**** BEGIN LOGGING AT Fri Apr 19 02:59:58 2013
Apr 19 05:02:19 <mranostay> hi Jefro
Apr 19 08:39:44 <bluelightning> morning all#
Apr 19 09:39:42 <david-e_> while I can successfully "bitbake core-image-minimal" from scratch, it is not clear to me if you can do it using a pre-built toolchain, thus avoiding to recompile it
Apr 19 10:22:38 <RP> david-e_: prebuilt from sstate or using an external toolchain?
Apr 19 10:22:47 <RP> david-e_: you can do both
Apr 19 10:26:12 <Saur> RP: Did someone screw what creating the release tar ball for 8.0.1, or did you consciously change the layout of it? In poky-danny-8.0.tar.bz2 there is a toplevel "poky-danny-8.0/" directory. However, in poky-danny-8.0.1.tar.bz2 there is no toplevel directory matching the name of the tar ball. Rather the contents is directly at the top...
Apr 19 10:26:27 <Saur> *what => up
Apr 19 10:26:37 <RP> Saur: I suspect someone screwed up
Apr 19 10:26:49 <RP> Saur: could you file a bug for that please?
Apr 19 10:27:06 <Saur> RP: Will do (after lunch)
Apr 19 11:14:09 <david-e_> RP: from sstate I guess: something like compiling it once and installing on a build machine so that others can use it
Apr 19 11:19:49 <Saur> RP: Bug submitted: https://bugzilla.yoctoproject.org/show_bug.cgi?id=4351
Apr 19 11:19:50 <yocti> Bug 4351: normal, Undecided, ---, jeffrey.osier-mixon, NEW , No top level directory in tar ball for Yocto release 1.3.1
Apr 19 11:20:03 <RP> david-e_: just share the sstate cache directory between the machines
Apr 19 11:20:20 <RP> Saur: thanks!
Apr 19 11:21:19 <david-e_> RP: not top prio at the moment, but I'll certainly look at it. Any docs/tutorials/examples I can take as a starting point?
Apr 19 11:23:09 <RP> david-e_: its covered in local.conf.sample, also more details in the manual http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#shared-state-cache
Apr 19 11:23:20 <david-e_> RP: cheers mate
Apr 19 14:31:25 <errordeveloper> hey
Apr 19 14:32:02 <errordeveloper> what would be the reason for getting packages built with -dev and -dbg
Apr 19 14:32:12 <errordeveloper> but no base package being built?
Apr 19 14:32:32 <errordeveloper> i had ${PN}_dev.bb at the start
Apr 19 14:32:37 <errordeveloper> sorry
Apr 19 14:32:45 <errordeveloper> that was ${PN}_git.bb
Apr 19 14:32:52 <errordeveloper> and it didn't set PV
Apr 19 14:33:16 <errordeveloper> then I tried setting PV = "+git${SRCPV}"
Apr 19 14:33:23 <errordeveloper> but that broke it
Apr 19 14:33:48 <errordeveloper> then I tried to pretend that it's ${PN}_0.0.1.bb
Apr 19 14:34:07 <errordeveloper> and removed previous ${PN}_git.bb
Apr 19 14:34:26 <errordeveloper> but still cannot get it to produce the base package
Apr 19 14:34:36 <errordeveloper> only -dev and -dbg get emitted
Apr 19 14:35:22 <errordeveloper> although, I had successfully built an image with the original ${PN}_git.bb which didn't set PVa
Apr 19 14:35:35 <errordeveloper> ah, I'm quite confused now
Apr 19 14:35:56 <bluelightning> errordeveloper: hmm, maybe leave it as _git and set PV = "0.0.1+git${SRCPV}" ?
Apr 19 14:41:21 <errordeveloper> bluelightning: tried that, didn't work
Apr 19 14:41:47 <errordeveloper> let me try again ...
Apr 19 14:44:31 <errordeveloper> ah, I can see one funny bug
Apr 19 14:53:20 <errordeveloper> was about to gist the actuall recipes, but figured I had a very silly thing there
Apr 19 14:53:42 <errordeveloper> yey, fixed now!!
Apr 19 14:53:42 <RP> errordeveloper: is this some kind of library called xxx.so?
Apr 19 14:54:12 <errordeveloper> no, I just had a attempt at having x.inc and x-header.inc
Apr 19 14:54:28 <RP> errordeveloper: glad you got it sorted...
Apr 19 14:54:42 <errordeveloper> which was a bullshit idea
Apr 19 14:57:05 <errordeveloper> so okay, package is built and image is built
Apr 19 14:57:40 <errordeveloper> but I now see the consecuqnce of the warning message I have asked about here: http://stackoverflow.com/questions/16090550/building-python-packages
Apr 19 14:57:56 <errordeveloper> while bitbake only gave me a warning
Apr 19 14:58:06 <errordeveloper> i get an error from opkg
Apr 19 14:59:57 <errordeveloper> i update the SO post now, cause that seems like a potential pitfal ...
Apr 19 15:00:09 <errordeveloper> any clues?
Apr 19 15:04:23 <rburton> errordeveloper: what did you hope to achieve with that that do_install_append fragment?
Apr 19 15:05:02 <rburton> because it's removing /usr/share/share/, which seems a bit odd
Apr 19 15:05:42 <rburton> you want something like rm -f ${D}${libdir}/python*/site-packages/site.py* to delete the overlapping files
Apr 19 15:05:50 <rburton> (python-requests shipping them is also wrong)
Apr 19 15:14:57 <errordeveloper> rburton: no, i actually don't use that one
Apr 19 15:15:15 <errordeveloper> rburton: I see ...
Apr 19 15:16:16 <rburton> afaik, site.py shouldn't be shipped
Apr 19 15:16:25 <errordeveloper> the do_install_append() was in pycurl
Apr 19 15:16:35 <errordeveloper> rburton: ah, so ok then
Apr 19 15:22:32 <errordeveloper> rburton: so I'm guessing it's a kind of a bug in distutils class
Apr 19 15:22:50 <errordeveloper> cause it's supposed to take care of these kind of thing
Apr 19 15:23:03 <rburton> errordeveloper: yeah it probably should ensure that file doesn't get packaged
Apr 19 15:28:33 <errordeveloper> rburton: so you are quite sure that removing site.py will not actually break anything?
Apr 19 15:30:03 <rburton> errordeveloper: an optional module shouldn't be shipping it, python itself should
Apr 19 15:30:12 <rburton> so delete it from your package
Apr 19 15:30:41 <errordeveloper> ok
Apr 19 15:35:18 <mcfrisk> sigh, I guess Ubuntu with encrypted home direcotry (ecryptfs) can not be used to build yocto since filenames are limited to 144 characters, sstate filenames can be longer..
Apr 19 15:35:51 <RP> mcfrisk: don't build on cryptfs, its a world of pain, not just that
Apr 19 15:37:05 <mcfrisk> I've never had issues with encrypted lvm, performance is 10% worse but on a laptop I'm not that picky, but this ecryptfs thingy is too much.
Apr 19 15:37:43 <RP> lvm may be ok, I know ecyrptfs isn't
Apr 19 15:37:49 <rburton> can't have long filenames?!
Apr 19 15:38:36 <rburton> man, failtastic
Apr 19 15:38:42 <errordeveloper> yeah, that's really #wtf
Apr 19 15:39:23 <errordeveloper> oh it's called "The enterprise cryptographic filesystem for Linux
Apr 19 15:39:27 <errordeveloper> "
Apr 19 15:39:38 * errordeveloper loves all things enterprise
Apr 19 15:40:58 <mcfrisk> yep, gory details at https://bugs.launchpad.net/ecryptfs/+bug/344878 and http://unix.stackexchange.com/questions/32795/what-is-the-maximum-allowed-filename-and-folder-size-with-ecryptfs
Apr 19 15:41:42 <rburton> that SX answer isn't entirely accurate though
Apr 19 15:41:54 <rburton> the guy talking about normal vs cjk is all confused
Apr 19 16:10:27 <errordeveloper> hm, so this doesn't work as expected:
Apr 19 16:10:27 <errordeveloper> do_install_append() { rm -f ${D}${libdir}/python*/site-packages/site.py[oc]
Apr 19 16:10:31 <errordeveloper> }
Apr 19 16:10:44 <errordeveloper> there is a new line after {
Apr 19 16:17:35 <rburton> you'll want to drop the site.py too?
Apr 19 16:20:24 <errordeveloper> ah, i sort of see what you mean, this doesn't actually match for site.py
Apr 19 16:20:25 <errordeveloper> ...
Apr 19 16:20:48 <rburton> *.py* will
Apr 19 16:21:32 <errordeveloper> ah, sure ... but actually it was wrong path
Apr 19 16:22:50 <errordeveloper> hm, there are /usr/lib/python2.7/site.py* and /usr/lib/python2.7/site-packages/site.py*  by the looks of it ...
Apr 19 16:23:42 <errordeveloper> another thing, how do you force package rebuild properly?
Apr 19 16:26:07 <errordeveloper> ah, I see it's `bitbake -c cleansstate` and then `bitbake -c package`
Apr 19 16:28:21 <errordeveloper> _not really_
Apr 19 16:28:34 <errordeveloper> that didn't actually give the package
Apr 19 16:28:57 <errordeveloper> i was doing find . -iname '*python[-_]requests*' | xargs rm -fr
Apr 19 16:29:03 <errordeveloper> but that's a bit nasty
Apr 19 16:33:27 <otavio> Hello
Apr 19 16:33:55 <otavio> Could someone help me with hob? How can I add a fstype in advanced settings
Apr 19 16:33:56 <otavio> ?
Apr 19 16:39:38 <otavio> bluelightning: ^
Apr 19 16:40:28 <bluelightning> otavio: it's controlled by IMAGE_FSTYPES I think
Apr 19 16:40:34 <bluelightning> er, I mean IMAGE_TYPES
Apr 19 16:56:44 <otavio> bluelightning: right
Apr 19 16:56:56 <otavio> bluelightning: it is not picking the sdcard image I added
Apr 19 16:57:08 <otavio> bluelightning: how it fetches this value?
Apr 19 16:57:46 <bluelightning> otavio: I'm not entirely sure, but the IMAGE_TYPES value is set in the image_types*.bbclass files and hob apparently reads that value
Apr 19 16:58:33 <bluelightning> for some reason I don't understand it does not show up in bitbake -e, and you can't extend it from local.conf
Apr 19 16:58:41 <bluelightning> (as an experiment)
Apr 19 16:59:13 <otavio> bluelightning: so it seems as a bug to me
Apr 19 16:59:18 <otavio> :)
Apr 19 16:59:34 <bluelightning> otavio: maybe... but does your image type class add to IMAGE_TYPES?
Apr 19 16:59:57 <feydrautha80> hello
Apr 19 17:00:01 <otavio> bluelightning: I think so
Apr 19 17:00:29 <otavio> bluelightning: IMAGE_TYPES? not defined
Apr 19 17:00:54 <bluelightning> otavio: then add to it with += in your class
Apr 19 17:02:44 <feydrautha80> is anybody using yocto/bitbake with jenkins ?
Apr 19 17:03:26 <bluelightning> feydrautha80: JaMa is
Apr 19 17:03:45 <otavio> bluelightning: doing it
Apr 19 17:04:49 <feydrautha80> do you know if he is "alive" more often than less ? :)
Apr 19 17:06:12 <JaMa> I'm right now, if the question is quick
Apr 19 17:07:26 <otavio> JaMa: do you have a place with the setup/scripts?
Apr 19 17:07:51 <otavio> JaMa: and how is your cycle to test things? use an internal GIT tree to push stuff for building?
Apr 19 17:18:35 <JaMa> I use the same scripts I'm using for shr
Apr 19 17:19:00 <JaMa> https://github.com/shr-distribution/shr-makefile/tree/jansa/test-all
Apr 19 17:27:12 <otavio> bluelightning: didn't work
Apr 19 17:30:16 <bluelightning> otavio: I would suggest filing a bug in that case
Apr 19 17:40:43 <errordeveloper> the distutils seems quite painfull
Apr 19 17:41:23 <errordeveloper> my packages seem to be missing some thing, need to figure it out
Apr 19 17:41:50 <errordeveloper> hopefully I will end-up submitting some pull-request to fix this :)
Apr 19 17:42:15 <errordeveloper> actually, any idea why python 3 is not in poky?
Apr 19 19:42:48 <errordeveloper> hm... are there logs for this channel?
Apr 19 19:45:45 <wmat> errordeveloper: http://logs.nslu2-linux.org/livelogs/yocto/
Apr 19 19:45:52 <errordeveloper> awesome!
Apr 19 19:47:34 <errordeveloper> ah, but it doesn't have stuff from today
Apr 19 19:48:18 <errordeveloper> I just wonder if anyone responded to me while I was offline
Apr 19 19:48:19 <wmat> errordeveloper: scroll down on the main listing, it's at the bottom
Apr 19 19:48:29 <wmat> http://logs.nslu2-linux.org/livelogs/yocto.txt
Apr 19 19:48:51 <errordeveloper> ah!
Apr 19 19:49:09 <errordeveloper> pretty sweet :)
Apr 19 19:49:58 <errordeveloper> ok, so nobody commented on distutils being sucky
Apr 19 19:52:02 <errordeveloper> i need to figure out why I only get one directory
Apr 19 19:52:03 <errordeveloper> cosm-demo-image-beaglebone/usr/lib/python2.7/site-packages/requests-1.2.0-py2.7.egg/
Apr 19 19:52:19 <errordeveloper> as in /usr/lib/python2.7/site-packages/requests-1.2.0-py2.7.egg/
Apr 19 19:52:33 <errordeveloper> and python doesn't seem to like it that way
Apr 19 19:53:57 <errordeveloper> so here what it's supposed to look like
Apr 19 19:53:57 <errordeveloper> https://www.archlinux.org/packages/community/any/python2-requests/files/
Apr 19 19:54:14 <errordeveloper> but I get just the egg dir ;9
Apr 19 19:54:47 <errordeveloper> ah ... may I just try to hack it
Apr 19 20:45:07 <errordeveloper> ok so replacing `inherit distutils` with `inherit setuptools` seems to work
Apr 19 20:45:35 <errordeveloper> setuptools is some tiny little class that passes different flags to destutils
Apr 19 20:45:46 <errordeveloper> i'm all confused with this python stuff now
Apr 19 20:46:04 <errordeveloper> anyhow, seems like I got the right dir structure now! wow!
Apr 19 23:37:41 <Circuitsoft> Hello - I'm getting an "Unable to resolve package <myapp>" error when I build.
Apr 19 23:38:03 <Circuitsoft> At the moment, <myapp>_0.1.bb only lists dependencies and doesn't build anything.
Apr 19 23:38:28 <Circuitsoft> Do I need to do something special on a dependency-only package?
Apr 19 23:52:12 <sgw1> Circuitsoft: do you list them as DEPENDS or RDEPENDS
Apr 19 23:52:25 <Circuitsoft> RDEPENDS
Apr 19 23:52:39 <Circuitsoft> RDEPENDS_${PN}
Apr 19 23:53:06 <sgw1> So are you trying to be like a packagegroup?
Apr 19 23:54:21 <Circuitsoft> For now. The app is written in python, and this is listing all the dependencies. At the moment it's easier to install it directly to the target than to go through Yocto.
Apr 19 23:54:35 <sgw1> Circuitsoft: ie, just creating a recipe to group other packages together?
Apr 19 23:54:40 <Circuitsoft> I do intend to have this recipe be the app, but not just yet.
Apr 19 23:55:16 <sgw1> you may want to mark it with ALLOW_EMPTY then
Apr 19 23:55:42 <sgw1> Circuitsoft: ALLOW_EMPTY = "1"
Apr 19 23:56:11 <Circuitsoft> Awesome. Thanks.
Apr 19 23:56:20 <sgw1> Circuitsoft: actually ALLOW_EMPTY_${PN} = "1" (I always for the ${PN}
**** ENDING LOGGING AT Sat Apr 20 02:59:57 2013