**** BEGIN LOGGING AT Tue Feb 19 02:59:58 2013 Feb 19 04:27:39 someone seeing a lot of rebuilds with today's bitbake + oe-core? Feb 19 04:32:17 I had image completely build with only systemd-machine-units failing, after cleansstate of systemd-machine-units it started with do_fetch for stuff like dbus/libcap etc. and now doing 9000 tasks again Feb 19 07:55:46 hi Feb 19 07:56:06 I've got a problem with my custom kernel Feb 19 07:56:28 it doesn't boot if I build it with yocto toolchain Feb 19 07:56:48 I see only something like 'Unpacking linux' and that's all Feb 19 07:57:31 it does work though if I build it with another toolchain (gcc-4.3) Feb 19 07:57:47 the hw is cortex-a8 Feb 19 07:58:51 it's vanilla 3.1.0 with some patches for my hw Feb 19 07:59:57 I've tried to build the yocto 3.4 kernel and it seems to work Feb 19 08:00:39 LunohoD: are you sure it is not your cmdline? Feb 19 08:02:48 well, the .config is the same in both cases Feb 19 08:03:34 LunohoD: just some general ideas.. compare the assembly.. or enable the low level debugging in the kernel so that even early code will print information Feb 19 08:03:56 as a data point.. I am using the edison toolchain to build a 3.2.x kernel fine Feb 19 08:04:26 bbl Feb 19 08:57:19 good morning Feb 19 09:20:05 morning all Feb 19 09:27:24 morning all Feb 19 09:28:02 How can I easily rebuild a Qt/QML app *in-situ* from the working directory (for short development cycles) Feb 19 09:28:35 I see only the do_configure and do_compile steps, but I think there is an additional step hidden somewhere that I cannot re-run. Feb 19 14:48:01 hi, i have a question about using yocto to build custom versions of the same package, but with different submodule versions - Is there a class to handle that? or an example i can look at? the closest i have seen is psplash. Feb 19 14:53:08 Hello :) Feb 19 14:53:32 Here is the new format for the weekly QA report : https://wiki.yoctoproject.org/wiki/1.4_QA_Status Feb 19 14:54:19 it will update in realtime so if you are interested in a certain feature, you can check it anytime Feb 19 15:31:27 Corneliu: typo: Meta-Yocyo -> Meta-Yocto Feb 19 15:39:15 jstashluk: thank you :D Feb 19 16:00:48 YPTM: welcome to the technical team meeting. Please let me know who's on the bridge Feb 19 16:01:20 YPTM: Paul Eggleton is on the bridge Feb 19 16:01:35 YPTM: Kevin Strasser is here Feb 19 16:02:02 YPTM: Björn Stenberg is on the call Feb 19 16:02:15 YPTM: belen is on the call Feb 19 16:03:07 YPTM: Amit is here Feb 19 16:03:22 YPTM: ross can't make it today. Feb 19 16:03:54 YPTM: Any opens? Feb 19 16:04:30 we're a silent bunch :-) Feb 19 16:05:50 :) Feb 19 16:07:14 * nitink YPTM is on the call Feb 19 16:07:54 YPTM: Corneliu joined Feb 19 16:12:48 Cristian joined Feb 19 16:14:43 https://wiki.yoctoproject.org/wiki/1.4_QA_Status Feb 19 16:21:05 Please send feedback for https://wiki.yoctoproject.org/wiki/1.4_QA_Status at corneliux.stoicescu@intel.com :) Feb 19 16:21:55 YPTM: thanks for joining the meeting. Have a nice day/evening Feb 19 17:04:35 Sane to you Song_Liu Feb 19 17:04:46 I meant "Same" Feb 19 17:07:12 * nitink will be at YPDD around 12:20pm today Feb 19 17:45:58 is it just me or did master start mucking with my git SRC_URI Feb 19 17:46:21 it appears to be removing the git:// Feb 19 17:46:22 git clone --bare --mirror //git.freescale.com/ppc/sdk/linux.git /local/home/mattsm/git/poky/build-master/downloads/git2/..git.freescale.com.ppc.sdk.linux.git Feb 19 17:46:27 from do_fetch log Feb 19 17:47:59 seemingly caused by 'bitbake: fetch2: Adapt encode/decode url to use URI class' Feb 19 17:50:17 but, really it's 'bitbake: fetch2: Add a class representing a generic URI' Feb 19 18:45:59 some days I really hate bitbake and oe, and others I love it, and some I think it's both Feb 19 18:47:27 :) Feb 19 18:47:52 I hate it when it removes almost all my stamp files without known reason Feb 19 20:18:58 hello all, i have question regarding the i586-pokymllib32-linux-g++ after "bitbake meta-toolchain" with multilib, when use that c++ compiler it always complains cannont find libstdc++, even though I have that 32bit libstdc++ installed user /usr/lib Feb 19 20:19:10 and /usr/lib is in search path ... Feb 19 20:19:30 anyone could point out why? Feb 19 20:20:05 and I was not crazy Feb 19 20:20:13 urlparse for git was broken on my version of python Feb 19 20:20:29 patch sent Feb 19 20:20:44 can anyone help? Feb 19 20:47:26 reeve: are you using it from a recipe or directly? Feb 19 20:47:34 you should not be using it directly Feb 19 20:48:20 reeve: meta-toolchain emits a toolchain archive in deploy you can use, you should not touch the sysroot contents directly ever as msm points out Feb 19 21:20:10 msm, sorry I was away for a while Feb 19 21:20:31 msm, I figured it out: I forgot to install lib32-libstdc++-dev package, that's the reason Feb 19 21:20:43 thanks though Feb 19 22:11:59 msm: Please report that on the bitbake list Feb 19 22:13:23 RP: hmm, thoughts on teaching exec_func to support passing arguments to the function at some point? of course, for python we'd have to store the argspec, but we already do that for def'd functions.. Feb 19 22:15:09 kergoth: Changing exec_func to pass the args is easy enough at least... Feb 19 22:17:14 msm: what python version are you using? Feb 19 22:19:53 RP: was just thinking there are times it could be useful to run a utility function directly, programmaticlaly. e.g. create_wrapper, or whatever Feb 19 22:20:14 and would need to pass arguments to make that at all useful Feb 19 22:21:26 kergoth: yes, true... Feb 19 22:21:55 kergoth: when I was improving package.bbclass I did something nasty with globals to work around this limitation :/ Feb 19 22:22:25 hehe, you know what would be cute/twisted, shove python wrappers into the namespace the way we do variables in inline python, so you could just run create_wrapper('foo', 'bar') Feb 19 22:22:55 zibri: 2.6 Feb 19 22:23:12 RP: already sent a patch Feb 19 22:23:15 kergoth: I like your description of that :) Feb 19 22:23:16 msm: right, sorry about that :) Feb 19 22:23:21 msm: just saw it, thanks Feb 19 22:24:29 RP: yeah, that kind of thing is common enough, really.. just look at how we handle command-like format strings, like we used to for the devshell cmd, and wget FETCHCMD, and.. any time we end up creating a copy of the metadata nad setting one or two vars and then getVar()ing just to implement a bitbake-based format-string, it's not ideal, i think Feb 19 22:24:40 * kergoth shrugs Feb 19 22:25:41 course, the new style python string formatting didn't exist when we started either, so our options were limited Feb 19 22:25:45 heh Feb 19 22:25:59 kergoth: things have changed a lot over the years... Feb 19 22:26:25 kergoth: the pool for parsing seemed stable btw so pulled it Feb 19 22:26:36 kergoth: passed tests on the autobuilder ok Feb 19 22:26:38 cool, glad to hear it Feb 19 22:36:23 RP, you awake? Feb 19 22:38:17 Crofton: yes Feb 19 22:38:29 heh Feb 19 22:38:31 I am in the back Feb 19 22:38:42 Crofton: I was thinking you must be after the email ;-) Feb 19 22:38:45 populate sdk makes an sdk and I added cheetah Feb 19 22:39:17 the good news is the sdk is resistant to using assuem native stuff Feb 19 22:39:28 the bad news is cheetah needs anothe rpython module Feb 19 22:39:38 that I need to figure out how to nativesdk Feb 19 22:39:48 Crofton: BBCLASSEXTEND = "nativesdk" ? Feb 19 22:40:10 yeah, I need to see where to put that Feb 19 22:41:12 Crofton: in the python module recipe ;-) Feb 19 22:41:20 we should probably start doing this by default for everything Feb 19 22:41:33 yeah, I ran into this last night on the way out Feb 19 23:04:55 rp, I had to do this: Feb 19 23:04:57 -TOOLCHAIN_HOST_TASK ?= "nativesdk-packagegroup-sdk-host packagegroup-cross-cana Feb 19 23:04:57 +TOOLCHAIN_HOST_TASK ?= "nativesdk-packagegroup-sdk-host packagegroup-cross-cana Feb 19 23:05:01 ah crap Feb 19 23:05:48 I had to use another var to insert something into TOOLCAHIN_HOST_TASK Feb 19 23:06:00 Crofton: why? Feb 19 23:06:26 it seems like setting it in the recipe erased the exisitng setting Feb 19 23:06:44 so I got cheetah, but not compiler Feb 19 23:06:50 Crofton: TOOLCHAIN_HOST_TASK_append = " xxx" Feb 19 23:06:58 I'll try that Feb 19 23:07:09 I was using += I guess they are not quite the same Feb 19 23:07:40 Crofton: they're not. ?= won't set if already set Feb 19 23:08:44 grr, it is subtle stuff that makes things tricky Feb 19 23:08:50 wtf, something i did in this class is causing KeyErrors from uihelper Feb 19 23:08:54 * kergoth mutters Feb 19 23:09:44 I know I have seen this discussion on the list and it still hasn't sunk in Feb 19 23:09:49 \ Feb 19 23:09:59 \\\\\\\\ Feb 19 23:10:04 gah Feb 19 23:10:23 RP, is being heroic atm Feb 19 23:14:20 I heard baserock handles this Feb 19 23:15:42 Crofton: handles what? Feb 19 23:16:10 repeatability and traceability Feb 19 23:16:30 RP, good job speaking into the mike Feb 19 23:16:43 although it seems like they must have turned the gain up for you :) Feb 19 23:17:03 Crofton: different mic Feb 19 23:17:14 heh Feb 19 23:18:06 We need to drop the Project and just say Yocto Feb 19 23:18:18 * Crofton is trolling Feb 19 23:18:44 Crofton: tell Tracey that ;-) Feb 19 23:19:03 I wish she was here for that segment Feb 19 23:24:35 how is the YP DevDay going? Feb 19 23:25:00 good Feb 19 23:25:02 fray, we're devday'ing up a storm here Feb 19 23:25:03 * fray won't be able to be at ELC until Friday morning.. (speaking with someone else Friday afternoon) Feb 19 23:25:06 ;) Feb 19 23:25:37 but I am staying hte night if anyone else is around and wants to do dinner Feb 19 23:25:57 * kergoth is sad he didn't make it this year Feb 19 23:29:20 RP, the headache is the the python-modules target has no nativesdk version Feb 19 23:29:37 cgi seems to be in a nativesdk-python-netserver Feb 19 23:29:55 the problem is I need to iterate sdks to find all the python depends Feb 19 23:30:05 this is a larger issue with python stuff Feb 19 23:52:17 https://gist.github.com/4991389 Feb 19 23:52:22 that's a new one.. Feb 19 23:54:06 hi , is there package available to show CPU temperature for a x86_64 CPU like i3status? Feb 19 23:57:40 kergoth: BB_ORIGENV? Feb 20 00:01:00 this class inherits terminal and runs oe_terminal, and that results in this failure Feb 20 00:01:14 so that wouldnt surprise me Feb 20 00:05:39 kergoth: right, I suspect that is likely cause... Feb 20 00:07:05 * kergoth adds BB_ORIGENV to oe_terminal's vardepsexclude and sees it go away **** ENDING LOGGING AT Wed Feb 20 02:59:58 2013