**** BEGIN LOGGING AT Tue Feb 12 02:59:58 2013 Feb 12 08:22:23 morning Feb 12 08:23:46 why does opkg keep all {pre,post}{inst,rm} scripts in /var/lib/opkg/info ? Feb 12 09:19:21 morning all Feb 12 09:20:00 good morning Feb 12 09:25:59 hi bluelightning florian all Feb 12 09:27:22 hi eFfeM_work Feb 12 10:20:22 i have errors with eglibc-initial_2.17 .. "OPTION_EGLIBC_NSSWITCH is disabled, but fixed config file" Feb 12 10:20:59 anyone an idea ? Feb 12 10:27:50 pwgen: not something I've seen, but maybe when khem wakes up he might be able to advise if nobody else has by then Feb 12 10:28:23 ok. Feb 12 10:51:04 pwgen: broken DISTRO_FEATURES Feb 12 10:51:23 pwgen: you probably have DISTRO_FEATURES += "something" in config Feb 12 10:55:49 . correct. Feb 12 10:56:13 how can i define a proper distro features ? Feb 12 11:03:22 pwgen: explicitly set it to include all of the features you want in your distro config Feb 12 11:04:00 you can include ${DISTRO_FEATURES_LIBC} or merge the values in it into your own DISTRO_FEATURES value, whichever you prefer Feb 12 11:06:53 by the way , is there a project that supports an basic initrd to set up the hardware proprer and the continues booting ? like the debian/ubuntu/redhat initrds ? Feb 12 11:25:42 hello Feb 12 11:43:19 hi all Feb 12 11:51:53 DJW|Home: ping Feb 12 12:00:06 the append syntax confuses me. I can do 'SRC_URI_append_ptest = " foo"' that is included but I can't seem to get a do_install_append_ptest() function ran Feb 12 12:03:40 | * satisfy_dependencies_for: Cannot satisfy the following dependencies for packagegroup-core-standalone-hhvm-sdk-target: Feb 12 12:03:43 | * eglibc-utils * libc6 * eglibc-thread-db * libc6 (= 2.17-r3) * libsegfault * Feb 12 12:04:02 LIBC_DEPENDENCIES in tclibc-eglibc.inc lists not existing entries? Feb 12 12:05:40 hrw: the do_install looks okay. So bitbake -e RECIPE.. doesn't show you the right function? Feb 12 12:06:10 zecke: my build lacks eglibc-utils package. have to find that why Feb 12 12:06:55 hrw: yeah.. i have one. :} Feb 12 12:07:32 bitbake eglibc -ccleansstate started Feb 12 12:14:19 eglibc-thread-db does not exists in any of my last builds ;( Feb 12 12:42:02 quick opkg question: the prerm, postrm scripts don't seem to run when a package is upgrade via 'opkg upgrade'... is there some way to have these scripts be run so that an application can be nicely stopped via systemd before upgrade and restarted via systemd after upgrade (kind of like doing 'opkg remove PKG; opkg install PKG'? Feb 12 12:45:09 jonibo why not stop the service in preinst et start it in postinst ? Feb 12 12:45:56 yeah, if that does what I described, then that sounds like a good way to solve the problem... i'll give that a try Feb 12 12:46:43 anyway the preinst may check it's started before stopping it Feb 12 12:47:11 i would hope that 'systemctl stop SERVICE' would suffice...??? Feb 12 12:48:01 in the case of installation (not upgrade) the service does not exist so the stop may fail Feb 12 12:48:24 that's why you have to check the service is running before stopping it Feb 12 12:48:45 ok... thanks Feb 12 12:48:57 preinst is used for install and upgrade Feb 12 12:49:22 someone tell me if i am wrong Feb 12 13:07:31 RP, bluelightning: what is the status of the PR fields now? do I bump them or remove them when submitting patches? Feb 12 13:11:22 Zagor: Its no longer necessary. I'm not rejecting patches that do it atm though Feb 12 13:11:45 RP: so should I just ignore PR, or remove it? Feb 12 13:11:57 Zagor: when upgrading it can be removed, otherwise ignore Feb 12 13:12:04 RP: ok, thanks Feb 12 13:12:13 Zagor: it can't go backwards Feb 12 13:12:26 (the overall version) Feb 12 13:13:37 right Feb 12 13:29:31 wakko: thanks for the upgrade tip advice... works like a charm Feb 12 13:38:59 ur welcome Feb 12 14:15:45 ok, time for ugly hacks Feb 12 14:16:02 or 'how to break depchain and stay OE compatible' ;D Feb 12 14:25:59 ok, one fix sent Feb 12 14:48:58 is there any way I can override do_package_qa for ptest only? some test suites contains data that package_qa won't allow. Feb 12 15:47:45 can we make ALLOW_EMPTY_${PN} = "1" default for libraries? Feb 12 15:48:07 that sounds dangerous Feb 12 15:48:32 kergoth: unless library is only shipped as static Feb 12 15:49:02 cause libXY-dev always depend on libXY - nevermind does it exist Feb 12 15:49:21 then it should only be default if there are static libs in -staticdev, or we should have a separate variable to indicate that the lib is only static, imo Feb 12 15:49:28 otherwise you could miss actual problems with do_install Feb 12 15:49:37 right Feb 12 15:50:43 another sdk image to push to external... Feb 12 15:52:16 I get a little worried whenever we add automagic behavior.. there's too much automatic behind the scenes stuff happening already.. sometimes explicit is better :) Feb 12 15:53:20 ok, will stay with current state ;) Feb 12 15:54:37 I could see an argument for a variable stating that this recipe is a static library, or even a class that handles it for you, though. that's still explicit, but more information is carried. just 'ALLOW_EMPTY_${PN} = "1"' doesn't tell the reader of the recipe *why* it's allowed Feb 12 16:07:52 and, anyway, if you do have a static-only library then the right answer is probably for PN-dev to only depend on -staticdev and not on the nonexistent PN at all. Feb 12 16:08:06 good point Feb 12 16:08:10 it seems a bit silly to have to produce lots of empty binary packages just to satisfy dependencies that oughtn't to be there in the first place. Feb 12 16:08:14 * kergoth nods Feb 12 16:08:30 maybe a staticlib class would be a good idea, just so it can do that and people don't mess it up Feb 12 16:08:33 heh Feb 12 16:08:42 right, I think that would be a fine idea Feb 12 16:10:11 and, if it doesn't already, insane.bbclass ought to be taught to spot intra-recipe depends on subpackages that have not been generated and produce a diagnostic. Feb 12 17:33:38 moin Feb 12 19:11:14 JaMa: hi... did you have any thoughts on the meta-oe overlayed/appended recipes RFC? Feb 12 19:12:04 yes, sorry I haven't replied yet Feb 12 19:12:42 scoutcamper: can you turn on the ConfirmAccount extension on the wiki to disable spammers from creating accounts? Feb 12 19:18:44 yep Feb 12 19:19:10 and soon i plan to update to the latest version Feb 12 19:22:04 scoutcamper: you may want to use CentralSiteNotice to tell potential users that they have to request an account, and why Feb 12 19:23:22 okay Feb 12 19:23:47 doing real work right now :( will get to this soon Feb 12 19:50:18 wmat, looks like confirm account no longer includes the bits for 1.16.x wiki's meaning i need to upgrade the wiki first.... Feb 12 19:50:58 hm Feb 12 19:55:24 scoutcamper: they still have instructions for setup with versions =< 1.16 Feb 12 19:55:55 instructions yes but no SpecialCreateAcoount.php Feb 12 19:56:01 errr Feb 12 19:56:05 ConfirmAccount Feb 12 19:56:31 ieehks php Feb 12 19:56:35 *g* Feb 12 19:59:46 scoutcamper: yes upgrade is required it would be nice if you can do that for oe-wiki Feb 12 20:00:08 ill try to get to it this week Feb 12 20:11:28 scoutcamper: ok Feb 12 20:26:36 wmat, perhaps you know of a replacement for IpbWiki Paypal MediaWiki extension Feb 12 20:26:36 as its the blocker to upgrading oe.org wiki Feb 12 20:27:00 scoutcamper: don't even know what that one is Feb 12 20:27:51 http://www.mediawiki.org/wiki/Extension:PayPal Feb 12 20:28:11 its the paypal donate button Feb 12 20:29:10 * wmat looks Feb 12 20:30:15 scoutcamper: it would seem there's a few Feb 12 20:30:25 SidebarDonateBox perhaps Feb 12 20:31:02 i like it, khem? http://www.mediawiki.org/wiki/Extension:SidebarDonateBox Feb 12 20:31:08 ka6sox-away, ^^ Feb 12 20:31:49 there's a MiniDonation extension as well Feb 12 20:31:52 sounds good. Feb 12 20:32:42 the MiniDonation extension looks more current and is authored by one of Mediawiki's maintainers Feb 12 20:34:27 JaMa, should we be hosting that Qt-4.8.0 source in sources.oe.org? Feb 12 20:36:03 ka6sox-away, khem, your choice, minidonation brings back the donation button like what is allready there, sidebar donate box puts a donate button on the sidebar Feb 12 20:36:13 ka6sox-away: no, it's on yp mirror now Feb 12 20:36:44 if the yp mirror is down and its an oe-core recipe...then should we not be hosting it? Feb 12 20:39:19 it's used only in denzil and only if you set preferred version, so I think there are better archives we should host and we're not Feb 12 20:45:28 speaking of wiki - can someone add Layer Index to sidebar? Feb 12 20:55:38 JaMa, okay let me know which ones and I'll add them. Feb 12 20:59:01 ka6sox-away: hello, I have another link for your crew http://forum.xda-developers.com/showthread.php?t=2132703 Feb 12 21:00:08 ant_home, thanks Feb 12 21:00:29 yw Feb 12 21:04:16 okay, trying to run opkg-cl configure in the rootfs_postprocess function coughs a lot of "foo.postinst returned 1" failures Feb 12 21:04:43 is that supposed to work in offline mode? Feb 12 21:07:04 mr_science yes Feb 12 21:30:01 okay, so here's my recipe command: Feb 12 21:30:17 ${STAGING_BINDIR_NATIVE}/opkg-cl --offline-root ${IMAGE_ROOTFS} --tmp-dir /tmp --force-postinstall configure Feb 12 21:30:38 doesn't work for many packages... Feb 12 21:34:56 hello Feb 12 22:01:06 woglinde: then i'm thinking we probably need a newer opkg, etc... Feb 12 22:01:42 we're still on arago 503SDK-ish due to dm816x hardware Feb 12 22:07:40 mr_science I cannot speak for all postinst script Feb 12 22:07:47 but most should work Feb 12 22:09:24 the list of "foo.postinst returned 1" was pretty long Feb 12 22:09:34 not just a few... Feb 12 22:10:19 which is why it sounds like we just need newer stuff, at least based on what you're saying Feb 12 22:11:38 i can work around it for now, but it would be a much cleaner/simpler build/deploy cycle if i could process all that stuff in the image recipe Feb 12 22:11:39 hm I am mostly at master Feb 12 22:11:59 good nite Feb 12 22:12:04 bed now Feb 12 22:12:23 we are definitely not anywhere near upstream master... Feb 12 22:12:31 and good night Feb 12 22:37:05 if a software builds a smalle helper tool which is supposed to run on the target to determine constants .. how is this going to happen inside oe-core toolchains .. as a crosscompiled binary cant run on the host ? Feb 13 00:53:59 does libudev still exist? Feb 13 00:54:24 hi all. Anybody tried narcissus lately ? I cannot build any image with it... Feb 13 00:54:28 Is it down ? Feb 13 00:54:34 does it come up? Feb 13 00:54:44 we had some hosting issues Feb 13 01:58:59 khem, mediawiki update complete, and ConfirmAccount setup **** ENDING LOGGING AT Wed Feb 13 02:59:58 2013