**** BEGIN LOGGING AT Thu Jun 10 02:59:57 2010 Jun 10 03:10:45 hmm, now.. Jun 10 03:10:48 bb.fatal Jun 10 03:10:57 this needs to just die Jun 10 03:11:00 period Jun 10 03:11:09 if shit hits the fan, raise a damn exception Jun 10 03:32:34 hi , i m currently having some problems on compiling with arm-angstrom-linux-gnueabi-g++ Jun 10 03:32:45 undefined reference to `feof@GLIBC_2.0' Jun 10 03:34:02 http://pastebin.com/QYQ213CU Jun 10 03:34:10 i got this error when i try to compile my project Jun 10 03:38:36 * kergoth has no idea what would cause that particular failure Jun 10 03:44:17 kergoth , glibc ? Jun 10 03:44:20 i need glibc ? Jun 10 03:45:32 you have it Jun 10 03:46:27 where is it ? Jun 10 03:46:32 what? Jun 10 03:46:38 this isn't linux 101 Jun 10 03:46:42 i cant do compilation Jun 10 03:47:17 undefined reference to `fputs@GLIBC_2.0' Jun 10 03:47:37 we heard you the first time Jun 10 03:47:42 repeating yourself isn't going to accomplish anything Jun 10 03:47:54 except perhaps getting you kicked and banned from the channel, eventually Jun 10 03:48:46 well okay but this is help channel Jun 10 03:51:10 i need glibc include and libs ? Jun 10 03:59:21 i ve found glibc location on : /home/kadirbasol/OE/angstrom-dev/staging/armv7a-angstrom-linux-gnueabi/usr/include Jun 10 03:59:26 but i donno how to compile Jun 10 05:24:01 I doubt he'll return, but if he does... OpenEmbed needs to add -mfpu=vfp Jun 10 06:05:36 hi , i m having problem on omapfbplay , when i try to play a video from framebuffer console , ( no x window system installed ) , it runs video perfectly but the problem is , when video play finishes , it doesnt return back to linux console , so i cant control machine anymore , how can i fix this problem ? Jun 10 06:11:39 BoardTester: there's a program called fbset that maybe can help you Jun 10 06:11:59 still, you'll have to try to run it via network/ssh or blindly typing it over the screwed console Jun 10 06:12:20 and you should read first its man Jun 10 07:10:30 morning Jun 10 08:22:17 * JaMa|Wrk is blind today.. added subpackages to navit recipe and they always stay empty, grr Jun 10 08:24:53 does anyone see obvious mistake/typo? http://gitorious.org/~jama/angstrom/jama-shr-experimental/commit/684b000d646d6bf0fbf74a9168eeb9c16b286b0c Jun 10 08:26:39 PACKAGES += Jun 10 08:26:47 do =+ instead Jun 10 08:27:02 otherwise PN package will take all Jun 10 08:27:29 ahh, thanks! Jun 10 08:27:56 JaMa|Wrk: and for future - try to split commits so packaging change is one and rest is in second Jun 10 08:28:01 easier to check changes ;D Jun 10 08:29:26 hrw: OK, I'll split it Jun 10 08:47:46 JaMa|Wrk: any news wrt Gentoo's binutils? Seems very old issue... Jun 10 08:48:14 morning Jun 10 08:55:10 ant_work: have you seen that reply on bug? Jun 10 08:55:48 yes, I even followed the links posted there Jun 10 08:55:59 well, we have to ask khem.... Jun 10 08:56:16 ant_work: from what I've read there I wasn't sure if it's the same issue Jun 10 08:56:34 seems so, but have you tried the bfd patch? http://sourceware.org/ml/binutils/2007-10/msg00019.html Jun 10 08:56:44 ant_work: because enabling plugins/threads "resolve" it.. so doesn't look like upstream issue Jun 10 08:57:15 ant_work: for binutils-cross you mean? no.. Jun 10 08:57:27 hmm...I understood that vapier did finally gave an evidence Jun 10 08:57:54 ant_work: I've already asked khem for ideas.. Jun 10 08:58:34 well, here I see two 'eek' http://sourceware.org/ml/binutils/2007-07/msg00407.html Jun 10 09:03:26 as far as I understand in 2007 vapier diagnosed "LD_LIBRARY_PATH .. not sure which legacy pulled this in, but the code didnt exist in older Jun 10 09:03:28 binutils ... it got merged as part of the recent libtool update i believe." Jun 10 12:03:11 hi sll Jun 10 12:03:14 hi all Jun 10 13:08:39 03Robert Schuster 07org.openembedded.dev * r8b29951fa7 10openembedded.git/recipes/db/ (db4.inc db5.inc): Jun 10 13:08:39 db4.inc: Change SRC_URI to a FreeBSD mirror (link on Oracle site does not work anymore) Jun 10 13:08:39 db5.inc: Dito. Jun 10 13:42:55 woglinde: navit packages splitted :) not pushed yet.. Jun 10 13:44:44 jama cool Jun 10 13:44:45 thanks Jun 10 13:45:05 add you RSUGESTIONS? Jun 10 13:45:10 for sound stuff Jun 10 13:45:22 morning Jun 10 13:45:30 jo kergoth Jun 10 13:47:00 * kergoth thinks about diving back into converting bitbake to the python logging module Jun 10 13:47:07 * kergoth ran into some totally insane behavior last night Jun 10 13:47:30 woglinde: RRECOMMENDS for stuff used in shipped config and RSUGGESTIONS for rest Jun 10 13:47:59 isn't it RSUGGESTS, not RSUGGESTIONS? :) Jun 10 13:48:14 RSUGGESTS Jun 10 13:48:31 yeah whatever it is calles Jun 10 13:48:33 called Jun 10 13:48:35 hi hrw Jun 10 13:48:47 yeah RSUGGESTS is in recipe :) Jun 10 14:09:07 okay, thoughts on me killing every single usage of bb.fatal and bb.msg.fatal in the bitbake codebase? Jun 10 14:09:26 they can all be replaced with either error messages and recovery, or they can raise a proper exception instead Jun 10 14:09:26 kergoth hm and than? Jun 10 14:09:31 okay Jun 10 14:09:34 fine for me Jun 10 14:09:37 i mean, if the shit has hit the fan Jun 10 14:09:47 best to retain as much info about why as we can Jun 10 14:09:53 heh Jun 10 14:10:01 this is the purpose of exceptions, after all Jun 10 14:10:39 was converting to the logging module, and realized that i can't just replace fatal with logger.critical(), cause the latter only prints a message, it doesn't actually exit :) Jun 10 14:11:47 kergoth: cannot you add bb.fatal( msg, exception ) method which also prints stacktrace and replace every bb.fatal() call with exception? Jun 10 14:12:00 i don't really see the point Jun 10 14:12:20 raise SomeException("foo") will do the job just as well as bb.fatal("foo") Jun 10 14:12:33 for now, we can replace with raise SystemExit(), which is exactly what it does today Jun 10 14:12:50 it prints, then runs sys.exit(), which raises SystemExit Jun 10 14:13:33 i , i need some framebuffer imageviewer program on my beagle , is there any software that do it ? Jun 10 14:13:35 but bb.fatal("Cannot count checksum for file %s" % file, exception) whould be easier to read in code (when exception has ie something about missing file) Jun 10 14:13:43 hi , i need some framebuffer imageviewer program on my beagle , is there any software that do it ? Jun 10 14:13:49 i don't really see how that's easier to read Jun 10 14:13:56 i'm not removing bb.fatal Jun 10 14:14:03 i expect it to be used by the metadata still, for convenience Jun 10 14:14:15 but inside of bitbake, we should be using the correct exception for the failure in question Jun 10 14:14:19 bb.fatal is generic Jun 10 14:14:30 imo, anyway Jun 10 14:15:03 remove bb.fatal files if its on bitbake conf Jun 10 14:15:42 mechman: what do you mean? Jun 10 14:15:44 does someone know any image dumping program for framebuffer ? Jun 10 14:15:49 * kergoth doesn't, sorry Jun 10 14:15:58 okay... Jun 10 14:16:09 i need some image display program on framebuffer Jun 10 14:16:21 zgv afaik can display images on fb Jun 10 14:16:21 oh, wait, i do know of one Jun 10 14:16:25 that will show screen in fullscreen mode Jun 10 14:16:27 i'm not sure if its in oe, but i always used to use 'feh' Jun 10 14:16:37 back when i didn't use X much Jun 10 14:16:46 feh ? Jun 10 14:16:50 opkg install feh Jun 10 14:16:52 i try now Jun 10 14:17:12 mechman: I use fbsplash, wich is standard in busybox iirc Jun 10 14:17:13 Collected errors: * Cannot find package feh. Jun 10 14:17:36 but it only shows ppm images Jun 10 14:17:41 ah , nice Jun 10 14:17:44 ppm Jun 10 14:17:45 hmm Jun 10 14:17:54 i need for png Jun 10 14:18:06 fbview works Jun 10 14:18:15 or fbv - do not remember Jun 10 14:19:18 there Jun 10 14:19:24 there is no packages on opkg Jun 10 14:19:28 fbview or fbv Jun 10 14:20:51 btw , i changed splash screen on boot but its background color is not 255,255,255 ( white ) Jun 10 14:21:02 whats the color on bsplash background ? Jun 10 14:21:22 and how can i change splash screen's background color on boot ? Jun 10 14:22:49 you can read about that in the repo in recipes/psplash/psplash.inc Jun 10 14:24:47 tsjsieb , i looked there but there is no color information ? Jun 10 14:25:56 well.. -native recipes will use patches laying in non-native recipe directory Jun 10 14:26:17 I'm not sure this is documented behaviour Jun 10 14:26:27 but I'll use it for patch cmake :) Jun 10 14:29:34 mckoan|away: it suggests to look in the patches, I guess these are in the psplash/files/ dir. Jun 10 14:30:36 he mickeyl Jun 10 14:30:46 sorry was there yesterday Jun 10 14:31:22 btw, what devices we have on LTG stand? :) Jun 10 14:31:34 the florian yes Jun 10 14:31:44 beagleboard, archos 7 Jun 10 14:31:48 ups has Jun 10 14:31:57 the booth isnt this big Jun 10 14:32:07 so thats proably fills up the desk Jun 10 14:32:07 * Jay7 have misread florian as device name ;) Jun 10 14:32:12 sorry Jun 10 14:39:29 I've got a prob- Trying to 'bitbake glibc-initial' fails on do_install because arm-angstrom-linux-gnueabi-gcc doesn't exist. But 'bitbake gcc-cross' fails because it tries baking glibc-initial first. Jun 10 14:40:02 glibc-initial runs after cross-initial, from what i recall. Jun 10 14:40:13 the crosscompiler build process is more than two steps Jun 10 14:41:17 JaMa|Wrk: what you sent to the ML is indeed a mega-patch! Jun 10 14:42:40 Anyone tried busybox >1.13.2 ? More recent recipes exist, but don't seem to be listed as PREFERRED_VERSIONS by distros; I'm wondering if that's because something isn't right with it. Jun 10 14:46:48 mwester: I'm not aware of anything wrong with it, but neither do I know of anything wrong enough with the old versions to make me want to go through the hassle of upgrading. Jun 10 14:47:12 upgrading busybox is always a headache and I suspect most folks would just rather stick with the versions that work for them. Jun 10 14:47:37 kergoth: silly question about buildall: does is still just extra packages the dependencies of PN ? Jun 10 14:48:06 and the rrecommends / rsuggests ? Jun 10 14:49:31 pb__: That's a reasonable answer; I'd like to see if I can get rid of some of the "full" utilities we still need to carry along on the base SlugOS image -- in particular it would be nice if basic things like "mount" finally got fixed in busybox so that we didn't need those pulled in. I'll try it, and see what works and not. Jun 10 14:50:58 (We've run out of space in the base image with newer versions of various utilities -- the software keeps getting bigger and bigger, but unfortunately existing hardware can't expand to match.) Jun 10 14:52:49 morning kergoth_ Jun 10 14:53:29 pb__: have you installed an ip-cam to watch the works? Jun 10 14:53:44 ant_work: heh, no, but my wife has been sending me pictures Jun 10 14:54:13 a webcam would be a fun idea though, maybe I will set one up at the weekend :-) Jun 10 14:54:17 important is the workers don't feel alone... Jun 10 14:54:37 yeah, luckily they have my daughter to keep them company Jun 10 14:54:41 ha ha Jun 10 14:55:22 we have 2 cams here Jun 10 14:55:25 ha ha Jun 10 14:55:34 but we need them for a product Jun 10 14:56:42 some of you have experienced this error during bitbake nano "ERROR: Please set the 'PERSISTENT_DIR' or 'CACHE' variable" Jun 10 14:57:12 Ive googled but there are nothing clear about the cause and solution Jun 10 14:57:25 your BBPATH isn't set up correctly Jun 10 14:57:33 it's falling back to the bitbake conf/bitbake.conf, rathe rthan using the OE one Jun 10 14:57:53 ant_work: this one was easy.. no need for manual changes.. just script monkeys played.. Jun 10 14:58:29 ant_work: rdepends/rrecommends needed some cleanup after monkeys and checksums in recipe was real mega-patch with stupid manual work Jun 10 14:59:09 ok my BBPATH was set in .bashrc "export BBPATH="/home/jaime/oe/build:/home/jaime/oe/openembedded"" Jun 10 14:59:29 echo $BBPATH, ensure its set the way you think it is Jun 10 14:59:30 where oe is a symbolinc link to another drive. Jun 10 14:59:36 then ls /home//jaime/oe/openembedded/conf/bitbake.conf Jun 10 14:59:39 okay, thats a problem too Jun 10 14:59:52 JaMa|Wrk: I see khem already ack'ed it, now let see what pb says Jun 10 14:59:53 if you do builds inside of a path with a symlink, autoconf can explode Jun 10 15:00:02 at least when built under oe Jun 10 15:00:27 ok, thanks, I will change that, Jun 10 15:00:28 JaMa|Wrk: for such patches better have solid ack's ;) Jun 10 15:00:39 np Jun 10 15:01:28 kergoth_: heh, should really fix that one day, it's pretty sad to have these weird restrictions Jun 10 15:01:36 agreed Jun 10 15:01:38 ant_work: I haven't read the whole patch but it looks basically fine to me Jun 10 15:01:43 low priority though, its too easy to work around :) Jun 10 15:01:52 almost wish it wasn't so easy Jun 10 15:02:21 pb__: yea, just dubious for some eventual strange border-case ...which I cannot imagine by myself.. Jun 10 15:02:50 okay, i'm really sick of these "new RPATH" messages Jun 10 15:02:57 http://en.pastebin.ca/1880453 Jun 10 15:03:11 this is my diff Jun 10 15:03:19 it it sane enough to commit to .dev? Jun 10 15:04:20 what's the purpose of that patch? Jun 10 15:04:53 cmake_2.6.4: Patch added to fix compilation when FORTIFY_SOURCE is enabled Jun 10 15:05:29 upstream notified is here http://public.kitware.com/mantis/view.php?id=9267 Jun 10 15:05:33 I made the changes as you suggested, even I changed the BBFILES path in build/conf/local.conf, but now the error persist. I tried using just "bitbake nano" without sudo but this error appears File "", line 18, in check_sanity NameError: global name 'os' is not defined Jun 10 15:05:35 ah, is your patch reversed? Jun 10 15:05:36 s/ is// Jun 10 15:05:45 it looks like you are _deleting_ that file, not adding it Jun 10 15:06:08 +++ b/recipes/cmake/cmake-2.6.4/fix_fortify_source_compilation.patch Jun 10 15:06:10 all this happend after a openembedded/git pull Jun 10 15:06:28 two hour ago everything was fine Jun 10 15:06:41 sorry for my english Jun 10 15:07:05 pb__: it's newly added patch Jun 10 15:07:11 jaime, you should never, ever build as root Jun 10 15:07:16 okay, so your diff in pastebin is backwards Jun 10 15:07:54 ups Jun 10 15:08:16 pb__: it's just commit in my local tree. not pushed yet Jun 10 15:08:35 I wish to collect some ack's :) Jun 10 15:08:56 jaime: unless you want to do "bitbake please-destroy-my-system-and-send-all-my-p0rn-to-ohter-place" Jun 10 15:08:57 * mwester hands Jay7 a large butterfly net Jun 10 15:09:24 mwester: thanks :) Jun 10 15:09:25 jjaja Jun 10 15:09:31 got ir Jun 10 15:09:53 hrw, hehe Jun 10 15:10:08 * mwester wonders when we added the "send-p0rn" feature, but knows that the "please-destroy-my-system-if-I-am-running-bitbake-as-root" mechanism usually works. Jun 10 15:10:41 mwester: ops.. probably forgot to merge some client branch... Jun 10 15:14:38 hmm, forgot to update the ncurses bitbake ui per these logging changes... Jun 10 15:14:39 * kergoth_ ponders Jun 10 15:17:16 hi - i have an OE system build with DBus for an ARM based platform (AT91RM9200) - when i try a little program using DBus i get a segmentation fault with an alignment trap - here is the backtrace from the dbus version with debugging symbols: http://pastebin.com/raw.php?i=fjEdvAuH - any ideas? Jun 10 15:18:05 JaMa|Wrk: I'm wondering if without BB_DEFAULT_TASK = "buildall" the packages in rdepends/rrecommends will not be packaged as before. Jun 10 15:19:29 neither can I remember if some sort of useflags are used to mask some undesidered deps (see gpsd>mysql) when using buildall Jun 10 15:19:56 they'll be packaged fine Jun 10 15:20:26 even without buildall? Jun 10 15:21:03 well, rdepends for sure, bitbake knows about those and includes them in its runqueue generation Jun 10 15:21:09 pretty sure rrecs too Jun 10 15:21:10 should be fine Jun 10 15:21:25 should double check rrecs, though Jun 10 15:21:26 hmm Jun 10 15:21:33 it's lot of time I don't trust opk and install all deps in my images ...should try my own feed again Jun 10 15:21:40 * kergoth_ 's never used buildall, never saw the need Jun 10 15:23:28 iirc comes from Poky Jun 10 15:24:35 we really need to figure out why cia notifications for bitbake don't work Jun 10 15:25:28 kergoth_: thx. I'll have to ask RP. http://www.mail-archive.com/bitbake-dev@lists.berlios.de/msg00490.html Jun 10 15:25:50 do_build shouldn't be executed. Jun 10 15:26:17 the image or package-index recipe will set the necessary flags to ensure that all dependent packages are generated to run its tasks Jun 10 15:26:41 so that link isn't a bug, it's working as designed Jun 10 15:26:46 afaik, anyway Jun 10 15:27:07 ah, rp explained it well in the resolve :) Jun 10 15:27:11 thought so Jun 10 15:27:30 so it depends on what you're trying to do Jun 10 15:27:47 'bitbake foo' won't package the deps of foo, but if you bitbake an image or whatever, then it would Jun 10 15:27:57 so if not using an image or similar recipe, buildall is the way to go Jun 10 15:28:15 well, border-case, then Jun 10 15:28:58 kergoth_: now I understand. thx Jun 10 15:29:57 not necessarily border case, i could see it being quite useful if you just need to build something to test it on your device, and its not something you build all the time.. in that case you definitely need all the packages Jun 10 15:29:59 it's a note for people building single binaries and wanting the deps Jun 10 15:30:02 * kergoth_ nods Jun 10 15:30:04 exactly Jun 10 15:30:14 hell, that should go in an FAQ Jun 10 15:32:27 yes, and I'd add: bitbake package-index afterward. Or? Jun 10 15:32:48 yeah, if they're using the feed rather than copying the packages manually Jun 10 15:32:50 absolutely Jun 10 15:33:07 * kergoth_ wonders if we even have an faq.. probably on the wiki somewhere :) Jun 10 15:34:06 * kergoth_ looks at prepping some more oe patches for submission Jun 10 15:34:29 oh, i remember someone in here mentioning they'd be willing to ack my 4 bitbake.conf patches, anyone recall who that was? i'd like to get those in Jun 10 15:34:46 blindvt, blindvt`: was it you? Jun 10 15:34:48 * kergoth_ scratches head Jun 10 15:35:22 well.. seems no one is against at least :) Jun 10 15:35:37 heh :) Jun 10 15:35:47 thats what usually happens with bitbake.conf and main class changes Jun 10 15:35:50 silence :) Jun 10 15:37:22 sorry, I'm still stuck trying to rebuild my gcc-cross. I was able to bake gcc-cross-initial. kergoth mentioned that gcc-cross is a multistep process. Where can I find the steps? Jun 10 15:38:06 read the recipes Jun 10 15:38:08 look at DEPENDS Jun 10 15:38:17 anyway this will do no harm.. I'll commit Jun 10 15:59:55 kergoth_: added to http://wiki.openembedded.net/index.php/Advanced_configuration Jun 10 16:01:49 hm a cool feature would be Jun 10 16:01:51 cool Jun 10 16:02:04 when the tasks could be split up more to threads Jun 10 16:02:18 for instance fetch threads could be 5 Jun 10 16:02:24 build threads 2 Jun 10 16:02:28 that would be cook Jun 10 16:02:31 yeah, i think thats on list Jun 10 16:02:34 on my list Jun 10 16:02:36 cool Jun 10 16:02:37 either by task or by resource Jun 10 16:02:39 gogo kergoth Jun 10 16:02:42 i.e. this task is disk bound Jun 10 16:03:12 woglinde, http://bit.ly/oe-wave Jun 10 16:03:24 wow nice Jun 10 16:04:00 for now I don't know what's the best setup because the thread could build,link,package,downloads etc... Jun 10 16:04:28 03Yuri Bushmelev 07org.openembedded.dev * rb8599e06e7 10openembedded.git/recipes/cmake/ (3 files in 2 dirs): cmake_2.6.4: Patch added to fix compilation when FORTIFY_SOURCE is enabled Jun 10 16:04:43 eah google-account needed? Jun 10 16:04:47 conceptually, its really the resource thats the concern, not the number of processes, so i kind of like the idea of marking up tasks with their primary resource and controlling the number by that Jun 10 16:05:23 i'd like to say okay, i have a fast cpu but mediocre disk, limit disk to 2 and cpu to 5, and i have good bandwidth, to set that to 6 Jun 10 16:06:18 cool Jun 10 16:07:25 * kergoth_ shrugs Jun 10 16:07:28 we can use num of cpu, incoming internet bandwidth, i/o bandwidth from some know test (hdparm -T?) Jun 10 16:07:30 something to give more thought to Jun 10 16:07:42 hmm thats an idea Jun 10 16:07:49 this will be easier to configure for ordinary user Jun 10 16:07:54 could actually add a separate script to run system tests and emit a site.conf Jun 10 16:08:00 so itd be one time operation Jun 10 16:08:09 could also emit ASSUME_PROVIDED from that Jun 10 16:08:12 hmmm :) Jun 10 16:08:15 ;) Jun 10 16:08:16 yeah Jun 10 16:08:21 cool idea jay7 Jun 10 16:08:29 I can test it even ;) Jun 10 16:08:38 pfft, start coding ;) Jun 10 16:08:46 seems I have strangest env here ;) Jun 10 16:08:49 *g* Jun 10 16:16:20 well.. you can get num of CPUs from /proc/cpuinfo (and look there for HT extension), i/o bandwidth from hdparm -t (requires root) Jun 10 16:16:59 jay7 right Jun 10 16:17:02 not sure about inet bandwidth Jun 10 16:17:14 hm maybee one could use iotop Jun 10 16:17:16 git pull? :) Jun 10 16:17:16 somehow Jun 10 16:17:24 untaring it Jun 10 16:17:28 kernelsource Jun 10 16:17:30 for instance Jun 10 16:17:33 hm Jun 10 16:17:48 I am at creating ideas too Jun 10 16:17:56 do timings for fetching kernel sources and for untaring it :) Jun 10 16:18:07 hehe Jun 10 16:18:17 to bad gsoc is over Jun 10 16:18:25 now we got a good idea Jun 10 16:19:04 bad thing that I don't know python :) Jun 10 16:19:09 uh Jun 10 16:19:12 learn it Jun 10 16:19:20 isnt that hard Jun 10 16:19:27 one time I'll do :) Jun 10 16:19:37 and there are tons of snipets how to do things the right way Jun 10 16:19:53 there are so many modules you mostly dont have to invent it Jun 10 16:20:04 only a sane netlink-wrapper is missing Jun 10 16:26:18 * kergoth_ experiments with changing the warnings module behavior for bitbake Jun 10 16:27:15 think i've got it so it only shows python warnings (i.e. DeprecationWarning) for the bb module, oe module, and ones from (i.e. python snippets, tasks in oe) Jun 10 16:28:12 and it sends them to a "BitBake.Warnings" logger with the logging module, so it can be controlled there from the ui Jun 10 16:28:14 hmmm Jun 10 16:34:13 well.. neon is still failing for me.. Jun 10 16:34:24 will investigate :( Jun 10 16:49:07 ah.. that was kerberos Jun 10 16:49:21 CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/krb5 Jun 10 16:52:08 ~lart adobe for not providing flash-previewplugin 10.1 for 64bit Jun 10 16:52:09 * ibot puts on some milking gloves. "All right, now, adobe for not providing flash-previewplugin 10.1, this won't hurt a bit...." for 64bit Jun 10 16:57:27 * Jay7 wonders why only he is having some problems with kerberos and neon Jun 10 16:57:47 maybee we dont have kerberos dev stuff installed Jun 10 16:58:06 all you here around? :) Jun 10 16:58:31 gentoo pplz should have imho Jun 10 17:00:17 kergoth_: mind acking my patch.py patch :) Jun 10 17:00:30 k Jun 10 17:01:55 sent Jun 10 17:05:25 hmm, bug Jun 10 17:05:36 if you chmod 000 local.conf, it doesn't error Jun 10 17:05:44 it just silently doesn't parse it, then errors later on due to shit being unset Jun 10 17:19:44 kergoth_: heh Jun 10 17:20:01 jo khem Jun 10 17:20:08 kergoth_: if (cant read local.conf) error please Jun 10 17:20:12 woglinde: hello Jun 10 17:20:38 yeah, exactly Jun 10 17:20:42 added it to the bitbake TODO Jun 10 17:21:04 * kergoth_ gets annoyed and finally fixes it to exit gracefully when you ^C during the initial parsing Jun 10 17:21:42 kergoth_: merci Jun 10 17:22:08 *trying* to make things a bit more sane while converting to the logging module Jun 10 17:22:13 but its a twisty maze Jun 10 17:22:56 kergoth_: you should put the todo in public so if someone dares can help you to fix some Jun 10 17:23:11 its in master Jun 10 17:23:14 there's a TODO file now Jun 10 17:23:18 oh cool Jun 10 17:23:21 added it yesterday or the day before :) Jun 10 17:23:44 http://git.openembedded.org/cgit.cgi/bitbake/tree/TODO Jun 10 17:23:55 I think I should put my compiler shoes back and hack on bitbake Jun 10 17:28:10 hmmm Jun 10 17:28:16 wonder how logger.exception() formats its output Jun 10 17:28:20 wonder if we should use it Jun 10 17:29:30 khem, 5 minutes due to the patch change? seriously? Jun 10 17:29:38 that's a pretty nice decrease Jun 10 17:29:41 kergoth_: I dont think Jun 10 17:29:45 its all that Jun 10 17:29:55 but I did simple time bitbake Jun 10 17:30:10 and it took 3hrs50min Jun 10 17:30:13 and 3hr45min Jun 10 17:30:24 * kergoth_ nods Jun 10 17:30:26 but its probably not all due to patch I am sure Jun 10 17:30:32 certain amount of run to run deviation always Jun 10 17:30:43 but it takes so long to run multiple timings Jun 10 17:30:54 hmm, could do bitbake -c patchall foo Jun 10 17:30:56 kergoth_: yeah Jun 10 17:30:56 and time that way Jun 10 17:31:02 no patchall exists atm, but its easy enough to add Jun 10 17:31:07 I am sure the automatic updater must have kicked in Jun 10 17:31:20 and other random automactic stuff Jun 10 17:31:23 * kergoth_ 'll see about fixing the unpack copying patches into workdir thing today Jun 10 17:31:27 but it does not make it worse Jun 10 17:31:39 kergoth_: that will be awsome Jun 10 17:39:33 okay, i'm also sick of knotty bitching about the log file not existing Jun 10 17:39:39 * kergoth_ fixes Jun 10 17:40:10 also: god, do we print the same shit enough? Jun 10 17:40:13 i know the function failed Jun 10 17:40:19 do i really need to see 6 error messages about it? Jun 10 17:43:56 okay, i'm going to put off investigating that one Jun 10 17:43:58 one step at a time.. Jun 10 17:43:59 heh Jun 10 18:33:32 okay, screw it, i'm pushing those bitbake.conf changes Jun 10 18:35:32 03Chris Larson 07org.openembedded.dev * r29e61b1d38 10openembedded.git/conf/bitbake.conf: Jun 10 18:35:32 bitbake.conf: don't filter out nonexisting paths from FILESPATH Jun 10 18:35:32 It's unnecessary, and can cause problems with amend.bbclass. Jun 10 18:35:32 Signed-off-by: Chris Larson Jun 10 18:35:39 03Chris Larson 07org.openembedded.dev * r52ef926522 10openembedded.git/conf/bitbake.conf: Jun 10 18:35:39 bitbake.conf: make the EXTENDPE bits slightly less disturbing Jun 10 18:35:39 Signed-off-by: Chris Larson Jun 10 18:35:40 03Chris Larson 07org.openembedded.dev * r4d060ecf3d 10openembedded.git/conf/bitbake.conf: Jun 10 18:35:40 bitbake.conf: Drop some unnecessary getVar usage in python snippets Jun 10 18:35:41 Signed-off-by: Chris Larson Jun 10 18:35:42 03Chris Larson 07org.openembedded.dev * r9004bd8f2e 10openembedded.git/conf/bitbake.conf: Jun 10 18:35:42 bitbake.conf: use := for BUILD_OS and BUILD_ARCH Jun 10 18:35:42 These don't change, so their initial values should be just fine. Apparently, Jun 10 18:35:43 in a typical build, those os.uname() snippets are called around 46k times, Jun 10 18:35:43 which seems a tad excessive :) Jun 10 18:35:44 Signed-off-by: Chris Larson