**** BEGIN LOGGING AT Wed Feb 11 02:59:58 2015
Feb 11 06:41:14 <miandonmenmian> where can i edit the sanity check?
Feb 11 06:41:37 <miandonmenmian> my local.conf does not have anything about checks
Feb 11 06:55:21 <AndersD> miandonmenmian, if I remember correctly, there shouln't be any problem using local.conf to control the sanity checks. What do you want to do?
Feb 11 06:56:04 <miandonmenmian> disable the TMPDIR change location, or find out how to overide this
Feb 11 06:56:36 <miandonmenmian> found the recipe.. maybe i just comment it out, and run bitbake once
Feb 11 07:00:31 <miandonmenmian> moving the folder wont break yocto, right?
Feb 11 07:16:05 <FrankSansC> Hello
Feb 11 07:26:57 <miandonmenmian> Hello Frank
Feb 11 08:20:24 <FrankSansC> Small question : I'm currently building an image from the master branch of poky. If I want to build another one from a specific tag, is it ok to just git checkout <tag> and launch bitbake again ? It's not going to screw up sstate cache or do something else really wrong ?
Feb 11 09:25:35 <bluelightning> morning all
Feb 11 09:37:31 <Aurele__> Hi again, I  wrote an "image_types_aurele.bbclass" (from meta-aurele) I would like it to override the "image_types_fsl.bbclass" (from meta-fsl-arm) both provide "IMAGE_CMD_sdcard"
Feb 11 09:37:43 <Aurele__> but it doesn't work at all
Feb 11 09:40:19 <Aurele__> I have a default include from freescale wich set "IMAGE_CLASSES = "image_types_fsl""
Feb 11 09:41:21 <Aurele__> If I use IMAGE_CLASSES ?= "image_types_fsl" my sdcard image is generated otherwise freescale sdcard is generated no matter I'm doing...
Feb 11 09:42:24 <Aurele__> maybe it is possible to ignore this include default file?
Feb 11 09:43:24 <sujith_h> Hi all. Any clue why this error comes: http://pastebin.ubuntu.com/10171026/ ?
Feb 11 09:43:47 <bluelightning> Aurele__: to start, use = not ?=
Feb 11 09:44:25 <bluelightning> sujith_h: you moved to an older branch and kept the same TMPDIR?
Feb 11 09:44:37 <Aurele__> bluelightning, it is written in the freescale layer not mine...
Feb 11 09:44:48 <bluelightning> Aurele__: I mean, when setting your value
Feb 11 09:45:04 <bluelightning> Aurele__: oh, right - sorry now I understand
Feb 11 09:45:20 <bluelightning> try IMAGE_CLASSES_append = " image_types_aurele"
Feb 11 09:45:32 <bluelightning> (note the leading space in the value, that's important)
Feb 11 09:46:02 <sujith_h> bluelightning: Well I guess I had not touched branches. Let me cross verify again
Feb 11 09:46:15 <Aurele__> bluelightning, in my machine.conf I set the "IMAGE_CLASSES = "image_types_aurele"", anyhow I will try your advice it seems a good thing to try
Feb 11 09:47:02 <bluelightning> sujith_h: you wouldn't get this error unless you were using a TMPDIR previously created with a recent version of the build system and now you're running an older version
Feb 11 09:47:09 <sujith_h> bluelightning: thanks for that pointer. I forgot that I had moved the poky to daisy branch :(
Feb 11 09:47:37 <sujith_h> bluelightning: after changing that to master its working
Feb 11 09:48:13 <bluelightning> sujith_h: great
Feb 11 09:48:50 <sujith_h> bluelightning: thanks
Feb 11 09:50:44 <bluelightning> sujith_h: btw, I have a question for you now :)
Feb 11 09:51:08 <bluelightning> sujith_h: regarding this patch: http://patchwork.openembedded.org/patch/77527/
Feb 11 09:51:57 <bluelightning> sujith_h: actually two questions - (1) how well does this actually work i.e. how early does the splash screen show? (2) did you ever get around to sending the patch for psplash itself?
Feb 11 09:52:39 * sujith_h looks
Feb 11 09:54:02 <Aurele__> bluelightning, I also tried to rename my IMAGE_CMD_sdcard to "IMAGE_CMD_aurelesdcard" but it leads to an error in bitbake code (to my mind the freescale include file override the "IMAGE_CLASSES" so I can't add anything to this value...)
Feb 11 09:54:31 <Aurele__> If my hypothesis is true maybe I could ask to modify this...
Feb 11 09:55:06 <bluelightning> Aurele__: if you use the _append, there's no way that is stopping your class from being in the include file, however if the types added by the two classes are the same then you may have issues
Feb 11 09:55:18 <sujith_h> bluelightning: I don't remember it. I would have to test it again for the same.
Feb 11 09:55:23 <bluelightning> Aurele__: what is the error in the second instance?
Feb 11 09:56:00 <bluelightning> sujith_h: ok, no worries... I added a pointer to your patch in the bug relating to this in any case: https://bugzilla.yoctoproject.org/show_bug.cgi?id=4339
Feb 11 09:56:01 <yocti> Bug 4339: normal, Medium, 1.8, Qi.Chen, NEW , psplash doesn't work with systemd
Feb 11 09:56:21 <sujith_h> bluelightning: Sure. I will check it out :)
Feb 11 09:57:00 <Aurele__> bluelightning, http://pastebin.com/vE8xj90h
Feb 11 09:58:12 <Aurele__> bluelightning, maybe I don't declare correctly my sdcard image...
Feb 11 10:01:14 <bluelightning> Aurele__: that says to me that IMAGE_CMD_<whatever is in IMAGE_FSTYPES> isn't able to be queried at that point in the code
Feb 11 10:10:48 <Aurele__> bluelightning, If I comment the IMAGE_CLASSES = "image_types_fsl" in the include file, my image is known otherwise it is unknown
Feb 11 10:11:01 <Aurele__> bluelightning, even with the "_append"
Feb 11 10:12:16 <Aurele__> bluelightning, maybe they should just do a "IMAGE_CLASSES += "image_types_fsl"" with "+=" I could add my image classes don't you think?
Feb 11 10:13:03 <bluelightning> Aurele__: oh, I think I know what's going on... the _append won't work in this case because of how IMAGE_CLASSES is being used
Feb 11 10:13:30 <bluelightning> Aurele__: yes they should do that - you should complain to the layer maintainer
Feb 11 10:15:52 <Aurele__> bluelightning, lets ask otavio for example ;) but they are organized by mchine
Feb 11 10:29:13 <abelloni> I'm currently replacing image_types_fsl by wic
Feb 11 10:40:34 <Aurele__> abelloni, haha interesting
Feb 11 10:41:36 <Aurele__> abelloni, should I post something to get the modification anyway or is it not needed
Feb 11 10:44:16 <abelloni> you probably should do it anyway
Feb 11 10:44:46 <abelloni> if your platform is i.mx based, I have wic fully working
Feb 11 10:45:03 <abelloni> I still need to convert the mxs ones
Feb 11 10:48:35 <Aurele__> I have i.mx based :)
Feb 11 10:49:21 <Aurele__> abelloni, where could I get your wic working version?
Feb 11 10:56:07 <abelloni> let me push it
Feb 11 10:58:15 <abelloni> Aurele__: https://github.com/alexandrebelloni/wic
Feb 11 10:59:40 <abelloni> if you use u-boot, that is the wks:
Feb 11 10:59:42 <abelloni> http://code.bulix.org/khiqqy-87865
Feb 11 11:05:32 <Aurele__> abelloni, thanks
Feb 11 11:06:09 <Aurele__> abelloni, May I ask you in the following line :
Feb 11 11:06:10 <Aurele__> part /boot --source bootimg-partition --ondisk mmcblk0 --fstype=vfat --label boot --active --align 4096 --size 8M --extra-space 0
Feb 11 11:06:33 <Aurele__> bootimg-partition refer to a file?
Feb 11 11:07:34 <Aurele__> will it deal automatically with dtd, uimage and bootscript?
Feb 11 11:12:27 <abelloni> ah yes, wait a sec
Feb 11 11:13:02 <abelloni> use IMAGE_BOOT_FILES="zImage zImage-imx6dl-<board>.dtb;imx6dl-<board>.dtb"
Feb 11 11:13:09 <abelloni> in your machine.conf
Feb 11 11:13:20 <abelloni> obviously, replace <board>
Feb 11 11:18:19 <Aurele__> thanks again :)
Feb 11 11:19:51 <abelloni> tell me if something doesn't work
Feb 11 11:20:10 <abelloni> it is always good to have testers :)
Feb 11 11:22:32 <Aurele__> abelloni, where should I install it?
Feb 11 11:22:50 <abelloni> you actually don't care
Feb 11 11:23:04 <abelloni> just put it in your path
Feb 11 11:23:15 <abelloni> or use the full path to it
Feb 11 11:23:37 <abelloni> it will complain about plugin existing twice but that doesn't do any harm
Feb 11 11:24:00 <abelloni> you can also put the full scripts folder in your layer
Feb 11 11:24:16 <abelloni> then add meta-aurele/scripts to your PATH
Feb 11 11:24:34 <abelloni> for fsl, I'm doing that in the setup-environment script
Feb 11 11:24:37 <Aurele__> ok but I will have 2 wic?
Feb 11 11:24:44 <Aurele__> I will deal with this
Feb 11 11:24:53 <abelloni> yeah, it is not an issue
Feb 11 11:25:03 <Aurele__> I will let you know I I have errors
Feb 11 11:25:24 <abelloni> after:
Feb 11 11:26:02 <abelloni> https://github.com/Freescale/fsl-community-bsp-base/blob/master/setup-environment#L147
Feb 11 11:26:05 <abelloni> I'm using:
Feb 11 11:26:22 <abelloni> export PATH=$BUILDDIR/../sources/meta-aurele/scripts:$PATH
Feb 11 11:29:17 <Aurele__> abelloni, I have extended partition to create and logical partitions in it did you tried this before?
Feb 11 11:29:56 <abelloni> it will be done automatically
Feb 11 11:30:29 <abelloni> note that it also supports gpt
Feb 11 12:50:02 <j105rob> anyone familiar w/ building xen-image-minimal? Getting the following error ERROR: Nothing RPROVIDES 'xen-base'  running poky dizzy
Feb 11 13:19:46 <Acheloos> hello
Feb 11 13:20:15 <Acheloos> I have difficulties to understand the differences between BSPs, layers and recipes.
Feb 11 13:20:25 <Acheloos> Is there a good link about that ?
Feb 11 13:20:41 <Acheloos> Or someone who have the motivation to explain ;)
Feb 11 13:21:51 <Acheloos> If I am right, recipes describe how to fetch, configure, compile and package applications and images, and layers are sets of recipes.
Feb 11 13:21:57 <Acheloos> Is it correct ?
Feb 11 13:22:53 <j105rob> Acheloos: here is the developers manual: http://www.yoctoproject.org/docs/1.6.1/dev-manual/dev-manual.html
Feb 11 13:23:10 <j105rob> Here is the quick start guide: http://www.yoctoproject.org/docs/1.6.1/yocto-project-qs/yocto-project-qs.html
Feb 11 13:23:25 <Acheloos> i've read a good part of them
Feb 11 13:23:30 <j105rob> and build system workflow video: http://vimeo.com/36450321
Feb 11 13:23:47 <Acheloos> I've already been running Yocto, and I have some working image.
Feb 11 13:24:04 <Acheloos> My problem is only concerning some vocabulary.
Feb 11 13:25:31 <bluelightning> Acheloos: your description of recipes and layers is generally accurate
Feb 11 13:26:07 <Acheloos> Cool. But I fail to see what is the BSP, and how it is different from the layers
Feb 11 13:26:30 <bluelightning> Acheloos: BSP = "Board Support Package"; in our system that means a layer specifically for enabling support for building for a particular machine (containing a machine configuration and recipes/bbappends as needed)
Feb 11 13:26:32 <Acheloos> The documentation mixes the two, so I'm confused : http://www.yoctoproject.org/docs/1.6.1/bsp-guide/bsp-guide.html#bsp-layers
Feb 11 13:26:42 <Acheloos> Ha !
Feb 11 13:26:51 <Acheloos> So the BSP *is* a layer.
Feb 11 13:26:53 <Acheloos> Ok.
Feb 11 13:27:32 <bluelightning> in our system it is yes... the term BSP is also used outside our system
Feb 11 13:46:33 <hugovs>  /j #oe
Feb 11 13:46:50 * hugovs :P
Feb 11 13:48:22 <abelloni> treason !
Feb 11 13:54:12 <pohly> Yocto uses a time-based release schedule with a new release every six months, doesn't it?
Feb 11 14:10:52 <vmeson> pohly: yes, late April, late October or thereabouts.
Feb 11 14:36:37 <kergoth> morning
Feb 11 14:43:08 <rich> quit
Feb 11 14:57:22 <bobdog555> Anyone know why genericx86 linux build does not work on Toshiba Tecra laptops?
Feb 11 15:01:14 <bluelightning> bobdog555: not really, no - what's the actual problem?
Feb 11 15:01:20 <bluelightning> i.e. where does it fail?
Feb 11 15:02:40 <bobdog555> It says Loading /vmlinuz ok ; Loading /initrd ok /booting kernel failed: invalid argument
Feb 11 15:02:47 <bobdog555> and keeps looping
Feb 11 15:03:31 <bobdog555> It works on a Dell
Feb 11 15:06:54 <bluelightning> bobdog555: is this with a live image? if not, how is it being deployed on the media booting on the device?
Feb 11 15:07:42 <bobdog555> Live image booting from a CD iso
Feb 11 15:08:05 <JEEBsv> hmm, what's the "proper" way to do multiple configurations for a recipe? multiple recipies or there's something else? I'm packetizing a thing that can either have very little dependencies, or a lot of them
Feb 11 15:08:12 <rich> hello, I have a question. I have an issue with qemu (archlinux) and I wondered if there are alteratives anyone is using? arm build that runs on my raspberrypi with no issues.
Feb 11 15:08:50 <bobdog555> bluelightning: live image booting from a CD (iso)
Feb 11 15:10:33 <bluelightning> bobdog555: ok, could you please file a bug on bugzilla.yoctoproject.org?
Feb 11 15:11:01 <bobdog555> blueligtning: ok. thanks.
Feb 11 15:11:03 <bluelightning> JEEBsv: typically PACKAGECONFIG for that sort of thing
Feb 11 15:11:34 <bobdog555> bluelightning: (can't type today) ok.... thanks
Feb 11 15:11:41 <JEEBsv> bluelightning: ah
Feb 11 15:11:58 <bluelightning> bobdog555: I can recommend tab completion if your IRC client supports it ;)
Feb 11 15:12:17 <bobdog555> bluelightning: Doh
Feb 11 15:12:24 <bluelightning> i.e. I just do bob<tab> to get your name
Feb 11 15:12:45 <JEEBsv> > pkg-config results when you google for PACKAGECONFIG
Feb 11 15:12:50 <JEEBsv> thank you, big G
Feb 11 15:13:06 <bluelightning> JEEBsv: I don't know why but our proper manuals don't seem to feature highly in google :(
Feb 11 15:13:22 <JEEBsv> yeah, but at least the ML results are on the top :)
Feb 11 15:56:07 <sgw_> armpit: good morning
Feb 11 15:56:42 <kergoth> Does anyone know why we duplicate OECORE_TARGET_SYSROOT and SDKTARGETSYSROOT? both are exported in the environment-setup script
Feb 11 15:56:47 <kergoth> seems a bit redundant
Feb 11 15:56:49 <Matulis> might be a stupid question, I am trying to do a "do_install" for a recipe with precompiled binaries, is there a way to specify all of the folders in my S directory to go to D?
Feb 11 15:56:54 <kergoth> compatibility perhaps?
Feb 11 15:57:05 <kergoth> Matulis: do_install is a shell script
Feb 11 15:57:12 <kergoth> do whatever you'd normally do to copy files from one place to another
Feb 11 15:57:15 <kergoth> e.g. cp -a
Feb 11 15:57:48 <Matulis> okay, thanks! I was reading something and it said it was preferred to do the install -d or whatever, but ill just use cp
Feb 11 16:01:11 <kergoth> install works well for single files, or just a couple, but for an entire tree, it doesn't make sense
Feb 11 16:09:29 <Matulis> hey kergoth, is there a way to skip do_package_qa?
Feb 11 16:09:55 <Matulis> Its picking up some dependencies that I don't have in my binaries
Feb 11 16:34:06 <ulf`> sgw_: ping
Feb 11 16:51:50 <realBigfoot> Is there a way to rewrite do_rootfs ?
Feb 11 16:57:15 <kergoth> any recipe can always replace a function from a class
Feb 11 16:57:16 <kergoth> images are no different
Feb 11 16:58:37 <realBigfoot> kergoth: so where do I begin ?
Feb 11 16:58:42 <kergoth> ?
Feb 11 16:58:51 <kergoth> have you actually read any recipes?
Feb 11 16:58:58 <kergoth> after hte inherit, do_rootfs () {
Feb 11 16:59:00 <kergoth> do something
Feb 11 16:59:00 <kergoth> }
Feb 11 16:59:07 <kergoth> the same way you define any other function, ever
Feb 11 16:59:52 <realBigfoot> nice... now I got it
Feb 11 17:00:12 <realBigfoot> i didn't know it was a possibility to put do_rootfs on image thanks
Feb 11 17:01:02 <kergoth> i'd suggest reading the bitbake manual at some point
Feb 11 17:01:21 <kergoth> images aren't special, they're just recipes that inherit a class
Feb 11 17:02:03 <realBigfoot> kergoth: will do
Feb 11 17:02:08 <realBigfoot> thank you very much
Feb 11 17:03:01 <kergoth> np
Feb 11 17:06:42 <abelloni> is there a way to ask the autotools class to do an in tree build ?
Feb 11 17:06:55 <kergoth> inherit autotools-brokensep
Feb 11 17:07:11 <abelloni> thanks :)
Feb 11 17:07:28 <ulf`> kergoth is on a roll :)
Feb 11 17:07:29 <abelloni> I actually fixed the sources but it may be handy at some point
Feb 11 17:16:35 <bluelightning> realBigfoot: depending on what you want to do, you may not want to rewrite all of do_rootfs... there are a bunch of different ways to extend it built-in
Feb 11 17:19:00 <abelloni> do we have a recipe for pandoc-native ?
Feb 11 17:19:24 <abelloni> for now it is relying on having pandoc installed on the build machine which is not really nice
Feb 11 17:19:54 <abelloni> (maybe it would be easier to stop building the man)
Feb 11 17:20:40 <realBigfoot> bluelightning: how would I extend it ?
Feb 11 17:22:41 <bluelightning> realBigfoot: e.g. if you want to run commands just after the image is created, you can define a shell function and add it to ROOTFS_POSTPROCESS_COMMAND
Feb 11 17:23:15 <realBigfoot> bluelightning: nice it should be better than rewriting it
Feb 11 17:23:25 <bluelightning> realBigfoot: there's also the IMAGE_CLASSES / IMAGE_TYPES / IMAGE_CMD / COMPRESS_CMD / etc. system for defining new image and compression types
Feb 11 17:24:05 <bluelightning> how-to documentation is a bit sparse on that but at least most of the variables should be in the reference manual
Feb 11 17:24:52 <realBigfoot> bluelightning: indeed, it is there. i was reading it but i haven't got in this part...
Feb 11 17:24:58 <abelloni> isn't using ROOTFS_POSTPROCESS_COMMAND disouraged ?
Feb 11 17:25:06 <realBigfoot> nice info
Feb 11 17:25:52 <bluelightning> abelloni: not especially - I mean, there are alternatives for some things you might consider doing (e.g. adding users/groups), but sometimes there's no other way
Feb 11 17:26:48 <bluelightning> abelloni: postinstall scripts in packages may be a better alternative for some types of things (especially if you are using package management at runtime)
Feb 11 17:28:07 <abelloni> the issue is that you end up with abuses like that: https://github.com/linux4sam/meta-atmel/blob/master/recipes-qt/images/atmel-qt4e-demo-image.bb#L92
Feb 11 17:28:34 <bluelightning> abelloni: right, that is a bit nasty
Feb 11 17:29:11 <abelloni> or:
Feb 11 17:29:12 <abelloni> https://github.com/linux4sam/meta-atmel/blob/master/recipes-core/images/atmel-demo-image.inc#L36
Feb 11 17:30:15 <bluelightning> abelloni: yeah, they aren't the only ones who have removed the kernel image that way... I think the proper mechanism may not even be documented
Feb 11 17:31:01 <abelloni> yeah, what is the proper mechanism ? :)
Feb 11 17:31:01 <bluelightning> there is a bug open suggesting that the default may not be the best: https://bugzilla.yoctoproject.org/show_bug.cgi?id=5990
Feb 11 17:31:03 <yocti> Bug 5990: enhancement, Medium+, 1.8, dvhart, ACCEPTED , kernel.bbclass: Revamp packaging and dependencies
Feb 11 17:32:01 <bluelightning> abelloni: basically set RDEPENDS_kernel-base = "" in your kernel recipe (the default, set with ?= in kernel.bbclass, is "kernel-image")
Feb 11 17:32:31 <abelloni> oh, right
Feb 11 17:32:59 <bluelightning> actually I just found that it is in our kernel manual: http://www.yoctoproject.org/docs/current/kernel-dev/kernel-dev.html#idm140599411945520
Feb 11 17:33:08 <bluelightning> not that everyone would think to look there necessarily
Feb 11 17:33:48 <abelloni> better than nothing
Feb 11 17:41:15 <abelloni> bluelightning: I'm solving https://lists.yoctoproject.org/pipermail/yocto/2014-November/022381.html
Feb 11 17:41:42 <abelloni> I mean, I'm doing it ustream
Feb 11 17:42:05 <abelloni> would you prefer having the -foreign option back or the AUTHORS file ? :)
Feb 11 17:43:01 <bluelightning> personally, I don't mind... whatever looks like it makes sense :)
Feb 11 17:43:34 <abelloni> ok :)
Feb 11 19:39:04 <hugovs> hi all, i'm working with meta-java and i'm hitting the following error when building cacao-native
Feb 11 19:39:13 <hugovs> | libtool: Version mismatch error.  This is libtool 2.4.4, but the
Feb 11 19:39:14 <hugovs> | libtool: definition of this LT_INIT comes from libtool 2.4.2.
Feb 11 19:39:14 <hugovs> | libtool: You should recreate aclocal.m4 with macros from libtool 2.4.4
Feb 11 19:39:14 <hugovs> | libtool: and run autoconf again.
Feb 11 19:39:14 <hugovs> | Makefile:1260: recipe for target 'cordprnt.lo' failed
Feb 11 19:39:16 <hugovs> | make[5]: *** [cordprnt.lo] Error 63
Feb 11 19:39:18 <hugovs> | make[5]: *** Waiting for unfinished jobs....
Feb 11 19:40:10 <hugovs> if i add a call to autoreconf --force --install in recipe's do_configure_prepend(), it gets built without problems.
Feb 11 19:40:53 <hugovs> the only thing is i'm not sure if this is the right thing to do.
Feb 11 19:42:57 <zecke> hugovs: normally if you inherit autotools it should run autoreconf
Feb 11 19:47:37 <hugovs> zecke: yes. and the recipe does it. autotools-brokensep, actually. autoreconf is being executed.
Feb 11 19:49:09 <hugovs> the error happens when building a subdir that seems a project inside a project...
Feb 11 20:15:55 <paulg> I think this is like the perl build failure race ; I'm seeing it say once in 5 builds (give or take)...
Feb 11 20:15:59 <paulg> NOTE: Checking autotools environment for common misconfiguration
Feb 11 20:15:59 <paulg> ERROR: This autoconf log indicates errors, it looked at host include and/or library paths while determining system capabilities.
Feb 11 20:15:59 <paulg> Rerun configure task after fixing this. The path was '/home/paul/poky/build/tmp/work/core2-64-poky-linux/mozjs/17.0.0-r0/build'
Feb 11 20:16:42 * paulg wishes it actually said what string / path it found that it didn't like....
Feb 11 20:17:31 <kergoth> read config.log
Feb 11 20:20:53 <hugovs> i think i got it
Feb 11 20:22:32 <hugovs> autotools.bbclass removes aclocal.m4 before calling autoreconf
Feb 11 20:23:07 <hugovs> but the package, as i said, has some others aclocal.m4 inside subdirs...
Feb 11 20:25:44 <paulg> kergoth, yep, rummaging through config.log already -- just saying it would have been nice to have a hint....
Feb 11 20:26:04 <kergoth> *shouldn't* matter, since afaik autoreconf -f should run aclocal -f which should overwrite any existing aclocal.m4.. unless i'ts not processing those subdirs, which would only happen if those subdirs aren't declared as such in the configure.{in,ac} (e.g. gcc/binutils/etc descent into the configure scripts of subdirs in makefile rules, rathe rthan the toplevel configure script calling the subdir configur escripts)
Feb 11 20:26:07 * kergoth shrugs
Feb 11 20:26:36 <kergoth> paulg: well, it'll be a host include or library path. e.g. not in the sysroot. so most likely /usr/include, /usr/lib, or /lib
Feb 11 20:26:46 <kergoth> i'd check for those before digging further
Feb 11 20:27:01 <kergoth> could be others, but those are te most likely suspects, afaik
Feb 11 20:27:58 <paulg> aha, here it is.
Feb 11 20:28:00 <paulg> configure:5494: checking for IceConnectionNumber in -lICE
Feb 11 20:28:00 <paulg> configure:5513: x86_64-poky-linux-gcc  -m64 -march=core2 -mtune=core2 -msse3 -mfpmath=sse --sysroot=/home/paul/poky/build/tmp/sysroots/genericx86-64 -o conftest  -O2 -pipe -g -feliminate-unused-debug-types  -Wl,-O1 -Wl,--hash-style=gnu -Wl,--as-needed -L/usr/lib conftest.c -lICE   1>&5
Feb 11 20:28:01 <paulg> /home/paul/poky/build/tmp/sysroots/x86_64-linux/usr/libexec/x86_64-poky-linux/gcc/x86_64-poky-linux/4.9.1/ld: warning: library search path "/usr/lib" is unsafe for cross-compilation
Feb 11 20:29:05 <paulg> now the mystery is why that only happens _sometimes_ ....
Feb 11 20:29:45 <kergoth> possibly an autodetected dependency
Feb 11 20:30:04 <kergoth> enable ice support if it exists, and whether it eixsts isn't deterministic if it doesn't 1) depend on it, or 2) disable it via the configure argument
Feb 11 20:32:28 <paulg> sounds like a reasonable plan.
Feb 11 20:32:58 <paulg> going to capture a config.log from a passed build 1st and then compare them.
Feb 11 20:39:14 <abelloni> I don't remember, how can I clean the stale sstate ?
Feb 11 20:42:16 <Crofton|work> bitbake -c cleansstate recipe
Feb 11 20:47:23 <zecke> hugovs: oh yeah.. then do_configure_append or such..
Feb 11 21:02:20 <abelloni> Crofton|work: no, I meant the sstate that doesn't correspond to any recipe anymore
Feb 11 21:02:35 <Crofton|work> ah
Feb 11 21:08:47 <JaMa> how different will be topics of Yocto Project Developer Day US 2015 from 2014 in Dusseldorf?
Feb 11 21:11:26 <Crofton|work> JaMa, I would guess not by much, especially since I would expect little overlap
Feb 11 21:12:24 <JaMa> ok, we should go to lemans karting in Fremont after that :)
Feb 11 21:13:04 <Crofton|work> I was just checking on the karting
Feb 11 21:42:15 <hugovs> kergoth: with acpaths="" in the recipe, package gets built
Feb 11 21:43:10 <hugovs> but then again, i'm not sure if this is the right thing to do...
Feb 11 21:53:23 <abelloni> JaMa: did you try to build qt5 without opengl recently ?
Feb 11 21:54:25 <JaMa> abelloni: no
Feb 11 21:55:22 <abelloni> qtdeclarative fails with "Project ERROR: Unknown module(s) in QT: quick-private"
Feb 11 21:55:43 <abelloni> I can find plenty of old reports of the same error but nothing seems to solve it
Feb 11 21:55:58 <Crofton|work> JaMa, I sent fix for pylxml
Feb 11 21:56:35 <JaMa> I know, but it wasn't included in built started 3 days ago
Feb 11 21:56:40 <Crofton|work> ok
Feb 11 21:56:51 <Crofton|work> just checking you knew the patch is out there already
Feb 11 21:57:10 <JaMa> I've updated its status on patchwork so I know :)
Feb 11 21:57:28 <JaMa> and it's in master-next now..
Feb 11 23:00:26 <darkhorse_> bluelightning: are you much familiar with kernel.bbclass? i build a kernel with a built-in initramfs. everything gets built fine. but the size of the kernel binary seems larger after the second pass through the kernel build (do_bundle_initramfs)
Feb 11 23:02:06 <darkhorse_> bluelightning: what i mean by larger size is this: kernel itself is 2M, initramfs.cpio.gz is 3.7M. but instead of ~5.7M i see the final kernel = ~8.2M
Feb 11 23:08:19 <paulg> deconstruct the final output and see what chunks make up what sizes?
Feb 11 23:19:57 <abelloni> darkhorse_: you shouldn't compress your intiramfs if your kernel is already compressed
Feb 11 23:21:01 <darkhorse_> abelloni: my kernel is not compressed
Feb 11 23:22:57 <abelloni> 2M for an uncompressed kernel is surprising
Feb 11 23:24:00 <darkhorse_> abelloni: yeah - it is a minimalist kernel
Feb 11 23:31:26 <darkhorse_> abelloni: what kernel config option can i use to compress it?
Feb 11 23:35:11 <kergoth> hugovs: that generally indicates that someone had copied old macros into the source tree. you could also have found and removed them (e.g. libtool.m4)
Feb 11 23:35:17 <abelloni> KERNEL_GZIP, KERNEL_LZMA, ...
Feb 11 23:35:17 <kergoth> libtoolize will do such copying
Feb 11 23:39:39 <Matulis> sorry to interrupt, is there a good way to check why a patch is failing? I am getting a "Patch does not apply (enforce with -f)
Feb 11 23:40:40 <rewitt> Matulis: I would expect a *.rej file in the workdir
Feb 11 23:42:14 <Matulis> cant find a rej file
Feb 11 23:43:27 <rewitt> what if you change PATCHTOOL to patch?
Feb 11 23:44:03 <Matulis> sorry, not sure how to do that?
Feb 11 23:44:58 <rewitt> PATCHTOOL = "patch" in the recipe doing the patching
Feb 11 23:45:34 <rewitt> quilt is used by default, but I can remember not having much information in the do_patch logs on failures as well
Feb 11 23:45:40 <Matulis> so its a kernel patch, so I would add that to the kernel-*.bb
Feb 11 23:45:53 <rewitt> Oh, the kernel has it's on patching code I believe :-/
Feb 11 23:45:58 <rewitt> s/on/own
Feb 11 23:46:40 <Matulis> so I am kind of baffled, I changed a line in the git repo, made a diff, added it to the patch, not sure why its saying it doesnt apply
Feb 11 23:48:50 <rewitt> Matulis: Sorry, I don't know much about the kernel patching mechanism
Feb 11 23:49:10 <Matulis> no problem! thanks for the help!
Feb 12 00:02:21 <paulg> kergoth,  the mozjs thing was a missing dependency on libxt ; patch sent.
Feb 12 00:02:36 <abelloni> Matulis: did you update the offsets in the patch ?
Feb 12 00:09:47 <Matulis> ughhhhhhhh
Feb 12 00:09:56 <Matulis> whitespace issue
**** ENDING LOGGING AT Thu Feb 12 02:59:58 2015