**** BEGIN LOGGING AT Mon Aug 25 02:59:58 2014 Aug 25 06:28:09 khem: ping Aug 25 06:31:48 otavio: setting the kernel-image in my local.conf did not work otherwise Aug 25 06:32:20 otavio: (for MACHINE="cubox-i") from meta-fsl-arm-extra Aug 25 06:35:32 do we have a way to specify additional kernel_imagetype btw? For this i.MX6 board it would be handy to build both uImage as well as (the default, non-working ?) zImage Aug 25 06:36:15 I.e. why is KERNEL_IMAGETYPE not a list? Aug 25 07:19:46 kergoth, ping Aug 25 07:22:43 RP, kergoth, since ages i have a tree lying around that virtualizes coreutils and util-linux so i can use busybox also as native providers of those (for checking and faster builds). Do you think it would be ok to spare me touching all those recipes by teaching bitbake to automatically handle virtual/ providers? That would make my life way easier.. Aug 25 07:33:16 Building 1.4 on a server with /bin/sh -> dash I get "ERROR: ld.so: object 'libpseudo.so' from LD_PRELOAD cannot be preloaded". Anyone remember any dash/bash fixes that could help here? Aug 25 07:33:46 Like I said pre-weekend I am stuck on 1.4 for this target, but would really like to build on the server. Aug 25 07:35:10 ln -s /bin/sh /bin/bash Aug 25 07:37:49 woglinde: would that I could, but my sysadmin does not want to :) Aug 25 07:38:57 blindvt`: whats up Aug 25 07:39:41 khem, could you bump uClibc srcrev please? Aug 25 07:39:59 what does !ubuntu use for /bin/sh. bash? Aug 25 07:40:06 blindvt`: sure, are you planning on release ? Aug 25 07:40:24 khem, Re eventfd_{read,write} what a crappy idea :) Let's just use a define Aug 25 07:40:28 tasslehoff: default is dash Aug 25 07:40:29 khem, yes Aug 25 07:41:05 blindvt`: OK. Let me run the mill tomorrow and test the builds out and send a patch Aug 25 07:41:24 khem: git rm compile-arm-fork-with-O2.patch;# applied something like that upstream Aug 25 07:41:25 khem: and you saw the '!' ? Aug 25 07:42:06 blindvt`: sure, I groked that Aug 25 07:42:35 tasslehoff: oh :) well most of them use bash yes Aug 25 07:42:54 khem: good! then I have ammo discussing with the sysadmin :) Aug 25 07:43:49 tasslehoff: ubuntu is special but helps in removing bashism something useful Aug 25 07:44:11 tasslehoff ask your admin why he hinders you to do your jobs Aug 25 07:44:18 and if that needs to get to the boss Aug 25 07:44:19 khem: :)' Aug 25 07:45:30 woglinde: there was no need. when given a proper reason for it he agreed to switch Aug 25 07:49:58 khem: hi. Any hint about klibc? What would you verify? Aug 25 07:50:16 btw...big earthquake I've heard... Aug 25 07:50:41 ant_work: I lost the connection to irc when you mentioned it so please write an email Aug 25 07:50:47 earthquake yes Aug 25 07:50:59 I am few mile away from where impact was seen Aug 25 07:51:12 6th degree is rather strong! Aug 25 07:52:42 khem, 0004-Add-clock_adjtime-syscall.patch is missing a stub fallback (please send a follow-up) Aug 25 07:55:55 blindvt`: I'm not 100% sure what you mean by automatically handling those provides Aug 25 07:57:05 khem: 6 earthquake sounds pretty scary. I hope you are all ok Aug 25 07:58:21 yes its quite strong Aug 25 07:58:45 * khem -> sleep("gn all") Aug 25 07:59:33 gn khem Aug 25 08:00:22 RP, i would want to use PREFERRED_PROVIDER_coreutils = busybox (let's say). ISTR that this does not work out of the box but that we have to workaround by using virtual/coreutils in all recipes and set that to a preferred provider Aug 25 08:02:19 RP, at least that's how it was some time ago and i assume that's still the case right? Aug 25 08:03:15 blindvt`: Something like the above should work if you get the right set of PREFERRED_PROVIDERs Aug 25 08:05:51 Hi all, Im trying to build qemuarm for cortex-a9, but didn't find the right config to do it, where can I look for help ? Aug 25 08:06:04 blindvt`: the thing you probably haven't accounted for is that recipes have dependencies like "util-linux-agetty" and therefore busybox would have to RPROVIDE those to ensure bitbake is satisfied that busybox can really provide these things Aug 25 08:06:31 RP, so iff that's still the case i would have to change all recipes mentioning coreutils resp coreutils-native to use the virtual/ prefix explicitly. To avoid that (especially coreutils and other such hardcoded packages) i'd have to somehow name a preferred provider for said packages and IIRC just stating that coreutils preferred provider is another package did not work at all a year ago or two Aug 25 08:08:22 khem: sent to oe-devel ML Aug 25 08:08:46 having one *libc in a separate layer is really a pain Aug 25 08:09:05 blindvt`: there are two namespaces, the "build" namespace (PROVIDE/DEPEND) and "runtime" (PACKAGES/RDEPENDS/RPROVIDES). You solve the first one with virtual/ usage, the second cannot be solved that way (grep for initscripts for the ugly way we solve that) Aug 25 08:09:56 blindvt`: I don't believe you have much of an issue with the first anyway as PREFERRED_PROVIDER *will* resolve that, your bigger problem is the second namespace since there are references like util-linux-agetty in recipes which are the real issue. Aug 25 08:10:39 blindvt`: You can resolve that second problem by having the busybox recipe RPROVIDE those things though, instead of the other types of change Aug 25 08:12:04 RP, mhm. I shall retry from scratch then Aug 25 08:30:44 RP: It seems to me there is a missing default-provider for libjpeg/jpeg-turbo in the case of nodistro Aug 25 08:36:43 RP, in 406aaccfd115ef189e4372f581bf19e5b0906a15 you unconditionally turned on UCLIBC_HAS_REENTRANT_RPC allegedly for quota. Is it really necessary to build and use rpc.rquotad? That strikes me as odd-ish for oe Aug 25 08:39:22 i mean quota is all nice and handy, but a remote quota server? :) Aug 25 08:55:51 blindvt`: Khem turned it on Aug 25 09:03:06 folks should really use libtirpc Aug 25 09:06:27 <_rink> hmm Aug 25 09:06:42 <_rink> i'm trying to add 'kernel-dev' to my SDK because I want to build kernel modules using the SDK yocto gives me Aug 25 09:06:52 <_rink> but SDK-EXTRAS += ' kernel-dev' does not seem to do anything Aug 25 09:06:56 <_rink> am I missing something ? Aug 25 09:15:09 <_rink> hmmm seems TOOLCHAIN_TARGET_TASK is the way to go here Aug 25 09:17:44 blindvt`: I don't disagree... Aug 25 11:13:19 Hi Aug 25 11:13:49 I want to use two repositories for a single component Aug 25 11:14:24 How do i clone both the repositories for the single component Aug 25 11:15:25 hmm, just add the URI to SRC_URI ? Aug 25 11:16:13 (or maybe that doesnt work because both are named git ... naming those might help then ?) Aug 25 11:25:24 but during do_unpack i will have problem Aug 25 11:30:36 I thought that the stuff I list in my IMAGE_INSTALL could be found as ipk's in deploy, but seems that doesn't hold for eglibc-dbg? Aug 25 11:43:18 or does adding eglibc-dbg to IMAGE_INSTALL change the eglibc package? Aug 25 11:56:07 pt: you need to set 'subdir' param for each git tree you want to clone. e.g. SRC_URI = "git://;subdir=something git://;subdir=anotherone" Aug 25 11:56:55 each tree will be unpacked in its 'subdir' folder Aug 25 11:59:31 and my question ends up as: if I do not add eglibc-dbg to my image, how can I add debug info with opkg later? Aug 25 12:11:43 pretend I haven't said anything today... Aug 25 13:08:11 blindvt`: uh? Aug 25 16:17:09 khem: are you there? Aug 25 16:40:45 yes Aug 25 16:49:40 khem: I'd like to try to figure out why he sdk failed Aug 25 16:49:47 khem: is you avail to help? Aug 25 16:50:07 sure Aug 25 18:08:11 blindvt`: you mean alternative impl for ! __NR_clock_adjtime case Aug 25 18:08:24 blindvt`: I think we should just error out Aug 25 18:08:32 there is no need for fall back Aug 25 19:54:17 guys, how to make bitbake ignore the following error: Aug 25 19:54:20 `ERROR: QA Issue: package jasper contains bad RPATH' ? Aug 25 19:54:24 Aug 25 19:56:24 INSANE_SKIP Aug 25 19:58:02 better yet, fix the buildsystem you're calling to stop passing -rpath :) **** BEGIN LOGGING AT Mon Aug 25 22:00:32 2014 Aug 25 22:09:24 khem, let me know if I can help with some more klibc debug before falling by drowsiness **** BEGIN LOGGING AT Mon Aug 25 22:13:09 2014 **** ENDING LOGGING AT Tue Aug 26 02:59:58 2014