**** BEGIN LOGGING AT Tue Jun 10 02:59:59 2014 Jun 10 07:04:16 ndec: getting this error : Can't install packagegroup-core-full-cmdline-1.0-r6@x86_generic_64: conflicted package packagegroup-core-basic-1.0-r6@x86_generic_64 is locked Jun 10 07:04:31 ndec: how to clean this? Jun 10 07:33:44 lsb_tester: hmm. not sure why you direct this to me, actually. Jun 10 08:09:55 good morning Jun 10 08:40:35 morning all Jun 10 08:43:42 morning Jun 10 08:44:12 hi all: i tried to build target gcc for armv7ab big-endian rootfs, but when compile i have to pass -Wl,--be8 to make the output work, if not output will show undefined instruction Jun 10 08:44:36 does any one know how to tell gcc add -Wl,--be8 by default ? Jun 10 08:49:48 ducuy, LDFLAGS=--be8 Jun 10 08:51:00 mago_: did you mean pass LDFLAG=--be8 when building gcc ? Jun 10 08:53:28 if your question is how to set linker arguments to GCC, you may use LDFLAGS environment variable Jun 10 08:54:00 it does not matter if what you are building is GCC or any other application, assuming you are using GCC to build it Jun 10 08:55:53 thanks mago_: i want to build native gcc that work for armv7 big-endian root file system, but now i see my gcc was not configured correctly Jun 10 08:55:57 I think you'd probably want TARGET_LDFLAGS += "--be8" Jun 10 08:56:24 it require to add -Wl,--be8 when i build hello.c to make it work Jun 10 08:56:41 root@mustangv7be:~# gcc -o hello hello.c root@mustangv7be:~# ./hello hello[850]: undefined instruction: pc=00000000000082f8 Code: e5bcf2d4 e3a0b000 e3a0e000 e49d1004 (e1a0200d) Illegal instruction Jun 10 08:56:57 root@mustangv7be:~# gcc -Wl,--be8 -o hello hello.c root@mustangv7be:~# ./hello hello root@mustangv7be:~# Jun 10 08:57:16 ducuy, oh you want to set the default LDFLAGS of your cross-compiler? then i misunderstood you. then what bluelightning says is the right way to go Jun 10 08:57:58 thanks, i make a try now Jun 10 08:58:00 TARGET_LDFLAGS is passed through to LDFLAGS only when cross-compiling for the target; I'm assuming that's what you want Jun 10 09:03:04 hi. i'm a bit confused. i've been away from yocto for a couple of weeks and now i'm having problem with a config file Jun 10 09:03:43 i set up the environment like this: TEMPLATECONFIG=meta-my-distro/conf source ./oe-init-build-env build Jun 10 09:04:10 in meta-my-distro/conf there is a file named "local.conf.sample" Jun 10 09:04:31 but oe-init-build-env complains that it does not find any local.conf ... Jun 10 09:05:10 should it be called local.conf rather than local.conf.sample now, or have i mixed things up? Jun 10 09:08:20 ionte: oe-int-build-env just said no local.conf availabe in build/conf directory, and it will create one, so should not any problem here Jun 10 09:08:32 i see Jun 10 09:08:57 but my local.conf.sample was not used anyway Jun 10 09:09:22 but that was because i've switched shell it seems! using zsh now Jun 10 09:09:42 so i need to "export TEMPLATECONFIG=..." first :( Jun 10 09:10:00 https://github.com/gumstix/Gumstix-YoctoProject-Repo/issues/5 Jun 10 09:14:45 Hi, I'm having trouble to run Qt5 applications on the AM335x, am I in the right channel? Jun 10 09:34:48 Hi all, i already tried TARGET_LDFLAGS += " -Wl,--be8" for gcc, but it does not help, Jun 10 09:35:47 that seem ask cross-compile to build target gcc as be8 format Jun 10 09:36:15 but what i want to ask target gcc, to build hello.c to be8 format Jun 10 09:38:02 so a little bit confuse here Jun 10 09:41:19 hi, can anyone point me to any links or documentation on how to use psplash bootscreen utility? Jun 10 09:45:01 i have a bbappend file that is not used with the current configuration. the build system complains that no such recipe is found (and there is none) Jun 10 09:45:16 is there any way to make a bbappend "optional"? Jun 10 09:45:34 if recipe is found, apply bbappend. if not found, ignore bbappend. Jun 10 09:46:56 i know about BB_DANGLINGAPPENDS_WARNONLY, but that is a global build setting Jun 10 09:47:10 i would like to treat just this .bbappend special Jun 10 10:36:42 blloyd: there are 4 versions for your geode upgrade, I'm taking latest to master-next, but please use PATCHvN to make it clear Jun 10 10:37:24 blloyd: also your full name is now in SOB line, but not in Author field, please fix it for next time (I'll change it manually for this geode upgrade) Jun 10 11:31:39 In general, how do people use Yocto in a Windows environment ? My current idea is useYocto for generating an SDK on a Linux host, and use VMWare++SDK Jun 10 11:37:19 kroon: that's the way to do it Jun 10 11:37:43 kroon: but be prepared for long compile times Jun 10 11:38:16 ionte, ok.. well hopefully we will switch to linux hosts on our dev. machines, but not for a while .. Jun 10 11:38:39 I want to try meta-mingw, and create at least the toolchain for native win32 Jun 10 11:40:43 RP: more info about grub do_package issue sent to ML (reply to last "status" e-mail) Jun 10 11:48:55 JaMa: thanks Jun 10 12:01:26 kroon, i tried the meta-mingw thing a few weeks ago. i had problems generating final tarball of the sdk, had to hack it to get it working Jun 10 12:02:32 and then i moved it to a windows machine, and wasn't entirely sure on how to use it. i think you have to bundle it with a cygwin and python to get it working Jun 10 12:02:56 it'd be nice if yocto could generate a complete windows installer for a native SDK Jun 10 12:46:55 Hi! Could someone advice me how to build yocto with softfp abi? Jun 10 13:04:52 thank you JaMa for the feedback. Jun 10 13:22:59 is there a config setting that allows setting the author for a patch? Jun 10 13:23:43 blloyd: —author to git commit, if its a one-time thing Jun 10 13:23:50 and how long are things typically on master-next before getting placed in master? Jun 10 13:23:53 stunpix: try posting your question to mailing list Jun 10 13:24:07 blloyd: "when they're ready" Jun 10 13:30:44 besides trying to make sure patch is solid before first commit (already done), is there much that I can do to keep this process moving forward? I don't want to submit a patch for diasy until master is done. Jun 10 13:54:05 mago_, ok, good to know. will try it out Jun 10 14:35:14 why would you ever get merge conflicts in a clean git repository? Tried a git pull and get conflicts (meta-oe). No local commits in that repository. In fact, I repulled it from scratch yesterday because my last copy did the same thing. Jun 10 14:36:01 blloyd: did you have any commits at all that weren't part of the upstream branch? Jun 10 14:36:32 not in that directory Jun 10 14:36:44 and no unstaged commits Jun 10 14:37:16 and it's conflicting on vim, a file I haven't had any reason to modify myself. Jun 10 14:37:19 hmm... well, on the face of it that should not happen Jun 10 14:37:33 I have to admit I did have had it happen a few times a long time ago, but not recently Jun 10 14:37:37 ok, so my understanding is not totally off. Just my experience. Jun 10 14:37:55 I had assumed it was a bug in an older version of git Jun 10 14:39:20 at least meta-oe is not that big Jun 10 14:42:35 * JaMa wish we have gcc-native Jun 10 14:43:08 blloyd: aren't you tracking master-next? Jun 10 14:48:35 JaMa: I was switching my master folder to master-next when the problem above occurred. Jun 10 14:49:08 master-next is often rebased, so it doesn't have linear history and git pull is expected to produce conflicts Jun 10 14:49:17 oh Jun 10 14:50:28 so how do you track it without redownloading the entire repository every few days? Jun 10 14:53:42 best advice is not to track master-next Jun 10 14:53:53 its literally WIP and stuff can and does disappear from it, because it was broken Jun 10 14:54:03 doesn't happen often but it has happened Jun 10 14:54:18 at least master doesn't get rebased underneath you Jun 10 14:54:38 also if something does get rebased that you were working on, just rebase your own branch Jun 10 14:54:56 ok, so how do you deal with your patches that have been put in master-next? Jun 10 14:55:09 and aren't in master yet? Jun 10 14:55:16 YPTM: Participant passcode: 42001078 Dial-in number: 1.972.995.7777 US Toll Free number: 1.877.561.6828 Jun 10 14:56:21 Stephen Joined YPTM Jun 10 14:56:28 YPTM: armin joined Jun 10 14:57:16 YPTM: Saul is here Jun 10 14:57:57 YPTM: Bruce is on the call. Jun 10 15:00:31 YPTM: hi all Jun 10 15:00:39 yup Jun 10 15:00:42 YPTM: tom z on Jun 10 15:00:53 YPTM: Participant passcode: 42001078 Dial-in number: 1.972.995.7777 US Toll Free number: 1.877.561.6828 Jun 10 15:01:12 * Jefro is on YPTM Jun 10 15:02:16 YPTM: nitin joined the call Jun 10 15:02:29 YPTM: sean joins the call Jun 10 15:02:30 David C is on Jun 10 15:02:48 YPTM: Jeff Polk is on Jun 10 15:05:33 YPTM: Matthew on call a bit late Jun 10 15:05:43 * zeddii listens to the crickets Jun 10 15:06:42 YPTM Michael here, Jun 10 15:09:13 halstead: webhook update? Jun 10 15:09:23 YPTM: ross is lurking here but not on the call, got another meeting Jun 10 15:10:44 darknighte, sure. I'm still working on them. All of the webhook servers I've looked at seem to want to manage all the repos in a directory. Jun 10 15:11:21 YPTM is over Jun 10 15:11:23 Thanks! Jun 10 15:11:51 halstead: thx Jun 10 15:12:25 darknighte, I'm testing out the two that have been recommended to see which changes might be possible to keep our systems working as expected. Jun 10 15:13:14 halstead: sounds good, i'll look for an update. Jun 10 15:13:30 Thanks darknighte. Jun 10 15:27:35 if you want to update meeting notes: Brian Lloyd was the person that couldn't be identified on the call earlier Jun 10 15:29:37 blloyd: as rburton said, just rebase your local branch on master-next Jun 10 15:29:47 conflicts are expected and you should be able to resolve them Jun 10 15:30:06 there is no need to re-download anything Jun 10 15:32:27 or just git reset —hard origin/master-next, if you don't have local-only commits to worry about Jun 10 15:32:28 * kergoth yawns Jun 10 15:42:58 In a comment on suggestability, I *read* that kergoth yawned and then yawned myself Jun 10 15:43:12 * darknighte yawns again Jun 10 15:55:57 * JaMa yawns to support group spirit Jun 10 16:04:50 kergoth: you don't by any chance have any OE metadata history prior to the first OE-Classic commit, do you? Jun 10 16:05:48 I don't *think* so. I do have the bkbits export to CVS which we broke up and imported to construct the early history for the bitbake and oe repositories, but i think there might have been more in early sf cvs before what got pulled into bk, I'm not certain offhand Jun 10 16:10:10 kergoth: if you could have a look around that would be great... I'm trying to trace the history of some files and at the moment it begins at the first import commit in OE-Classic Jun 10 16:11:42 we should be able to check the ibot logs of the original irc discussions from back when the repositories were created, and see any pasted cvs urls to confirm that we did have it on sf before hh and before bkbits. my memory isn't 100% clear on it, just due to how long ago it was now :) i'll check that, and also confirm that the bkbits export doesn't go back any farther Jun 10 16:41:50 bluelightning, why do you need to go that far back? Jun 10 16:43:04 Crofton: I'm assembling a proper git repository out of the patch series we have for Qt/E 2.x, and I wanted ideally to have author + date + commit message for each one if at all possible Jun 10 16:43:13 ah Jun 10 16:55:57 some of the early patchsets might well have come from the oz buildroot before they went into the oe repo. worth checking, at least Jun 10 17:02:41 kergoth: yep, koen pointed me there as well; the patches there don't match up with the ones that go back to the first import though so there still appears to be a gap Jun 10 17:02:48 fyi, http://community.validation.linaro.org/dashboard/image-charts/OpenEmbedded-BootTests Jun 10 17:02:48 ah Jun 10 17:02:57 akbennett, has some ptest questions also Jun 10 17:03:23 bluelightning, interesting piece of archaeology :) Jun 10 17:03:26 FWIW though I did do a git import of the buildroot-oz repository which I'll probably publish for posterity Jun 10 17:03:42 pidge, http://community.validation.linaro.org/dashboard/image-charts/OpenEmbedded-BootTests Jun 10 17:04:21 akbennett, has some questions about ptest and the auto builder (or advisory board, but that make sno sense ) Jun 10 17:04:31 and that page loads slowly Jun 10 17:05:43 software archeology is always a pain in the ass. it's amazing how much early stuff is just nowhere to be found online. e.g. good luck finding a version of sysvinit earlier than 2.0 Jun 10 17:06:21 sil Jun 10 17:06:29 ya.. which makes license stuff sometimes a pain, because people thought "took this code from foo".. but foo is long gone Jun 10 17:06:36 yeah Jun 10 17:07:24 the gaps are interesting, too. automake 1.0 and 1.2 are easily found, but not so automake 1.1 Jun 10 17:07:50 GNU yanked various things that had precompiled bits w/o corresponding sources.. Jun 10 17:08:04 but they usually replaced them with an 'a' version.. i.e. 1.1a Jun 10 17:08:06 that's a good point Jun 10 17:09:03 one useful resource that I at least tended to forget about was newsgroups. you can still dig up old text encoded posts of ancient releases of software there Jun 10 17:09:30 up.. and what I've found is there are a lot of old repositories still available that are tagged right Jun 10 17:09:34 pidge, Crofton, page load times will be addressed soon, may require some re-architecting this 'beta' feature Jun 10 17:10:06 but I have (intentionally) forgotten how to use CVS, so it's sometimes challenging.. ) Jun 10 17:12:10 fray: my first step there is cvs2git ;) Jun 10 17:12:13 otavio: you around? I have a FSL-Arm question for you Jun 10 17:12:30 akbennett: sure, question away. Jun 10 17:12:37 Crofton, yeah, the question around ptests is auto builder, essentially I seem to find ptests included in the auto builder artifacts every once in a while, not enough to predict Jun 10 17:12:43 Crofton: that page isn't loading. Jun 10 17:12:43 fray: preceded by cvssuck if you can't get the real repo Jun 10 17:12:56 pidge, be patient :) Jun 10 17:13:05 pidge, ^^ wondering Jun 10 17:13:09 Crofton: waa, want it now ;) Jun 10 17:13:17 ditto Jun 10 17:13:20 patience is a virtue Jun 10 17:13:56 akbennett: where are you finding them? link? my ab instance or a different one? Jun 10 17:15:17 the answer should be in here http://community.validation.linaro.org/scheduler/job/13638 Jun 10 17:15:49 http://autobuilder.yoctoproject.org/pub/nightly/CURRENT/machines/qemu/qemux86-64/* Jun 10 17:16:05 bbl Jun 10 17:16:27 I think there used to be a perl-ptest, and I've seen a few others come and go Jun 10 17:18:59 huh, http://invisible-island.net/ncurses/ncurses-license.html is rather interesting Jun 10 17:21:19 akbennett: give me a bit to look at this. just to be clear, you're saying that ptests are installed on the images? Jun 10 17:34:34 pidge, to tell you the truth, I don't know, I've seen them in the past, but haven't seen them for awhile. The end goal is we would like a build that has all ptests installed so we can run all ptests each day across the boards Jun 10 17:36:37 akbennett: understood. if you could open a bug on this, I'll take a look at it (but not until ~next week) Jun 10 18:17:07 Jefro, we should capture a summary of the why use "Yocto" thread for advocacy purposes Jun 10 18:19:07 Crofton I have it marked, good discussion Jun 10 18:19:27 yeah, I've never read the Ubuntu legaleses Jun 10 18:19:39 we should also check on the freescale dev day Jun 10 20:38:33 What is considered the current most recent released version of Yocto? Is 1.6 complete? Jun 10 20:42:38 I guess the tag "yocto-1.6" is solid? Jun 10 20:48:18 1.6 is released Jun 10 20:49:34 skfax: we are working on a 1.6.1 that address so of the new CVE issues btw! Jun 10 20:51:05 Are tags such as "yocto-1.6" fixed, such that I can trust them not to be changed to another commit? Jun 10 22:04:55 hmm Jun 10 22:05:49 If I have a "recipes-bps/u-boot" directory in two meta-layers; are they both used? Jun 10 22:06:01 recipes-bsp* Jun 10 22:07:10 wonder if the historical OE.mtn is floating around anywhere anymore, for posterity Jun 10 22:07:28 skfax: that depends on what's in there. Jun 10 22:10:10 kergoth: It's http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-bsp/u-boot?id=yocto-1.6 and https://github.com/netmodule/meta-netmodule/tree/master/recipes-bsp/u-boot Jun 10 22:10:51 i'm assuming oyu just linked the wrong branch and aren't actually trying to use master from one layer and yocto-1.6 from another, because that's a recipe for failure Jun 10 22:11:06 but yes, both will be available Jun 10 22:11:25 The linking was intentional, but it's probably misguided Jun 10 22:11:47 you can't mismatch your branches Jun 10 22:12:00 if you want to use the release, then you'll need to use the stable branch corresponding to that release in your other layers (daisy) Jun 10 22:12:14 But meta-netmodule doesn't offer any branches or tags, and is intended to be used with 1.3 Jun 10 22:12:23 good luck Jun 10 22:12:25 :) Jun 10 22:12:45 It's not a huge meta-layer, so I was able to fix some other build issues with minor tweaks Jun 10 22:12:47 that is, its doable, but might require work Jun 10 22:12:49 yeah, true Jun 10 22:12:54 small amount of metadata to deal with Jun 10 22:13:27 if two layers have recipes with the same filename, only the one from the higher priority layer will be used Jun 10 22:13:41 but if there are recipes iwth different versions/filenames, then they'll all be available Jun 10 22:14:10 Well the directories have the same names, but they contain bb-files etc with different names Jun 10 22:14:10 thats over-simplifying it, of course, but that's basically the idea Jun 10 22:14:20 bitbake doesn't care about directories Jun 10 22:14:20 So those would qualify as different recipes? Jun 10 22:14:23 ok Jun 10 22:14:50 angstrom does builds with like 15 bsp layers all included in the config at once Jun 10 22:14:54 so yeah, its fine :) Jun 10 22:18:35 well, i fixed the current issue in a very naive fashion Jun 10 22:18:40 on to the next one :P Jun 10 22:19:20 another option is to downgrade my tar to 1.26 and use 1.3 branches for everything Jun 10 22:30:00 what is a good way to identify if an network interface link is up/down. /etc/network/if-up.d does not work if the cable is plugged/unplugged Jun 10 22:46:03 woot. stuff is building Jun 10 23:25:25 Is the directory structure of having a "files" folder which file:// refers to a new thing? **** ENDING LOGGING AT Wed Jun 11 02:59:59 2014