**** BEGIN LOGGING AT Wed Apr 03 02:59:58 2013 Apr 03 07:15:02 good morning Apr 03 08:00:22 bluelightning: woglinde Apr 03 08:00:35 morning all Apr 03 08:00:38 JaMa: ? Apr 03 08:00:47 bluelightning: woglinde reported yesterday that he sees a lot of do_fetch tasks being reexecuted Apr 03 08:00:55 I belive it's because of do_write_srcrev :/ Apr 03 08:01:07 at least with rm_work I see the same now Apr 03 08:01:57 should we be using vardepsexclude for that then? Apr 03 08:02:46 rm_work should probably keep do_write_srcrev stamp file Apr 03 08:03:18 I don't have confirmation if woglinde is also using rm_work Apr 03 08:03:51 but it could be the case, I see do_fetch,do_write_srcrev,do_rmwork for every recipe included in image I'm building Apr 03 08:07:50 JaMa: it's kind of what you'd expect though since do_write_srcrev is "after do_fetch before do_build" Apr 03 08:08:11 gm Apr 03 08:08:16 yes, I'll send patch for rm_work Apr 03 08:08:27 woglinde: do you have some spies here? :) Apr 03 08:08:38 jama hm why? Apr 03 08:09:06 you spawned just after we started talking about do_fetch reexecution you reported yesterday Apr 03 08:09:08 jama but good you see the bitbake master stuff too Apr 03 08:09:14 woglinde: are you using rm_work? Apr 03 08:09:18 yes Apr 03 08:09:58 good morning Apr 03 08:10:07 gm apelete Apr 03 08:11:33 hi woglinde Apr 03 08:11:47 thanks again for staying late last night Apr 03 08:12:38 JaMa: Darren is fixing rootless xserver-nodm-init. Maybe now is the right moment to unify the recipes? Apr 03 08:12:59 maybe yes Apr 03 08:14:18 apelete hm when switching xmlpull to ant you maybee need to set JAVAC to the openjdk native Apr 03 08:14:22 one Apr 03 08:14:28 it might not work with jamvm Apr 03 08:14:31 bluelightning: keeping stamp file will resolve reexecution, but there is also "issue" that when reusing do_package from sstate you will still need to fetch it first (for do_write_srcrev task) Apr 03 08:14:53 no idea if that's issue or feature Apr 03 08:15:05 JaMa: buildhistory does not really work well when just building from sstate; it's something I hope to address in the future Apr 03 08:15:53 JaMa: (requires that we have somewhere else to gather the info from, eg. pkgdata, but that currently doesn't have a file list and there may be other missing things as well) Apr 03 08:16:03 maybe do_write_srcrev should be also covered by do_package_setscene task (just to ignore srcrevs when you're building from sstate) Apr 03 08:17:28 woglinde: I don't get it when you say "switching xmlpull to ant". do you mean compile xmlpull using ant ? Apr 03 08:18:13 or use ant instead of xmlpull in my layer ? Apr 03 08:25:05 gm florian Apr 03 08:25:18 can you pls kick linuxtogo ? Apr 03 08:34:15 apelete using ant for compiling xmlpull Apr 03 09:02:05 ok, that's what I thought. will read the code of other java recipes when I get the time, to see how it works. Apr 03 09:05:17 woglinde: meta-java has openjdk? Apr 03 09:05:47 hrw: yes, openjdk-6 and openjdk-7 Apr 03 09:06:02 cool Apr 03 09:06:38 suihkulokki: do you know how it goes for Andy with his aarch64 work? Apr 03 09:07:33 hrw: well I'm assuming that will be the easiest path for him Apr 03 09:08:52 suihkulokki: if he will not be able to use OE then we can provide him meta-toolchain-openjdk (like we have meta-toolchain-hhvm) and let he work outside of OE Apr 03 09:12:07 hi Stefan Apr 03 09:12:13 hi hrw Apr 03 09:12:29 Holger Apr 03 09:15:45 gm zecke Apr 03 09:27:51 argh... build failed on Linaro's CI but works locally ;( Apr 03 09:30:40 hrw haha you bad Apr 03 09:32:08 woglinde: ;D Apr 03 09:49:50 hi Apr 03 09:50:08 hi Apr 03 09:50:31 why do we get do_package_qa error?\ Apr 03 09:52:21 i'm getting ncurses-libpanel rdepends on dev package of my toolchain Apr 03 09:52:24 alone why is the sea wet? Apr 03 09:52:54 and why crystalballs have no answers Apr 03 09:53:04 what? Apr 03 09:54:02 exactly Apr 03 09:54:16 you did not tell us how you configured your system Apr 03 09:54:23 and what you used at all Apr 03 09:54:31 so crystballs have no answers Apr 03 09:54:48 crystal even Apr 03 09:55:33 i'm using my own toolchain Apr 03 09:57:02 i think there is no fish in sea Apr 03 09:57:07 woglinde: you are lucky... my OE crystalball never returned from service Apr 03 09:58:16 alone: to make things easier: which exactly toolchain you are using ('my own' is nothing), show recipe/config of your external toolchain Apr 03 10:03:24 i'm building for arm Apr 03 10:03:56 in recipe i'm writing FILES_$PN} and FILES_$PN}-dev Apr 03 10:05:25 alone: to make things easier: which exactly toolchain you are using ('my own' is nothing), show recipe/config of your external toolchain Apr 03 10:05:34 do not describe but show Apr 03 10:06:24 FILES_${PN}-dev += "${libdir}/lib* ${libdir}/libncurses*" FILES_${PN} += "${libdir}/libiconv*${SOLIBS} ${libdir}/libc*${SOLIBS}" Apr 03 10:06:46 alone: read my lips: show us recipe Apr 03 10:06:51 whole Apr 03 10:22:14 hi all Apr 03 10:23:22 I'm trying to build a toolchain for Qt4X11, with custom dev packages, I was able to put all qt stuff and custom packages inside the tar.bz2, but I lost gcc.... how come this ? any suggestion on where to search ? thanks in advance Apr 03 10:23:42 linuxtogo.org is again unresponsive .. and several hours ago I requested a password reminder, and never got it Apr 03 11:06:49 very nice work on the layers.openembedded.org page .... i am impressed Apr 03 11:07:31 indeed Apr 03 11:07:45 bluelightning: was it you who made that? Apr 03 11:08:20 hrw: mostly, with some help from others on the design and code review Apr 03 11:08:43 bluelightning: great job - much easier to search for stuff Apr 03 11:08:43 pwgen: hrw: thanks Apr 03 11:10:11 django based ? Apr 03 11:15:09 hmm Apr 03 11:15:17 I would have said yes... I guess he'll be back Apr 03 11:36:03 I lost eglibc headers and gcc in my sdk . anyone has ideas on how to let them back ? Apr 03 11:40:05 silvio build again? Apr 03 12:43:11 hi Apr 03 12:43:25 if we do not write PROVIDES in bb file Apr 03 12:43:36 what does it take by default Apr 03 12:46:34 alone: the filename before .bb Apr 03 12:46:59 bcoz i'm getting error like multiple bb file for a single module Apr 03 12:47:36 i've given different filenames for the two bb files Apr 03 12:47:48 but it is still giving error Apr 03 12:48:48 alone: a more detailed explanation would help Apr 03 12:50:27 the bb file names are enyo.bb and enyo-1.0.bb Apr 03 12:50:44 alone: do you mean enyo_1.0.bb? Apr 03 12:51:44 meta-webos/recipes-webos/frameworks/enyo-1.0.bb Apr 03 12:51:48 it's not PV == 1.0 :/ Apr 03 12:52:18 I see.. not naming I would choose in that case Apr 03 12:53:09 true.. it's version 1.0, but submission 128.2 so PV is 1.0-128.2 Apr 03 12:53:45 woglinde, I add Qt4 and I break something with gcc Apr 03 12:54:02 silvio__: details please, how did you add it? Apr 03 12:57:38 I can't build my svn reciepes anymore (OE classic). Apr 03 12:57:41 BB tries to download the code from the angstrom server instead of my local svn server. Apr 03 12:57:41 http://pastebin.com/XTLUDJti Apr 03 12:57:41 Any idea? Apr 03 12:59:27 ya Apr 03 12:59:34 enyo-1.0.bb Apr 03 13:10:00 cubi try http://www.angstrom-distribution.org in browser Apr 03 13:10:09 the server is not reliable at the moment Apr 03 13:11:07 Cubi_ and maybee ipmon-netcom is not public source Apr 03 13:12:56 bluelightning, I create a series of .bb and task based on the GMAA Apr 03 13:13:06 GMAE Apr 03 13:13:34 woglinde, it's my own reciepe. Why wants BB download it from the angstrom server and not from the server given in the .bb file? Apr 03 13:13:53 silvio__: I mean further details; can we see the recipe? Apr 03 13:14:06 Cubi_: probably PREMIRRORS are set to be tried first Apr 03 13:14:25 bluelightning, Yes, there are many, now I paste Apr 03 13:14:27 Cubi_: this will be in the angstrom distro config, I'm assuming you are using that Apr 03 13:16:55 I start from this that is equivalent of meta-toolchain.bb Apr 03 13:17:00 http://paste.ubuntu.com/5673553/ Apr 03 13:19:29 bluelightning, Now I do a trick, I miss base toolchain with working gcc with my Apr 03 13:19:34 Cubi_: the idea with PREMIRRORS is to try to fetch from there first and if that fails then download from the location specified in SRC_URI Apr 03 13:20:27 but I had an error on "feature.h", but I cant give right headers Apr 03 13:22:19 http://paste.ubuntu.com/5673572/ Apr 03 13:23:17 silvio__, have you tried bitbake -c populate_sdk image_name? Apr 03 13:25:17 Crofton|work, Image became populated, but not with all packages, anyway now i m tryng Apr 03 13:26:10 bluelightning, thank you, this seems to be the problem. Apr 03 13:26:55 Now rebuilding the cache fails, but I think I can fix it. Apr 03 13:32:04 task-core-standalone-btw-sdk-target Apr 03 13:32:12 http://paste.ubuntu.com/5673563/ Apr 03 13:58:06 otavio: is there a way to install a linux-firmware component selectively? Apr 03 14:29:53 has "src_distribute_local.bbclass" been deprecated in oe-core ? Apr 03 14:29:56 I still have a " INHERIT += "src_distribute_local" " in a distro config file and bitbake is complaining about it Apr 03 14:31:21 mckoan: ? what do you mean? Apr 03 14:31:45 apelete: the better method nowadays is to use archiver.bbclass. Apr 03 14:32:40 kergoth: so I should replace INHERIT += "src_distribute_local" with INHERIT += "archiver" ? Apr 03 14:33:50 it's not a drop in replacement, it uses entirely different variables to control its behavior Apr 03 14:36:55 kergoth: okay, commenting it should be safer then. (will find a solution to that if a problem arises) Apr 03 14:37:00 thanks Apr 03 14:37:19 np Apr 03 14:45:36 otavio: I'd need ony one of the .fw installed by the recipe Apr 03 14:46:09 mckoan: I've sent some patches which has splitted some (which I use) Apr 03 14:46:19 mckoan: I don't know if I covered your need Apr 03 14:46:29 mckoan: otherwise you may send a patch similar to mine Apr 03 14:47:35 otavio: yep, I was simply courious, thx Apr 03 14:47:49 mckoan: which fw you need? Apr 03 14:49:09 otavio: actually I already have it but I'm installing it by hand, ath9k Apr 03 14:49:49 mckoan: I did the split of it, take a look Apr 03 16:33:39 JaMa: so do you want to do this rm_work + do_write_srcrevs fix or should I? Apr 03 16:33:49 JaMa: I think you may be better placed to test it at least... Apr 03 16:34:21 it does seem like making rm_work preserve the stamp is the best solution Apr 03 16:34:29 bluelightning: I've it locally already Apr 03 16:34:39 JaMa: ah, great! Apr 03 16:35:36 sent Apr 03 16:35:56 btw: there is updated patch for those tag-srcrev mappings, please review it when you have time Apr 03 16:36:51 here is output from it https://github.com/webOS-ports/buildhistory/commit/65dbe7ee83124df0c634d3ae19a1993d6b033ce7 Apr 03 16:36:54 yep I'm looking at it right now as it happens :) Apr 03 16:38:26 I was hit by #4102 when I was testing it, that's why I forgot that I've it locally (I've sent 2 patches for #4102 instead) Apr 03 16:38:45 JaMa: what's with the slashes in the tag_xxx ? Apr 03 16:38:52 I guess that comes from the recipe? Apr 03 16:39:40 yes is't in tag name Apr 03 16:40:14 ah ok Apr 03 16:41:23 and in this case I haven't optimized it to drop 'default' SRC_URI name if there is only one, I can update it if you want, but for functionality it's not imporant Apr 03 16:42:07 actually it's a bit easier to check name_tag then to add another if to check if there is only one Apr 03 16:47:22 JaMa: yeah it's probably fine to leave it that way I think Apr 03 21:23:56 hello Apr 03 21:48:36 hi Crofton|work and bot Apr 03 21:51:09 Umm, I found that libxcb_1.1.91.bb and liberation-fonts_1.06 didn't work.. Apr 03 21:51:30 they are missing the SRC_URI[...] checksums.. has anyone tested them in the last few years, or should they be removed/fixed? Apr 03 22:35:23 khem, any official plans on expanding qemu machines to cover some other arches or provide bare-metal qemu MACHINEs? Apr 03 22:36:05 khem has a branch with qemumips64 on it, we should get those bits merged at some point Apr 03 22:38:24 given how oe seems to work today (everybody not employed by intel or win-driver gets his merge-requests ignored forever), is there any way i could get my -contrib aldot/fixes applied anytime soon? If that doesn't happen i don't see a point to polish up my stuff against micro nor add new MACHINEs 8( Apr 03 22:39:16 but anyway. late here, /me is tired; i'd better not rant aloud Apr 03 22:39:19 * blindvt & Apr 03 22:39:48 I get stuff merged almost every time i submit patches to the list, and i work for mentor Apr 03 22:39:51 so.. Apr 03 22:41:20 kergoth, fair enough. make that any big vendor as opposed to honks playing around in their precious spare-time :P Apr 03 22:43:18 khem seems to get stuff in pretty often too, iirc he works for a company that doens't use oe at all Apr 03 22:43:21 * kergoth shrugs Apr 03 22:45:02 perhaps i should try more often or ping frequently Apr 03 22:45:40 ah, yeah, definitely have to babysit the patchsets, stuff does get missed from time to time. resends seem to help quite a bit Apr 03 22:45:47 usually not too bad though Apr 03 22:49:27 kergoth, btw, what's up with linuxtogo.org? Looking at https://www.google.at/search?hl=en&as_q=bernhard+oe+ping results ends nowhere, i.e. times out on me Apr 03 22:51:14 rather annoying not even being able to point to alot of pings when there are no working archives ;) Apr 03 22:51:52 but i find myself ranting again Apr 03 22:51:56 blindvt: I said it in the TSC meeting and I'll say it again, I think that allegation is unfair and untrue Apr 03 22:52:10 * blindvt waves -- nighty night, all Apr 03 22:55:28 blindvt: you may wish to read http://permalink.gmane.org/gmane.linux.embedded.poky/8807 **** ENDING LOGGING AT Thu Apr 04 02:59:58 2013