**** BEGIN LOGGING AT Tue Nov 17 02:59:58 2015 Nov 17 06:49:45 hi Nov 17 08:38:38 JaMa: I've seen you're samba patch on oe-devel which seems on top of mine - do you have a dedicated branch for them to let me work on? Nov 17 08:52:55 good morning Nov 17 08:56:44 sno: "master" :) Nov 17 08:57:06 sno: your and mine patch was merged, there are few patches on ML which aren't in master-next Nov 17 08:57:35 sno: I think at least one of your talloc patches is one of them, rebasing it on top of master-next and resending would be highly appreciated Nov 17 08:58:00 JaMa: I need to wash my eyes, I didn't see it yesterday in master nor master-next during scanning ^^ Nov 17 08:58:21 I found a typo in my samba-patch and rebased -i it - have to redo ;) Nov 17 08:58:37 JaMa: any opionion about talloc vs. libtalloc? Nov 17 08:58:51 1b2a08a samba: fix dependencies and QA issues Nov 17 08:59:32 no strong opinion about recipe name, but that PACKAGing looks wrong, so it should be fixed one way or another Nov 17 08:59:48 renaming recipe and leaving package rename to debian.bbclass looks very reasonable solution Nov 17 08:59:54 do you have a hint (I think, talloc and tevent should be done similar) Nov 17 09:00:22 has anyone experience this kind of perl issue: SSL connect attempt failed error:14090086:SSL routines:ssl3_get_server_certificate:certificate verify failed? i'm not sure about my configuration, i added libio-socket-ssl-perl & libnet-ssleay-perl. it seems that im missing something to verify the certificates on my raspberrypi Nov 17 09:00:32 JaMa: I grep for examples of debian.bbclass Nov 17 09:00:43 it's included by default in distro config Nov 17 09:00:48 you don't need to worry about it Nov 17 09:01:00 I have to know how to use ;) Nov 17 09:01:00 it does the right thing (TM) automagically on background Nov 17 09:01:16 I renamed the recipe and it still seems wrong Nov 17 09:01:22 you probably use it already Nov 17 09:01:32 unless you override DISTRO_INHERITS Nov 17 09:01:45 check bitbake -e and log.do_package for talloc Nov 17 09:02:22 that's what I mean - is there an example where it's done right to have something to compare with? Nov 17 09:02:27 but there is nothing special about using it, all you can find is "LEAD_SONAME" used for packages with more than one library included Nov 17 09:04:39 I'd like to understand what's wrong with my talloc patch Nov 17 09:17:17 any idea? :( Nov 17 09:18:42 drou: I'm using only meta-cpan - so no :/ Nov 17 09:19:12 drou: looks like you're Net::SSLeay is missing something Nov 17 09:20:57 i use the libnet-ssleay-perl from meta-perl actually, it should work with this one. Nov 17 09:21:10 drou: realized http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#migration-1.8-ssl ? Nov 17 09:21:37 drou: I have no clue about meta-perl - it has to many outdated modules, as said, I rely on meta-cpan Nov 17 09:23:11 it has it's own kind of problems - but there I know how to solve Nov 17 09:23:24 i'll take a lot at the meta-cpan then Nov 17 09:23:33 it's closer at the way the perl module authors want's there module Nov 17 09:23:50 you need jethro or master for that (required poky patches weren't in fido) Nov 17 09:24:06 for fido I can give you a patchset Nov 17 09:24:33 but drou, your problem looks like an issue in ssl, not Net::SSLeay Nov 17 09:26:11 sno: first i'll try to enable SSL 3.0 to see if it helps Nov 17 09:26:17 thanks Nov 17 09:33:07 another question, is there any way to force the rebuild of a dependant package. for example, i have a recipe called toast which depends on toast-native, is there a way, when i do a bitbake toast, to force toast-native to be built from scratch (even re-fetched)? Nov 17 09:33:37 why would you want to do that? Nov 17 09:35:48 not sure if SIGGEN_EXCLUDE_SAFE_RECIPE_DEPS is the proper way to do it Nov 17 09:40:31 joshuagl: because since i use ${AUTOREV} to fetch sources in my -native recipe, the code may have changed and can have an impact on the build of the 'toast' build Nov 17 09:48:08 yeah, that makes sense - but I would have expected an AUTOREV recipe to rebuild each bitbake invocation anyway? Nov 17 10:49:11 anyone know whether newly opened programs in matchbox will get focus? Nov 17 11:02:02 maybe rburton knows, if he don't mind me asking :) Nov 17 11:16:02 yes they will Nov 17 11:16:08 its a standard window manager in that respect Nov 17 13:18:02 hey, I am getting this: NOTE: Runtime target 'RDEPENDS_${PN}-core' is unbuildable, removing... -> I have this in my recipe: RDEPENDS_${PN} = "RDEPENDS_${PN}-core" Nov 17 13:20:19 I am probably missing this? http://www.yoctoproject.org/docs/2.0/ref-manual/ref-manual.html#var-PACKAGES Nov 17 13:30:09 also, is there a way to specify optional dependencies? Nov 17 13:30:19 what happened to bluelightning? Vacation? :) Nov 17 13:32:19 lpapp__: using overrides, like RDEPENDS_append_machine = "package" Nov 17 13:35:42 lpapp__: moved to the other side of the planet Nov 17 13:35:51 US? Nov 17 13:36:57 joseppc: do I really need to make it arch dependent or can I get along without hardcoding that? Nov 17 13:43:41 lpapp__: you can use whatever you have in OVERRIDES... Nov 17 13:46:45 joseppc: hmm, I do not get it. Nov 17 13:58:41 lpapp__: what exactly are you trying to accopmlish with that rdepends? copy the rdeps from one package to another? if that's the case, you need to acutally reference the variable. ${RDEPENDS_${PN}-core}, RDEPENDS_${PN}-core will make it look for a nonexistent 'RDEPENDS_${PN}-core' package, which is what that error is indicating Nov 17 13:58:54 i.e. you forgot the ${} around it Nov 17 14:03:07 hmm, thank you very much! Nov 17 14:03:55 or perhaps just use ${PN}-core? Nov 17 14:05:41 if you want the PN package to depend on the PN-core package, rather than what the PN-core package depends on, yes, that'd be correct Nov 17 14:06:42 ah, ok, thanks. Nov 17 14:06:47 makes a lot of sense now. Nov 17 14:07:24 np Nov 17 14:10:37 Hi, Nov 17 14:10:37 i want to install libpcsclite.so.1 with the recipe pcsc-lite. Nov 17 14:10:37 Somehow it gets build, it lies in the sysroot, but it wont get installed on the target. Nov 17 14:12:49 The recipe seems to split the package with: Nov 17 14:12:49 PACKAGES =+ "${PN}-lib" Nov 17 14:12:59 and Nov 17 14:12:59 FILES_${PN}-lib = "${libdir}/lib*${SOLIBS}" Nov 17 14:13:49 but i dont know what to do with that. bitbake pcsc-lite-lib doesnt work :/ Nov 17 14:14:49 recipes can emit any number of packages Nov 17 14:14:56 but bitbake accepts a recipe name, not a package name, on its commandline Nov 17 14:15:07 bitbake pcsc-lite will result in a pcsc-lite-lib binary package being built Nov 17 14:15:14 you'd want to add pcsc-lite-lib to your IMAGE_INSTALL or CORE_IMAGE_EXTRA_INSTALL to install it into your rootfs/image Nov 17 14:15:26 building reicpes doesn't magically put them into your filesystem Nov 17 14:16:27 fray: around? Nov 17 14:17:52 Thanks! Nov 17 14:19:20 np Nov 17 14:26:46 hmm, I had a package. Nov 17 14:27:09 I split it up into that package name with less content and put the content removed from there into further small packages depending on this one. Nov 17 14:27:34 but now when I try to upgrade the software, I am getting a conflict that some new split packages would like to provide the same as the monolithic before. Nov 17 14:27:42 What is the recommended way to resolve this package splitting? Nov 17 14:30:09 simply RPROVIDES or something? Nov 17 14:31:31 i think this is where RREPLACES is best used. not sure if oyu'd also need RCONFLICTS/RPROVIDES or not Nov 17 14:31:59 rconflicts would uninstall the other, but if the old big package gets uninstalled upon installing just part of it, then you'd lose content, so nto sure that wuld be ideal Nov 17 14:32:22 probably each component would RREPLACES, then the main one that pulls in all the individual pieces could RCONFLICTS? Nov 17 14:32:24 * kergoth shrugs Nov 17 14:41:45 Hi Guys, I'm working with a freescale bsp, how would I go about adding elfutils to the meta-toolchain recipe? Nov 17 14:43:46 Or how would I find out if any of the toolchains I can build has elfutils in it? Nov 17 14:45:51 TOOLCHAIN_TARGET_TASK += "whatevewr packages you want to add to the target sysroot in the toolchain" Nov 17 14:46:31 kergoth> that would be in local.conf? Nov 17 14:52:02 lpapp__: sorry I was caught in a meeting. I may have misled you a bit but it seems you've got a good answer in the end Nov 17 15:04:24 joseppc: yes, I did, thanks. Nov 17 15:14:10 rburton: https://github.com/kergoth/meta-mentor/commit/2125799 seem reasonable to you? just want to confirm that the behavioral change wasn't intentional Nov 17 15:17:56 meh, head is killing me today Nov 17 15:18:43 kergoth, thanks for noticing the typo. I've emailed halstead Nov 17 15:18:49 np Nov 17 15:47:07 Hey Yocto-Team, i want set up yocto for some guys in the university so i decided to use the sstate_mirror to accelerate the building process. This works pretty fine for me but every other user get the following error message: ls: cannot access /build/tmp/sysroots/zedboard-zynq7/pkgdata/runtime-reverse/warning:: No such file or directory. When they delete the sstate_mirror entry in the local.conf file yocto works fine for all other use Nov 17 15:47:07 r. Did anybody know the problem? Nov 17 15:58:04 kergoth: change wasn't intentional. didn't paul send that already? Nov 17 15:58:43 kergoth: (either way, yes, needs to be in jethro) Nov 17 16:01:52 JaMa: can you point me out where in https://github.com/rehsack/meta-openembedded/commit/4fca1c50fceb7a91c2d067475eb47510672856a2 is a PACKAGing mess? Nov 17 16:02:10 I tried to understand but didn't get the point Nov 17 16:48:55 is there a way with Yocto to reinforce rm_work? Nov 17 16:49:10 I accidentally left it in after debugging one particular build and started building an image... run out of space. Nov 17 16:49:19 I could surely remove the whole build directory, but then I would lose everything. Nov 17 16:49:54 there's a clean-workdirs script in the scripts/ dir in oe-core/poky to remove them after the fact Nov 17 16:50:09 more often than not i just wipe tmp and let things reconstruct from sstate, but it can be useful from time to time Nov 17 16:51:01 so just rm -r tmp in my build directory, then Nov 17 16:52:49 even from sstate, it usually takes me about 40 minutes to get the image built Nov 17 16:56:55 hmm, removing only ./tmp as opposed to ./build is much much faster, thanks :) Nov 17 17:05:08 :) Nov 17 17:09:53 rburton: I am at lttng-tools issue Nov 17 20:25:17 kergoth: you commented on your logging color patches to scriptutils… did you want to retract those or just thinking about the future? Nov 17 21:04:53 the latter. i think we need them, just eventually bitbake should provide those functions, since it has to configure its own loggers and we're duplicating the code, and i'll likely propose moving warnings to stderr too, but that'd likely be more controversial Nov 17 21:09:22 if I have more than 1 SRC_URI urls for a recipe and I would like to override the SRCREV for both of them in a .conf file SRCREV_pn- = "xxx" wont be enough Nov 17 21:09:32 is there a way to specify both ? Nov 17 21:23:19 you can use the _pn- override with any variable name Nov 17 21:23:24 including SRCREV_name Nov 17 21:23:37 i.e. SRCREV_main_pn-foo = .. Nov 17 21:23:48 the name based SRCREV doesn't use overrides Nov 17 21:23:54 even if it did, though, you can combine multiple overrides Nov 17 21:38:29 kergoth: I see so if I have SRCREV Nov 17 21:38:49 SRCREV_foo = "xx" and SRCREV_bar = "yyy" Nov 17 21:39:26 then I can use SRCREV_foo_pn- = "aa" and SRCREV_bar_pn- = "bbb" Nov 17 21:39:39 yep Nov 17 21:39:40 cool I wil try that out Nov 17 21:56:42 ha! Nov 17 21:56:43 whoops Nov 17 21:57:14 ah balls, NAS full Nov 17 21:57:33 halstead: nas01:/vhosts/ is full up :( Nov 17 22:03:17 rburton, Cleared some. You can queue a new build. Nov 17 22:36:43 halstead: thanks! Nov 17 22:50:18 WARNING: The free space of /data/poky-master/tmp (/dev/sdb1) is running low (0.067GB left) Nov 17 22:50:24 damnit now i've filled up my disks too Nov 17 22:50:41 /dev/sdb1 1.8T 1.8T 71M 100% /data Nov 17 22:50:47 * rburton is proud Nov 17 23:28:50 impressive **** ENDING LOGGING AT Wed Nov 18 02:59:59 2015