**** BEGIN LOGGING AT Tue Jan 25 02:59:57 2011 Jan 25 05:41:02 can anyone tell me how to switch to a different kernel version for a specific machine..is there anything besides specifying it in conf file..like 2.6.31? Jan 25 05:41:54 anyone? Jan 25 07:33:38 has anyone knows an efficient ARM emulator so that I can run ARM based filesystem Jan 25 07:40:06 qemu Jan 25 07:45:35 coreutils is still checksum broken Jan 25 07:45:41 any developer in here? Jan 25 07:49:05 #pupnik, can you help me how to install and use qemu on ubuntu-linux Jan 25 07:52:30 03Martin Jansa  07master * r757adb6342 10openembedded.git/recipes/xorg-lib/libxp_1.0.0.bb: Jan 25 07:52:30 libxp: move from 1.0.0 to 1.0.1 Jan 25 07:52:30 Signed-off-by: Martin Jansa Jan 25 07:52:30 sedrik: can you compare your source archive with wrong checksum with right one (from some source mirror)? Jan 25 07:52:30 sedrik: without diff and confirmation that upstream just repackaged release, nobody will change checksums Jan 25 07:56:12 JaMa: sure, see http://pastebin.com/Bqri9CpZ Jan 25 07:58:18 hmm, nevermind. seems to been fixed already just me who did not remove the old source *facepalm* Jan 25 08:00:19 sanket: you can download a plug-n-play runnable arm-linux execution environment in the form of a Maemo scratchbox VMware SDK Jan 25 08:00:51 then compile "arm programs" in the SDK and run the native arm executeables in your computer, using Xephyr to provide an X-server Jan 25 08:01:10 there's probably a way to do it with openembedded as well Jan 25 08:01:38 the components would be 1) a chroot, 2) qemu-arm and 3) xephyr to display the output Jan 25 08:15:52 sedrik: I meant diff of _content_ inside that source archive, but good that was just broken download not problem in our metadata Jan 25 08:20:32 good morning Jan 25 08:32:30 kergoth: ping Jan 25 08:35:41 kergoth: once I asked you about broken -b virtual:native: in current master, after bisect I can say, last working rev is 73b40f06444cb877a5960b2aa66abf7dacbd88f0, then couple of revisions are not bisectable due to "AttributeError: 'Cache' object has no attribute 'getVar'", which was resolved in dd0ec2f7b18e2a9ab06c499b775670516bd06ac8 Jan 25 08:36:04 and after that I got "IOError: file virtual:native:/OE/dev/recipes/efl1/ecore_svn.bb not found in /OE/tmpdir-shr:/OE/dev" also in bitbake master HEAD Jan 25 09:11:43 morning Jan 25 09:18:27 morning Jan 25 09:57:34 I m getting an error while using qemu .... qemu: fatal: Trying to execute code outside RAM or ROM at 0x30008000.......can anyone help me what is this error and how to solve it? Jan 25 10:01:08 many google results for that... Jan 25 10:23:53 sanket, maybe give more surrounding infos Jan 25 10:26:35 GNUtoo|laptop, I had use my rootfs using following command ... Jan 25 10:27:45 GNUtoo|laptop, qemu-system-arm -kernel -initrd Jan 25 10:29:23 does the kernel argument expect an uImage Jan 25 10:29:24 ? Jan 25 10:29:33 doesn't it expect a zImage instead? Jan 25 10:31:04 I had zImage but nothing happens just a blank screen... Jan 25 10:33:29 GNUtoo|laptop, now I had use this command..... Jan 25 10:33:58 GNUtoo|laptop, ./qemu-img create 100M Jan 25 10:34:42 GNUtoo|laptop, but now I m getting another message..... Jan 25 10:34:49 GNUtoo|laptop, rom: requested regions overlap (rom /home/sanket/Desktop/X_rootfs/gpe/rootfs.jffs2. free=0x06410000, addr=0x00800000) Jan 25 10:34:50 rom loading failed Jan 25 11:17:32 03Koen Kooi  07org.openembedded.dev * re276a19e1c 10openembedded.git/recipes/linux/linux-omap4/omap4430-panda/defconfig: Jan 25 11:17:32 linux-omap4: tweak panda defconfig a bit Jan 25 11:17:32 Signed-off-by: Koen Kooi Jan 25 11:17:38 03Koen Kooi  07org.openembedded.dev * r37e9f0230b 10openembedded.git/recipes/htop/ (files/remove-proc-test.patch htop_0.7.bb): Jan 25 11:17:38 htop: update to 0.9 Jan 25 11:17:38 Signed-off-by: Koen Kooi Jan 25 11:24:42 03Leon Woestenberg  07master * r1f8d2e0f28 10openembedded.git/ (5 files in 4 dirs): Jan 25 11:24:42 p2020rdb: New Freescale QorQ P2 machine, based on p1020rdb (P1) Jan 25 11:24:42 Signed-off-by: Leon Woestenberg Jan 25 11:28:34 03Stefan Schmidt  07master * r45377629e0 10openembedded.git/recipes/buglabs/ (10 files in 2 dirs): buglabs-ppp-scripts: Some scripts to enable BUG for different broadband providers Jan 25 11:28:55 he stefan Jan 25 11:29:07 hi woglinde Jan 25 11:29:25 stefan sorry something crazy is going on my sid unstable with openjdk Jan 25 11:29:49 woglinde: hmm, works ok here on my sid Jan 25 11:29:50 maybee .jar creation or fastjar is somehow broken and is choking bad at updating rt.jar Jan 25 11:30:53 so I didnt yet commit 1.8.4 Jan 25 11:31:55 stefan did you made some tests with libtool2.4 again? Jan 25 11:32:00 woglinde: no problem Jan 25 11:32:15 woglinde: not yet, trying to get our patchstack smaller Jan 25 11:32:27 and b20-1.9 working Jan 25 11:32:30 yeah Jan 25 11:32:35 not much time lfet for libtool 2.4 Jan 25 11:32:54 hm as you know b20-1.9 has no thumbee support Jan 25 11:33:20 did not know that. Jan 25 11:33:51 hmm Jan 25 11:34:25 why does it has no thumbee support anymore? Jan 25 11:34:43 because some internal method calling has changed Jan 25 11:35:03 hmm, ok Jan 25 11:35:09 and the asm stuff needs to reflect this Jan 25 11:35:13 This openjdk thing is a real beast Jan 25 11:35:13 and nobody did it Jan 25 11:35:21 got it Jan 25 11:35:35 cannonical/linaro hired someone to make it Jan 25 11:35:43 but that was utopical for one month Jan 25 11:36:06 well, kgilmr asked for a newer openjdk build and I will prepare one for him. Lets see what he will like dislike about it. Jan 25 11:36:27 oh, linearo is also supporting openjdk work on ARM. Good to know Jan 25 11:37:06 *sigh* I still have in mind to make java test recipe Jan 25 11:37:14 which covers serval tests and benchmarks Jan 25 11:37:58 so you can compare btw. releases Jan 25 11:38:23 woglinde: _that_ would be great :) Jan 25 11:41:40 * stefan_schmidt -> lunch Jan 25 11:41:57 03Alex Ferguson  07org.openembedded.dev * r997b012d01 10openembedded.git/recipes/midpath/ (files/sdl-fixes.patch midpath-core_0.2+0.3rc2.bb): (log message trimmed) Jan 25 11:41:57 midpath: Make sdl-fixes patch external Jan 25 11:41:57 Deleted sdl-fixes patch and uploaded it to an external repository Jan 25 11:41:57 because it needs CRLF line endings to apply and git automatically Jan 25 11:41:57 converts CRLF to LF. This should provide a working solution. Jan 25 11:41:57 Bumped PR. Jan 25 11:41:58 Signed-off-by: Alex Ferguson Jan 25 11:59:45 qemu: fatal: Trying to execute code outside RAM or ROM at 0x08000000 Jan 25 11:59:56 I m getting this error Jan 25 12:00:03 while using qemu Jan 25 13:25:04 Co mments on ASSUME_PROVIDED for qemu-native? (host is Fedora 14) Good idea? Problematic? Or doesn't save enough time to be worth the trouble? Jan 25 13:46:26 re Jan 25 13:51:28 JaMa: thanks for the details, can you start a thread on bitbake-dev about it, or open a bug? Jan 25 13:51:51 JaMa: oh, you did, nevermind :) Jan 25 14:01:46 * kergoth goes to find caffeine Jan 25 14:02:24 till later kergoth Jan 25 15:26:47 Anybody who has a smooth way of finding out wha a new package depends on? Jan 25 15:27:43 When writing a entire new receipt for a new src package. Any smooth wayy of finding out the dependancies to put into the recept? Jan 25 15:28:06 Other than digging into the configure file? Jan 25 15:31:05 No, that's it really Jan 25 15:31:11 run configure, watch what it checks for Jan 25 15:32:25 Thought so. Jan 25 15:55:01 Or read the configure.ac file. The fact I can read m4 should probably worry me :/ Jan 25 15:58:32 I can *read* m4. Jan 25 15:58:38 Comprehension is an issue but I can read it. Jan 25 16:02:41 quick theoretical Linux driver question: if I have a LED display with push buttons on my board, and I push a button that should do something eventually rather than something right away, should I be looking at tasklets or softIRQs or neither? Jan 25 16:06:36 bertszoghy: What do you mean by "do something eventually"? Jan 25 16:06:59 If you want it do so something in task context use a threaded IRQ handler. Jan 25 16:07:40 If you want to actually delay then schedule_delayed_work() is normally what you're looking for. Jan 25 16:15:12 03Martin Jansa  07master * r13f761d770 10openembedded.git/recipes/xorg-lib/ (25 files in 2 dirs): Jan 25 16:15:13 pixman: upgrade from 0.21.2 to 0.21.4 Jan 25 16:15:13 Acked-by: Koen Kooi Jan 25 16:15:13 Signed-off-by: Martin Jansa Jan 25 16:15:15 thanks broonie; yes, a threaded IRQ handler is what I need to look into. Jan 25 16:15:33 Depending on how much later, perhaps you should consider user-space. Jan 25 16:16:10 ok, updated testing matrix w/ this weeks results Jan 25 16:17:41 khem, any chance you've got the uclibc arm+samba or ppc e500v2+gcc-cross-sdk things fixed? :) Jan 25 16:18:00 * Tartarus is nearly 100% building, aside from perl-native races Jan 25 16:19:22 understood mwester Jan 25 16:43:35 What is OE's license? Jan 25 16:43:45 It doesn't seem to be listed anywhere on the OE wiki. Jan 25 16:44:49 o.O? Jan 25 16:44:58 recipes are MIT Jan 25 16:45:12 jeremiah: see http://git.openembedded.org/cgit.cgi/openembedded/tree/COPYING.MIT Jan 25 16:45:13 hm classes Jan 25 16:45:25 mario-goulart: Ah thanks. :) Jan 25 16:45:50 you're welcome Jan 25 16:53:45 ugh. insert_feed_uris() tries to populate tmp/rootfs/micro-image/etc/opkg/${feed_name}-feed.conf even no online-manager is chosen (and as such no etc/opkg dir exists) Jan 25 17:06:43 did anyone other than khem have any thoughts on the removal of the coreutils-native build and avoiding a recipe for stagemanager-native? (4 patches in patchwork) Jan 25 17:11:52 03Tom Rini  07org.openembedded.dev * r2cb0d78be2 10openembedded.git/conf/machine/p2020rdb.conf: Jan 25 17:11:52 p2020rdb: Add xserver-xorg stuff Jan 25 17:11:53 Signed-off-by: Tom Rini Jan 25 17:13:41 So, I'm in favor of that too of course, but trying to avoid acking each others work :) Jan 25 17:13:54 * kergoth_ nods Jan 25 17:14:17 i'm kind of surprised it didn't get many comments, considering its one less thing we have to build, but maybe i should have posted a 0/4 email summarizing the intent Jan 25 17:15:38 So, hmm Jan 25 17:15:50 My attempt at moving perl-native to the front of the line hasn't appeared to go too haywire Jan 25 17:16:02 Will need to, uh, check the logs, and make sure it really used the right perl next Jan 25 17:16:14 did you see that the poky guys shifted the components of the toolchain bulid into separate parts of sysroots, to avoid issues? Jan 25 17:16:28 seems like we should use that pattern for a number of different similar situations, including perl-native/python-native Jan 25 17:17:31 Ah yes, I missed you having signed off when I said this: Jan 25 17:18:13 What I just mis-pasted into #poky again, but in short, some folks really need us to provide perl-native as the perl we use Jan 25 17:18:40 ah, right, makes sense Jan 25 17:20:01 so, i was thinking, for something like ncurses where native is only built for tic, if the app supports multiple objdirs, we could just build a native one right there Jan 25 17:20:14 rather than re-unpacking/patching/etc and adding more tasks in the runqeuue Jan 25 17:20:31 Doing that cleanly would be good even more so for things like newer icu Jan 25 17:20:48 icu-4.6 cross demands the build dir for the -native Jan 25 17:20:55 which isn't unpossible if you set aside rm_work Jan 25 17:21:24 And, yeah, OK, autoconf/automake used the perl-native I built Jan 25 17:23:11 * Tartarus gets a little scared Jan 25 17:25:12 03Andrea Adami  07org.openembedded.dev * r284a274dd7 10openembedded.git/recipes/linux/linux-kexecboot_2.6.37.bb: Jan 25 17:25:12 linux-kexecboot: reset DEFAULT_PREFERENCE = "-1" Jan 25 17:25:12 * changed by mistake with commit 9c72d93981212efdbfb87ad2b3c1ac4a8f43daf7 Jan 25 17:25:12 * def. pref. will be set only for specific machines (Zaurus) soon Jan 25 17:25:12 * sorry for the noise Jan 25 17:36:03 hrw, ping Jan 25 17:36:55 pong Crofton Jan 25 17:37:08 hey Jan 25 17:37:14 did you buy a nexus s? Jan 25 17:37:28 Crofton: I did. in best buy Jan 25 17:37:58 does it work in .pl? Jan 25 17:38:02 If I add a new machine, can I put the machine.conf file into a layer? Jan 25 17:38:09 Crofton: yes, it does Jan 25 17:38:15 atl1971, yes Jan 25 17:38:19 data and everything ok Jan 25 17:38:28 Crofton: it has 900/1900/2100 umts and 1700 is not used in europe Jan 25 17:38:33 atl1971, I did exactly that Jan 25 17:38:55 cool, then its just me, ERROR: Unable to determine endianness for architecture 'INVALID' Jan 25 17:39:11 My layer.conf is wrong maybe? Jan 25 17:39:11 does it work at the highest speed it can handle Jan 25 17:39:19 hspda work ok? Jan 25 17:39:27 atl1971, pastebin your machine.conf Jan 25 17:39:32 also, your layer.conf Jan 25 17:40:10 hrw why not a htc hd? Jan 25 17:40:23 woglinde: I want fast phone with latest os Jan 25 17:40:40 hm the htc hd's are the fastest Jan 25 17:41:00 woglinde: and by "htc hd" which you mean? "htc hd2" "htc desire hd"? Jan 25 17:41:04 ot was it dream Jan 25 17:41:09 hd2 Jan 25 17:41:29 dream was g1 and is obsolete now Jan 25 17:41:38 yes right Jan 25 17:41:43 machine.conf http://pastebin.com/NWxhtN6V Jan 25 17:41:44 hd2 is at froyo (2.2) only Jan 25 17:42:29 atl1971 which machine that should be? Jan 25 17:42:59 hrw hm I maybee had waited for 3.0 to come out Jan 25 17:43:07 its a beagleborad Jan 25 17:43:13 layer.conf http://pastebin.com/eBrh2t95 Jan 25 17:43:21 and a phone with tegra Jan 25 17:44:19 woglinde: next time I will be in US during autumn Jan 25 17:44:41 hm ah Jan 25 17:44:59 I already wondered that best buy is in .pl Jan 25 17:45:14 woglinde: nexus s so far is in uk and us only Jan 25 17:45:26 and in uk it is more expensive then in us Jan 25 17:45:31 sure Jan 25 17:46:15 Crofton: doing speedtest on 3G now: 2929/803 kbps got Jan 25 17:49:58 atl1971: are u using master ? Jan 25 17:52:18 atl1971: show bblayers.conf it should be in parallel to your local.conf Jan 25 17:53:58 khem http://pastebin.com/PAWHeXAM Jan 25 17:56:02 I created a peer directory mustang/conf and added a file layer.conf Jan 25 17:56:33 then created mustang/conf/machine/mustang.conf Jan 25 17:57:12 b/c there is a require in my overlay, do I need to bring in the openembedded/conf/machine/include too? Jan 25 18:02:19 atl1971: prolly no Jan 25 18:04:06 atl1971: try something like this http://pastey.net/145334 Jan 25 18:04:11 for your bblayers.conf Jan 25 18:08:48 khem, see my Q above about uclibc issues? Jan 25 18:08:57 * Tartarus linked pastebins to the testing page for this week Jan 25 18:09:12 hm whats with uclibc? Jan 25 18:09:24 Tartarus: you win by 2-0 Jan 25 18:09:28 heh Jan 25 18:09:47 I have been trying to evade ipkg-utils Jan 25 18:10:00 look into those patches and give some acks Jan 25 18:10:04 and fix the fallouts Jan 25 18:10:12 although I expect not many Jan 25 18:10:32 samba one will take time Jan 25 18:10:55 Tartarus: if you can provide login to one of your fast servers I can fix things faster Jan 25 18:10:59 :) Jan 25 18:11:11 * Tartarus looks for an open req Jan 25 18:11:12 right now a simple toolchain build from scratch for me is closely 3 hrs Jan 25 18:11:36 haha Jan 25 18:11:39 I'll see if I can find a chance to try your suggestion of ARM_whatever="arm" Jan 25 18:11:44 k Jan 25 18:12:03 * Tartarus is kicking perl-native stuff around right now since that's hurting in both community and internally Jan 25 18:12:47 you are brave I meant death to perl back when I has to modify test driver for intel compiler Jan 25 18:13:00 oh yes, death to perl Jan 25 18:13:08 But still, I need to stop having race condition build failures Jan 25 18:13:35 all that said, need to bug maybe fray and see if he can release some WRS make perl relocatable patches Jan 25 18:13:54 out of the blue: can the ARM_whatever be tried with unsupported machines? Jan 25 18:14:06 Since the big downside to making everything use perl-native is it kills packaged stagings win there (same w/ sstate) Jan 25 18:14:19 hi gnutoo Jan 25 18:14:25 dinner now Jan 25 18:14:26 bertszoghy, is the thing to turn off thumb mode and isn't usually needed Jan 25 18:14:28 till later Jan 25 18:14:51 woglinde, hi Jan 25 18:14:53 mwester: our qemu has patches to get it working on some obscure arches like SH4 etc and for ppc they dont provide the roms. depending upon what you want it could be safe or could not be Jan 25 18:18:58 k Jan 25 18:42:14 * cbrake adds qt4 images to testing combinations Jan 25 18:43:58 its interesting, with a new disk, my system hardly slows down at all during builds (user response). I wonder if having nearly full disks causes a lot of I/O problems in Linux that slows things down ... Jan 25 18:48:46 heh Jan 25 18:48:57 what're some good qt targets to add, for minimal? Jan 25 18:50:33 qt4-x11-demo-image Jan 25 18:50:42 qt4e-demo-image Jan 25 18:50:50 is what I'm building under Angstrom Jan 25 18:51:02 I've not tried them under minimal, but that would be a good test Jan 25 18:52:19 both? ok Jan 25 18:52:30 any clue if they should work on uclibc? Jan 25 18:52:31 or nfw Jan 25 18:54:11 * Tartarus puts them in uclibc for the moment, and turns them on in the angstrom boards he's been donig too Jan 25 18:56:53 Tartarus: yeah, I've never tried them under uclibc Jan 25 18:57:44 cbrake: qt4-x11-demo-image and qt4e-demo-image build under minimal too but it takes whole day on my box to link webkit :) Jan 25 18:58:04 Tartarus: last time I tried they did not work on uclibc Jan 25 18:58:15 Tartarus: but I discounted uclibc Jan 25 18:59:04 OK Jan 25 18:59:09 i'll give 'em a run this week Jan 25 18:59:11 khem: yeah, qt takes forever to build Jan 25 18:59:11 and see what happens Jan 25 18:59:28 if they all fail pretty badly, i'll pastebin the logs and you can tell me if that's a hard no :) Jan 25 18:59:31 or a lemme fix Jan 25 19:00:14 I wonder why openwrt does not use OE Jan 25 19:00:24 * Tartarus does a quick test to see if he's got the matrix configured right Jan 25 19:02:40 is gpe still used actively Jan 25 19:02:42 khem dont know either Jan 25 19:02:48 they like there makesystem Jan 25 19:03:17 hmm is it offshoot of buildroot somehow Jan 25 19:03:53 now that uclibc is well supported in OE they can exploit that Jan 25 19:04:13 yep and I like their init/config(uci) system also Jan 25 19:04:27 hmm Jan 25 19:04:55 was referring to owrt :) Jan 25 19:14:52 ok, got qemu* + those qt images added Jan 25 19:15:07 along w/ p2020rdb/p2020ds/mpc8315e-rdb (on minimal and minimal-uclibc) Jan 25 19:15:20 and added them to angstrom 2008.1 and 2010.x and mini*/beagleboard/hawkboard Jan 25 19:15:31 Maybe that'll push this build cycle over 15h finally Jan 25 19:15:46 * Tartarus refrains from doing the math on how many combos he's up to, now Jan 25 19:24:03 is there some way to force OE building a recipe each time I build an image? Jan 25 19:24:27 I mean, even if there's no new version available Jan 25 19:29:31 pespin: why would you want to do that Jan 25 19:30:48 khem, cause I have a svn repo locally without write access, and I'm using local-builds Jan 25 19:31:15 and I want to be able to compile whenever I want without having to change config files Jan 25 19:32:16 will bitbake -c clean ; bitbake be acceptable ? Jan 25 19:34:18 pespin: if you use AUTOREV, bitbake will compile the recipe whenever the SVN rev changes Jan 25 19:34:43 pespin: I do that often for custom applications that need to be included in an image Jan 25 19:35:05 cbrake, but I don't have write access to svn repo, so I can't push new revs Jan 25 19:36:58 khem, acceptable but not optimum :P Jan 25 19:37:39 khem, would be great to have something similar to AUTOREV, called for example ALWAYS. Which would build it always :P Jan 25 19:54:39 03Víctor Manuel Jáquez Leal  07org.openembedded.dev * ree0f1b93ec 10openembedded.git/conf/distro/minimal.conf: Jan 25 19:54:39 minimal.conf: inherit devshell task Jan 25 19:54:39 Everybody loves devshell Jan 25 19:54:39 Signed-off-by: Víctor Manuel Jáquez Leal Jan 25 19:54:39 Acked-by: Michael 'Mickey' Lauer Jan 25 19:54:39 Acked-by: Frans Meulenbroeks Jan 25 20:00:41 does the angstom_2010.X console-image work on a beagleboardXm ? Jan 25 20:02:03 it better should Jan 25 20:03:26 ;) Jan 25 20:03:50 with the same tmp dir can you change machine type ? Jan 25 20:04:22 about to bump tzcode and tzdata, but I am wondering why the native recipe is in a different dir than the target one Jan 25 20:06:32 khem: hi, yes I survived my own snowboarding skills Jan 25 20:06:59 heh Jan 25 20:08:10 eFfeM: which native recipe ? Jan 25 20:11:34 khem, tzcode package has a recipe tzcode-native; this builds tzcode and tzdata native, but we also have a dir tzdata which builds tzdata for target, both seem to use teh same tzdata sources (but tzcode also uses tzcode src); seems somewhat odd Jan 25 20:12:36 btw if TZDATA_PV is updated do I need to bump PR ? Jan 25 20:12:49 nope, back to r0 i think Jan 25 20:12:54 it's only like r0 or r1 now isn't it? Jan 25 20:12:59 tzcode does not build tzdata Jan 25 20:13:00 does it Jan 25 20:14:32 nope, false alarm, sorry, noticed it pulls tzdata, that got me confused Jan 25 20:14:39 Tartarus: it is at r0 Jan 25 20:16:18 then r0 it stays :) Jan 25 20:17:55 Tartarus: ok there it is Jan 25 20:18:03 03Frans Meulenbroeks  07org.openembedded.dev * r62ee559f8d 10openembedded.git/recipes/tzcode/tzcode-native_2010n.bb: Jan 25 20:18:04 tzcode: moved to 2011a Jan 25 20:18:04 Signed-off-by: Frans Meulenbroeks Jan 25 20:21:49 testing tzdata, will push it in 30 mins or so had to rebuild a fw deps too, didn't have time to update my tree last week Jan 25 20:52:28 hm, eglibc-initial fails for me Jan 25 20:52:29 | /home/frans/oe/tmp_minimal/sysroots/i686-linux/usr/bin/install: cannot stat `/home/frans/oe/tmp_minimal/work/armv5te-oe-linux-gnueabi/eglibc-initial-2.11-r21.8+svnr12231/build-arm-oe-linux-gnueabi/gnu/lib-names.h': No such file or directory Jan 25 20:53:39 actually it already seems to fail before Jan 25 20:54:27 http://www.pastie.org/1497281 Jan 25 20:55:24 seems cc1 is not found, but just before that it did build gcc-cross-initial: Jan 25 20:55:25 NOTE: package gcc-cross-initial-4.5-r31.1+svnr168622: task do_rm_work: Succeeded Jan 25 20:56:04 * eFfeM is puzzled Jan 25 20:56:23 khem any idea ? Jan 25 20:58:00 Anyone rfc'd killing perl 5.8.8 yet? Jan 25 20:58:13 or at least dropping the D_P = -1 from 5.10? Jan 25 21:00:26 khem posted a patch iirc Jan 25 21:02:15 if he did it's archived Jan 25 21:02:19 not sure if it got applied Jan 25 21:02:22 let me check Jan 25 21:02:42 Don't think so, unless it was this week Jan 25 21:02:47 on last weeks testing-next atm Jan 25 21:02:48 eFfeM call me after you are done with this thread here. Jan 25 21:03:17 Tartarus: can't find it, guess I'm mistaken with something else Jan 25 21:03:21 k Jan 25 21:03:27 i'll rfc it soon enough then i hpoe Jan 25 21:03:29 ka6sox: feel free to speak up or pm Jan 25 21:03:39 * Tartarus crosses fingers Jan 25 21:04:24 eFfeM I think we should ask if anyone has got a backup of the hh.org source repo on the ML. Jan 25 21:05:07 ka6sox: yes Jan 25 21:05:11 let me check what I have Jan 25 21:05:46 Tartarus: found it, khem did commit a patch for perl 5.10.1 for minimal; this was acked by you and me Jan 25 21:06:14 so only slightly crazy, thanks ;) Jan 25 21:07:46 ka6sox: I have 43 handhelds.org tarballs here but no git or so Jan 25 21:08:07 eFfeM okay I just sent an email to the ML. Jan 25 21:08:29 ka6sox: list: http://www.pastie.org/1497322 Jan 25 21:09:27 eFfeM I'm pretty sure that bluelighting can provide the opie stuff. Jan 25 21:09:40 that only leaves like 5 that are missing Jan 25 21:09:55 eFfeM reply to my message with that list of "wants" Jan 25 21:11:23 ka6sox: it has not arrived here yet; guess zecke can provide src for qpe-games Jan 25 21:12:01 eFfeM ya, if we ask maybe we can get it all. Jan 25 21:27:32 Tartarus: my patch was to use it in minimal Jan 25 21:27:36 not DP -1 Jan 25 21:27:42 so go ahead Jan 25 21:27:44 even delete it Jan 25 21:27:50 well no Jan 25 21:27:53 make it DP =1 Jan 25 21:28:48 eFfeM: whenever you have to rebuild toolchain do this bitbake -c clean eglibc eglibc-initial gcc-cross gcc-cross-initial gcc-cross-intermediate; bitbake gcc-cross Jan 25 21:28:59 toolchain bootstrap steps on each others toes Jan 25 21:29:13 hi Jan 25 21:29:27 jo obi Jan 25 21:30:21 what's the best way to have OpenEmbedded download all the packages without doing a build? Jan 25 21:31:12 misham: bitbake -c fetchall Jan 25 21:31:57 Tonight: 0330-0430UTC melo planned outage: need to configure drbd and ganeti for better reliability of our cluster. Jan 25 21:32:03 khem: thank you Jan 25 21:40:46 khem, will try that thought i got it all fixed Jan 25 21:41:17 khem, does fetchall fetch all or only the "active" version of a recipe Jan 25 21:41:22 I thougth the latter Jan 25 22:50:53 khem, with your command gcc builds fine Jan 25 22:55:29 nite all Jan 26 00:46:48 03Khem Raj  07master * rd8d40b0158 10openembedded.git/classes/ (3 files): (log message trimmed) Jan 26 00:46:49 classes: Replace ipkg-make-index with opkg-make-index Jan 26 00:46:49 * STAGE_PKGMANAGER use opkg Jan 26 00:46:49 * rename IPKGBUILDCMD to OPKGBUILDCMD Jan 26 00:46:49 * Use opkg-build instead of ipkg-build Jan 26 00:46:49 Signed-off-by: Khem Raj Jan 26 00:46:49 Acked-by: Koen Kooi Jan 26 00:46:55 03Khem Raj  07master * r7de073be0e 10openembedded.git/conf/ (bitbake.conf distro/nylon.conf distro/openwrt-sdk.conf): Jan 26 00:46:55 bitbake.conf,nylon.conf,openwrt-sdk.conf: Use opkg-build instead of ipkg-build Jan 26 00:46:55 * Rename IPKGBUILDCMD to OPKGBUILDCMD Jan 26 00:46:55 Signed-off-by: Khem Raj Jan 26 00:46:55 Acked-by: Koen Kooi Jan 26 00:46:55 Acked-by: Graham Gower Jan 26 00:46:56 03Khem Raj  07master * r5bd4e2057d 10openembedded.git/conf/distro/include/slugos.inc: Jan 26 00:46:57 slugos.inc: Dont lock SRCDATE for ipkg, its not used anymore Jan 26 00:46:57 Signed-off-by: Khem Raj Jan 26 00:46:57 Acked-by: Koen Kooi Jan 26 00:46:58 Acked-by: Graham Gower Jan 26 00:47:01 03Khem Raj  07master * rd1c65e9a7d 10openembedded.git/recipes/ipkg-utils/ (12 files in 3 dirs): Jan 26 00:47:01 ipkg-utils: Remove the recipes Jan 26 00:47:01 * So that they are not accidently used in future Jan 26 00:47:01 Signed-off-by: Khem Raj Jan 26 00:47:03 Acked-by: Koen Kooi Jan 26 00:47:03 Acked-by: Graham Gower Jan 26 00:47:08 uh-oh. Jan 26 00:47:18 stage-manager: Use opkg utilities Jan 26 00:47:18 * Rename stage-manager-ipkg -> stage-manager-opkg Jan 26 00:47:18 * Rename stage-manager-ipkg-build -> stage-manager-opkg-build Jan 26 00:47:18 Signed-off-by: Khem Raj Jan 26 00:47:18 Acked-by: Koen Kooi Jan 26 00:47:18 Acked-by: Graham Gower Jan 26 00:47:19 03Khem Raj  07master * r2433813d40 10openembedded.git/recipes/ (9 files in 2 dirs): Jan 26 00:47:19 recipes/meta recipes/tasks: Replace ipkg-utils with opkg-utils Jan 26 00:47:20 Signed-off-by: Khem Raj Jan 26 00:47:20 Acked-by: Koen Kooi Jan 26 00:47:21 Acked-by: Graham Gower Jan 26 00:47:21 03Khem Raj  07master * r22399a2c45 10openembedded.git/classes/ (package_ipk.bbclass sourceipk.bbclass): Jan 26 00:47:22 classes/package_ipk.bbclass,classes/sourceipk.bbclass: Use opkg-utils-native instead of ipkg-utils-native Jan 26 00:47:23 Signed-off-by: Khem Raj Jan 26 00:47:24 Acked-by: Koen Kooi Jan 26 00:47:24 Acked-by: Graham Gower Jan 26 00:47:56 mwester: Do you know how can I boot kernel over network on slug Jan 26 00:48:18 bootloader can't do it. Jan 26 00:48:31 So you have to use kexec Jan 26 00:49:29 hrmmm Jan 26 00:49:37 But there's a bug in the ixp4xx network driver that doesn't reset the PHY, so kexec will result in no network if you use the current kernel. No idea if this was ever fixed in a newer kernel -- I reported it but my patch was rejected. Jan 26 00:49:51 k Jan 26 00:50:01 we still use 2.6.27.x for slug Jan 26 00:50:12 did the kernel got rebased to something newer Jan 26 00:50:31 I was trying tftp option on redboot Jan 26 00:50:34 only because I've not gotten around to testing newer kernels -- mainly because it's so painful to boot them and test them! Jan 26 00:51:00 yeah, the tftp option doesn't work in the linksys version of redboot. Jan 26 00:51:28 And the apex bootloader that we use as a second-stage bootloader doesn't have a driver for the ixp4xx nic Jan 26 00:51:31 so may be we could provide a redboot to slug which can boot from network Jan 26 00:51:43 why do we need apex Jan 26 00:51:56 what does it do that redboot wont Jan 26 00:52:26 You'd need to write a driver for the NIC to make redboot work -- that's the root problem. (we use apex to initialize extra RAM for modified slugs, and to switch endianness if necessary) Jan 26 00:53:11 There's a recipe in OE for "arm-kernel-shim" IIRC; that was what we used to use before apex. Very hacky, very limited. Jan 26 00:55:16 I think I got 2.6.31 to boot at one point, and I may have even committed the patchset... that was a long time ago, before I ended up on "the-contract-from-hell" and sorta lost touch with SlugOS... Jan 26 01:24:48 mwester: ok, can you try to remove those two options about sysfs I mentioned and boot that kernel ? Jan 26 01:25:11 I am interested to either apply or discard the budybox modification Jan 26 01:25:20 I'm on my way to Detroit; won't be back until the weekend. Jan 26 01:25:29 ah the auto city Jan 26 01:25:50 Well, that's what it used to be. It's now reverting to agriculture. Jan 26 01:26:00 even better :) Jan 26 01:26:06 I'll try a kernel this weekend. Jan 26 01:26:18 mwester: can I somehow ask it to have the kernel on sdcard Jan 26 01:26:21 or usb stick Jan 26 01:26:44 Nope - redboot knows only about the internal 8MB of flash. Jan 26 01:26:45 so I can replace the kernel on usb stick and if it fails to boot then I can plugout usb stick Jan 26 01:26:52 hmmm Jan 26 01:27:08 But kexec will boot a kernel from anywhere... Jan 26 01:27:16 I was trying to avoid bricking Jan 26 01:27:38 hmm kexec hmm Jan 26 01:27:42 has it worked on slug Jan 26 01:27:48 and is there some docs Jan 26 01:27:55 on how to set it up on slug Jan 26 01:28:04 You won't "brick" it -- you just have to use the reset-button technique to enter upgrade mode, and reflash the entire device. Jan 26 01:28:25 hmmm Jan 26 01:28:46 kexec has worked; I use it to do all my kernel work on the slug. Might even be sample code in /boot. Jan 26 01:29:50 The only big issue with it is a flaw in the kernel; the PHY is not reset so the network won't come back up. Jan 26 01:30:21 thats fine Jan 26 01:30:26 as long as kernel boots Jan 26 01:30:48 there is nothing apparent in /boot wrt kexect Jan 26 01:32:31 Hmm... there was at one time. Probably cut removed from teh final release images. Check to see if it's still in OE in the slugos-init recipe (under files) **** ENDING LOGGING AT Wed Jan 26 02:59:57 2011