**** BEGIN LOGGING AT Wed Jul 31 02:59:58 2013 Jul 31 03:04:35 yes but i don't like replacing.. it totally since i need to add all of them that previously included Jul 31 03:05:13 some variable .. EXTRA_OECONF Jul 31 03:05:20 for configurations Jul 31 03:09:53 are you making an append and the parent recipe sets it? Jul 31 03:11:32 in that sort of scenario i don't think you have a choice Jul 31 03:13:03 hmm, are you more comfortable with a sed or grep hack? Jul 31 03:13:05 yes i'm using .bbappend Jul 31 03:13:43 can you paste the EXTRA_OECONF you want to change? Jul 31 03:17:15 then again, i'm not sure this would even work in python... Jul 31 03:18:39 ok there no option like removeing some thing from a variable.. but it would be helpfull if it is there Jul 31 03:19:28 i have some concern, when i launch my image built using Oe-core,it is dead slow.!!! is there any specific reason..? normal build output fast Jul 31 03:19:29 yup, you're not the first one to express that... Jul 31 03:19:38 ohh is it..! Jul 31 03:19:55 i'm using some xcb graphic plugin in qt Jul 31 03:20:18 but it is very slow!! and it is not rendering the application properly Jul 31 03:20:39 just graphics are slow? Jul 31 03:54:24 i didn't really want to try that hack anyway... Jul 31 03:55:59 so for kergoth and other internals folk, would an anonymous python function in the recipe be the right/proper way to remove a string from EXTRA_OECONF in a bbappend? Jul 31 03:56:34 EXTRA_OECONF := "${@oe_filter_out('—something-i-dont-want$', EXTRA_OECONF, d)}" Jul 31 03:56:43 lots of examples of usage around, grep fro oe_filter_out Jul 31 03:57:43 cool, not seen that one yet... Jul 31 03:58:30 only works because := forces it to be evaluated then and there, and oe_filter_out() is a just a convenience wrapper around splitting the value, matching the pattern on each, and returning everything that doesn't match the pattern, as a string Jul 31 04:03:52 you must have some layers i don't... Jul 31 04:04:12 i only see it in a couple of recipes besides eglibc Jul 31 04:04:25 but thanks for the tip Jul 31 04:05:45 * nerdboy hands kergoth the trout Jul 31 04:06:03 i actually found one in my own mupdf recipe... Jul 31 04:06:32 not that i actually wrote it from scratch, but still Jul 31 04:06:52 heh, np Jul 31 04:07:19 i keep a local clone of every layer in the layer archive/index handy locally for grepping/finding Jul 31 04:07:23 really handy from time to time Jul 31 04:07:33 apparently i should look at my own crap more often... Jul 31 05:25:32 morning #oe Jul 31 06:47:34 good morning Jul 31 06:48:07 hi mckoan, all Jul 31 06:55:51 #angstrom Jul 31 06:57:52 How do one build a i686-mingw32 SDK on a x86_64 linux machine? Jul 31 06:58:53 I have tried "SDKMACHINE=i686-mingw32 MACHINE=beaglebone bitbake core-image-minimal -c provides_sdk" without success. Jul 31 07:29:45 The compile log indicates that host include and/or library paths were used. is it dangerous? Jul 31 07:30:07 becasue after building i need to copy the rootfs to some other machin and use Jul 31 07:30:22 why it is using my host libs? Jul 31 07:30:24 any idea Jul 31 07:30:25 ? Jul 31 07:42:34 johny_: yes it's dangerous, read log file to find where it found these references and also check config.log for -L/usr/lib or -I/usr/include etc, it's often wrong configure check for some libs/headers Jul 31 07:42:48 it shouldn't, but someone isn't getting sysroot Jul 31 07:43:29 hmm ok Jul 31 07:43:33 thanq Jul 31 07:44:55 even though -L/usr/lib is there, we wil give prefix to our sysroot right? why becasue for all other modules i'm not getting this type of warning Jul 31 07:45:02 any way i wil check it once Jul 31 07:47:31 any body know where i can get some help (like in IRCs) for graphics related things? such as wayland, xcb, mesa modules Jul 31 07:57:41 morning all Jul 31 08:49:21 morning bluelightning, all Jul 31 08:49:28 morning mckoan Jul 31 09:21:03 morning all Jul 31 09:28:47 hi eren Jul 31 09:41:53 morning all Jul 31 11:23:59 hi, I am using update-alternatives.bbclass, and it seems that it doesn't cope with well with ALTERNATIVE_TARGET_${PN}[name] - I have to spell out ${PN} myself in the recipe. Is this expected? Why is that so? Jul 31 11:24:37 It's a bit counter-intuitive, since e.g. ALTERNATIVE_${PN} works fine Jul 31 12:12:54 bluelightning: ah, it seems that the live image that I produced is not direct bootable Jul 31 12:13:12 I dd'ed into CF card, and now I get "no default or ui configuration found" error Jul 31 12:16:34 ah, maybe I dd'ed the wrong image Jul 31 12:44:11 bluelightning: oh, it was a wrong image :) I tried with "hddimage" and now it's ok Jul 31 12:44:33 however, I have really big problem about VGA resolution. It is just 80x60, which my monitor cannot display of course Jul 31 12:44:36 eren: ah ok, great Jul 31 12:45:00 when I try "scan", I only see 80x60 resolution Jul 31 12:45:03 eren: hmm, I'm not sure - is that something influenced from kernel configuration or the kernel command line maybe? Jul 31 12:45:20 well I've put additional kernel config line, vga=318 Jul 31 12:45:47 it complained that it cannot find that mode and presented me some options. Unfortunately, they were really low res Jul 31 12:46:53 eren: that may be dictated by which framebuffer driver is enabled, but this is not an area I'm particularly expert in Jul 31 12:46:59 APPEND initrd=/initrd LABEL=boot root=/dev/ram0 video=vesafb vga=0x318 console=tty0 Jul 31 12:47:11 this is the config line Jul 31 12:47:13 bluelightning: ah, okkie Jul 31 12:47:24 geode framebuffer is a little bit problematic, huh Jul 31 15:23:16 hm khem broke uclibc on oe-core master and nobody noticed until now Jul 31 15:24:17 I know uclibc isn't tested by the YP. but how long ago did it get broken? Jul 31 15:24:30 lack of users of master, or lack of users of uClibc or? Jul 31 15:25:53 probably a bit of both Jul 31 15:26:07 woglinde: what's wrong with it? Jul 31 15:26:30 interesting.. I know I don't use uClibc, nor do I advocate its use.. (with limited exceptions) most modern machines don't need it anymore.. Jul 31 15:27:49 pb he forgot to checkin merge_config.sh Jul 31 15:27:59 ah, heh Jul 31 15:28:45 fray believe it or not uclibc feels a lot faster than glibc in some cases Jul 31 15:29:01 besides faster buildtime Jul 31 15:29:12 I am not surprised by that... Jul 31 15:29:13 and smaller packages Jul 31 15:29:32 but in the past I've fought enough bugs, not implemented features, that it was no longer worth my time/effort to work with it.. Jul 31 15:29:38 I suspect it's better now then it was 3 years ago.. Jul 31 15:29:51 but even so, I'm not sure the potential incompatibilities/headaches are worth it for me.. Jul 31 15:30:32 if glibc ever moves to a (L)GPLv3... I may have no other choice then uClibc.. but we'll see.. (I am somewhat interested in Bionic as a general embedded Libc.. but I havn't had time to see if that is practical or not Jul 31 15:31:56 hmm Jul 31 15:32:08 fray it has enough features to run openjdk Jul 31 15:32:25 ya... and it's BSD based which has some definite advantages.. Jul 31 15:32:34 no I meant uclibc Jul 31 15:32:38 not bionic Jul 31 15:32:39 somehow I'd say exactly the same thing about bionic ... "I suspect it's better now then it was 3 years ago.." Jul 31 15:32:40 but lack of time has prevented me from seeing if it's useful or not.. Jul 31 15:32:43 ahh I was talking Bionic Jul 31 15:32:57 because 3years ago it was plain awful/incomplete Jul 31 15:32:58 last time I looked at the source for Bionic was about a year ago.. Jul 31 15:33:03 it has definitely advanced a lot Jul 31 15:33:05 bencoh yes 3 years ago it has not support vor armv7 Jul 31 15:33:20 ohmy, I wasn't even thinking about armv7 support :) Jul 31 17:04:19 ka6sox, Do you have a moment to chat about bandwidth? Aug 01 02:38:56 missing a sed-native dep in the sdk recipes? Aug 01 02:39:15 my toolchain build just died... Aug 01 02:40:11 first time i built it since i pulled on master the other night, maybe it's fixed already? **** ENDING LOGGING AT Thu Aug 01 02:59:58 2013