**** BEGIN LOGGING AT Tue Oct 13 02:59:59 2015
Oct 13 07:05:44 <bluelightning> morning all
Oct 13 07:06:00 <torqueware> Hello everyone.
Oct 13 07:07:52 <torqueware> I'm having a bit of trouble building java for a Freescale architecture. I'm looking for guidance. I'm using angstrom, and I've joined #angstrom to look for help as well, but I don't know if anyone is on this late.
Oct 13 07:09:01 <bencoh> it's not that late (9am) in europe ;)
Oct 13 07:10:48 <torqueware> : ) good to hear
Oct 13 07:11:12 <torqueware> (I'm in California - being a bit of a night owl)
Oct 13 07:11:44 <torqueware> Should I post a log?
Oct 13 07:12:08 <bencoh> I guess so, someone might be able to help
Oct 13 07:13:56 <torqueware> Alright then! Well, for starters, I'm working on reproducing the problem on my end.
Oct 13 07:14:27 <torqueware> Specifically, I'm having trouble with jamvm-initial-native segfaulting
Oct 13 07:17:14 <torqueware> Build machine is Fedora 22 (8Gb ram, x86 dual core hyperthreaded), and the build target is a freescale p4080ds (ppc)
Oct 13 07:20:24 <torqueware> should I post logs to pastebin?
Oct 13 07:23:22 <torqueware> http://pastebin.com/HqJtPbgD
Oct 13 07:23:46 <torqueware> Line 326 is where the offending ecj-initial segfaults
Oct 13 07:28:56 <torqueware> http://pastebin.com/Cyhm0f2r
Oct 13 07:29:32 <torqueware> Line 2909 is where ecj-initial shows that it is segfaulting in classpath-initial's config.log
Oct 13 07:32:08 <mckoan> good morning
Oct 13 07:32:30 <torqueware> Good morning!
Oct 13 07:33:19 <maxin> torqueware: similar problem reported there : https://github.com/woglinde/meta-java/issues/49
Oct 13 07:34:52 <maxin> torqueware: It doesnt happen while building in ubuntu .. "could" be gcc version
Oct 13 07:34:52 <torqueware> Thank you maxin! I read that thread, and I've tried switching to the native openjdk jvm to continue the build process (one of the workarounds in the thread). No joy so far.
Oct 13 07:35:17 <torqueware> hmmmm, alright! let me grab my gcc version
Oct 13 07:35:19 <maxin> torqueware: ah..
Oct 13 07:36:00 <torqueware> switching jvms is a really janky workaround, I didn't honestly expect it to work.
Oct 13 07:36:31 <torqueware> what worries me is that issue was closed awhile ago
Oct 13 07:37:03 <torqueware> I have gcc version 5.1.1 20150618 (Red Hat 5.1.1-4) (GCC)
Oct 13 07:37:14 <torqueware> Checking the thread
Oct 13 07:38:16 <torqueware> grabbing gcc 4.7
Oct 13 07:41:41 <torqueware> silly question, how much can a gcc version difference affect a oe build - in your experience?
Oct 13 07:43:01 <torqueware> (I'm curious, I've never cross compiled for an embedded system outside the openwrt project. So, I'm pretty green.)
Oct 13 07:43:02 <maxin> torqueware: I have experienced weird kernel bugs in arm arch with some gcc versions..
Oct 13 07:44:35 <torqueware> thank you! will keep you posted about the gcc switcharoo. Do I specify the gcc version in conf/local.conf?
Oct 13 07:44:46 <maxin> torqueware:  https://lkml.org/lkml/2014/10/10/272
Oct 13 07:46:35 <maxin> torqueware:  yes.. GCCVERSION="4.7%"  or similar
Oct 13 07:47:16 <torqueware> sure hope i dont have to build gcc 4.7 myself
Oct 13 07:47:29 <bluelightning> that's for the cross compiler though - it's the host compiler involved here surely?
Oct 13 07:48:44 <torqueware> I believe so
Oct 13 07:48:49 <torqueware> yes, I'm certain
Oct 13 07:48:58 <torqueware> the sysroot is x86_64
Oct 13 07:49:11 <torqueware> so the natives are crashing - build with my gcc
Oct 13 07:49:15 <torqueware> built*
Oct 13 07:50:45 <torqueware> jamvm-initial-native is a native binary built to cross compile openjdk-7-jre for the target platform
Oct 13 08:01:57 <torqueware> looks like I have to build it
Oct 13 08:01:59 <torqueware> *sigh*
Oct 13 08:11:06 <t0mmy____> exit
Oct 13 08:11:10 <t0mmy____> exit
Oct 13 08:12:56 * torqueware twiddles thumbs while building gcc
Oct 13 08:17:21 <torqueware> Another question: I know the dtb file is a way to pass device information to the kernel, but how do I know the dtb file I build is correct? The device I'm building for has a p4080 processor in it, but it's a SBC built by a manufacturer. Could I use their provided dtb to have access to all the boards devices?
Oct 13 08:19:50 <maxin> torqueware: may be decompile the dtb files and diff those files ?  
Oct 13 08:20:31 <torqueware> good idea - but are they human readable files? I thought dtbs were binary blobs
Oct 13 08:20:39 <torqueware> ill try it now
Oct 13 08:20:44 <torqueware> ill let you know what I find
Oct 13 08:22:26 <maxin> torqueware: dts files are "sort of" readable.. depends on the documentation of SBC ..
Oct 13 08:23:18 <LetoThe2nd> torqueware: dtbs can be easily decompiled, just have a short look at the dtcs man/help text
Oct 13 08:23:34 <torqueware> thanks!!
Oct 13 08:36:46 <torqueware> could i run strings on the dtb files and compare the output? would that be sane?
Oct 13 08:37:03 <LetoThe2nd> why not just disassemble + read/diff?
Oct 13 08:39:45 <Jin^eLD> does Paul Barker use IRC, can't remember his nick?
Oct 13 08:41:18 <LetoThe2nd> Jin^eLD: tried typing "/who" and reading the list?
Oct 13 08:41:25 <Jin^eLD> yes I did actually
Oct 13 08:42:11 <torqueware> don't know how to disassemble (green)
Oct 13 08:42:12 <LetoThe2nd> \o/ if there's nothing hidden there, then no idea.
Oct 13 08:42:13 <maxin> torqueware:  dtc -I dtb -O dts -o <devicetree name>.dts <devicetree name>.dtb
Oct 13 08:42:41 <Jin^eLD> info in who is sparse and I did not want to /whois everyone for more details, figured maybe someone just knows him :)
Oct 13 08:43:07 <LetoThe2nd> torqueware: plus, just googling "dtb disassembly" would also have given a lot of hints.
Oct 13 08:43:16 <torqueware> reading up on dtc. I really appreciate the guidance maxin, Leto, everyone
Oct 13 08:43:17 <Jin^eLD> trying to figure out whats up with opkg and discuss some future patches
Oct 13 08:43:33 <Jin^eLD> the readme says ml is on google, but I think google groups got shut down
Oct 13 08:44:21 <LetoThe2nd> Jin^eLD: i saw him posting stuff to the ml, so he's certainly active.
Oct 13 08:45:35 <Jin^eLD> I guess he is, I'm just trying to figure out where to discuss opkg issues
Oct 13 08:45:53 <LetoThe2nd> Jin^eLD: if in doubt i'd suggest starting on the ml
Oct 13 08:45:55 <LetoThe2nd> http://blog.gmane.org/gmane.comp.handhelds.openembedded
Oct 13 08:46:30 <Jin^eLD> for some reason I never get a confirmation when trying to subscribe to the oe dev ml
Oct 13 08:52:14 <Jin^eLD> hey rburton
Oct 13 08:52:31 <Jin^eLD> from IRC logs I see that you were looking for Paul Barker on IRC :)
Oct 13 08:52:35 <Jin^eLD> did you manage to find out his nickname?
Oct 13 08:52:37 <rburton> hi Jin^eLD
Oct 13 08:52:40 <rburton> no, sorry
Oct 13 08:52:47 <Jin^eLD> :)
Oct 13 08:53:02 <Jin^eLD> are you also doing stuff with opkg?
Oct 13 08:53:05 <rburton> that was likely before he stopped being the opkg maintainer too - if you want to talk opkg there's a new maintainer
Oct 13 08:53:16 <Jin^eLD> oh ok
Oct 13 08:53:17 <Jin^eLD> who is the new one?
Oct 13 08:54:03 <rburton> +  * Alejandro del Castillo <alejandro.delcastillo@ni.com>
Oct 13 08:54:11 <rburton> says opkg AUTHORS file
Oct 13 08:54:33 <Jin^eLD> but they did not move away from google groups yet? 
Oct 13 08:54:44 <Jin^eLD> afaik google groups is being shut down so I was not sure where to write
Oct 13 08:55:27 <bluelightning> google groups being shut down?
Oct 13 08:55:28 <rburton> its on git.yoctoproject.org now
Oct 13 08:55:44 <bluelightning> google code is, but google groups would be the first I've heard...
Oct 13 08:55:46 <Jin^eLD> bluelightning: is it not? I believe reading that?
Oct 13 08:55:47 <Jin^eLD> aaah
Oct 13 08:55:54 <Jin^eLD> ok so I mixed up stuff then
Oct 13 08:56:01 <Jin^eLD> :)
Oct 13 09:03:01 <namra> # The hostname can be changed outside of this recipe by using
Oct 13 09:03:03 <namra> # hostname_pn-base-files = "my-host-name".
Oct 13 09:03:14 <namra> where? in a recipe or in a conf file?
Oct 13 09:09:34 <XorA> namra: local.conf or the such
Oct 13 09:25:08 <bluelightning> preferably your custom distro config
Oct 13 09:25:18 <bluelightning> local.conf will work temporarily though
Oct 13 09:33:52 <namra> XorA: yup that works
Oct 13 09:34:25 <namra> bluelightning: thanks for the advice, certainly will implement a custom distro config. but still to novice
Oct 13 09:35:55 <belen> Crofton, bluelightning (an anybody else for that matter): stand submission for FOSDEM here http://piratepad.nl/ahWPO0iG9b Let me know if you want to change anything
Oct 13 09:37:31 <bluelightning> belen: looks good to me FWIW
Oct 13 09:37:55 <belen> bluelightning: thanks! Let's see what Crofton says
Oct 13 09:38:00 <bluelightning> good that they've moved on from using email
Oct 13 09:38:43 <belen> bluelightning: yes, I think so to. I am also hoping they have replaced pentabarf (chuckles), but I might be hoping in vain
Oct 13 09:39:25 <bluelightning> pentabarf... it's 5 times sicker than the alternative (TM)
Oct 13 09:39:48 <belen> bluelightning: it really is ;)
Oct 13 11:19:25 <Crofton|work> oh I must read my email
Oct 13 11:22:35 <XorA> Crofton|work: the rest of us let the NSA do that these days, saves so much time
Oct 13 11:22:45 <Crofton|work> heh
Oct 13 12:37:01 <mago_> can anyone point me to an example opkg config that works with opkg-cl in the sdk? i just can't get it to run, either it tries to write stuff in my host root fs (like /sysroot), or it complains about something else. I'm trying to follow this example: https://www.yoctoproject.org/docs/current/adt-manual/adt-manual.html#configuring-the-pms
Oct 13 12:54:12 <mckoan> otavio: ping
Oct 13 12:54:20 <otavio> mckoan: hello
Oct 13 12:54:40 <mckoan> otavio: hi, just a quick question
Oct 13 12:55:23 <mckoan> otavio: IIRC if you use Fido you MUST use kernel 3.14.28 because 3.10.17 doesn't support vivante?
Oct 13 12:56:21 <otavio> mckoan: yes
Oct 13 12:56:27 <mckoan> otavio: or in any case 3.10.17 is incompatible with Fido?
Oct 13 12:56:31 <otavio> or 3.10.53
Oct 13 12:57:41 <mckoan> otavio: having to move my patches from 3.10.17, qould you suggest to use 4.x ?
Oct 13 12:57:52 <mckoan> s/qould/would
Oct 13 12:58:23 <otavio> mckoan: 4.x won't offer gpu-viv
Oct 13 12:58:35 <otavio> mckoan: well, not soon
Oct 13 12:59:02 <mckoan> otavio: ar ELCE was told there is a new GPL driver for that
Oct 13 12:59:18 <otavio> mckoan: yes, etnaviv
Oct 13 12:59:49 <otavio> mckoan: I did not try it myself but I would love if someone could integrate it and send patches
Oct 13 13:00:48 <mckoan> otavio: not sure I can afford it, but I could spend a bit of time on it next week
Oct 13 13:01:04 <mckoan> otavio: thank you
Oct 13 13:01:08 <otavio> mckoan: yw
Oct 13 13:22:24 <joshuagl> iirc etnaviv hasn't even landed in staging yet
Oct 13 13:26:44 <otavio> joshuagl: yes, it did not
Oct 13 13:27:00 <otavio> joshuagl: it is still very experimental afaik and has some features lacking
Oct 13 13:27:12 <otavio> joshuagl: however I would love to have it as an option
Oct 13 13:36:15 <spaszPL> otavio, JaMa which recipe should I compile to have a QtNetwork class in my embedded image?
Oct 13 13:36:51 <spaszPL> Is is qtconnectivity?
Oct 13 13:38:28 <otavio> spaszPL: IIRC it is part of core
Oct 13 13:39:21 <spaszPL> so it's qtbase recipe? Should i change the packageconfig ?
Oct 13 13:43:03 <AndersD> spaszPL, no, it's part of the default build from qtbase
Oct 13 13:45:58 <spaszPL> ok, thx a lot
Oct 13 13:46:03 <spaszPL> :)
Oct 13 15:21:38 <ndec> hi, in a .conf file, i need to do something like that FOO += "${@if MYVAR == 1} blabla". I am not sure what syntax to use..
Oct 13 15:24:28 <bluelightning> ndec: FOO += "${@'blabla' if d.getVar('MYVAR', True) == '1' else ''}"
Oct 13 15:32:48 <ndec> bluelightning: hmm. ok. thx..
Oct 13 15:33:24 <kergoth> it's just python :) see the python documentation for details
Oct 13 15:33:27 * kergoth yawns
Oct 13 15:38:43 <rburton> in the past there was an idiom to do a test and then use the implict coercion to int as an index into an array
Oct 13 15:38:54 * kergoth twitches
Oct 13 15:39:00 <rburton> yeah that was hideous
Oct 13 15:39:21 <rburton> [if enabled, if disabled][MYVAR==1]
Oct 13 15:39:25 <rburton> no, other way around
Oct 13 15:39:27 <rburton> see, hideous
Oct 13 15:39:45 <bluelightning> yes, we probably still have one or two of those
Oct 13 15:39:56 <kergoth> ah, it was added in 2.5, always wondered when that came in - https://mail.python.org/pipermail/python-dev/2005-September/056846.html
**** ENDING LOGGING AT Wed Oct 14 02:59:59 2015