnt looked at the graphs Mar 29 17:36:40 maybee java-target Mar 29 17:36:47 is defined for cacao Mar 29 17:36:52 instead of openjdk Mar 29 17:41:00 i think it's time to make devshell suck less Mar 29 17:41:32 hehe Mar 29 17:41:39 I want bitbake -i back Mar 29 17:41:55 this manual selection crap is just lame. there's no reason it can't check for an existing shell in a priority order and spwan the 'best' one we have Mar 29 17:42:42 woglinde: shouldn't be terribly hard, but not using knotty. the ncurses UI already has a text input box which is intended to be used as input for bitbake commands, but it's not hooked up / implemented :) Mar 29 17:43:10 evening Mar 29 17:43:36 * Jay7 hope to run builder tonight Mar 29 17:43:46 er, wrong button. Mar 29 17:58:50 Hrmm, I am connected up to an embeddedarm TS-7553 over the UART console. I can see the boot procedure, I can type stuff and it is echoed back... but at the login I enter "root" hit enter, and all it does is wrap to the beginning of the line. It doesn't ask for my password. It is like it did not recognize the enter key. Mar 29 18:13:42 ericben: ping Mar 29 18:36:17 hm.. Mar 29 18:36:42 from POV of client/server bitbake yocto's runtime QA-testing looking strange.. Mar 29 18:36:59 on which side bb should run qemu?.. Mar 29 18:41:08 ~seen ericben Mar 29 18:41:12 ericben is currently on #oe (15h 16m 54s), last said: 'mickey|zzZZzz: btw as I've not studied phone stacks, do you think a phone stack could help in this usage while not adding to much dependecies ?'. Mar 29 18:54:20 03Michael 'Mickey' Lauer  07org.openembedded.dev * r0def694f73 10openembedded.git/recipes/vala/ (vala_0.11.7.2.bb vala_0.11.7.3.bb): vala[-native]: 0.11.7.2 -> 0.11.7.3 Mar 29 19:19:25 http://kergoth.pastey.net/148458 Mar 29 19:19:31 this seem sane to people? Mar 29 19:19:54 you set it with OE_TERMINAL = "gnome" or OE_TERMINAL = "screen" or whatever. OE_TERMINAL is a typed variable, 'choice' type, only allows the ones we support Mar 29 19:20:01 by default, it uses 'auto' Mar 29 19:21:05 is "none" supported? I.e. I'm going to have a config where I dont want the devshell to come up.. ever Mar 29 19:21:23 but otherwise the behavior looks reasonable Mar 29 19:24:00 should be easy enough to add Mar 29 19:24:48 class None(Terminal): def __init__(self, cmd, title=None): Terminal.__init__('true') Mar 29 19:24:53 there will always be a case in the devshell, patcher, etc where that behavior can't be enabled for some reason (usually an autobuilder or something).... even though the devshell should never be invoked on an autobuilder, it makes sense to me to avoid it.. Mar 29 19:24:56 make it run /bin/true :) Mar 29 19:25:00 yeah, agreed Mar 29 19:25:05 /bin/false might be better Mar 29 19:25:09 true Mar 29 19:25:11 okay then Mar 29 19:25:36 kergoth_: implement true client-server model :) Mar 29 19:25:47 w/o terminal Mar 29 19:26:01 just pass command to server and report output back to client Mar 29 19:27:02 and wrt screen Mar 29 19:27:18 if possible please print command with tty Mar 29 19:27:30 yeah, and while we're dreaming i'll take a million dollars Mar 29 19:27:51 because people are running bb inside screen already frequently Mar 29 19:28:01 so? Mar 29 19:28:02 so screen -r is not enough to get right screen Mar 29 19:28:11 so i'll spawn it with a named session, nothing to it Mar 29 19:28:15 or may be I'm too old :) Mar 29 19:28:18 or whatever Mar 29 19:28:22 painless either way Mar 29 19:28:30 worst case you screen -l and screen -r the id Mar 29 19:28:45 I mean screen -r [[pid.]tty[.host]] Mar 29 19:28:49 what you're talking about is likely a week of work. Mar 29 19:28:54 i did this in about 15 minutes. Mar 29 19:28:59 so no, thanks, i'll stick to this for right now Mar 29 19:29:11 well, I'm just asking :) Mar 29 19:29:29 yeah, it'd be nice Mar 29 19:29:34 along with about a billion other long term tasks Mar 29 19:30:11 yeah.. Mar 29 19:36:58 Jay7: all that said, i'd say this is an improvement to the usability, no? at least it will try to find a terminal for you, rather than forcing everyone to set it manually Mar 29 19:37:15 kergoth_: sure, it's better than we have before :) Mar 29 19:37:16 kergoth thank you Mar 29 19:37:30 kergoth_: I respect your work Mar 29 19:37:38 just looking for better way :) Mar 29 19:38:08 baby steps :P Mar 29 19:38:12 we have a long way to go Mar 29 19:41:00 Jay7: just implemented what i mentioned a few minutes ago. Mar 29 19:41:11 it spawns screen -S devshell, and you do screen -r devshell Mar 29 19:41:18 fine :) Mar 29 19:41:22 so there's no problem with running under screen Mar 29 19:42:53 hmm Mar 29 19:42:59 running under screen doesn't get us a useful exit code Mar 29 19:43:05 that's a problem for things like patch resolving Mar 29 19:43:09 hrm Mar 29 19:43:19 well, potentially Mar 29 19:43:39 * kergoth_ hacks on the noop/none/invalid Terminal Mar 29 19:44:17 also spawns $SHELL now rather than 'bash' Mar 29 19:44:17 heh Mar 29 19:45:15 $SHELL may be empty.. :( Mar 29 19:45:23 e.g. under FreeBSD's /bin/sh Mar 29 19:45:35 so provide fallback there Mar 29 19:52:06 i do, devshell.bbclass sets SHELL ?= 'bash' Mar 29 19:52:18 but maybe sh is better, more portable Mar 29 19:52:24 course we require bash anyway today.. Mar 29 19:53:12 yeah.. I should run testbuilder with dash as /bin/sh one time.. Mar 29 19:53:55 somewhere i have a bbclass that runs checkbashisms against everything in ${S}.. it's not pretty Mar 29 19:54:41 kergoth_: the script or the results. :0 Mar 29 19:55:29 both :) Mar 29 19:55:36 Jay7: I am sure that it will fail quite badly Mar 29 19:55:54 it yelled about autoconf generated ./configure, which was worrisome. i don't know how much is checkbashisms not handling syntax and how much is truly bash, though Mar 29 19:56:03 might be better to enhance pysh to spot the constructs Mar 29 19:56:13 since we already use it for signature generation in poky/oe-core Mar 29 19:56:24 course, it's not exactly speedy Mar 29 19:56:36 kergoth_: Debian's checkbashims works well. Might be worth a look Mar 29 19:58:21 is that http://sourceforge.net/projects/checkbaskisms/ or something else? Mar 29 19:58:36 ah, http://qa.debian.org/daca/checkbashisms/ Mar 29 19:58:47 nifty Mar 29 19:58:53 wonder if it is the same tool though, will have to check Mar 29 20:05:48 okay, there we are. https://github.com/kergoth/oe-core/commit/78ba056 Mar 29 20:09:58 hrm. Mar 29 20:21:56 03Jeff Lance  07master * r50bf1ef7be 10openembedded.git/recipes/ti/ (3 files): Mar 29 20:21:56 matrix-gui: update SRCREV Mar 29 20:21:56 * am1808 - pull in better description for WiFi. Mar 29 20:21:56 Signed-off-by: Jeff Lance Mar 29 20:21:56 Acked-by: Chase Maupin Mar 29 20:21:56 Signed-off-by: Denys Dmytriyenko Mar 29 21:13:08 Any FSO folks awake? Mar 29 21:15:09 Tartarus, sed'ing resolved my autofoo problems Mar 29 21:15:23 good Mar 29 21:15:28 I'm going to push them, but you might want to look at them "quickly" Mar 29 21:16:07 since they only override for the target, that should minimize the risk of collatoral damage :) Mar 29 21:17:20 sure Mar 29 21:57:38 03Andrea Adami  07org.openembedded.dev * r1df09e229a 10openembedded.git/conf/machine/include/zaurus.inc: Mar 29 21:57:38 zaurus.inc: fix typo ( s/device_table-add/device_table_add/ ) Mar 29 21:57:38 Signed-off-by: Andrea Adami Mar 29 22:10:24 https://github.com/kergoth/oe-core/compare/oe-sync...devshell - okay, i think the metaclass method is a fair bit cleaner, will switch the variable typing to use the same mechanism Mar 29 22:11:06 kergoth: devshell thing looks ok Mar 29 22:11:19 will it support TERM=xterm-color too :) Mar 29 22:11:29 wow Mar 29 22:11:29 or xterm-256color Mar 29 22:11:36 bells and whistles :) Mar 29 22:11:44 khem: adding it is easy enough, just add another class with the cmd Mar 29 22:11:59 * khem usually have xterm-256color Mar 29 22:12:00 oh, TERM, not actual terminals Mar 29 22:12:09 yes TERM Mar 29 22:13:01 it's a bit easier in the oe-core world, since we can rely on using bitbake master or something similar, which includes the change so that all tasks get the exported variables into their environment Mar 29 22:13:08 but yeah, should be fine afaik Mar 29 22:16:01 oe-core has better dep checking too Mar 29 22:16:15 so if I change elfutils bb file it rebuilds gcc from beginning Mar 29 22:16:23 punishment of goodness as you take it Mar 29 22:18:19 yeah.. more rebuilds, but things are more predictable Mar 29 22:18:43 if you use the basichash signature generator, it includes the signature in the stamp, so even changing a variable will result in rebuilds all the way down :) Mar 29 22:18:49 (a var thats used, that is) Mar 29 22:19:43 khem: would you pls review otavio's patch about bitbake.conf compression handling? Mar 29 22:22:17 hi JaMa Mar 29 22:23:44 ka6sox-away: don't kill me...testing Narcissus (trying to, server slow as hell) Mar 29 22:25:42 hi Mar 29 22:36:38 ant__: will do Mar 29 22:36:41 right now no time Mar 29 23:00:03 elfutils is a PITA for uclibc Mar 29 23:14:29 well... Mar 29 23:14:37 * Jay7 fired testing build on new setup Mar 29 23:15:04 ant__, your killing *me* Mar 29 23:16:53 dual hdd setup is much better Mar 29 23:16:57 kergoth: I want to ignore require a.inc in a recipe if libc is uclibc Mar 29 23:17:04 Jay7, I'll bet Mar 29 23:17:14 system is responsible, build is faster at same time :) Mar 29 23:17:18 kergoth: overrides wont work I thing Mar 29 23:17:20 think Mar 29 23:17:39 16Mb/sec out of the narcissus server....bleh Mar 29 23:20:10 well Mar 29 23:20:17 * Jay7 is happy with current setup Mar 29 23:21:16 Jay7, good...more disks are better... Mar 29 23:21:36 ka6sox-away: building on raid0 is fast :) Mar 29 23:22:22 Jay7, once I get the controller fixed I should have somewhere around 400MB/sec of RAID0 available for building. Mar 29 23:23:02 * Jay7 now using md raid0 with SATA3 WD Black Mar 29 23:23:20 that HDD model is nice Mar 29 23:23:30 SATA3 + 64Mb cache Mar 29 23:23:41 and silent enough for home use Mar 29 23:25:11 there is some place for improvement still.. e.g. separate hdd for sources Mar 29 23:29:06 once I get the controller fixed we should have 2 SSD's to build on...ought to be really fast. Mar 29 23:29:50 ka6sox-away: yeah, SSD should be nice :) Mar 29 23:57:50 * Jay7 -> sleep Mar 29 23:58:17 testbuilder is running against master Mar 30 00:06:10 Jay7, I think Yocto plans on implementing the builder changes in Q2 before next release right? **** ENDING LOGGING AT Wed Mar 30 02:59:57 2011