**** BEGIN LOGGING AT Fri Aug 10 02:59:57 2007 Aug 10 07:18:25 * xota saluda! Aug 10 08:54:03 koen: what are you crazy Dutch upto now? http://community.livejournal.com/legos/191707.html Aug 10 08:56:52 nice Aug 10 08:57:12 thats mad, but blackfin and uclibc in OE ... crazy talk, koen! Aug 10 13:20:10 koen: ping Aug 10 13:20:26 pong Aug 10 13:20:47 I want to change mokoslug.conf into angstrom.conf + angstrom-mokogateway-image Aug 10 13:21:46 what do you think of the following patch to angstrom-2007.1.conf? http://pastebin.ca/652345 Aug 10 13:22:29 * Crofton|home wants a general way to extend the base angstrom images, without copying from existing bb files :) Aug 10 13:22:57 Crofton: in this case, angstrom-mokogateway-image already exists Aug 10 13:23:01 rwhitby: I don't like that patch Aug 10 13:23:18 koen: good - that's why I'm asking :-) Aug 10 13:23:27 it stops people from buiding an image they can chroot into and debug stuff Aug 10 13:23:53 is there another way to remove those debug apps from the jffs2 image? Aug 10 13:24:00 you can put 'DEBUG_APPS_nslu2 = ""' in the -image.bb Aug 10 13:24:00 (since it won't fit with them) Aug 10 13:24:15 ok, that works for me. Aug 10 13:24:27 how about the second bit (the addition of the sysconf package) Aug 10 13:24:38 I'm not sure about that Aug 10 13:24:54 could you ask about that on the mailinglist? Aug 10 13:25:13 ok, Aug 10 13:25:19 What is the new way of doing the following: Aug 10 13:25:26 IMAGE_PREPROCESS_COMMAND_append_nslu2 = "rm ${IMAGE_ROOTFS}/boot/zImage*;" Aug 10 13:26:01 not sure about that either, ask RP or psokolovsky Aug 10 13:26:34 koen: putting the DEBUG_APPS line in angstrom-mokogateway-image.bb then breaks the orthogonality between distros and images - i.e. it ties that image to angstrom - it that intended? Aug 10 13:26:37 I assume you want that for jffs2 images, but not ready-for-hd images (with kexec or something like that) Aug 10 13:26:46 correct. Aug 10 13:27:48 RP: ping Aug 10 13:27:53 in this case DEBUG_APPS does nothing (AFAIK) in other distros Aug 10 13:28:03 ok, that's true. Aug 10 13:28:18 but for the general case you have a point Aug 10 13:30:19 I think I can remove the DISTRO_FEATURES_nslu2 line from angstrom.inc now too. Aug 10 13:31:23 I presume you don't mind if I make changes to the contents of angstrom-mokogateway-image.bb (i.e. adding some more stuff to the image) Aug 10 13:31:42 rwhitby: That used to work and doesn't anymore? Aug 10 13:31:57 RP: that works fine, but I thought there was a new way to do it now? Aug 10 13:32:09 rwhitby: images are mostly fair game Aug 10 13:32:22 rwhitby: Ah, you mean through making the kernel image empty? Aug 10 13:32:38 RP: yes - I don't remember how to do that Aug 10 13:32:42 rwhitby: I can't honestly remember. I did mention it on the list though Aug 10 13:32:50 rwhitby: I want to encourage people to add new images, we can always tweak/delete/etc the ones that don't make sense Aug 10 13:33:24 koen: yeah, I saw that strong recommendation of yours, which is why I'm looking to see whether mokoslug.conf distro can be removed. Aug 10 13:33:29 rwhitby: a quick search suggests RDEPENDS_kernel-base = "" Aug 10 13:33:45 rwhitby: (in the machine.conf file) Aug 10 13:34:05 rwhitby: I did commit the code for that, I don't think I got around to activating it Aug 10 13:34:33 RP: Hmm - how does that work for both jffs2 images (where I don't want a kernel) and kexec to hd images (where you do want a kernel) ? Aug 10 13:35:22 rwhitby: It doesn't. It was assumed either the image needs a kernel or it doesn't Aug 10 13:35:49 right, but putting that in the machine.conf means you can't choose by the image Aug 10 13:36:03 rwhitby: Nobody mentioned that contraint when we discussed this last time :/ Aug 10 13:36:23 RP: agreed - not complaining here - just raising a new scenario none of us considered before. Aug 10 13:37:00 rwhitby: I guess you will have to do that at image generation time Aug 10 13:37:33 will that RDEPENDS line work in the image.bb ? Aug 10 13:37:33 (I'm expecting not) Aug 10 13:37:50 rwhitby: No, its needed by the kernel package Aug 10 13:38:10 ok, I'll put the original line in the image file then :-) Aug 10 13:38:30 rwhitby: probably easiest :) Aug 10 13:38:38 Are lines like the following appropriate for a linux.bb file: Aug 10 13:38:53 module_autoload_ixp4xx_mac = "ixp4xx_mac" Aug 10 13:38:53 module_conf_ixp4xx_mac = "alias eth0 ixp4xx_mac" Aug 10 13:39:11 (the ixp4xx needs that module autoloaded for eth0 to come up on boot) Aug 10 13:39:41 (it will be in ixp4xx-linux_2.6.21.6.bb in this case) Aug 10 13:39:51 Yes, although I think that will break if MACHINE=ipx4xx :/ Aug 10 13:40:08 did you mean != ? Aug 10 13:40:36 rwhitby: No, MACHINE is an override so module_autoload_ipx4xx_mac will become module_autoload_mac Aug 10 13:40:48 oh, damn. Aug 10 13:41:01 Which highlights a nice problem with the existing infrastructure Aug 10 13:41:20 Personally, I'd like to throw away that autoload syntax anyway Aug 10 13:41:38 any way I can inhibit the substitution? What's the alternative you want to replace it with? Aug 10 13:42:08 rwhitby: A sane variable like AUTOLOAD_MODULES = "ixp4xx" Aug 10 13:42:18 rwhitby: You can't inhibit it Aug 10 13:42:31 can I do some python setvar or something? Aug 10 13:42:58 rwhitby: no, overrides are unavoidable Aug 10 13:43:45 * koen mentions we only have MACHINE=ixp4xx{le,be} Aug 10 13:44:06 koen: Do they not set MACHINE=ixp4xx internally? Aug 10 13:44:17 I don't hope so Aug 10 13:45:15 no, ixp4xx stuff is clean. nslu2le and nslu2be both define nslu2 and ixp4xx overrides however. Aug 10 13:46:55 and in this case it would be nslu2le, and it would define an ixp4xx override, and it would trigger the problem. Aug 10 13:47:03 rwhitby: The current module loading code is blackmagic to me but I really don't understand those autoload variables, the syntax seems a bit mad... Aug 10 13:47:59 and worse, you have to rebuild the kernel for it to take effect, even if the autoload statements are scattered all over the metadata Aug 10 13:49:58 is it a new thing, or has it been there forever? Aug 10 13:50:16 (and is there a better way to do it in an image file for instance?) Aug 10 13:50:32 its been there as long as I remember Aug 10 15:08:16 hi everyone. Aug 10 15:08:56 does anyone know the way boost the zaurus sound a bit ? Aug 10 15:44:28 magnet_: play in alsamixer Aug 10 16:13:04 hi all Aug 10 17:04:20 hello Aug 10 17:04:35 what is the present state of the N800 build? Aug 10 17:05:09 does it work the same way as Poky? Aug 10 18:32:18 has any one tried angstrom on the n800? Aug 10 22:52:59 hi all; I'm having trouble playing sound on my asus730w. The device seems to be recognised; I have a huge array of controls in alsamixer, but when I try and play sounds I can't hear anything, either from the speaker or through headphones Aug 10 23:08:17 lmm: maybe Aug 10 23:08:20 !oebug 2667 Aug 10 23:08:22 * * Bug 2667, Status: NEW, Created: 2007-07-19 08:07 Aug 10 23:08:23 * * uncle.grog(AT)gmail.com: Sound problems with recent kernels a.k.a. asound statefiles need updating Aug 10 23:08:23 is something for you Aug 10 23:08:24 * * http://bugs.openembedded.org/show_bug.cgi?id=2667 Aug 10 23:08:33 thanks Aug 10 23:08:41 lmm: how did you get angstrom on your asus?? Aug 10 23:08:51 are there images Aug 10 23:08:58 I've bitbaked my own Aug 10 23:09:16 but how Aug 10 23:09:26 I'm not aware of any prebuilt images being available Aug 10 23:09:36 just using the openembedded tree Aug 10 23:09:52 lmm my brother's working for asus and I'd like to try angstrom on one mypal Aug 10 23:10:04 lmm what did you set as machine type? Aug 10 23:10:09 asus730 Aug 10 23:10:11 it's in there Aug 10 23:10:16 ah i see Aug 10 23:10:27 though the kernel config is wrong Aug 10 23:10:31 are there also definitions for other mypals? Aug 10 23:11:42 IIRC the 716 port is pretty mature but the others are barely started Aug 10 23:12:42 ok you said the kernel config is wrong. did you have to change much in order to make it boot? Aug 10 23:12:46 I can see a config file for a 620 Aug 10 23:12:49 nothing to make it boot Aug 10 23:12:58 but the touchscreen driver isn't enabled by default Aug 10 23:13:06 which makes it pretty useless Aug 10 23:13:38 but the right one is in oe? Aug 10 23:14:48 not as far as I know Aug 10 23:15:03 it may have been fixed since I last looked though Aug 10 23:15:22 maybe you could commit it? Aug 10 23:16:16 do the other defined asus machines in oe use the same driver? Aug 10 23:16:36 I'll submit a bugreport when I get round to it; I have no idea what driver the other asus machines use Aug 10 23:16:38 would be very useful thing then I guess. Or at least submit to the bugtracker Aug 10 23:17:21 is there any way to contact you? You registred on the bugs.openembedded.org yet? Aug 10 23:17:33 no, I haven't Aug 10 23:17:40 I'll try and hang around here Aug 10 23:17:54 and I'll register soon, probably under this name Aug 10 23:20:39 ok because I'm sure I will try one mypal in the nex weeks Aug 10 23:20:48 so we could share experiences Aug 10 23:29:57 bah, it's not just the asound.state Aug 10 23:30:06 or at least, I don't think it is. There are so many controls Aug 10 23:34:21 I think the many control are pretty normal Aug 10 23:34:31 I also have like 20 or so Aug 10 23:34:39 on my zaurus sl-C1000 Aug 10 23:38:31 yeah, but it means I don't know what they should be set to because I don't know what they mean Aug 11 01:37:11 is it possible that a machine's name can vary through ipkg upgrades? Aug 11 01:37:43 I could swear it was akita before, now it is ixp4xxle Aug 11 01:55:24 dcordes: it could if base-files had /etc/hostname in it, and didn't mark it as a conffile, or you accepted the upgrade of the conffile question. Aug 11 01:56:40 ok could be that I didn't notice and accepted last time on upgrade then Aug 11 01:56:49 was just paranoid ^^ Aug 11 01:57:05 rwhitby: you are the www.nslu2-linux.org guy aren't you? Aug 11 01:57:10 yep Aug 11 01:57:38 I just had a conversation about slug OSs with an owner of a nslu2 and we came to angstrom Aug 11 01:57:47 is it any useful on the slug? Aug 11 01:58:54 it's not ready for the slug yet. it runs fine, but doesn't have the features to turnup to disk or automatically start the right modules for networking out of the box. you need a serial cable. Aug 11 02:00:19 so you can't boot it from usb attached volumes ulike the other slugs OS? Aug 11 02:00:34 not unless you get altboot working Aug 11 02:00:52 (all the bits are there, someone just has to glue them all together into something that works) Aug 11 02:01:16 I see thanks for info Aug 11 02:01:30 I'd like to help with that but I have no slug Aug 11 02:01:32 a lot of work went into slugos to get that stuff working. but it was all done slugos-specific. we want to redo it in angstrom, but in a generic way so that all OE distros can use the features. Aug 11 02:01:59 (e.g. we have sysconf reimplemented and working in angstrom so far) Aug 11 02:02:04 yea I'm a big fan of OE compatibility Aug 11 02:02:21 I was worrying about openmoko not being synced with OE Aug 11 02:03:29 but that changed didn't it? Aug 11 02:05:15 soon Aug 11 02:05:34 i.e. they expect to announce the switch this coming week Aug 11 02:16:36 rwhitby: great to hear :) Aug 11 02:18:30 however it will be interesting to see how OE reacts to a whole new set of users, most of which have never used OE before. Aug 11 02:18:38 (at least not OE .dev) Aug 11 02:20:52 anyway I bet it is the best way to go Aug 11 02:22:39 so OE can participate from all the openmoko progress Aug 11 02:22:43 and vice versa Aug 11 02:26:07 yeah, there will be teething pains, and it will be a test of the OE core devs to see if they can handle all the new developers without getting angry at them **** ENDING LOGGING AT Sat Aug 11 02:59:57 2007