**** BEGIN LOGGING AT Thu Sep 06 02:59:58 2012 Sep 06 07:31:32 good morning Sep 06 08:40:31 morning all Sep 06 08:41:03 bluelightning: hi Sep 06 08:58:19 g m all Sep 06 09:05:39 hi mckoan , hi bluelightning Sep 06 09:33:56 cordial group Sep 06 10:19:11 JaMa|Off: btw, thanks very much for getting those x11 package group changes through for both OE-Core and meta-oe Sep 06 10:26:54 morning all Sep 06 10:32:47 hi pb_ Sep 06 10:36:36 hi bl Sep 06 10:36:38 and pb Sep 06 10:37:46 hi woglinde Sep 06 10:39:38 is bitbake core-image-minimal-initramfs correct way for creating initramfs image? Sep 06 10:39:48 usable with haret? Sep 06 10:41:20 hi bluelightning, woglinde Sep 06 10:42:21 ventyl: it's one way yes, not sure about haret, haven't used it in a while Sep 06 10:42:36 ventyl: which machine are you working with btw? Sep 06 10:43:31 i'm trying to boot it on something which currently isn't supported but it's basically generic PXA270 machine Sep 06 10:44:53 ventyl: ah ok, I was curious if it was actually an old handheld (ipaq, etc.) Sep 06 10:45:05 so as starting point i selected h2200 kernel which may or may not be bootable and i am familiar with it Sep 06 10:45:30 is a little old unfortunately though :( Sep 06 10:45:50 oh how I wish that work had been done against mainline Sep 06 10:46:39 i have h2200 kernel and root filesystem built but core-image-minimal-initramfs gives me error that architecture is unsupported or something similar Sep 06 10:48:21 are you giving it the right architecture number when you boot? Sep 06 10:48:25 hmm... but i don't need it to check if kernel boots at least Sep 06 10:48:49 oh, I see, you mean the recipe gives you that error, not the kernel Sep 06 10:48:56 yep Sep 06 10:50:59 maybe mine assumptions are totally wrong Sep 06 10:54:07 ventyl: can you pastebin the exact error? Sep 06 10:55:23 it's single line Sep 06 10:55:45 initramfs-live-install was skipped: incompatible with host arm-oe-linux-gnueabi (not in COMPATIBLE_HOST) Sep 06 10:59:57 oh, that's a bit odd. initramfs-live-install is apparently for use with grub, so it makes some sense that it wouldn't build on arm, but it isn't immediately obvious to me why core-image-minimal-initramfs would need it. Sep 06 11:00:24 however, the naming of the core-* things has long been slightly obscure to me so there might well be some entirely rational explanation. Sep 06 11:03:11 it could be that it has not been used on anything but x86 Sep 06 11:03:23 since before Intel was involved, before anyone suggests it ;) Sep 06 11:03:55 now i realized that it won't be so easy Sep 06 11:04:11 h2200 kernel won't boot here Sep 06 11:04:46 ventyl: FYI there is some intramfs support added in meta-initramfs in the meta-openembedded repo Sep 06 11:05:10 it's geared towards kexecboot, but is used on zaurus for example Sep 06 11:05:55 it looks like core-image-minimal-initramfs was previously generic but has been x86-only since 4bce3417917a3e88ba6529db394525fba82e0699. Sep 06 11:08:21 worse is that i have to create kernel support for my machine Sep 06 11:09:55 that is, indeed, worse Sep 06 11:10:19 (though, presumably, not altogether surprising) Sep 06 11:12:51 the good - more building blocks/examples in the kernel these days; the bad - ARM kernel code is tied up with devicetree now which makes coding harder (or so an ARM kernel developer friend tells me) Sep 06 11:13:10 well, harder than it used to be, at any rate Sep 06 11:14:25 heh. well, the hardest part of arm kernel development always used to be dealing with russell king. all other issues were trivial in comparison. Sep 06 11:15:20 *g* Sep 06 11:15:26 heh, one of the reasons the handhelds.org kernel development was done separately, no doubt Sep 06 11:16:13 i wonder how much work has to be done to get bootable kernel itself Sep 06 11:16:30 and how much PXA270 machines differ each one Sep 06 11:16:58 probably not much. iirc, pxa270 is a fairly constrained architecture, so it's most likely just a case of sorting out gpio allocations and suchlike. Sep 06 11:18:14 so i have chance to boot kernel to stage where i can see some of it's output Sep 06 11:18:31 yeah, I would have thought so Sep 06 11:18:39 assuming that graphic controller is sort of internal part of PXA270 Sep 06 11:19:27 and LCD should be in state requiring no fiddling with GPIO to get some output Sep 06 11:19:36 after booting throught haret Sep 06 11:52:49 bluelightning: yeah, I think that was a big part of it Sep 06 12:09:35 how can I filter out GPLv3 code from openembedded. I know I can do INCOMPATIBLE_LICENSE="GPLv3", but this filters out everything mentioning GPLv3, also gcc which has an exception with libgcc, as well as xz where only the build system is GPLv3. As far as I know there should be a "sane" way to do this, without removing libgcc and a couple of other packages. Sep 06 12:36:54 mertsas: INCOMPATIBLE_LICENSE is the only mechanism we have for this I think Sep 06 12:37:21 mertsas: it could be that the code needs to be tweaked to allow such things through Sep 06 12:40:31 you can whitelist individual packages; I think that's about the only sensible way to deal with the things you mentioned given the infrastructure that exists today. Sep 06 12:40:51 the license checker is indeed fairly simple-minded and could be made a lot smarter, but someone would have to do the work to make it so. Sep 06 12:42:13 in the particular case of xz, I think you could fix the problem by declaring more detailed license information in the .bb file. Sep 06 12:42:45 actually, come to think of it, that would probably work for gcc as well. Sep 06 12:42:51 so, maybe there is no need for a whilelist in that situation after all. Sep 06 12:43:00 ok Sep 06 12:43:09 how can I whitelist a package then? Sep 06 12:48:26 mertsas: WHITELIST_GPLv3 = "something" Sep 06 12:48:59 ok, thanks bluelightning. I'll do that then, and see how much trouble I get into:P Sep 06 12:50:21 new question btw: We have some problems with the wget module. Is there a good reason why it uses --passive-ftp? We have a couple of build servers here, which don't like passive-ftp, so I'm thinking we either going to have a local patch changing that to --no-passive-ftp, or I can push that upstream if there is no problems related to it? Sep 06 12:54:18 mertsas: it would be worth digging back through the history to try to find out why it was added in the first place Sep 06 13:33:57 bluelightning: found it out as well. Could use a config flag to fix it. Sep 06 22:34:57 http://wbe002.mibbit.com/?server=irc.uk.mibbit.net&nick=enb%3F%3F%3F&channel=%23enbgta4&settings=c76462e5055bace06e32d325963b39f2 Sep 06 22:35:00 oops **** ENDING LOGGING AT Fri Sep 07 02:59:59 2012