**** BEGIN LOGGING AT Mon Jul 18 02:59:57 2011 Jul 18 10:13:56 hi all, in my local.conf I add DISTRO_EXTRA_RDEPENDS += "boa grub" Jul 18 10:13:57 for DISTRO = "minimal" but no grub, no boa compiled and add to the system, any idea ? thx in advance Jul 18 10:35:22 morning all Jul 18 10:39:29 hi bluelightning Jul 18 11:34:08 pb_: surprisingly repackaging klibc with PACKAGES_DNAMIC + do_split_packages() worked at first try Jul 18 11:34:35 I have just to filter out unwanted -dev and -dbg empty packages Jul 18 11:35:06 how...well, that's another question Jul 18 11:48:05 why does gcc-configure-cross.inc in oe-core defines do_compile_prepend() but not do_configure_prepend()? Old OE modified configure() and new variant makes build fail when '-mthumb' is required for target arch Jul 18 11:51:54 ok... oe-core code is from 2005 and oe fixes from 2009 have not been merged :( Jul 18 12:15:02 ensc|w: ? Jul 18 12:15:16 http://cgit.openembedded.org/cgit.cgi/openembedded-core/log/meta/recipes-devtools/gcc/gcc-configure-cross.inc Jul 18 12:16:50 ensc|w: That isn't entirely true as the gcc bootstrap process is much improved in oe-core for example Jul 18 12:18:07 RP__: I have supposedly built in thumb mode for armv5te Jul 18 12:18:22 can't check right now here in office Jul 18 12:21:09 my problems might be caused by gcc 4.6 vs. 4.5... gcc 4.5 checked for 'gcc -E' while 4.6 uses the cross compile '.../xgcc -E' (which misses -mthumb from TARGET_CC_ARCH) Jul 18 12:26:09 http://pastebin.com/W9Rr8ngm lines 7 + 20 Jul 18 12:29:33 caused by zecke-xgcc-cpp.patch Jul 18 12:30:15 patch should probably set $(XGCC_FLAGS_FOR_TARGET) for CPP too Jul 18 12:51:53 ant_work: fwiw, 'git blame meta/recipes-devtools/gcc/gcc-configure-cross.inc' gives more reliable information about history of this file; the do_compile_prepend is from 4b46c1f6e891b1ddd5968536440b888661fade3e (Wed Aug 31 10:45:47 2005) Jul 18 12:52:44 ensc|w: coming from oe-dev and testing a beta-layer I thought I' was still compiling in thumb mode...youmake me doubt about it ;) Jul 18 12:55:22 ant_work: you might be compiling in thumb mode. The problem is, that configure calls gcc -E without -mthumb which breaks build for thumb-only archs Jul 18 12:58:06 mmmh... with oe-core I have now a bash shell by default, where oe installed only the busybox sh applet? Jul 18 13:07:27 oe-core does not strip kernel modules? Jul 18 13:12:16 ant_work: oh, very good Jul 18 13:13:28 ensc|w: yeah, there probably is a bunch of stuff that still has not been merged to oe-core. as you say, it was forked from oe.dev about six years ago and a lot of stuff has changed in the meantime. Jul 18 13:14:18 as for bash, that's probably because you are installing somor other package that has bash in RDEPENDS. Jul 18 13:14:33 "opkg whatdepends bash" might reveal the truth Jul 18 13:17:55 opkg does not work anymore with nfs rootfs because it places pkg info into the /var tmpfs :( Jul 18 13:18:10 ah, that is a bit sad Jul 18 13:39:20 03Martin Jansa  07master * rb10f3d7a10 10openembedded.git/recipes/freesmartphone/libgisi_git.bb: Jul 18 13:39:20 libgisi: bump SRCREV Jul 18 13:39:20 Signed-off-by: Martin Jansa Jul 18 14:05:25 can someone confirm that www.eglibc.org is down? Jul 18 14:05:38 * JaMa cannot update eglibc checkout :/ Jul 18 14:06:58 JaMa: seems to be broken for me too Jul 18 14:09:07 bluelightning: thanks Jul 18 14:18:38 morning Jul 18 14:21:34 g'day kergoth Jul 18 15:11:39 heh, crazy linaro and their generic mailing list names Jul 18 15:45:11 nice... with minor adaptations, oe-core allows me to develop for cortex-m3 again. Jul 18 15:46:40 nice Jul 18 15:47:47 and the gcc 4.6 compiled binary seems to run :) Jul 18 17:04:49 hi all (back @ work) :) Jul 18 17:11:01 I'm still fighting with a ttySAC0 not being created :( Jul 18 17:19:06 how to debug why ttySAC0 isn't created ? kernel debug shows the device but there is not device file node Jul 18 17:24:24 cool ... now ERROR: Nothing PROVIDES 'virtual/INVALID-angstrom-linux-gcc' :( Jul 18 17:38:39 alx_: device node is created by udev/mdev or by the devtmpfs Jul 18 17:46:25 ensc|w, mmm how to discover what is it using ? Jul 18 17:46:35 udev has the rule for ttySAC0 Jul 18 17:56:22 alx_: will node be created by 'echo add > /sys/class/tty/ttySAC0/uevent'? Jul 18 17:57:44 ensc|w, how to know ? I can't login Jul 18 17:57:44 you might miss a cold plug mechanism (udevtrigger) Jul 18 17:58:12 I don't know. Jul 18 17:58:16 alx_: try 'init=/bin/sh' kernel param Jul 18 17:58:29 should use initial console Jul 18 17:58:59 * Tartarus kicks btrfs-tools, double checks where oe-core does uuid stuff Jul 18 17:59:12 trying Jul 18 17:59:22 ok Jul 18 17:59:36 wondering why... /dev/ttySAC0 exists Jul 18 17:59:42 lol no Jul 18 17:59:49 ok Jul 18 18:00:09 it exists: /dev/ttySAC0 exists : I see it using init=/bin/sh Jul 18 18:03:10 ensc|w, what if init=/bin/sh let me ls /dev/ttySAC0 ? Jul 18 18:05:05 alx_: your initsystem can mount tmpfs at /dev Jul 18 18:05:18 how do you know that this node does not exist? Jul 18 18:10:54 ensc|w, by init respawning getty Jul 18 18:11:25 ensc|w, I discovered the problem: inittab search for ttySAC0; mdev creates s3c2410_serial0 Jul 18 18:11:47 and deletes the /dev/ttySAC0 created by device_table Jul 18 18:12:01 ensc|w, check http://code.google.com/p/cfgrepo/source/browse/linux_config/embedded_gentoo/armv4tl_etc/portage/my_fix_overlay_2440/etc/mdev.conf_backup?r=9f8b75c180b05cc6213a55f46b0cf45b944eb1f1 Jul 18 18:12:11 sorry for long link :( Jul 18 18:25:57 yay, ../oe-core/scripts/create-pull-request just works Jul 18 18:25:57 almost like it's planned or something ;) Jul 18 18:26:17 Tartarus: hi Jul 18 18:26:26 hey khem Jul 18 18:26:38 Tartarus: btw. I have to patch oe-core to build standalone with uclibc Jul 18 18:26:46 but in the end x86-64 booted Jul 18 18:26:52 fine like angstrom Jul 18 18:27:05 I see you posted uclibc stuff Jul 18 18:27:06 haven't gotten back to that yet, but am making progress on deep path builds Jul 18 18:27:08 so there must be some foo in your env that sent it crashing Jul 18 18:27:36 actually I will post another update of that patch series Jul 18 18:27:46 I am now able to build core-image-sato with uclibc Jul 18 18:28:05 Did you use the test framework Jul 18 18:28:05 or just boot by hand? Jul 18 18:28:19 I use runqemu script Jul 18 18:28:22 k Jul 18 18:28:22 sec Jul 18 18:28:28 https://wiki.yoctoproject.org/wiki/Enabling_Automation_Test_in_Poky Jul 18 18:28:33 also just works for oe-core Jul 18 18:28:36 havent tried that Jul 18 18:29:35 * khem tries runqemu qemuarm core-image-sato Jul 18 18:30:01 it's an apt-get install and a line to local.conf, easy to try ;) Jul 18 18:30:08 for uclibc I have to say TCLIBC=uclibc Jul 18 18:31:33 I will try it eventually just trying to keep variables minimum Jul 18 18:31:41 use what works :) Jul 18 18:31:50 for me Jul 18 18:45:21 yeah, this just kicks off the testsuite Jul 18 18:45:24 which is kinda nice :) Jul 18 18:54:29 is it ok that the javac wrapper script, hides legitimate errors from programs calling it? Jul 18 18:56:27 Sounds like sub-optimal behavior to me, but I just don't know the history there Jul 18 18:59:20 how would hiding errors help ? Jul 18 18:59:27 it only helps lawyers Jul 18 19:00:08 when the script was modified to try multiple times, the error code was no longer returned from the shell script Jul 18 19:00:35 this is a bug in script imp Jul 18 19:00:37 imo Jul 18 19:00:43 I think I submitted a patch last week, but have seen no activity on it. Jul 18 19:02:49 should I just reply to the mail with a ping message? Jul 18 19:03:22 I don't want to be a nuisance on the mailing list. Jul 18 19:06:20 ok that was actually 8 days ago. Jul 18 19:07:09 is it for oe.dev ? Jul 18 19:07:40 uff Jul 18 19:07:44 and if the problem is there in 2011.03 release branch you might get better audience Jul 18 19:08:13 why OE (angstrom-2010.x) don't create a ttyUSB0 if I connect a ftdi device ? :( Jul 18 19:08:20 on pc it works :( Jul 18 19:09:04 the problem is in the 2011.03 release, but I thought it had to go to oe.dev before it could be in the maintenance branch Jul 18 19:10:04 alx_: do u know if your kernel has ftdi drivers ? Jul 18 19:10:23 gchiii: did you mention the problem for both branches ? Jul 18 19:10:31 gchiii: if not then do that Jul 18 19:10:50 ok I am doing this now. Jul 18 19:10:52 thanks Jul 18 19:11:10 khem, good question... I was already going to check :) Jul 18 19:11:33 if a patch is solely for oe.dev it might not get much audience Jul 18 19:12:05 ok Jul 18 19:13:27 since further release of oe will be based on new layering model Jul 18 19:13:55 I did a reply and added the 2011.03-maintenance to the subject Jul 18 19:13:56 its better to port the patches to layers Jul 18 19:14:56 yes, I want to move to layers, but I just got my customer to agree to migrate to the maintenance branch Jul 18 19:15:22 thats fine I dont expect customers to move to bleeding edge Jul 18 19:15:27 and they don't seem to like changes Jul 18 19:15:29 its just forward looking Jul 18 19:15:39 well they just don't like paying for changes Jul 18 19:15:48 yes stability is what they want and thats why you are there :) Jul 18 19:16:23 unfortunately they have met me and realize I am anything but stable! :) Jul 18 19:16:56 khem, it has, but no autoload modules :) Jul 18 19:38:13 gchiii, your last sentence will become my motto :) Jul 18 19:38:46 :-P Jul 18 20:34:20 fray: ping re rpm and oe-core and DEPENDS Jul 18 20:35:46 sure Jul 18 20:36:07 is git.openembedded.org down? Jul 18 20:37:02 git was hanging for me trying to update bitbake.. I switched to github and it loaded immediately Jul 18 20:37:23 git is hanging for me as well. Jul 18 21:32:22 is anyone else having problems accessing git.openembedded.org right now? Jul 18 21:36:19 yeah, melo looks hosed again Jul 18 21:36:27 presumably the old problem with disks going walkabout Jul 18 21:36:41 we need to neat the Gerbils Jul 18 21:36:47 er beat Jul 18 21:37:12 anything we (the community) can do to help out? Jul 18 21:39:17 probably not a great deal Jul 18 21:39:36 ^_^ Jul 18 21:39:43 ok, then i'll go back to drooling on my keyboard. :-) good luck beating on melo. Jul 18 21:39:48 if you happen to work for one of the billion-dollar companies that use oe then maybe you could try to persuade your bosses to donate a new computer to the cause :- Jul 18 21:39:58 what are the specs needed? Jul 18 21:40:23 good question. melo is actually a vm and I'm not entirely sure what the host hardware is like. Jul 18 21:40:41 write me up a spec list and let's talk. Jul 18 21:41:01 ok, cool Jul 18 21:41:38 can I reach you by email? Jul 18 21:41:59 I think after github morrorring started melo has started to fail more often Jul 18 21:42:07 definitely, preferred in fact. kg4ysn@gmail.com Jul 18 21:42:18 great, thanks Jul 18 21:42:26 khem: yeah, it seems, but I can't find any causal link Jul 18 21:42:27 np, thanks for all you guys do! :-) Jul 18 21:42:52 pb_: I'm almost ready to send the damned klibc for inclusion in meta-oe Jul 18 21:43:02 every time I have managed to catch melo before it goes so far out to lunch that logins become completely impossible, the main symptom seems to be that the disk system has just stopped working Jul 18 21:43:05 ant__: very good Jul 18 21:43:17 I'm a bit shy..still a couple of doubts... Jul 18 21:43:57 pb_: something fires up ntop Jul 18 21:44:10 pb_: and that kills it IIRC Jul 18 21:45:38 that doesn't appear to be the actual cause from what I can see Jul 18 21:46:05 it's true that when it dies you get loads of processes running and a load of about 800 (literally) but i think that is a symptom not a cause Jul 18 21:51:04 fray: incoming Jul 18 21:51:06 may be we can stop mirroring to github and see if that helps and try to see if the mirroring is somehow causing it Jul 18 21:58:49 while you're there masters...http://paste.debian.net/123287/ Jul 18 21:59:41 the first recipes spits out empty -dev and -debug plus the correct static utils (splitted) Jul 18 21:59:58 why -dev and -dbg? Jul 18 22:00:15 They're ALLOW_EMPTY, iirc Jul 18 22:02:57 I'll post a list of ipk in a second Jul 18 22:04:02 http://paste.debian.net/123289/ Jul 18 22:05:40 (in the meanttime I re-nuked klibc-utils-klcc_1.5.23-r2.0_poodle.ipk) Jul 18 22:19:19 Tartarus: strangely the empty packages are emitted only for the static utils. Same inc... Jul 18 22:19:28 must be the packaging... Jul 18 22:30:13 cryptk|offline: there ? Jul 18 22:34:27 pb_: can you CC ka6sox (Tom king) too when you write to kg4ysn Jul 18 22:41:54 anyone else having trouble connecting to git.oe? Jul 18 22:53:08 sgw its down Jul 18 22:53:19 work is in progress to get it back on its feet Jul 18 22:53:25 khem: well that would explain things then! Jul 18 22:54:03 pb_: is github mirror invoking ntop somehow ? Jul 18 22:54:28 sgw: but it might take a while (hours) Jul 18 22:55:15 no worries, I will deal with things locally and on git.yp.org, I am trying to get the next Consolidated request out to RP. Jul 18 23:09:11 sgw: do u have my uclibc patches in there ? Jul 18 23:09:37 khem which ones, sent when? Jul 18 23:09:42 sgw: I have actually more patches on top to get core-image-sato built Jul 18 23:09:46 sgw: last week Jul 18 23:10:48 http://lists.linuxtogo.org/pipermail/openembedded-core/2011-July/006130.html Jul 18 23:10:50 yes, I am reviewing then and they are pending Jul 18 23:10:54 that thread Jul 18 23:11:00 ignore that series Jul 18 23:11:07 I will send again today Jul 18 23:11:12 which has more fixes Jul 18 23:11:27 and I can sqaush couple of patches that way Jul 18 23:11:40 Ok, they will miss today's request, because I like to "try" and get a build or two in. Jul 18 23:11:55 yes Jul 18 23:12:07 thats ok Jul 18 23:12:22 you need to do bitbake world with them Jul 18 23:12:31 since I touched gettext.bbclass Jul 18 23:13:45 khem: when you'll have 5 mins pls look at the revamped http://paste.debian.net/123287/ Jul 18 23:14:09 I feel ashamed for klcc-cross and I need help before sending it for meta-oe Jul 18 23:18:27 ant__: those comments inside SRC_URI should be deleted Jul 18 23:19:08 the whole patchset need review Jul 18 23:19:41 yes Jul 18 23:19:47 those are obvious ones Jul 18 23:20:01 you should send this as a patch so we can try it out Jul 18 23:20:39 the stopper is that if I don't have klcc together with gcc-cross this is not found on compile Jul 18 23:20:55 and to do this I have to change PACKAGE_ARCH Jul 18 23:21:09 there must be a wiser way Jul 18 23:21:21 khem, still dead? Jul 18 23:21:38 (originally there was a do_stage in the main klibc recipe) Jul 18 23:22:19 ka6sox: yes dead Jul 18 23:27:56 who keeps restarting ntop...20 instances of it... Jul 18 23:28:18 no wonder your server keeps DYING Jul 18 23:28:45 I have no idea Jul 18 23:29:02 would you kindly apt-get remove ntop plz Jul 18 23:29:06 then we will see what fails Jul 18 23:29:17 and we will find the culprit Jul 18 23:29:32 we did that..the LAST time. Jul 18 23:29:58 got logwatch installed? Jul 18 23:30:26 pop that in so you know when someone installs ntop and you can go parse the logs to see who sudo'd over around then Jul 18 23:30:28 think so...tripwire too. Jul 18 23:30:29 next time Jul 18 23:30:38 Oh, well, where are the logs being sent? Jul 18 23:30:40 just root@? Jul 18 23:30:52 let me ask the guy who set it up... Jul 18 23:30:53 Tartarus: its always root Jul 18 23:30:54 (Note you need to unbreak debian/ubuntu, or have int he past, to make it really email them) Jul 18 23:30:56 so we dont know Jul 18 23:30:59 he isn't here right now. Jul 18 23:31:38 khem: are they deleted? Otherwise pop up a mail client :P Jul 18 23:31:46 * Tartarus has his sent to his gmail, fwiw Jul 18 23:34:37 let me do some sleuthing Jul 18 23:37:23 'nite Jul 18 23:38:13 ka6sox: melo was upgraded not installed fresh 10.04 right ? Jul 18 23:38:24 yup... Jul 18 23:38:52 ok so it was installed before upgrade **** ENDING LOGGING AT Tue Jul 19 02:59:56 2011