**** BEGIN LOGGING AT Mon Sep 26 02:59:56 2011 Sep 26 06:36:44 good morning Sep 26 09:30:52 gm Sep 26 09:31:16 SRC_TARBALL_STASH="" doesn't seem to work any more for preventing bitbake from trying to download everything from Angstrom. Is that right? Sep 26 09:33:30 hi all Sep 26 09:36:56 hello pb_ Sep 26 09:37:39 pb_: btw, didn't know you wrote update-rc.d Sep 26 09:38:40 the one in oe, yeah Sep 26 09:39:10 debian's update-rc.d is written in perl, which is no good Sep 26 09:44:56 * ant_work remembers chkconfig Sep 26 09:51:57 hmm. build is now ignoring our svn server completely Sep 26 09:58:36 I added the package 'dhcp' to my image, but I'm getting errors now: resolve_conffiles: Existing conffile /home/kobe/svn/baseunit_gen1/build/tmp/rootfs/image/etc/device_table is different from the conffile in the new package. The new conffile will be placed at /home/kobe/svn/baseunit_gen1/build/tmp/rootfs/image/etc/device_table-opkg. But I don't see the file device_table in the dhcp folders nor recipes... Sep 26 10:12:47 kobe, did you -c clean yourimage.bb after that ? Sep 26 10:14:42 what would make bitbake suddenly start only looking on the Angstrom servers for SVN images, completely ignoring the actual SVN server itself? Sep 26 10:14:57 in a recipe that worked fine the other day Sep 26 10:37:08 hello Sep 26 10:37:39 can anyone help me how to compile dosemu package for mini2440 (ARM) on oe... Sep 26 10:38:51 um, dosemu on arm? that sounds like a doomed proposition. Sep 26 10:46:15 pb_, i didn't get you... Sep 26 10:47:10 well, what DOS binaries do you have that can run on arm? Sep 26 10:48:50 pb_, I want to compile only DOSEMU.... Sep 26 10:48:59 right, but what do you want to use it for? Sep 26 10:54:07 logrotate has a do_install_append. I want to create a .bbappend that uses my logrotate.conf. can I use do_install_append_beagleboard for that? Sep 26 10:54:22 how can I append to do_install_append? Sep 26 11:03:15 tasslehoff: You can have multiple do_install_append entries Sep 26 11:10:14 RP__: ah. can I know which one will be ran last? I want to overwrite the logrotate.conf that the .bb creates. Sep 26 11:10:43 tasslehoff: I think bbappend should run last Sep 26 11:17:02 what's wrong with http://pastebin.com/z19yVFe3 ? Sep 26 11:19:20 tasslehoff: I think you want SRC_URI_append_beagleboard Sep 26 11:19:38 tasslehoff: also it will need to have a space at the start of the string Sep 26 11:23:24 bluelightning: tried both, but my logrotate.conf isn't copied down to WORKDIR Sep 26 11:25:16 tasslehoff: it definitely needs to be _append_beagleboard Sep 26 11:25:22 tasslehoff: if you "bitbake -c cleansstate logrotate" then "bitbake logrotate" does it work? Sep 26 11:26:38 bluelightning: that did the trick. thanks :) Sep 26 11:26:48 hi all Sep 26 11:27:40 tasslehoff: really that shouldn't be necessary though... -c clean should be enough, I presume that's what you were doing in between builds before? Sep 26 11:27:51 hi alx_ Sep 26 11:27:59 * tasslehoff whistles innocently Sep 26 11:27:59 hi bluelightning Sep 26 11:28:00 :) Sep 26 11:28:16 bluelightning: no. I should always clean after a build fails before I retry? Sep 26 11:28:43 tasslehoff: well, if you've changed the recipe either you should bump PR or use -c clean Sep 26 11:28:50 I'm fighting with 32bit/64bit... is it possible to extract a sdk from oe build somehow ? Sep 26 11:29:07 bluelightning: ok. thanks. Sep 26 11:30:35 tasslehoff: sometimes you can get away with not doing it... but e.g. if you change SRC_URI and it's already done do_fetch/do_unpack, it's not going to rerun those tasks Sep 26 11:30:51 unless you make it Sep 26 11:32:56 bluelightning: which was just the case here :) Sep 26 11:36:39 hmm. there's no longer a mount.sh and a mount.blacklist in the udev-recipe, only in udev-extra-rules that nobody seems to use? Sep 26 11:39:38 hi all, if I have layer A->B->C such as A is the lowest layer and overridden by B and C can I have a .bbappend in layer C which can change the behaviour of layer A? Sep 26 11:40:07 bbappend still puzzles me a bit :) Sep 26 11:46:13 mr_nice: yes that will work just fine Sep 26 11:48:10 bluelightning: cool :) thx Sep 26 11:50:42 mr_nice: easy way to imagine it is if the bb and all of the bbappends for it were just concatenated together in the layer priority order Sep 26 11:54:58 bluelightning: ah, ok Sep 26 11:57:07 bbappend is awesome it solves my #1 problem with layers - duplicated code Sep 26 11:58:18 the only thing I dislike about it is that I did not notice it until today :D Sep 26 12:00:21 I want my machine to use u-boot_2011.09.bb. What's the right place to set that? Sep 26 12:06:34 tasslehoff: probalby local.conf is the best place for it Sep 26 12:08:17 tasslehoff: something like PREFERRED_VERSION_u-boot = "2011.09" should work Sep 26 12:17:28 I added iperf to the IMAGE_INSTALL variable of my image recipe, but the iperf executable is nowhere to be found in the final image... Sep 26 12:19:22 kobe: did it really get added? what does "bitbake -e your-image | grep IMAGE_INSTALL" report Sep 26 12:19:23 ? Sep 26 12:21:44 mr_nice: thanks. Sep 26 12:43:26 @bluelightning: if I do that, iperf is in the list Sep 26 13:05:38 kobe: then it ought to be in the image Sep 26 13:05:53 kobe: are you sure the iperf *package* is not installed in the image? Sep 26 13:15:36 I will try to clean and rebuild the image. Sep 26 13:43:07 hi all Sep 26 13:43:27 ping bluelightning Sep 26 13:45:22 Noor: pong Sep 26 13:45:37 Hi, my build always hangs at: 0: gcc-cross-initial-4.5-r43+svnr178923 do_fetch (pid 3807) Sep 26 13:45:41 Anyone having a similar issue? Sep 26 13:45:43 bluelightning: sorry to bother you but just need to ask one thing Sep 26 13:46:22 if I need to exclude a variable from cache which is being used in do_install function of base-files how i will do that Sep 26 13:46:51 bluelightning: I tried this do_install_pn-base_files[vardepsexclude] = "DISTRO_VERSION" but it did not work Sep 26 13:47:00 Noor: the real question is how is it being used in do_install? Sep 26 13:47:25 Noor: you need to set the vardepsexclude on the variable/procedure that *directly* uses it Sep 26 13:48:37 bluelightning: do_install directly useses it but this is only used in base-files recipes Sep 26 13:48:49 so I want to make it specific to base-files Sep 26 13:48:57 am i doing wrong Sep 26 13:48:59 ? Sep 26 13:49:31 Noor: why do you want to do this btw? Sep 26 13:49:32 should I ignore the recipe and make it like do_install[vardepsexclude] = "DISTRO_VERSION" Sep 26 13:50:03 bluelightning: DISTRO_VERSION was depending on DATE Sep 26 13:50:22 so my cache was invalidated which was build a day earlier Sep 26 13:50:56 Noor: if you do what you are asking then your /etc/issue will not report the correct DISTRO_VERSION Sep 26 13:50:59 that could be very bad Sep 26 13:51:50 the DISTRO_VERIOSN is defined as DISTRO_VERSION=4.0-${DISTRO_TYPE}-${DATE} Sep 26 13:52:08 right, so it legitimately depends on DATE Sep 26 13:52:14 so the cache for basefiles will only be valid for one day Sep 26 13:52:19 yeah Sep 26 13:52:49 so next day the cache will be invalid for this recipe Sep 26 13:52:52 that's expected, if you define DISTRO_VERSION in that way Sep 26 13:53:24 anything else would result in the incorrect date being reported after the first time it is built Sep 26 13:55:06 wow, I think it was a .lock file in /downloads, and flock was hanging. wonder how that happened Sep 26 13:57:42 Noor: so either you live with it, or change DISTRO_VERSION... Sep 26 13:58:10 bluelightning: looking into DISTRO_VERSION now Sep 26 13:58:25 I think I need to change it Sep 26 13:58:27 thanks Sep 26 14:02:58 bluelightning: so just for may information if I use DATE (or any varaible that in do_install of any recipe) which is changed daily and that change is not important to the recipe that we build it daily so how we will ignore that varaible using vardepexclude Sep 26 14:04:07 Noor: assuming the value of DATE is not material to the output, you can exclude it using vardepsexclude Sep 26 14:04:41 however vardepsexclude *must* be specified on the variable that directly depends on the variable Sep 26 14:04:57 i.e. here it's DISTRO_VERSION, not do_install Sep 26 14:05:19 ahan so we cant remove the dependency of task Sep 26 14:05:37 bluelightning: we can just remove the dependency of varaibles Sep 26 14:05:40 Noor: the dependency of the task will be removed via the variable Sep 26 14:05:48 since that's the only way it got added in the first place Sep 26 14:06:09 but you have to be careful, this is not designed for hacking around legitimate dependencies Sep 26 14:07:20 bluelightning: what if there is a statement like this in do_install echo -n "${DATE} " >> ${D}${sysconfdir}/my_date Sep 26 14:07:52 Noor: then that's an additional dependency Sep 26 14:08:10 but this time for the task in which that appears Sep 26 14:08:29 although it might not be picking that up automatically Sep 26 14:09:29 bluelightning: so in this scenation task is depenedent on DATE .... am I right? Sep 26 14:09:51 what so u mean by ur last line? Sep 26 14:09:52 Noor: well, since it uses DATE, yes... Sep 26 14:10:19 Noor: I don't know if bitbake can look through shell script code and figure out dependencies automatically Sep 26 14:10:50 Noor: but even if it doesn't that script still uses DATE, so it will do something different dependent on the DATE, naturally Sep 26 14:15:11 bluelightning: so my question is can we exlude DATE from do_install function? Sep 26 14:16:00 Noor: to do it properly I think you'd need to provide a new do_install without those lines Sep 26 14:17:22 should things without license checksums make it into meta-openembedded? Sep 26 14:19:14 bluelightning: sorry I think could not word my qestion properly .... my question was can we exlcude DATE from cache by exclude do_install depency on DATE using vardepexclude? Sep 26 14:19:15 Zoxc: not really, because they won't build... Sep 26 14:20:01 Noor: well, yes you can but what I'm trying to tell you is you should not do it that way Sep 26 14:21:53 libsdl-ttf lacks one, but it also seem to have conflicting licensing Sep 26 14:21:57 bluelightning: using vardepsexclude? you are saying that I should provide such a function which does not use DATE in it? Sep 26 14:22:29 Noor: that's the most correct way of fixing it, yes Sep 26 14:23:30 Zoxc: I'm guessing that's probably the way it was in OE classic, just needs correcting now that we're a little more strict about licensing Sep 26 14:31:29 the lua static library gives undefined references to floor and pow Sep 26 14:41:45 Zoxc: hmm, guess it's not correctly linked to libm (missing -lm in link command) Sep 26 14:42:59 It worked when I added that, too bad nothing figures out that lua requires that Sep 26 14:43:48 Zoxc: well, that's really up to lua's makefile (or whatever build system it uses, I don't know much about it) Sep 26 14:44:40 doesn't help when that information isn't passed along to packages using lua Sep 26 15:07:33 does oe use a chroot jail? Sep 26 15:18:15 Zoxc: no Sep 26 15:30:29 how can I get config scripts to install to crossscripts? Sep 26 15:51:33 or how can I get /usr/bin to go in the dev part of the package? FILES_${PN}-dev += "${bindir}" didn't work Sep 26 15:53:39 you probably need to change the order of the PACKAGES, or remove ${bindir} from FILES_${PN}. Sep 26 16:17:47 how do you escape " in bitbake files? Sep 26 16:18:18 escaping is really limited. if you need to use ' in a variable, best to use " to quote it, and vice versa Sep 26 16:18:36 a la python Sep 26 16:18:44 but, that said, due to how the parser works, it doesn't really care that much what's inside. it strips off the leading ' and " and ending ' and " and keeps the rest Sep 26 16:18:53 iirc, anyway. Sep 26 16:22:51 ' doesn't work =P Sep 26 16:23:06 or nvm Sep 26 16:28:47 how can you get autotools to use crossscripts? Sep 26 16:30:17 that probably is impossible, but in koen's defence (!!) there is no reason that somebody else couldn't do that testingmorning kergoth Sep 26 16:30:26 er, stupid ircii Sep 26 16:30:36 morning kergoth, is what i meant to say Sep 26 16:30:40 hey pb_ Sep 26 16:31:30 Hm, OE wiki is broke. I think the webserver URL rewrite is busted. Sep 26 16:32:15 ka6sox: would you be able to look at that ^ ? Sep 26 16:32:54 mewyn: if you access it under http://www.openembedded.org/wiki/ everything should be fine Sep 26 16:33:05 other URL prefixes may still be semi-broken Sep 26 16:33:16 Ah, yeah. openembedded.org/wiki is what's busted. Sep 26 16:33:51 May want a rewrite rule to fix the raw URL to be www. :) Sep 26 16:34:57 mewyn, okay asking the web gurus to fix that issue and make a nice rewrite rule. Sep 26 16:37:01 if i have a open embedded distribution that lacks native gcc in its final distribution image, what would i need to do to get it to add it in? Sep 26 16:38:48 ka6sox: there's also the issue of index.php vs wiki/ and wiki.openembedded.org vs. www.openembedded.org Sep 26 16:40:03 okay let me look @ the nginx config and make sure its answering up to all of those. Sep 26 16:41:38 www.openembedded.org openembedded.org www.openembedded.net openembedded.net wiki.openembedded.org wiki.openembedded.net Sep 26 16:41:47 it aswers to all of those. Sep 26 16:41:54 but not openembedded.org I see Sep 26 16:42:07 nm...it does Sep 26 16:42:37 ka6sox: go to http://wiki.openembedded.org, then click on any link Sep 26 16:42:50 that's still broken Sep 26 16:43:08 ya, thats the rewrite thing... Sep 26 16:43:25 still working on that. Sep 26 16:43:29 ok, cool Sep 26 16:50:53 bluelightning, try now please? Sep 26 16:51:02 let me see if I got these rewrites correct. Sep 26 16:52:10 ka6sox: sorry, wiki.openembedded.org is not working still although I can see it is rewriting Sep 26 16:52:55 maybe wiki.openembedded.org itself should just redirect to www? Sep 26 16:53:25 ~dig wiki.openembedded.org Sep 26 16:53:32 hmmm my bot isnt' here... Sep 26 16:53:33 brb Sep 26 16:54:48 okay, this is a DNS issue Sep 26 16:55:01 wiki and @ are still pointed to the old IP. Sep 26 16:55:25 so my rewrites on the new server are ineffective. Sep 26 16:55:39 Crofton, Crofton|work ping? Sep 26 16:57:21 pong Sep 26 16:58:10 trying to fix the wiki issue here when you have a few minutes. Sep 26 16:58:25 lets do it fast Sep 26 16:58:52 @ IN A 140.211.169.152 Sep 26 16:59:06 wiki IN A 140.211.169.152 Sep 26 16:59:26 the .nets we will need mickeyl to fix. Sep 26 17:00:14 right Sep 26 17:00:20 ok, I think they are fixed Sep 26 17:00:33 okay so its just dns timeouts and it should all be good. Sep 26 17:00:43 the rewrites are in place. Sep 26 17:00:45 now going for lunch Sep 26 17:00:49 laters Sep 26 17:00:52 thanks Sep 26 17:01:09 bluelightning, thanks for bringing that to my attention :D Sep 26 17:01:29 mewyn, 3-5hrs and it should all be good. Sep 26 17:01:31 ka6sox: np, I wanted to do so sooner but I forgot... Sep 26 17:17:43 pb_, ping? Sep 26 17:46:00 ka6sox: Regarding the rewrite issue I had sent that to the list. What way do you prefer to contact you? Sep 26 17:46:22 PaulePanter, the list is fine. Sep 26 17:46:54 the primary issue is that we are moving servers to a new one and things were in a state of flux. Sep 26 17:48:38 Understood. I just want to make sure that work is not duplicated. Sep 26 17:49:43 ericben: around? Sep 26 17:50:05 ericben: it seems our sdk has been broken with 4.7.4 and I am trying to figure out WHY Sep 26 17:50:23 ericben: qt4 ;) Sep 26 17:50:58 what's the meaning of "referenced by DSO" ? Sep 26 17:52:35 errorlog --> http://pastebin.com/yDEwTktr Sep 26 18:02:01 ericben: it seems it is not using the QT_CONF_PATH env variable to look for the qt.conf file Sep 26 18:12:12 ericben: it ignores the environment variable completely Sep 26 18:15:20 ericben: I fixed it here and dunno why it changed. But I found a way to make it work. I'll post a patch for review. Sep 26 18:47:56 MrCurious: by native gcc do u mean the gcc that should run on your target or do you mean native gcc for your build machine ? Sep 26 18:52:43 in order to include native gcc for your target you should look into task-core-sdk.bb if you are using oe-core Sep 26 18:53:56 webkit link step is a complete ld bashing Sep 26 18:54:32 right now ld is eating up 2G memory and running linking for past 30 mins Sep 26 18:54:57 we should definitely think of forcing gold for such bad ass apps Sep 26 18:55:07 since gold takes 7 mins to link this Sep 26 18:56:16 IIRC in oe-core pb had a patch where per recipe linker could be selected Sep 26 18:58:44 ld-is-gold hmmm but its a DISTRO_FEATURE Sep 26 20:04:30 khem: i mean when i boot the created image on arm machine, there will be a native gcc there that can be used' Sep 26 20:05:05 i am sure its a 1 line thing to get a working gcc in the final product, just dont know the 1 line Sep 26 20:06:40 bitbake core-image-lsb-sdk Sep 26 20:06:44 has a native gcc toolchain Sep 26 20:06:52 in general you want the '-sdk' Sep 26 20:07:27 i have a recipe for a chumby one that i want to retrofit gcc into, rather than the other way around... Sep 26 20:08:39 i tried adding PREFERRED_VERSION_glibc-gcc or something similar... Sep 26 20:08:59 but i didnt see that a gcc bin appeared in the final image Sep 26 20:24:33 whats the best way to have an init script installed but not configured to run - INITSCRIPT_PARAMS="" ? Sep 26 20:40:43 MrCurious: Try something like EXTRA_IMAGE_FEATURES += "tools-sdk" Sep 26 20:54:27 RP__ : thanks! will add that to the top level? recipe tonight Sep 26 20:54:46 MrCurious_: Add that to your image recipe Sep 26 20:55:00 ok Sep 26 20:55:06 how do I override a particular autoconf variable for a specific package? Sep 26 20:55:14 on a specific arch Sep 26 20:55:29 msm: which variable? Sep 26 20:55:38 host_cpu Sep 26 20:55:51 i can make a patch for configure.ac which seems to be done a lot Sep 26 20:55:52 msm: passed to configure? Sep 26 20:56:26 EXTRA_OECONF_powerpc64 = "host_cpu=ppc64" Sep 26 20:56:28 ? Sep 26 20:56:38 RP__ thanks! Sep 26 20:56:41 If its embedded within configure you might have to patch it, depends how its being set Sep 26 20:59:33 RP_: shouldnt I add that to meta/site/powerpc64-linux though? Sep 26 21:05:22 msm: those are usually used for site config tests, we've not usually had to set cpu data there :/ Sep 26 21:18:35 I can't access the oe web page: Error 310 (net::ERR_TOO_MANY_REDIRECTS): There were too many redirects. Sep 27 01:13:40 Is there any word on when http://openembedded.org/ and http://wiki.openembedded.org/ will be back up? **** ENDING LOGGING AT Tue Sep 27 02:59:57 2011