**** BEGIN LOGGING AT Tue Oct 29 02:59:58 2013 Oct 29 05:00:02 afternoon all! Oct 29 05:01:07 quick question (hopefully), if i have a xorg.conf file in my layer under recipes-graphics/xorg-server/xserver-xf86-config/, whats the proper way to get that included in my image? Oct 29 05:01:59 ive got a bb append file 1 folder back, shouldnt the xorg.conf file sort of overwrite the default one? Oct 29 06:01:45 nevermind, figured it out i think, was using x11-base, changed to full fat X11 and it picked up my custom xorg Oct 29 06:01:51 .conf Oct 29 08:05:24 sstate mirror logic appears broken. it reads from a different path than it writes. Oct 29 08:31:02 rp: are there any issues with the patch about adding debian-7.2 to sanity tested distros i sent to the poky list a couple of weeks ago? or you just haven't gotten around to it yet? Oct 29 08:36:36 I created a bug about the sstate issue: https://bugzilla.yoctoproject.org/show_bug.cgi?id=5405 Oct 29 08:36:37 Bug 5405: normal, Undecided, ---, richard.purdie, NEW , SSTATE_MIRRORS search path differs from sstate-cache layout Oct 29 09:01:26 hi Oct 29 09:02:14 I'm trying to figure out why varable defined in image.bb (e.g. UNIT_TEST = "1") isn't visible in recipe where creating package Oct 29 09:02:21 do I need some import or something? Oct 29 09:02:31 when defined in local.conf it works fine Oct 29 09:03:06 I need that variable to distinguish between image.bb and imageunittest.bb Oct 29 09:14:55 mbelisko: You cannot affect the recipes by setting image specific flags. An image just gathers the packages. How they are built are determined by the distribution (e.g., distro.conf and local.conf). Oct 29 09:18:37 Saur: I see thanks for reply Oct 29 09:38:11 zibri: not got around to it yet, sorry Oct 29 09:38:17 zibri: conferences etc Oct 29 09:45:46 rp: yeah, i understand. don't worry. i hope you had a nice time at elce :) Oct 29 09:56:25 good morning Oct 29 09:57:25 morning all Oct 29 10:06:08 bluelightning, morning. Oct 29 10:06:18 mckoan, morning. Oct 29 10:06:37 hi Stygia Oct 29 10:06:41 We'd say formiddag, before-noon, but still, that doesn't translate into English, so 'morning'. Oct 29 10:19:07 zibri: it was good but tiring, thanks :) Oct 29 10:33:52 Hi all, Any idea why qemu throws this error after my build is success "spl: error reading image u-boot.img, err - -1" Oct 29 10:39:53 Hello. I'm trying to use Dora, and getting a build error of: "Fixup Perms: Unable to correct directory link, target already exists: /var/run -> /run" Oct 29 10:40:39 On my hardware, I'm getting a very hard lockup when the cdc_acm module loads, so I though I'd try a newer version for that reason. Oct 29 10:42:18 Circuitsoft: is the recipe that's failing installing something to /run ? Oct 29 10:42:33 The recipe in question is samba Oct 29 10:42:55 But I can't find any references to /run or /var/run in the samba recipe. Oct 29 10:43:14 I'm getting samba from meta-openembedded/meta-oe Oct 29 10:44:14 Circuitsoft, try to look in its package-split directory and such, see what it installs. Oct 29 10:44:25 Circuitsoft: are you also using the dora branch of meta-oe? Oct 29 10:44:29 yes Oct 29 10:44:46 And of meta-intel/fri2 Oct 29 10:46:18 Samba, itself, is making a /var/run, but there's nothing in it. Oct 29 10:47:57 * Circuitsoft has been awake for 21 hours, so is now going to get some breakfast. Oct 29 10:58:30 Circuitsoft: this may be relevant: http://permalink.gmane.org/gmane.comp.handhelds.openembedded/58530 Oct 29 12:28:46 bluelightning: That fixed the symlink error, but now it's saying that "/run" and "/run/lock" were installed but not shipped. Oct 29 12:32:50 Circuitsoft: You do not want to create files under /run (or /var/run) directly. You should do it indirectly by writing a configuration file to "${D}${sysconfdir}/default/volatiles/" and/or "${D}${sysconfdir}/tmpfiles.d/". Oct 29 12:35:22 I figured so, but I removed them all in do_install_append, and it still giving that error. Oct 29 12:35:42 Adding them to FILES_${PN}, while probably wrong, does appear to have fixed it, though I'd prefer to not have them there. Oct 29 12:36:06 Oh... I want it in do_install_prepend, I suppose... Oct 29 12:36:34 Circuitsoft: Then my guess is that you are not removing them even though you think are... Oct 29 12:37:01 RP: maybe it is easier to patch sumtool ... Oct 29 12:38:15 Saur: http://paste.ubuntu.com/6323741/ Oct 29 12:41:16 Circuitsoft: No, _append should be correct as you want to do it after the make install. Oct 29 12:41:36 Well, with that file, it still complains about /run and /run/lock being there. Oct 29 12:43:12 But, it does stop complaining about a symlink target existing. Oct 29 12:44:36 Circuitsoft: Do you get any errors in the log.do_install file (e.g., from rmdir failing to remove directories)? Oct 29 12:45:33 Haven't checked, but previously when I didn't have the [ -d file ] tests, it did fail, saying those folders weren't there to remove. Oct 29 12:59:02 Is there a way to make a do_package_qa step not a catastrophic failure? Oct 29 13:17:11 Circuitsoft: you can use INSANE_SKIP, but it is better to just fix the actual problem Oct 29 13:48:27 Right now, I'm trying to do a tech demo, not an actual project, so I think I'll go with INSANE_SKIP. Oct 29 14:13:27 And now, it's failing on a line that says "mv $infile $outfile || true" Oct 29 14:13:42 I wonder how it can fail with the "|| true" in there. Oct 29 14:22:29 So, what do I put in INSANE_SKIP to not get the errors about /run and /var/run? Oct 29 14:28:03 Circuitsoft: http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#ref-classes-insane Oct 29 14:35:25 It's still failing with INSANE_SKIP_${PN} = "${WARN_QA} ${ERROR_QA}" Oct 29 14:37:54 bluelightning: I appear to not fully understand INSANE_SKIP. Oct 29 14:38:16 that is not how you should use it, no Oct 29 14:38:42 INSANE_SKIP_${PN} += "some-check-name" Oct 29 14:38:56 Well, I can't figure out what check to disable, so I'm trying to disable all of them. Oct 29 14:39:10 This is a one-time-use tech demo, not a project, so I really don't care if it's unstable. Oct 29 14:39:11 assuming ${PN} (i.e. the main package for the recipe) is the specific *package* from the recipe that the check fails on Oct 29 14:39:48 log.do_package is not making it obvious where the failure is. Oct 29 14:40:03 perhaps you could pastebin the error you're getting? Oct 29 14:40:16 Well, it starts with DEBUG: Executing python function fixup_perms Oct 29 14:40:24 So, do I put fixup_perms in INSANE_SKIP? Oct 29 14:40:44 Should I pastebin the bitbake output, or log.do_package? Oct 29 14:40:53 the bitbake output Oct 29 14:41:37 http://paste.ubuntu.com/6324304/ Oct 29 14:42:29 hello i'm trying to build 1.5 for a gumstix (overo) i'm reusing the machine/overo.conf that previously worked with 1.4 but i get this error : the PACKAGE_ARCHS variable does not contain TUNE_PKGARCH (armv7a-vfp-neon) Oct 29 14:42:45 what can i do ? :) Oct 29 14:44:59 Circuitsoft: ok, this particular error cannot be skipped using INSANE_SKIP; if you want to bypass it you'll need to modify ERROR_QA from your local.conf Oct 29 14:45:10 Circuitsoft: apologies for leading you up the wrong path Oct 29 14:45:44 tida: I would suggest complaining to the meta-gumstix maintainer Oct 29 14:49:16 less fun :) Oct 29 14:50:04 bluelightning: ERROR = "" in local.conf did nothing. Oct 29 14:50:43 Circuitsoft: try ERROR_QA_forcevariable = "" Oct 29 14:51:02 this is of course using a sledgehammer to crack a nut... Oct 29 14:51:52 and if I catch anyone recommending anyone else to use the above then there'll be trouble ;) Oct 29 14:57:18 Did adding the _forcevariable invalidate my whole sstate-cache? Oct 29 14:59:18 no, changing ERROR_QA might have done that though (it shouldn't ideally, but I think it was recently discovered that that variable was erroneously being included in checksum values) Oct 29 15:00:08 FWIW I am now building samba here to see what is going on Oct 29 15:00:17 and, it succeeded Oct 29 15:00:20 :/ Oct 29 15:00:23 brb Oct 29 15:00:49 YPTM: Matthew Weigel on the phone Oct 29 15:00:53 YPTM: Scott Rifenbark joined the call Oct 29 15:01:09 YPTM: Nitin joined the call Oct 29 15:01:11 YPTM: Björn is on the call Oct 29 15:01:33 YPTM: Polk is on Oct 29 15:01:52 YPTM: Cristiana is on the call Oct 29 15:02:12 YPTM: Belén joined Oct 29 15:02:14 The Yocto Project Tech Meeting Con-Call is starting Oct 29 15:02:14 Dial-in number: 1.972.995.7777 / Participant passcode: 42001078 Oct 29 15:02:14 This call is open to all and the channel remains open to discuss any topic Oct 29 15:02:32 YPTM: Paul Eggleton joined Oct 29 15:03:12 YPTM: Saul is on Oct 29 15:03:38 YPTM: ALexG on. don[t hear anything on the bridge Oct 29 15:03:46 Circuitsoft: can you verify you have 82f7bdc63e084e8482be0914d9b46fc10cada185 in your meta-oe checkout? Oct 29 15:03:54 YPTM: welcome to the technical team meeting, please let me know who's on the bridge. Oct 29 15:04:05 YPTM: Tom Z here Oct 29 15:04:16 YPTM: LaurentiuP joined Oct 29 15:04:37 bluelightning: I have 513e7ca20ddd0a5c3b649bf292a67c3e0473d3a8 Oct 29 15:04:43 Song_Liu: you might have skipped over me Oct 29 15:04:55 Song_Liu: np Oct 29 15:05:47 alex d here Oct 29 15:05:53 git branch --contains 82f7bdc63e084e8482be0914d9b46fc10cada185 lists my current branch. Oct 29 15:06:52 Circuitsoft: I'm a bit puzzled as to how you're getting this failure and I'm not then... Oct 29 15:07:17 YPTM: any opens? Oct 29 15:07:50 Test Oct 29 15:08:20 Circuitsoft: if you run "bitbake-layers show-recipes samba" the meta-oe version 3.6.8 one is the one listed first? Oct 29 15:08:51 * fray will be joining the call soon.. Oct 29 15:09:05 YPTM: Bruce Ashfield joined late. Oct 29 15:09:05 3.6.8 is the only one listed. Oct 29 15:09:12 https://www.google.com/calendar/embed?height=600&wkst=1&bgcolor=%23FFFFFF&src=mg0m150m30gs891eqqhtfu5jgg%40group.calendar.google.com&color=%238C500B&src=sc7rov4ck552t2f0pav120t7r0%40group.calendar.google.com&color=%23528800&src=theyoctoproject%40gmail.com&color=%23182C57&ctz=America%2FLos_Angeles Oct 29 15:09:47 https://wiki.yoctoproject.org/wiki/Yocto_1.6_Features Oct 29 15:20:16 * RP suspects everyone is "conferenced out" :) Oct 29 15:20:25 lol.. Oct 29 15:20:34 I'm still on another call.. should have ended 20 minutes ago Oct 29 15:21:08 seems like a problem with the bridge this morning; the first time I dialed in there wasn't any audio Oct 29 15:22:44 anything left for me to call in for? Oct 29 15:25:56 fray: no Oct 29 15:26:11 thanks Oct 29 15:30:38 RP: question about the usage of package naming for packages containing CONFFILES, should they be PN or BPN? Oct 29 15:31:03 BPN... Oct 29 15:31:14 assumign the same conf file could/should be used for both the 32-bit and 64-bit versions Oct 29 15:32:03 fray: that's what I was thinking also, but I can't find anything actually doing that. Oct 29 15:32:16 may very well be a bug nobody has noticed before Oct 29 15:32:42 RP and I where talking about clean-up stuff, this looks like another place that needs some work! Oct 29 15:32:47 yup.. Oct 29 15:32:49 I will file and own this one. Oct 29 15:32:55 verifying conf files would be a big deal.. something I'd like to see.. Oct 29 15:33:08 the other is man page compression (and hell generation for that matter) Oct 29 15:33:18 there is already a bug open on the compression issue.. Oct 29 15:35:59 fray, yes, I am the owner of the man page one also, and honestly right now that's low priority for me. If you have someone to assign it to, that would be great! Oct 29 15:36:09 janitor.. ;) Oct 29 15:36:56 yeah Oct 29 15:37:13 might aswell call me the user space janitor also! Oct 29 15:37:28 -if- people want man pages on the target, that will get attention.. until it does, it's a desire, nothing more Oct 29 15:51:37 When are files in build/tmp/work/my_arch/package/version/package/ made? Oct 29 15:51:53 do_package Oct 29 15:52:24 Which happens after do_install, so there's no way that a file that is removed in do_install_append should be in that folder? Oct 29 15:52:56 It's only in /package/ not in /image/ Oct 29 15:54:02 hmm, that shouldn't be possible Oct 29 15:54:29 I've tried running cleansstate on the package, but it hasn't helped. Oct 29 15:54:30 Circuitsoft: can you try "bitbake -c cleansstate samba" then "bitbake samba" ? Oct 29 15:57:32 I decided to try copying samba and talloc over to my own overlay so I could get rid of the rest of meta-oe to make sure it wasn't causing problems, and it's still the same error. Oct 29 15:58:23 When I reapplied ERROR_QA_forcevariable = "", it again seemed to invalidate the sstate cache and is rebuilding everything. Oct 29 16:04:36 Circuitsoft: we fixed that recently in master... Oct 29 16:08:03 So, switching from Dora to Master may fix this? Oct 29 16:08:26 And, should I do so on all my layers? Oct 29 16:09:02 if you see things in master that are not yet fixed in Dora, be sure to let us know.. (since RP mentioned this, it's likely already known) ;) Oct 29 16:09:45 Circuitsoft: RP is referring to changing ERROR_QA affecting signatures, not the underlying problem you are experiencing Oct 29 16:09:51 At the moment, I just need something working ASAP, and the kernel in Dylan locks my machine up when it loads cdc_Acm. Oct 29 16:10:15 I'm tempted to say "screw it" and put something other than yocto on it for the moment. Oct 29 16:10:26 which arch? Oct 29 16:10:31 fri2 Oct 29 16:10:46 (superset of crownbay) Oct 29 16:10:48 odd, I wouldn't have expected that Oct 29 16:10:53 Circuitsoft: for samba, did you try cleansstate as I suggested above? Oct 29 16:10:57 Yes. Oct 29 16:10:59 No effect. Oct 29 16:11:04 ok Oct 29 16:11:23 I'm running a second build here, this time with MACHINE = "fri2" Oct 29 16:11:45 I tried running "bitbake samba base-files -c cleansstate" in case base-files might have done something with /run Oct 29 16:12:26 I /msg'ed you a link to a Github repo with my outer skeleton. Oct 29 16:16:45 right, I got that Oct 29 16:17:07 I looked at what you have, there is nothing unusual that I can see Oct 29 16:24:09 and again, no error Oct 29 16:24:46 Odd. Oct 29 16:26:27 very Oct 29 16:43:16 Circuitsoft: Where are you at now? I tried reading the scrollback but got a bit lost :/ Oct 29 16:43:31 Circuitsoft: You still have errors since these directories exist yet you're deleting them? Oct 29 16:44:48 Yes. Since what I'm working on is a technology demo for a potential customer rather than an ongoing project, and I don't need any yocto-specific stuff, I'm giving up on it for now. Oct 29 16:45:02 I've used it successfully on other projects, but it's just not working for now. Oct 29 16:46:14 Circuitsoft: :( Seems odd nobody else can even reproduce it... Oct 29 16:46:51 I'm giving it one last try, this time with the repo Circuitsoft sent me on an Ubuntu 12.04 64-bit VM Oct 29 16:47:22 My one idea is that I do have a system-global sstate-cache, but I was under the impression that it's perfectly safe to mix that across different projects and versions of yocto because it looks like each filename in there has a hash of a config. Oct 29 16:47:42 Circuitsoft: are you using SSTATE_MIRRORS to point to that? Oct 29 16:47:44 it is safe Oct 29 16:47:48 Circuitsoft: it is safe... Oct 29 16:48:07 Mostly, I think it speeds up building the compiler, pseudo, and most of the native- utilities that don't really change. Oct 29 16:48:35 Circuitsoft: are you using SSTATE_MIRRORS to point at that? This could be the key part... Oct 29 16:49:19 DL_DIR and SSTATE_DIR Oct 29 16:49:35 It's local, so no need for http Oct 29 16:49:39 Hmm, I see BB_URI_HEADREVS is only used for autorev, not for srcrev set to a ref, or tag=, or branch=. that's unfortunate. I see why, but it means there's no way to supply the mapping for the latter 3 up front to support BB_NO_NETWORK when using them Oct 29 16:53:58 Circuitsoft: So its still failing in do_package with ""Fixup Perms: Unable to correct directory link, target already exists: /var/run -> /run" ? Oct 29 16:54:17 Yes Oct 29 16:54:52 But, at this point I've been awake for over 27 hours, and just want something to work. So far the Arch install is going smoothly, and I think I'll have my demo working within half an hour. Oct 29 16:55:09 Circuitsoft: If you want to quickly hack this, change your do_install_append () to rm -r ${D}/var/ Oct 29 16:55:21 I think I tried that. Oct 29 16:55:52 Circuitsoft: hmm. Where are you putting the do_install_append ? Oct 29 16:56:54 Circuitsoft: if you're looking for a quick hack, you could add the line to /meta-oe/recipes-connectivity/samba/samba.inc which already has an append Oct 29 16:57:29 It does seem to be getting run, because I had a few other removes there that failed until I made them depend on existence tests. Oct 29 16:58:11 Circuitsoft: I'm wondering about ordering, maybe if somehow it was coming after some of the other appends Oct 29 16:58:16 er, before them even Oct 29 16:58:21 Hmm. Oct 29 16:59:56 Circuitsoft: look at the WORKDIR/temp/run.do_install.* file see what it ran Oct 29 17:00:26 My append was the last thing in the do_install in the install.do. file. Oct 29 17:01:07 I have a reasonable idea about how Yocto works, and I really am thoroughly stumped by this. But, again, I think my best bet is to abandon Yocto for now, and revisit it if we actually get this project. Oct 29 17:02:17 Circuitsoft: At points like this, I tend to try and soft reset, I'd try an install append which blows away /var and /run entirely. If you still get build errors then, there is something very very wrong Oct 29 17:02:59 Circuitsoft: Is this DISTRO=poky or is DISTRO unset btw? Oct 29 17:03:32 Not sure. I would have to go look, which I really don't feel like doing at the moment, since I have something else working now. Oct 29 17:03:44 Circuitsoft: ok Oct 29 17:03:48 Thank you all for your help up to this point. Oct 29 17:03:49 * RP stops asking questions Oct 29 17:04:35 This level of support is one among many reasons that I do intend to use Yocto/Poky if we do actually get this customer/project. Oct 29 17:05:21 Circuitsoft: If you do have some time to look into this properly, it would be good to understand what the problem is/was... Oct 29 17:05:38 I may do that tomorrow, after I've had some sleep. Oct 29 17:05:55 Circuitsoft: The worry is there is some lurking bug we could do with fixing, I appreciate you have other issues right now though! Oct 29 17:07:05 RP: I restarted the AB because of the debug issue, no use in building all red! Oct 29 17:28:04 Someone used to use 'wic' ? Oct 29 17:28:17 I'd like to give it a try and see how it goes Oct 29 17:37:25 otavio: i've been eating my own dog food and booting all my test images lately using wic Oct 29 17:37:56 tomz: though I should say other people have had problems booting on their hardware Oct 29 17:38:49 otavio: anyway, i just gave scottr some documentation that he'll be putting somewhere Oct 29 17:39:58 otavio: http://git.yoctoproject.org/cgit/cgit.cgi/poky-contrib/log/?h=tzanussi/wic-doc Oct 29 17:40:43 otavio: there's also one patch outstanding that hasn't been pulled in yet Oct 29 17:41:28 tomz: anyway, wic is a wip at this point and needs a lot of work that will happen in 1.6 Oct 29 17:58:24 tomz: we're intending to use it for a customer image Oct 29 17:58:28 tomz: based on u-boot Oct 29 17:58:41 tomz: mario-goulart is working on it with us Oct 29 18:00:37 Does wic support u-boot as bootloader? Oct 29 18:01:16 mario-goulart: in theory it should be easy to do; we need to find how to write in a specific offset of the image Oct 29 18:03:55 mario-goulart: no, u-boot support would have to be added Oct 29 18:04:28 otavio: right, basically if you look at similar code in partition.py Oct 29 18:05:32 otavio: e.g. the prepare()/prepare_bootimg_pcbios()/prepare_bootimage_efi() in /scripts/lib/mic/kickstart/custom_commands/partition.py Oct 29 18:06:48 otavio: that should give you a start, but I may have to look at some uboot-based images to get a better idea. any examples? Oct 29 18:06:49 tomz: right; we'll look at this Oct 29 18:07:55 otavio: ok, let me know if you have questions and i'll take a look Oct 29 18:08:56 tomz: sure; will take a look on this Oct 29 18:09:34 tomz: do you have the doc generated somewhere? Oct 29 18:10:45 otavio: no, at this point, just the wic-doc link above - scott will be working on getting the content into the official docs, dunno where yet Oct 29 18:11:56 otavio: also, if there's some functionality that you need that isn't mentioned in the doc, let me know and i can work on adding it - right now there's just a subset of what will eventually be needed Oct 29 18:13:19 otavio: what's there essentially is what enable me to get my images working - but there's a lot more e.g. in the kickstart syntax that wasn't needed and therefore skipped over Oct 29 18:13:55 otavio: that will be added in 1.6, but we should start with what people actually need now Oct 29 19:04:14 tomz: are you keeping a branch, against dora, with 'wip' patches? Oct 29 19:05:38 otavio: no, just what's in master, but I should do that - give me a bit and I can do that and point you to it Oct 29 19:06:13 tomz: that'd help; as wic is not 'widely used' it could be fixed in dora I guess Oct 29 19:06:32 mario-goulart: wait for tomz branch so we avoid rebuilds Oct 29 19:06:59 otavio: yeah, definitely, we'll need that branch for 1.5.1, just hadn't gotten to it. Oct 29 19:07:43 ok Oct 29 19:08:17 who will be in charge of dora stable maintenance? Oct 29 19:17:13 Not directly topical, but I am sure a room full of programmers will appreciate it: Oct 29 19:17:14 http://blogs.msdn.com/b/oldnewthing/archive/2013/10/29/10461148.aspx Oct 29 19:25:36 hello all. very new to yocto and have a basic question. i've downloaded and built an image using instructions provided by Digi Embedded Linux. I'd like to "overlay" (?) another meta package on my existing image, probably in a different directory (in an effort to keep things tidy). i haven't the foggiest idea how to go about that. is there a yocto-for-dummies? I've read all about how to create my own recipies and whatnot, b Oct 29 19:27:20 you want to create your own layer.. (overlay was used in older versions of openembedded so we tend to stay away from that term) Oct 29 19:27:44 within the new layer, you need to provide a basic configuration file, and either a whole new version of the recipe you are replacing or a '.bbappend' file that will append actions onto the existing one Oct 29 19:28:03 @fray, thanks. so the intent is to create my own layer that "depends" (?) on both the digi-provided stuff and this other meta package? Oct 29 19:28:36 a layer can depend on other layers.. Oct 29 19:28:43 alright Oct 29 19:28:51 anything you want to override can be done in your layer.. Oct 29 19:29:48 there is a tool in the newer YP releases that will help create a layer for you to start working in.. Oct 29 19:30:08 convention is that when you replace or append to a layer, you should use the same directory structure as the item you are replacing/amending Oct 29 19:33:14 alright. i'm not entirely sure what the last part means, but i'm sure it'll click when i see it Oct 29 19:33:30 tomz: it seems wic is still quite hddimg specific Oct 29 19:33:42 tomz: it asks for it hardcoded Oct 29 19:35:00 otavio: there are currently two kickstart files, directdisk and mkefidisk Oct 29 19:35:20 otavio: they're just kickstart files in the canned-wks dir Oct 29 19:35:58 otavio: available via 'wic image list'. you can define your own and put them in there, or invoke them directly Oct 29 19:36:27 tomz: yes but it looks for 'hddimg' file Oct 29 19:39:21 otavio: it is based on artifacts from a current built image, in the case of the current set, that's what they look for - like i said some things will have to change for a u-boot image Oct 29 19:40:56 tomz: is getenforce really required, even if we are not running selinux? Oct 29 19:41:45 mario-goulart: the patch i referred to that wasn't pulled in fixes that. let me get you the dora branch in a minute Oct 29 19:42:06 tomz: ah, ok. Oct 29 19:42:45 mario-goulart: i did my development on fedora and didn't see that, anyway, the dora branch should contain the fixes for the things picked up in other host distro testing Oct 29 19:43:12 Alright. Thanks. Oct 29 19:44:36 mario-goulart, otavio: ok, here's the wic-dora branch: Oct 29 19:44:39 http://git.yoctoproject.org/cgit/cgit.cgi/poky-contrib/log/?h=tzanussi/wic-dora Oct 29 19:51:02 Thanks tomz. Oct 29 19:51:05 (noob) where is this mythical yocto-layer script? Oct 29 19:51:07 * mario-goulart fetches Oct 29 19:51:29 dustingooding: poky/scripts, IIRC Oct 29 19:52:08 i'm guessing not in the danny branch, then Oct 29 19:54:02 You are right Oct 29 19:57:08 so i'm guessing dora is the recommended branch? Oct 29 20:01:59 Yeah Oct 29 20:56:24 bluelightning: ping Oct 29 20:57:10 bluelightning: un-ping (it was about debug in oe-pkgdata-util and now I see that fix was already merged) Oct 29 20:57:30 JaMa: right Oct 29 20:58:04 sgw_ seems to think it's still broken, I'm just double-checking Oct 29 21:00:29 did someone else receive oe-core ML e-mails from sgw with date around Dec 2011 today? Oct 29 21:01:24 JaMa, yes..I've seen kdrive..must be Halloween Oct 29 21:13:51 JaMa: yeah, somehow those got lost in an unsent queue! Oct 29 21:14:02 bluelightning: yes the AB is still failing Oct 29 21:19:19 sgw_: see my email Oct 29 21:20:44 bluelightning: I looked at ab01 and this is the revision: d69931da48b9cc5082feb46c654b9861f30244d9 Oct 29 21:21:15 sgw_: I looked at what the autobuilder reported as the resolved revision Oct 29 21:22:04 sgw_: it was 76751a03b1475f22499c488395274453096b9c38 Oct 29 21:22:34 bluelightning: I restarted it after I saw the first set of failures, since I had started before RP added your patch, so maybe we have an AB issue? Oct 29 21:23:11 bluelightning: where are you looking, I logged into AB01 and did a git log Oct 29 21:23:23 sgw_: the failure email Oct 29 21:24:13 commit_resolvedhead_poky to be more specific Oct 29 21:26:51 bluelightning: strange, the only thing I can think of is that is picked up the prior build from before I reset. AB weirdness at this point Oct 29 21:27:38 bluelightning: take a look at : http://autobuilder.yoctoproject.org:8011/builders/nightly/builds/300 Oct 29 21:27:49 it uses the correct revision, strange Oct 29 21:28:27 I'm not sure I trust that Oct 29 21:28:57 bluelightning: yeah the Revision on all these builds is d6, but "Got Revision" is 76. that seems to be the problem Oct 29 21:29:37 I chaulk it up to AB strangeness, I will build MUT next Oct 29 21:35:40 bluelightning: did you see the rpm / perl failures? These seem to be new Oct 29 22:21:52 can anyone point me to where x11-base, x11, and x11-sato are sort of defined, ie what those features include? Oct 29 22:36:30 GusBricker: x11-base and x11-sato are packagegroups defined in the packagegroup-core-x11-[base,sato] files located in either recipes-graphics or recipes-sato packagegroups directories Oct 29 22:39:30 cheers sgw_ will check that out now! Oct 29 23:55:46 f7409a9fe83ba2535a43f39ed57cd78242a88557 breaks the install path of some headers of qt-mobility-x11 Oct 29 23:56:14 which is kind weird Oct 29 23:59:37 it's generating different makefiles... here is one example Oct 29 23:59:39 http://pastebin.com/yQzXB4Fn **** ENDING LOGGING AT Wed Oct 30 02:59:58 2013