**** BEGIN LOGGING AT Fri Jun 18 02:59:57 2010 Jun 18 05:49:51 JaMa, when you install a package with --force-reinstall, would you expect new dependencies to be installed as well? (they currently are not) Jun 18 05:51:19 i'm considering making 'opkg --force-reinstall install foo' an alias to 'opkg --force-depends remove foo && opkg install foo' Jun 18 05:54:39 grg: well maybe it would be good too (I just didn't have the case for this yet), but this alias is what I usually expect to happen Jun 18 05:55:42 JaMa, there are currently a bunch of special cases in the install code paths to handle the --force-reinstall case. Jun 18 05:56:03 i think i'll remove them and see if things behave Jun 18 05:57:19 if this alias makes things easier I agree it's good idea (I used exactly this to simulate expected behavior when --force-reinstall failed..) Jun 18 06:00:40 the only other way i see of fixing the force reinstall bug is to add a bunch more special cases - and hope i don't forget any Jun 18 06:08:51 hi, I'm currently bisecting: Jun 18 06:09:11 in images MACHINE_EXTRA_RRECOMANDS seem broken Jun 18 06:10:54 I hope I'll find the commit soon Jun 18 06:25:11 JaMa,hi, jama do you have all modules in images you ship for the freerunner Jun 18 06:27:50 I think bisect failed again Jun 18 06:28:00 I need to bisect building from scratch Jun 18 06:30:15 GNUtoo|laptop: no, removed them a while ago Jun 18 06:30:29 GNUtoo|laptop: I once bisected the kernel with ubuntu configs, each build took a hour Jun 18 06:30:40 I've a core i.7 now Jun 18 06:30:43 It took me 4 days to find the offending commit, so good luck to you :) Jun 18 06:30:47 but.... Jun 18 06:30:56 this time I need to build from scratch each time Jun 18 06:31:07 else I've commits that are both good or bad Jun 18 06:31:12 s/or/and Jun 18 06:31:27 GNUtoo|laptop: there is only few modules really needed on fr and opkg upgrade (for slightly bumped kernel) is now much faster without upgrading tons of never used modules Jun 18 06:31:53 ok Jun 18 06:32:56 JaMa, are MACHINE_EXTRA_RRECOMMENDS included in ultra-recent images? Jun 18 06:33:23 seems so (it was in image built yesterday) and I did rebuild from scratch last weekend Jun 18 06:35:20 ok Jun 18 06:35:25 thanks for the info Jun 18 07:08:12 Hi All, I have a doubt.. How to restrict bitbake to try only http:// links for downloading source etc. and not ftp:// links..? Jun 18 07:11:10 Nikunj: I have a doubt it's possible and sane way.. if the source is available only on ftp:// then how should bitbake deal with http-only? Jun 18 07:13:41 morning Jun 18 07:16:09 good morning Jun 18 07:18:31 gm all Jun 18 07:33:01 JaMa : Can I make like the 1st priority is to look for http:// link... if its not thr then go for ftp:// link..? Jun 18 07:38:51 Nikunj: it usually tries distro source mirror first (usually http://) and if it's not found then it tries SRC_URI from recipe and I don't think there is a way to force it try http:// first for SRC_URI in recipe Jun 18 07:41:17 JaMa : My organisation doesnot allow to use ftp://... What I can see is, Its retrying for some 5-6 times to download source using a particular ftp:// link.. Can I restrict this retrying to only once..? Jun 18 07:41:52 JaMa : It will save a lotta time.. and bandwidth. Jun 18 07:44:46 Nikunj: grep wget conf/bitbake.conf Jun 18 07:46:37 JaMa : ohh nice.. got it.. the " -t " parameter for wget..? Jun 18 07:47:27 JaMa: Currently My build is running.. should I edit the bitbake.conf file in between..?? Jun 18 07:53:20 JaMa: Man, where are you?? Jun 18 07:57:09 Can anyone tell, If I can edit the bitbake.conf file , while the build is currently running..?? Jun 18 07:58:12 Nikunj: you can not edit bitbake.conf at all Jun 18 08:00:12 hrw: If I interrupt the currently running build, and then edit bitbake.conf file, and then start the build again.. will it resume the build from whr it left off..?? Jun 18 08:01:50 it will Jun 18 08:02:26 Nikunj: I'm at work doing job completely unrelated to OE.. Jun 18 08:06:10 JaMa: Thanks a Ton for your advice.. Bye. Jun 18 08:07:08 Nikunj: bye Jun 18 08:19:48 what is a policy of handling missing gettext warnings? Jun 18 08:20:16 other then ignore which looks like is in use now (as core components has warnings) Jun 18 08:20:32 hrw: It seems to have stuck after resuming from the place whr it left off... http://pastebin.org/338540 Jun 18 08:21:05 Nikunj: break, run with -DDDDD to get debugs Jun 18 08:22:01 can I use an if like construction in a machine .conf file? I want to specify a different TARGET_PREFIX if LIBC is uclibc Jun 18 08:22:15 distro is after machine Jun 18 08:22:17 good morning Jun 18 08:22:23 and LIBC is distro thing Jun 18 08:22:28 hrw: I have set the wget 's "-t " parameter in bitbake.conf file to 1. It is 5 by default. I dont want to try ftp:// links, since Its banned here.. And so to save time I edited the same. Jun 18 08:23:42 Nikunj: copy those wget lines to your local.conf instead Jun 18 08:26:14 hrw, not sure if I uinderstand you properly. atm I am using angstrom distro, in local.conf user can say ANGSTROMLIBC = glibc or uclibc, in case of uclibc I want to add _uclibc to TARGET_PREFIX, your answer seems to indicate to me: not possible. right? Jun 18 08:26:33 eglibc Jun 18 08:27:04 woglinde_: true, forgot about that one, didn't even try to build it for my arch Jun 18 08:27:09 eFfeM_work: ok, if angstrom allows for setting distro vars by user then it will work Jun 18 08:27:29 eFfeM_work: but for umbaumba distro which does not allow for user to set distro vars it will not work Jun 18 08:27:40 uclibc TARGET_PREFIX is set default to uclibcsometjing Jun 18 08:28:03 hrw, ok Jun 18 08:28:07 /devel/arm/oetmp-ang/sysroots/armv7a-angstrom-linux-uclibceabi/ Jun 18 08:28:19 woglinde_: where? siteinfo.bbclass ? Jun 18 08:28:39 hrw: Line 32. and " FETCHCOMMAND_wget = "/usr/bin/env wget -t 1 --passive-ftp -P ${DL_DIR} ${URI}" " "RESUMECOMMAND_wget = "/usr/bin/env wget -c -t 1 --passive-ftp -P ${DL_DIR} ${URI}" " Jun 18 08:28:42 woglinde_: but eFfeM_work wants to set armv7a-angstrom-linux-uclibceffemeabi instead Jun 18 08:28:58 iirc Jun 18 08:29:02 Nikunj: and? Jun 18 08:29:29 hrw: should i copy these lines to local.conf.. and remove from bitbake.conf..? Jun 18 08:29:43 copy to local.conf and revert changes to bitbake.conf Jun 18 08:30:03 hrw: Ohh.. okay.. Letme try. Jun 18 08:30:29 hrw, woglinde_I'm working on nios2, siteinfo.bbclass contains both nios2-linux and nios2-linux-uclibc, I seem to recall if I didn't set TARGET_PREFIX explicitly binutils would end up with the wrong names Jun 18 08:30:35 or is gnuconfig involved here Jun 18 08:34:05 Has anybody managed to run a LAMP stack on the beagleboard ? Jun 18 08:37:09 screwgoth1: apache and php on beagle is a pain, better use lighttpd, I had lighttpd, php5, mysql running on beagle (but better ask this in #beagle) Jun 18 08:43:21 hrw: man.. thanks a ton.. its working fine..!! Jun 18 08:45:08 Nikunj: nearly any setting in bitbake.conf can be overriden by local.conf Jun 18 08:46:05 hrw: Ohh.. That is cool.. Is there anything I can speed up building..? Jun 18 08:47:06 Nikunj: http://marcin.juszkiewicz.com.pl/2008/04/07/speeding-up-bitbake-builds/ Jun 18 08:47:48 effem iiiehkkkss Jun 18 08:48:06 woglinde_: ieks what ? Jun 18 08:48:14 nios2? Jun 18 08:49:20 php mysql Jun 18 08:49:54 woglinde_: mythtv needed it Jun 18 08:50:06 I thought only mysql Jun 18 08:50:33 actually mythtv needs mysql for its database, but if you want to run mythweb you need php Jun 18 08:50:45 ieeehkkskss Jun 18 08:52:00 RDEPENDS_mythweb_lighttpd = "lighttpd lighttpd-module-cgi lighttpd-module-fastcgi \ Jun 18 08:52:00 lighttpd-module-rewrite php-cgi lighttpd-module-auth" Jun 18 08:52:01 :-) Jun 18 08:52:30 so again: what is a policy of handling missing gettext warnings? Jun 18 08:52:47 woglinde_: actually it was not too fast at some times Jun 18 08:55:04 hrw, no idea Jun 18 09:11:55 03Koen Kooi  07org.openembedded.dev * rd5f753d8a5 10openembedded.git/recipes/xbmc/libmicrohttpd_0.4.6.bb: libmicrohttpd: fix typo Jun 18 09:12:06 03Koen Kooi  07org.openembedded.dev * rfc07994de9 10openembedded.git/recipes/swig/swig_1.3.39.bb: swig 1.3.39: bump PR Jun 18 09:37:41 eFfeM_work: lighty is plan B for me... first, I want to see if I can get a proper LAMP stack installed for angstrom Jun 18 09:43:31 screwgoth1: I had quite some issues with apache and php_mod, eventually gave up Jun 18 09:43:47 consider yourself warned :-) Jun 18 09:43:54 eFfeM_work: :-) Jun 18 09:44:36 eFfeM_work: hmm... but can you tell me did configuring apache with DSO support (using --enable-so configure option) work ? Jun 18 09:51:40 screwgoth1: been > half a year ago since I tried, so don't recall exact details. Tried quite some things. basically there were issues with cross and native paths and things that needed to be run on the host (like axfs). Forgot the details. I recall XorA also peeked into this, not sure about the current status Jun 18 09:52:30 I thought apxs might give some problems.... Jun 18 09:52:52 it will Jun 18 09:52:52 and you say you managed to get lighty to work with mysql and php ? Jun 18 09:52:57 yes Jun 18 09:53:11 the recipe for those things is in the tree Jun 18 09:54:19 hi likewise Jun 18 09:54:49 hi eFfeM_work Jun 18 09:55:44 likewise: nios2+mmu seems to work (at least I have a busybox prompt), full oe built toolchain, kernel etc Jun 18 09:56:02 hrw: If I edit local.conf in between currently running build, Do I have to restart the build to see the changes or it will automatically take the changes into account.? Jun 18 09:56:03 cleaning up things etc will try to push something next week or so Jun 18 09:57:09 eFfeM_work: I never touched apache, just cleaned up php and lighty Jun 18 09:57:23 03Koen Kooi  07org.openembedded.dev * r7cdfbc23dc 10openembedded.git/recipes/systemd/systemd_git.bb: systemd: fix PV and a typo in FILES Jun 18 09:57:27 Nikunj: have Jun 18 09:58:20 XorA: ok, thanks for the clarification (and I must say: a very wise decision) Jun 18 09:58:54 eFfeM_work: well actually I did look at apache, turned pale and decided Rum was a better option Jun 18 09:59:00 hrw: I have set PARALLEL_MAKE = "-j 2" and BB_NUMBER_THREADS = "4" to speed up building. I have core2duo processor and 2GB RAM.. Is this okay.? Jun 18 09:59:09 XorA: Lol !!! Jun 18 09:59:53 aaaahhhh Rum :-) good choice Jun 18 10:00:25 eFfeM_work: sometimes :-) Jun 18 10:00:27 likewise: I think we are under attack ..... Jun 18 10:01:21 Nikunj: sure Jun 18 10:04:03 03Koen Kooi  07org.openembedded.dev * rcdfe641ca1 10openembedded.git/recipes/linux/linux.inc: Jun 18 10:04:03 linux.inc: add config mangler to enable cgroups support, disabled by default Jun 18 10:04:03 * set KERNEL_ENABLE_CGROUPS = "1" in your distro/local.conf to enable this Jun 18 10:04:03 * CGROUPS were added in 2.6.24 Jun 18 10:04:04 khem, can't you pass -fomit-frame-pointer only if gcc-4.5 is really used? Jun 18 10:04:56 blindvt`: there is no way to check PV of other recipes used Jun 18 10:05:43 oh?! Jun 18 10:06:18 i remember now, yea. How inconvenient Jun 18 10:08:21 we talked about storing more if not all of recipes in the DB, this could potentially fix the missing info wrt PVs too Jun 18 10:10:25 eFfeM_work: haha, F-16's doing a low-pass over the building? Jun 18 10:10:45 eFfeM_work: I'm sure some of the car alarms will be triggering on the low-passes. Jun 18 10:11:15 yes and some choppers; flight show today and tomorrow .... had to make a detour to get here Jun 18 10:11:37 our office windows are facing the wrong direction :-( Jun 18 10:12:02 eFfeM_work: infiltrate the R&D manager office :-) Jun 18 10:12:32 :-) Jun 18 10:15:23 eFfeM_work: Did you try cross-compiling directly, or did you try installing apache by meddling with the recipes ? Jun 18 10:15:35 recipes Jun 18 10:15:58 screwgoth1: the simplest way if you just want to make it work is to compile apache natively Jun 18 10:16:31 Yup. That's the plan. I was wondering what you had tried Jun 18 10:16:45 hrw: yup its working fine.. I can see the difference.. Nicee :) Jun 18 10:17:08 screwgoth1: as i wanted others to be able to have a different setup I didn't go for native Jun 18 11:00:07 eFfeM_work: on what nios platform was the shell prompt succes? Jun 18 11:04:16 likewise: sygeg1 hw, nios2+mmu (processor produced by walter) Jun 18 11:04:32 likewise: 2.6.34 kernel from mmu git Jun 18 11:04:49 eFfeM_work: ok, the same platform we tested the nios2+nommu toolchain on. Jun 18 11:05:28 eFfeM_work: I think for OE, we need to target a standard dev board, cyclone iii and iv Jun 18 11:05:54 likewise: planning to get it running on the neek baseboard (whatever its name is) as well Jun 18 11:06:35 eFfeM_work: cool, the're two versions of the NEEK: the starter board (no display) and the NEEK (with display), AFAIK the NEEK is a superset of the starter board, though I need to check that Jun 18 11:07:11 likewise: that is also what I've been told, we have the starter board here, so I can probably get it running on that one Jun 18 11:07:17 might try that later today Jun 18 11:16:03 eFfeM_work: that might be the best way to get it into OE. I think I can revisit the nios2 branch and get it up-to-date with current, so that we can merge that as well. Jun 18 11:17:43 likewise: you want to keep the non-mmu branch too? Jun 18 11:18:28 eFfeM_work: well, if we support nios2, why not non-mmu as well? I think MMU is the future, but for real-time, nommu can be beneficial. Jun 18 11:18:40 eFfeM_work: we should focus on mmu though Jun 18 11:19:25 likewise: I haven't really tried the non-mmu stuff, not sure whether none of my changes conflict with it Jun 18 11:19:56 eFfeM_work: no problem, let's get mmu in first, I will merge nommu later. Jun 18 11:20:09 likewise: ok will keep you posted Jun 18 11:22:12 eFfeM_work: thanks Jun 18 11:22:19 eFfeM_work, hi I saw a mail about python, I do not know all bitbake's functions but I hope I remember python enough, is it the " ".join that you don't understand? Jun 18 11:22:45 eFfeM_work: and after we have this in, altera probably announces their Cortex-A9 core FPGA :-) Jun 18 11:23:08 GNUtoo|laptop: I think the join just concatenates, the issue is that my python is not good enough to come up with a fix Jun 18 11:23:11 likewise: :-) Jun 18 11:23:17 ah ok Jun 18 11:23:24 indeed it concatenates Jun 18 11:23:45 likewise: at least in that case the compiler and tools will be there Jun 18 11:24:28 eFfeM_work: true :-) Spoke with Altera ppl yesterday, they didn't know what ARM will be in there yet (talked to the wrong ppl probably). Jun 18 11:31:28 likewise, you serious about Altera coming out with a part including an arm? Jun 18 11:31:45 you did see the Xilinx announcement. Jun 18 11:32:19 Crofton: yes, serious Jun 18 11:32:24 cool Jun 18 11:32:29 is there an url Jun 18 11:32:44 Crofton: yes I know the Xilinx announcement. However, Xilinx always has their announcements way ahead of their silicon Jun 18 11:32:53 yeah Jun 18 11:32:55 I laughed Jun 18 11:33:11 we will tell you real details in early 2011 Jun 18 11:33:46 Crofton: no I got it from Altera, but they were then still deciding on what ARM core exactly. Jun 18 11:34:32 Crofton: they already tried it with the Excalibur device (years ago), but that was probably way ahead (and superseded by softcores). Jun 18 11:39:42 03Koen Kooi  07org.openembedded.dev * r1d7eeddd2a 10openembedded.git/conf/distro/angstrom-2008.1.conf: angstrom: bump distro version Jun 18 13:46:35 03Koen Kooi  07org.openembedded.dev * r3981cd6da9 10openembedded.git/recipes/angstrom/ (angstrom-version.bb angstrom-version/lsb_release): angstrom-version: add lsb_release script Jun 18 14:00:14 morning Jun 18 14:13:46 How do I fix the CROSS-COMPILE BADNESS error when cross-compiling from source Jun 18 14:14:26 depends. you'll need to provide more details Jun 18 14:14:31 specifically, what buildsystem? Jun 18 14:15:59 03Philip Balister  07org.openembedded.dev * r1fcb25aff7 10openembedded.git/recipes/swig/ (swig.inc swig_1.3.39.bb swig_2.0.0.bb): Jun 18 14:15:59 swig : Add swig-2.0.0, add DEPENDS on python, and remove DEF PREF -1 on .39 Jun 18 14:15:59 * Without the DEPENDS on python it is possible to build swig without python Jun 18 14:15:59 support, which breaks gnuradio builds. Jun 18 14:35:49 hey, is the new staging mechanism available in oe-stable? Jun 18 14:37:57 good morning people Jun 18 14:38:01 hi, Jun 18 14:38:11 I don't know for the new staging btw Jun 18 14:38:59 I have an Embest SBC6020 here, and I want to get rid of it :) Jun 18 14:39:14 ? Jun 18 14:39:15 anyone here interested? Jun 18 14:39:24 what kind of board is it? Jun 18 14:39:30 ARM? Jun 18 14:39:33 yes Jun 18 14:39:40 wait, i'll look it up Jun 18 14:39:43 ok Jun 18 14:40:03 http://armkits.com/Product/sbc6020.asp Jun 18 14:40:06 this one Jun 18 14:40:10 I'm not shure that I'll be interested but I'm interested in making that the board find an owner Jun 18 14:40:39 50 euros, incl. shipping and packing and stuff, would that be a good price? Jun 18 14:41:20 looks nice but I've already too much machines Jun 18 14:41:23 it's not yet supported in OE Jun 18 14:41:26 ah ok Jun 18 14:41:36 but I don't have the knowledge to make it supported Jun 18 14:42:03 it's not that hard,the fist thing to do is to create a recipe for the kernel Jun 18 14:42:08 If I had, I would have used the board myself Jun 18 14:42:24 yeah, but the problem is to get the hardware working ;) Jun 18 14:42:28 and a machine config Jun 18 14:42:59 after that you bitbake a basic image and then try to findout what doesn't work and fix it Jun 18 14:43:11 CF card works after some trouble, but not all of the serial ports do and there are some mysteries about the network ports too Jun 18 14:43:13 it's not that hard Jun 18 14:43:16 and another question: if I want to add a new task to a recipe, and that task depends on another package has been staged already, how should I configure it (before/after config) Jun 18 14:43:38 no time, not enough knowledge, not enough money to create the time and knowledge :P Jun 18 14:43:54 thebohemian: you need intertask dependency Jun 18 14:44:11 hi hrw Jun 18 14:44:13 hrw: how does that look like? Jun 18 14:44:19 thebohemian: grepping Jun 18 14:44:39 hrw: what is a good keyword? Jun 18 14:44:57 maybe using bitbake -g to find the good recipes Jun 18 14:44:57 ? Jun 18 14:45:00 and grep that? Jun 18 14:45:03 not shure tough Jun 18 14:45:31 but again what to grep in bitbake -g Jun 18 14:46:11 Romke, maybe ask the mailing list Jun 18 14:47:34 thebohemian: if I would remember... Jun 18 14:47:56 thebohemian: juce/juce_1.29.bb:do_unpack[depends] += "unzip-native:do_populate_sysroot" Jun 18 14:48:04 I can't try because I'm already compiling and bisecting the tree Jun 18 14:48:12 ok Jun 18 14:48:32 hrw: thanks! you are saving my day :) Jun 18 14:48:44 np Jun 18 15:08:27 mickey|wm: you there? Jun 18 15:08:39 mickey|wm: you there? ping Jun 18 15:12:16 is the something special about the intertask dependencies? I ask because I get a nice python stacktrace from bitbake as soon as I have task dependency in the recipe Jun 18 15:12:23 the stacktrace looks like this: Jun 18 15:12:23 mark_active(depend, depth+1) Jun 18 15:12:23 File "/home/rschus/oe/com.buglabs.build.oe/oe/bitbake/lib/bb/runqueue.py", line 546, in mark_active Jun 18 15:12:23 if runq_build[listid] == 1: Jun 18 15:12:23 IndexError: list index out of range Jun 18 15:13:13 (however looking into task-depends.dot all is fine) Jun 18 15:20:32 ok, this issue showed up only with my current 'tmp' folder Jun 18 15:20:40 starting from scratch solved it ... Jun 18 16:32:02 likewise, I've been looking for mickeyl for several days Jun 18 16:32:30 likewise, I hope that he will connect to irc the week-end Jun 18 16:33:07 thebohemian, indeed during git bisect I faced the same issue Jun 18 16:33:26 GNUtoo|laptop: what does the "wm" suffix mean on him? Jun 18 16:33:36 I've no idea Jun 18 16:33:40 maybe windows mobile? Jun 18 16:33:49 he has a windows mobile htc Jun 18 16:34:14 someone told him that everything worked on GNU/Linux in this phone,so he bought it Jun 18 16:34:30 but that didn't reflect the reality well Jun 18 16:34:41 so I don't know if he still uses it Jun 18 16:35:11 likewise, you know fso userland a bit? Jun 18 16:38:24 GNUtoo|laptop: nope, don't know what fso means. I am (not yet) in mobile developments, but that's exactly why I need mickey :-) Jun 18 16:38:55 likewise, ah ok,maybe other people could answer your questions Jun 18 16:39:01 what would be your question? Jun 18 16:41:03 GNUtoo|laptop: I need iPhone devs Jun 18 16:41:20 (oh, and I need an iPhone as well :-) ) Jun 18 16:42:12 likewise, so sorry battery problem Jun 18 16:42:43 [18:41] GNUtoo|laptop: I need iPhone devs Jun 18 16:42:45 [18:41] (oh, and I need an iPhone as well :-) ) Jun 18 16:43:06 ah ok Jun 18 16:43:14 so you need mickeyl Jun 18 16:44:05 ouch present rate: 0 mW Jun 18 16:44:09 what kind of battery is that Jun 18 16:44:14 brand new computer Jun 18 16:44:29 less than 1 month old Jun 18 16:46:10 ah no it's present rate Jun 18 16:46:17 anyway I've a battery issue Jun 18 16:52:52 I'm trying to cross-compile a package called "sofsip_cli", using a toolchain I compiled using "bitbake console-image" Jun 18 16:53:13 But I get the following error "CROSS-COMPILE BADNESS" Jun 18 16:53:20 ring any bells ? Jun 18 16:56:31 screwgoth, look at the logs Jun 18 16:57:22 especially the configure log Jun 18 16:57:31 s/log/logs Jun 18 16:58:30 screwgoth, apparently you didn't read what i said when you asked the last time. that's not enough information to help you fix it Jun 18 16:58:58 hi kergoth_ Jun 18 16:59:03 I'm still git bisecting Jun 18 16:59:10 fun fun Jun 18 16:59:10 even if the number of commits are small Jun 18 16:59:18 I must rebuild from scratch each time Jun 18 16:59:20 well i guess each test takes a while Jun 18 16:59:21 * kergoth_ nods Jun 18 16:59:22 that sucks Jun 18 16:59:40 there is also another approach : Jun 18 16:59:45 test commits one by one Jun 18 16:59:51 but do not rebuild from scratch Jun 18 16:59:58 I wonder which one takes more time Jun 18 17:00:09 one by one starting from older to newer Jun 18 17:00:16 or just dig in and try to find the root cause directly, but that could be painful Jun 18 17:01:01 I've already failed at that Jun 18 17:01:20 :\ Jun 18 17:03:02 kergoth_: oops, I really didn't read Jun 18 17:04:44 screwgoth, iirc that message means the crosscompiler was called with a -I or -L pointing at build machine paths (i.e. /usr). but we can't tell you how to make that stop without knowing what buildsystem you're dealing with Jun 18 17:05:24 The actual error is as follows : "CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/sofsip_cli cc1plus: internal compiler error: in add_path, at c-incpath.c:362 ..." Jun 18 17:05:37 right, so it tells you the problem Jun 18 17:05:45 the crosscompiler can't be using shit from /usr/include Jun 18 17:05:54 that'll be the wrong arch Jun 18 17:07:25 yeah, but the files it is referencing in /usr/include are the same as in tmp/sysroot/arm.../usr/include path Jun 18 17:07:36 so why the fuck is it causing an error ? Jun 18 17:09:30 screwgoth, please be polite,some of us are doing it without beeing paid Jun 18 17:09:43 and no software is perfect Jun 18 17:10:15 screwgoth, basically what does it include? Jun 18 17:10:20 /usr/include/? Jun 18 17:10:29 ah ok sorry Jun 18 17:10:36 /usr/include/sofsip_cli Jun 18 17:11:33 hmmm you have sofsip_cli on your host? Jun 18 17:11:47 screwgoth, and you can guarantee they're the same, can you? Jun 18 17:11:53 same glibc version? same 'asm' directory? Jun 18 17:12:04 same versions of the librarieswhose headers live there? Jun 18 17:12:06 not likely. Jun 18 17:12:15 and furthermore, how do i get configure to generate a makefile to include from my tmp/sysroot/arm.../usr/include path ? Jun 18 17:12:26 if you're using an autoconf based buildsystem, say so Jun 18 17:12:33 i've told you we need to know what the buildsystem is like 4 times now Jun 18 17:12:37 kergoth_: oops, sorry ... apologise ppl, i was angry on the failed compilation sir. Jun 18 17:12:39 i'm not sure how many more times i need to say it Jun 18 17:13:19 the most common build system is autotools Jun 18 17:13:25 other includes: Jun 18 17:13:27 cmake Jun 18 17:13:31 bjam Jun 18 17:13:33 etc... Jun 18 17:13:42 or plain Makefiles without ./configure Jun 18 17:14:09 and how you fix it varies, it depends on where it's coming from. the most common case is a -I@includedir@ in a Makefile.am, if it uses automake, but it's not the only way it could be pulled in Jun 18 17:14:19 or -I$(includedir), that is Jun 18 17:16:44 kergoth_: Yes, it's autoconf, I missed your messages somehow. And I'll try your suggestion of modifying Makefile.am and let you guys know.... Thanks, that helped Jun 18 17:17:06 03Koen Kooi  07org.openembedded.dev * reaf6ed8af8 10openembedded.git/recipes/bluez/ (5 files): bluez4: add 4.66, clean up DP in older versions Jun 18 17:17:33 the crosscompile badness error exists because we prefer an immediate, obvious error to inexplicable failures and odd behavior later on Jun 18 17:17:40 better to error right away so we can fix it Jun 18 17:17:47 or not Jun 18 17:18:04 yes because after you must go with gdb Jun 18 17:18:10 or disassemble the binary etc... Jun 18 17:18:18 you would get an illegal instruction Jun 18 17:18:30 and it would be a lot harder to know why Jun 18 17:18:44 he disconnected :) Jun 18 17:18:47 and the fix would just be the same Jun 18 17:18:47 ok Jun 18 17:18:51 thanks Jun 18 17:18:54 hehe Jun 18 17:20:30 I'm thinking of puting a segfault or illegal instruction message in my quit message...lol Jun 18 17:21:22 * gnutoo is attracted by debugging Jun 18 17:21:48 like going to look what happens with gdb+gdbserver Jun 18 17:22:53 gnutoo: that's a fantastic idea (about the quit message). :-) Jun 18 17:23:25 lol Jun 18 17:23:27 Maybe an oops followed by a "kernel panic". Jun 18 17:23:54 in that case you must do something like that Jun 18 17:24:02 init segfault -> kernel panick Jun 18 17:25:38 "pernel kanic" would make things even funnier. Jun 18 17:25:40 maybe a python module for irc that says a random crash message in /contrib Jun 18 17:25:56 yes but the kernel panic must be caused by something Jun 18 17:25:57 :-D Jun 18 17:26:05 NULL pointer exception? Jun 18 17:26:36 General Protection Failure on win32.dll Jun 18 17:27:07 I prefer things I know Jun 18 17:27:30 NULL pointer exeption is a common failure in kernel oops Jun 18 17:27:35 indeed Jun 18 17:36:48 maybe you should make the message dynamic: blame the last IRC message for the error Jun 18 17:38:05 seem to work Jun 18 17:38:19 lol Jun 18 17:38:41 lol Jun 18 17:38:42 re. My kernel crashed. Jun 18 17:38:48 :-) Jun 18 17:39:08 I had my hand ready over the Enter key to trigger on your response :-) Jun 18 17:39:28 ok Jun 18 17:39:28 now you need to automate that (don't know how) Jun 18 17:39:39 there are bindings Jun 18 17:39:44 for irc clients Jun 18 17:40:09 the message that I've put is a real one from gdb Jun 18 17:40:38 more explicit than the usual segmentation fault Jun 18 17:42:11 compiling with -mthumb -O1 and gcc chickens out on every other recipe hmmmm Jun 18 17:44:25 hi likewise Jun 18 18:12:08 http://tree.celinuxforum.org/CelfPubWiki/OpenEmbedded Jun 18 18:12:19 someone should fix this page.. Jun 18 18:17:58 03Khem Raj  07org.openembedded.dev * r7565ef3e4a 10openembedded.git/ (4 files in 3 dirs): Jun 18 18:17:58 omap5912osk: Use 2.6.34 as PREFERRED_KERNEL Jun 18 18:17:59 * TMPFS is disabled it does not boot with tmpfs. Jun 18 18:17:59 needs to be investigated. Jun 18 18:17:59 Signed-off-by: Khem Raj Jun 18 18:18:00 03Khem Raj  07org.openembedded.dev * r097131c401 10openembedded.git/recipes/busybox/busybox_1.15.3.bb: Jun 18 18:18:00 busybox_1.15.3.bb: Work around the gcc 4.5 ICE on thumb Jun 18 18:18:00 Signed-off-by: Khem Raj Jun 18 18:18:03 03Andreas Mueller  07org.openembedded.dev * r19bd6d851c 10openembedded.git/recipes/xfce-base/ (14 files): Jun 18 18:18:03 xfce-base: Bump xfce recipes to version 4.6.2 Jun 18 18:18:03 Signed-off-by: Andreas Mueller Jun 18 18:18:03 Signed-off-by: Khem Raj Jun 18 18:18:03 03Khem Raj  07org.openembedded.dev * r83be2bd634 10openembedded.git/recipes/liboil/liboil_0.3.15.bb: Jun 18 18:18:03 liboil_0.3.15.bb: Get rid of do_stage Jun 18 18:18:04 Signed-off-by: Khem Raj Jun 18 18:22:06 hi khem Jun 18 18:22:17 how is it going Jun 18 18:29:49 khem: good, playing with bare-metal cortex m3 today, just for fun Jun 18 18:34:50 hope to run linux on such a device later: http://www.linux-arm.org/LinuxKernel/LinuxM3 Jun 18 18:36:14 likewise: cool Jun 18 18:36:31 likewise: although you must be looking for uclinux on that Jun 18 18:38:36 http://www.linux-arm.org/git?p=linux-2.6-stable.git;a=summary Jun 18 18:38:39 hmm Jun 18 18:39:14 likewise: are you gonna use OE ? Jun 18 18:39:51 khem: probably, yes Jun 18 18:40:06 likewise: nice Jun 18 18:40:09 khem: it's just a fun project, I see no real use yet. Jun 18 18:40:28 khem: I'm playing with the small RTOSes a bit Jun 18 18:40:53 3 us thread switch times are kinda of cool :-) Jun 18 18:41:53 heh nice Jun 18 18:47:00 UFO: Alien Invasion 2.3 released Jun 18 18:47:11 hm.. good game for openpandora :) Jun 18 19:08:59 likewise, you mean OS or execution environment? Jun 18 19:09:16 git bisect is still going... Jun 18 19:15:47 * GNUtoo is realizing that he is adicted to cross-compilation Jun 18 19:16:15 git bisect takes too much time and I don't know what to do while waiting for it to finish Jun 18 19:18:31 too much time? Jun 18 19:18:43 yes it's been a day that I'm running it Jun 18 19:18:51 heh Jun 18 19:19:01 yesterday I was running it too but it failed because I didn't recompile from scratch every time Jun 18 19:19:02 faster box or smaller testcase is what you need :) Jun 18 19:19:13 I can't get faster than that: Jun 18 19:19:16 core i7 Jun 18 19:19:33 I think I need a good occupation Jun 18 19:20:58 I'm afraid of smaller test case Jun 18 19:21:07 I can't afford it to fail a new time Jun 18 19:31:36 GNUtoo: Was it you or denix that notes packaged staging stuff have the wrong arch? Jun 18 19:33:01 it was not me Jun 18 19:33:27 I'm curently bisecting the tree to find why kernel modules are not shipped anymore Jun 18 19:37:04 what board/image? Jun 18 19:40:22 bug/angstrom Jun 18 19:40:32 the images include base-something as task Jun 18 19:41:02 which include MACHINE_EXTRA_RRECOMMAND Jun 18 19:41:13 which include linux-modules in my case + some firmwares Jun 18 19:48:16 GNUtoo: on the thread switch? Jun 18 19:48:49 GNUtoo: gnutoo_> likewise, you mean OS or execution environment? Jun 18 19:50:01 likewise, by execution environment I mean something that looks like an OS when you program for it but on the target it's a single binary Jun 18 20:04:18 re Jun 18 20:14:07 GNUtoo: thats what most rtoses are :) Jun 18 20:15:54 indeed Jun 18 20:16:01 but he said OS Jun 18 20:16:43 11:41 < likewise> khem: I'm playing with the small RTOSes a bit Jun 18 21:30:48 03Khem Raj  07org.openembedded.dev * rafcaed5d09 10openembedded.git/recipes/netkit-telnet/netkit-telnet_0.17.bb: Jun 18 21:30:48 netkit-telnet_0.17.bb: Add -lncurse to LDFLAGS on uclibc bases builds. Jun 18 21:30:48 Signed-off-by: Khem Raj Jun 18 21:40:57 can someone enlighten me about the difference between RRECOMMENDS and RDEPENDS ? Jun 18 21:42:06 so far: Jun 18 21:42:14 725c273d05d9e3e940abfe5da7d045bd61ce02d6 bad Jun 18 21:42:28 7dc8f297689cdbba9021c2c620bd025ca234c58a good Jun 18 21:43:36 maybe it's opkg then Jun 18 22:00:29 hi JaMa Jun 19 00:43:18 gpm do_compile http://tinderbox.openembedded.net/public/logs/task/6283284.txt there are other build reports with same error. any pointers on how to fix this ? **** ENDING LOGGING AT Sat Jun 19 02:59:57 2010