**** BEGIN LOGGING AT Tue Sep 03 02:59:58 2013 Sep 03 07:26:38 good morning Sep 03 07:36:47 morning all Sep 03 07:37:55 hi silvio_l_, all Sep 03 07:39:30 hi mckoan Sep 03 07:56:25 morning #oe Sep 03 08:01:02 morning silviof, all Sep 03 12:12:10 hi. why are bitbake commits not sent to openembedded-commits@ mailing list? Sep 03 12:12:33 i was monitoring the list to see when my bitbake patch would be merged, and just noticed that. Sep 03 12:14:50 bluelightning: my bitbake patch for run.do_xxx symlink was merged in master, do you think it could be backported to dylan? when i wrote the patch it applied to both master and dylan branch. i suspect it would still apply. Sep 03 12:14:50 a672b39 build.py: create symlink for run.do_xxx scripts Sep 03 12:17:38 ndec, because OE is not bitbake :) Sep 03 12:17:43 or ebcuase no one asked Sep 03 12:17:57 well, meta-oe commits are sent there too. Sep 03 12:18:07 then reason #2 Sep 03 12:18:10 ;-) Sep 03 12:18:20 maybe i fixed #2 just now, then? Sep 03 12:18:59 I'd ask in the bitbake list Sep 03 12:19:28 I do not think anyone would have a problem adding bitbake commits Sep 03 12:19:38 the volume should not impact list traffic Sep 03 12:42:53 ndec: so, generally we backport only bugfixes Sep 03 12:43:23 ndec: we might be able to take this one as it's fairly trivial, I'll add it to my list anyway Sep 03 12:43:34 ah. ok. thx. Sep 03 13:12:56 bluelightning: JaMa: please be so kind to merge the klibc changes. There is an endianness bug with mips and 3.10 headers exposed by the recipe. I've gathered some attention by the maintainers, you know, it's sunny outside and they will disappear soon as snow ;) Sep 03 13:29:29 ant_work: looks like JaMa is gone; I can do it if you like Sep 03 13:30:19 heh, he left after sending that scary queue of 66 patches :p Sep 03 13:51:51 morning Sep 03 13:54:16 morning chouimat Sep 03 13:54:35 ant_work: I've merged the meta-initramfs changes (3 patches) Sep 03 13:54:59 ant_work: havent done the meta-filesystems one, will need to be the maintainer of that layer Sep 03 14:16:15 bluelightning: many thx. I'll ping koen Sep 03 14:16:57 ant_work: I think it's JaMa for that layer Sep 03 14:17:31 ok Sep 03 15:04:35 Stygia: for future patches can you add [meta-oe] in the subject line i.e. --subject-prefix="meta-oe][PATCH" Sep 03 15:04:49 Stygia: or meta-perl in this case Sep 03 15:07:15 bluelightning, Oh. Yea, sure. Sep 03 15:07:31 bluelightning, Aah, hmm. This means it evades people's mail filters... yea, sorry. Sep 03 15:08:59 Stygia: I already wrote it in review of your first patch Sep 03 15:09:39 Stygia: also your patches doesn't seem to be final version, you add recipe and immediately remove it in next patch Sep 03 15:10:26 JaMa, Hmm? I don't remove them AFAIK? Sep 03 15:10:58 d .../perl-module-test-harness_3.28.bb | 18 ++++++++++++++++++ Sep 03 15:10:58 .../perl-module-test-more_0.98.bb | 19 +++++++++++++++++++ Sep 03 15:11:04 .../perl-module-test-more_0.98.bb | 19 ------------------- Sep 03 15:11:04 .../perl-module-test-simple_0.98.bb | 19 +++++++++++++++++++ Sep 03 15:11:22 rename meta-perl/recipes-perl/{perl-module-test-simple/perl-module-test-simple_0.98.bb => libtest-simple-perl/libtest-simple-perl_0.98.bb} (56%) Sep 03 15:11:33 JaMa, God dammit. I deleted the repo and cloned it again to avoid those old patches that I never pushed. Sep 03 15:11:40 JaMa, How the hell did they still get in there... Sep 03 15:11:46 rename meta-perl/recipes-perl/{perl-module-test-harness/perl-module-test-harness_3.28.bb => libtest-harness-perl/libtest-harness-perl_3.28.bb} (96%) Sep 03 15:12:56 JaMa, Weird... Well, fuck. Those weren't supposed to be part of this at all. Sep 03 15:13:12 Stygia: I'd honestly suggest getting familiar with git rebase -i origin/master Sep 03 15:13:14 well it's what you sent to ML Sep 03 15:13:26 bluelightning, I probably should yea... Sep 03 15:13:42 Stygia: if you check out a new branch beforehand you can play around with it and not risk losing anything Sep 03 15:13:50 JaMa, Hmm so it seem anyway. Sep 03 15:14:10 bluelightning, Hmm right, but I deleted all of my meta-openembedded repo, and I figured that would reset everything I did earlier. Sep 03 15:14:20 Somehow it still knows about the incomplete versions of my recipes and such... Sep 03 15:14:29 Well. That was definitively unintended. Sep 03 15:14:37 Stygia: hmm, I'm not quite sure how that has happened, but in any case deleting isn't necssary Sep 03 15:15:01 bluelightning, Yea I know there's a better way, I just didn't want to bother - I deleted it and cloned it again... so I figured that I had a completely clean setup here. Sep 03 15:15:37 Stygia: I would have thought that would have done it yes, I can only assume some sort of directory mix-up Sep 03 15:16:14 bluelightning, I suppose. I'm 100% I deleted meta-openembeddded and cloned it again. Sep 03 15:16:17 Ah, either way. Sep 03 15:16:24 From now on, I'll use git status to see what I've done before I send-email. Sep 03 15:16:34 Stygia: honestly though "git rebase -i" is so useful it'll be worth the time to learn how to use it - at least when I did my git productivity went up hugely Sep 03 15:16:34 That'll show me if there's some rename or something else unwelcome, yea? Sep 03 15:16:58 bluelightning, Yes I know. Learning to use git properly is on my list of things... unfortunately not at the top. Sep 03 15:17:03 well git status only tells you what branch you're on and what hasn't yet been committed Sep 03 15:17:45 bluelightning, But if it says, "New file: ..." that _should_ mean I can safely commit it assuming the recipe itself is good, yea? Sep 03 15:18:01 Stygia: yes Sep 03 15:18:09 bluelightning, Then it servers my purpose for now. :) Sep 03 15:18:16 bluelightning, But yea. I know. I suck at git and I should learn. Sep 03 15:18:23 Stygia: ok Sep 03 15:18:27 But for now I'd rather just push these patches and get back to work. :) Sep 03 15:19:00 Stygia: I'd say if you have time for nothing else, definitely play around with git rebase -i :) Sep 03 15:19:16 I'll get down off my soapbox now :) Sep 03 15:19:19 bluelightning, Yup, I'll note that down. Sep 03 15:30:43 also, don't just add [meta-perl] or whatever, keep the PATCH bit too, please :) Sep 03 15:34:54 kergoth, Ah, so... [meta-perl][PATCH] should be the prefix? Sep 03 15:35:17 Ah shit, I missed that in bluelightning's message to me originally... dammit. Sep 03 15:35:30 indeed, yes. --subject-prefix='meta-perl][PATCH' or so Sep 03 15:35:36 not a problem, just a reminder for the future Sep 03 15:35:44 * kergoth uses the [PATCH bit in his filters too :) Sep 03 15:35:47 kergoth, Alright... Well now it's in my bash history, heh. Sep 03 15:36:11 kergoth, That's how I usually remember long commands. Sep 03 15:38:07 heh, a friend of mine set up a shell prompt command to email him new shell history entries in blocks, so even bits that are outside the maximum history size get archived, for that reason. seemed a bit overly complex to me, but interesting nonetheless Sep 03 15:39:31 Completely off topic. Isn't there like a sequence to store your own work then rebase and readd your work? Like git stash, git rebase ... ?? Sep 03 15:40:36 if you have just local changes, thats what stash is for. if you have multuiple commits, you don't need a stash, you can put it on a branch Sep 03 15:42:22 yeah, but someone said it was useful for collecting multiple changes in one commit, can't remember exactly how the workflow went Sep 03 15:42:31 into one* Sep 03 15:43:14 yes, rebase -i lets you rearrange, merge, etc commits Sep 03 15:43:22 but there's no need for stash for that, rebase alone does htat job just fine Sep 03 15:43:45 stash is just for temporarily setting aside local, non-checked-in changes, usually when switching branches or updating your local branch from upstream before checking in Sep 03 15:44:27 very good explanation, thanks ! Sep 03 15:45:04 looks like git stash pop is the reversal Sep 03 15:45:23 indeed Sep 03 15:47:37 "git stash list" and "git stash show -p" become useful as you start to rely on stashing more and more :) Sep 03 15:51:10 As I interprete kergoth I'm a bad boy that don't commit. I should really find a better workflow instead of relying on stash. Sep 03 16:03:28 It is very tempting to misuse stash and base a local workflow off it. Reading the examples almost encourages misuse due to limited knowledge of other ways: https://www.kernel.org/pub/software/scm/git/docs/git-stash.html Sep 03 16:05:50 the git docs are only about 50% useful IMO :/ Sep 03 16:08:06 JaMa: are you the ad interim maintainer of meta-filesystems? Sep 03 16:09:36 honestly I don't know :) I was waiting for Hongxu to ack that patch before merging it (the same for meta-perl patches) Sep 03 16:16:20 ant_work: ping with CC to Hongxu I would suggest Sep 03 16:17:25 see, koen appears as maintainer Sep 03 16:17:33 well, of the recipe Sep 03 16:17:54 I suppose he can remove it Sep 03 16:25:00 bluelightning: btw I have small fix for mariadb transition (PROVIDES in mariadb-native) Sep 03 16:25:54 bluelightning: and there is small problem in incremental builds, because mysql5 isn't cleared from sysroot before mariadb is built, so all header files are still from mysql5 not mariadb Sep 03 16:26:25 not only header files all files with the same name and path Sep 03 16:26:26 JaMa: I'm ok with the former, feel free to apply that Sep 03 16:26:33 JaMa: is there any way to actually fix the latter? Sep 03 16:27:10 I can only think of implementing my suggestions from https://bugzilla.yoctoproject.org/show_bug.cgi?id=4102 Sep 03 16:27:28 alias REPLACES variables Sep 03 16:32:40 JaMa: right, unfortunately we don't have that mechanism at the moment though Sep 03 16:33:53 JaMa: that bug is marked RESOLVED FIXED, is that really correct? Sep 03 16:34:17 or do we need a separate feature request? Sep 03 16:35:08 bluelightning: well I'm not sure if we want to bump OELAYOUT_ABI for every recipe rename or package_arch change Sep 03 16:35:36 er, no, that would be excessive Sep 03 16:35:45 with more layers I see changes like this quite often Sep 03 16:36:10 usually it's easy for me to spot it and manually clean what's needed to be cleaned Sep 03 16:36:30 but for new commers it just means that incremental builds aren't reliable Sep 03 16:40:07 JaMa: I know; FWIW I would like to see a mechanism present that would help solve this at least where it's practical Sep 03 16:40:25 agreed Sep 03 16:40:41 something we can look at in 1.6 Sep 03 16:40:53 ant_work: did you forgot to Cc Hongxu in last oe-devel e-mail? Sep 03 16:41:34 I really didn't see his e-mail as maintainer Sep 03 16:42:31 JaMa: CCs get dropped on mails sent out by the list server, so not necessarily Sep 03 16:42:53 JaMa: no, cc: Hongxu Jia Sep 03 16:43:08 bluelightning_: yes Sep 03 16:43:51 aha Sep 03 16:45:50 it's confusing, but I suspect there's no way for the list server to show CC on the email it sends out or it would CC the person again :) Sep 03 16:46:07 true Sep 03 16:46:28 it's funny I've been using mailing lists for years but I only noticed that recently Sep 03 16:47:20 I was even thinking that I've seen Cc: headers there, but that was probably in e-mails sent by me :) Sep 03 16:49:14 if the CC is the mailing list itself it'll show up, it seems Sep 03 16:51:25 how do I create a snapshot of all sources used for building a specific image? Need this for GPL compliance reasons. Sep 03 16:54:46 I think you need archiver.bbclass, but I am not certain how to use it Sep 03 16:55:24 there's a section of the manual on that Sep 03 16:55:53 the Yocto Project dev manual that is Sep 03 16:55:54 http://www.yoctoproject.org/docs/current/dev-manual/dev-manual.html#maintaining-open-source-license-compliance-during-your-products-lifecycle Sep 03 16:55:57 hno: ^ Sep 03 16:58:03 bbl Sep 03 17:09:12 i wonder what he's like at home... Sep 03 17:15:53 bluelightning, thanks, Just found the same via another user hinting about archiver.bbclass Sep 03 18:24:19 heh looks like axis is using OE too :) Sep 03 18:47:09 axis? Sep 03 18:47:14 the webcam people? Sep 03 18:47:38 JaMa: unbelievable! Sep 03 18:48:50 JaMa: how did you know that? Sep 03 18:49:34 got e-mail from their buildbot :) Sep 03 18:49:58 eren: ^ Sep 03 18:52:23 JaMa: BTW, I suppose you will be at ELCE this year, don't you? Sep 03 18:52:23 :) Sep 03 18:52:52 yes, hotel and flight is already booked Sep 03 18:53:08 only ELCE registration is remaining :) Sep 03 18:53:20 JaMa: but aren't you in Edi? Sep 03 18:53:42 florian: xserver-common ping Sep 03 18:53:45 JaMa: damn! He's XorA Sep 03 18:55:17 unfortunately Yocto day has been added the day before elce, so I had to book an hotel more for the 22 only Sep 03 18:56:30 re Sep 03 18:56:35 JaMa: thx Sep 03 19:03:15 anyone know much about using external sourcery toolchains? Sep 03 19:04:07 should external-sourcery-toolahin work in dylan or is there a better way Sep 03 19:04:14 * Crofton|work hopes kergoth is awake :) Sep 03 19:18:29 can anyone explain why/when you would use meta-sourcery over external-sourcery-toolchain? Sep 03 19:18:57 meta-sourcery includes better support for tunes, multilibs and other tidbits.. Sep 03 19:18:59 the one in oe-core isn't really maintained right now Sep 03 19:19:03 I really think the external-sourcery should be removed Sep 03 19:19:14 ah Sep 03 19:19:25 that reminds me, i need to finish reviewing seebs's multilibs pull request Sep 03 19:19:32 A friend of mine is using the one in oe-core Sep 03 19:19:40 I've pointed him a meta-sourcery Sep 03 19:19:46 should it work with Dylan? Sep 03 19:20:56 yep, should. Sep 03 19:21:50 thanks kergoth Sep 03 19:22:06 I also suggested they can use the issue tracker there Sep 03 22:18:45 jama: haha... sorry about the axis buildbot mails... a bit embarassing. i disabled all mail notifications for the time being. Sep 03 22:22:59 zibri: no problem at all, at least we know another company using OE :) Sep 03 22:23:08 jama: :) Sep 03 22:23:30 and I like the name of that layer :) Sep 03 22:23:31 (it was all a trick to promote axis' use of oe!) Sep 03 22:24:17 jama: heh, our line manager thought poky was called porky during a meeting, and the name stuck :) Sep 03 22:24:39 :) Sep 03 22:25:10 reminds me manager in previous job who called meeting to decide whether to use SuSe or ReadHead Sep 03 22:25:58 :D Sep 03 22:26:16 until now I don't know if it was honest mistake or funny nickname for less favourite option :) Sep 03 22:37:55 JaMa, another discovery...klcc-cross is machine specific as it is. Somehow rebuilds in sysroot so I could not see failures rebuilding armv5te after qemuarm Sep 03 22:38:16 jeez Sep 03 23:00:09 zibri, welcome aboard Sep 03 23:14:09 otavio, thanks for the PR http://www.ossystems.com.br/blog/2013/09/01/o-s-systems-at-esc-brazil-2013.html **** ENDING LOGGING AT Wed Sep 04 02:59:58 2013