**** BEGIN LOGGING AT Wed Feb 24 02:59:58 2016 Feb 24 04:34:33 Hi bitbake fails at last step complaining " do_rootfs) failed with exit code '1'" can anyone please help Feb 24 04:44:11 the complete error is here http://pastebin.com/U4PixfYh it says KeyError: 'getpwuid(): uid not found: 0 Feb 24 06:44:23 Hey guys. Where can I get a list of possible " EXTRA_IMAGE_FEATURES " for Yocto? Feb 24 06:50:14 LetoThe2nd: ^^ Feb 24 06:50:25 Sorry you are the only person on here that has ever responded lol Feb 24 06:54:00 cart_man: hehe, maybe you asked the wrong questions - or at the wrong time ;-) Feb 24 06:54:40 cart_man: basically there's a list of known IMAGE_FEATURES in the yocto documentation. that however is not exhaustive, because any distro or layer can add whatever it wants. Feb 24 07:05:45 LetoThe2nd: I do not see the list of things that you can actually put behind " EXTRA_IMAGE_FEATURES " Feb 24 07:08:45 cart_man: http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#usingpoky-extend-customimage-imagefeatures Feb 24 07:09:15 cart_man: and therefore, http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#ref-features-image Feb 24 07:37:50 LetoThe2nd: Ok thanks allot! Is it possible to build APT-GET into the package? Feb 24 07:40:17 cart_man: we have dpkg, ipkg and rpm. no idea about more 'advanced' package management tools, as i've never needed them Feb 24 07:42:44 cart_man: http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#ref-classes-package Feb 24 07:43:02 cart_man: as well as http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#using-runtime-package-management Feb 24 07:43:37 i think especially the latter is of interest for you. Feb 24 07:44:11 hi all, how to tell bitbake to create an debug image out of the package-split debug folders (dbg ipks)? Feb 24 07:46:19 btooth: do you need anything that extends beyond just building with IMAGE_FEATURES dbg_pkgs and dev_pkgs? Feb 24 07:49:52 LetoThe2nd: i need a full debug image, all parts should be unstripped, so the resulted image under tmp/deploy/images should contain only debug versions Feb 24 07:52:06 btooth: is that just your requirement reworded, or have you actually verified that said feature does not fulfill your needs? Feb 24 07:55:06 LetoThe2nd: okay :) I am going to add them to IMAGE_FEATURES and test.. thanks Feb 24 07:55:37 yw. Feb 24 08:05:39 good morning Feb 24 10:24:08 <_gpg_> hello everybdy Feb 24 10:24:33 <_gpg_> i'm stuck with an issue (meta-fsl-arm) and i'd really appreciate if some can give me an insight Feb 24 10:25:42 <_gpg_> when building my image http://pastebin.com/RBfD82bM when building sgml, it complains that openjade-1.3.2.cat" does not exist or is empty. Feb 24 10:25:52 <_gpg_> i'm really stuck and dont know how to approach this issue Feb 24 10:26:40 <_gpg_> ive read some old posts reparding opensp dependency i've add it and it builds fine Feb 24 13:08:09 hello, is it possible to somehow verifiy if a recipe will compile before running bitbake? Feb 24 13:18:11 cygnus: may be you can check the environment of the recipe with bitbake -e recipe. And then check for do_compile function Feb 24 13:19:07 so I would just type bitbake -e and it would verifiy? Feb 24 13:19:56 and it knows where the recipe it, without giving it the "layer" or something>? Feb 24 13:20:01 cygnus: It would help us know the complete environment recipe has got. So in that you can search for do_compile and see if there are instructions to compile Feb 24 13:21:48 cygnus: If the bblayers has the information about the layer where the recipe is located, then bitbake will automatically pick it up. Feb 24 14:32:12 ok... so I verified the package with bitbake -e Feb 24 14:32:23 seems like... it compiled or something? Feb 24 14:32:36 doesn't seem to output any log message... like a "success" or something? Feb 24 14:32:53 is that how it's supposed to run? Feb 24 14:32:57 :)) Feb 24 14:34:25 hi everybody, i'm new of yocto and i need a recipse to add to my project nagios. Can someone help me? [the one on yocto project website had me some problems] Feb 24 14:34:39 so after the usual .bb recipe configs,I added "inherit autotools" Feb 24 14:43:02 I'm trying to get ModemManager to work, but it doesn't seem to auto start on my system, anything I'm missing? Feb 24 15:28:12 could anyone help me? Feb 24 16:54:40 * armpit more blacklisted recipes i care about Feb 24 16:57:29 armpit, I think i got the ones I care about fixed Feb 24 16:57:38 JaMa, is making us work Feb 24 16:58:16 sorry for not doing your work :) Feb 24 16:58:37 :) Feb 24 16:58:52 I think I got the ones I care and couple I don't care about fixed Feb 24 16:58:53 better to keep builds going and mark troubelsome recipes Feb 24 16:59:03 exactly Feb 24 16:59:27 yep Feb 24 16:59:30 and if it stays broken, then at least my bitbake world wont take so long Feb 24 17:33:41 QUESTION - Has anyone installed mcafee solidcore (scsrvc) or any other virus scan (clamav) to their image? Is it a package that is readily available? Don't ask why :( Feb 24 17:37:19 clamav is in meta-security Feb 24 17:39:44 armpit - Thanks! Feb 24 17:44:09 kergoth: https://bugzilla.yoctoproject.org/show_bug.cgi?id=9020 FWIW. May or may not be related to what you were asking about Feb 24 17:44:10 Bug 9020: normal, Medium+, 2.1 M3, richard.purdie, NEW , populate_sdk fails when I have another version of gcc in my own layer Feb 24 18:28:41 ALL: I have a binary that requires libstdc++.so.5 and my jethro image uses libstdc++.so.6. Can both co-exists in image, and be installed through a simple variable/option?? Feb 24 20:25:21 morning all Feb 24 20:42:09 gah Feb 24 20:42:26 The more I look at the pseudo startup code, the more I wonder why it's taken this long to notice problems. Feb 24 20:42:48 bluelightning: morning, UGT or did you move to different timezone? Feb 24 20:42:49 Like, I am *sure* this can't be totally reliable, and yet, apparently failures have been rare enough that, before we added an abort-on-fail, it wasn't noticed failing particularly often. Feb 24 20:43:19 * armpit bluelightning from the future Feb 24 20:43:31 As in, I don't actually know of a known case of a failure caused by the server restarting code. Which is very odd. Feb 24 20:43:49 And I can't make it happen outside of the heavily-loaded build machines, but it doesn't look like resource failure, and... gah. Feb 24 20:49:51 okay so Feb 24 20:50:07 For reference, when I find "No, really. It works." in a comment, in code I know I wrote but that I am not convinced I understand. Feb 24 20:50:11 It is *not* a big confidence booster. Feb 24 20:51:13 * kergoth chuckles Feb 24 20:51:14 been there Feb 24 20:54:19 (For those just joining us, apparently under heavy load, pseudo can fail to spawn/connect-to a server, and we can't immediately see why. It looks like the server's not logging as-expected in these cases, or something, and we're trying to figure out what's failing. Unfortunately, it won't reproduce in anything reasonably-small.) Feb 24 21:01:38 seebs: Did you recently make a change so the issue becomes more visible? Feb 24 21:02:04 Or rather: What are the symptoms? We run a couple of heavily loaded buildservers and haven't seen issues with pseudo at all. Feb 24 21:10:50 JaMa: I'm living in NZ now Feb 24 21:11:43 i've seen failures contacting the bitbake server more than i've seen failures contacting the pseudo server, but then the build servers i use aren't massive in scale Feb 24 21:16:26 There is a recent change, which is that if the client can't get a server, it calls abort(). Feb 24 21:16:35 Which makes transient failures Much More Visible. Feb 24 21:16:44 But it seems to me that if there have been failures all along, we should have seen SOME of them. Feb 24 21:19:57 seebs: so what does it do if it isn't able to connect to the server? Feb 24 21:20:55 (before that change, that is, we likely don't have it yet) Feb 24 21:21:33 Before that, it would in general just silently-fail. Well, it should have emitted diagnostics, but I think only if debugging was on. Feb 24 21:21:43 So it'd just use raw filesystem data, or not update the database. Feb 24 21:21:58 Thing is, I'd guess 80% or more of pseudo messages end up not mattering. Like, you create a file, then delete it, net result no one cares. Feb 24 21:22:20 Or you call stat, but you're only checking size or whether rwx are available, not setuid bits or ownership. Feb 24 21:22:30 So it isn't totally impossible that it just hasn't happened to matter, but it seems odd. Feb 24 21:23:12 OK, but if we start rolling out permission bits and users on a large scale, we should eventually see it (a) fail, (b) not use the uid/gid/mode we want the files to be? Feb 24 21:27:43 Yes. In theory. But we haven't, and that's what's bugging me. Feb 24 21:27:59 It's conceptually possible that, say, prior to the recentish respawn changes, we *weren't* hitting it. But I can't imagine why, since those changes fixed something that was clearly a bug. Feb 24 21:29:09 Thus: More comprehensive logging (an event-log thing to let me record things and then dump them retroactively if something goes wrong), and thinking about reworking the server startup so that server exit codes can be more informative when daemonizing. Feb 24 22:07:33 Has anyone used the wic image fstype with hddimg/live/bootimg? pretty sure it's not possible at the moment, since live isn't handled like others, and the wic wks for efi requires the hddimg bits, which aren't populated until do_bootimg Feb 24 22:11:33 Well, hmm. I can find at least one plausible-looking race condition, but I can't make it happen without intentionally adding a sleep in the server on shutdown. Feb 25 01:42:14 anyone debugged code in gdb that forks? **** ENDING LOGGING AT Thu Feb 25 02:59:59 2016