**** BEGIN LOGGING AT Fri Jun 07 02:59:58 2013 Jun 07 06:21:33 <_alex_kag_> hello, i whant delete some files from initscripts (mountnfs.sh, umountnfs.sh). What is the better way to do it? (I don't whant modify original files) Jun 07 07:04:02 _alex_kag_: own layer with an .bbappend of the initscript-recipe Jun 07 07:09:19 <_alex_kag_> , i don't know, how to use .bbappend to remove. I know only how to add. Jun 07 07:23:59 has anyone had libpseudo-related issues recently? I get an error that libpseudo can't be preloaded during do_rootfs when installing udev-hwdb Jun 07 07:25:16 how can i force yocto to rebuild my hard drive (.ext3 file) on bitbake Jun 07 07:29:13 erbo: One other person recently reported that. Jun 07 07:29:27 It appears to be host-specific but I don't have details. I can run the same build he's running and not see it. Jun 07 07:30:02 If you come up with steps to reproduce it, or details on how you're doing it, please send me email (peter.seebach@windriver.com), because I want to track that one down. Jun 07 07:31:15 seebs: I'll try to get some time to narrow it down, if I find anything useful I'll email you Jun 07 07:31:39 Thanks. Even just, say, a bitbake.conf and host system info (distro, version, etc.) might help. Jun 07 07:31:40 happens to a colleague as well, so we're up to three :) Jun 07 07:32:04 *ahem* "It's user error." Jun 07 07:32:14 (I always say this on pseudo bugs, because then if it is I look really smart.) Jun 07 07:32:41 hehe Jun 07 07:34:30 How can i force bitbake to rebuild my rootfs? Jun 07 07:34:58 <_alex_kag_> try -f Jun 07 07:36:08 thanks Jun 07 07:38:53 _alex_kag_: generally: the recipes name is a.bb you can create in your layer a a.bbappend and in this you can write do_install_append. here you can remove your files Jun 07 07:40:45 <_alex_kag_> Thanks. Jun 07 07:54:10 worked. didn't update my symlinks to it but it's fine :) thank you Jun 07 07:54:54 good morning Jun 07 08:16:58 morning all Jun 07 08:24:01 bluelightning: hi Jun 07 08:24:09 hi zecke Jun 07 08:24:19 bluelightning: I found a funny dylan issue. I added a PACKAGE_ARCH = "all" to a package and bumped the PR Jun 07 08:24:53 bluelightning: now it tries to look for a -rNEW-1 in the all dir.. while it is still in the armv5te direcyory Jun 07 08:25:00 cleanall doesn't remove that 'state' either Jun 07 08:26:06 hmm Jun 07 08:26:15 I guess that's a case that hasn't been tested Jun 07 08:26:24 zecke: can you please file a bug? Jun 07 08:27:10 zecke: there is one already, mmt Jun 07 08:27:24 zecke: https://bugzilla.yoctoproject.org/show_bug.cgi?id=4102 Jun 07 08:27:25 Bug 4102: normal, Medium+, 1.5, richard.purdie, NEW , Incremental builds do not work well when renaming recipes or changing architecture. Jun 07 08:27:39 zecke: you need to remove old pkgdata for it to work Jun 07 08:28:15 JaMa: yes I resorted to find tmp -name "*package-name*" -delete Jun 07 08:30:40 JaMa: thanks! Jun 07 08:50:05 has anyone actually tried using qmake on the target?cd Jun 07 08:50:25 I get "sh: -d: command not found" Jun 07 08:52:24 Net147: I know at least one person has used it in the recent past Jun 07 08:52:50 it's not something we test regularly though Jun 07 08:53:14 with the introduction of automated testing in 1.5 we should certainly look at it though Jun 07 08:53:37 s/introduction of/introduction of enhanced/ Jun 07 08:53:50 bluelightning: I am trying to get Qt Creator running on target. I have submitted all the patches I needed to get it compiled. Jun 07 08:53:59 bluelightning: it's running now, just that qmake isn't working. Jun 07 09:03:55 root Jun 07 09:50:06 quit Jun 07 09:50:16 :) Jun 07 10:28:16 Is there a way to install closed-source software on yocto? Jun 07 10:28:21 Like adobe flash? Jun 07 10:28:34 to install - to build a recipe for Jun 07 10:37:20 ivali: sure, you can write a recipe to package anything, it doesn't have to be something built from source Jun 07 11:25:08 thanks Jun 07 12:32:11 what is a/the swabber-report? see meta/classes/image-swab.bbclass Jun 07 12:36:09 silviof: https://www.yoctoproject.org/tools-resources/projects/swabber Jun 07 12:36:51 silviof: it's not really maintained atm as far as I am aware Jun 07 12:50:40 bluelightning: thx for the link Jun 07 14:03:19 kergoth: I assume you're ok with the bitbake-worker split from runqueue? Jun 07 15:22:57 RP: maybe I've missed it somewhere, but pseudo-native check and build will be also resolved without wrapper, right? Jun 07 15:23:30 JaMa: it has been in that changeset yes Jun 07 15:26:45 JaMa: newer bitbake will be able to handle that internally. I've not posted the patches yet but see http://git.yoctoproject.org/cgit.cgi/poky-contrib/log/?h=rpurdie/t8 Jun 07 15:27:12 JaMa: its working well locally, just ironing out a few minor issues Jun 07 15:28:49 RP, the fialure case (spawn) where we occasionally get the can't LD_PRELOAD. Will any of that be hidden by the way it's being spawned now.. (I suspect no, but just curious if anythings changed there) Jun 07 15:29:34 fray: it should still hit the console/logs, just later in the build Jun 07 15:29:40 ok Jun 07 15:30:02 that was the one issue that I've noticed, it doesn't always his the logs (today).. if we could get it into the logs, it would help us spot failures Jun 07 15:30:27 fray: I'd expect that would be easier now at the very least Jun 07 15:30:56 nice Jun 07 15:33:40 RP: ah ok, I was just surprised that your e-mail was talking mostly about python version and native replacements, I really like this idea (one less reason for my occassional hangs) Jun 07 15:34:36 JaMa: btw, I did see a hang that resembled your report where I broke the PATH manipulation in that script and it infinitely executed itself with no output :) Jun 07 15:35:19 JaMa: I haven't forgotten about those hangs, I'm just struggling to understand what is going on there :/ Jun 07 15:36:03 I'm fine with having them until this new bitbake-worker is implemented Jun 07 15:36:41 I'm still inclined to think that it has something to do with /dev/shm shared between multiple bitbake calls Jun 07 15:38:46 JaMa: could well be, I tend only to run one build at a time, the autobuilders only run two Jun 07 15:39:14 JaMa: I'm not sure bitbake-worker is going to fix this though as you appear to be seeing it really early :/ Jun 07 15:39:24 unless its pseudo related... Jun 07 15:39:39 the new code does move pseudo into a much more manageable place Jun 07 15:42:08 yes I'm running them even in different chroots (one gentoo based shr-chroot, 2nd ubuntu chroot for work) but both sharing the same /dev/shm from host Jun 07 15:42:26 and I don't remember seeing hang when I was only in one chroot (running one build) Jun 07 15:50:57 RP: not opposed to bitbake-worker as long as it doesn't impact performance terribly, and I'm sure you're testing that Jun 07 15:57:35 kergoth: it actually improves it! Jun 07 15:57:54 kergoth: 3% speedup on a core-image-sato Jun 07 15:57:58 I'm not surprised by that.. especially on the newer truely multithreaded machines Jun 07 16:18:27 Is there a way to find the absolute path to a '.inc' file? Jun 07 16:18:41 Hmm wrong channel.. but someone might know here as well Jun 07 16:35:29 fray: In what context? Jun 07 16:35:37 would BBINCLUDED help? Jun 07 16:35:45 RP: see #oe Jun 07 16:35:49 I ended up just using ${COREBASE}.. Jun 07 16:36:08 but in short.. I have one recipe using the .inc from another recipe (in another layer).. and I need a way to get to that path so I can reference the files.. Jun 07 16:36:23 qemu_custom.bb -> require recipes-devtools/qemu/qemu.inc .... Jun 07 16:36:39 FILESEXTRAPATHS_prepend := "${COREBASE}/meta/recipes-devtools/qemu/files:" Jun 07 16:37:17 that is probably reasonable if not ideal Jun 07 16:37:52 ya, what I was looking for is something that would tell me the path of 'recipes-devtools/qemu/qemu.inc' Jun 07 16:38:07 fray: hmm... is /files necessary? Jun 07 16:38:25 I'll have to go back and look, I think it was.. Jun 07 16:38:34 fray: I think by virtue of using FILESEXTRAPATHS you'll get /files and /${BPN}/ etc. added automatically Jun 07 16:38:34 I'm modifying some stuff someone else hacked up.. at at least works now Jun 07 16:39:39 hmm Jun 07 16:39:42 that might not be right Jun 07 16:39:57 my head hurts at the moment... Jun 07 16:39:57 fray: bb.utils.which() against BBPATH Jun 07 16:40:01 ya, it's needed.. Jun 07 16:40:09 ok, my mistake Jun 07 16:40:15 kergoth.. thats what I was thinking, there had to be something like that.. Jun 07 16:40:15 * kergoth ponders Jun 07 16:40:51 anyway, the corebase is fine for this.. and it makes sure that no other layers going to get in the way.. Jun 07 17:24:27 hi hollisb Jun 07 17:24:47 hey mranostay Jun 07 17:31:58 halstead: around? Jun 07 17:35:08 sgw_: yes. Grabbing coffee right now. Jun 07 17:37:09 halstead: when you are back, we need to do a update to the AutoBuilders python, want to get an estimate from you. Jun 07 17:42:38 sgw_, I saw the e-mail this morning. ab06 meets the requirements but none of the other builders do. Jun 07 17:44:46 halstead: How hard will it be to update those? Jun 07 17:46:45 sgw_, If we can use the local versions built in ~pokybuild/bin that it's easy. If we have to upgrade system wide that will be an issue. Jun 07 17:47:34 halstead: let's work with pokybuild then. Jun 07 17:48:01 sgw_, I'll need to add "source /home/pokybuild/sdk/environment-setup-i586-poky-linux" to the biuld steps. Jun 07 17:49:14 sgw_, I'll reply to the e-mail and then get started. Jun 07 17:59:54 Anyone given any thought to how to handle selection of gnutls vs openssl and the like for recipes that support such things? separate variable? packageconfig? the problem with the latter is it doesn't really grasp this-or-that choices, so you'd want to add some sort of 'conflicts' handling Jun 07 18:22:46 hollisb: so be honest how many days do we get like this a year? :P Jun 07 19:22:47 kergoth / RP -- I found a case where I have to manually remove the bitbake cache to trigger a reparse.. Jun 07 19:22:56 configure a project, inherit the blacklist.bbclass.. Jun 07 19:23:02 build something (setup the cache) Jun 07 19:23:18 then go in and add: PNBLACKLIST[recipe] = "reason" to the local.conf, try to build the recipe.. Jun 07 19:23:25 it goes right on and tries to build.. Jun 07 19:23:43 (that is bitbake as of Monday) Jun 07 19:38:56 mranostay: almost guaranteed 4th July through October is like this Jun 07 19:39:06 mranostay: June can go either way Jun 07 19:41:35 hollisb: at least there is beer for the rain :P Jun 07 19:47:02 fray: not entirely surprising, please file a bug Jun 07 19:49:47 will do Jun 07 19:53:12 RP -- 4627 **** ENDING LOGGING AT Sat Jun 08 02:59:58 2013