**** BEGIN LOGGING AT Thu Jan 30 02:59:58 2014 Jan 30 07:31:12 hi Jan 30 07:31:35 pb_: great idea with 'Richard Purdie' nametag ;D Jan 30 08:00:39 gm Jan 30 08:04:21 hi zecke Jan 30 08:05:11 hi JaMa , i didn't follow the Qt5 SDK discussions lately. i am seeing 2 patches from denix in master. does that mean it now works? anything missing? i am interested to get that in dora branch, if you confirm it's working in master, i can try to backport. Jan 30 08:08:22 ndec, fwiw, the only issue I have with meta-qt5 master is that the Qt5 .la files are broken. other than that, I can create an sdk and build qt5.2 apps for the target with it Jan 30 08:08:57 kroon: ah nice. what's broken with .la? Jan 30 08:11:25 ndec, somewith with the path where the libs should be installed, host path leaks into the .la files. the sed:ing that is going on to create the .la files are kind of tricky.. Jan 30 08:11:50 hmm. ok. i will play with that then. Jan 30 08:14:05 ndec, JaMa suggested we get a small test-case package to easily fix and handle regressions of these kind of problems in the future, so if you have any idea please shout Jan 30 08:14:25 ndec, requilding qtbase every time is a pain.. Jan 30 08:14:30 hm first armboard from amd Jan 30 08:14:54 woglinde: that's old news... from yesterday ;-) Jan 30 08:14:58 haha Jan 30 08:14:59 yes Jan 30 08:15:10 I did not read news yesterday Jan 30 08:15:21 might be the perfect platform to make arm native OE builds... Jan 30 08:16:02 hm Jan 30 08:16:10 tons of error will comes out Jan 30 08:17:17 kroon: you mean you want to generate the SDK and test building a Qt5 application with it? Jan 30 08:18:40 ndec, well.. somehow test that for instance the .la files are correct. but we would probably want to test .pro files (?) aswell. I dunno, I'm an autotools guy, not qmake project Jan 30 08:19:06 ok Jan 30 08:29:52 kroon: delete the .la files Jan 30 08:29:59 you should be still ok Jan 30 08:30:09 khem, yeah, that works. .la files are still broken though Jan 30 08:30:34 thats ok they dont serve any practical purpose anyway Jan 30 08:30:41 on a cross build env Jan 30 08:31:02 infact in my not so long TODO list there is one item where I want to nuke all .la files in OE Jan 30 08:32:35 khem, sounds good if that works :) Jan 30 08:33:13 khem, my desperate final solution was to locally patch the sdk bbclass to remove libQt5*.la files.. Jan 30 08:33:33 k Jan 30 08:40:52 hi khem Jan 30 08:41:00 woglinde: wie gehts Jan 30 08:41:08 khem fine fine Jan 30 08:41:35 I have to practice my german skills if I have to visit ELCE this year Jan 30 08:41:41 yeah Jan 30 08:41:46 we will finally meet Jan 30 08:41:48 nice Jan 30 08:52:23 ndec: aarch64 hardware is fast Jan 30 08:53:13 hrw: i have no doubt. that's why we need to make sure we can run OE builds on them Jan 30 08:53:22 to cross compile x86 stuff ;-) Jan 30 08:53:25 ndec: :) Jan 30 08:53:36 ndec: I may try it one day Jan 30 08:53:59 buy the new AMD boxes Jan 30 08:54:18 khem`: I have access to AMCC ones Jan 30 08:54:51 they are FPGAs isnt it Jan 30 08:55:16 khem`: no. silicon Jan 30 08:55:28 well, i can do that on a pandaboard ;-) Jan 30 08:55:51 ndec: panda has just 2 cores and 1GB of ram. Jan 30 08:56:00 not to mention lack of any usable storage ;d Jan 30 09:00:34 ndec: how long does gcc take to compile on pandaboard Jan 30 09:01:07 or webkit ;) Jan 30 09:03:23 khem`: chrome-browser needed 18h for linking final binary Jan 30 09:03:32 and 2.5GB ram Jan 30 09:03:44 s/ram/memory Jan 30 09:05:59 heh I thought so Jan 30 09:06:27 well AMD boxes are a) 64bit so will have more RAM so linker will be lot happier Jan 30 09:06:33 b 8 core CPU Jan 30 09:06:34 khem`: up to 128GB Jan 30 09:06:53 and may be we should use gold for linking that anyway Jan 30 09:07:11 that will bring it down to may minutes Jan 30 09:09:09 khem->sleep() Jan 30 09:09:30 How can i link OpenCV to Intel's TBB with OE? http://docs.opencv.org/doc/tutorials/introduction/crosscompilation/arm_crosscompile_with_cmake.html Based on this link, i added the -DWITH_TBB=ON and -DBUILD_TBB=ON to EXTREA_OECMAKE in bitbake recipe. The opencv libs do not link to TBB still. Jan 30 09:09:31 night khem Jan 30 09:09:53 khem`: gnight Jan 30 09:15:20 I am OE building armv7a on armv8a Jan 30 09:15:42 BUILD_SYS = "aarch64-linux" Jan 30 09:18:36 hrw *g* Jan 30 09:18:44 hrw whats redhat doing with oe? Jan 30 09:19:05 woglinde: nothing. I just started build to test how it works Jan 30 09:19:11 core-image-minimal Jan 30 09:19:36 fetching will take huge amount of time probably ;( Jan 30 09:19:59 * hrw hugs 250/20Mbps connection at home Jan 30 09:20:05 good morning Jan 30 09:21:20 hrw 250? Jan 30 09:21:37 I am only getting 2 mbit Jan 30 09:21:45 hi mckoan Jan 30 09:23:48 woglinde: yes, 250 Jan 30 09:24:20 hrw what connection is this cable, dsl, fiber to home? Jan 30 09:24:29 woglinde: cable Jan 30 09:24:38 okay Jan 30 09:25:06 woglinde: 20€ per month Jan 30 09:27:26 * tbr only gets 100/10 through in house VDSL Jan 30 09:27:52 I wish I could get cat5 to the basement, there is a fibre and I could get 1G/100 or such Jan 30 09:28:12 tbr: beer tomorrow? Jan 30 09:28:24 hrw: yes! Jan 30 09:28:37 tbr: what time you arrive? Jan 30 09:29:11 1340, SAS1591 from CPH Jan 30 09:29:41 decided to take an earlier flight and maybe have a chance for some sightseeing this time... Jan 30 09:30:01 dinner with AV500 and the trolls and then delirium Jan 30 09:30:14 08:35 for me Jan 30 09:30:54 hm I will land 21:55 Jan 30 09:31:59 woglinde: beer event ends at 0600 so you will manage Jan 30 09:32:58 no beer event is to crowded for me Jan 30 09:33:15 and too much smoking Jan 30 09:34:53 ok Jan 30 09:36:43 | /bin/ld: bin/pseudo: No symbol version section for versioned symbol `memcpy@GLIBC_2.0' Jan 30 09:36:47 | /bin/ld: final link failed: Nonrepresentable section on output Jan 30 09:36:49 | collect2: error: ld returned 1 exit status Jan 30 09:37:08 pseudo-native do_compile failed Jan 30 09:37:10 smoking happens outside nowadays Jan 30 09:37:14 but yes, it's crowded Jan 30 09:37:17 bluelightning: hi Jan 30 09:37:22 unless you manage to find a corner somewhere Jan 30 09:37:29 bluelightning: do not forget to take your nametag tomorrow Jan 30 09:37:37 but then you still need to convince someone to bring you beer ;) Jan 30 09:37:47 mmmm, nametag, good point! Jan 30 09:38:03 time to warm up the department printer Jan 30 09:38:37 tbr: 18:47 < bluelightning> I suppose I could find one of my old conference badges and wear that Jan 30 09:38:41 18:47 < bluelightning> I think that's what hrw does sometimes... Jan 30 09:38:43 18:47 < pb_> bluelightning: he wears your old badges? what a prankster. Jan 30 09:38:46 18:48 < bluelightning> pb_: heh, but who would want to be me? ;) Jan 30 09:39:06 :D Jan 30 09:39:08 morning all Jan 30 09:39:13 moaning Jan 30 09:39:13 18:49 < pb_> yeah, I guess the bigger risk is that you'd arrive at the oe stand and there would be 20 people there all wearing badges saying "Richard Purdie". Jan 30 09:39:14 ;) Jan 30 09:47:26 bl *g* Jan 30 09:47:41 hm right I will take an older badge with me too Jan 30 09:48:28 hrw finally 22:39 < RP> pb_: heh, I'm not sure why everyone would pretend to be me, that sounds rather scary Jan 30 09:49:35 ;D Jan 30 09:50:20 but faking his accent would be hard Jan 30 09:50:56 two years ago my wife told that Richard was easiest to understand when everyone was talking English Jan 30 09:51:36 eiY1quoo Jan 30 09:52:35 jackmitchell? Jan 30 09:53:19 woglinde: morning Jan 30 09:53:34 oh heh Jan 30 09:53:51 heh typing passwords into irc again ;) Jan 30 09:55:31 loool Jan 30 09:56:31 this is what happens when you IRC with caffeine Jan 30 09:56:38 s/with/without Jan 30 09:56:45 I'll just go get some and shut up Jan 30 10:06:33 are there known issue with GCC 4.8 and Linux 3.10? Jan 30 10:06:57 From time to time a process will get -821712832 on a read syscall Jan 30 10:07:06 (or other equally small values) Jan 30 10:15:23 koen: (how) can I check the Angstrom version on a system I have ssh-access to? Jan 30 10:16:08 tasslehoff /etc/issue? Jan 30 10:16:54 woglinde: yes. thanks :) Jan 30 10:17:45 good Jan 30 10:24:15 lsb_release -a Jan 30 10:33:47 kroon: cool. i bet you'll have a nice time Jan 30 10:43:20 morning all Jan 30 10:43:36 hi pb_ Jan 30 10:43:44 hi bluelightning Jan 30 10:46:18 crumbs, the fosdem schedule is a lot bigger than last time I was there. Jan 30 10:47:04 yeah, there are many, many parallel tracks Jan 30 10:47:04 hi pb Jan 30 10:47:15 but some tracks don't have that many talks in them Jan 30 10:47:40 yeah Jan 30 10:48:22 * pb_ trying to plan which talks to attend Jan 30 10:49:01 I guess I should make a cut down schedule with only the bits I'm interested in for ease of understanding Jan 30 10:49:17 * woglinde will attend some java talks Jan 30 10:49:43 need to make sure that I don't accidentally end up in the ada devroom for example Jan 30 10:50:02 or java Jan 30 10:50:04 :-} Jan 30 10:50:27 actually, some of the java talks do sound quite interesting. Jan 30 10:50:58 yeah the garabge collector of course Jan 30 10:51:21 pb_: if you have an android phone, use the app :) Jan 30 10:51:41 you can star talks you're interested in, then get a cut-down list of those Jan 30 10:52:11 oh, right. I do have an android phone, I should try that. Jan 30 11:02:11 pb_: I cut to 42 entries Jan 30 11:04:20 * tbr needs to start selecting those, has the app for weeks now Jan 30 11:05:15 hm, I wish they wouldn't schedule talks at 9am on sunday Jan 30 11:05:24 the chances of me actually getting to those seems to be low or nonexistent Jan 30 11:05:58 I loaded ics to thunderbird and did mass removal of anything that looked not interesting Jan 30 11:06:23 final ics imported as 'fosdem 2014' calendar to the phone Jan 30 11:06:41 hrw what phone you have now? Jan 30 11:06:48 woglinde: nexus 5 Jan 30 11:07:08 btw - someone want galaxy s4? 350€ Jan 30 11:07:48 hrw: the one from the titzen contest? Jan 30 11:08:14 tbr: ja Jan 30 11:08:31 hehe Jan 30 11:09:00 tbr: samsung phones are not for me Jan 30 11:09:17 no surprise there, I've been avoiding those Jan 30 11:15:55 people are crazy... friend just shared instructions how to order <25GBP from amazon.co.uk without paying for posting. Jan 30 11:21:40 is there any magic to fake os.uname in OE? Jan 30 11:22:05 and if not, what's the preferred replacement? Jan 30 11:22:16 * koen stabs scons Jan 30 13:27:09 * pb_ tax return completed! Jan 30 13:27:14 a day early this year, hurrah Jan 30 13:31:03 bluelightning: there seem to be 3 fosdem apps for android. is one of them in particular recommended? Jan 30 13:31:28 hmm, I didn't notice that... Jan 30 13:31:33 I installed it last year though Jan 30 13:32:19 "FOSDEM Companion", "fosdem-android" and "Giggity" Jan 30 13:32:33 I think it's fosdem-android that I'm using, FWIW Jan 30 13:32:39 it's this one: https://play.google.com/store/apps/details?id=org.fosdem Jan 30 13:32:48 righto, thanks Jan 30 13:38:04 getting errors when using populate_sdk...has someone seen this Jan 30 13:38:42 any particular errors? Jan 30 13:44:42 pb hehe gratz on tax Jan 30 13:44:58 I did the last years in may or june Jan 30 13:46:34 getting errors when using populate_sdk...has someone seen this Jan 30 13:46:58 ARNING: Checksum failure encountered with download of http://alioth.debian.org/frs/download.php/file/3648/chrpath-0.14.tar.gz - will attempt other sources if available WARNING: Renaming /home/ridish/RDK-RI-NICO-2/build-qemux86/downloads/chrpath-0.14.tar.gz to /home/ridish/RDK-RI-NICO-2/build-qemux86/downloads/chrpath-0.14.tar.gz_bad-checksum_c4fbd0a8dc7c3d3302b915dfee2b3489 NOTE: Applying patch 'relocatable_sdk.patch' (../meta-linaro/meta Jan 30 13:47:27 NOTE: Applying patch 'relocatable_sdk.patch' (../meta-linaro/meta-linaro/recipes-devtools/qemu/files/relocatable_sdk.patch) ERROR: Command Error: exit status: 1 Output: Applying patch relocatable_sdk.patch can't find file to patch at input line 13 Perhaps you used the wrong -p or --strip option? The text leading up to this was: Jan 30 13:47:54 did any one got this kind of error? Jan 30 13:48:08 while doing poupulate_sdk Jan 30 13:48:40 meta-linaro? I guess you should take it up with those dudes. Jan 30 13:49:25 k Jan 30 13:50:20 it looks some patch applying issue? Jan 30 13:50:29 it seems Jan 30 13:50:46 if I had to guess, I would say that you have changed the version of qemu that you're using and linaro have not yet caught up with it. Jan 30 13:51:53 we freshly created the build... Jan 30 13:52:13 it is failing when i do populate_sdk Jan 30 13:52:26 right, you said that already Jan 30 13:52:26 but full build is compiling fine Jan 30 13:53:09 "full build" meaning what? sdk, target, both? Jan 30 13:53:29 that failure is nothing to do with "-c populate_sdk" per se. I imagine it would happen if you did a regular build of qemu-nativesdk. Jan 30 13:54:01 target Jan 30 13:55:43 okay, so, that's probably unsurprising since a target build probably doesn't include qemu at all, let alone nativesdk. Jan 30 13:56:04 the patch that's failing was dropped from oe-core back in November when qemu was updated to 1.6.1. Jan 30 13:57:34 I don't really know why qemu is in meta-linaro at all so there is presumably some linaro-specific craziness going on there. I think you probably need to either debug this yourself or talk to the linaro h4x0rs. Jan 30 13:59:06 The text leading up to this was: -------------------------- |Upstream-Status: Inappropriate [SDK specific] | |In order to be able to change the dynamic loader path when relocating |binaries, the interp section has to be made big enough to accomodate |the new path (4096 is the maximum path length in Linux). | |Signed-off-by: Laurentiu Palcu | |Index: qemu-1.4.0/ldscripts/i386.ld |================================ Jan 30 13:59:56 --- qemu-1.4.0.orig/ldscripts/i386.ld 2013-02-15 15:05:35.000000000 -0800 |+++ qemu-1.4.0/ldscripts/i386.ld 2013-02-28 22:55:36.138816418 -0800 -------------------------- No file to patch. Skipping patch. 1 out of 1 hunk ignored can't find file to patch at input line 26 Jan 30 14:00:29 pb -c populate_sdk should not used anyway Jan 30 14:02:21 to get toolchain... Jan 30 14:02:47 which command we have to use to get toolchain? Jan 30 14:03:33 sorry meeting Jan 30 14:04:06 woglinde: "should not _be_ used anyway"? Jan 30 14:18:43 qemu-1.4.0 folder is not available in build folder Jan 30 14:19:00 what might be the reason? Jan 30 14:19:19 you're building a version of qemu other than 1.4.0? Jan 30 14:36:19 khem: ping? Jan 30 15:03:06 hi zecke Jan 30 16:48:35 right then everyone, I might not get chance to IRC again anytime soon, so for those of you who are going, see you at FOSDEM! Jan 30 16:56:46 gm all Jan 30 16:58:22 khem`: morning. somebody pointed me to this: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58854. At least in dora this doesn't appear to be backported? Can you confirm that it should be included? Jan 30 17:01:06 yes this should be included Jan 30 17:01:43 4.8.3 is due soon although we might not be able to include it in 1.6 release Jan 30 17:04:30 zecke: if you were using linaro gcc you wont have this problem since the patch is applied there to 4.8 Jan 30 17:07:30 khem`: I am using dora. So it doesn't appear to use the linaro gcc. :( Jan 30 17:28:36 khem`: which layer do I need for the linaro toolchain? Jan 30 17:30:16 meta-linaro-toolchain Jan 30 17:30:25 morning Jan 30 17:30:32 or simple use angstrom-2012.13 Jan 30 18:09:54 how are we doing for power strips at FOSDEM? Jan 30 18:14:20 Crofton: if florian brings the box I think we should be OK Jan 30 21:39:22 * mr_science gets out his red pen and headphones **** ENDING LOGGING AT Fri Jan 31 02:59:58 2014