**** BEGIN LOGGING AT Fri May 31 02:59:58 2013
May 31 03:08:01 * nerdboy enjoys the occasionaly "drive-by" flogging
May 31 03:08:15 <nerdboy> at least faceplant is good for something...
May 31 08:24:33 <hrw> morning
May 31 08:40:25 <bluelightning> morning hrw, all
May 31 08:40:28 <ynezz> flood morning
May 31 08:40:34 <bluelightning> heh
May 31 08:40:36 <bluelightning> hi ynezz
May 31 08:40:42 <ynezz> hi
May 31 08:40:44 <hrw> ynezz: ;D
May 31 08:40:50 <hrw> hi bluelightning ynezz 
May 31 08:41:02 <hrw> today is my last day at Linaro
May 31 08:41:13 <hrw> no more work on meta-linaro from me then
May 31 08:41:25 <ynezz> it's not your hobby? :p
May 31 08:41:45 <CruX|> hrw: and your next working position will be ?
May 31 08:42:03 <hrw> bluelightning: can you cchange maintainer of meta-linaro(-toolchain) and meta-aarch64 to suihkulokki?
May 31 08:42:21 <hrw> CruX|: ask me in a week - it is still in discussion
May 31 08:42:22 <bluelightning> hrw: sure, one sec
May 31 08:43:09 <hrw> ynezz: OE stopped being hobby in February 2007. It became a job.
May 31 08:43:49 <ynezz> nice
May 31 08:44:11 <hrw> bluelightning: Riku Voipio <riku.voipio@linaro.org>
May 31 08:44:15 <suihkulokki> aye
May 31 08:44:33 <hrw> ops, had to go to /query
May 31 08:44:59 <bluelightning> hrw: thanks... also meta-aarch64?
May 31 08:45:07 <hrw> yes
May 31 08:45:17 <bluelightning> ok
May 31 08:45:20 <hrw> thanks
May 31 08:45:37 <hrw> I will probably idle again in OE community
May 31 08:48:01 <bluelightning> hrw: suihkulokki: done
May 31 08:48:07 <hrw> cool
May 31 09:43:37 <tsjsieb> Goodmorning
May 31 09:45:35 <tsjsieb> Is it possible to create from an image recipes 2 ubi images (1 for rootfs and 2 for 'user partition' with only some files in it)?
May 31 09:45:59 <hrw> anything is possible - question how much code you want to write for it
May 31 09:46:54 <tsjsieb> more trying to find out what would be a 'standard approach' to do this. I guess more people want to use it in such a way
May 31 09:48:39 <tsjsieb> If it's an option to just have a second image recipe with only some files in it (no complete rootfs) than thats fine too, but I haven't find yet if that is possible
May 31 10:24:46 <mckoan> good morning
May 31 10:26:01 <mckoan> in my image I get an error when launch an exec
May 31 10:26:06 <mckoan> Unable to load library icui18n "Cannot load library icui18n: (icui18n: cannot open shared object file: No such file or directory)" 
May 31 10:26:17 <mckoan> which recipe provides icui18n?
May 31 10:36:47 <pb_> icu
May 31 12:21:52 <tasslehoff> Is it sane to use openembedded just to create a toolchain I can use to develop for cortex m3?
May 31 12:23:45 <mckoan> tasslehoff: why not. OTOH you can use a code sourcey one
May 31 12:24:27 <mckoan> In core-minimal-image there is not libstdc++, which package provides it?
May 31 12:30:33 <bluelightning> mckoan: funnily enough it's called "libstdc++" :)
May 31 12:36:13 <mckoan> bluelightning: sorry, I searched a recipe with that name without success
May 31 12:36:43 <bluelightning> it comes from the gcc recipe I think
May 31 12:41:05 <mckoan> bluelightning: it works (of course) thank you
May 31 12:43:53 <hrw> gcc-runtime iir
May 31 12:43:54 <hrw> c
May 31 12:52:10 <mckoan> does exist into generated image a timestamp  identifying the image which generated the fs itself?
May 31 12:52:18 <mckoan> something like /etv/version
May 31 12:53:51 <mckoan> I have /etc/timestamp:201305311239 but minimal-image-20130531123838.rootfs.tar.bz2
May 31 13:02:55 <mckoan> if I launch bitbake minimal-image which variable identify the "minimal-image" string?
May 31 13:03:33 <mckoan> maybe {PN} 
May 31 13:13:49 <bluelightning> ${PN} yes
May 31 13:15:44 <{md}> you can see all the variables which hold it by looking at the output of `bitbake -e core-image-minimal`, piping through grep is a good idea though `bitbake -e core-image-minimal | grep '"core-image-minimal"'`
May 31 13:16:22 <mckoan> {md}: I did, thx
May 31 13:23:18 <mckoan> bluelightning: actually {PN} gives you the name of the recipe inside it is called
May 31 13:24:06 <mckoan> I thought to modify base-files do_install() creting my image-timestamp
May 31 13:24:46 <mckoan> so I got /etc/VERSIONA-base-files instead of /etc/VERSIONA-minimal-image
May 31 13:25:43 <bluelightning> mckoan: you can't do that sort of thing from a recipe installed into the image like base-files
May 31 13:26:06 <bluelightning> mckoan: probably you should do it using a shell function called from ROOTFS_POSTPROCESS_COMMAND
May 31 13:29:39 <mckoan> bluelightning: ok, I could emulate the behaviour of ROOTFS_POSTPROCESS_COMMAND += "remove_packaging_data_files ; " but do I need to add my function into a class?
May 31 13:30:32 <bluelightning> mckoan: you can, or it can be added to your custom image recipe
May 31 13:44:57 <hrw> libpci depends on pciutils which depends on pciutils-ids which adds ~200KB even to core-image-minimal
May 31 13:45:01 <hrw> similar with usbutils
May 31 13:49:57 <bluelightning> hrw: presumably there is an optimisation to be made there
May 31 13:50:51 <hrw> also ldconfig can be dropped 
May 31 13:50:56 <hrw> another 630KB
May 31 13:50:57 <bluelightning> hrw: there is a pci DISTRO_FEATURES feature, does that not affect that?
May 31 13:51:16 <hrw> bluelightning: what if you want pci feature?
May 31 13:51:41 <bluelightning> hrw: why would you drop pci* if you did?
May 31 13:51:45 <hrw> core-image-minimal is useful for kvm tests
May 31 13:52:03 <hrw> but you do not need names for lspci/lsusb names in such case
May 31 13:52:57 <bluelightning> the question would be how do you ensure lspci/lsusb work in other images?
May 31 13:53:12 <bluelightning> assuming it's true that you don't want these in core-image-minimal
May 31 13:53:54 <bluelightning> udev depends on *-ids btw
May 31 13:53:59 <bluelightning> not sure what it needs them fore
May 31 13:54:00 <bluelightning> for
May 31 13:54:31 <hrw> found!
May 31 13:54:36 <hrw> it is udev fault
May 31 13:54:42 <hrw> udev recommends pciutils-ids
May 31 13:54:48 <hrw> and usbutils-ids
May 31 13:55:03 <hrw> ops, you wrote that too ;D
May 31 13:55:20 <bluelightning> assuming udev can work without them, BAD_RECOMMENDATIONS could be used here, although that only works for ipk
May 31 13:55:21 <hrw> looks like case for BAD_RECOMMMENDATIONS or sth like that
May 31 13:55:32 <bluelightning> ditto again :)
May 31 13:55:37 <hrw> udev should not need them
May 31 13:57:11 <hrw> systemd 197 does not even have pci.ids or usb.ids mentioned in code - only in makefiles/datafiles
May 31 14:01:47 <bluelightning> hmm...
May 31 14:02:32 <hrw> will send patch for discussion
May 31 14:10:56 <mckoan> bluelightning: seems more difficult than expected
May 31 14:11:08 <mckoan> I added ROOTFS_POSTPROCESS_COMMAND += "add-image-timestamp ; "
May 31 14:11:23 <mckoan> add-image-timestamp() { printf "${DISTRO_NAME} " >> ${D}${sysconfdir}/VERSION-${PN} }
May 31 14:11:27 <hrw> sent
May 31 14:14:24 <bluelightning> mckoan: that's what I was thinking... except it's not really a timestamp
May 31 14:19:57 <mckoan> bluelightning: run.do_rootfs.8865: line 249: add-image-timestamp: command not found
May 31 14:20:17 <mckoan> add-image-timestamp() is in the image.bb file
May 31 14:21:22 <kergoth> mckoan: - isn't a valid character in a shell function
May 31 14:21:51 <kergoth> afaik, anyway
May 31 14:36:38 <mckoan> kergoth: I replaced with underscore, but the problem was a missing fakeroot before add_image_timestamp() function declaration
May 31 14:41:10 <mckoan> kergoth: actually the fakeroot was not necessary, underscore solved the issue
May 31 14:54:56 <n|west> I'm trying to understand the populate_sysroot command and when/where TOOLCHAIN_HOST_TASK[_append] is required. Any good references? Just about everything I've found on teh googol is from mailing lists about very specific things
May 31 14:57:00 <n|west> err sorry, populate_sdk command!
May 31 14:57:41 <tgall> is there a way screen out a package from being built or figure out where in the dependency chain a package gets pulled in ?
May 31 14:58:06 <n|west> bitbake -g shows dependency graph (in dot format)
May 31 14:58:24 <tgall> I've got a broken package that I don't really need in my image so really would like to just avoid it somehow
May 31 14:58:51 <n|west> also BBMASK variable can hide recipes
May 31 14:59:47 <tgall> awesome
May 31 15:00:43 <kergoth> if something is being pulled into an image, then it's being pulled in by a dependency, and masking out what it depends on will just cause bitbake to abort before the build even begins :)
May 31 15:01:13 <tgall> yeah need to get to the bottom of the nature of the dependency 
May 31 15:01:19 <tgall> the package in question is perf
May 31 15:01:29 <tgall> and it's blowing chunks on aarch64 
May 31 15:01:33 <n|west> true, so bitbake -g <image-name> will show you how it gets included
May 31 15:01:46 <n|west> it creates a few different .dot files you can look at
May 31 15:02:46 <n|west> now any tips on populate_sdk documentation for moi? ;-)
May 31 15:05:58 <n|west> specifically I'm looking to know if I have to include in my custom image some TOOLCHAIN_HOST_TASK_APPEND to get some libs does that mean I should be patching up some recipes or is it a normal thing to have to do?
May 31 21:19:45 <n|west_> I'm having trouble getting swig-native in my SDK. It's a DEPENDS of a package I'm trying to work on (gnuradio), but for some reason doesn't get put in the sdk. Listing swig-native in TOOLCHAIN_HOST_TASK results in opkg_install_cmd: Cannot install package swig-native.
Jun 01 02:08:00 <nerdboy> hei
Jun 01 02:12:37 * nerdboy hands signal11 a soldering gun
**** ENDING LOGGING AT Sat Jun 01 02:59:58 2013