**** BEGIN LOGGING AT Wed Aug 30 02:59:57 2006 Aug 30 05:23:12 * v8jlene is back Aug 30 06:08:21 hi Aug 30 06:08:36 good morning Aug 30 06:08:57 I finally got it to build "something", heh Aug 30 06:09:12 it boots too! Aug 30 06:10:08 great Aug 30 06:10:15 :P Aug 30 06:10:26 now to do the customizing Aug 30 06:10:37 T0mW: can you write kind of success story when you will get all done? Aug 30 06:11:08 ha! If I do that in the wiki, you guys will change stuff again and break it! Aug 30 06:11:21 then I'll get blamed for being a dummie Aug 30 06:11:25 ;) Aug 30 06:11:44 T0mW: at oedem we plan to discuss how to 'fix' it Aug 30 06:11:47 surprised you had that much difficulty with it. please feel free to modify the wiki if anything could be clarified to make it easier in the future Aug 30 06:12:12 hrw: should be a good place to talk about some of the long term improvements that can be made Aug 30 06:12:31 kergoth: feel free to add any ideas to oedem page Aug 30 06:13:01 I think that the major two things that I did was that I didn't understand that *-image was the thing to build, also, that I tried to build stuff that was too complex and it broke (e.g. gpe-image). Aug 30 06:14:26 whatever, the goal for now is to get a serial console running so I can login to start fixing stuff that isn't working in my image. Aug 30 06:15:16 i like starting off a fresh config with a bootstrap-image Aug 30 06:15:20 hard to go wrong there Aug 30 06:16:46 that definately worked for me. so I'm going to stick with that for now. There are some things I need to learn, for example, I'm still doing dev on the kernel sources for this board. Aug 30 06:16:55 hmm... 1.8G athlonxp and kde apps are slow.. Aug 30 06:17:12 what distro? Aug 30 06:17:42 T0mW: oe is great for building and maintaining distros for the target, unfortunately its weak on actually actively developing code for the packages being included. doesnt include a lot of tools fostering a development environment Aug 30 06:17:49 thats one of the main things that needs improving long term, imo Aug 30 06:18:43 T0mW: debian Aug 30 06:18:56 at this point, I'm wondering how I'm going to make changes to sources. For example, while tuning the cchanges to the kernel source in the /tmp tree, how do I recompile the kernel? Aug 30 06:20:40 T0mW: bitbake -f -ccompile virtual/kernel Aug 30 06:20:50 hrw: dunno, mandriva had some app that would be installed in the kde bundle, it would slow the desktop down. Sort of like a fast-file-find thingy Aug 30 06:21:35 hrw: thanks, I was going to try "rebuild", but that probably would blow the source dir away and redo it. Aug 30 06:22:01 exactly Aug 30 06:23:17 I will repeat myself... I want tool which show real memory usage of apps... Aug 30 06:23:34 total used free shared buffers cached Aug 30 06:23:34 Mem: 906416 852084 54332 0 205120 313768 Aug 30 06:24:15 I too often feel like having 768/1024M of ram is not enough in this days.. Aug 30 06:24:52 kde desktop, firefox, amarok, kmail, few terminals in konsole, kopete = 750M used... Aug 30 06:29:23 is there like a skeleton (skel.tar.gz)? Aug 30 06:30:44 or, does each package / task define vars such as "SERIAL_CONSOLE" to affect behavior of how the rootfs gets constructed (other than actual package name callout in a recipe)? Aug 30 06:31:48 T0mW: each .bb emits binary packages, which are then installed into the image Aug 30 06:31:56 seems silly to edit the rootfs image only to have those disappear the next time I do a bitbake build Aug 30 06:32:00 there are some fixups and such done outside of package management, but not much, at least generally Aug 30 06:32:15 most distros (if not all) use ipkg for package management on the device Aug 30 06:32:27 you can certainly emit tarballs, but i dont know if oe has an image creator that uses them, yet Aug 30 06:32:57 well, for example, /etc/securetty is missing from this image. Aug 30 06:33:22 kergoth: does not have Aug 30 06:33:46 time to rebuild 2.6.17 for tosa/poodle/collie Aug 30 06:33:53 under buildroot, I would let the rootfs be built, then explode a tarball across the rootfs image tree to fixup stuff. Aug 30 06:34:11 T0mW: you can create tarball with rootfs Aug 30 06:34:25 you could just create a .bb that unpacks that tarball and installs it, then put that ipk into the image Aug 30 06:34:31 lots of options Aug 30 06:35:13 * T0mW will have to learn what ipackage does Aug 30 06:35:50 .ipk files are nearly the exact same format as .deb Aug 30 06:35:52 is ipackage an integral part of constructing the final rootfs image tree? Aug 30 06:37:21 IOW, you build into staging having an ipk file created in the process, then some other part of bitbake says "ok, let's construct a rootfs image, time to run ipackage-native to construct the tree, then tar it, ext2 it or jffs2, ..." Aug 30 06:39:34 yep Aug 30 06:39:51 pretty much, except that its the package task that populates the stuff for the ipk creation, not staging, at least as things stand today :) Aug 30 06:39:59 ok, I was wondering how pervasive this ipkg thing was, thanks. Aug 30 06:40:20 bitbake use ipkg to install packages into tmp/rootfs/ and then tarballs, ext2, jfs2, cramfs, etc are created from it Aug 30 06:41:07 T0mW: Take a look at the rootfs_ipkg and image_ipkg in the reference section of the user manual. It gives the low level details on how the root filesystem and images are created. Aug 30 06:41:27 ok Aug 30 06:41:48 * T0mW thinks man / info / doc pages are a "good thing" Aug 30 06:42:16 wow, the user manual actually has that stuff now? Aug 30 06:42:23 impressive Aug 30 06:42:29 ~lart kergoth Aug 30 06:42:30 * ibot blasts kergoth with a huge firehose then strangles kergoth with it Aug 30 06:42:41 T0mW: I haven't completed the higher level documentation stuff yet, but I'm slowly working towards having all that in there. Aug 30 06:42:48 kergoth: iirc v8jlene merged own OE docs with official one Aug 30 06:42:52 nice Aug 30 06:42:56 v8jlene: nice job then :) Aug 30 06:43:52 Some of my stuff.. I have loads more partly done. The stuff on writing .bb recipes is probably the most useful, but I want the reference section there so I can refer to it for detailed explanations of stuff... Aug 30 06:47:19 yo Aug 30 06:48:51 kergoth: I'll going to check in my change to classes/patch.bbclass to create the patches dir for quilt if you have no complaints (as sent to mailing list). You were the one who made the changes there that broke it I think. Aug 30 06:52:37 FSCK.. Aug 30 06:52:42 koen: /dev/sda2 60G 60G 20K 100% /home Aug 30 06:54:08 I had a bitbake build loop on something and fill my root partition a week or so ago.. took me ages to figure out what the hell was going on ;) Aug 30 06:54:42 v8jlene: ah, feel free. quilt needs an option to disable the search up the tree, imo Aug 30 06:54:54 were there any other issues with the patch changes? Aug 30 06:55:53 kergoth: lack of 'I do not want to be dropped into shell'? Aug 30 06:55:54 kergoth: With my patch included I rebuilt from script for sh4 and for epia last night with no issues. Aug 30 06:56:27 from script? from scratch ... Aug 30 06:56:43 hrw: as i said on irc earlier today, i'm gonna check in a no-op resolver class to be able to do so Aug 30 06:56:47 v8jlene: cool Aug 30 06:57:49 hrw: bitbake foo -c -f doesn't work Aug 30 06:57:59 hrw: -f only works when used together with -b Aug 30 06:58:10 koen: ok Aug 30 06:58:17 I think it's a bug Aug 30 06:58:18 I'm working on a router distro though, so I don't do and/or build any UI related stuff. Aug 30 07:02:58 xsystem: ERROR: fork failed before execution: `wget --passive-ftp Aug 30 07:03:00 hmm.. Aug 30 07:03:10 simple 'ipkg update' fails.. Aug 30 07:05:19 'fork failed' means 'yarrrrrr, no more RAM' most of the times Aug 30 07:06:27 03lenehan 07org.oe.dev * r4b6ef124... 10/ (1 classes/patch.bbclass): (log message trimmed) Aug 30 07:06:27 classes/patch.bbclass: Create a "patches" directory when initialising Aug 30 07:06:27 the quilt patcher class. Without this quilt will search for a patches Aug 30 07:06:27 directory - starting from the current directory up to the root Aug 30 07:06:27 directory. If it finds an existing patches directory it will use it Aug 30 07:06:27 for its patches. This causes all sorts of problems since it is not Aug 30 07:06:31 where the patches are expected to be. Prior to the recent patcher Aug 30 07:06:58 koen: I plan to do some tests today on tosa (remotely) Aug 30 07:07:50 hrw: Xephyr comes in quite handy when doing that Aug 30 07:08:06 ok. 2.3G free on ewi now. can build stuff Aug 30 07:08:34 koen: it does not emulate machine running OZ rather Aug 30 07:08:55 I meant starting Xephyr on the Z and xforward it to your workstation Aug 30 07:09:08 koen: I need shell only Aug 30 07:09:36 and with 64M I prefer to do run only minimal set of soft Aug 30 07:11:47 ~kill ipkg for it slowness Aug 30 07:11:50 * ibot shoots a hyper-charged proton gun at ipkg for it slowness Aug 30 07:13:28 ipkg list *links* = 1m2s egrep "^Package:.*links.*" /var/lib/ipkg/*" = 19s Aug 30 07:14:42 time to go to work - bbl Aug 30 07:18:26 03clarson 07org.oe.dev * raa23e3ef... 10/ (1 classes/patch.bbclass): Aug 30 07:18:26 patch.bbclass: Aug 30 07:18:26 * Add NOOPResolver class, which simply passes the patch failure on up, not Aug 30 07:18:26 doing any actual patch resolution. Set PATCHRESOLVE = "noop" to make use of Aug 30 07:18:26 it. Most useful for unattended builds. Aug 30 07:21:31 udev sux Aug 30 07:21:43 * kergoth disagrees Aug 30 07:21:58 udev sucks differently as devfs Aug 30 07:22:13 sucks how? Aug 30 07:22:15 my /dev/ttySAC0 is not there. I had to mknod it in the init scripts Aug 30 07:22:24 i've yet to have an issue with it. the flexibility it gives you is incredible Aug 30 07:22:32 sounds like you should learn about how it works. Aug 30 07:22:52 * T0mW already knows how it doesn't work, so... Aug 30 07:23:05 clearly you don't know much about it at all. Aug 30 07:23:11 i suggest you do so before commenting further Aug 30 07:24:35 I'm looking the rule set over now. Aug 30 07:27:29 03koen 07org.oe.dev * r7ff3301c... 10/ (1 packages/xorg-proto/calibrateproto_git.bb): calibrateproto: add git version Aug 30 07:37:44 03koen 07org.oe.dev * r422a2eb1... 10/ (1 packages/xorg-proto/calibrateproto_git.bb): calibrateproto: add a PV Aug 30 07:37:54 03koen 07org.oe.dev * r1c574776... 10/ (1 packages/xorg-lib/libxcalibrate_git.bb): libxcalibrate: add git version Aug 30 07:42:11 udev's startup time sucks but it is useful. /dev/input/touchscreen0 being my favourite example atm :) Aug 30 07:42:24 morning all btw Aug 30 07:42:44 i love using it for persistant device naming for all sorts of things, based on all sorts of sysfs attributes :) Aug 30 07:43:39 with the right mount.blacklist startup is quite fast Aug 30 07:44:00 koen: Is it mount.blacklist slowing things down? Aug 30 07:44:21 I'd love to have a udev rule that automatically swaps on detected swapdrives Aug 30 07:44:40 RP: if you don't blacklist ramdisks and stuff, startup is quite slow Aug 30 07:45:16 koen: We should switch to whitelisting Aug 30 07:45:28 It should always have been a whitelist... Aug 30 07:46:13 swapdrives would be beyond udevs skills - there's no sysfs attribute for that Aug 30 07:48:25 RP: can't a rule check partition type? Aug 30 07:48:44 koen: you do not have to set partition type to swap to have swap onit Aug 30 07:49:04 koen: and solaris partitions has same type as linux swap... Aug 30 07:49:07 no, but I assume most people set it to 'linux swap' :) Aug 30 07:49:35 koen: It it was a syfs attribute, it could check it. I don't know if the partition code knows about sysfs yet Aug 30 07:50:23 there is no type in /sys/block/hda/hda1/ Aug 30 07:57:11 kergoth: http://rafb.net/paste/results/KLMHWO50.html Aug 30 07:58:14 koen: yes, that looks familiar.. consdiering it looks exactly like the post on the oe mailing list Aug 30 07:58:20 :P Aug 30 07:58:34 this has a twist Aug 30 07:58:57 of lemon? Aug 30 07:58:59 this .bb doesn't have a SRC_URI :) Aug 30 08:01:16 morning all Aug 30 08:01:24 hi dirk Aug 30 08:02:22 hey Marcin Aug 30 08:03:16 hey Dirk Aug 30 08:04:25 kergoth: and more to the point, this recipe has no ${S} Aug 30 08:05:15 hi, i have a problem creating the bootstrap-image for an arm-architecture. While linking the arm-linux-uclibc it crashs with: bla/bla/bla/ld: crt1.o: No such file: No such file or directory. does anyone knows this problem? Aug 30 08:05:16 hey Koen Aug 30 08:06:58 03koen 07org.oe.dev * r9f5352e4... 10/ (3 files in 3 dirs): xserver-kdrive git: adjust for new calibrate stuff, package Xsdl, enable build of epson server Aug 30 08:11:30 good morning Aug 30 08:13:13 * v8jlene is away Aug 30 08:20:53 03clarson 07org.oe.dev * r16c62009... 10/ (1 classes/patch.bbclass): Aug 30 08:20:53 patch.bbclass: Aug 30 08:20:53 * switch os.mkdir to os.makedirs. Aug 30 08:20:53 * pass on all errors from QuiltTree.Clean(), as it can fail in ways that do Aug 30 08:20:53 not need to be reported to the user, and a failure will end up being seen Aug 30 08:20:54 again during the Import/Push of the patches. Aug 30 08:26:47 morning Aug 30 08:27:35 hey XorA Aug 30 08:29:44 argh... tosa does not connect via usbnet with my work desktop.. Aug 30 08:33:17 so no usage of tosa today for development testing Aug 30 08:35:11 hrw|work: even after rmmod g-ether, modprobe g-ether? Aug 30 08:41:13 XorA: even Aug 30 08:41:55 psokolovsky_: xserver-kdrive_git is now up to date with the calibrate and tslib stuff Aug 30 08:43:55 koen: Hi Koen, great work, thanks! And h4000 is now able to boot GPE image an dit works somehow. So, feel free to add it to generated snapshots next time ;-) . More changes pending, of course. Aug 30 08:48:41 hrw|work: strange, I checked this month ago and it worked. I don't use usb here Aug 30 08:50:02 do13_: I will check it other day at home Aug 30 08:50:20 ~seen mikearthur Aug 30 08:50:33 ~summon mikearthur and his TS patch Aug 30 08:50:33 mikearthur was last seen on IRC in channel #oe, 16h 11m 42s ago, saying: 'lrg: slaters'. Aug 30 08:50:36 apt takes out 20 clean, identical-looking phones, some extra hands, and pretends to be a telemarketer for a large corporation, so he gets delivered a phonelist containing mikearthur and his TS patch's coordinates Aug 30 08:51:15 poushag: I am subscribed to my iCal calendars with gpe-calendar Aug 30 08:51:26 poushag: I can even update them via webdav Aug 30 08:52:26 psokolovsky_: does the h4000 kernel compile in cvs head Aug 30 08:52:44 ? Aug 30 08:54:18 koen: ah, yes, thanks for reminding. it should use 2.6.16-hh5 for now. I'll add recipe for that. Aug 30 08:57:29 psokolovsky_: and packages/linux/handhelds-pxa-2.6/ is missing a defconfig for h4000 by the looks of it Aug 30 08:59:33 koen: Ok, so I was a bit quick at it. Let's say: h4000 should be in shape within next few days. ;-) It's just I'm out of monthly traffic, and it's a bit pain to update. Should be ok after Sep 1st ;-) Aug 30 09:08:52 hi Crofton Aug 30 09:09:09 the build succeeded last night. building 'bootstrap-image' now. :) Aug 30 09:13:37 23:20 < Crofton> need to get it booting ok tomorrow, bigdog will want to do the same thing very Aug 30 09:13:37 soon I expect Aug 30 09:13:39 :) Aug 30 09:13:54 aww <3 Aug 30 09:19:32 03pfalcon 07org.oe.dev * r1a0fa553... 10/ (1 packages/alsa/alsa-utils_1.0.11.bb): alda-utils: speaker-test doesn't really work without alsa-conf. Aug 30 09:19:49 ~lart vim for not applyable patches Aug 30 09:19:49 * ibot shoves a crumpet down vim's throat, happy now?! Huh? Want some JAM with that? for not applyable patches Aug 30 09:23:49 hrm. that's odd. Aug 30 09:23:51 ERROR: Exception:exceptions.OSError Message:[Errno 2] No such file or directory: '/home/bigd0g/stuff/tmp/work/ipkg-collateral-1.0-r3/ipkg-collateral-1.0/patches' Aug 30 09:24:38 03pfalcon 07org.oe.dev * rdef4216e... 10/ (3 files in 3 dirs): handhelds-pxa-2.6: Add defconfig for h4000. Aug 30 09:24:52 vim_7.0.bb will beat even zaurus kernels/SRC_URI Aug 30 09:26:09 hehe Aug 30 09:27:03 ~70 patches Aug 30 09:27:15 Is that a challenge? Aug 30 09:27:23 ~lart bitbake do_patch anyway Aug 30 09:27:23 * ibot lowers bitbake do_patch anyway's priority Aug 30 09:27:59 I can add patches with quilt import but have problems with bitbaker... Aug 30 09:28:52 I will probably prepare summarized patch... Aug 30 09:28:56 bigd0g: update OE to get rid of that message, kergoth should have fixed is a few minutes ago Aug 30 09:29:16 koen: thanks. doing it now... figured that might be it Aug 30 09:33:33 03pfalcon 07org.oe.dev * r82565cfc... 10/ (1 packages/linux/handhelds-pxa-2.6_2.6.16-hh5.bb): Aug 30 09:33:33 handhelds-pxa-2.6_2.6.16-hh5: Recipe for the latest HH.org 2.6.16 tag. Aug 30 09:33:33 Based on handhelds-pxa-2.6_2.6.15-hh0.bb. Aug 30 09:42:47 hrw|work: vim's packaging is a pain in the ass, mainly because you cant split out the different pieces of functionality from the binary. have to have seperate fscking builds Aug 30 09:43:12 :\ Aug 30 09:43:20 ~lart bram Aug 30 09:43:21 * ibot frags bram with his BFG9000 Aug 30 09:43:53 it is Aug 30 09:44:07 ~lart combinediff for not be able to combine 76 patches.. Aug 30 09:44:07 * ibot accelerates a free AOL cd to 50,000 rpm and lets combinediff feel it for not be able to combine 76 patches.. Aug 30 09:44:13 hrw|work: i have a script for that Aug 30 09:44:14 its ugly Aug 30 09:44:30 but its called multicombinediff, it calls combinediff repeatedly, gathering them up into one big patch :P Aug 30 09:44:39 give me... Aug 30 09:44:59 * kergoth plugs in the usb stick that has all that random stuff on it Aug 30 09:44:59 I gave up on adding all 7.0.001 -> 7.0.076 patches into SRC_URI Aug 30 09:45:13 bitbake is not able to handle those patches Aug 30 09:46:05 ~lart context diffs Aug 30 09:46:06 * ibot whips context diffs with a wet and grimy noodle just because Aug 30 09:47:56 Hi Aug 30 09:50:55 kergoth: does patchutils can change diff from p0 to p1? Aug 30 09:51:15 when I cross compile qtopia core with OpenEmbedded I get the error libQtCore_debug.so undefined reference to __dlsym, __dlclose as well as __dlopen. Is this the problem from qtopia core? With another cross compiler I could compile it without problem, but I would like to use OpenEmbedded, because it has everything what my OpenZaurus need. Aug 30 09:51:37 PolicyBased: qtopiacore - which ver? Aug 30 09:51:41 hrw|work: hmmm not that i know of. itd have to add another level to all of the paths.. hmm Aug 30 09:51:57 4.1.4 Aug 30 09:52:29 PolicyBased: add -ldl into LDFLAGS Aug 30 09:52:48 kergoth: I regenerated it with quilt Aug 30 09:52:53 thanks Aug 30 09:53:05 quilt snapshot;quilt push -a;quilt diff --snapshot -p1>file.diff Aug 30 09:54:03 ah nice Aug 30 09:55:33 I would lart bitbake again... Aug 30 09:57:07 if I add patch to SRC_URI then it does not apply.. but apply by hand.. Aug 30 09:57:24 weird Aug 30 09:59:03 quilt by hand also apply it.. I have to dig for patches/ dir... Aug 30 10:00:24 arghh again Aug 30 10:05:43 yay! bootstrap-image built. Aug 30 10:11:35 03koen 07org.oe.dev * r5ef29a16... 10/ (1 packages/linux/handhelds-pxa-2.6_2.6.16-hh5.bb): hh-pxa-2.6 2.6.16 hh5: unbreak do_configure for EABI, add proper COMPATIBLE_MACHINES and do a version check for good measure Aug 30 10:12:04 * remove patches that won't apply Aug 30 10:12:13 * koen always forgets stuff in the Changelog Aug 30 10:12:50 hrm... Aug 30 10:13:03 linux-omap1, bootstrap-image, devshell ... anything else i could need? Aug 30 10:13:14 hi, i have a problem creating the bootstrap-image for an arm-architecture. While linking the arm-linux-uclibc it crashs with: bla/bla/bla/ld: crt1.o: No such file: No such file or directory. does anyone knows this problem? Aug 30 10:28:28 RP: koen hrw|work should I patch Mardy's Xw100 driver into out xserver-kdrive .bb? Aug 30 10:28:51 if it works :) Aug 30 10:29:09 koen: Well I have a screen, but no-one will ever test if its not avalable Aug 30 10:29:23 I had to make some changes for X11R7.1 Aug 30 10:29:30 some auto* stuff changed Aug 30 10:29:43 XorA: if it works then go ahead Aug 30 10:30:09 I mostly use tosa. c7x0 usually as text terminal Aug 30 10:34:19 ok, I shall test for rest of morning and incorporate this afternoon Aug 30 10:34:45 jadefalcon: you update all your local repositories and try again yet? Aug 30 10:48:29 bigd0g: ok i udating org.openembedded.dev and try again... revision is now 5ef29a16dd96c72ef0f33bd0f53e06381ba285dd Aug 30 10:49:18 why is there a symlink in /usr/local/arm/oe/arm-inux/lib/libdl.so which points to ../../lib/libdl.so.2 which doesn't exist? Doesn't it have to point to /usr/local/arm/oe/arm-linux/lib/libdl.so.2? Or doesn't this matter? Aug 30 10:49:40 jadefalcon: great. run again and see if you get any errors. Aug 30 10:51:39 bigd0g: same result, is there a channel where i can post the complete error or should i do that on the mailinglist? Aug 30 10:52:22 jadefalcon: post the complete error in a pastebin and you can paste the url in the channel Aug 30 10:56:43 gm bigd0g Aug 30 10:57:04 ~pastebin Aug 30 10:57:12 methinks pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.com/ (BROKEN AND SUCKING NUTS), or http://pastebin.ca, or http://channels.debian.net/paste, or for #oe use http://oe.pastebin.com, or http://bzflag.pastebin.ca/, or http://paste.lisp.org/ for the lisp/scheme nuts, or http://paste-it.net Aug 30 10:57:32 hiya Crofton Aug 30 10:57:45 hear the news? successful build Aug 30 10:58:06 linux-omap1, bootstrap-image, and devshell ... all built and compiled successfully Aug 30 11:00:22 cool Aug 30 11:00:34 I have booted using the linux-omap1 .... git Aug 30 11:00:49 and booted the filesystem, but it is hanign starting udev :( Aug 30 11:03:35 03koen 07org.oe.dev * r708df891... 10/ (5 files in 4 dirs): Aug 30 11:03:35 horizon: add cvs version of this cool sketch toy Aug 30 11:03:35 * still has /media/mmc1/horizon hardcoded Aug 30 11:06:00 bigd0g: http://pastebin.com/779611 Aug 30 11:07:19 jadefalcon: i don't know if i can necessarily help you, but i'll take a look Aug 30 11:07:42 jadefalcon: i've only started working on an OE project of my own these past couple of weeks Aug 30 11:08:12 bigd0g: thanks, i started yesterday :-) Aug 30 11:10:30 jadefalcon: check the /home/azuber/arm/openembedded/stuff/tmp/work/arm-linux-uclibc/uclibc-0.9.28-r6/temp/log.do_install.16440 text file Aug 30 11:10:40 it may give you more hints as to what's going on Aug 30 11:10:55 your build of uclibc is what's failing... so the update did appear to fix your original error Aug 30 11:17:46 ok, i'll try a rebuild. deleting the tmp-dir and running again Aug 30 11:19:34 ok. if it doesn't work on the clean build, then a configuration change probably has to be made somewhere Aug 30 11:21:03 ok, thanks Aug 30 11:29:25 anyone have any thoughts why udev might be haning on my OSK? Aug 30 11:29:46 since I don't know much, bvious suggestions are welcome Aug 30 11:32:56 hrw|work: koen ah that xserver upsites gtk+ mightilly :-) Aug 30 11:34:43 Crofton: I added a couple of lines to the end of /etc/init.d/finish (e.g. ls -F /dev) to see what was going on with udev creating the ttySAC0, actually didn't make that node, just a "s3c2410-serial0" instead. Aug 30 11:35:20 Crofton: udev might not be doing what you expect, try the ls in the finish script Aug 30 11:35:35 ok Aug 30 11:35:47 Crofton: worked for me! :) Aug 30 11:35:52 I am not at the board right now, just trying to get some ideas Aug 30 11:37:18 Crofton: you might sprinkle some "echo" statements into the various starup scripts (runlevel scripts) to see if you are actually hung at some specific point. Aug 30 11:37:55 I think it is hanging when it tries to start udevd Aug 30 11:38:27 echo "starting the udev" ; sleep 5 ; ... Aug 30 11:38:50 echo "udev done" ; Aug 30 11:39:05 I use init=/bin/sh Aug 30 11:39:10 not pretty, but who cares if it works. Aug 30 11:39:17 and try starting udevd by hand Aug 30 11:39:22 heh Aug 30 11:39:44 yeah, I guess that would work too. Didn Aug 30 11:39:49 I am wondering if the kernel doesn't hae the right options or something Aug 30 11:39:51 't think about that Aug 30 11:40:29 Crofton: hey, you got it processing the init scripts, that is a big plus! Aug 30 11:40:34 yeah Aug 30 11:40:40 lol Aug 30 11:40:54 sometimes just getting up that far can be so painful Aug 30 11:40:55 I just need to understand udev know Aug 30 11:41:02 * T0mW ducks Aug 30 11:41:08 heh Aug 30 11:41:33 someone had experience with Fujitsu-Siemens notebooks? Aug 30 11:42:04 do want to buy one? Aug 30 11:42:18 Ania want to buy notebook, I have to select which one Aug 30 11:42:21 hrw|work: avoid plague are words that come to mind Aug 30 11:43:00 I bought a Samsung last year Aug 30 11:43:03 XorA: so I should add F-S to same list as Acer? Aug 30 11:43:31 hrw|work: my GF had one, it broke, work has them they just plain suck Aug 30 11:43:39 hrw|work: erm wife now :-) Aug 30 11:44:29 in fact my inlaws have a F-S desktop as well and that s appallng Aug 30 11:44:44 by spec it should match one of my machine, it does not Aug 30 11:44:54 XorA: other option is Toshiba L30-115 Aug 30 11:45:36 generally it has to be CeleronM 1.5/1.6, 512M ram, 40G, wifi and 1280x800 screen - all <600EUR new Aug 30 11:45:41 ~change 600 eur to pln Aug 30 11:45:46 600.00 Euro (EUR) makes 2,373.77 Zloty (PLN) (from http://www.xe.com/) Aug 30 11:45:56 ~change 2500 pln to eur Aug 30 11:46:01 2,500.00 Zloty (PLN) makes 631.907 Euro (EUR) (from http://www.xe.com/) Aug 30 11:46:11 ok. 600 - 650eur Aug 30 11:46:32 hrw|work: my tosh lasted 3 years of me banging it about on public transport and is now owned by my brother who still has broken it, it was a S3000-X4 (really made by compal) Aug 30 11:47:02 hasnt broken it Aug 30 11:47:17 bakc soon Aug 30 11:47:24 morning all Aug 30 11:47:49 hi Liam Aug 30 11:47:55 lrg: where is your slave? Aug 30 11:48:12 hrw|work: he is ill :( Aug 30 11:48:47 ah. I hope he will get well soon Aug 30 11:48:53 hrw|work: I will send his patches to you, do13 and RP for review shortly. Aug 30 11:49:12 hey Liam Aug 30 11:49:29 hrw|work: so do I, but I am getting some peace and quite today ;) Aug 30 11:50:36 who or what recipe / bb / ?? builds the work/deploy/image ? Aug 30 11:50:42 lrg: which brand of notebooks (to 600-650 eur) you will suggest? Aug 30 11:50:52 T0mW: any from packages/images/ Aug 30 11:50:53 I cannot find anything in packages, conf that does this. Aug 30 11:51:32 hrw|work: not sure, we have a contract with Dell, so all ours are Dell (and all are ok with Linux) Aug 30 11:51:40 T0mW: bootstrap-image? Aug 30 11:51:44 hrw|work: well, bootstrap-image preps the image, but who make the actual jffs2, ext2, tar, etc. binaries. Aug 30 11:52:00 T0mW: thats job of bbimage (part of bitbake) iirc Aug 30 11:52:10 T0mW: IMAGE_FSTYPES is var which you need Aug 30 11:52:17 got that Aug 30 11:52:31 lrg: good part of job ;) Aug 30 11:53:14 I want to change the name of the "bootstrap-image-" to simple "disk.img" Aug 30 11:53:15 why would I build udev_092 instead of udev_097? Aug 30 11:53:26 I do not think I prefer 092 anywhere? Aug 30 11:53:31 hrw|work: I have Latitude D410 atm, and it's pretty good. Aug 30 11:53:53 T0mW: then do own tomw-image which will have IMAGE_NAME="disk.img"? Aug 30 11:54:05 yeah Aug 30 11:54:27 AH, "IMAGE_BASENAME" Aug 30 11:54:54 hrw|work: sorry, you were right, I was moving too fast again to think Aug 30 11:55:33 T0mW, I usuallly shorten the name when I copy it Aug 30 11:55:54 but leave the long name to make it easier when I accumulate too many images Aug 30 11:56:23 well, having a constantly changing name is not convenient for what I'm doing right now. Aug 30 11:57:14 conf/bitbake.conf Aug 30 12:00:24 I can understand that :) Aug 30 12:01:35 hrw|work: with all these changes I'm doing right now, adding new recipes. Is this stuff going to be lost if I do an update on the OE.mtn database, or sync the org.openembedded.dev branch? Aug 30 12:01:49 T0mW: will not get lost Aug 30 12:02:01 T0mW: you can have few dirs in BBFILES too Aug 30 12:02:07 Crofton: then there is the confusion as to which is the most current. Aug 30 12:02:33 not if they have time/date stamps :) Aug 30 12:02:34 T0mW: BBFILES="/stuff/org.openembedded.org/packages/*/*.bb /stuff/my-own-recipes/*/*.bb" Aug 30 12:02:36 hrw|work: oh! So, I can create my own private recipe branch, cool! Aug 30 12:02:50 thats much better Aug 30 12:03:11 T0mW: there are also BB_COLLECTIONS which can have higher priority with OE ones... but I never used them Aug 30 12:14:38 re XorA Aug 30 12:14:59 re people Aug 30 12:15:13 XorA ^ XorA Aug 30 12:15:26 T0mW: same result :-) Aug 30 12:16:17 heh Aug 30 12:16:33 XorA: I got to build now, and something that boots too. Aug 30 12:16:46 T0mW: cool Aug 30 12:17:01 I hope that this pain is going to be worth the time Aug 30 12:17:20 getting easier thought Aug 30 12:17:23 getting easier though Aug 30 12:17:41 T0mW: its worth it in 3 months time when your boss goes, can you do a change to that image and all you have to do is bitbake swanky-image :-) Aug 30 12:20:03 IOW, revert back to a specific system release? Aug 30 12:20:43 that's like an integrated cvs but on a system wide basis Aug 30 12:21:51 I've only been viewing bitbake / OE as a way to control / co-ordinate the build of a complex system. But, the ability to revert could be useful, "oops, let me go back" Aug 30 12:21:52 heh Aug 30 12:34:44 morning Aug 30 12:58:22 ~hail mikearthur Aug 30 12:58:26 * ibot bows down to mikearthur and chants, "I'M NOT WORTHY!!" Aug 30 13:11:16 03hrw 07org.oe.oz354x * r1d134e2a... 10/ (4 files in 3 dirs): linux-openzaurus: added patches from mikearthur to fix TS on tosa Aug 30 13:11:17 ~seen zecke Aug 30 13:11:30 zecke was last seen on IRC in channel #oe, 1d 1h 26m 19s ago, saying: 'later'. Aug 30 13:12:08 hrw|work: you are very fast Aug 30 13:12:19 03hrw 07org.oe.dev * rf07209e1... 10/ (4 files in 3 dirs): linux-openzaurus: added patches from mikearthur to fix TS on tosa Aug 30 13:13:05 do13_: built, flashed, tested Aug 30 13:13:14 hrw|work: these patches needs love Aug 30 13:13:41 do13_: I know but even in this moment tosa behave better with them Aug 30 13:14:50 hrw|work: it's ok for the moment. But this needs to be implemented in a cleaner way Aug 30 13:14:59 hrw|work, do13: mike will hopefully be better tomorrow and will clean up as required. Aug 30 13:15:26 * lrg hopes mike is better tomorrow Aug 30 13:16:40 lrg: np, we have a working solution. we can do the cleanup later Aug 30 13:18:41 do13: :) Aug 30 13:21:55 hey Koen Aug 30 13:23:20 koen - good morning - thx for the feedback about ical/webdav - you must have read something besides the build-on-osx page of the wiki to know im interested in that Aug 30 13:27:35 poushag: the OZ wiki recentchanges is in my newsreader :) Aug 30 13:28:16 brb ... updating xchat Aug 30 13:33:08 ah - makes sense now :) glad to know im not hte only one reading that page - speaking of which - oz.org isnt loading for me - is it down now? Aug 30 13:33:39 loads for me but slow Aug 30 13:34:50 XorA: how does it upset gtk+? Aug 30 13:35:08 Crofton|lab: you might want to attach udevmonitor to see what udev is doing Aug 30 13:38:47 /* The 25 and 44 'magic numbers' are from Sharp's 2.4 patches */ Aug 30 13:40:11 anyone of you ever seen: Aug 30 13:40:13 touch: setting times of `pm_to_blib': Permission denied Aug 30 13:40:16 while building perl-native ? Aug 30 13:40:26 hi mickeyl Aug 30 13:40:49 hey mickey|work Aug 30 13:40:59 morning, btw. Aug 30 13:41:00 mickey|work: did the world-record thing work out? Aug 30 13:41:28 ya. actually the guiness book wasn't interested, but we did it despite of them. Aug 30 13:42:21 hi mickey|work Aug 30 13:42:43 mickey|work: did you read oedem page? I added some PR links Aug 30 13:43:12 not yet. cool - thanks Aug 30 13:43:35 lunch Aug 30 13:49:46 koen: total renderng falure, bg black boxes, corruption, wierd colors Aug 30 13:50:12 koen: and slows down to a crawl Aug 30 13:53:14 I could have lived with it being fast and ugly ;) Aug 30 13:55:36 * XorA was hopng for some nice fast pdf scrolling, but the epdfview doesnt display Aug 30 13:58:10 03koen 07org.oe.dev * r2e20cd2e... 10/ (1 packages/xorg-lib/libx11_X11R7.1-1.0.1.bb): libx11: bump PR to keep up with diet-x11 Aug 30 13:58:49 XorA: maybe mardy or sirfred will know the cause Aug 30 13:59:17 XorA: have you seen http://libw100.sourceforge.net/index.php/Overview ? Aug 30 13:59:46 (as soon as sf.net wakes up) Aug 30 14:03:24 03mickeyl 07org.oe.dev * r9a707a50... 10/ (5 files in 3 dirs): ltrace: add mvista-toolchain specific version, fix building w/ gnueabi Aug 30 14:03:24 eww Aug 30 14:03:26 mvc Aug 30 14:03:43 * mickey|work hides in shame Aug 30 14:03:50 hahaha Aug 30 14:04:16 Why does S03sysfs not mount /proc and /sysfs ? Aug 30 14:04:21 * koen sets DISTRO=mvmd Aug 30 14:04:29 :) Aug 30 14:04:37 mickey|work, has sinned Aug 30 14:04:51 only for those crazy moto-ezx folks Aug 30 14:05:01 who want to expand the MOTO image Aug 30 14:06:06 koen: Did you ever get Firefox 1.5 running on your device? Aug 30 14:07:07 ~praise mtn pluck Aug 30 14:07:12 All hail mtn pluck! Aug 30 14:08:33 * chouimat goes back swearing at cirrus kernel code ... Aug 30 14:08:50 chouimat: ep93xx? Aug 30 14:09:00 mattcamp: nope Aug 30 14:09:10 XorA: http://pippin.gimp.org/horizon/ Aug 30 14:09:12 koen ep9315 ... Aug 30 14:09:36 chouimat: that's supported in mainline now Aug 30 14:10:03 koen I don't see the spi code for the edb9315a board Aug 30 14:10:38 ah, right Aug 30 14:10:45 http://ep93xx.giantshoulderinc.com:8001/wiki/wiki/KernelStatus lists spi as missing Aug 30 14:11:03 cirrus has a git tree nowadays Aug 30 14:11:11 so you can do distributed swearing Aug 30 14:11:17 koen hehe Aug 30 14:11:21 koen: Did Firefox 1.0.7 work for you? Aug 30 14:11:34 mattcamp: also didn't try that Aug 30 14:11:44 koen the 2.6.8.1 code is ugly ... Aug 30 14:12:16 chouimat: I know Aug 30 14:12:36 chouimat: I don't need spi, so I run derevo on my ep9312 Aug 30 14:13:00 koen and all the missing stuff I need those ... stupid hardware guys who seems to think I will be able to "fix" this in 2 weeks Aug 30 14:13:21 is the oe.org web site down? Aug 30 14:15:11 Is Firefox 1.5 working for anyone on an XScale-based device? For me, it starts and doesn't show any errors, but just sits there and doesn't open any windows. Aug 30 14:15:36 bigd0g: it works over here Aug 30 14:18:04 koen: hrm... was loading, but now is not. can't ping the site, either Aug 30 14:18:18 dns issue? hrm. using opendns.com servers Aug 30 14:18:36 koen look like I will have fun this weekend ... Aug 30 14:18:37 Aside from the automatically computed shared library dependencies, and of course an X server, Does firefox 1.5 need anything else in order to run? Aug 30 14:19:42 strace it Aug 30 14:21:19 site works now. odd. Aug 30 14:21:21 * bigd0g shrugs Aug 30 14:23:23 03mickeyl 07org.oe.dev * r3c1d5bb4... 10/ (4 files in 2 dirs): lsof: remove 4.74, remove 4.75, add 4.76, add gnueabi support Aug 30 14:25:29 mattcamp: all firefox 1.5 series + minimo do that Aug 30 14:27:02 Any idea why? Aug 30 14:28:12 mattcamp: no, never figured it out, and as ff 1.5 series is broken for cross compile, I gave up Aug 30 14:28:28 03tmbinc 07org.oe.dreambox * rb763cae3... 10/ (1 packages/enigma/enigma_cvs.bb): enigma: update to fix dm600pvr issues Aug 30 14:28:42 hm than koen compiles ff native? Aug 30 14:28:52 03tmbinc 07org.oe.dreambox * r8e6f26df... 10/ (3 files in 3 dirs): initscripts-opendreambox/dm600pvr: create required links for demux and dvr Aug 30 14:29:08 woglinde: I don't run the opterons in 64bit Aug 30 14:29:11 03tmbinc 07org.oe.dreambox * rebf3cc87... 10/ (1 packages/linux/linux-dm600pvr.bb): linux-dm600pvr: update to patch r3 Aug 30 14:29:23 koen ah Aug 30 14:29:39 arm compiler doesnt like -m32 as an arguement Aug 30 14:30:06 sed Aug 30 14:30:16 woglinde: one day when I have time Aug 30 14:30:21 hehe Aug 30 14:30:22 yes time Aug 30 14:30:32 never ever have enough Aug 30 14:30:56 and I you think you will have, you have more work to do than ever imagined Aug 30 14:31:00 if Aug 30 14:31:09 woglinde: I just use links2 instead Aug 30 14:32:31 koen - when using gpe-calendar for webdav/ical remote calendars, do you use password? and if so, do you secure it with openvpn or anything? Aug 30 14:34:05 any tmpfs gurus about Aug 30 14:34:19 mount -t tmpfs /var Aug 30 14:34:30 fails with invalid argument Aug 30 14:35:03 hm Aug 30 14:35:05 tmpfs /tmp tmpfs size=700M,mode=1777 0 0 Aug 30 14:35:09 is from my fstab Aug 30 14:35:20 I am trying to do it manually first Aug 30 14:36:12 Crofton: mount -t tmpfs tmpfs /var Aug 30 14:36:36 still fails with invalid arg Aug 30 14:36:47 hm mount ist busybox? Aug 30 14:36:53 yes Aug 30 14:37:01 to be sure, check for it in /proc/filesystems Aug 30 14:37:13 giel, it is there Aug 30 14:41:46 poushag: plain webdav Aug 30 14:41:58 ah Aug 30 14:42:09 looks like tmpfs isn't in kernel config Aug 30 14:42:24 it does show in /proc/filesystems ... Aug 30 14:42:26 | make[1]: Leaving directory `/space/fic/openmoko/build/tmp/work/i686-linux/libxml-parser-perl-native-2.34-r0/XML-Parser-2.34/Expat' Aug 30 14:42:26 | mkdir /usr/local/lib/perl: Permission denied at /usr/share/perl/5.8/ExtUtils/Install.pm line 112 Aug 30 14:42:28 grrr Aug 30 14:42:30 anyone seen thatß Aug 30 14:42:31 ? Aug 30 14:43:17 seems it's not picking up the proper perl isn't it? Aug 30 14:43:33 * mickey|work tries with ASSUME_PROVIDED Aug 30 14:45:09 koen - so you are using 'security by obscurity'; that is anyone could change your calendars if knew where to find them? Aug 30 14:45:27 poushag: they'd have to authenticate to apacke Aug 30 14:45:34 apache Aug 30 14:45:47 https? Aug 30 14:46:06 http to local server Aug 30 14:46:26 * koen loves ssh portforwarding Aug 30 14:46:39 * poushag too Aug 30 14:46:41 * koen hints mickey|work at http://www.openembedded.org/people Aug 30 14:47:33 can't see any perl hacker listed there Aug 30 14:49:51 koen - i running apache-1.3.something on my osx box - using webdav-digest to authenticate calendar sharing - can you share details of yours please? Aug 30 14:50:06 s/i/i am Aug 30 14:50:11 mickey|work: you can't see a lot of people listed there Aug 30 14:50:15 poushag: debian stable + moddav Aug 30 14:50:35 is that apache 2 or 1? Aug 30 14:50:44 koen: we should keep that more uptodate, right Aug 30 14:53:14 hi polyonymous Aug 30 14:53:18 hi Aug 30 14:54:17 someone want to dig out his collie? Aug 30 14:54:26 http://ewi546.ewi.utwente.nl/tmp/hrw/3.5.4x-experimental/collie/ - 3.5.4.1/collie images Aug 30 14:55:54 hi hrw Aug 30 14:57:10 what's new? Aug 30 14:59:05 mickey|work: nothing - its 3.5.4.1+collie/3.5.4 stuff Aug 30 14:59:19 2.6 based? Aug 30 14:59:21 mickey|work: but it give binary locales, hostap 0.4.7 etc Aug 30 14:59:41 mickey|work: no - 2.4.18 for those who has problems with 3.5.4+upgrades Aug 30 14:59:45 ah right. Aug 30 14:59:51 cool, i'll check the opie-image later tonite Aug 30 15:00:24 if I copy a config file into a kernel tree, I must run make defconfig? Aug 30 15:00:35 hrw|work: no 5000d kernel? Aug 30 15:00:51 crofton yes Aug 30 15:00:52 chouimat: can build it Aug 30 15:01:08 hmm Aug 30 15:01:27 config_prepend for nslu2 doesn't ... Aug 30 15:01:35 hm.. is anyone doing something on qtopia core 4 or shall I start to get a first version of a .bb working? Aug 30 15:02:48 go ahead. Aug 30 15:03:08 i had it working once in snapshot-mode Aug 30 15:03:27 but they changed too much, so i gave up catching up Aug 30 15:03:32 k Aug 30 15:03:37 'cause i need it anyway Aug 30 15:03:50 i'll try to fiddle with it and make something which works for arm Aug 30 15:03:54 good. it'd be neat to check its performance Aug 30 15:03:59 i don't have any other systems to try it on anyway Aug 30 15:04:40 i expect to have something somewhere next week, it's something that i need to do in between other stuff... i'll let you guys know... Aug 30 15:06:13 chouimat: added 32-0 Aug 30 15:06:14 bootstrap-image complains about missing tail, cut sort, ipkg-cl, id Aug 30 15:07:53 hrw|work: thanks Aug 30 15:21:02 what specifies which "locale"-directory is used? ( /lib/locale vs. /usr/lib/locale ) Aug 30 15:29:07 03hrw 07org.oe.oz354x * rcdf2b3c7... 10/ (3 files in 2 dirs): oz 2.4 kernels: move some tasks to get them build with rm_work Aug 30 15:36:22 tmbinc hm I thinl /usr/lib/locale should be used Aug 30 15:36:45 yes but my most recent image uses /lib/locale Aug 30 15:37:01 hm Aug 30 15:37:09 at least it searches that directory for locales Aug 30 15:37:25 hm Aug 30 15:50:29 http://ewi546.ewi.utwente.nl/tmp/hrw/fb/?name=bluez&action=search Aug 30 15:51:28 koen: will finish it later and then push Aug 30 15:52:20 cu Aug 30 15:54:44 hey pH5 Aug 30 15:55:13 hi lrg. any news on the dbus-native problem? Aug 30 15:55:17 pH5: dbus-glib-native builds fine now thanks Aug 30 15:55:21 ah, good Aug 30 16:03:43 lrg: Do you know what pieces would be needed to get asoc to work with the opie volume controls? Aug 30 16:05:00 hvontres|poodle: opie volume will need alsa support, I'm afarid I don't know much about opie-mixer Aug 30 16:05:58 hvontres|poodle: If it has alsa support then main volume should be mapped to "PCM Volume" or "Playback Volume" Aug 30 16:08:28 lrg: thanks. I might take a look at it... not much of a coder tho.. still learning Aug 30 16:09:34 hvontres|poodle: maybe it's something that can be done in a config file ? Aug 30 16:10:59 lrg: hmm, maybe. I'll have to take a peek at things tonight when I get home. Aug 30 16:15:07 zecke: Do you know if the opie mixer supports alsa yet? Aug 30 16:16:06 hey mikearthur is alive ! Aug 30 16:16:20 lrg: am now Aug 30 16:16:27 been in best most of the day :s Aug 30 16:16:37 lrg: any problems with the patch Aug 30 16:16:48 s/been in best/been in bed Aug 30 16:16:55 mikearthur: tosa touch needs a little clean up. Will do this tomorrow. Aug 30 16:17:00 sure Aug 30 16:17:22 mikearthur, Marcin has commited in to oe atm. Aug 30 16:17:36 hvontres|poodle: unlikely Aug 30 16:17:53 lrg: cool, both patches? Aug 30 16:18:21 mikearthur: afaik, yes. Aug 30 16:18:29 k Aug 30 16:18:38 what sort of stuff needs cleaned on tosa, out of interest Aug 30 16:19:13 mikearthur: I've not had a chance to look yet. do13 has looked and spotted a few areas that need love Aug 30 16:19:19 ok sure Aug 30 16:19:45 hey mikearthur Aug 30 16:19:58 hey koen Aug 30 16:20:25 mikearthur: GWS Aug 30 16:20:31 gws? Aug 30 16:21:16 mikearthur: get well soon Aug 30 16:21:23 ah, k, thanks Aug 30 16:21:39 who knows when that will be :s Aug 30 16:21:55 normally it's fine, but this summer my medication has been screwing around Aug 30 16:22:06 so I think I've had 4 days ill in 3 months, which isn't great :s Aug 30 16:22:13 normally I'm fine Aug 30 16:29:39 * koen is trying to think of a clean way to make TARGET_FPU overridable Aug 30 16:31:25 koen: good idea, this should make adding mx31 easier Aug 30 16:31:49 exactly :) Aug 30 16:31:51 koen: :) Aug 30 16:32:21 +#make it overridable for platforms with FPU, like ep93xx or i.mx31 Aug 30 16:32:21 +TARGET_FPU_arm ?= "soft" Aug 30 16:32:52 and put TARGET_FPU="" in conf/machine/include/tune-imx31.conf Aug 30 16:33:24 koen: fine by me Aug 30 16:33:37 as soon as I verify it work Aug 30 16:33:42 * koen kicks bitbake Aug 30 16:37:04 * pH5 wishes somebody would write a battery driver for his magician Aug 30 16:39:41 03koen 07org.oe.dev * rc8d646c1... 10/ (1 conf/distro/include/angstrom.inc): Aug 30 16:39:41 angsgrom: make TARET_FPU overridable Aug 30 16:39:41 * put TARGET_FPU="" in conf/machine/include/tune-.conf to build a hardfloat toolchain Aug 30 16:39:41 * mixing soft and hardfloat code is safe, BUT hardfloat binaries still don't run onf FPU-less systems Aug 30 16:40:33 lrg: there you go Aug 30 16:41:06 hehe Aug 30 16:41:09 ~hail koen Aug 30 16:41:18 * ibot bows down to koen and chants, "I'M NOT WORTHY!!" Aug 30 16:42:35 mikearthur: hey, Are you feeling better? Aug 30 16:43:40 hi Aug 30 16:43:50 hey hrw Aug 30 16:44:37 a+ Aug 30 16:44:39 koen: thanks Aug 30 16:45:27 mikearthur: get well soon Aug 30 16:46:02 hrw: that search thingy looks neat Aug 30 16:46:08 mv: cannot stat `hostname': Invalid argument Aug 30 16:46:16 ~lart xfs and my home machine Aug 30 16:46:16 * ibot executes killall -KILL xfs and my home machine Aug 30 16:46:31 koen: it need adding sections, searching in provides... Aug 30 16:46:48 koen: using Packages.filelist would be nice too Aug 30 16:47:12 hrw: ipkg-utils produces broken Packages.filelists :( Aug 30 16:47:32 koen: even when you remove them before ipkg-make-index? Aug 30 16:47:59 that should work Aug 30 16:48:28 koen: thats how I refresh feeds Aug 30 16:49:02 koen: rm Packages*;ipkg-make-index;grep -v ^Source: Packages|gzip -c9>Packages.gz ;gunzip -c Packages.gz>Packages Aug 30 16:51:27 Hello Aug 30 16:51:36 hey sirfred Aug 30 16:51:56 Hi, koen Aug 30 16:52:06 hi sirfred Aug 30 16:52:27 hrw: Hi. Aug 30 16:53:03 hi sirfred Aug 30 16:53:50 ooops Aug 30 16:54:05 * koen forgot -name in 'find . foo | xargs rm' Aug 30 16:54:16 glubs Aug 30 16:54:28 That's not good, is it? Aug 30 16:55:21 * koen hugs backups Aug 30 16:58:48 * chouimat is away: lunch and getting ready for the Long weekend Aug 30 17:11:41 ibot: 713 - 337 Aug 30 17:11:42 376 Aug 30 17:12:04 ibot: 0xff - 0x55 Aug 30 17:12:06 koen: update.php parse ~10 packages per second.. Aug 30 17:12:21 ... Aug 30 17:12:26 ibot: 2b or !2b? Aug 30 17:12:38 hrw: could be, I didn't benchmark it (yet) Aug 30 17:12:45 koen: I just did Aug 30 17:12:50 0.44072000 1156958337 loaded Packages.gz Aug 30 17:12:54 0.02924500 1156958713 new feed start Aug 30 17:13:11 koen: want to rewrite it in C? Aug 30 17:13:36 heh Aug 30 17:13:52 that would involve pointers Aug 30 17:29:36 03tmbinc 07org.oe.dreambox * r15f5ed52... 10/ (1 packages/glibc/glibc_2.3.5+cvs20050627.bb): glibc-2.3.5+cvs20050627: fix localedir path Aug 30 17:29:44 03tmbinc 07org.oe.dreambox * r5e485420... 10/ (1 packages/tuxbox/tuxbox-plugins.bb): tuxbox-plugins: update package to fix include path Aug 30 17:30:39 I'm trying to debug Firefox 1.5 inside gdb, and after I let it run for about a minute, it apparently kills itself, because I see "Killed" and then a shell prompt. Is there a way to catch the KILL signal, or at least dump core when it happens? Aug 30 17:31:47 mattcamp: Perhaps what gets killed is gdb itself? Aug 30 17:32:10 mattcamp: welcome to ARM Linux :) Aug 30 17:33:16 Oh, is this common under ARM Linux? Aug 30 17:33:50 mattcamp: I see that regulary, and always lacked the time to look into the source Aug 30 17:36:02 mattcamp: I guess that perhaps gdb is running out of memory. Aug 30 17:36:09 mattcamp: And gets killed by the kernel. Aug 30 17:36:22 That makes sense; Firefox is rather large. Aug 30 17:36:30 check the dmesg for the OOM Aug 30 17:36:40 This device has 64 MB of RAM. It does have a hard disk so I could turn on swap if needed. Aug 30 17:37:04 yeah, out of memory Aug 30 17:37:14 mattcamp: Don't you have access to a flash or something? Aug 30 17:37:34 mattcamp: Sure it will be slow, and perhaps the flash gets damaged, but they're cheap nowadays. :) Aug 30 17:38:03 mattcamp: I mean an external flash, of course. A CF card or something so. Aug 30 17:38:12 The device I'm working on has a built-in hard disk so I'll use that. Aug 30 17:38:40 mattcamp: Oh, excuse me. I understood it hasn't it. Aug 30 17:38:52 mattcamp: Good luck. Aug 30 17:53:21 psokolovsky_: http://angstrom-distribution.org/unstable/images/20060829/h4000/ Aug 30 17:55:22 koen: Cool! Thanks! Aug 30 17:55:55 I'm having some troubles with convincing bitbake to only build diet-x11, though Aug 30 17:56:10 * koen stabs libx11_X11R7.1*.bb Aug 30 17:58:09 koen: yes, full was built for me too Aug 30 17:58:44 I can spare the extra 400k, but I do like to sort out the situation Aug 30 18:02:33 evening all Aug 30 18:02:39 hey Ifaistos Aug 30 18:04:34 koen : I have an old ipaq 3630 (or 3360 not sure) which image is better ? Aug 30 18:04:49 ? Aug 30 18:05:36 koen : from the targets opie/gpe ? Aug 30 18:05:38 I'd say install a bootstrap-image and go from there Aug 30 18:07:43 koen : ipaqs have a bootloader if i remember correctly (called parrot or something similar) in case things screw up Aug 30 18:28:00 'evening all ! Aug 30 18:36:25 * chouimat is back. Aug 30 18:38:26 hmmm Aug 30 18:38:35 Now I can't seem to flash a kernel Aug 30 18:38:48 when I run imls I get bad checksums .... Aug 30 18:56:32 Does anyone had trouble theses days to build quilt-native ? I have a strange error on do_unpack http://rafb.net/paste/results/5mYwNl34.html ... Aug 30 18:57:16 Right now I am regretting not saving the last working kernel image I had ..... Aug 30 19:02:26 03pfalcon 07org.oe.dev * rcc5a7825... 10/ (1 packages/initscripts/initscripts-1.0/populate-volatile.sh): Aug 30 19:02:26 initscripts: populate-volatile.sh: Add explicit command separators. Aug 30 19:02:26 * There was issue seen when populate-volatile.sh executed something like Aug 30 19:02:26 "mkdir chown" when configuration was not full right. This fixed it. Aug 30 19:02:26 Closes #1355. Aug 30 19:05:34 03pfalcon 07org.oe.dev * rb8233d3f... 10/ (1 packages/linux/handhelds-pxa-2.6_2.6.16-hh5.bb): handhelds-pxa-2.6_2.6.16-hh5: Add patches from handhelds-pxa-2.6_cvs . Aug 30 19:08:09 cyrilRomain: Did mtn pull? There were fixes for that. Aug 30 19:08:14 hum... it works with bitbake 1.4.5 but not with 1.6.1 ... Aug 30 19:08:31 psokolovsky_: I updated this evening Aug 30 19:08:50 psokolovsky_: let's update again :) Aug 30 19:10:08 I'm on 1.6.1, but didn't have issues today with build. maybe I'm just lucky and it didn't rebuild ;-) Aug 30 19:11:00 psokolovsky_: I'm starting all again from scratch (otherwise quilt would not rebuild) Aug 30 19:12:13 03koen 07org.oe.dev * r8605731b... 10/ (1 packages/linux/handhelds-pxa-2.6_2.6.16-hh5.bb): hh-pxa-2.6 2.6.16-hh5: remove maxdate from iwmmxt patch Aug 30 19:14:29 has anything changed in OE over the past day or so, that would cause kernels to get crc errors when flashed? Aug 30 19:14:34 this is driving me nuts Aug 30 19:16:39 argh. still don't work even with lastest database... so I have to investigate more debuging bitbake Aug 30 19:25:15 psokolovsky_: are the fixes you talk been pushed on .dev ? Aug 30 19:26:11 cyrilRomain: I just saw some discussion and commits regarding that, but not sure to what exact issue they were related. Aug 30 19:27:13 psokolovsky_: ok thx anyway Aug 30 19:37:46 hi Aug 30 19:38:33 hi hrw Aug 30 19:53:22 I'm having problems with do_rootfs... Or more specifically, log_check Aug 30 19:53:25 log_check: Using /home/repvik/slug/releases/slugos-3.10-beta/debianslug-nslu2.tmp/work/slugos-image-1.0-r28/temp/log.do_rootfs.783 as logfile Aug 30 19:53:28 log_check: There were error messages in the logfile Aug 30 19:53:31 -e log_check: Matched keyword: [Cannot find package] Aug 30 19:53:48 The only problem is that there is *no* errors in the logfile :( Aug 30 19:58:31 anyone got an idea why? Aug 30 20:37:49 NAiL: search for "Cannot find package" Aug 30 21:06:59 finally got this stupid OSK to boot Aug 30 21:07:10 a chain of stupid mistakes cost me hours Aug 30 21:07:45 welcome to our world Aug 30 21:08:32 JustinP: I grepped for "annot". No hits. Aug 30 21:10:01 NAiL: strange. Well, it *says* it's finding that...perhaps find the code for log_check? Aug 30 21:10:49 now I should check in some crap Aug 30 21:11:18 synthesizing initial hotplug events takes a long time Aug 30 21:12:23 koen: any ideas on http://rafb.net/paste/results/5mYwNl34.html ? I'm still looking for my chain of stupid mistakes but ... Aug 30 21:13:15 the strange thing is that the unpacking command "os.system(cmd)" returns 0 when executed by hand in a python shell but -1 with bitbake :-/ Aug 30 21:13:58 * koen has ASSUME_PROVIDED += "quilt-native" Aug 30 21:14:13 ah Aug 30 21:14:54 * cyrilRomain gives it a try Aug 30 21:15:14 any quick hints on how to mount a cf card in a udev system? Aug 30 21:15:26 Crofton: it should automount Aug 30 21:15:46 Crofton: otherwise 'mount /dev/hda /mountpoint' Aug 30 21:15:47 what would I need to setup to make this happen :) Aug 30 21:16:16 I may have issues Aug 30 21:16:27 may??? Aug 30 21:16:34 I get registering new device pcmcia0.0 Aug 30 21:17:03 Crofton: so it detects a new pcmcia device Aug 30 21:17:14 this is progress Aug 30 21:17:17 Crofton: but it doesn't load ide_cs Aug 30 21:17:27 last time I did this I hacked the support together :) Aug 30 21:17:44 ah module not found Aug 30 21:18:40 likely becasue they are iinstalled in the wrong dirctory Aug 30 21:18:45 kernel version stuff Aug 30 21:18:47 check your kernel config if it has the ide bits enabled and the right filesystem Aug 30 21:19:43 I think they are Aug 30 21:19:57 but in lib/modules I have two directories Aug 30 21:20:22 2.6.18-rc4-omap1 and 2.6.x+git Aug 30 21:20:34 try installing the module for your kernel version :-/ Aug 30 21:20:55 hrm Aug 30 21:21:06 pfalcon@colinux:~/linux-ppc/oe-build$ PATH="/home/pfalcon/linux-ppc/oe-build/tmp/staging/i686-linux/bin/arm-angstrom-linux-gnueabi:/home/pfalcon/linux-ppc/oe-build/tmp/staging/i686-linux/bin:/home/pfalcon/linux-ppc/oe-build/tmp/cross/bin:/usr/local/arm/3.4.1/bin:/home/pfalcon/linux-ppc/bitbake/bin:/home/pfalcon/linux-ppc/bin:/usr/local/wince/cross-tools/bin:/usr/local/bin:/usr/bin:/bin:/usr/bin/X11:/usr/games" I18NPATH="/usr/share/i18n" qemu-ar Aug 30 21:21:06 m -L /home/pfalcon/linux-ppc/oe-build/tmp/work/armv5te-linux-gnueabi/glibc-2.4-r10/locale-tree /home/pfalcon/linux-ppc/oe-build/tmp/work/armv5te-linux-gnueabi/glibc-2.4-r10/locale-tree/bin/localedef --force --old-style --no-archive --prefix=/home/pfalcon/linux-ppc/oe-build/tmp/work/armv5te-linux-gnueabi/glibc-2.4-r10/locale-tree --inputfile=/usr/share/i18n/locales/es_NI --charmap=UTF-8 es_NI Aug 30 21:21:06 I need to look at bb file Aug 30 21:21:08 FATAL: kernel too old Aug 30 21:21:10 pfalcon@colinux:~/linux-ppc/oe-build$ uname -a Aug 30 21:21:11 I think ipkg-make-index -m caught you Aug 30 21:21:12 Linux colinux 2.6.11-co-0.6.3-pre13 #1 Tue May 24 14:58:16 CDT 2005 i686 GNU/Linux Aug 30 21:21:15 args Aug 30 21:21:18 pastebin Aug 30 21:21:20 koen, me? Aug 30 21:21:34 Crofton: yes Aug 30 21:21:35 that sucks. now what, it won't work on kernel <2.6.16? Aug 30 21:21:45 koen: any idea ^ Aug 30 21:21:58 Crofton: I suspect you had both versions in deploy/ipk Aug 30 21:22:22 psokolovsky_: Linux bitbake 2.6.16.1 #2 SMP PREEMPT Tue Jun 27 10:27:18 CEST 2006 i686 GNU/Linux Aug 30 21:22:28 psokolovsky_: it works on that system Aug 30 21:23:16 koen: I guess. Would you think that EABI build would check for kernel to be at least 2.6.15, where EABI started to be supported? Aug 30 21:23:41 psokolovsky_: I wouldn't be surprised Aug 30 21:24:03 psokolovsky_: glibc and gcc all have a lot of braindeadness when it comes to EABI and crosscompiling Aug 30 21:24:17 I wish I didn't have to go on vacation for four days Aug 30 21:25:38 the 2.6.18... stuff comes from the kernel build Aug 30 21:25:51 2.6.x (empty btw) is the name of the bb file Aug 30 21:26:23 ah Aug 30 21:27:03 that could only screw up module dependencies, but is harmless otherwise Aug 30 21:27:15 #define __LINUX_KERNEL_VERSION (2 * 65536 + 6 * 256 + 14) Aug 30 21:27:26 ^ glibc 2.4 config.h. crap. Aug 30 21:29:02 koen: same error even with ASSUME_PROVIDED += "quilt-native" :-/ Aug 30 21:29:31 cyrilRomain: it shouldn't build quilt-native if y ou have that in local.conf Aug 30 21:29:39 I need to check in what I have so far Aug 30 21:29:43 yes, really strange Aug 30 21:29:51 the fstab will need fooling with .... Aug 30 21:31:06 * cyrilRomain feels like in the 4th dimension Aug 30 21:33:13 I'm too stupid (or tired I hope): I was using 'bitbake quilt-native'. But I get the same problem on gnu-config-native now. Aug 30 21:34:13 so I'll add gnu-config-native to the ASSUME_PROVIDED too Aug 30 21:37:05 koen: thanks btw ! Aug 30 21:37:14 * cyrilRomain really need some sleep ... Aug 30 21:42:30 03crofton 07org.oe.dev * re9ef89a3... 10/ (6 files in 5 dirs): Aug 30 21:42:30 Update openomap distro and omap5912osk machine file with fixes to Aug 30 21:42:30 allow the OSK to boot. Includes changes to the kernel config to add Aug 30 21:42:30 TMPFS to the kernel config. Aug 30 21:42:30 The deployed fstab may need some tweaking still. Aug 30 21:43:42 hum.. something is wrong anyway. I will investigate further on that.. Aug 30 21:43:44 well, if bigdog fools with it, it shoudl at least boot Aug 30 21:43:58 tweaking next week Aug 30 21:44:05 Crofton: we are standing by to help :) Aug 30 21:44:14 heh Aug 30 21:44:34 I am headed for a field in the boonies to drinkn beer with about 2000 people Aug 30 21:44:44 most of who I can't stand :) Aug 30 21:44:55 rumour has it someone has set up wireless though Aug 30 21:45:03 no AC power though Aug 30 21:45:46 later guys Aug 30 21:46:44 night. Aug 30 21:50:40 koen: is it intented with bitbake 1.6 to use ASSUME_PROVIDED for tools that are available on the host system ? (I guess no) Aug 30 21:56:55 cyrilRomain: you can assume_provide them if you want Aug 30 21:57:28 koen: well, but this should also work without, right ? Aug 30 21:57:36 yes Aug 30 21:58:22 koen: I don't know why it works with bitbake 1.4.5 and not with bitbake 1.6.1. Aug 30 21:58:40 hmm Aug 30 21:58:40 Am I the first how have this issue ? Aug 30 21:58:49 do these quotes have to be escaped? Aug 30 21:58:51 for keyword_die in "Cannot find package" "exit 1" ERR Fail Aug 30 21:59:30 s/how/who/ Aug 30 22:01:08 NAiL: for bash script no need to escape iirc Aug 30 22:07:21 hmm Aug 30 22:07:53 there's something in log_check that fails on my box :-( Aug 30 22:08:41 http://pastebin.ca/155233 <-- code from log_check Aug 30 22:08:57 somehow it erroneously finds "Cannot find package" Aug 30 22:10:24 03koen 07org.oe.dev * r370dc8d1... 10/ (1 packages/angstrom/task-angstrom-x11.bb): task-angstrom-x11: mask of thcrut Aug 30 22:15:15 In your logfile you have the "Cannot find package" string. As JustinP said search for this string in your logfiles (log.do_something.number) Aug 30 22:16:34 NAiL: it should be dump after the 'Matched keyword: ["Cannot find package"]' line Aug 30 22:16:40 hi Aug 30 22:57:35 gaaah Aug 30 22:57:36 no Aug 30 22:57:40 there's no match at all Aug 30 22:57:56 there's nothing on the line after "Matched keyword:" Aug 31 00:07:39 Hi all Aug 31 00:08:57 having problem compiling mysql in oz-3.5.4.2 Aug 31 00:09:08 cp: cannot stat `/OE/OZ/org.openembedded.oz354x/packages/mysql/files/my.cnf': No such file or directory **** BEGIN LOGGING AT Thu Aug 31 01:17:15 2006 Aug 31 01:19:58 anyone else see a decompression error for a kernel compiled with gcc-4.1.1 ? Aug 31 01:20:42 if I compile it with 3.4.5 it works fine, but the gcc bitbake built (4.1.1) fails to decompress Aug 31 01:26:20 hmm, gcc-4.0.2 does it as well. Aug 31 01:49:00 hmm Aug 31 01:49:09 I was having issues earlier Aug 31 01:49:17 what version kernel? Aug 31 01:49:47 my issues were related to not erasing enough flash ... Aug 31 01:59:45 Crofton: good morning, it is 2.6.17.8 Aug 31 02:01:45 I decide to move all the 4.x stuff out of the way in packages/gcc and let bitbake only build for a gcc-3.4.4 and binutils 2.16.9x Aug 31 02:02:15 I moved all the binutils > 2.16 out of the way. Time to rebuild again. :( Aug 31 02:04:48 are you trying to build for OZ? Aug 31 02:04:55 no Aug 31 02:05:12 and you dont want any of the eabi stuff? Aug 31 02:05:19 correct Aug 31 02:05:42 not sure if any of the newer 2.6 kernels will build for you then Aug 31 02:06:56 it seems that gemu or whatever that is, won't build under a 4.x compiler anyway. My target system will have either opie / gpe, so, dropping back to 3.4.4 may be a better thing to do? Aug 31 02:07:14 qemu builds fine now Aug 31 02:07:23 with 4.x? ok Aug 31 02:07:28 its used in building both the oz and angstrom images Aug 31 02:07:55 and i belive they use 4.1.1 so they can use all the eabi stuff Aug 31 02:08:13 just verifying my bootloader is not at issue for some reason with the decompression error Aug 31 02:08:26 * emte shrugges Aug 31 02:08:44 i get that error on known good images Aug 31 02:08:52 just redownload/repack it Aug 31 02:09:24 I use mostly use sh4 and that doesn't work at all with gcc 4.x. Aug 31 02:09:25 um, it is the kernel expansion (boot) that gives the error Aug 31 02:09:27 and use the test tools so i dont find out 40min after i transfered it to my device Aug 31 02:09:47 oh not the image expansion? Aug 31 02:09:58 If you need to you can make OE use a different gcc to build the kernel than it uses for everything else. Aug 31 02:10:04 no, on the kernel boot I get a crc error Aug 31 02:10:07 sounds like a miscompiled kernel or module Aug 31 02:10:12 nod Aug 31 02:10:35 you using your own patches or proviced ones? Aug 31 02:10:43 provided* Aug 31 02:11:13 3.4.5 was what I used for the basic kernel dev, I tried compiling with a known-good 4.0.2 and that fails as well. It appears to be something in the 4.x compiler vs my kernel source Aug 31 02:11:38 those version numbers refer to ?? Aug 31 02:11:44 gcc? Aug 31 02:11:45 gcc Aug 31 02:12:04 the kernel patches, are they from existing recipies or yours? Aug 31 02:12:14 heh, when anything fails, suspect gcc as the culprit Aug 31 02:12:20 mine Aug 31 02:12:27 they are my patches Aug 31 02:12:35 and your using a binutils without the eabi modifications? Aug 31 02:13:07 if I take that kernel source from within tmp/work/ and compile using gcc-3.4.5, it boots ok. Aug 31 02:13:09 actually strike that... that would show up when compiling glibc Aug 31 02:13:47 IOW, CROSS_COMPILE= what advantage would there be in using EABI? Aug 31 02:16:29 I've been away from kernel development for a while Aug 31 02:17:21 the past few years mostly 8051 work and recently ARM7TDMI Aug 31 02:18:33 ; Aug 31 02:19:56 you know, I can't see how the kernel patches would disrupt the kernel decompression. I've no changes in arch/arm/boot/* Aug 31 02:20:34 so, it would appear to be some gcc version related issue Aug 31 02:37:09 error 404: /home/tom/OpenEmbedded/sources/arm-2005q3-2-arm-none-linux-gnueabi.src.tar.bz2 Aug 31 02:37:26 maybe I'll have to step back further than 3.4.x Aug 31 02:37:39 bah Aug 31 02:37:56 error 404: http://www.oesources.org/source/current/arm-2005q3-2-arm-none-linux-gnueabi.src.tar.bz2 Aug 31 02:38:25 that was the failure, oesources appears to no longer have the files for bitbake build **** ENDING LOGGING AT Thu Aug 31 02:59:57 2006