**** BEGIN LOGGING AT Fri Jun 28 02:59:58 2013
Jun 28 07:51:11 <mckoan> good morning
Jun 28 07:58:21 <florian> hi mckoan 
Jun 28 08:53:40 <juhe> Hi, I have a question about a *.bbappend files. I have a problem with including a file into bbappend using include "features.inc". Bitbake says that features.inc is not a bitbake file although it contains just plain assignment to FEATURE variable. Is it possible to include files from bbappend files? Any help would be greatly appreciated.
Jun 28 09:13:03 <bluelightning> morning all
Jun 28 09:32:58 <silviof> morning all
Jun 28 09:43:28 <mckoan> hi florian, bluelightning, silviof, all
Jun 28 09:44:18 <bluelightning> hi mckoan, silviof
Jun 28 10:18:20 <ant_work> bluelightning: JaMa: so what's the plan for PR / PRINC in meta-openembedded layers?
Jun 28 10:21:12 <JaMa> I haven't seen any plan
Jun 28 10:26:38 <bluelightning> I think the only choice is a flag day for the changeover
Jun 28 10:28:25 <JaMa> btw: that's right ML for meta-qt3 changes?
Jun 28 10:28:43 <JaMa> 2 are in oe-devel ML but I don't know who maintains meta-qt3
Jun 28 10:29:09 <JaMa> bluelightning: ah layerindex says it's you :)
Jun 28 10:29:25 <bluelightning> JaMa: er... yes :)
Jun 28 10:29:34 <ant_work> heh
Jun 28 10:29:43 <bluelightning> I think the correct ML is the yocto one
Jun 28 10:29:52 <bluelightning> but I can pick them up from there
Jun 28 10:30:25 <bluelightning> that layer doesn't see too many patches as you might guess
Jun 28 10:30:32 <ant_work> maybe on the next Yocto release we bump PE and get rid of PR stuff
Jun 28 10:30:54 <JaMa> I was just thinking if I should create new bundle for them or reply to submitter that it's wrong ML
Jun 28 10:30:58 <bluelightning> I don't think bumping PE really helps though if you have bbappends
Jun 28 10:31:21 <ant_work> ok, then at layer level ?
Jun 28 10:31:21 <JaMa> right, bumping PE isn't better then bumping PR
Jun 28 10:31:56 <ant_work> afais the issue is only for feeds
Jun 28 10:39:18 <bluelightning> JaMa: I think I'll just merge it now and reply for future reference
Jun 28 10:39:23 <bluelightning> JaMa: thanks for the reminder
Jun 28 10:39:54 <bluelightning> and tut tut, the README doesn't say where to send patches
Jun 28 10:40:05 <bluelightning> I'll fix that as well
Jun 28 11:16:23 <pb_> hi all
Jun 28 11:17:53 <bluelightning> hi pb_
Jun 28 11:33:47 <Crofton|work> it looks like our cat is letting a local cat in the house to eat her food
Jun 28 11:34:21 <broonie> That's nice and friendly.
Jun 28 11:34:48 <Crofton|work> yes
Jun 28 11:34:56 <Crofton|work> I have a cam on the cat door
Jun 28 11:35:11 <Crofton|work> and have video of them coming and going around the same time
Jun 28 12:14:37 <Crofton|work> ant_work, is it easy for foreigners to by SIM's in Italy?
Jun 28 12:15:05 <ant_work> well, I've asked someone, seems straightforward
Jun 28 12:15:17 <ant_work> still your ID will be registered
Jun 28 12:15:38 <ant_work> you can go in any shop and buy a 'rechargeable'
Jun 28 12:15:38 <Crofton|work> as in they copy my passport?
Jun 28 12:15:49 <Crofton|work> no special id needed
Jun 28 12:15:52 <ant_work> seems so
Jun 28 12:15:54 <Crofton|work> ok
Jun 28 12:16:03 <Crofton|work> I had seen references to needing a tax number
Jun 28 12:16:10 <Crofton|work> spain wanted a passport
Jun 28 12:16:13 <ant_work> I've asked the shop on the corner, dealing with tourists
Jun 28 12:16:27 <Crofton|work> Belgium mailed to the hotel from internet form :)
Jun 28 12:16:29 <Crofton|work> thanks
Jun 28 12:16:50 <ant_work> the gray men have to know who you're
Jun 28 12:16:56 <Crofton|work> heh
Jun 28 12:17:16 <ant_work> maybe you'll get some tablet for free those days...
Jun 28 12:17:17 <Crofton|work> they can match my IMSI to other records :)
Jun 28 12:20:00 <XorA> Crofton|work: use a GTA02 with IMSI 000000000000 :-D
Jun 28 12:20:31 <Crofton|work> rofl
Jun 28 12:20:36 <Crofton|work> that would be awesome to have
Jun 28 12:20:45 <XorA> there were about 400 of them I beleive
Jun 28 12:20:49 <XorA> I had 2
Jun 28 12:21:07 <XorA> of course the IMSI on GTA02 in in flash so you can just change it
Jun 28 12:22:00 * broonie keeps on meaning to look at getting his running.
Jun 28 12:24:11 * stefan_schmidt_w wonders if he can comment on that with techincal details. Don't havce my NDA around to look it up. ;)
Jun 28 12:24:26 <ant_work> Crofton|work: http://www.vodafone.it/engl/living/italy_pack.html
Jun 28 12:24:55 <XorA> stefan_schmidt_w: I lost the doc so I cant remember the AT commands
Jun 28 12:26:13 <Crofton|work> hmm, they talk about a fiscal code card
Jun 28 12:26:15 * bluelightning still has his GTA02 but the thing was so buggy it wasn't much use as a phone
Jun 28 12:26:17 <Crofton|work> weird
Jun 28 12:28:08 <stefan_schmidt_w> bluelightning: Don't tell me. I used it 6 months as my main phone. I know the pain. :)
Jun 28 12:28:27 <stefan_schmidt_w> XorA: I still should have all the docs somewhere. Don't know where so. :)
Jun 28 12:28:43 <XorA> I gave my GTA01/GTA02 devices to the local hacklab
Jun 28 12:28:48 <XorA> and my GTA03
Jun 28 12:29:14 <bluelightning> I have a GTA04 as well, haven't used it much but it does seem functional and impressive as an effort from a small team
Jun 28 12:29:26 <stefan_schmidt_w> I kept the gta02 for dfu-util testing
Jun 28 12:31:50 <XorA> its the only dfu device I ever saw :-D
Jun 28 12:32:33 <broonie> Everything else seems to have adopted the fastboot protocol instead.
Jun 28 12:34:38 <stefan_schmidt_w> Not really. Only if you think about the android world of devices.
Jun 28 12:34:49 <stefan_schmidt_w> Still tons of smaller devices that use dfu
Jun 28 12:35:14 <stefan_schmidt_w> The bluetooth usb chip in your Thinkpad might use it.
Jun 28 12:35:50 <stefan_schmidt_w> PHones or tablet or other "big" devices don't use it.
Jun 28 12:36:21 <stefan_schmidt_w> But smaller things do. Just seen it mentioned on some Sony site about a wrist watch, etc
Jun 28 12:36:53 <stefan_schmidt_w> And if we talk about number one have to see that Samsung is not using fastboot either.
Jun 28 12:37:06 <stefan_schmidt_w> At leats not when I look around in the office here.
Jun 28 12:37:42 <stefan_schmidt_w> But I think thats not OE related at all :)
Jun 28 12:37:42 <broonie> I've never actually seen any of these other devices using DFU.
Jun 28 12:37:58 <broonie> The Samsung stuff is their own custom protocol that predates Android.
Jun 28 12:38:25 <stefan_schmidt_w> And they also have one that comes after android and is not fastboot
Jun 28 12:38:43 <broonie> Oh, they changed? Nice :)
Jun 28 12:38:56 <stefan_schmidt_w> I don't think they changed
Jun 28 12:39:16 <stefan_schmidt_w> Can't comment on devices that are here compared to the one available in the market
Jun 28 12:40:09 <stefan_schmidt_w> Anyway, there are DFU devices. Just not stuff you guys work with it seems. Quite some microcontroller projects are using it.
Jun 28 12:40:28 <stefan_schmidt_w> Its better on smaller firmware images and chips with less power anyway
Jun 28 12:41:11 <stefan_schmidt_w> Flashing the gta02 was quite a pain already with 60 MB or so. Nowadays it would be even more painful. Spec was never written for such big amounts of data.
Jun 28 12:52:45 <ant_work> XorA: that iphone$ still have DFU
Jun 28 12:53:52 <XorA> ant_work: cant stand apple kit
Jun 28 12:54:45 <stefan_schmidt_w> ant_work: not by the spec though. Would be to easy. :)
Jun 28 13:51:02 <RP> I only have bits of GTA01 somewhere, not sure that was ever very functional :/
Jun 28 15:08:54 <mckoan|away> 
Jun 28 16:23:09 <jkroon_> Hi. I'm trying to build a cross gcc for arm, that runs on Win32, using the sysroot created by OpenEmbedded. Problem is when I try to compile an application using the resulting gcc, ld.exe complains that it "cannot represent machine arm:beaglebone". Anyone who knows whats wrong ?
Jun 28 16:24:04 <jkridner> jkroon_: sounds ambitious.  
Jun 28 16:24:31 <jkridner> does Windows have something like strace/
Jun 28 16:24:32 <jkridner> ?
Jun 28 16:24:38 <jkroon_> jkridner, why ambitious .. ?
Jun 28 16:24:56 <jkroon_> jkridner, dunno.. I'm testing it using wine
Jun 28 16:25:11 <jkridner> compiling anything for Windows is always difficult for me....
Jun 28 16:25:30 <jkridner> especially a compiler.
Jun 28 16:26:14 <jkroon_> yeah
Jun 28 16:26:40 <jkroon_> still, I just think I'm missing some linker-scripts or something, then it should work
Jun 28 17:40:28 <khem> pb_: the gihub sync for meta-openembedded repo is not happening 
Jun 28 17:41:17 <khem> pb_: well the push is happenin but it seems the repo its syncing from is not uptodate
Jun 28 17:52:41 <khem> pb_: never mind it seems to be ok 
Jun 28 17:54:51 <jkroon_> for some reason, all my linker-scripts contain "OUTPUT_ARCH(arm:beaglebone)", and my gcc doesnt understand that
Jun 28 17:56:44 <mr_science> jkroon_: maybe you're just lucky today?
Jun 28 17:57:06 <jkroon_> lucky ? :-p
Jun 28 17:57:56 <mr_science> i got nothin' better than sarcasm right now...
Jun 28 17:58:07 <mr_science> all my new stuff is at home
Jun 28 18:18:56 <jkroon_> oh. I think binutils uses "$MACHINE" from the environment. And I was setting it for OpenEmbedded
Jun 28 18:27:56 <jkroon_> yes. that was pretty sneaky :-p
Jun 28 18:37:35 <jkroon_> Using bitbake, and given the image bb recipe, is there a command I can give to populate the sysroot with files from all packages that is pulled in by the image ?
Jun 28 18:37:56 <fray> bitbake -c populate_sdk <image>
Jun 28 18:39:29 <jkroon_> doesnt that also create the SDK "package" ?
Jun 28 18:39:55 <fray> yes
Jun 28 18:40:05 <fray> end result is a full SDK..
Jun 28 18:40:14 <fray> based on the image vs a "meta-toolchain" recipe
Jun 28 18:41:52 <jkroon_> fray, yeah I guess I kind of want this, thanks
Jun 28 18:43:10 * fray explains to people at least once a week the difference between populate_sdka nd meta-toolchain... SDKs..
Jun 28 18:43:19 <fray> "generic image based" vs "highly targeted"
Jun 28 18:47:38 <jkroon_> fray, are there any major differences between the target system sysroot in tmp-eglibc/sysroots and the one shipped in the SDK ?
Jun 28 18:47:47 <fray> yes
Jun 28 18:48:14 <fray> the sysroot in the buildsystem is specific to the instances of the build system.. it will likely contain a lot of things that could build software that is incapable of running on the target..
Jun 28 18:48:38 <fray> many libraires and headers are produced that are never used in the software put into the image recipe..
Jun 28 18:48:55 <fray> so if you directly build against the build system sysroot, you can allow your SDK users to generate software that is broken..
Jun 28 18:49:20 <fray> the SDK on the otherhand constructs the SDK based only on the software that is available to the end user to develop against, to hopefully ensure that apps will run on the target image.
Jun 28 18:49:39 <fray> for the image recipe, it does this by simply adding in -dev packages to each of the things insalled.. (simplified, but thats basically what it does)
Jun 28 18:49:59 <fray> for the meta-toolchain (targeted SDK).. its up to the author of the recipe to state precisely what the app developer is allowed to use..
Jun 28 18:50:14 <fray> both are useful, it all depends on how much you want to hide from the application developer..
Jun 28 18:50:14 <jkroon_> i would guess that would be the same dev packages as the ones in the build sysroot, no.. ?
Jun 28 18:50:29 <fray> build sysroot has more then what was instaled onto the image..
Jun 28 18:50:53 <fray> i.e. if you build bash, you need ncurses..  ncurses has a number of libraries available to it.. but only one of them is needed to run bash.
Jun 28 18:51:01 <fray> so you build 5 libraries, your image only contains 1..
Jun 28 18:51:13 <jkroon_> fray, aha
Jun 28 18:51:19 <jkroon_> fray, makes sense
Jun 28 18:51:22 <fray> if you build against the sysroot, you have access to link to any of the 5.. if you build against the SDK.. you will only have access to link to the 1 that is available
Jun 28 18:51:50 <fray> and with the targeted SDK, you can say I may have ncurses, but it's forbidden for apps to use it.. only the system can.. so you just don't make it available at all
Jun 28 18:52:11 <fray> (thats pretty normal behavior for a GUI based SDK... where people are limited to well defined GUI interfaces)
Jun 28 18:52:25 <jkroon_> fray, so, im creating a gcc toolchain which shall run on win32, for creating an application that runs on my image. Im better of extracting the sysroot in the SDK package, than using the one in tmp-eglibc/sysroots, right ?
Jun 28 18:52:50 <fray> in that case, you need to generate the win32 toolchain externally.. since the build system can't..
Jun 28 18:53:07 <fray> you then want to generate an SDK (using either approach).. install it on a Linux machine, and then copy the sysroot from the SDK itself..
Jun 28 18:53:11 <jkroon_> fray, yeah, i just got it working, but ive been using the sysroot in tmp-eglibc
Jun 28 18:53:20 <fray> combine the Win32 and the extracted SDK stuff and you now have a windows sdk
Jun 28 18:53:34 <jkroon_> cool
Jun 28 18:53:53 <jkroon_> there really ought to be a recipe for that..
Jun 28 18:54:41 <fray> at this point people have avoided any Win involvement..  the future is to be able ot run a VM running oe/Yocto Project code..
Jun 28 18:54:54 <fray> and then simply doing your builds (controlled by the windows side) in the Linux VM..
Jun 28 18:55:18 <fray> initial prototyping show that is actually faster then using a windows cross compiler, due to windows and it's horrible CoW filesystem issues
Jun 28 18:56:11 <jkroon_> yeah.. i would guess windows not having a proper fork() not helping the performance
Jun 28 18:56:20 <fray> yup
Jun 28 18:57:12 <jkroon_> the company i work for i still using windows xp
Jun 28 18:57:22 <jkroon_> I want to use linux in our next project
Jun 28 18:57:26 <fray> tell them to invest in a free 'VM'.. ;)
Jun 28 18:57:37 <jkroon_> there arent any good ones..
Jun 28 18:57:52 <fray> they're still better then building with windows though... serious performance will improve
Jun 28 18:57:54 <jkroon_> Qemu doesnt support full virtualization on win32 AFAIK
Jun 28 18:58:03 <fray> There are others..
Jun 28 18:58:17 <jkroon_> which one are you thinking of ?
Jun 28 18:58:19 <fray> basically they're 'player' versions thata re free.. you just need to generate an image from somewhere..
Jun 28 18:58:29 * fray tries to remember I don't do much with that stuff anymore
Jun 28 18:58:35 <jkroon_> vmplayer is not free for commercial use
Jun 28 18:58:44 <jkroon_> which is a shame..
Jun 28 18:59:09 <jkroon_> i tried the one oracle owns... dont remember the name... but it didnt even startup :-(
Jun 28 18:59:15 <fray> https://www.yoctoproject.org/documentation/build-appliance-manual
Jun 28 18:59:24 <fray> YP recommends VM Ware player..
Jun 28 18:59:35 <fray> the oracle one is the other ones I've seen people use
Jun 28 19:00:34 <jkroon_> oh well, win32 version of gcc should work for us I hope
Jun 28 19:00:52 <fray> if you do that, then the right approach is to extract the sysroot from the SDK..
Jun 28 19:01:03 <jkroon_> will do. thanks for the info
Jun 28 19:01:12 <fray> that is how the company I work for ships the stuff to our unfortunate Win32 customers
Jun 28 19:01:48 <jkroon_> i find it pretty ironic that even the win32 gcc binary is compiled on a linux machine
Jun 28 19:02:34 <jkroon_> at least thats how I do it
Jun 28 19:02:45 <fray> ya, same here
Jun 28 19:03:26 <jkroon_> everyone is struggling with working around win32 limitations :-)
Jun 28 19:03:49 <jkroon_> and still almost all companys out there buy windows year after yeat
Jun 28 19:04:05 <jkroon_> that is fucked up
Jun 28 20:25:24 <mr_science> windows was never a platform intended for developers or engineers...
Jun 28 20:25:54 <mr_science> it's a marketing tool intended to make "normal" people dumb and dependent
Jun 29 01:28:19 <khem> Crofton|work: ah yes I see it :)
Jun 29 01:28:30 <khem> Crofton|work: not running OE yet ;)
Jun 29 01:29:49 <khem> we must have charged an arm and leg for it
Jun 29 02:10:40 <Crofton|work> khem, those guys get free shit :)
**** ENDING LOGGING AT Sat Jun 29 02:59:58 2013