**** BEGIN LOGGING AT Fri Jan 06 02:59:57 2012 Jan 06 03:18:45 nitink: got it, thx Jan 06 06:13:34 hello all Jan 06 06:13:41 I m getting this error while building strace..... QA Issue with strace: No GNU_HASH in the elf binary : Jan 06 06:58:38 I am getting uclibc compilation errors ,uclibc version 0.9.30.2.-35-r1 , http://pastebin.com/V23BfkHM Jan 06 09:15:44 morning all Jan 06 09:21:25 morning all Jan 06 09:26:45 hi RP__ Jan 06 10:45:13 hey guies i have compiled a x11-image using uclibc ,but my task fails at the last task package x11-image-1.0-r0: task do_rootfs: Failed an anyone tell me whts the problem with the last task http://pastebin.com/APxke9Dc Jan 06 10:48:58 erwt: bluez4 and udev are both prodoviding the file lib/udev/hid2hci Jan 06 10:49:15 RP__, so wht to do Jan 06 10:49:36 erwt: stop one of these packages providing that file Jan 06 10:49:47 bluez is basically a bluetooth package right ? Jan 06 10:49:55 erwt: yes Jan 06 10:50:10 RP__, ok so i will remove it Jan 06 13:05:42 RP__: what you have in mind with the udev-cache patch? Jan 06 14:24:09 otavio: I just don't like the number of execs in that patch Jan 06 14:25:08 otavio: for example we could just concat all the files Jan 06 14:25:21 otavio: also use the shell's read function rather than cat Jan 06 14:59:49 RP__, how do i remove the package bluez from compiling Jan 06 15:18:27 RP__: I will try that Jan 06 15:18:38 Is someone having problem building udisks? Jan 06 15:18:42 it fails to build here Jan 06 15:18:45 in two builders Jan 06 16:23:05 anyone know if joshua will be around today? Jan 06 16:26:09 davest: yesterday in the meeting some status was given regarding Michael Halstead and his joining Linux. Is he no longer available now for support? Jan 06 16:46:49 scottrif, I am here to support the Yocto project as much as possible. Jan 06 16:47:22 halstead: Hi - was hoping later to get some help on setting up my linux email from Belize Jan 06 16:48:04 scottrif, I can probably help with that from memory. Jan 06 17:30:57 fray: are you around? Jan 06 17:45:12 bluelightning I am now.. had to get the kid from school Jan 06 17:45:32 fray: ah ok no worries Jan 06 17:46:00 fray: I wanted to ask you about bug #1654, specifically about how to reproduce it Jan 06 17:46:12 http://bugzilla.pokylinux.org/show_bug.cgi?id=1654 Jan 06 17:47:08 looking Jan 06 17:47:58 FYI I have an implementation of your feature request in that bug ready to go as well, just wanted to make sure the issue got addressed if it's reproducible Jan 06 17:48:10 create two layers.. the first meta-one/recipes-one/recipe/recipe_ver.bb Jan 06 17:48:22 the second meta-two/recipe/recipe_ver.bbappend Jan 06 17:48:37 merge the layers and what I saw was the new merged layer contained: Jan 06 17:48:50 merged/recipes-one/recipe/recipe_ver.bb & merged/recipe/recipe_ver.bbappend Jan 06 17:48:59 yep, that's what I tried, but I did not get that Jan 06 17:49:01 the tool should detect this and add the .bbappend to the bb Jan 06 17:49:06 ahh maybe it's been fixed? Jan 06 17:49:51 what I ran was a merge of meta, meta-yocto, meta-x32 and meta-intel.. and I observed that issue Jan 06 17:49:55 I'm not sure... the odd thing is even the suggested failure mode in the comment I made wouldn't happen because the (original) code explicitly does not copy .bbappend files to the output Jan 06 17:49:59 give that a try.. (I jdid it purely as a test) Jan 06 17:50:10 ok will do Jan 06 17:50:13 odd then.. Jan 06 17:50:45 maybe I misclassified the problem -- or it's been fixed as a result of other work Jan 06 17:55:20 fray: nope, with those four layers and the old code I get no .bbappends Jan 06 17:55:37 however I did just notice it's not skipping .git, should probably fix that one :) Jan 06 17:56:14 ahh ya.. good idea.. so within that, there are no duplicate directories? things that exist both in recipes-foo and recipes-bar.. same version? Jan 06 17:56:37 there won't be duplicate recipes, that I can promise Jan 06 17:56:46 dunno then.. maybe something glitched when I tried it Jan 06 17:56:57 if the structure differs between the layers however then you can get a broken output layer Jan 06 17:57:14 an additional patch I put together today shows warnings when it finds that situation however Jan 06 17:57:39 my recollection was the meta-x32 stuff was in a different order then the meta.. so that was where I thought the problem came from.. but I don't really remember anymore.. Jan 06 17:58:44 fray: yeah, it's still like that - it doesn't have recipes-*/, just the recipe dirs Jan 06 17:59:15 ya.. I expect that to be a fairly common situation.. Jan 06 17:59:40 so merging the second/third/fourth into the directory structure of the first is my suggetion.. if there is no equivalent then it goes wherever it is set Jan 06 17:59:55 hmm, without hacks that might be a little tricky Jan 06 18:00:00 i.e. if first/recipe-something/foo/foo_ver.bb is ammended in second/third/etc we use the path of first Jan 06 18:00:18 if second/recipe-something/foo/foo_ver.bb is ammended by third/fourth we use that path.. Jan 06 18:00:25 basically set an order of precidence for the merge Jan 06 18:02:31 my layer assumption is later layers always add, amend or override (replace) earlier layers Jan 06 18:02:38 so thats the order to manage the directories in.. Jan 06 18:04:31 hmm, that should work fine for recipes, but what about the associated files? what if two recipes are in the same dir in one layer and separate dirs in another - how do you automatically handle the associated patches/defconfigs etc.? Jan 06 18:05:03 for any recipe set, I believe those rules will work.. Jan 06 18:05:13 simply due to the conventional naming schemes used today.. Jan 06 18:05:48 for the class, conf, files directories.. I'm not familiar with them being embedded under subdirs.. so any merging using current rules should be fine.. if there is a conflict a warning/error can be noted.. Jan 06 18:05:55 or am I missing a case? Jan 06 18:07:46 fray: no, I'm talking about patches that are referred to by recipes Jan 06 18:08:05 logically they have to move with the recipe Jan 06 18:08:17 are there patches referred outside of the normal conventions.. i.e. is anyone doing something like ../foo/file? or in all cases are they setting the BBPATH (or similar?) Jan 06 18:08:26 yes, absolutely.. they have to move w/ the recipe Jan 06 18:08:46 so if the location of foo_ver.bbappend is "changed", the patches move w/ it to the location of the .bb Jan 06 18:08:53 | make[2]: *** Waiting for unfinished jobs.... Jan 06 18:08:53 | swaprep.c: In function 'ReplyNotSwappd': Jan 06 18:08:53 | swaprep.c:1300:1: warning: function might be candidate for attribute 'noreturn' [-Wsuggest-attribute=noreturn] Jan 06 18:08:53 | make[2]: Leaving directory `/home/bruce/poky/build/tmp/work/ppc603e-poky-linux/xserver-kdrive-1_1.7.99.2-r27/xorg-server-1.7.99.2/dix' Jan 06 18:08:54 | make[1]: *** [all] Error 2 Jan 06 18:09:00 odd .. Jan 06 18:09:07 * zeddii scratches his head and refreshes things. Jan 06 18:09:15 zeddii, there is a patch for that.. Jan 06 18:09:21 on the list ? Jan 06 18:09:25 check the yocto list, I believe Gary THomas sent it yesterday Jan 06 18:09:27 * zeddii hasn't read it yet today. Jan 06 18:09:42 let me find it.. Jan 06 18:09:52 fray: right, the tricky part would be how to programmatically determine which files are referred to by which recipe Jan 06 18:10:00 0001-xserver-kdrive-work-around-issues-with-Werror-addres.patch Jan 06 18:10:08 cool. thanks fray. Jan 06 18:10:22 bluelightning I think it's reasonable to assume all of the files in a directory can be referenced by all of the recipes (or bbappends) within that directory.. Jan 06 18:10:36 I'm just trying to test the new ppc emulation .. and sort of need sato :) Jan 06 18:10:37 if the path of the bbappend and bb (being appended) change.. then you carry all of the associated files with it.. Jan 06 18:10:48 yup.. I hit this on MIPS yesterday testing the prelinker.. Jan 06 18:11:01 Subject: [OE-core] [PATCH] xserver-kdrive: work around issues with Jan 06 18:11:01 -Werror=address Jan 06 18:11:03 thats the subject Jan 06 18:11:45 someone just proposed an upstream xorg patch to remove -Winline from the default error-warnings Jan 06 18:12:19 probably a good move Jan 06 18:16:02 fray: hmm, I'll have to put a bit more thinking into this one; at the moment it still seems a bit dangerous though Jan 06 18:16:33 I think the first logical step is simply to warn the user htey have a recipe/directory mismatch between the two merged layers and they'll have to hand resolve it Jan 06 18:16:38 auto resolving is step 2 Jan 06 18:17:14 right, I have most of step 1 done here Jan 06 18:26:19 zeddii: if you want more help with testing that, let me know I can fire up a build here also. Jan 06 18:29:01 sgw. will do, I'm almost about to boot test. I pulled master again and picked up the kdrive fix, so sato is almost done building now. Jan 06 18:40:17 Anyone know if there is a way to disable RRECOMMENDS? core-image-minimal has grown slightly and I think I am picking some RRECOMMENDS that built from a previous build. Jan 06 18:44:22 I don't think there is, but it sounds like a reasonable thing to me.. Jan 06 18:48:02 fray, I saw a couple of files like usb.ids.gz and pci.ids.gz slip in recently (about 400K) enough to notice. Jan 06 18:48:31 ya.. those are the files for some of the usb-utils and pci-utils.. (and perhaps udev is using them?) Jan 06 18:48:41 fray: exactly. Jan 06 18:49:09 fray they are RRECOMMENDS so only pulled in if they are built Jan 06 18:49:31 BTW this is the exact problem that RPM developer has mentioned why he rejects the idea of "soft" depdnencies (i.e. recommends) Jan 06 18:49:35 either something is needed or it's not Jan 06 18:49:51 in this case, either those id files are needed or they're not.. so what is the right (automatic) policy? Jan 06 18:49:55 **** ENDING LOGGING AT Fri Jan 06 22:06:39 2012 **** BEGIN LOGGING AT Fri Jan 06 22:07:27 2012 **** ENDING LOGGING AT Fri Jan 06 22:08:09 2012 **** BEGIN LOGGING AT Fri Jan 06 22:08:41 2012 Jan 06 23:41:32 Note to the community - this is the final day of the Call for Participation for the Embedded Linux Conference in February. If you are planning to attend and/or give a Yocto-oriented presentation, please let me know. thanks **** ENDING LOGGING AT Sat Jan 07 02:59:57 2012