**** BEGIN LOGGING AT Fri Oct 14 02:59:58 2016 Oct 14 06:33:58 koen: been a while since I asked this now, but do you recall the latest Angstrom version that could/should work on Beagle C3/C4? Oct 14 06:34:26 thought I was done with this old product... Oct 14 06:38:38 I guess my first question should be: should angstrom-v2013.06-yocto-1.4 compile on Ubuntu 16.04? Oct 14 06:44:46 tasslehoff: I fixed it a while ago Oct 14 06:44:54 tasslehoff: but it might have broken again Oct 14 06:45:03 tasslehoff: it has old thumb2 silicon Oct 14 06:49:53 koen: my current compile error is | gutils.c:(.text+0x70): multiple definition of `g_bit_storage' when compiling glib. Oct 14 06:50:15 I'll see if I'm missing some patches from upstream. Oct 14 07:44:46 Compilation of gmp-native fails. Can I force use of gcc-4.9 instead of gcc-5.0 when it uses my host compiler? Oct 14 07:47:42 tasslehoff: yes, set BUILD_CC etc. Only known to work reliably in master, but should be enough for gmp-native. alternatively fiddle your $PATH / symlink so that 'gcc' is gcc-4.9. Oct 14 07:48:56 hello berlin! Oct 14 07:49:50 where's the google doc? Oct 14 07:50:37 https://docs.google.com/document/d/1fDu0lw_BRh0fNXBByqM6OD_iiHwlsoA9D185Xpe1cUg/edit Oct 14 08:06:58 tlwoerner_: hello !berlin Oct 14 08:07:39 LetoThe2nd: it's very much !berlin here too ;-) Oct 14 08:08:41 tlwoerner_: i know :) it was menat like - you greeted us here in berlin, and i greeted you back in !berlin. Oct 14 08:09:06 LetoThe2nd: :-) Oct 14 08:10:42 wunderbar Oct 14 08:11:55 from the notes: "they took devtool and half of devtool" ? Oct 14 08:12:11 tlwoerner_: not sure what was being said there Oct 14 08:14:43 ok. so my problem is that an sdk built for 16.04 does not run so well on 14.04 (wrong libc.so.6 is the current issue). is there another approach than compiling a new sdk for 16.04? Oct 14 08:15:16 since the ancient oe-version the sdk was built from does not compile on 16.04 Oct 14 08:31:50 something just happened which has made the bridge fully useless, rather than just mostly useless Oct 14 08:35:12 https://wiki.yoctoproject.org/wiki/Cookbook Oct 14 08:41:50 joshuagl: hey, be happy. At least you have the bridge. I didn't even get that ;) Oct 14 08:42:57 belen: no? you're not an Ev member? Oct 14 08:43:20 joshuagl: o Oct 14 08:43:25 the quality of the zoom app/plugin wasn't great, so i dialed into the phone number (via gmail) and it works better for me Oct 14 08:43:33 joshuagl: that was supposed to be "no" :) Oct 14 08:45:25 belen: the minutes are being written in google docs in real time, that's how i'm following Oct 14 08:46:13 yes, a huge thanks to Jan who is the one mostly filling the google doc!! :-) Oct 14 08:46:28 \o/ Oct 14 08:46:33 dl9pf_: ^ Oct 14 08:47:31 rburton: yes, that's how I am following too Oct 14 08:53:02 i was on break too :-) Oct 14 08:53:26 thanks for moving the mic :-) Oct 14 09:30:27 shoragan, rocks! Oct 14 09:36:59 Did koen really just say he'd create a fork but contribute back and that you need a single repo for users? Oct 14 09:37:40 Yes, I did Oct 14 09:37:50 but a top level repo, not a combolayer thing Oct 14 09:38:29 if anyone is on he call and would like a demo of the stuff, let me know... I can setup a webex later and give another demo Oct 14 09:39:02 fray: the notes captured it pretty well, awesome demo, thanks! Oct 14 09:40:29 thanks Oct 14 09:40:50 fray: i'm looking forward to testing it out Oct 14 09:41:35 I do expect people to find lots of problems, coding issues, etc... Oct 14 09:41:37 fray: sounds really neat, I want to play with it! Oct 14 09:41:46 but I'd rather get it out early and fix issues that polish it and have to redo it later Oct 14 09:41:58 definitely, do that Oct 14 09:46:43 Crofton: +1 shoragan rocks (too many Jan's I had them confused) Oct 14 09:47:45 * LetoThe2nd hands shoragan some extr metal döts for the great job Oct 14 10:18:24 darknighte: link to my slides: https://goo.gl/qt3lKp Oct 14 10:18:34 #20 is the good/bad/ugly Oct 14 10:19:52 just dialled in, I can barely make out anything of what's being said :( Oct 14 10:20:26 that's better, thanks :) Oct 14 10:20:29 bluelightning: big room :-( Oct 14 10:20:37 bluelightning: jus tmoved the pickup Oct 14 10:20:40 just let us know, we're happy to move it Oct 14 10:21:36 bluelightning, you have the google doc Oct 14 10:21:38 ? Oct 14 10:21:49 Crofton: I haven't opened it yet but I should Oct 14 10:22:06 yes Oct 14 10:22:17 shoragan, is doing a good job keeping it up to date Oct 14 10:22:55 a very good job, much appreciated shoragan Oct 14 10:23:14 shoragan: +1 Oct 14 10:23:16 indeed. +1 shoragan Oct 14 10:24:03 sorry I missed the devtool discussion Oct 14 10:24:58 bluelightning, we understand people have lives :) Oct 14 10:25:12 We are sorry we can't make things work for everyone Oct 14 10:25:19 not your fault :) Oct 14 10:25:29 clearly we need quantum meetings Oct 14 10:26:07 bluelightning: we just gave you all the actino items :) Oct 14 10:26:22 darknighte: great, I love challenges :D Oct 14 10:26:50 having a hard time hearing Richard Oct 14 10:27:44 +1 for live meeting notes.. Oct 14 10:28:29 we will move phone in one sec.. I think conv is temporarily on otherside of room Oct 14 10:30:05 sound any better right now? Oct 14 10:30:24 (talking about clear linux , extended attributes, etc) Oct 14 10:31:05 I think folks are talking about things I did, but I can't follow it :-( Oct 14 10:31:13 * joshuagl clearly needs to check out the mender layer Oct 14 10:33:42 catching up on minutes now, re the discussion on site.conf i've an evil hack where meta-ross has a site.conf with my global assignments. Oct 14 10:37:29 no sound? Oct 14 10:37:43 oh... just me Oct 14 10:38:40 RAUC is worth considering as well https://github.com/jluebbe/rauc Oct 14 10:39:30 ok, I have to go right now so I guess that's it, but I'll read over the minutes later Oct 14 10:39:32 did me walking around help? Oct 14 10:39:35 we're breaking for lunch Oct 14 10:39:45 ok good.. off to lunch, back soon.. Oct 14 10:39:49 fray: yes, immensely! Oct 14 10:39:55 when are you back? Oct 14 10:43:59 fray: FYI I am reviewing that layer index branch - looks ok, there are a few very minor issues with the commit messages but I'll fix those when merging Oct 14 11:10:49 Anyone know what causes the "relocation R_X86_64_PC32 against protected symbol" error? Oct 14 11:13:05 please let us know when you're back from lunch and setup for the 2nd half of the OEDEM, will the Zoom meeting ID change? or is it the same? (578998354) Oct 14 11:15:50 Thought it was gcc version, but it happens with 4.8 as well. Oct 14 11:28:16 tasslehoff: link order bad? Oct 14 11:28:28 (especially with security flags or --as-needed) Oct 14 11:51:06 rburton: I wonder if this could be related https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65248 Oct 14 11:56:23 tlwoerner_, I assume someone warned you we are back ? Oct 14 11:56:36 zeddii: yes, thank you :-) Oct 14 11:56:50 zeddii: they did it on the google doc comments Oct 14 11:56:56 cool Oct 14 11:57:33 tlwoerner_ do I need to start walking with mic? Oct 14 11:58:00 oh, hadn't seen that folks are back Oct 14 11:58:08 fray: the mic is louder now, it's fine so far :-) Oct 14 11:58:21 ok, ping me if you want me to walk around.. person talking is currently very close Oct 14 11:59:04 fray: ok, thanks Oct 14 12:03:06 my opinion re py/perl distro features is that sounds like a vast amount of work for little gain. Oct 14 12:03:18 that is my concern as well.. Oct 14 12:03:53 if the goal was "don't have python in the image" then isn't there a way to abort a build if a specific package appears in the image? Oct 14 12:04:13 for "still a build dep", live with it. or add/use packageconfig to turn it off. Oct 14 12:04:40 ya Oct 14 12:05:29 uhoh snipes from irc go into the minutes ;) Oct 14 12:05:41 only limited.. feel free to change.. ;) Oct 14 12:05:49 fray announced your comments into the room rburton Oct 14 12:05:52 (since I read your comment) ;) Oct 14 12:05:54 he's some kind of IRC medium Oct 14 12:09:42 gnome is debating meson right now and their conclusion seemed to be that every new upstream you package needs patching meson, so it's very much in development still Oct 14 12:10:08 nice Oct 14 12:11:58 esoteric/new build systems are the ban of build people :-( Oct 14 12:12:14 s/ban/bane Oct 14 12:12:21 tlwoerner_: bane of all developers? Oct 14 12:12:26 we could ban them! Oct 14 12:12:40 +1 ban the bane! Oct 14 12:12:41 and all use emacs Oct 14 12:12:56 darknighte: +1 Oct 14 12:13:06 hippies Oct 14 12:13:27 I always knew I liked you zeddii! Oct 14 12:13:39 tlwoerner_: There are many build systems Oct 14 12:13:49 nih syndrome Oct 14 12:13:56 zeddii: Zealot Oct 14 12:14:20 tlwoerner_: There are many use cases. ;) Oct 14 12:14:24 hmm, should troll with -EDUPLICATEIRCCHANNEL Oct 14 12:14:24 zap-vi-users Oct 14 12:14:49 I'll defend my use of emacs to the death, with my crippled, tendonitis-ridden claws that used to be fists of fury Oct 14 12:15:43 zeddii: So many keystrokes to do simple things. ;) Oct 14 12:16:13 people saying emacs gives you tendonitus haven't used atom on mac Oct 14 12:16:18 ALL THE MODIFIERS Oct 14 12:16:27 Jefro: There are many editors. Oct 14 12:16:37 Jefro: angry bcs. of caffeine deprivation? Oct 14 12:16:51 i thought sugar intake during lunch was sufficient. Oct 14 12:17:19 * behanw is of the opinion that editors are personal things. Whatever makes you code the best is the right editor for you. Oct 14 12:17:23 emacs made me what I am! Oct 14 12:17:39 Far be it for me to tell you what your editor should be. Oct 14 12:17:56 Jefro: is that a warning? :-o Oct 14 12:18:27 joshuagl: ha.. more like a confession Oct 14 12:19:04 * behanw now knows how to DDOS Jefro. Remove emacs from his computer >:) Oct 14 12:19:26 rather computers (since DDOS) Oct 14 12:19:36 behanw: was just about to point that out. Oct 14 12:20:06 echo "rm -f $(which emacs)" >> ~/.bashrc Oct 14 12:20:21 bulldozer through your house -- need to make the universe a better place with a by-way Oct 14 12:20:45 behanw this is why I always keep a backup copy of emacs Oct 14 12:20:46 fray: you probably meant s/bulldozer/RP/ Oct 14 12:21:17 Jefro: Hence "which emacs". If it is in your $PATH, then it will be removed. ;) Oct 14 12:21:26 lol... na, RP doesn't hide the documents at the bottom of a drawer only visible at certain times a few glaxies away Oct 14 12:21:27 I always write my scripts to be self healing. Oct 14 12:21:52 fray: ++ Oct 14 12:24:39 fray: You forgot the sign which says "Beware of The Leopard" Oct 14 12:24:47 yes I did Oct 14 12:25:56 http://layers.openembedded.org/layerindex/branch/master/recipes/?q=numpy Oct 14 12:25:58 dafuq Oct 14 12:27:10 http://layers.openembedded.org/layerindex/branch/master/recipes/?q=nodejs Oct 14 12:27:36 *eyeroll* Oct 14 12:27:43 numpy is fine - the pypy recipes *are* different, and oe-core has py2 and py3 forms, so its just meta-redhawk-sdr with a old copy Oct 14 12:28:03 looks like about 50% of the nodejs copies are jama's problem Oct 14 12:28:41 Crofton: actually the nodejs thing is not as bad as it seems. at least half is only containing the string as part of the name, and actually is not nodejs itself Oct 14 12:28:54 redhawk is all my fault Oct 14 12:29:09 well, the overall project, not that specific Oct 14 12:29:12 case Oct 14 12:29:26 Crofton: and another 4 or 5 are intentionally providing specific versions in meta-nodejs by imyller. so AFAICS i see only like 2 replications, which is not too bad. Oct 14 12:31:00 http://layers.openembedded.org/layerindex/branch/master/recipes/?q=linux Oct 14 12:31:11 Hi! I sent a patch for daisy branch (an URL fix for python-cython) 2 weeks ago. I don't see it anymore on https://patchwork.openembedded.org/project/oe/patches/ ; Is it rejected or should I wait more? Oct 14 12:31:16 http://layers.openembedded.org/layerindex/branch/master/recipes/?q=u-boot Oct 14 12:58:38 Tartarus: I suppose we should talk here instead. ;) Oct 14 12:58:49 Oh fine :) Oct 14 12:59:02 av500 was getting grumpy Oct 14 13:00:22 he's just upset that bearborad doesn't run u-boot or oe :) Oct 14 13:00:34 RP: is there a reason why the timestamps are in zulu (?) time and not localized time? Oct 14 13:00:48 So, if we want oe-core to provide latest release, always, for U-Boot, it just needs to work on the qemu targets? Oct 14 13:01:13 I have a few HW platforms in my automated test farm Oct 14 13:01:21 And I'm trying to get more diversity in that Oct 14 13:01:36 koen is leaving. Time to volunteer him for $RANDOMSTUFF Oct 14 13:01:42 Adding more qemu is "easy" once I update the test hook framework to allow for that Oct 14 13:07:55 I guess you guys have moved on, please cc me on whatever comes out :) Thanks Oct 14 13:09:10 tlwoerner_: never noticed they were not local, the joy of working in the one true timezone Oct 14 13:12:11 rburton: haha! Oct 14 13:21:50 one TZ to rule them all Oct 14 13:24:00 TZ: You mean PDT right? Oct 14 13:24:40 LetoThe2nd, tlwoerner_ : ^^^ Oct 14 13:27:18 behanw: MT Oct 14 13:28:33 LetoThe2nd: And yet most of the world's tech companies/engineers are concentrated in PDT. :P Oct 14 13:29:17 Just wait for the big one when the West coast disappears... Then MT will be on the coast. Oct 14 13:30:09 behanw: MetalTime is not location based ;-) Oct 14 13:30:57 Tartarus: I was looking for a web reference for yocto ref boards, but can't find it. Oct 14 13:31:05 Should be minnowboard, beagleboard, edgerouter for mips, and the freescale mpc8315e-rdb Oct 14 13:31:58 behanw: https://layers.openembedded.org/layerindex/branch/master/layer/meta-yocto-bsp/ Oct 14 13:32:16 and click on "Machines" Oct 14 13:33:45 http://git.yoctoproject.org/cgit/cgit.cgi/meta-yocto/tree/meta-yocto-bsp/conf/machine?h=krogoth Oct 14 13:33:50 tlwoerner_: LOL. Oct 14 13:34:04 That too Oct 14 13:34:24 tlwoerner_: I was looking for something in the wiki or docs. Oct 14 13:35:17 behanw: OK. Of those, minnow isn't usually a U-Boot one (but can be) and I have to look into edgerouter Oct 14 13:35:22 I know the 8315e-rdb is a pita to get Oct 14 13:35:43 Tartarus: U-boot supports x86 right? ;) Oct 14 13:35:44 But if "yocto" can test on it for me, it ought to still work Oct 14 13:36:02 behanw: Yes. I just don't wanna wear out my SPI flash re-flashing it a few times a day Oct 14 13:36:06 Tartarus: I'd suggest testing on arm, mips and aarch64 and call it done. Oct 14 13:36:39 Tartarus: Test via pxe. Oct 14 13:36:46 Though that doesn't test the SPL I suppose. Oct 14 13:37:13 behanw: PXE doesn't integrate well enough with what I'm testing automatically Oct 14 13:42:05 Tartarus: wanna take up u-boot recipe and keep it up-to-date ? :) Oct 14 13:42:18 use a MAINTAINERs file! ;-) Oct 14 13:42:49 tlwoerner_: which one ? Oct 14 13:45:12 Marex: a top-level linux-kernel-style MAINTAINER file in a layer that describes who is responsible for what Oct 14 13:45:28 Tartarus: Pity. Oct 14 13:45:51 Marex: That's what I was hoping as well. Oct 14 13:48:52 hello all Oct 14 13:49:16 is there a way to check which encrypted packages are included in the build Oct 14 13:49:39 any utility or any kind of thing that can be used with yocto Oct 14 14:01:06 Noor (IRC): What do you mean by encrypted packages? Oct 14 14:04:19 bachp: the packages that uses encryption for example openssl Oct 14 14:08:31 Noor (IRC): I don't think there is any declaration of crypto. I assume you ask because of export regulations? Oct 14 14:09:31 thanks for the phone bridge!! :-D Oct 14 14:13:51 w00t Oct 14 14:17:44 fray: +1 Oct 14 14:17:52 bachp: sort of Oct 14 14:18:37 in our office management is asking for a list of packages which uses strong encryption Oct 14 14:18:59 I was googling it to find something which I can use but could not find anything Oct 14 14:25:04 Noor (IRC): Sorry I'm not aware of anything Oct 14 14:25:27 bachp: no problem :) Oct 14 15:03:05 Anyone have any experience with dpkg-native? The recipe appears to put Dpkg.pm in a directory that perlnative does not check, making dpkg-architecure and friends unusable. Oct 14 15:03:20 Unless I'm missing something. Oct 14 15:05:05 perlnative won't show as up 'perl' unless the one running it inherits perlnative, of course Oct 14 15:05:06 * kergoth yawns Oct 14 15:07:25 That's not the issue. Take dpkg-architecture for example, when called from the native sysroot it invokes perlnative. The issue is that it places Dpkg.pm in ${STAGING_DIR_NATIVE}/usr/lib/perl-native/perl. Oct 14 15:07:42 when perlnative is created, this path is not included as part of PERL5LIB Oct 14 15:08:19 You end up with an error: "Can't locate Dpkg.pm in @INC" Oct 14 15:08:37 huh, interesting Oct 14 15:08:55 sadly dpkg is a bit of a second class citizen in oe/yocto today, not a lot of folks using package_deb Oct 14 15:09:03 needs more love :) Oct 14 15:10:22 See http://cgit.openembedded.org/cgit.cgi/openembedded-core/tree/meta/recipes-devtools/perl/perl_5.22.1.bb?h=master Oct 14 15:10:33 PERL5LIB='$PERL5LIB:$OECORE_NATIVE_SYSROOT/${libdir_nativesdk}/perl/site_perl/${PV}:$OECORE_NATIVE_SYSROOT/${libdir_nativesdk}/perl/vendor_perl/${PV}:$OECORE_NATIVE_SYSROOT/${libdir_nativesdk}/perl/${PV}' Oct 14 15:11:30 ah, you're talking about problems in the sdk? Oct 14 15:12:12 whoops, wrong paste Oct 14 15:12:49 I meant http://cgit.openembedded.org/cgit.cgi/openembedded-core/tree/meta/recipes-devtools/perl/perl-native_5.22.1.bb?h=master Oct 14 15:13:02 create_wrapper ${D}${bindir}/perl PERL5LIB='$PERL5LIB:${STAGING_LIBDIR}/perl/site_perl/${PV}:${STAGING_LIBDIR}/perl/vendor_perl/${PV}:${STAGING_LIBDIR}/perl/${PV}' Oct 14 15:13:51 I suspect the sdk suffers the same issue Oct 14 15:14:57 We don't use package_db but a user of our meta layer had dpkg-native installed. Our build system uses dpkg-architecure, hence the issue. Oct 14 15:18:25 that sounds problematic in general in a cross-compilation environment, but beyond that, yeah, ti's certainly possible the library search paths don't match up, but we do manage to create root filesytems with dpkg-native, so i'd investigate why that works if that was the case. Oct 14 15:22:22 tlwoerner: from the notes: "they took devtool and half of devtool" Oct 14 15:22:31 that was half of bitbake, i fixed it now Oct 14 15:22:41 shoragan: ah, thanks! Oct 14 15:25:58 * RP wonders whether that bit or bake Oct 14 15:26:14 whether that was bit or bake Oct 14 15:26:24 my brain is fried... Oct 14 15:27:18 RP: Any objection to making an addtask referencing a non-existent task fatal, for 2.3? if not at parse time, at least at build time. right now you can addtask bar after do_some_invalid_function and it ignores that silently Oct 14 15:27:50 behanw, did you meet me about rauc at ELCE? or was that someone else? Oct 14 15:28:02 kergoth: no objection. Its hard as addtask can be before the function is declared but there are ways to deal with that Oct 14 15:28:52 yeah, i figured either at the end of parsing when we addtask from __BBTASKS, or later on when doing a build_dependencies Oct 14 15:28:55 tlwoerner, also thanks for fixing up my typos :) Oct 14 15:29:04 kergoth: right, makes sense Oct 14 15:29:35 k, thanks, will add to my list Oct 14 15:29:36 shoragan: haha, your notes were what was keeping me fully up-to-date with all the mumblings :-) Oct 14 15:30:18 :) Oct 14 15:30:29 i didn't want you to stop to fix typos, it was best you plough on ahead with the conversation while i follow behind, reading, and fixing up Oct 14 15:31:33 yes, that workflow usually works pretty well Oct 14 15:31:35 after lunch the mic seemed to pickup more of the conversation, before lunch i was only getting about 60% of the conversation Oct 14 17:14:11 * vmeson suspects that they took the 'bike' and left the rest 'tba' -- ha, ha! Oct 14 19:43:04 shoragan: no not me. But I suspect you maybe spoke to mdp (Matt Porter from Konsulko) since he told me about rauc Oct 14 21:16:50 what's rauc? Oct 14 22:44:43 tlwoerner: https://github.com/jluebbe/rauc Oct 14 22:45:57 rauc - Robust Auto-Update Controller **** ENDING LOGGING AT Sat Oct 15 02:59:58 2016