**** BEGIN LOGGING AT Tue Apr 07 02:59:58 2009 Apr 07 06:50:35 morning Apr 07 06:56:27 ~logs Apr 07 06:56:30 All conversations are logged to http://ibot.rikers.org/channel, where "channel" is replaced by the URL-encoded channel name, such as %23freenode for #freenode. Lines starting with spaces are not logged. Apr 07 07:15:43 good morning Apr 07 07:43:19 is there anyone who recently compiled a version of gsoap newer than 2.7.7 with bitbake? Apr 07 07:48:13 03Koen Kooi  07org.openembedded.dev * rfbf22c52a2 10openembedded.git/ (8 files in 3 dirs): linux-omap 2.6.29: add patches to fix some bugs in the display driver, XV now works again Apr 07 08:25:09 http://docs.openembedded.org/usermanual/usermanual.html seems to be offline? could someone have a look at this? Apr 07 08:25:43 yes, i see a blank page Apr 07 08:25:57 http://docs.openembedded.org/usermanual/html/ works fine Apr 07 08:26:09 I gues that should be the same Apr 07 08:27:16 tsjsieb: yes it's the same, but the wiki pages link to the usermanual.html one Apr 07 08:27:26 tsjsieb: anyway, thanks! Apr 07 08:42:25 good morning Apr 07 08:42:30 morning Apr 07 08:42:42 03Michael 'Mickey' Lauer  07org.openembedded.dev * r6c5e86faf0 10openembedded.git/recipes/libssh/ (libssh-native_0.2.bb libssh_0.11.bb libssh_0.2.bb): libssh: add libssh-native Apr 07 08:42:44 03Michael 'Mickey' Lauer  07org.openembedded.dev * r8658973a56 10openembedded.git/ (5 files in 3 dirs): socat: catch up SRC_URI Apr 07 08:42:55 03Michael 'Mickey' Lauer  07org.openembedded.dev * rb4c72a2368 10openembedded.git/ (3 files in 2 dirs): gabriel[-native]: new recipe; proxy for remote dbus access Apr 07 08:44:06 03Michael 'Mickey' Lauer  07org.openembedded.dev * r6fb6092676 10openembedded.git/recipes/tasks/task-cli-tools.bb: task-cli-tools: add socat Apr 07 08:46:28 has someone already tried kergoth's patch regarding collections management? Apr 07 09:00:44 florian: good morning Apr 07 09:00:47 hi lrg_ Apr 07 09:01:02 mpoullet|work: He was talking about that a few days ago, but I don't know whether he released that yet. Apr 07 09:06:29 hey pb_ Apr 07 09:09:31 has anyone already tried to bitbake gsoap > 2.7.7 ? Apr 07 09:11:11 is there a way to tell the fetchers to bypass SRC_TARBALL_STASH? Apr 07 09:11:49 because on an intranet it makes no sense pulling tarballs Apr 07 09:12:12 I just want the recipes to fetch from the git/svn server directly Apr 07 09:21:00 03Michael 'Mickey' Lauer  07org.openembedded.dev * r1157bf509e 10openembedded.git/ (5 files in 2 dirs): vala-native: bump to 0.6.0 and make git version the default; we need a bootstrap version now though. Apr 07 09:40:47 mickeyl: Was there any problem with my key? Apr 07 09:41:32 only in so far that i never received one Apr 07 09:41:37 where did you send it ot? Apr 07 09:41:43 oh. keys.. Apr 07 09:42:08 I sent you a mail on Saturday. Apr 07 09:42:16 <20090404144041.GB6474@sirena.org.uk> Apr 07 09:42:20 mickeyl: I'll send my key to you now :) Apr 07 09:43:34 03Koen Kooi  07org.openembedded.dev * rf2b04954af 10openembedded.git/recipes/e17/systray_svn.bb: systray: as the name implies, a systray module for e17 Apr 07 09:44:40 03Koen Kooi  07org.openembedded.dev * ra28b666653 10openembedded.git/contrib/angstrom/build-feeds.sh: angstrom feed builder: add systray Apr 07 09:52:26 hmm Apr 07 09:52:37 did anything change wrt toolchain? Apr 07 09:52:51 after updating i no longer can build from scratch Apr 07 09:52:55 looks like it doesn't build a toolchain Apr 07 09:53:00 and then complains about missing arm-foo-gcc Apr 07 09:53:40 ok, it's something wrt. DEPLOY_DIR Apr 07 09:53:53 packaged staging still doesn't quit cut it here Apr 07 09:53:56 hi Apr 07 09:53:59 mickeyl: ping :-) Apr 07 09:54:05 yo? Apr 07 09:54:18 mickeyl: Did you see what I said about the efl UTF input issues we had? Apr 07 09:54:19 * mickeyl sighs and removes DEPLOY_DIR completely Apr 07 09:54:21 I don't know if you have them Apr 07 09:54:31 i'm not doing much UI these days Apr 07 09:54:41 so, no, i don't have any problems i can't remember Apr 07 09:54:45 s/n't// Apr 07 09:54:51 Well we weren't able to input non ASCII characters Apr 07 09:54:59 Which was doable on GTK based stuff Apr 07 09:55:14 This was due to the lack of libx11-locale & localedef packages Apr 07 09:55:23 Which e seems to use to do his mapping Apr 07 09:56:03 mickeyl: just thought it might help you anyway :-) Apr 07 09:56:22 mickeyl: and when you'll have some free time, I'd like to talk about opimd state (as told on byndmko) Apr 07 09:57:06 ok, so we just have to install the as part of the x11 task? Apr 07 09:57:21 mickeyl: Well I'd say it'd be more logical to put it as a dependency of e Apr 07 09:57:32 I could have done that but as it has impact, I didn't Apr 07 09:57:36 Plus I don't have net at home right now :-) Apr 07 09:57:38 hmm Apr 07 09:57:44 i think x11 shoudl RRECOMMEND them Apr 07 09:58:01 Well it's not a necessary dependency Apr 07 09:58:18 yep Apr 07 09:58:23 RRECOMMENDS therefore, not RDEPENDS Apr 07 09:59:41 what are you guys doing to keep your toolchain intact when removing tmp? Apr 07 10:00:29 Errr we don't :-) Apr 07 10:00:51 anyone else does? Apr 07 10:01:00 i thought that's what we had packaged staging for Apr 07 10:01:11 but my crossdir still seems to be inside tmp Apr 07 10:01:17 so it's gone Apr 07 10:01:51 moreoever, if I move DEPLOY_DIR, it looks like it _thinks_ it already has a toolchain and no longer builds one Apr 07 10:01:58 which means nothing builds Apr 07 10:02:01 after rm -rf tmp Apr 07 10:02:17 mickeyl: could you have a look at renaming zopeinterface to python-zoperinterface? I tried, but did not succeed and I guess for you it is a simple job Apr 07 10:03:13 it looks someone already did that Apr 07 10:03:20 right now everything is broken here Apr 07 10:03:28 and i have no idea why Apr 07 10:03:34 hmmm Apr 07 10:03:39 i'd like to fix that first Apr 07 10:03:51 good plan, thanks anyway Apr 07 10:05:13 CROSS_DIR is hardset in bitbake.cof Apr 07 10:05:18 is anyone moving this? Apr 07 10:05:27 args how I disbale the blinking cursor in gnome-terminal? Apr 07 10:05:31 to keep the toolchain intact? Apr 07 10:06:01 hmm Apr 07 10:06:03 oh the silence Apr 07 10:06:05 ah well Apr 07 10:06:43 hi, anyone interested in testing the recipe for gsoap version 2.7.13 I created just yet? Apr 07 10:19:53 hi Apr 07 10:20:38 mickey|bbl: good morning Apr 07 10:20:42 ah, doh, too late Apr 07 10:21:38 woglinde: im trying to get gxine working, nothing provides libxine-plugin-decode-image, isnt libxine supposed to provide libxine-plugin-decode-image? Apr 07 10:24:16 hi pb Apr 07 10:25:43 03Raffaele Recalcati  07org.openembedded.dev * re7f45a4a42 10openembedded.git/ (5 files in 4 dirs): linux-omap: add modeline for 720p60 to .28 and .29 Apr 07 10:26:27 vsr hm dont know Apr 07 10:26:39 maybe its an older plugin which git deleted Apr 07 10:26:53 vsr are there newer versions of gxine? Apr 07 10:29:11 hi woglinde Apr 07 10:32:18 woglindge: there are new version of gxine as well Apr 07 10:33:09 woglindge: i will try with that and keep you updated Apr 07 10:34:07 woglinde: i removed that dependency and got gxine built and installed, but it gives me segmentation fault when i run it Apr 07 10:34:16 vsr uh oh Apr 07 10:34:37 hm Apr 07 10:34:41 could you run gdb Apr 07 10:34:46 to see where it crashs Apr 07 10:34:49 i did that as well Apr 07 10:35:03 hm but I have to go hunting lunch Apr 07 10:35:15 till later Apr 07 10:35:22 okay Apr 07 10:35:41 it crashes in pthread_mutex call Apr 07 10:36:03 uclibc? Apr 07 10:36:11 libc Apr 07 10:37:06 for some reason i could not get debug symbols built into gxine/libxine even though i specified CFLAGS+= "-g" in .bb files Apr 07 10:38:16 okay, later Apr 07 10:55:16 morning Apr 07 10:56:30 'morning Apr 07 10:56:34 hi hrw Apr 07 11:01:59 woglinde: I've created a recipe and a patch for gsoap 2.7.13, what do I do with it now (I mean, could this possible end up in the repo?) Apr 07 11:03:14 virk send it to -devel mailinglist for review Apr 07 11:05:18 by the way whats gsoap doing? Apr 07 11:07:40 is is a C/C++ toolkit for soap: http://gsoap2.sourceforge.net/ Apr 07 11:12:09 btw, I am quite new to mailing lists, do I need to attach a tarball, or just append every file in plaintext to the email? Apr 07 11:13:29 virk please read the git page Apr 07 11:13:38 git format-patch is your friend Apr 07 11:13:42 hi mrnic Apr 07 11:13:43 e Apr 07 11:14:15 woglinde hi Apr 07 11:14:56 woglinde: gxine error: Program received signal SIGSEGV, Segmentation fault. 0x40037354 in pthread_mutex_lock () from /lib/libpthread.so.0 Apr 07 11:15:14 do you think moving to a new version of gxine will solve the problem Apr 07 11:16:14 hms 130 vsr hehe probably yes Apr 07 11:16:21 but cant promise Apr 07 11:20:42 what is hms 130? Apr 07 11:21:15 aeh Apr 07 11:21:16 sorry Apr 07 11:21:23 only hms Apr 07 11:21:33 something like hmmmmmm Apr 07 11:24:12 how do tell oe not to build perl-native but use the host perl ?? Apr 07 11:26:23 ah think i got it Apr 07 11:26:35 rob_w: you know that it is not granted to work? Apr 07 11:26:35 03Koen Kooi  07org.openembedded.dev * r43fbe6ad1d 10openembedded.git/recipes/kexecboot/initramfs-kexecboot-image.bb: initramfs-kexecboot-image: fix parse warning Apr 07 11:26:54 local perl you mean ? Apr 07 11:27:27 rob_w: building perl without our perl-native Apr 07 11:27:35 damn Apr 07 11:27:37 ok Apr 07 11:27:49 perl build system is sick Apr 07 11:28:15 hrw definitly Apr 07 11:28:58 hmm perl-native broke in hte first place missing a parameter at a dbmclose() call .. so i made it dbmclose(NULL) which worked .. then i was missing XML:something for intltool Apr 07 11:34:10 03Jeremy Lainé  07stable/2009 * r8f3a7a8b73 10openembedded.git/recipes/cups/ (cups.inc cups_1.2.12.bb cups_1.3.8.bb): Apr 07 11:34:10 cups: use LDFLAGS for shared libs (fixes QA error) Apr 07 11:34:10 Signed-off-by: Jeremy Lainé Apr 07 11:34:10 Acked-by: Koen Kooi Apr 07 11:34:12 03Jeremy Lainé  07stable/2009 * r0a8ecc41d5 10openembedded.git/recipes/libmng/ (libmng/ldflags.patch libmng_1.0.8.bb): Apr 07 11:34:15 libmng: use LDFLAGS (fixes QA error) Apr 07 11:34:17 Signed-off-by: Jeremy Lainé Apr 07 11:34:19 Acked-by: Koen Kooi Apr 07 11:34:21 03Jeremy Lainé  07stable/2009 * rb8ca80f286 10openembedded.git/recipes/dhclient/dhclient_2.0pl5.bb: Apr 07 11:34:24 dhclient: use LDFLAGS (fixes QA error) Apr 07 11:34:26 Signed-off-by: Jeremy Lainé Apr 07 11:34:28 Acked-by: Koen Kooi Apr 07 11:34:32 03Jeremy Lainé  07stable/2009 * r4c864ef77d 10openembedded.git/recipes/pth/ (files/ldflags.patch pth_2.0.2.bb): Apr 07 11:34:35 pth: use LDFLAGS (fixes QA error) Apr 07 11:34:37 Signed-off-by: Jeremy Lainé Apr 07 11:34:39 Acked-by: Koen Kooi Apr 07 11:35:10 Jeremy 'The LDFLAGS fix' Laine :) Apr 07 11:47:18 wow, that git thing is huge... Apr 07 11:47:44 hm? Apr 07 11:47:55 are you sure you did it the right way? Apr 07 11:47:59 hi booxter Apr 07 11:48:18 hi woglinde Apr 07 11:48:49 any docs on SRCREV_FORMAT thing? It makes me cry, I think I don't understand the concept Apr 07 11:48:53 woglinde: I haven't done anything yet, I'm just reading the git docs, and it is taking me ages Apr 07 11:50:13 virk hehe okay Apr 07 11:53:15 for some reason package gpe-image-1.0-r0: task do_rootfs: started takes more than a hour & still not over, i clean and tried. i dont know what could be the problem Apr 07 11:53:20 I'm stuck now, I'm getting confused on this checkout, push, branche, ... stuff Apr 07 11:53:31 hm? Apr 07 11:53:43 read the site on openembedded.org Apr 07 11:53:48 gitphrasebook Apr 07 11:53:52 should be enough Apr 07 11:53:53 that's what I am doing Apr 07 11:53:57 for the first steps Apr 07 11:54:50 03Koen Kooi  07org.openembedded.dev * re1d6691193 10openembedded.git/ (conf/distro/include/sane-srcrevs.inc recipes/e17/e-wm_svn.bb): Apr 07 11:54:50 e17: bump SRCREV to 39886 Apr 07 11:54:50 * runtime tested on beagleboard, feels a bit faster as before Apr 07 11:54:50 * also gets us the blingy new wallpaper2 dialog Apr 07 11:57:36 so i do have XML/Parser.pm in my staging/x86_64-linux/usr/lib/perl/5.8.8/ but intltool cant see it ?! Apr 07 11:58:19 any clue what could be wrong, what would make task do_rootfs go on and on for a long time? Apr 07 12:01:38 any way to know what task_do_rootfs is doing at present, i dont see any log files in there Apr 07 12:02:48 vsr: is it doing the testlab? Apr 07 12:04:31 i dont see a testlab in deploy directory Apr 07 12:05:41 my log.do_package_stage_all is empty Apr 07 12:07:25 i dont see a run.do_rootfs either Apr 07 12:11:08 * virk has clearly done something wrong: git format-patch is spawning patches as it was raining cats and dogs... Apr 07 12:12:29 virk: git format-patch .. ; where commit id was the last commit before you made the change Apr 07 12:12:56 how to force regeneration of ipks for every package? Accidentally I deleted some of them and can't enforce rootfs rebuilding without it... Apr 07 12:16:38 booxter hm Apr 07 12:16:44 only rebuild the package Apr 07 12:19:20 booxter: rm `find tmp/stamps/ -name *package_write`;bitbake image? Apr 07 12:19:41 yeap! tnx Apr 07 12:29:37 question: how many serial ports your computers have? Apr 07 12:30:12 question to everyone? Apr 07 12:30:21 yes Apr 07 12:30:24 2 Apr 07 12:30:26 but Apr 07 12:30:36 only headers on the motherboard Apr 07 12:30:37 unfortunately only 1 Apr 07 12:30:41 and one is only for ir Apr 07 12:30:54 none, I use a usb<->rs232 converter Apr 07 12:31:08 I have one on mainboard, 2 on pci card and one usb<>rs232 Apr 07 12:31:51 so had 4 of them (took mainboard one for beagleboard) Apr 07 12:32:14 and now I am considering buying 6xrs232 pci card instead of current 2xcom+1xlpt one Apr 07 12:32:27 insane... Apr 07 12:37:11 * virk is stupid: format-patch does only work when committing locally (which I thought was impossible) Apr 07 12:43:51 vsr: can I just send a patch like this git-send-email --to openembedded-devel@lists.openembedded.org 0001-Added-gsoap-version-2.7.13.patch Apr 07 12:46:30 03Koen Kooi  07stable/2009 * rdeaae0aca8 10openembedded.git/contrib/angstrom/local.conf: Apr 07 12:46:30 angstrom local.conf: change /OE -> ${OETREE} Apr 07 12:46:30 Signed-off-by: Koen Kooi Apr 07 12:48:48 virk: sorry i dont use git-send-email Apr 07 12:49:34 haha, me neither, it seems I do nt even have that command on ubuntu (neither after installing git-email) Apr 07 13:06:59 virk, remove the hyphens. Hyphenated git commands have been depreciated Apr 07 13:07:29 virk, i.e. "git send email ..." Apr 07 13:07:31 03Koen Kooi  07org.openembedded.dev * r687a7700ea 10openembedded.git/recipes/linux/ (2 files in 2 dirs): linux-omap 2.6.2[89]: make 720p mode conform to CEA 681-E instead of CEA 681-D Apr 07 13:11:40 MWelchUK: thanks, I noticed just yet, though there's still something wrong with my setup, I'm getting this message: Can't call method "command" on an undefined value at /usr/lib/git-core/git-send-email line 770. Apr 07 13:25:55 dammit, took me almost 3 hours to send that patch for gsoap Apr 07 13:35:38 I've just sent a patch to the -devel mailing list but it says I've been rejected.. Apr 07 13:41:13 virk: try to send from subscribed address :) Apr 07 13:51:14 Jay7: thanks, it should however be noted that opensource-land has an extremely steep learning curve Apr 07 13:51:54 hehe.. that's true :) Apr 07 14:09:42 hi, is anyone planning to merge qt-extended-improved into OE ? Apr 07 14:09:52 or rather ... is anyone actually working on it ? Apr 07 14:10:59 dang sorry ... did I miss something ? Apr 07 14:14:18 "ImportError: No module named _md5" - is it that error with python-native? Apr 07 14:14:30 Jay7: rebuild openssl-native Apr 07 14:14:57 hrw: thanks :) Apr 07 14:19:32 hrw, hi ... any idea about that ? Apr 07 14:19:37 (QtEI) Apr 07 14:21:33 marex: nope Apr 07 14:27:26 hi mickeyl Apr 07 14:27:44 hey hrw Apr 07 14:27:46 how's life? Apr 07 14:28:09 mickeyl: fine Apr 07 14:28:30 mickeyl: I am wondering about buying 6xCOM pci card due to lack of serial ports Apr 07 14:28:36 hmm Apr 07 14:28:38 you can do that Apr 07 14:28:41 i solved it differently Apr 07 14:28:49 usb hub + usb->serial Apr 07 14:28:57 a bit more flexible when you travel a lot Apr 07 14:29:24 mickeyl: I have one ftdi adapter for travels Apr 07 14:29:29 * Crofton|work is using usb serial almost exclusively now Apr 07 14:29:55 mickeyl: I have at91sam9263ek, atngw100, beagleboard, bug on desk and each of them needs serial Apr 07 14:30:08 hrw, usb :) Apr 07 14:30:10 hrw: heh, that's quite a lot Apr 07 14:30:28 Crofton|work: 6xcom is cheaper then 6 usb<>rs232 ;D Apr 07 14:30:39 ah Apr 07 14:30:50 mickeyl: need to buy AT/ATX psu to power all of them ;D Apr 07 14:31:43 usb-serial dongles tend to be teh suck anyway. Apr 07 14:31:57 about half the ones that I've used seem to malfunction in various weird ways. Apr 07 14:32:11 hrw: how much does that 6xcom cost? Apr 07 14:32:22 hehe Apr 07 14:32:28 and I read too many complains about pl2303 ones... two of them works but not 3 at same time Apr 07 14:32:38 virk: 89pln which is ~20€ Apr 07 14:33:08 i have a strange problem, when i try to build images, it compiles, packages, when gets struck up in task_rootfs, im not sure what could be the problem, each time is just hangs and does nothing in task_rootfs. is there a way to debug this Apr 07 14:33:21 that is cheap indeed, I always buy my usb->serial dongles at dealextreme where they're less than 5$ each Apr 07 14:33:32 cause i dont want to delete my tmp directory as it will put me back by a day Apr 07 14:34:48 i've been using these pl2303 thingies most of the time. In fact, i find them a lot better than those ftdi ones Apr 07 14:37:37 morning Apr 07 14:37:51 hi kergoth Apr 07 14:38:48 * kergoth thinks about pulling in ncurses 5.7 Apr 07 14:39:23 is there any way to see what task_dorootfs is exactly doing? Apr 07 14:39:54 look at the run. scripts in the temp/ dir in the workdir for the recipe in question, or build with -D Apr 07 14:41:36 anyone ever seen 'tic' go crazy during an ncurses do_install? Apr 07 14:41:44 i dont see a run.do_rootfs getting created even though i get a messages saying task do_rootfs started Apr 07 14:41:47 * kergoth thinks this seems familiar Apr 07 14:42:00 vsr: in that case, it's probably a python script. read the .bbclass Apr 07 14:42:15 kergoth: yeah, though I can't remember what the cause was Apr 07 14:42:16 heh Apr 07 14:42:22 heh, me neither Apr 07 14:42:23 hrmph Apr 07 14:42:47 ncurses 5.4 has some autoconf annoyances, uses a macro that no longer exists in autoconf.. upgrading gets me lovely 100% cpu usage by tic :) Apr 07 14:48:12 hmm, i havent made any change to openembedded/classes, i really dont know Apr 07 14:48:32 will rebuilding python affect this? Apr 07 14:50:14 gcc 4.3.3 doesn't seem to be on the angstrom sources site yet... Apr 07 14:50:50 how do I make angstrom use a different source site for it? Apr 07 14:51:17 dhd: OE fetch first from normal sites so why it is a problem? Apr 07 14:51:36 oh, I see, gcc.releasenotes.org is unresolvable Apr 07 14:51:46 that's the only other mirror the bbfile has for it Apr 07 14:52:09 then use other GNU_MIRROR? Apr 07 14:52:18 gcc-4.3.3.inc:SRC_URI = "${GNU_MIRROR}/gcc/gcc-${PV}/gcc-${PV}.tar.bz2 \ Apr 07 14:52:19 yeah I will just edit the bbfile, ok Apr 07 14:52:26 ah Apr 07 14:52:36 the version in the stable/2009 branch hardcodes gcc.releasenotes.org Apr 07 14:52:42 that doesn't seem right Apr 07 14:53:24 pb_: ah, tic goes nuts if the wrong version of tic gets run. i.e. ncurses 5.4 tic against ncurses 5.7. good to know Apr 07 14:53:39 03Koen Kooi  07org.openembedded.dev * r936d7a45a1 10openembedded.git/ (3 files in 2 dirs): Apr 07 14:53:39 gnome-bluetooth: add 2.27.1 Apr 07 14:53:39 libunique: needed for gnome-bluetooth, add 1.0.8 Apr 07 14:53:40 * kergoth thinks they should make it at least error out in that case rather than hanging at 100% cpu Apr 07 14:54:04 aha Apr 07 14:54:11 yeah, that is a bit sad Apr 07 14:54:30 now I get this problem: ERROR: gcc-cross-initial-4.3.3: ftp://ftp.gnu.org/gnu/gcc/gcc-4.3.3/gcc-4.3.3.tar.bz2 has no entry in conf/checksums.ini, not checking URI Apr 07 14:54:42 (sorry, I'm not super experienced with bitbake yet) Apr 07 14:55:08 I suppose I should just add an entry to checksums.ini Apr 07 14:55:13 or merge the one from the .dev branch Apr 07 14:55:51 oh, the checksums.ini uses /pub/gnu rather than /gnu, ok Apr 07 14:56:25 should bitbake.conf or checksums.ini be fixed? Apr 07 15:06:19 experimented with a recipe in the .git source tree of a package.. works pretty well once you override a few vars Apr 07 15:16:16 kergoth: what about your patch for managing collections, any chance that it goes upstream? Apr 07 15:16:54 aye, haven't had time to mess with it. how many +1's are you supposed to get on a proposal before pushing it, anyway? Apr 07 15:17:26 khem: gcc 4.3.3 builds uclibc in thumb mode without errors Apr 07 15:17:36 kergoth: I find this very useful, but I don't think that my +1 counts :) Apr 07 15:17:46 hehe Apr 07 15:18:09 kergoth: I'd like to have it in stable too Apr 07 15:18:21 zecke seemed to have concerns about the implementation, but he didn't reply to my latest comments Apr 07 15:18:54 * kergoth thinks collections.inc is a solid usability improvement, and we need as many of those as we can get Apr 07 15:18:58 because in my company's setup I've to use different overlays and it could help a lot Apr 07 15:19:37 anyone have any good diagrams of bitbake/oe, how the pieces fit together and such? guess i should check the presentations in the repo, maybe they have that Apr 07 15:20:28 whoops, that was the power cord Apr 07 15:20:34 hehe Apr 07 15:20:48 hehe Apr 07 15:21:05 kergoth: there's an old one in my first fosdem presentation Apr 07 15:21:22 and a new one in the one for foss.aalborg09 Apr 07 15:21:31 cool, thanks Apr 07 15:22:21 * mickeyl checks download site Apr 07 15:23:02 ~lart ltg Apr 07 15:23:02 * ibot sends a legion of lawyers after ltg's head Apr 07 15:23:50 * kergoth wonders where to find those presentations Apr 07 15:24:02 mickeyl: can i get your opinion on http://thread.gmane.org/gmane.comp.handhelds.openembedded/22305 ? wondering whether or not to merge it to dev Apr 07 15:24:12 * kergoth taps foot while waiting for test builds Apr 07 15:25:48 kergoth: i deleted my open embedded repo and tried to build an image all over again, it again gets struck at task do_rootfs: started and i dont see a run_do rootfs file in temp directory Apr 07 15:25:53 what could be wrong? Apr 07 15:26:16 i just told you that. it's a python script, not shell, so there won't be a run. script in the temp directory. once again, to see what its doing, read the bbclass Apr 07 15:27:10 .bb class are stored in classes directory and they are not modified Apr 07 15:27:40 yes, and? Apr 07 15:27:58 so how will something go wrong Apr 07 15:28:10 http://linuxtogo.org/~mickeyl/tools/ Apr 07 15:28:12 FOSDEM2005.pdf Apr 07 15:28:14 FOSDEM2007.pdf Apr 07 15:28:16 FossAalborg.pdf Apr 07 15:28:27 first of all, you haven't even said what *is* going wrong, all you've done is asked how to tell what do_rootfs is doing Apr 07 15:28:36 mickeyl: thanks much Apr 07 15:28:48 http://thread.gmane.org/gmane.comp.handhelds.openembedded/22305 looks pretty good, actually Apr 07 15:28:54 mickeyl: what license are those under? Apr 07 15:28:56 i think it would solve some of the problems people have with collections Apr 07 15:28:58 cc-by-sa Apr 07 15:29:07 okay, i asked that since do_rootfs takes a lot of time and does nothing Apr 07 15:29:22 ugh actually gcc 4.3.3 is not building - it fails with the "Link tests are not allowed after GCC_NO_EXECUTABLES" error Apr 07 15:29:56 i wanted to know since now i reliaze that my python skills are zero, i would like to fix the problem Apr 07 15:30:09 mickeyl: news from packaged-stage? Apr 07 15:30:32 vsr: ahhh, i see. hmm. that's odd. which image is this you're buliding? Apr 07 15:30:34 ant_work: none that i know of. i'm afraid my mail goes under now that oe-devel is a patch list Apr 07 15:30:54 i tried a few gpe-image, base-image Apr 07 15:31:02 i need to add a mail filter for oe patches, otherwise there's no way to have discussions going on there Apr 07 15:31:12 it was working all while something happened in between i dont remember Apr 07 15:31:15 mickeyl: this was partially discussed in march Apr 07 15:31:21 btw, mickeyl, did you recived my key? :) Apr 07 15:31:25 Jay7: got it Apr 07 15:31:29 ok :) Apr 07 15:31:34 mickey: before last patch-flood Apr 07 15:31:41 kergoth: i just dont want to delete my tmp directory cause it will put me back by a few days Apr 07 15:32:32 it compiles, packages and gets struck in do_rootfs Apr 07 15:32:47 mickeyl: after all these QA fixes only a few remain: http://rafb.net/p/eyM7tJ91.html Apr 07 15:33:06 that's not too bad! Apr 07 15:33:23 bbiab, my turn to clean the bathroom *sigh* Apr 07 15:33:27 ant_work: nice Apr 07 15:33:47 kergoth: we'll have to hire Jeremy for another morning ;-) Apr 07 15:33:59 another beer... Apr 07 15:42:10 any ideas i could try? Apr 07 15:54:36 03Koen Kooi  07org.openembedded.dev * r30cdb3a5af 10openembedded.git/recipes/gnome/gnome-bluetooth_2.27.1.bb: gnome-bluetooth: add RCONFLICTS with bluez-gnome Apr 07 16:20:51 Hello Apr 07 16:20:57 he otavio Apr 07 16:21:00 hey otavio Apr 07 16:21:00 Did someone saw my RFH on ml? Apr 07 16:21:19 kergoth: :-) Apr 07 16:21:22 woglinde: :-) Apr 07 16:21:33 Tartarus: ping Apr 07 16:22:04 hi kergoth Apr 07 16:22:32 http://article.gmane.org/gmane.comp.handhelds.openembedded/22510 Apr 07 16:22:35 Tartarus: I was able to figure the curl issue few days back, but got preempted by personal matters... Apr 07 16:22:39 Could someone take a look at it? Apr 07 16:22:53 It is quite strange for me and I have no guess where is the cause of it Apr 07 16:24:18 denix: hey.. can I talk to you private? :) Apr 07 16:24:40 Jay7: sure Apr 07 16:24:49 Jay7: pm me Apr 07 16:25:52 hi, which xserver are we suppossed to choose for an embedded device with touch, mouse and keyboard support, xserver-kdriver or x.org Apr 07 16:26:04 any of them Apr 07 16:26:19 kdrive will be easier to get running if you never did that before Apr 07 16:26:26 wont it make a difference in size? Apr 07 16:26:49 kdrive is lighter i guess Apr 07 16:26:50 grr: not too much Apr 07 16:27:41 i got this doubt cause some use x.org and some use xserver-kdrive Apr 07 16:35:18 denix, yay on fixing that curl issue. I see it locally too on various builds Apr 07 16:36:28 vsr you could even test XDirectfb Apr 07 16:38:44 denix: oh Apr 07 16:38:53 denix: your issue is the same issue I have too Apr 07 16:38:59 denix: http://article.gmane.org/gmane.comp.handhelds.openembedded/22510 Apr 07 16:39:06 denix: take a look and confirm it plz Apr 07 16:39:21 denix: do you have a patch that I could test/ Apr 07 16:39:23 denix: ? Apr 07 16:40:04 otavio, I don't know. denix and I see a problem making a lock file Apr 07 16:40:49 otavio: maybe similar, but not exactly, as we see lockfile() failing Apr 07 16:40:57 Tartarus: I believe it is the same issue since it looks to "work" on the wrong package Apr 07 16:41:55 but the thing is it also fails in do_package_write_ipk()... Apr 07 16:42:44 otavio: ok, let me finish the patch and we'll see if it fixes your issue too Apr 07 16:43:51 I found the root-cause for the issue, but haven't looked yet which commit causes it Apr 07 16:44:27 denix, can you please please post an interm patch to the list? Apr 07 16:45:14 will do Apr 07 16:47:58 Hi, I'm tweaking the recipe for gsoap 2.7.13 and I got so far that the ipk packages of this version contain exactly the same as the ipkg packages of version 2.7.7 Apr 07 16:48:21 however; i am now getting these messages: the following files were installed but not shipped in any package: Apr 07 16:48:56 how can I prevent files from being installed? Apr 07 16:51:31 03Koen Kooi  07org.openembedded.dev * r2fe33f3c7f 10openembedded.git/ (2 files in 2 dirs): linux-oma 2.6.29: reenable BT for beagleboard Apr 07 16:53:12 * otavio curious to see why it fails Apr 07 17:32:49 will selecting xserver-kdrive-fbdev as my xserver allow me to use my keyboard and mouse? or touch screen is the only input Apr 07 17:35:22 denix: please ping me when you mail ml; so I can test it here Apr 07 17:35:35 denix: I can do a full build for test propouse Apr 07 17:36:13 otavio: absolutely! I just need some time to finish it... Apr 07 17:36:39 denix: right; :-) take your time Apr 07 17:44:42 hrw|gone: heh, i worked for digi, a serial communications company.. they make boards with external concentrators that let you hook up 300+ serial ports from a single pci card :) Apr 07 17:47:36 kergoth: who haven't you worked for? :) Apr 07 17:50:19 it was basically a warehouse, then office max & office depot retail, then digi in tech support, then TI, then TI again, then astrocom, then CSC/NCS/Geco Inc, then MontaVista. not that many really.. it was over a relatively short period of time though Apr 07 17:52:12 Some of us get around a lot :-) Apr 07 17:52:50 I'm starting to get sick of it actually.. it seems like embedded Linux folk are in demand, but scattered.. seems like i almost always end up relocating when i switch jobs, pain in the ass :) Apr 07 18:40:06 re Apr 07 18:40:21 hey hrw Apr 07 18:40:26 kergoth: you still have a car which you bought during TI work or not? Apr 07 18:40:40 yep. though I'm trying to get rid of it, too expensive Apr 07 18:40:41 hrw: buy console server or old cisco 2511/2509 :) Apr 07 18:41:03 hrw: it's about your latest post :) Apr 07 18:41:38 I know Apr 07 18:45:43 kergoth: what car is it? Apr 07 18:45:57 '04 Chrysler Crossfire Apr 07 18:46:12 in addition to being more than i want to spend, I'm starting to want something with a back seat. Apr 07 18:46:13 heh Apr 07 18:47:55 ah, and how's the rear view? Apr 07 18:48:58 Jay7: cisco 2511 on ebay costs ~250usd.. for that cash I can buy buildmachine with 2 8xcom cards... Apr 07 18:49:13 hrw: well you are right.. Apr 07 18:49:33 Jay7: 6xcom is ~20€. 8xcom is ~30€ Apr 07 18:51:05 but 8xcom gives me octopus to which I connect serial cables... 6xcom occupies 3 slots with 2com each Apr 07 18:51:09 bbl Apr 07 19:02:03 re Apr 07 19:03:58 re Apr 07 19:04:14 re hrw Apr 07 19:04:18 daughter is sleeping finally Apr 07 19:04:22 hehe Apr 07 19:04:35 tobi is still fighting Apr 07 19:10:09 #oe.parents (c) ant :) Apr 07 19:10:33 my daughter is sleeping too already :) Apr 07 19:12:20 OE parser + AUTOREV kills Apr 07 19:12:44 hi, when compiling latest OE http://pastebin.com/m1e815dfb Apr 07 19:13:05 I used gcc-4.3.3, binutils 2.19.1 and glibc 2.7 Apr 07 19:13:15 target machine is palmtt5 Apr 07 19:13:21 * hrw unpacks ngw100 to check how it works Apr 07 19:13:35 florian, hi Apr 07 19:15:04 hi marex Apr 07 19:15:14 hrw: you need to feed it... :) Apr 07 19:15:48 anyone interested in checking the above ? Apr 07 19:16:52 hm ... if I compile that file by hard (with different compiler), it works Apr 07 19:17:14 could it be gcc-4.3.3 bug ? Apr 07 19:17:20 urks Apr 07 19:17:54 Blah Apr 07 19:18:03 socat 1.3.2.1-r1 doesn't build on x86 Apr 07 19:18:08 yup ... looks like the gcc-4.3.3 issue Apr 07 19:18:10 vala-0.6.0 isn't in checksums.ini Apr 07 19:22:21 tartarus so mickeyl forget maybee Apr 07 19:22:25 ~seen phb Apr 07 19:22:28 heh.. ngw100 does not like to appear ;( Apr 07 19:22:29 phb was last seen on IRC in channel #uclibc, 819d 20h 43m 33s ago, saying: 'CU; thanks, all !'. Apr 07 19:23:39 woglinde, perhaps, yes :) Apr 07 19:24:13 * Tartarus wonders what set of images should in fact build for MACHINE=qemux86 DISTRO=minimal in .dev Apr 07 19:24:24 Tartarus: any Apr 07 19:25:31 owch, then Apr 07 19:26:02 yesterday, out of minimal-image console-image fso-{console,illume}-image gpe-image x11-image only minimal built Apr 07 19:26:11 rest had the lock file thing happen, or something else blow up Apr 07 19:26:36 pulled an hour ago and chugging along, with the problems i've said thus far Apr 07 19:30:16 someone remember does ngw100 require cross nullmodem or straight cable? Apr 07 19:30:45 hm Apr 07 19:30:48 we have one Apr 07 19:30:54 but I dont know it Apr 07 19:31:01 maybe usb/serielle Apr 07 19:31:43 it has female rs232 port so should rather use straight. but plugged directly to ftdi usb<>serial do not work Apr 07 19:33:00 ok. now it works Apr 07 19:33:05 *g* Apr 07 19:33:08 false baurate? Apr 07 19:33:58 rahter not connecting Apr 07 19:34:05 will show you photo Apr 07 19:37:31 florian about the gsoc. It would really be great to get this going Apr 07 19:38:08 wow Apr 07 19:38:21 we are providing gsoc tasks? Apr 07 19:39:20 http://marcin.juszkiewicz.com.pl/tmp/ngw100.jpg Apr 07 19:39:53 woglinde: check that Apr 07 19:40:14 hrw lol Apr 07 19:40:27 keesj: yes indeed... and the best thing is that it is some "train" people might want to jump on Apr 07 19:40:37 yay for bitbake -k letting me get some stuff built at least Apr 07 19:41:02 woglinde: the worst thing is that I got board configured to boot tftp kernel... Apr 07 19:41:36 florian: it has been on my wish list for so long Apr 07 19:41:58 Jay7: not directly, but mameo has one oe task Apr 07 19:42:36 hrw: *g* Apr 07 19:42:38 did you mean maemo? or mameo? :) Apr 07 19:42:56 ooops maemo Apr 07 19:43:00 florian: but I never really understood why mamona and poky didn't really pick up Apr 07 19:43:10 hm.. Apr 07 19:43:12 uf. got to uboot console Apr 07 19:43:21 * Jay7 is interested in maemo integration Apr 07 19:43:41 jay7 check chinookcompat Apr 07 19:44:26 woglinde: seems we have a little of maemo apps Apr 07 19:44:34 but even many open-source minded people don't know what maemo is so talking to them about maemo on OE is like explaining SQL to my dog *my GF does have a SQL interface* Apr 07 19:45:22 hehe.. well Apr 07 19:45:29 keesj: you are right Apr 07 19:45:51 * Jay7 will just have hildon on zauruses Apr 07 19:46:32 how does bitbake/oe decide which kernel version to use in the absence of a PREFERRED_VERSION_linux or a DEFAULT_PREFERENCE_${machine} ? Apr 07 19:46:48 may be fixed a bit to support 640x480 Apr 07 19:47:06 I'm currently using 2.6.29 but there is an ARM patch in 2.6.29.1 that I need, and I don't care to create an entire new directory in recipes/linux Apr 07 19:47:37 (I suppose I can just include the patch in question) Apr 07 19:47:37 Well I guess part of the strength of maemo is the finger friendlyness and that does not translate good everywhere Apr 07 19:48:17 my point that strength of maemo is good integrated interface Apr 07 19:48:21 'solid' Apr 07 19:48:41 we can't have this with just of bunch of misc. x11 apps Apr 07 19:49:05 (that we have in most of x11 distros) Apr 07 19:50:16 bye all Apr 07 19:50:21 bye hrw Apr 07 19:53:26 Jay7: indeed and I like the kind of app that are created for maemo, Apr 07 19:53:53 well.. I will bring this to zauruses at least Apr 07 20:11:19 I'll give it a go with gcc 4.3.2 to see if there's some difference Apr 07 20:23:04 kergoth: just last week we talked about readynas/drobo... Apr 07 20:24:39 kergoth: and today drobo released an 8-drive monster - it's nice with eth/iSCSI and their expandable "RAID" alternative, but $1300 w/o drives is just outrageous! :) Apr 07 20:25:34 btw, it is apparently still runs vxworks... Apr 07 20:27:23 oom killer has to die! Apr 07 20:28:12 hrw|gone: amen! Apr 07 20:28:24 4GB ram and oom killer randomly kills processes so it took down x11 session with 2 builds Apr 07 20:29:03 whoa, I thought you were talking about embedded systems... :) Apr 07 20:30:35 in moments like this I am considering adding 4GB just for situations like this Apr 07 20:33:43 what about swap? Apr 07 20:34:29 normally I do not use any swap as system do not need it and I do not like to wait few seconds on desktop switch Apr 07 20:38:17 at least it won't bomb like that... Apr 07 20:43:23 What's the easiest way to test if bb.data.getVar('SDK_OS'), d, 1) contains either mingw32 or mingw64? Apr 07 20:43:50 in [bla, foo] Apr 07 20:43:56 Ah, in, thanks Apr 07 20:44:13 look in the conf for angstroem and oabi Apr 07 20:45:47 yes [] makes a list Apr 07 20:50:07 is there a way to tell the SVC fetchers not to try to use a mirror, but checkout from SVC directly instead? Apr 07 20:50:17 OK, this works but seems backwards Apr 07 20:50:22 SYSROOT = "${@['${SDK_REALPATH}/${TARGET_SYS}', '${SDK_REALPATH_MINGW}${SDK_REALPATH}/${TARGET_SYS}'][bb.data.getVar('SDK_OS', d, 1) in ['mingw32', 'mingw64']]}" Apr 07 20:50:47 For SDK_OS==mingw32, that does give me SYSROOT=${SDK_REALPATH_MINGW}${SDK_REALPATH}/${TARGET_SYS} Apr 07 20:50:59 But I thought it would be the other way around, ie [ Apr 07 20:51:13 [true, false][test] Apr 07 20:51:19 * Jay7 consider installing ccache.. Apr 07 20:52:57 Or should I do getVar('..', d, 0) for ['mingw path', '!mingw path Apr 07 20:52:59 '] ? Apr 07 20:58:49 woglinde, hi do you know what ./boost/detail/atomic_count_sync.hpp:54: error: impossible constraint in 'asm' means with that code : http://rafb.net/p/N17tr051.html ? Apr 07 20:59:38 ah maybe I've forgetten the + Apr 07 21:03:18 hm sorry Apr 07 21:03:26 mpoullet|home: you mean like checking out from svn directly? Apr 07 21:04:36 mpoullet|home: If that's what you're looking for, this is how it's done in (for example) tslib Apr 07 21:04:37 SRC_URI = "svn://svn.berlios.de/tslib/trunk;module=tslib" Apr 07 21:05:44 re ant Apr 07 21:06:18 virk: but in the SVN fetcher in bitbake, it seems that it always called try_mirror first? Apr 07 21:07:29 virk: I'd like to fetch from SVN servers on an intranet and it always tries to fetch a tarball first, then fails, then svn fetch Apr 07 21:09:23 hmm, seems I did not fully understand, and as I am quite new here, I cannot help any more, sorry Apr 07 21:09:50 virk: no problem, thanks :)! Apr 07 21:12:54 Hey, sorry to be a bother. We want to ship a java toolchain in our distro, is openjdk-langtools-native the right recipe to include? Apr 07 21:12:54 is there a specific channel for bitbake? Apr 07 21:13:04 nope Apr 07 21:13:07 EHLO woglinde Apr 07 21:15:22 woglinde: ok, will try its m-l then, thanks Apr 07 21:18:09 mpoullet|home: these bb files accept python, isn't it? you could perhaps somehow define your own do_fetch in which you can test whether svn is accesable or not? Apr 07 21:18:23 03Stanislav Brabec  07org.openembedded.dev * r9497284236 10openembedded.git/ (4 files in 3 dirs): autoconf213-native: Added old autoconf for fennec recipe. Apr 07 21:18:25 03Stanislav Brabec  07org.openembedded.dev * r67949fd395 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Apr 07 21:28:42 hi does anyone knows well inline assembly? Apr 07 21:38:25 03Andrea Adami  07org.openembedded.dev * r65c025a2e2 10openembedded.git/conf/machine/include/zaurus-2.6.inc: Apr 07 21:38:25 zaurus-2.6.inc: add ZAURUS_KERNEL_IMAGETYPE Apr 07 21:38:25 - switch kernel types without re-editing the machine.conf files Apr 07 21:38:25 - using KERNEL_IMAGETYPE ?= here obliged to edit machine.conf Apr 07 21:39:07 just a stupid question; what does this do: oe_libinstall -C gsoap $lib ${D}${libdir} Apr 07 21:44:59 ant_home: hi, do you know what's the status of http://patchwork.openembedded.org/patch/181/ ? Apr 07 21:52:21 mpoullet|home: can you unset the SRC_TARBALL_STASH? Apr 07 21:53:39 mpoullet|home: mickey|bbiab is investigating Apr 07 21:53:52 mpoullet|home: also, you can set SRC_TARBALL_STASH_foo per package... Apr 07 21:54:03 but I haven't played with it myself... :) Apr 07 21:56:13 Tartarus, otavio, kergoth: http://news.gmane.org/gmane.comp.handhelds.openembedded Apr 07 21:56:57 Hmm Apr 07 21:58:10 I'm pretty sure that's mine, but, curl-sdk doesn't make packages? that seems wrong too :) Apr 07 21:58:57 lemme guess, opkg builds curl support in Apr 07 21:58:57 denix: good idea, I'll try the SRC_TARBALL_STASH_foo method tomorrow! Apr 07 21:59:11 kicking a build of my image targets now Apr 07 21:59:16 ant_home: that's good news Apr 07 22:00:39 denix: if I don't set SRC_TARBALL_STASH it uses Angstrom's default one, which won't work behind a firewall... Apr 07 22:01:15 mpoullet|home: if your recipes don't use package_stagefile_shell you're almost done Apr 07 22:01:21 have you tried setting it to nothing, i.e. empty? Apr 07 22:01:39 ant_home: I'm tired of coworkers asking why can't we put that stuff out of TMPDIR? Apr 07 22:01:39 mpoullet|home: it seems mickeyl thinks about pulling /cross out of /tmp Apr 07 22:01:56 mpoullet|home: that's indeed nasty issue Apr 07 22:02:07 time-consuming Apr 07 22:02:46 collections è packaged staging are two 'must work' for professionals imho Apr 07 22:02:58 (I'm hobbyst :-)) Apr 07 22:03:00 denix: I suppose it then always tries to fetch directly from the SVC repos Apr 07 22:03:35 denix: and then it works only for the SVC repos I'm mirroring not the others one Apr 07 22:03:55 CVS? Apr 07 22:04:36 ant_home: what's mickey's issue with CROSS_DIR? I'm confused Apr 07 22:04:45 SVC: Source Version Control = git/svn/cvs Apr 07 22:04:53 well, same happened to me Apr 07 22:05:03 no crosscompilers found... Apr 07 22:05:20 or rather...cannot create executable Apr 07 22:05:33 two spices of same thing Apr 07 22:05:35 right, but bb only has SRC_TARBALL_STASH and CVS_TARBALL_STASH :) Apr 07 22:05:52 I confirm that packaged-staging must work if I want to convince pros to use oe Apr 07 22:06:23 ant_home: so, why aren't they installed from pstage? Apr 07 22:06:43 Hm. I tried adding openjdk-langtools-native to my image's IMAGE_INSTALL and DEPENDS, but it doesn't work, saying that there's no such package openjdk-langtools-native. Any idea? Apr 07 22:06:51 denix: I you look at the try_mirror method, SRC_TARBALL_STASH is treated as being equivalent to CVS_TARBALL_STASH Apr 07 22:07:18 denix: there are two behaviours if 1) you leave whole /deploy or 2) just /deploy/pstage Apr 07 22:07:50 mpoullet|home: pretty much Apr 07 22:09:05 ant_home: hmm. I need to try it both ways here... Apr 07 22:10:24 denix: bonus is the package_stagefile_Shell thing... Apr 07 22:11:00 huh, that one been sitting in bugzilla for over 6 months now... Apr 07 22:12:26 * mpoullet|home 'd like the package_stagefile_shell bug to be fixed too Apr 07 22:13:09 * mpoullet|home must learn more python to fix things instead of complaining in #oe Apr 07 22:15:45 denix: ant_home: thanks for your help, gb! Apr 07 22:16:20 * Jay7 catch strange error in linux-kexecboot with ccache enabled Apr 07 22:16:43 | CC kernel/bounds.s Apr 07 22:16:45 | make[1]: *** [kernel/bounds.s] Error 1 Apr 07 22:17:14 seems I can't get any profit from ccache.. Apr 07 22:17:31 * ant_home sticks with Gentoo Apr 07 22:17:52 ant_home: what version of ccache you have/ Apr 07 22:18:15 ccache -V Apr 07 22:18:53 Jay7: 2.4 Apr 07 22:19:05 the same.. Apr 07 22:19:24 * kergoth thinks its past time that we stop the automatic use of ccache in oe.. explicit is better than implicit Apr 07 22:19:38 Tartarus: I guess curl-sdk is only envisioned to be consumed by opkg-sdk, hence it only stages its files, but doesn't package anything... Apr 07 22:20:42 wouldn't work if someone actually wants a real curl-sdk - would anyone? :) Apr 07 22:21:07 I just should buy faster buildbox :) Apr 07 22:21:11 denix, yeah.. Apr 07 22:22:39 Tartarus: have you seen the patch? does it make sense/work for you? Apr 07 22:22:51 tcp-wrappers issue wasn't the same Apr 07 22:23:07 PR got bumped and somehow it didn't fully clean or build everything, or something Apr 07 22:24:31 Tartarus: yeah, I always try killing TMPDIR first, if something is broken :) Apr 07 22:24:55 dang ... gcc 4.3.2 is no good ... segfault as well Apr 07 22:25:57 03Tom Rini  07org.openembedded.dev * r9f0f45d82f 10openembedded.git/conf/checksums.ini: checksums.ini: Add vala-0.6.0 Apr 07 22:26:01 denix yeah, i should have too I guess Apr 07 23:06:55 any objection to adding ncurses 5.7 to the tree? Apr 07 23:12:36 kergoth: do you see THE obious error here ? Apr 07 23:12:38 http://rafb.net/p/zV99Rk89.html Apr 07 23:12:56 nothing immediately obvious. what's the behavior? Apr 07 23:13:06 he..variables are empty Apr 07 23:13:41 so mkimage bombs out Apr 07 23:13:52 is it the sed commands? ar ethe vars nonempty if you just do the objdumps? Apr 07 23:14:03 from console seems ok Apr 07 23:14:24 in kernel.bbclass seems not working as expected Apr 07 23:14:25 I'd stuff an "oefatal $UBOOT_LOADADDRESS" right before the package_stagefile_shell Apr 07 23:15:27 kergoth: doing Apr 07 23:18:59 hm..-c deploy would be faster... Apr 07 23:20:09 heh, indeed Apr 07 23:20:14 maybe with -b too :) Apr 07 23:20:43 kergoth: | FATAL: 0x00000000 0x001498cc Apr 07 23:21:08 and the run-log: Apr 07 23:21:29 uboot-mkimage -A arm -O linux -T kernel -C none -a ${UBOOT_LOADADDRESS} -e ${UBOOT_ENTRYPOINT} -n "Angstrom/2.6.26/c7x0" -d linux.bin /oe/build/tmp/angstrom Apr 07 23:21:29 Apr 07 23:21:38 unexpanded... Apr 07 23:22:18 03Tom Rini  07org.openembedded.dev * r69c119a2b9 10openembedded.git/classes/canadian-sdk.bbclass: canadian-sdk: Add SDK_REALPATH_MINGW variable, so that we can prepend a drive letter for sysroot. Apr 07 23:22:28 03Tom Rini  07org.openembedded.dev * rec31872ec8 10openembedded.git/recipes/binutils/binutils-canadian-sdk_2.18.bb: binutils-canadian-sdk: Use add SYSROOT, use SDK_REALPATH_MINGW, bump PR Apr 07 23:23:36 03Tom Rini  07org.openembedded.dev * rae48ce4854 10openembedded.git/recipes/gcc/ (gcc-canadian-sdk_4.2.4.bb gcc-configure-canadian-sdk.inc): gcc-canadian-sdk: Fix C++ relocation, add SYSROOT, use SDK_REALPATH_MINGW, bump PR. Apr 07 23:34:27 I'm creating a recipe for gsoap v 2.7.13, but I geet ketting this message :NOTE: the following files were installed but not shipped in any package: Apr 07 23:35:05 is there any way to get rid of this message (thus not installing these files), without rewriting do_install? Apr 07 23:47:15 hmm, ml is very sloooow again... Apr 07 23:49:01 'nite all Apr 07 23:49:56 'night Apr 07 23:53:51 finally... Apr 07 23:54:18 Tartarus: I resent the patch with common/target .inc files Apr 08 00:01:38 looks good, thanks & ack'd Apr 08 00:02:03 and, blah, vala-bootstrap blew up on qemux96 Apr 08 00:02:17 * Tartarus waits for his loop and bitbake -k to finish up Apr 08 00:04:45 virk, what files? Apr 08 00:05:40 a bunch of c, c++ and header files, which aren't needed Apr 08 00:06:05 but they get installed Apr 08 00:06:48 ignore the message, or add them to a package Apr 08 00:06:48 hmm, sounds like a weird package... Apr 08 00:07:03 -examples -de v that sort of thing? Apr 08 00:07:04 unless the these files need to end up in the dbg package, they arent needed anywhere, the dev package already contains the libs, packageconfig files and header Apr 08 00:07:10 * Jay7 -> sleep Apr 08 00:07:15 Jay7: gn Apr 08 00:07:24 'nite pplz :) Apr 08 00:07:24 ${PN}-garbage :) Apr 08 00:08:31 folders like 'import' , 'plugin', 'extras' Apr 08 00:10:09 is ${PN}-garbage an accepted solution? :) Apr 08 00:11:45 Well, if they aren't needed, kill 'em in do_install Apr 08 00:11:56 or if you're using autotools do_install, add do_install_append() { .. } Apr 08 00:12:01 and note why this stuff is garbage Apr 08 00:12:35 But be really sure they shouldn't end up in -dev or -dbg via FILES_${PN}-dev += "...." Apr 08 00:14:30 I've checked the ipk's of the previous version, and except for this surplus on files, exactly the same is packaged Apr 08 00:17:06 In fact, all these files end up in .../image/usr/share Apr 08 00:19:00 OK Apr 08 00:19:22 I'm just saying, adding a -garbage package isn't the right fix. Either they need to be removed or packages Apr 08 00:19:26 s/s$/d/ Apr 08 00:20:08 hmm, gsoap install the soapcpp compiler, which needs headers located in /usr/share/gsoap/import Apr 08 00:22:58 the more I look at those files, the more I get the feeling they belong in the ${PN} package. after all gsoap is a devel package. Apr 08 00:23:45 btw, -garbage was obviously a joke... :) Apr 08 00:24:18 hmm, is there a libgsoap? Apr 08 00:24:33 not that I am aware of Apr 08 00:25:15 but I am not that familiar with gsoap Apr 08 00:25:21 so, why is it needed? who depends on it? what does it provide? Apr 08 00:25:59 * denix needs to actually look in the recipe... Apr 08 00:26:29 ehm, yes, I understood incorectly I guess Apr 08 00:27:28 gsoap actually provides some binaries like soapcpp2 and wsdl2h, which can be used to do some soap related stuff Apr 08 00:28:00 hmm, it's mostly bindings Apr 08 00:28:17 but, it does also provide stdsoap2.h and some libs with which you can create custom projects Apr 08 00:28:30 that's in fact all I know about it Apr 08 00:29:02 normally things like headers go into -dev packages, as they are only required if one wants to do development on the target Apr 08 00:29:40 yes I know, but these headers in /usr/share are needed by soapcpp2 Apr 08 00:29:57 which is some kind of compiler itself Apr 08 00:30:22 I just figured that out with the readme's in those folders Apr 08 00:30:33 if those are needed for the binary to work, then you need to package them Apr 08 00:30:42 ah, ok :) Apr 08 00:31:18 the last time I looked at SOAP it was 9-10 years ago and I only used it in perl... Apr 08 00:33:12 I'm not so into soap either, but I'm obliged to wotk with gsoap for now Apr 08 00:34:03 In fact I only needed the staging package to be able to compile my project, but I cannot use 2.7.7, the header has changed since version 2.7.8 I think Apr 08 00:35:01 So I thought I could submit a patch for teh newest version 2.7.13, and I've been strugling with that the whole day :) Apr 08 00:35:24 every start is hard :) Apr 08 00:42:43 * kergoth ponders Apr 08 00:44:48 denix: can I only package files which where installed with do_install? Apr 08 00:45:27 yep Apr 08 00:46:24 either change do_install() or add do_install_append() Apr 08 00:48:17 so I will package these files in usr/share into ${PN}. is there anything like ${libdir} for /usr/share ? Apr 08 00:49:42 datadir? Apr 08 00:49:45 check conf/bitbake.conf Apr 08 00:49:48 datadir Apr 08 00:50:19 yeah, I'm too slow :) Apr 08 00:52:08 thanks Apr 08 00:53:06 * virk should really go to bed! Apr 08 00:59:06 If ${PN} contains those files from /usr/share, should the dbg package contain them as well?µ Apr 08 01:10:59 virk: nope Apr 08 01:11:15 -dbg and -dev depend on main package Apr 08 01:12:01 ok Apr 08 01:12:25 almost done then, only the doc package left Apr 08 01:16:40 actually, -dbg recommends, not depends on the main package. but -dbg only provides "debug" versions of the binaries... Apr 08 01:20:41 but not the headers that are needed for this soap compiler Apr 08 01:22:21 right. at least that's how it is in other packages Apr 08 01:23:13 I mean the headers found in /usr/share Apr 08 01:30:23 denix: can install -m work recursively? Apr 08 01:32:20 I don't think install supports recursive option at all... Apr 08 01:32:27 Crofton: ping Apr 08 01:50:25 denix: another probably stupid question: gsoap-native requires the bb recipe of gsoap. how can I let gsoap-native override do_install_append from gsoap? Apr 08 01:52:07 or should I add an .inc file and seperate them from eachother? Apr 08 01:53:44 yeah, I was about to suggest an .inc file... Apr 08 01:54:05 I'm getting really slow... :) Apr 08 01:54:18 could be the time of day? Apr 08 01:55:39 just not paying attention to irc too much Apr 08 02:05:47 denix: which commit? Apr 08 02:06:00 ok, if only there was a recursive install. e.g. I want to install every file and filder in given dir Apr 08 02:06:33 otavio: start here - http://thread.gmane.org/gmane.comp.handhelds.openembedded/22376 Apr 08 02:08:01 denix: ok; got it Apr 08 02:08:04 denix: let me do it Apr 08 02:09:02 otavio: try reverting the commit I mentioned, because the fix is actually for curl only - you are having similar issue with something else Apr 08 02:09:16 denix: reverted; doing the build now Apr 08 02:09:25 thanks Apr 08 02:10:29 Is a -doc package obliged? Apr 08 02:11:48 If there's doc type things installed by the install target Apr 08 02:11:51 man pages and such Apr 08 02:12:38 hmm, in fact there is'nt any Apr 08 02:13:01 I'm so tired I was trying to pack the doxygen docs :p Apr 08 02:18:50 heh Apr 08 02:31:58 voila, hopefully this time the gsoap recipe is correct. patch sent Apr 08 02:32:03 * virk is off to sleep Apr 08 02:41:34 virk|sleep: and you forgot to update the checksums.ini... :) Apr 08 02:41:45 virk|sleep: good night anyway :) Apr 08 02:42:09 what was that again, don't commit if you're tired? Apr 08 02:42:12 :) Apr 08 02:44:03 right :) Apr 08 02:48:47 denix: is there any rule on where to add th checksum, top, bottom, grouped with similars ? Apr 08 02:49:12 sorted Apr 08 02:50:25 so just below the other gsoap should do Apr 08 02:50:25 there is contrib/source-checker/oe-checksums-sorter.py Apr 08 02:51:29 I wonder if it's anywhere on the wiki - it is really a Frequently Asked Question :) Apr 08 02:52:24 damn, that thing just spawns a whol new one Apr 08 02:52:29 basically, run your recipe, let it fetch and fail on the checksum - it will generate a missing portion for you in TMPDIR, then append it to the main checksums.ini and run the script above to sort it Apr 08 02:54:35 So I have been here before looking for help with a very odd issue. When trying to build anything, it gets to the do_compile phase of shasum-native and fails. There is nothing in the log.do_compile.xxx file, but if I run the run.do_compile.xxx file, the do_compile runs just fine and generates the file. Is this file exactly the same thing that the script runs or is it different in some way? Apr 08 02:58:37 denix: it's not in the wiki Apr 08 02:59:04 it probably should :) Apr 08 02:59:07 mekius: it is the same thing that bitbake runs... that's very strange Apr 08 02:59:31 kergoth: yeah, I agree it's strange heh Apr 08 02:59:48 ok, patch is finally done **** ENDING LOGGING AT Wed Apr 08 02:59:57 2009