**** BEGIN LOGGING AT Fri Apr 04 02:59:59 2014 Apr 04 07:35:38 good morning Apr 04 08:19:14 mckoan, morning Apr 04 08:27:24 hi kroon Apr 04 08:32:37 morning all Apr 04 08:47:07 hi bluelightning Apr 04 08:47:38 hi mckoan Apr 04 09:07:40 bluelightning: I'm trying to run buildhistory-diff on an OE-built distro ('dora' based) and I get 'NameError: global name 'report_ver' is not defined' Apr 04 09:08:03 bluelightning: http://lists.openembedded.org/pipermail/openembedded-core/2013-October/085321.html and the backport https://github.com/openembedded/oe-core/commit/a51d96c44e6feac8322284c54bfc01ef598f8821 don't seem to match Apr 04 09:10:03 bluelightning: 'git cherry-pick b7de1eaac9eed559b2d68058f5de67de74a6cb58' in the dora branch fixes it for me Apr 04 09:10:09 are you seeing the same? Apr 04 09:10:30 I don't think I've tried buildhistory-diff on dora recently Apr 04 09:10:47 I'll follow up on this with Robert though Apr 04 09:10:49 thanks Apr 04 09:15:22 I've sent the cherry-pick as patch to the list Apr 04 09:22:37 koen: thanks Apr 04 09:26:18 IMAGESIZE changed from 15780 to 13448 (-14%) Apr 04 09:26:20 \o/ Apr 04 09:26:30 still 7MiB too big Apr 04 09:29:51 * koen notices his .gitconfig files have the wrong user.email on the work laptop Apr 04 09:49:49 hehe Apr 04 09:50:27 koen: squeezing into flash? Apr 04 10:29:20 XorA: that's the private goal Apr 04 10:29:42 XorA: the work related goal is to have a systemd based initramfs that isn't a gazillion megs Apr 04 10:35:17 koen: cool, I non systemd one once, used it for years debugging android devices without needing working mmc drivers Apr 04 10:35:54 so much nicer to boot into a real shell in initramfs than braindead android shell Apr 04 10:38:03 I wish I could get a card assigned to integrate dracut properly Apr 04 10:38:10 that would save a lot of reinventing the wheel Apr 04 10:38:25 that would be nice Apr 04 10:38:50 I think its essential for LEG BTW ;-) Apr 04 10:39:37 it does work currently :) Apr 04 10:39:40 but online Apr 04 10:43:13 which incidentally maps quite well to the current way we do tests Apr 04 11:03:49 koen: do you plan to add the image to meta-initramfs? Apr 04 11:04:45 not sure yet Apr 04 11:04:59 that layer was not exactly of wider use so I've been thinking to remove it but maybe we could put there the 'small' initramfs frameworks Apr 04 11:05:15 initramfs is a product of distro in my mind Apr 04 11:05:58 we could not find any better place for klibc Apr 04 11:06:51 must say I just dropped elibc into my initramfs :-) Apr 04 14:42:36 gah I am out of date with bitbake, how do I force it to actually do a re-compile instead of just re-using old build results? Apr 04 14:43:05 XorA: -f -c compile just like the old days... Apr 04 14:43:30 XorA: or -C compile to force compile then run through the rest of the tasks for the recipe Apr 04 14:43:57 or -c cleansstate to clean the recipe and get rid of sstate for it at the same time Apr 04 14:44:04 depends on what you're trying to do really Apr 04 14:44:28 I messed up a kernel build and it has the wrong uImage parameters Apr 04 14:44:50 if you're modifying something in the recipe or files referred to in SRC_URI and it's not rebuilding, that's almost certainly a bug Apr 04 14:45:09 no, its the UBOOT_LOADADDR etc vars Apr 04 14:47:21 that feeds into the u-boot recipe though right? Apr 04 14:47:29 kernel recipe Apr 04 14:47:42 it may be it did rebuild correctly, and something else insane is happening Apr 04 14:47:51 right, the kernel Apr 04 14:48:08 from what I can tell it ought to be picking up the dependency on that... Apr 04 14:48:14 my compiled externally kernel boots, one compiled in OE with same defconfig, git repo, SRCREV does not Apr 04 14:51:23 * koen would suggest adding bootz support to u-boot and switching to zImage Apr 04 14:52:04 koen: I would if it wasnt a google device with evil vendor trees Apr 04 14:52:13 ah Apr 04 14:52:24 * koen just ordered a samsung chromebook Apr 04 14:52:44 I have Angstrom running, but only with an externally compiled kernel Apr 04 14:56:51 the only difference I can think of is toolchain Apr 04 14:57:10 ls Apr 04 14:59:27 . Apr 04 14:59:29 .. Apr 04 14:59:31 uImage Apr 04 15:00:47 scp koen/uImage xora/uImage Apr 04 15:04:08 koen: its in meta-exynos now anyway Apr 04 15:04:33 I'll have a look on monday when the chromebook arrives :) Apr 04 15:05:11 basically if you compile the same kernel with linaro 2013-10 toolchain and defconfig it works Apr 04 15:08:06 hmm, I'm on .11 Apr 04 15:08:17 I guess I could update it to something more recent Apr 04 15:08:32 not sure what Angstrom is on Apr 04 15:08:49 v2013.12 is using linaro-2013.11 Apr 04 15:08:52 bluelightning: wrt what we were chatting about yesterday, i'm not so sure... i did a "find . -name "*avahi*" -print" after a bitbake avahi, then after a clean, cleanall, and cleansstate cache of avahi Apr 04 15:08:53 for arm Apr 04 15:09:04 bluelightning: and there was no change whatsoever in tmp/qemuarm-eglibc/sysroots Apr 04 15:09:17 i don't think any of those targets touches the sysroot Apr 04 15:09:43 tlwoerner: they definitely do, it's clean that does it and the others do a clean Apr 04 15:11:25 heh, the launchpad got stuck at 2013.10 Apr 04 15:13:43 lp is being phased out Apr 04 15:13:47 (much rejoicing) Apr 04 15:13:56 \o/ Apr 04 15:14:55 bluelightning: staging.bbclass, from the comments, appears to clean out the sysroot "if we rerun configure" as a prefunc to do_configure Apr 04 15:14:59 koen: +1 Apr 04 15:15:33 bluelightning: and SYSROOT_DESTDIR is added to the cleandirs for do_populate_sysroot Apr 04 15:15:49 tlwoerner: that code path is not what should be at work here Apr 04 15:17:29 bluelightning: do you have a suggestion of where i should be looking? Apr 04 15:17:56 tlwoerner: I don't know what's going on but I guess I'm going to have to try it here... remind me which branch is this? Apr 04 15:18:55 master Apr 04 15:23:34 bluelightning: maybe if i explained why i'm doing this it might help? (or lead to a different solution?) Apr 04 15:24:28 tlwoerner: sure, that might be useful, but what you're describing does not match up with my expectations of the system's behaviour Apr 04 15:24:56 while working on JaMa "state of bb world" list, i was doing some world builds Apr 04 15:25:09 my build of gvfs failed because a depends wasn't in place Apr 04 15:25:39 the current DEPENDS list for gvfs is "libsoup-2.4 gnome-keyring glib-2.0 avahi gconf libgphoto2" and i discovered that i needed to add libgnome-keyring Apr 04 15:25:58 so now i'm suspicious of the non-lib gnome-keyring Apr 04 15:26:16 so i wanted to clean out my sysroot of libsoup-2.4 gnome-keyring glib-2.0 avahi gconf libgphoto2 Apr 04 15:26:33 so i could try building up the DEPENDS list for gvfs from scratch Apr 04 15:35:42 ping Apr 04 15:35:49 duh, wrong window ;) Apr 04 16:23:29 tlwoerner: hmm I've confirmed the behaviour you're seeing Apr 04 16:23:34 it seems wrong to me... Apr 04 16:31:33 bluelightning: okay, thank you for the confirmation Apr 04 16:32:28 bluelightning: i don't have the history that you have so i can't say that it *feels* wrong, but if there were a mechanism to remove a given package/recipe from the sysroot that would be nice Apr 04 16:33:20 tlwoerner: as far as I know this should be it Apr 04 16:34:15 bluelightning: i was looking at staging.bbclass and thought if there were a converse to "sysroot_stage_dir()" it could be useful Apr 04 16:36:57 it could walk through a package's sysroot-destdir and instead of doing a "find ... | cpio" maybe a "find ... | rm -f"? Apr 04 16:39:04 tlwoerner: this is a bug, so I expect it'll be fixed shortly Apr 04 16:49:21 should i open an issue? Apr 04 16:50:33 I think it might be worth doing that; RP is looking at the problem now though FWIW Apr 04 16:51:28 okay. i'll hold off opening an issue Apr 04 16:54:48 bluelightning: looks like ss['task'] is getting corrupted, digging... Apr 04 17:04:56 tlwoerner: http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/t2&id=086ea6f6bfdc2ddf7faae32f28743ef2eedcc5e7 Apr 04 17:05:10 tlwoerner: the next patch in the tree makes a better job of cleaning things up Apr 04 17:06:46 gah one thing about OE I hate, I always need a faster machine :-) Apr 04 17:06:57 tlwoerner, bluelightning: patches also sent to the list. It will take me a while to test them properly but I'm fairly confident they're right Apr 04 17:08:17 RP: awesome thanks! Apr 04 17:08:20 of course I find the parse error after I send the patch :/ Apr 04 17:09:19 heh, the code in question is completely broken too. If its not been used for this long, I wonder if we need it :} Apr 04 17:11:06 RP: i would certainly be useful for the use-case i described, and i believe that's a valid activity to want to do (?) Apr 04 17:12:13 tlwoerner: your use case is fine, see my next patch for the code I'm talking about (siteconfig) Apr 04 17:42:09 you lot start work to early Apr 04 17:47:43 Crofton: you need to move more to the East :-) Apr 04 17:49:39 yeah Apr 04 17:49:57 a week from tomorrow Apr 04 18:15:41 koen: the minor changes your linux.inc does to .config is what kills the kernel on Chromebook Apr 04 18:16:22 * XorA just copied the .config out back out of OE work/ and into local build Apr 04 18:18:19 bad koen Apr 04 18:19:16 they are not incorrect changes, must expose some bug in the kernel Apr 04 18:19:32 magic Kconfig is always the issue with linux :-( Apr 04 18:21:29 I guess a line by line is needed :-( Apr 04 18:39:03 WTF does CONFIG_AEABI mean? Apr 04 18:40:34 It's got Kconfig help text :) Apr 04 18:41:38 thats what Im failing for find :-D Apr 04 18:42:24 weirder and weirder Apr 04 18:43:40 arch/arm/Kconfig Apr 04 18:43:42 XorA, google it: http://cateee.net/lkddb/web-lkddb/AEABI.html Apr 04 18:44:09 broonie: yeah found it, could have sworn that used to be called CONFIG_EABI :-) Apr 04 18:44:48 XorA: That was a different thing IIRC. Apr 04 18:46:32 anybody know who is handling openembedded.org account requests? Apr 04 18:48:17 on the wiki? Apr 04 18:48:36 correct. JaMa just pinged Khem for me... Apr 04 18:52:07 I suspect I can do that. I should remember how Apr 04 18:52:42 ha Apr 04 19:12:17 or Jefro Apr 04 19:15:03 RP: are you still available? Apr 04 19:54:20 goddamit Im getting to old for this OE stuff Apr 04 19:57:47 Leave it to the kids and go back to MMA like the other oldsters. Apr 04 19:59:51 really need to get round to stop being ill and get a new gym :-( Apr 04 20:00:13 kicking people is more enjoyable than code :-) Apr 04 20:00:27 :) Apr 04 20:01:48 I still, in the back of my mind, wnat to create a bitbake alternative, but the older i get, the harder it gets to find enough motivation for such a project :) Apr 04 20:03:02 kergoth, I think you have been redesigning bitbake for as long as I have know you Apr 04 20:03:26 never have found enough motivation to actually make it happen :) Apr 04 20:03:43 this sounds interesting... meh, i'm going to go kill something in a game instead Apr 04 20:03:48 Apr 04 20:03:51 :) Apr 04 20:12:16 kergoth: thats the trouble of software thats "good enough" :-) Apr 04 20:51:53 JaMa, hmm, my sstate-cache is still growing after each rebuild I do after updating from master. Ideally, the cache should stay at roughly th same size after sstate cache cleaning, right ? Apr 04 20:53:31 kroon: check extra-archs param if you aren't missing any Apr 04 21:51:48 JaMa, i dont think i am Apr 04 21:53:22 ill poke around Apr 04 22:36:30 tlwoerner: I'm back around for a few mins... Apr 04 22:50:07 building a souped up console image for beagleboneblack found another bug Apr 04 22:51:50 nerdboy: what's the bug? Apr 04 22:52:02 gpm install fail Apr 04 22:52:11 just sent a patch to ml Apr 04 22:52:29 ahh yes. Excellent. Apr 04 22:52:55 thanks for the gnuplot license fix, BTW Apr 04 22:53:50 i normally only see the fails building my own stuff for a couple of arm devices Apr 04 22:55:01 er, QA warnings too, but i get a little absorbed trying to get a working image... Apr 04 22:55:54 indeed Apr 04 22:56:24 have to remind myself to push fixes through the process Apr 04 22:56:55 step 1: local commits :) Apr 04 23:35:44 koen: is there a readme on the linux-rt recipe? Apr 04 23:51:17 i take it there's a linux.inc somewhere that goes with that? Apr 05 00:57:39 nm, looks like it needs some preprocessing **** ENDING LOGGING AT Sat Apr 05 02:59:58 2014