**** BEGIN LOGGING AT Tue Apr 23 02:59:58 2013 Apr 23 06:47:47 #morning Apr 23 06:56:24 http://pastebin.com/zqk1wLfi - uncatched exception in bitbake? Apr 23 07:26:08 morning all Apr 23 07:33:32 hi silvio Apr 23 07:34:18 hi hrw Apr 23 07:36:42 WARNING: llvm2.8 does not support aarch64 yet Apr 23 07:36:53 bb.warn() on parsing time should be forbidden Apr 23 07:37:24 hrw jama has now taken over llvm Apr 23 07:37:56 JaMa|Zzz: can this bb.warn() be dropped? it is useless Apr 23 07:38:21 JaMa|Zzz: aarch64 for example will be in 3.3 Apr 23 07:38:48 Good morning Apr 23 07:40:04 hi apelete Apr 23 07:41:13 File "/home/hrw/HDD/devel/canonical/aarch64/openembedded/repos/openembedded-core/bitbake/lib/bb/cache.py", line 814, in save_merge Apr 23 07:41:16 extradata, version = p.load() Apr 23 07:41:19 ValueError: could not convert string to int Apr 23 07:41:21 argh Apr 23 07:41:50 uf. removed cache dir and it started Apr 23 07:42:28 Hi woglinde, hrw Apr 23 08:28:27 hi hrw , woglinde , apelete Apr 23 08:29:29 Hi silvio_ Apr 23 08:35:56 hi friends Apr 23 09:41:41 woglinde: ping Apr 23 09:52:01 ant Apr 23 09:52:42 hi, quick question (forgive me, it's naive) Apr 23 09:52:59 Hello all Apr 23 09:53:10 ant I have many of them too Apr 23 09:53:12 I added a new package in a recipe PACKAGES += "${PN}-systemd ${PN}-srv" Apr 23 09:53:26 how does "export ac_cv_func_vfscanf=no" would help if the libc lacks vfscanf ? Apr 23 09:53:50 but when I add FILES_${PN}-srv = "/srv/test.h" as a last line of the recipe this variable is not evaluated Apr 23 09:54:00 ant_work you have to check configure and than the code where it is used Apr 23 09:54:18 ant often configure files do not get updated Apr 23 09:54:28 and than they have a lot of non used stuff Apr 23 09:54:31 but when I move it further up to the recipe FILE_${PN}-srv starts showing in bitbake -e Apr 23 09:54:42 ok, but the func is used in the code Apr 23 09:54:52 (x86 / x86_64 only) Apr 23 09:54:57 is there a restriction that FILES_ var can be the last thing in recipe Apr 23 09:55:07 noor no Apr 23 09:55:22 yes that was my assumtion as well Apr 23 09:55:26 assumption Apr 23 09:55:27 ant what is your problem? Apr 23 09:55:39 but this is happening on danny Apr 23 09:55:55 woglinde: http://www.zytor.com/pipermail/klibc/2013-April/003416.html Apr 23 09:58:21 ant hm the commit is 5 years old Apr 23 09:58:38 heh, I've never cared about x86 before ;) Apr 23 09:59:19 ant the comment you got is valid Apr 23 10:00:16 yes, I can trust hpa :) Apr 23 10:02:10 i.e. in ubi-utils I've just commented out the problematic code Apr 23 10:04:14 like that http://tinyurl.com/cu6lfut Apr 23 10:10:06 morning all Apr 23 10:12:05 hi pb Apr 23 11:17:37 <_rink> otavio, hmm, i built fsl-image-gui on sabrelite imx.6 - it works fine, but the performance is abysmal ... Apr 23 11:18:04 <_rink> everything uses like 100% CPU, even the simplest examples - and opengl in qt looks distorted Apr 23 11:18:07 <_rink> is this known ? Apr 23 11:26:06 _rink: Hello Apr 23 11:26:41 _rink: did you apply the patches we sent to ml yesterday? Apr 23 11:28:10 <_rink> otavio, hi Apr 23 11:28:17 <_rink> didn't see any patches... ? Apr 23 11:28:49 _rink: in ml Apr 23 11:29:47 <_rink> is that the PATCH v2 thing ? Apr 23 11:30:03 _rink: BINGO! lol :) Apr 23 11:30:21 <_rink> hmm guess not - is there a diff that contains them all somewhere? Apr 23 11:31:29 <_rink> they aren't in git due to the freeze ? Apr 23 11:31:55 <_rink> (i only have the web archives, isolating patches from there is a little tedious) Apr 23 11:37:24 otavio: does there exist some tool for uploading boot images over usb to imx6 already? I hacked one for myself and wondered whether I should spent more time into it Apr 23 11:37:42 ensc|w: sure it does Apr 23 11:37:53 ensc|w: Boundary did it Apr 23 11:38:01 ensc|w: and I use it daily :) Apr 23 11:38:10 ensc|w: hold on Apr 23 11:38:33 ensc|w: https://github.com/boundarydevices/imx_usb_loader Apr 23 11:38:47 thx Apr 23 11:44:01 * _rink discovers patches.openembedded.org :) Apr 23 11:44:11 rink lol Apr 23 11:54:58 :) Apr 23 13:12:51 hii Apr 23 13:13:39 any body has any idea how to use bitbake variables in script file. i need to run a script file through bb file and ${WORKDIR} i need to use in shell script Apr 23 13:14:37 ? Apr 23 14:05:19 Sj: not sure i understand, could you maybe explain again? Apr 23 14:06:27 are you trying to install a shell script into the resulting package? or need to use a shell script sometime during the build process? Apr 23 14:24:35 hello all! Apr 23 14:30:23 (hopefully) quick question: we build serveral images based on a core image. all images can be build with several BSPs. one of the BSPs needs to overwrite a task in the images (in recipes-/image/core-image-.bbappend). what's the the best practice to do so without having to write a .bbappend for each image and hence the need to change X files if we need to change the task? Apr 23 14:47:56 mhh...nobody? Apr 23 14:52:36 hsychla: if it's just one BSP that needs to change the task, then just write a bbappend for that one BSP Apr 23 14:53:20 or asked differently. all our images are based on a core image. we try overwriting the task in that core image and have a "inherit in the different images but the overwritten task does not seem to be inherited? Apr 23 14:54:27 waynr: we do that for one image now but that way we would need to write recipes-/image/.bbappend for every image with which we want to use that BSP? or are we doing it wrong? Apr 23 14:55:35 ah, i think i see what you are asking now Apr 23 14:58:50 hsychla: do you mean "inherit" or "include"? Apr 23 14:59:28 inherit right now (I'm quite new to all this) Apr 23 15:00:57 okay i am confused right now...there is a bbclass called "core-image" is that also the name of a .bb file in your recipes-/image/ directory? so for example recipes-core/image/core-image.bb Apr 23 15:02:28 because if what you want is to extend some base image recipe in another image recipe you should "require recipes-core/image/core-image.bb" (in keeping with my previous example) Apr 23 15:04:37 if you write "inherit core-image" you are telling that particular recipe to "inherit" functionality from a bitbake class, in some folder called classes/core-image.bbclass, which is necessary in your base image recipe Apr 23 15:05:55 ah... would that classes/core-image.bbclass be in my layer or in the layer of the core-image? Apr 23 15:06:27 it would be in whatever base layer you are using Apr 23 15:07:04 so for example if you were using poky, the base layer containing this core-image.bbclass would be poky/meta so the full path would be poky/meta/classes/core-image.bbclass Apr 23 15:07:38 I am currently working on core-image-A which is based on core-image. In the layer defining core-image, there is a classes/core-image.bblcass and a recipes-core-image/image/core-image.bb Apr 23 15:08:34 and in my layer I have classes/core-image-A.bbclass and recipes-core-image-A/image/core-image-A.bb Apr 23 15:08:38 but anyway, you only need to "inherit core-image" in your base recipe. so for the layout you just described, recipes-core-image/image/core-image.bb would have to have a line "inherit core-image" somewhere Apr 23 15:08:54 core-image-A would ahve to "require recipes-core-image/image/core-image.bb" Apr 23 15:09:26 oh, i see Apr 23 15:09:40 I see. will try with and require and come back later. thanks a lot for your help so far! Apr 23 15:09:46 then maybe you need to "inherit core-image-A" in core-image-A.bb Apr 23 15:11:04 I do have inherit core-image-A in core-image-A.bb Apr 23 15:11:41 okay...can i ask what you are trying to do with core-image-A.bbclass? Apr 23 15:12:04 if you want to include orther recipe then use include (failure accepted) or require (failure not accepted) Apr 23 15:12:37 have a nice day Apr 23 15:13:16 in core-image-A.bblass I currently only set IMAGE_FEATURES Apr 23 15:13:40 _IMAGE_FEATURES = "package-management ssh-server-dropbear" Apr 23 15:13:58 these can differ from the core-image Apr 23 15:15:47 hsychla: i believe it is more common to do that using local.conf, or even a distro configuration Apr 23 15:16:21 so in the default local.conf there is a variable EXTRA_IMAGE_FEATURES which can be populated with your desired features Apr 23 15:17:23 you could also append to this variable in the image recipe with a line like IMAGE_FEATURES += " package-management" Apr 23 15:19:51 ah.. I forgot to mention that core-image-A is also a distro with a conf/distro/core-image-A.conf Apr 23 15:20:46 or rather conf/distro/A.conf Apr 23 15:37:49 * pb_ stabs openssl Apr 23 15:38:03 this afternoon for some reason I keep getting "../libcrypto: file truncated" all the time. Apr 23 15:38:40 disk out of space? Apr 23 15:39:04 nope, 50G available Apr 23 15:39:25 I think it's probably some sort of PARALLEL_MAKE lossage. Apr 23 15:39:32 * pb_ turns that off Apr 23 15:39:40 oh, and now poxy eglibc is broken again, grr. Apr 23 15:40:28 * pb_ gives in and sets UNWINDING_pn-eglibc = "" Apr 23 15:46:40 hm, anybody know why icu is setting FULL_OPTIMIZATION_arm = "-Os"? That seems rather bogus. Apr 23 15:56:16 pb_: looks like it was added for icu-3.6 in 5c8340 Apr 23 15:56:58 waynr: that commit just moved it from icu_3.6.bb to the .inc file Apr 23 15:57:32 d'oh Apr 23 15:57:37 ah, here we are, it was added in 9b547ecf849b66aa883328fb6b26d9b412a4d2ec Apr 23 15:57:47 "icu: use -Os for arm to make it build with gcc 4.3.1" Apr 23 15:57:53 so, seems it was a workaround for some old gcc issue Apr 23 15:58:52 the gcc bug in question was fixed five years ago so I guess this is no longer needed :-} Apr 23 15:58:59 : ) Apr 23 16:04:08 so, next question, why doesn't eglibc DEPEND on gettext-native? Apr 23 16:04:58 hm really? Apr 23 16:05:32 well, as far as I can tell it doesn't, and I get build failures due to msgfmt not being present. Apr 23 16:05:47 so, I guess the options are: Apr 23 16:05:58 a) there is something broken about my configuration which causes it to need msgfmt when it shouldn't Apr 23 16:06:11 b) there is something broken about my configuration which causes gettext-native not to have been build when it should have been Apr 23 16:06:22 c) everybody else has msgfmt installed in their host environment and is happily using that one Apr 23 16:06:30 I bet c Apr 23 16:06:44 pb_: I've seen the same issue today on jenkins Apr 23 16:07:13 ah right, not just me then Apr 23 16:07:29 ah hm host quilt depends on gettext Apr 23 16:07:35 yes and I've seen it today for first time, so something new is causing that Apr 23 16:07:48 it's been happening for me for a while Apr 23 16:07:58 certainly since last week Apr 23 16:08:08 dpkg: dependency problems prevent removal of gettext: quilt depends on gettext. Apr 23 16:08:08 (my current oe-core snapshot is from april 16) Apr 23 16:08:37 pb_: I had eglibc in sstate-cache from april 10th or something like that Apr 23 16:08:37 and I wonder why quilt needs it Apr 23 17:03:19 hey! yesterday i was asking how to add a module to auto-load (on danny), then there was a netsplit Apr 23 17:03:50 i tried adding 'module_autoload_mymodule = "mymodule"' to the recipe, but the file system did not end up with that anywhere in etc/ Apr 23 17:04:22 did you rebuild kernel? Apr 23 17:04:32 JaMa, did a whole new build Apr 23 17:04:43 but kernel, you say? it's an out of tree module Apr 23 17:05:43 it currently inherits module Apr 23 17:06:46 arago-project.org/git/?p=meta-ti.git;a=blob;f=recipes-ti/dsplink/ti-dsplink.inc;hb=HEAD Apr 23 17:06:53 out of tree modules do not support module_autoload until master/dylan (with kernel-module.bbclass) Apr 23 17:07:04 JaMa, ahh, i see Apr 23 17:07:41 i guess for now i should append it manually on postinst Apr 23 17:08:17 you can backport kernel-module.bbclass, something like https://github.com/openwebos/meta-webos/commit/c5a3676e69ab54d62f29d7bbeb6a4ac517eb1c33 Apr 23 17:09:15 https://github.com/openwebos/meta-webos/commit/c9f91ecf31ebb602a8a86c8e2017ca431e2623e8 is how to use it Apr 23 17:09:31 and s/kernel-module/kernel-module-split/g in all I've said :) Apr 23 17:13:03 JaMa, that looks very cool. thanks for the pointers! Apr 23 17:25:18 waynr: me again ;) Apr 23 17:26:34 I changed my image to require the core image to which the BSP adds the task bundle_files (addtask bundle_files after do_vmdkimg before do_build) but it is still not done in my image... any idea, why? Apr 23 17:29:10 btw. sorry i mixed up overwritung and adding of a task earlier Apr 23 17:38:04 hsychla: in my admittedly limited experience, a "task" function name must take the form "do_" where in your case would be "bundle_files" Apr 23 17:38:29 so you may want to try renaming the function to do_bundle_files Apr 23 17:38:58 but you would still keep the "addtask bundle_files after do_vmdkimg before do_build" Apr 23 17:39:41 the "addtask" syntax expects just the portion of the function name if i remember correctly Apr 23 17:42:15 meh... , was wrong again, sorry. it is called via "python do_bundle_files() { bb.build.exec_func('create_bundle_files', d) }. and it does work if I build the image my image is based on... Apr 23 17:46:25 bitbake -c bundle_files core-image-A Apr 23 17:46:31 Task do_bundle_files does not exist for target core-image-A Apr 23 17:48:54 hsychla: can you pastebin the core-image-A recipe? Apr 23 17:52:13 sure, but it's not much in there: http://pastebin.com/VSQBU7Xh Apr 23 17:53:38 hmm try moving your "inherit" line before the "require" line Apr 23 17:54:00 actually Apr 23 17:54:12 try just inheriting core-image-A in the core-image-A recipe Apr 23 17:54:34 you probably already "inherit core-image" in the core-image.bb recipe Apr 23 17:54:50 inheriting it again could overwrite the task list or something Apr 23 17:55:03 ok, trying a build Apr 23 17:56:43 still does not know the task Apr 23 17:59:36 by the way, i meant for you to remove the "core-image" from the inherit statement instead of swapping the require and inherit lines Apr 23 17:59:54 so your inherit core-imageA should still follow the require line Apr 23 18:00:38 ah ok, I did both Apr 23 18:01:35 still nothing: ERROR: Task do_bundle_files does not exist for target core-image-tposs Apr 23 18:02:16 core-image-tposs == core-image-A, tired of censoring... Apr 23 18:03:27 what does bitbake -c listtasks core-image-tposs show? Apr 23 18:06:03 http://pastebin.com/tP25P7cp Apr 23 18:11:09 quick question, I have a recipe using externalsrc.bbclass Apr 23 18:11:26 and I always want to execute do_compile Apr 23 18:11:31 is that possible ? Apr 23 18:14:56 abelloni: as far as i can tell from reading externalsrc.bbclass do_compile won't be affected by that class Apr 23 18:16:06 but if you want do_compile to run unconditionally every time (even if a previous compile step has been successful, then do_compile[nostamp] = "1" might work Apr 23 18:17:17 hsychla: is the "python do_bundle_files...." code defined in the core-image-tposs.bbclass file? Apr 23 18:18:20 no, it is not Apr 23 18:18:29 where is it defined? Apr 23 18:19:01 in the core-image.bbappend Apr 23 18:19:11 ah right, sorry Apr 23 18:19:37 yeah, basically, the customer has like 40 recipes using externalsrc and they would prefer to detect automatically what sources has changed and what to rebuild, I don't know if there is a way to do that Apr 23 18:20:58 hsychla: what are the layer priorities? Apr 23 18:23:07 waynr: all 3 layers (the base image, my image and the BSP have prio 6) Apr 23 18:23:19 waynr: all 3 layers (the base image, my image and the BSP) have prio 6 Apr 23 18:25:56 try setting the prioritis such that base < my < BSP Apr 23 18:26:47 otherwise I believe that recipes will be parsed in the order they are found in the BBLAYERS variable in the build directory's conf/bblayers.conf Apr 23 18:28:19 maybe that doesn't explain why youre core-image builds with the bbappend but not the core-image-A but it can't hurt to be more explicit with the layer prirorities Apr 23 18:31:01 well, I think the explanation, why it works with the core-image is that the task is appended in recipes-core-image/images/core-image.bbappend Apr 23 18:32:29 yeah i'll admit i have no idea whether an included recipe's bbappend files are accounted for Apr 23 18:36:06 changing the priorities did not help either Apr 23 18:39:41 hsychla: one more thing i would try, just after the "require" line in core-image-A.bb try "include recipes-core-image/image/core-image.bbappend" Apr 23 18:40:04 ok, just a sec Apr 23 18:43:27 waynr: ERROR: ParseError in recipes-core-image/image/core-image.bbappend: not a BitBake file Apr 23 18:51:27 it says the same if I intentionally make a typo. so it seems like it does not find the file Apr 23 19:24:13 waynr: thanks for your help, I have to go (9:30 pm here). I think we will return to writing a recipe for each image for now (or I#ll be back tomorrow ;) ). again, thanks! Apr 23 21:22:08 Is there no recipe for the Adwaita gtk theme ? Apr 23 21:23:17 for gtk3.4 i mean Apr 23 22:08:55 Hey is there any virtual memory/swapping concept in embedded devices? Apr 23 22:26:58 other than swap partitions/files, there's the compressed ram swap thing Apr 23 22:28:39 compcache i think it's called... Apr 23 22:29:12 seems to work pretty well on efikamx Apr 23 22:30:09 well, the compcache thing came up in a google hit Apr 23 22:30:34 the actual kernel module on my efika is called ramzswap Apr 23 22:30:49 ^^ wooo ^^ Apr 23 23:50:11 okay, if i have a tree of directories with small files (that are not machine-specific) does it make sense to put those under recipes/foo/files and install them from there? Apr 23 23:50:39 or should it really be a tarball in SRC_URI ? Apr 23 23:51:28 i guess i can think about it on the way home... **** ENDING LOGGING AT Wed Apr 24 02:59:58 2013