**** BEGIN LOGGING AT Fri Feb 08 02:59:59 2013 Feb 08 08:05:22 Hi all Feb 08 08:09:37 good morning Feb 08 09:12:55 morning all Feb 08 09:43:58 'morning Feb 08 10:08:07 hello Feb 08 11:33:11 morning Feb 08 11:34:06 bluelightning: funny thing I found. conf/site.conf has DISTRO_FEATURES += "ptest" and eglibc-initial fails. bitbake -e tells: DISTRO_FEATURES="ptest pulseaudio sysvinit" but when I remove my += then I have proper list Feb 08 11:35:17 bluelightning: ideas? Feb 08 11:51:34 hrw: I suspect you'll find the rest of DISTRO_FEATURES is being set elsewhere using ?= Feb 08 11:51:54 hrw: if the += is evaluated before the ?=, then the ?= will not do anything because the value is already set Feb 08 11:51:54 probably Feb 08 11:52:09 append works so I am covered Feb 08 11:52:32 want to test that ptest stuff so maybe we will have some more thigns to run on aarch64 models Feb 08 11:52:42 with current master bitbake -e will show variable tracking info so it should be clearer when these situations happen Feb 08 11:53:44 I usually do bitbake -e|grep VARWHICHIWANT but will take a look once at full log Feb 08 12:52:40 are the lists behaving today? Feb 08 13:00:37 systemd booted on aarch64 but I will stay with sysvinit cause it gives me shell in shorter time Feb 08 13:34:44 bluelightning: you just wrote something about += and ?=, which give different result depending on the order of evaluation. Feb 08 13:34:59 Is there a way to track assignment ? Feb 08 13:35:23 ascor_: with latest master, bitbake -e will show a trace of the actions performed for each variable Feb 08 13:35:30 I mean DISTRO_FEATURES = Feb 08 13:35:31 Crofton|work: I think so Feb 08 13:35:35 argl Feb 08 13:36:21 I mean in toto.BB DISTRO_FEATURES = wifi, then in titi.BB DISTRO_FEATURES = wifi bluetooth Feb 08 13:37:13 from what I experienced bitbake -e gives the final variable value, I am more interested in how that variable got its value. Feb 08 13:38:47 ascor_: as I mentioned, the tracing is only available in latest master; previously bitbake -e did not collect or report that information Feb 08 13:39:42 it will be in the next stable release naturally Feb 08 13:40:39 ahhh Feb 08 13:41:16 Sorry. I just got the HEAD of master branch of bitbake this morning. I'll test that Feb 08 13:48:24 That reminds me another question : is the HEAD revision always ok to build, or can it be broken sometimes ? Feb 08 14:38:03 ascor_: it can be broken from time to time, so if you are after stability it's probably best to stick to one of the release branches Feb 08 14:39:35 usually if there are breakages in master they get fixed pretty quickly though Feb 08 14:45:58 ok Feb 08 14:46:26 the new bitbake -e is muuuch more informative ! Feb 08 15:04:14 can someone point gd recipe to existing source? Feb 08 15:04:49 or should I just poke at each OE derived distro source mirrors for it? Feb 08 15:08:43 ascor_: master bitbake would probably work with danny, though likely not with dnezil, but no guarantees Feb 08 15:47:53 when doing this in php : I get a "PHP Notice: iconv(): Unknown error (22)" that is EINVAL aka "An incomplete multibyte sequence has been encountered in the input." Could this be related to eglibc iconv implementation ? Feb 08 16:02:38 afournier: or a php misconfiguration, it's a bit of a monster to configure Feb 08 16:02:47 kinda Feb 08 16:03:02 i found this EGLIBC_OPTION_EGLIBC_CHARSETS Feb 08 16:03:12 but it looks like it's set Feb 08 16:03:31 I'm no expert I'm afraid Feb 08 16:04:02 maybe you know if it's possible to get eglibc configuration at runtime ? Feb 08 16:04:22 I don't, sorry! Feb 08 16:25:34 despite DISTRO_FEATURES does not contain bluetooth, pulseaudio configuration still enable bluetooth. How can this happen ? Feb 08 16:35:46 hi, what's the right way of making a package compatible with a family of machines? shall I list them all in the COMPATIBLE_MACHINE ? Feb 08 16:36:36 or I can export something in the machine's .conf that will allow a simpler COMPATIBLE_MACHINE ? Feb 08 17:01:22 hm why is create-pull-request bash only? Feb 08 17:06:58 bye bye ! Feb 08 18:01:38 bye Feb 08 18:42:22 from a recent oe-core build: /etc/rc5.d/S20udev: line 82: /usr/sbin/udevadm: No such file or directory Feb 08 18:42:44 looks like udevadm is in /usr/bin Feb 08 19:22:52 crofton hm thats wrong Feb 08 19:22:58 should be sbin Feb 08 21:05:38 alright, i have a newly built image with IMAGE_FEATURES += "read-only-rootfs" Feb 08 21:06:33 so for the first-boot configuration stuff, i still want ROOTFS_READ_ONLY=no in default/rcS, correct? Feb 08 21:07:04 * mr_science throws a guinness in kergoth's direction Feb 08 21:07:12 hmm then how is it RO Feb 08 21:07:48 hi khem Feb 08 21:07:48 doesn't it need to be read-write for the opkg postinst stuff to run? Feb 08 21:08:16 i'm talking only for a first config boot, then set it to ro Feb 08 21:09:38 yea or nay Feb 08 21:15:51 well, a quick experimental boot says the answer is an emphatic yes Feb 08 21:18:49 case solved Feb 08 21:18:54 very good Feb 08 21:28:30 i take that back... Feb 08 21:29:10 i thought the huge boot-log spew was errors like i saw yesterday, but it looks like the config stuff actually worked Feb 08 21:29:27 the spew was a glibc stack trace... Feb 08 21:29:27 ? Feb 08 21:29:40 look at do_rootfs Feb 08 21:29:52 in your image temp folder Feb 08 21:30:43 looks like afew errors in the populate volatile stuff too Feb 08 21:31:10 couple of "cp: missing destination file operand..." Feb 09 02:50:36 hello Feb 09 02:51:05 I'm cleaning up a recipe for redis for submission, anyone know a good list of SECTION=""s to use? Feb 09 02:54:14 ooh found it: http://docs.openembedded.org/usermanual/html/section_variable.html Feb 09 02:55:47 boo, mysql is under "libs", but I guess that's close enough **** ENDING LOGGING AT Sat Feb 09 02:59:57 2013