**** BEGIN LOGGING AT Mon Nov 16 02:59:56 2009 Nov 16 07:21:19 good morning Nov 16 07:51:22 good morning Nov 16 08:40:00 morning Nov 16 08:56:19 I was able to make tslib depending on override dm365-evm, but I don't understand which is the ipk to be re-installed Nov 16 09:01:22 I'm trying to use usb mouse on arago, but I need to change tslib Nov 16 09:05:30 hi Nov 16 09:06:49 hrw: hi Nov 16 09:06:59 I was able to make tslib depending on override dm365-evm, but I don't understand which is the ipk to be re-installed Nov 16 09:07:26 recalcati__: I dont understand your question Nov 16 09:09:41 I change the recipe, but I don't understand which is the ipk to be re-installed, it seems libts-dev, but I get Package libts-dev md5sum mismatch. Either the opkg or the package index are . Nov 16 09:10:07 recalcati__: rm /usr/lib/opkg/lists/* and then install Nov 16 09:11:23 I did package-index Nov 16 09:13:02 but have also angstrom feeds there? Nov 16 09:13:14 now I'm re-compiling all qt, after I'll remember all your helps Nov 16 09:13:28 hrw: I'm in arago overlay for TI PSP Nov 16 09:13:45 my apache provide the ipk Nov 16 09:13:55 strange then Nov 16 09:14:04 what? Nov 16 09:15:26 ok, from begining then Nov 16 09:15:44 1. you run 'bitbake package-index' on your devmachine Nov 16 09:16:02 2. you run 'opkg update;opkg install libts-dev' on devboard Nov 16 09:16:11 3. you got md5 mismatch? Nov 16 09:22:20 03Jeremy Lainé  07org.openembedded.dev * r5f13a271b0 10openembedded.git/recipes/linux/linux-2.6.31/boc01/005-091110-isl12024.patch: linux-2.6.31 : fix email for isl12024 driver author Nov 16 09:27:04 good morning Nov 16 09:27:48 hrw: yes, I did exactly that Nov 16 09:27:50 hi florian Nov 16 09:28:05 grrrrrrrr Im getting more and more annoyed at rday who is earning money from OE but too fscking lazy to help Nov 16 09:28:25 recalcati__: then no idea what can be broken Nov 16 09:28:37 recalcati__: rm /usr/lib/opkg/lists/*;opkg update;opkg install libts-dev Nov 16 09:29:18 "/usr/lib/opkg/lists/*': No such file or directory " Nov 16 09:29:19 XorA: yep. Nov 16 09:29:30 recalcati__: so check where opkg stores it Nov 16 09:29:38 * florian is back to work... finally Nov 16 09:29:47 florian: morning :) Nov 16 09:29:52 florian: I've read your message Nov 16 09:30:13 florian: what database settings should I pass to wiki? Nov 16 09:30:15 * XorA shoulda stayed epically drunk Nov 16 09:30:36 Jay7: Great - so the plugin works now, I have to set up a database for it only. Nov 16 09:30:39 mornng all Nov 16 09:30:44 hi RP Nov 16 09:30:47 hey RP Nov 16 09:31:19 florian: ah.. then I'll just wait some time :) Nov 16 09:31:30 btw is there someone who is using multiple buildhost and needs to have SRCPVs in sync? Nov 16 09:31:42 hrw: /usr/lib/opkg/lists is empty Nov 16 09:32:00 as Koen asked in "Re: [oe] SRCPV migration", I'm not sure what is easiest way to do this.. Nov 16 09:32:24 JaMa|Wrk: Angstrom Nov 16 09:32:44 JaMa|Wrk: koen and I both do different builds for Angstrom feeds on different machines Nov 16 09:33:03 JaMa|Wrk: the persistant cache can be implmented as xmlrpc Nov 16 09:33:07 i guess you can sync bb_persist_data.sqlite3 before building.. but seems a bit error-prone Nov 16 09:33:07 recalcati__: check where 'opkg update' stores Nov 16 09:33:13 JaMa|Wrk: but my python is useless Nov 16 09:33:41 XorA: There is no change from the present with the persist_data cache? Nov 16 09:33:44 JaMa|Wrk: but how I, as a random developer, will sync with them? I do angstrom feed build rarely Nov 16 09:34:29 RP: sorry didnt parse that Nov 16 09:35:08 I have no idea how to sync it properly, when I'm building image for myself then I'll always use only my feeds Nov 16 09:35:31 rm var/lib/opkg/armv5te same result Nov 16 09:35:34 and If I would like to start building locally from image installed from other builder I would download hist bb_persist_data.sqlite3 before building locally Nov 16 09:35:38 JaMa|Wrk: the best fix would be to code up some network database capable persisteant methods Nov 16 09:35:58 JaMa|Wrk: which RP told me should be simple the way he designed the code Nov 16 09:36:13 XorA: How do ou currently solve that problem? Nov 16 09:36:23 XorA: probably.. but then every builder should be able to push his updated revs there Nov 16 09:36:29 RP: persistant cache isnt used in any Angstrom package Nov 16 09:36:50 JaMa|Wrk: thats not an issue I think Nov 16 09:36:55 XorA: ok, so the move to SRCPV doesn't change anything really... Nov 16 09:37:17 RP: well the PV will now change depending on the machine its built on Nov 16 09:37:36 RP: where as before it didnt Nov 16 09:38:07 before it did only for packages with AUTOREV, didn't it? Nov 16 09:38:22 JaMa|Wrk: yes Nov 16 09:38:35 which Angstrom douesnt use Nov 16 09:38:58 hrw: the opkg verbose is here http://pastebin.com/f660f67c5 Nov 16 09:42:04 no idea recalcati__ Nov 16 09:42:37 ok, after my full recompilation I hope it will be ok Nov 16 09:43:16 maybe -c clean of packe-index can help , I will try Nov 16 10:09:19 pb__: I'm slowly catching up with my mail... did you have success with the greylist whilteliat at ltg? Nov 16 10:11:59 florian: ah, not yet. it turns out that the greylistd config doesn't currently have provision for a per-domain whitelist and I didn't get around to adding one. Nov 16 10:12:16 if you wanted to do that, that'd be awesome Nov 16 10:24:56 pb__: just found it... "whitelist-hosts" is the magic file and it takes domains as well. Nov 16 10:25:17 florian: oh, does it? very good Nov 16 10:25:23 when I looked at that before, it seemed just to be for hosts. Nov 16 10:26:29 03Richard Purdie  07org.openembedded.dev * rd762c75040 10openembedded.git/classes/ (gconf.bbclass gtk-icon-cache.bbclass mime.bbclass): Nov 16 10:26:29 gconf/gtk-icon-theme/mime: Really fix PKGDEST references Nov 16 10:26:29 Signed-off-by: Richard Purdie Nov 16 10:27:19 pb__: I ran into the same before... but I found a hint in a german description and looked at the documentation in the file itself. Nov 16 10:28:46 pb__: It supports both simple wildcards and regular expressions. Nov 16 10:29:02 so we should have fixed this :) Nov 16 10:30:34 okay, excellent Nov 16 10:47:36 * XorA wonders what the issue is with guile-native Nov 16 10:48:10 * JaMa|Wrk wonders what will happen with SRCPV now :/ Nov 16 10:50:13 03Richard Purdie  07org.openembedded.dev * rbc465d83ad 10openembedded.git/classes/ (31 files): Nov 16 10:50:13 classes: Drop a number of unneeded import calls (from Poky) Nov 16 10:50:13 Signed-off-by: Richard Purdie Nov 16 11:04:13 my build failed I demand that you fix it!! Nov 16 11:19:53 XorA: The issue... might be FC11 :) Nov 16 11:21:07 florian: its all OEs fault, Out of Diskspace Error Nov 16 11:21:38 heh Nov 16 11:21:53 actually 100G of tmp/ directories Nov 16 11:21:54 wow Nov 16 11:22:23 XorA: You are building too many beagleboard-demo-images with Angstrom :) Nov 16 11:29:15 just 100GB? Nov 16 11:42:09 xora, http://pastebin.ca/1673327 Nov 16 11:47:00 hrw: I did bitbake -c clean package-index; bitbake package-index; and now works Nov 16 11:47:22 Crofton|work: been seeing that on x86 for a while also (was Ubunutu 8.04, now 9.10). Also see something similar in OPKG (but the resulting images are fine). Nov 16 11:48:02 hi zecke Nov 16 11:49:44 Crofton|work: heh thats anew Nov 16 11:52:12 crap I forgot about the libsdl-mixer breakage Nov 16 11:55:33 pb__: master! Nov 16 11:55:42 pb__: thanks for the check sum patch Nov 16 11:56:08 RP: I have managed to build a meta-toolchain for MIPS el... but didn't try to build something with it yet Nov 16 12:05:02 zecke: yeah, I think this way of doing checksums should be an improvement in general. Nov 16 12:05:43 thanks for your review of the patch. Nov 16 12:26:09 zecke: cool :) Nov 16 12:39:16 RP: another minor issue after last-week changes...for some reasons 3 recipes having QA (non -dev containing .so) did build without QA....magically? Nov 16 12:39:35 well, recipes were libpcre, libopieobex0, libgtkstylus. The latter was revisited in the last days and got PR bump, but the other two not yet. Nov 16 12:41:03 RP: can we get legacy staging recipes logged into tmp/legacy-staging.log? Nov 16 12:41:52 RP: I was ready to fix the packaging of the recipes one by one but now I'm a bit confused.. Nov 16 12:46:48 ant_work: Some QA messages are warnings, some are fatal Nov 16 12:47:00 hrw: Yes, we can. Someone just needs to write the patch Nov 16 12:47:44 I mean only about that specific QA: -dev packaging Nov 16 12:49:20 it seems QA stopped after 2009-11-10 afternoon Nov 16 12:49:42 ant_work: all QA? Nov 16 12:51:02 03Richard Purdie  07org.openembedded.dev * r9cbd93cd04 10openembedded.git/classes/package.bbclass: Nov 16 12:51:02 package.bbclass: Add back import os until further metadata cleanups happen Nov 16 12:51:02 Signed-off-by: Richard Purdie Nov 16 12:54:22 RP: http://tinderbox.openembedded.net/packages/libpcre/ Nov 16 12:56:13 libgtkstylus was probably fixed Nov 16 12:57:05 libopieobex0 needs just a fix in palmtop.bbclass Nov 16 12:57:19 but I still don't see it.. Nov 16 12:57:33 so QA should be present Nov 16 13:02:14 ant_work: Try my last commit Nov 16 13:02:25 03Richard Purdie  07org.openembedded.dev * r777b4f6c6b 10openembedded.git/classes/insane.bbclass: Nov 16 13:02:25 insane.bbclass: Remove hardcoded path PKGDEST issue Nov 16 13:02:25 Signed-off-by: Richard Purdie Nov 16 13:02:25 RP: remaining QA are only about 'invalid' desktop files. These 3 were the last 'serious' Nov 16 13:02:29 ah, thx Nov 16 13:02:42 I'll do in a few hours @home Nov 16 13:03:09 RP: any gift you want from iceland? Nov 16 13:03:48 zecke: its ok thanks :) Nov 16 13:31:44 RP: patch for LOCALCOUNT ready.. I'll send to ML for review.. Nov 16 13:36:54 I want to load the g_file_storage module on boot. Which recipe/module should I fiddle with? Nov 16 13:39:41 tasslehoff: kernel recipe or machine config Nov 16 13:40:16 JaMa|Wrk: ok. thanks. Nov 16 13:40:34 tasslehoff: you need just module_autoload_g_file_storage = "g_file_storage" Nov 16 13:41:11 JaMa|Wrk: and with the arguments/parameters at the end? Nov 16 13:44:05 module_autoload_g_file_storage = "g_file_storage" Nov 16 13:44:10 with parameter at the end Nov 16 13:44:22 module_autoload_g_file_storage = "g_file_storage blah=blub" Nov 16 13:44:25 tasslehoff: parameters for modprobe? Nov 16 13:45:00 module_conf_g_file_storage = "params" Nov 16 13:45:17 JaMa|Wrk: what spaetz said :) "g_file_storage file=/dev/foo removable=1" Nov 16 13:45:52 tasslehoff: module_conf_g_file_storage will create /etc/modutils/g_file_storage.conf with "params" Nov 16 13:46:20 module_autoload_g_file_storage = "g_file_storage blah=blub" will create /etc/modutils/g_file_storage with "g_file_storage blah=blub" Nov 16 13:46:21 ah. I didn't see the difference in your lines until now. thanks :) Nov 16 13:46:38 and I can't add this to my local.conf? Nov 16 13:47:19 tasslehoff: I guess so.. Nov 16 13:48:05 JaMa|Wrk: but the '..' means that you don't think it is good practice? :) Nov 16 13:53:11 03Marcin Juszkiewicz  07org.openembedded.dev * rd20b32225e 10openembedded.git/recipes/xorg-xserver/ (xserver-xorg-conf/bug/xorg.conf xserver-xorg-conf_0.1.bb): xserver-xorg-conf: added BUG support Nov 16 13:53:11 03Marcin Juszkiewicz  07org.openembedded.dev * r740361d57e 10openembedded.git/conf/machine/bug.conf: bug: ship SDIO firmware for WiFi Nov 16 13:53:12 03Marcin Juszkiewicz  07org.openembedded.dev * red895c630f 10openembedded.git/MAINTAINERS: MAINTAINERS: added BUG to my devices Nov 16 13:54:53 JaMa|Wrk: only reason I have for putting it in local.conf is that I know it stays with me even if the machine or kernel recipes changes. Nov 16 13:55:01 anyway. thanks for the help Nov 16 13:56:19 tasslehoff: it depends what you want to accomplish and if it would be usefull for all users of that recipe/machine :) Nov 16 13:58:06 JaMa|Wrk: yep. I think the machine.conf is the right place :) Nov 16 14:04:39 RP: isn't ${SRCPV} created as gitr${LOCALCOUNT}+HASH, so with my patch you have constatnt gitr0+? seems working here.. Nov 16 14:05:33 RP: or you want to preserve count for othere usages somewhere else and modify just FORMAT of ${SRCPV} if you don't want to include LOCALCOUNT? Nov 16 14:07:37 hrw: how do you find SDIO performance on the BUG with WiFi? Nov 16 14:09:51 DJWillis: I have builtin ethernet which do 50Mbps so I do not use wifi Nov 16 14:10:37 hrw: ahhh, I just noticed the commits regarding the SDIO firmware for WiFi on the BUG so wondered what speed was like. Nov 16 14:17:45 bye Nov 16 14:18:37 RP: ping Nov 16 14:27:22 RP: my clean build of gnome didn't work :-( -- same postinst issue Nov 16 14:27:50 RP: I see you redid one of the fixes -- I'll try again Nov 16 14:36:09 if I change my machine.conf, do I have to force a rebuild of something to make it take effect? Nov 16 14:36:32 depends exactly on what you changed Nov 16 14:36:38 JaMa|Wrk: ^. I added the module-stuff you gave me, but lsmod shows nothing loaded Nov 16 14:36:59 tasslehoff: you need to rebuild kernel you're using Nov 16 14:37:18 ah. of course.. Nov 16 14:37:30 tasslehoff: and kernel-module-g-file-storage should be changed (and containt those 2 files) Nov 16 14:38:50 JaMa|Wrk: that last one went over my head :) Nov 16 14:47:24 JaMa|Wrk: what is kernel-module-g-file-storage? I only see it referenced in task-base.bb. and which 2 files did you mean? Nov 16 14:51:12 tasslehoff: it should be .ipk file generated from kernel recipe Nov 16 14:51:21 tasslehoff: if you have that module enabled in kernel config Nov 16 14:51:39 ah. you meant it should be changed after my build? Nov 16 14:51:46 tasslehoff: and inside is control.tar.gz and data.tar.gz Nov 16 14:51:48 sakoman: yes, I screwed up the patches, sorry :( Nov 16 14:51:52 I thought it was something I had to modify :) Nov 16 14:52:03 and data.tar.gz should contain actuall module and those 2 /etc/modutils/ filed Nov 16 14:52:05 files Nov 16 14:52:42 JaMa|Wrk: My point was that I'd like the distro to be able to set the localcount for a given package should it wish to Nov 16 14:54:11 *** glibc detected *** opkg-cl: realloc(): invalid next size: 0x00000000025e6880 *** Nov 16 14:54:18 this hangs the build Nov 16 14:54:32 RP: so no global config entry, but for every package if needed? Nov 16 14:54:51 RP: no worries, I'm trying again with your new patches Nov 16 14:54:56 JaMa|Wrk, it seems like this is too make it easier for people using OE to develop SW? Nov 16 14:56:07 JaMa|Wrk: Basically bitbake should just check a variable like LOCALCOUNT_OVERRIDE and use that if its set Nov 16 14:56:13 Crofton|work: I find that I get opkg-cl errors simialr to that on my gnome image -- just after the image is created, so I've never spent too much energy tracking down why Nov 16 14:56:44 JaMa|Wrk: We can let the distro worry about setting it (BB_LOCALCOUNT_pn-somepackage = "12") ? Nov 16 14:57:00 the build doesn't hang in my case, though it feels like it. If I just wait it completes sometimes with another similar error or two. Nov 16 14:57:05 also the do_rootfs log shows: opkg: Cannot create directory `/var/lib/opkg': Permission denied Nov 16 14:57:20 I was seeing those messages, I don't know why :/ Nov 16 14:57:24 Crofton|work: that message has been there for ges Nov 16 14:57:28 ages Nov 16 14:57:31 I think its an opkg problem though Nov 16 14:57:31 this looks like it is trying to write in to the building machine var Nov 16 14:58:10 RP: Ah I see your point but is it really needed? (will someone use it that way?) I hoped that with gitr0+Hash for ALL SRCPV recipes will make Koen happy and thats all Nov 16 15:01:06 RP: or we can change that localcount increment configurable option (default 1, 0 for angstrom) and then he can set 12 for somepackage with sqlite and it will stay 12 forever Nov 16 15:01:20 RP: a quick test rebuild of metacity shows that it now has a good postinst! Nov 16 15:01:43 RP: I'll fire off a new clean build now Nov 16 15:04:17 JaMa|Wrk: Setting 12 with sqlite is not what is required Nov 16 15:05:22 I mean if he or someone else really wants Nov 16 15:10:30 JaMa|Wrk: I put the lines you suggested in beagleboard.conf, cleaned and built the kernel, but data.tar.gz only contains the module (and it is not loaded) Nov 16 15:14:00 tasslehoff: do you have updated oe.dev? Nov 16 15:14:53 tasslehoff: it was fixed in 2c09164fd4c11832865f18b72f46266b7abffbc7 Nov 16 15:20:57 JaMa|Wrk: ah. that's just a couple of days ago :) Nov 16 15:26:12 tasslehoff: and it was broken by 29c7d3351f43678c6e93b707b301832009f64b31 also just few days before fix.. Nov 16 15:26:43 JaMa|Wrk: thanks again :) Nov 16 15:34:28 re Nov 16 15:35:25 * mwester is quite happy; the SlugOS image boots (from flash at least) just fine on real hardware, after all the OE changes recently. ;) Nov 16 15:36:08 Woo! And the package feeds built in their entirety, without any failures or restarts. Nov 16 15:36:24 Quick! Tag that point on the dev branch! :D Nov 16 15:38:34 mwester: cool :) Nov 16 15:38:44 mwester: You sound surprised? ;-) Nov 16 15:39:56 hehe! I've ignored the SlugOS autobuilder for the most part, since early this summer -- I'm really surprised that it took so little effort to make it all work over the past months. Nice job by everyone to keep .dev working so well. Nov 16 15:47:59 Hi!. I am trying to build oe (angstrom distribution) and have some build problems: http://dpaste.com/120148/ and http://dpaste.com/120149/ (that's the log) Nov 16 15:48:17 Is this a known problem? Nov 16 15:53:37 hi how can i use external toolchain in OE?? Nov 16 15:53:58 i tried to add the following line in local.conf Nov 16 15:54:00 "TARGET_PREFIX = arm-none-linux-gnueabi- Nov 16 15:54:00 ASSUME_PROVIDED += "virtual/${TARGET_PREFIX}gcc virtual/libc"" Nov 16 15:54:18 but OE seems to build the cross compiler again Nov 16 16:05:06 Is MACHINE_FEATURES a free-for-all? Nov 16 16:08:07 montamer, there was a thread in the mailing list about that recently Nov 16 16:08:27 montamer, so the best option is to find this thread Nov 16 16:40:43 What is more correct to put in a machine conf file: Nov 16 16:41:37 PREFERRED_PROVIDER_xserver = "xserver-kdrive" or PREFFERED_PROVIDER_virtual/xserver ="xserver-kdrive" Nov 16 16:41:39 ? Nov 16 16:42:35 Does xserver and virtual/xserver have any difference? Nov 16 16:42:35 neither of those is really correct Nov 16 16:42:40 Oh Nov 16 16:42:46 anything that starts "PREFFERED" is pretty much a dead loss Nov 16 16:42:54 :) Nov 16 16:43:03 it needs to be PREFERRED_PROVIDER, with that number of Fs and that number of Rs, to have any useful effect at all Nov 16 16:43:12 and, afaik, only virtual/xserver is used as a virtual Nov 16 16:43:29 so, I think "PREFERRED_PROVIDER_virtual/xserver" is probably the magic combination Nov 16 16:43:36 With one F? Nov 16 16:43:43 And two Rs? Nov 16 16:43:55 well, four if you count the Rs in "PROVIDER" :-} Nov 16 16:43:56 but, yes Nov 16 16:44:03 Thanks, dude. ;D Nov 16 17:01:30 florian: heh, I see you spoke to Nils about those cpu modules Nov 16 17:02:18 pb__: heh yes, I remebered the fact we forgot them when I saw Nils at the office :-} Nov 16 17:03:01 RP: after the pull this morning my kernel build fails! Nov 16 17:03:29 florian: yeah, that was dumb of me. oh well. Nov 16 17:03:47 * florian was not much better Nov 16 17:03:54 RP: http://pastebin.com/m1f0104a3 Nov 16 17:04:34 RP: I think this might be related to your kernel.bbclass changes Nov 16 17:04:37 florian: morning again ;) Nov 16 17:05:06 Jay7: hehe... yes right, the database... Nov 16 17:07:13 sakoman: alsmost certainly :( Nov 16 17:09:46 hi, is there a way to generate an initramfs and a rootfs in the same build process, or do I have to create a different task for the initramfs, build it, then build the rootfs? Nov 16 17:10:11 sakoman: fix pushed (I hope) Nov 16 17:10:17 03Richard Purdie  07org.openembedded.dev * r2eb99471ec 10openembedded.git/classes/kernel.bbclass: Nov 16 17:10:17 kernel.bbclass: Revert import os removal for now Nov 16 17:10:17 Signed-off-by: Richard Purdie Nov 16 17:15:25 Anyone? I can't udnerstand gcc-cross-initial-4.2.4-r5 can't be built on my system. Error: http://dpaste.com/120148/ Log: http://dpaste.com/120149/ Nov 16 17:16:25 Why the makefile is missing "install" target? I don't think it is correct at all Nov 16 17:21:03 Jay7: done Nov 16 17:21:33 florian: thank you! I will check a bit later Nov 16 17:22:01 Jay7: yw Nov 16 17:29:23 Hmm, my script foo is weak today. How do I make sed remove each and every -L option on a line? :/ Nov 16 17:32:52 RP: sed 's/-L//g'? Nov 16 17:33:35 RP: linux built now :-) thanks Nov 16 17:33:59 or did you mean that you want it to swallow -L plus its following argument? obviously that would require a slightly more complex regex although the idea is the same. Nov 16 17:34:17 pb__: I guess I need to be more specific - its a .la file and I want to remove any -Lfoo options from the dependency_libs = line :/ Nov 16 17:34:45 pb__: yes, its the complex bit that is evading me. Or the regexp isn't running and I have a bigger problem :/ Nov 16 17:36:37 RP: remember that sed uses basic regex syntax by default, you need "sed -E" to enable the extended things. Nov 16 17:36:42 that's what usually catches me out, heh Nov 16 17:37:25 sed -E 's/-L([^ ]+|( +[^ ]+))//g' seems to do the job for me. give that a go and see if it does what you need. Nov 16 17:38:53 if you're confident that it will always be "-Lfoo" and not "-L foo" then you can omit the bit after the | Nov 16 17:44:52 pb__: Thanks, I need to do this on a specific line which complicates things but you've given me some further things to think about :) Nov 16 17:45:43 pb__: sed claims it doesn't have a -E option here btw Nov 16 17:46:08 doh Nov 16 17:46:11 what sed is this? Nov 16 17:46:40 pb__: GNU sed 4.1.5 Nov 16 17:47:20 ah, I have 4.2.1 Nov 16 17:47:24 does "sed -r" work for you? Nov 16 17:47:49 pb__: That gives different interesting errors :} Nov 16 17:47:54 failing that you can obviously convert the regexp to basic form, it just means adding about a hundred backslashes :-} Nov 16 17:48:51 or use awk, which (iirc) always uses extended regexps Nov 16 17:50:46 perl Nov 16 17:50:49 :) Nov 16 17:51:14 pb__: I think I should have gone for awk. I always end up lost in backslashes :/ Nov 16 17:51:23 mwester: python? :) Nov 16 17:51:37 Well, yeah, I suppose... :D Nov 16 18:05:21 RP: hm, weird, sed -r does work fine for me. what errors were you getting? Nov 16 18:05:28 fwiw, the equivalent backslashed one appears to be: Nov 16 18:05:38 sed 's/-L\([^ ]\+\|\( \+[^ ]\+\)\)//g' Nov 16 18:05:53 pb__: no, it didn't like my own expression :) Nov 16 18:05:59 RP: ah, heh Nov 16 18:06:04 pb__: I have something ugly that works now :) Nov 16 18:06:10 jolly good Nov 16 18:06:21 * RP is trying to clean up more of gcc (and convert to the new staging world order) Nov 16 18:06:47 This is a challenge for cross packages Nov 16 18:06:58 (which also contain things that become target packages) Nov 16 18:08:23 yeah, that whole thing is a bit unsatisfactory Nov 16 18:08:34 it's quite tempting to just unbundle the libs and make them be built by a separate recipe Nov 16 18:09:01 pb__: I wondered about that, it'd be a lot nicer. I probably have it working now of course Nov 16 18:09:14 and cleaner than it was, at least to me Nov 16 18:09:20 * RP -> food Nov 16 18:09:25 good-o Nov 16 18:09:28 bon appetit Nov 16 18:35:32 mickey|sofa: good evening Nov 16 18:37:46 hey Nov 16 18:37:51 just tried to compile OE Nov 16 18:38:03 and discovered that it takes about 18G ! Nov 16 18:38:22 actually it wanted more, but I ran out of disk space Nov 16 18:38:48 how much disk space this so called embedded distro needs ? for build I mean Nov 16 18:39:17 * vasyapupkin considers to buy a new HDD in order to compile OE Nov 16 18:40:58 greetings pb__ Nov 16 18:41:33 RP: what about the bbextends roadmap? Nov 16 18:41:39 bbclassextends, that is Nov 16 18:51:42 vasyapupkin: yep, quite a lot, generally when I build mamona it takes around 20-25G Nov 16 18:54:15 rsalveti: it's a bit insane, don't you think ? Nov 16 18:54:35 my good old uClinux distro needs less than 1G Nov 16 18:55:47 vasyapupkin: yeah, but because with oe you end up building a lot of things Nov 16 18:56:10 source + compiled objects + installed binaries + metadata Nov 16 18:56:12 and so on Nov 16 19:01:21 mickey|sofa: What about it? Nov 16 19:02:36 RP: when is it going to be in? Nov 16 19:02:52 or is it already? Nov 16 19:03:28 mickey|sofa: We can start using it, its in Nov 16 19:03:51 ah, awesome, that's what i was asking for Nov 16 19:03:55 mickey|sofa: We've not converted many recipes yet Nov 16 19:04:04 *nod* i want to have a go at some, Nov 16 19:04:22 mickey|sofa: Have a go, if you have a problem let me know Nov 16 19:04:49 oki Nov 16 19:04:50 mickey|sofa: There could be some subtle fixes in class/conf files that I've missed Nov 16 19:05:00 but I should recognise those :) Nov 16 19:05:07 rsalveti: maybe... still, I feel that 20GB build size for a console image is too much. Nov 16 19:05:16 too much space for bugs, etc... Nov 16 19:05:32 I wish they went with uClinux Nov 16 19:05:42 beagleboard guys. I mean Nov 16 19:06:41 vasyapupkin: The reason is we have debug symbols enabled Nov 16 19:06:47 vasyapupkin: 20Gb taken by sources, toolchain, staging libraries and headers, resulting packages and, finally, rootfs images Nov 16 19:07:08 vasyapupkin: That makes things large but when things break it helps a lot Nov 16 19:07:29 if you just want to compile hello_world.c for arm, seems you are just needed any arm toolchain Nov 16 19:07:50 OE is more than just toolchain Nov 16 19:08:27 I know, I know. Still, I felt much more comfortable with uClinux when I could actually understand each and every step of the build process Nov 16 19:09:40 I don't need "just hello world", I need "just" a simple embedded linux distro for beagleboard Nov 16 19:09:51 apparently I'm stuck with OE Nov 16 19:10:01 take one from narcissus :) Nov 16 19:10:13 http://www.angstrom-distribution.org/narcissus/ Nov 16 19:10:48 interesting Nov 16 19:11:10 * vasyapupkin checks http://www.angstrom-distribution.org/narcissus/ Nov 16 19:11:54 not good enough, I want to be able to modify the sources Nov 16 19:12:14 that's what real embedded distro is all about Nov 16 19:12:31 I have to be able to modify all sources easily, kernel, bootloader, libc, everything Nov 16 19:12:33 well... then use OE and build one Nov 16 19:12:46 * vasyapupkin compiles OE in the background Nov 16 19:13:16 btw, there was manual in russian somewhere on openembedded.ru Nov 16 19:13:31 I have a few hours to kill till OE build process finishes, so I'm hanging around here complaining about OE Nov 16 19:13:34 norguthar is translating it slowly Nov 16 19:14:26 you mean there is some mysterious way to compile OE in 15 minutes described in russian version of the manual :) ? Nov 16 19:14:38 vasyapupkin: hehe :) Nov 16 19:14:54 just for your knowledge :) Nov 16 19:15:05 I will keep it in mind :) Nov 16 19:17:41 re Nov 16 19:17:44 wb Nov 16 19:19:26 i think we should silence the 'NOTE: Unpacking' Nov 16 19:19:44 it's the only non-task related note we still show Nov 16 19:20:18 florian: is wiki ready? :) Nov 16 19:20:43 Jay7: http://projects.linuxtogo.org/plugins/mediawiki/index.php?group_id=55 Nov 16 19:21:20 florian: btw, delete ML zaurus-test please :) Nov 16 19:22:35 Jay7: oh yes right Nov 16 19:22:39 Jay7: done Nov 16 19:22:42 10x Nov 16 19:22:48 last problem is dns :) Nov 16 19:22:57 but it is not critical Nov 16 19:23:50 btw Nov 16 19:24:14 florian: can userbase of wiki be in sync with main fusion forge userbase? Nov 16 19:24:28 just to skip one registration Nov 16 19:25:01 Jay7: I hope so... in fact I don't know if the plugin does this trick already. Nov 16 19:25:37 mickey|sofa: I totally disagree with all this "note silencing" thing Nov 16 19:25:47 mickey|sofa: Why didn't we just change the default log level? Nov 16 19:26:03 RP: isn't that what we did? Nov 16 19:26:09 mickey|sofa: "note" was always about giving an indication of progress Nov 16 19:26:20 mickey|sofa: no, every note appears to have been turned into a debug Nov 16 19:26:26 oh Nov 16 19:26:33 must confess i was not aware of that Nov 16 19:26:43 * RP is not happy about that :/ Nov 16 19:28:48 florian: seems no.. I can't login with my FF's account Nov 16 19:30:03 Jay7: we should check some other site using FF if it is meant to work :) Nov 16 19:30:14 hm.. Nov 16 19:30:19 did you know any? Nov 16 19:36:48 RP: at some point of time we could go over those notes again, put them back into the level we think is appropriate and silence the default level instead Nov 16 19:37:21 * kergoth thinks there are a lot of notes which supply useless information,, stuff the user is never going to care about, in general.. all the output everywhere should be audited Nov 16 19:38:13 * mwester agrees; OE seems to oppose the *nix principle that silence is goodness. Nov 16 19:42:33 it could be useful to provide a generic "progress" / update mechanism to indicate that, yes, we're still doing something (for tasks to use) Nov 16 19:44:42 well, it is nice to see that things are moving along :) Nov 16 19:45:23 that's why i was thinking about a generic progress indication method.. for when you want to let the user know that we're doing things, but the information you have isn't particularly useful to them Nov 16 19:45:25 * kergoth shrugs Nov 16 19:45:38 * kergoth should see about pushing the commit that makes bb.msg use the logging module under the hood Nov 16 19:46:24 OT: cars suck. just replaced the battery in mine, the battery cost $150 Nov 16 19:46:26 hrmph Nov 16 19:46:38 wow, that must be quite some battery Nov 16 19:46:50 The nice thing about a "generic" progress mechanism is that a wrapper tool can use that as well -- e.g. a GUI interface might use that to animate a progress indicator. Nov 16 19:47:04 mwester: yeah, exactly Nov 16 19:47:05 kergoth must have a BMW or something like that! Nov 16 19:47:21 it's a chrysler crossfire, but sadly almost everything under the hood is mercedes parts Nov 16 19:47:23 * mwester paid $75 for his last battery, in his Saturn. Nov 16 19:47:25 so pretty close to that, yes Nov 16 19:47:38 Crossfire is a really nice vehicle :) Nov 16 19:47:54 its fun... at times i wish i had a back seat, though Nov 16 19:47:56 Saturn was basic transport, and nothing more. Nov 16 19:48:55 yah, I just had to sell my two-seater Nov 16 19:49:04 rather a sad day. Nov 16 19:49:09 aw Nov 16 19:50:05 * Crofton mourns Nov 16 19:50:13 pb__, minivan now? Nov 16 19:50:16 bought an audi a3 instead which, as it turns out, I never get to drive because my wife uses it. doh. Nov 16 19:50:55 * RP is having the usual fun with winter in a 37 year old car Nov 16 19:51:07 so, not quite a minivan yet. :-} Nov 16 19:51:44 pb__: yet :) Nov 16 19:51:44 RP: heh, apparently I am supposed to be buying my wife an MGA to replace her car now. Nov 16 19:51:53 the minvan was great fun, you could load it up with stuff and go on road trips! Nov 16 19:52:22 driving the 2 seater in a minnesota winter was pretty amusing. i backed out of the driveway one day, and never got any further than that. thankfully, i had an olympic weight set to throw in the trunk :) Nov 16 19:52:27 pb__: I've heard tales of my Dad's various MGAs. My Mum knows how to put the spark plugs in the oven from one to get it going... Nov 16 19:52:53 * RP hopes spark plug design has improved in the intervening years Nov 16 19:54:04 yeah, I think spark plugs are better nowdays. also, you can get aftermarket electronic ignition which helps quite a bit. Nov 16 19:54:29 pb__: Points aren't that bad if you set them up right Nov 16 19:54:45 * Crofton knows RP is crazy now Nov 16 19:54:49 pb__: The MGB's downfall is twin 6V batteries so twice as many terminals to corrode Nov 16 19:54:56 the plug wires were problematic on the older vehicles as well -- I recall drying many sets of plug wires in Canadian winters... Nov 16 19:55:12 kergoth: heh, I had that problem a few years ago when it snowed here. had to drive my brother to telford to meet his girlfriend, who turned out to live at the bottom of a valley. getting down there was fine, if a little bit exciting at times, but getting back up again was virtually impossible. Nov 16 19:55:24 mwester: I have silicon leads (which are a different set of problems) Nov 16 19:55:48 had to have his girlfriend's dad sit on the lid of the trunk to provide enough weight on the rear wheels Nov 16 19:56:02 pb__, kergoth: Try snow on a motorcycle. It means you'll never complain about the car again :) Nov 16 19:56:09 pb__, did you also have to reverse up the hill? Nov 16 19:56:10 hehe Nov 16 19:56:15 Of course I now have the ideal motorcycle for this :) Nov 16 19:57:09 Of course, many of us can work from home now, so not so many troubles in bad weather anymore. :) Nov 16 19:57:20 mwester: there is that too Nov 16 19:57:37 mwester: the sad thing is I kind of miss the crazy bits... Nov 16 19:58:44 The MGB has van tyres running at low pressures on it. All bets are off when it rains, let alone snows Nov 16 20:38:11 Any good mips toolchain versions? Nov 16 20:49:03 good evening, how do you do when you are compiling the same kernel several times with slight changes on .config? Nov 16 20:49:24 i'm using bitbake linux-rp -c menuconfig, and then bitbake linux-rp Nov 16 20:49:46 but after compiling the first time, it wont ever compile again, unless i do a -c clean... is there another way? Nov 16 20:58:45 bitbake never runs tasks it has already run, generally. you can use -f to override Nov 16 20:58:56 bitbake -c compile -f linux-rp; bitbake linux-rp Nov 16 20:58:59 or what have you Nov 16 21:11:33 kergoth, thank you! :-) i need the -f switch then. i used the previous process i mentioned before for this compilation already, but next one will be easier with -f. :-) Nov 16 21:12:59 03Mike Westerhof  07org.openembedded.dev * r10e531444f 10openembedded.git/recipes/owfs/owfs_2.7p24.bb: Nov 16 21:12:59 owfs: disable swig to prevent doomed attempt to build a perl module. Nov 16 21:12:59 Signed-off-by: Dirk Opfer Nov 16 21:16:41 03Mike Westerhof  07org.openembedded.dev * r3d48522d84 10openembedded.git/recipes/meta/slugos-packages.bb: SlugOS: add owfs to the package feeds Nov 16 21:44:12 Hi, I'm trying to get the Omnia touchscreen to work properly with X... I'm having some issues Nov 16 21:44:53 Whenever I press something it returns 3 or more screentaps Nov 16 21:45:21 Basically, it thinks I'm pressing down on the mouse key or something Nov 16 21:45:31 No such issues in OPIE or ts_test Nov 16 21:45:55 I've tried playing around with ReportDelay and UntouchDelay, but nothing... Nov 16 21:46:58 Is there any way I can have X not to auto-repeat the screentaps like that? Nov 16 22:02:29 Sometimes the state of the metadata makes me want to cry :/ Nov 16 22:02:47 update-rc.d - installs into native and target in one go :/ Nov 16 22:02:55 now I can see why but... :/ Nov 16 22:03:02 B_Lizzard: EVDEV or TSLib for the X driver? And what sort of event flooding are you seeing. Nov 16 22:03:18 tslib Nov 16 22:03:25 RP: lots of things in life have that effect on people ;-) Nov 16 22:03:32 B_Lizzard: version? Nov 16 22:03:37 Under X, whenever I press a button it reports multiple keypresses Nov 16 22:03:39 Wait a sec... Nov 16 22:04:01 1.0 Nov 16 22:04:12 And xf86-input-tslib version 0.0.6 Nov 16 22:04:14 B_Lizzard: that is the X TSLib driver, not TSLib its self that I am interested in, 0.5 or 0.6? Nov 16 22:04:21 Ahhh, okies Nov 16 22:05:03 It basically seems like autorepeat has no delay Nov 16 22:07:58 B_Lizzard: try something like http://pastebin.com/d6d572c5d I have seen that before and used to carry that patch (a colleague of mine knocked it up) on the 0.0.5 xf86-input-tslib in the my OpenPandora overlay to fix a similar issue I see on that touchscreen (not sure that 0.0.6 shows the issue for me however). Nov 16 22:08:33 Thanks a lot, dude. Nov 16 22:08:42 I'll apply the patch and try it out. Nov 16 22:09:46 B_Lizzard: use something like SRC_URI_append_omap3-pandora = "file://dont_flood_events.diff;patch=1" so you don't upset other users of the recipe, if you find it fixes things let me know as that would be 2 platforms that seem to exhibit the snag. Nov 16 22:10:11 is there anyone who knows about building angstrom with external toolchain? Nov 16 22:11:00 tilarids_: done it a long time ago. Don't recall it was that hard to do. Nov 16 22:13:19 DJWillis, can you help me with some links or manuals? I'm new to OE and angstrom and can't build it in usual way, so I am looking for the other ways. Nov 16 22:13:50 I've tried to use user-manual, but it tries to build gcc-cross anyway Nov 16 22:14:15 tilarids_: any reason you can't build it in the normal way? Nov 16 22:15:03 DJWillis, it says that '/home/sergey/dev/OE/openembedded/recipes/gcc/gcc-cross-initial_4.2.4.bb' failed . Nov 16 22:15:11 And in logs I can see only make: *** No rule to make target `install'. Stop. Nov 16 22:16:15 I was asking for help in #angstrom and #oe recently, with no respond Nov 16 22:16:55 tilarids_: that sort of implies something is borked, I am not sure an external toolchain is the way to get up and running ;-). So you have done a bitbake -cclean gcc-cross-initial && bitbake gcc-cross-initial ? Are you on .dev or stable? Nov 16 22:18:46 DJWillis, I'm on stable/2009 branch. I was trying to make bitbake base-image ; bitbake console-image ; bitbake x11-image (as said in angstrom build manual) Nov 16 22:18:58 will give a try to bitbake -cclean gcc-cross-initial && bitbake gcc-cross-initial now, thanks Nov 16 22:20:06 tilarids_: and of course the usual git pull ;-) Nov 16 22:20:52 tilarids_: -cclean and then a rebuild can be a good check just to get rid of guff sometimes, esp. if you run into the odd timing issues that sometimes crop up. Nov 16 22:25:04 DJWillis, hmm, as I can see that helped! Thank you very much! btw, bitbake -cclean world should clean all the packages, right? Nov 16 22:26:17 tilarids_: if your going down that road you may as well blow away your tmp folder ;-), world is not really the keyword of the novice, lots of time and a fair bit of pain (esp. if you ever use it on the .dev tree). Nov 16 22:27:20 tilarids_: assuming your keeping your source downloads out of your build tmp of course, if you blow away your sources that's going to add a lot to the next build ;). Nov 16 22:29:49 DJWillis, thanks a lot! Nov 16 22:50:27 DJWillis: The patch doesn't seem to change anything. Nov 16 22:50:48 I was thinking of some kind of delay between the first keypress and the subsequent ones Nov 16 23:14:32 Does openmoko-image depend on a non-existant package? Nov 16 23:14:33 matchbox-panel-2-applets Nov 17 00:00:44 kergoth: Does prefixing our libtool actually gain us much? Nov 17 00:00:54 (the script name) Nov 17 00:01:09 kergoth: Just looking at the mess that is libtool-cross :/ Nov 17 00:01:29 RP: hi, QA's Revenge this evening :) Nov 17 00:01:56 all fine Nov 17 00:02:05 ant__: great Nov 17 00:02:42 thx again for the quick fix Nov 17 00:03:17 good evening, as some of you know i've been struggle to build linux-rp kernel for my c700. solution is to compile it without cpufreq governors.. is there a good way to share this so that others can skip these troubles i had? Nov 17 00:04:18 kergoth: and in answer to my question, yes there is a point :/ Nov 17 00:04:30 ant__: np Nov 17 00:04:41 qsp: patch to the mailing list? Nov 17 00:06:07 dont know how to make a patch yet... i[m kinda new on all this.. anyway, i successfully compiled and booted the 2.6.26 kernel, with defconfig, only change that needs to be done is disabling the governors section in menuconfig Nov 17 00:06:23 i'll research how to make a patch then :-) Nov 17 00:06:39 RP, not sure what may have been fixed, but it seems that as a side effect of the staging changes or other recent changes, a number of other obscure packages are building for SlugOS now. :) Nov 17 00:08:27 mwester: excellent :) Nov 17 00:08:46 mwester: The plan is to simplify the metadata a bit, perhaps that is helping Nov 17 00:09:18 RP: qsp's issue on c7x0..very strange, isnt? Nov 17 00:09:38 RP: same kernel is ok on C860 Nov 17 00:10:09 PXA specifics? Nov 17 00:10:47 arm CPU is the only real diif iirc Nov 17 00:11:29 btw, the error with 'stock' kernel-rp is that it hangs after uncompressing the kernel, before any other message. Nov 17 00:11:39 ant__: yes, pxa250 and 255 were quite different Nov 17 00:12:00 possibly that code is not safe... Nov 17 00:12:12 we'll ask pb :) Nov 17 00:12:22 * qsp thinks he needs a soldering iron to instal a pxa255 :-D Nov 17 00:12:45 qsp: I would be interested in exchanging a C860 with your C700 Nov 17 00:12:52 I have two Nov 17 00:13:14 :-) i suppose yours more valuable :-) Nov 17 00:13:25 doh Nov 17 00:13:42 6 yrs old hardware... Nov 17 00:13:58 nice petty :) Nov 17 00:14:14 why not, i could use the extra ram, really Nov 17 00:14:32 :-) Nov 17 00:14:48 I bought the second on e-bay to rip it off, in case I 'dneed have spare part for the first... Nov 17 00:15:16 well, this one was ebay'd as well :-) Nov 17 00:15:44 eh..that perhaps explains 'won't boot' ...:=) Nov 17 00:16:20 well, it doesnt have nand errors, if you remember, when we were debugging kexec :-) Nov 17 00:17:19 both mines have 2 bad sectors, ex works Nov 17 02:11:35 yo Nov 17 02:12:07 anyone gotten OE on a HP Jornada, like an old one? And how the heck can you do it without ripping out the internal flash thing? **** ENDING LOGGING AT Tue Nov 17 02:59:56 2009