**** BEGIN LOGGING AT Thu Aug 12 02:59:57 2010 Aug 12 05:37:05 gm Aug 12 05:49:30 oops...too late Aug 12 06:37:17 03Jason Kridner  07org.openembedded.dev * r49ddf7eeda 10openembedded.git/recipes/u-boot/ (2 files in 2 dirs): Aug 12 06:37:17 u-boot: beagleboard: adjusted clocks to match PSP Aug 12 06:37:17 Signed-off-by: Jason Kridner Aug 12 06:37:17 Signed-off-by: Koen Kooi Aug 12 06:58:05 Which QT application will give me a GUI to configure my wifi and ethernet interfaces, dhcp, netmask and so on **** ENDING LOGGING AT Thu Aug 12 07:18:02 2010 **** BEGIN LOGGING AT Thu Aug 12 07:24:15 2010 **** ENDING LOGGING AT Thu Aug 12 07:26:10 2010 **** BEGIN LOGGING AT Thu Aug 12 07:28:14 2010 **** BEGIN LOGGING AT Thu Aug 12 08:22:14 2010 **** ENDING LOGGING AT Thu Aug 12 08:26:17 2010 **** BEGIN LOGGING AT Thu Aug 12 08:28:15 2010 Aug 12 08:57:32 zub: JaMa|Wrk: when i go to that website i find http://hostap.epitest.fi/wpa_supplicant - which are wpa_supplicant sources, does the GUI part be in it? Aug 12 08:59:24 rsv: S = "${WORKDIR}/wpa_supplicant-${PV}/wpa_supplicant/wpa_gui-qt4" Aug 12 08:59:34 rsv: that's path to wpa_gui sources **** ENDING LOGGING AT Thu Aug 12 09:08:57 2010 **** BEGIN LOGGING AT Thu Aug 12 13:58:16 2010 Aug 12 14:19:31 jconnolly, hmmm hi Aug 12 14:19:49 jconnolly, I don't remember but there is a variable to disable paralell make Aug 12 14:19:54 in recipes Aug 12 14:19:57 I just modded the recipe directly Aug 12 14:20:14 GNUtoo|laptop: PARALLEL_MAKE = "" you mean? Aug 12 14:20:14 PARALLEL_MAKE="" Aug 12 14:20:17 indeed Aug 12 14:20:20 hi hrw|gprs Aug 12 14:20:24 nice n900 post Aug 12 14:20:25 ;D Aug 12 14:20:26 jconnolly, I don't remember but something in theses lines Aug 12 14:20:28 thx Aug 12 14:22:01 morning Aug 12 14:22:48 hi Chris **** ENDING LOGGING AT Thu Aug 12 14:26:17 2010 **** BEGIN LOGGING AT Thu Aug 12 14:38:15 2010 Aug 12 14:38:30 i'm just glad it was that simple to fix Aug 12 14:38:39 final test is to build a kernel Aug 12 14:38:47 that he will do today Aug 12 14:38:59 kergoth: yeah Aug 12 14:39:10 i wonder if you can combine amend.bbclass + INHERIT += "srctree" safely, or .bbappend and srctree. woudl be handy to redirect an existing recipe to a local tree Aug 12 14:39:15 probably will work fine Aug 12 14:39:26 Little Chips is a board manufacturer. I acquired an ARM based development platform, and I was wondering. Thanks. Aug 12 14:39:35 ah Aug 12 14:40:01 kergoth: yeah interesting Aug 12 14:40:14 I havent played with bbappend so much yet Aug 12 14:40:21 but will try this combo Aug 12 14:40:25 the kernel particularly would be nice for that. take the stock linux-2.6 and point it at your local development git tree Aug 12 14:40:36 * khem nods Aug 12 14:41:12 I think my logging module support for bitbake is finally cleaned up enough to be merged. opens up some cool things you can do Aug 12 14:41:25 nice Aug 12 14:41:47 * RP__ is still chasing bugs in the task based prebuilts code Aug 12 14:42:01 hehe, i'm sure thats nontrivial Aug 12 14:42:17 hah - http://zeus.ugent.be:4011/ Aug 12 14:42:17 hrw|gprs: linaro gcc 4.5 is from CSL ? Aug 12 14:42:35 khem: Linaro gcc 4.4 4.5 have many csl parts Aug 12 14:42:48 khem: CodeSourcery is Linaro member Aug 12 14:42:57 kergoth: Its highlighted some nice problems in the metadata Aug 12 14:43:16 hrw|gprs: I see. Aug 12 14:43:26 hrw|gprs: do you use there 4.5 branch ? Aug 12 14:43:36 I'm not surprised, there's some crazy stuff happening in there :) Aug 12 14:43:42 khem: I provide 4.4 and 4.5 cross compilers Aug 12 14:43:54 hacks that never got cleaned up.. Aug 12 14:43:56 heh Aug 12 14:44:17 khem: but my work is mostly packaging related - I do not even use those compilers for more then helloworld checks Aug 12 14:44:43 hrw|gprs: are the sources available somehwere for public Aug 12 14:44:49 yes Aug 12 14:45:05 http://launchpad.net/ search for gcc-linaro Aug 12 14:45:24 there are bzr branches and recently there was 2008.08 release of them Aug 12 14:45:44 khem: we do as much as possible in open way Aug 12 14:46:33 khem: Ubuntu gcc 4.4 uses Linaro patches, do not remember how is with 4.5 Aug 12 14:47:31 * kergoth thinks about releasing the current 1.10 branch as 1.10.0, can always backport more fix commits from master after that.. Aug 12 14:48:09 do it, do it Aug 12 14:48:41 * kergoth tries to recall the process... git-archive, git-tag, upload to berlios, email lists Aug 12 14:48:46 thats it, isn't it? Aug 12 14:48:48 and bump OE to require it ;D **** BEGIN LOGGING AT Thu Aug 12 15:18:14 2010 Aug 12 15:18:34 its hard to highlight just what is major, of course Aug 12 15:24:20 03Martin Jansa  07org.openembedded.dev * r9e1bbdf672 10openembedded.git/ (22 files in 7 dirs): Aug 12 15:24:20 xorg: add latest versions for 2010-08-12 Aug 12 15:24:20 Signed-off-by: Martin Jansa Aug 12 15:24:23 03Martin Jansa  07org.openembedded.dev * r367d28ff1b 10openembedded.git/ (17 files in 4 dirs): Aug 12 15:24:23 qt-4.7.0-beta1: move to beta2 Aug 12 15:24:23 Signed-off-by: Martin Jansa Aug 12 15:24:23 03Martin Jansa  07org.openembedded.dev * r686b238437 10openembedded.git/recipes/navit/ (files/navit.xml navit-icons_svn.bb navit.inc navit_svn.bb): Aug 12 15:24:24 navit: bump SRCREV a bit Aug 12 15:24:24 Signed-off-by: Martin Jansa Aug 12 15:24:51 wait, events didn't exist initially.. they were added pretty quickly, but that was an addition.. lets see.. Aug 12 15:31:01 wtf Aug 12 15:31:08 this worked fine just a week or two ago Aug 12 15:31:10 now it doesn't Aug 12 15:32:14 http://web.archive.org/web/20030513111311/www.openembedded.org/phpwiki/index.php/BuildSystem - robots.txt retrieval exclusion :( Aug 12 15:33:37 first I had to ask 'we used phpwiki?' but then I looked at date. at that time it was rather you then we Aug 12 15:34:20 heh, indeed Aug 12 15:34:22 damnit Aug 12 15:34:29 i wonder if its using the *current* robots.txt, not the old one Aug 12 15:34:40 the current http://openembedded.org/robots.txt blocks everything Aug 12 15:34:50 who runs that? Aug 12 15:35:55 linuxtogo.org? Aug 12 15:36:00 no idea **** BEGIN LOGGING AT Thu Aug 12 15:38:15 2010 Aug 12 15:38:35 damn. Aug 12 15:38:44 i should have sucked that down while it was still accessible Aug 12 15:38:51 now i have to go from memory Aug 12 15:41:33 hrw|gprs: no Aug 12 15:42:15 hrw|gprs: where is your repo to get gcc 4.5 Aug 12 15:42:22 hrw|gprs: I mean on launchpad Aug 12 15:42:58 khem: moent Aug 12 15:43:43 i hope I can remember enough Aug 12 15:43:46 perhaps Aug 12 15:43:56 khem: https://edge.launchpad.net/gcc-linaro Aug 12 15:44:47 khem: 'Code' page has bzr branches Aug 12 15:45:36 you will there more infos Aug 12 15:46:04 I am on gprs most of time today - train area does not have 3G coverage too often Aug 12 15:48:57 * jconnolly just found another hrw|gprs gem... Aug 12 15:48:58 11:05 < hrw> if [ ethernet-on-usb ] then if [ no mac address ] then usb0 else eth0 fi fi Aug 12 15:49:10 just sorted out smsc95xx mac issue on bug20 finally Aug 12 15:49:41 mysteriously came up eth0 when i gave it a valid mac, without udev rename rule... i was mystified, but remembered that Aug 12 16:06:51 03Frans Meulenbroeks  07org.openembedded.dev * r1e3f1f1225 10openembedded.git/recipes/mythtv/ (3 files): Aug 12 16:06:51 mythtv: upgraded to SRCREV 25609 Aug 12 16:06:51 This includes the 0.23.1 patches Aug 12 16:06:51 Signed-off-by: Frans Meulenbroeks Aug 12 16:16:06 anyone familiar with the oe patch submission process... would you be so kind as to look here and see if it looks right? http://bugcommunity.com/wiki/index.php/Submitting_a_patch_to_Openembedded Aug 12 16:17:07 it's intended for the n00b **** BEGIN LOGGING AT Thu Aug 12 16:44:14 2010 Aug 12 16:44:18 maybe putting that on oe's wiki would be better Aug 12 16:44:19 no? Aug 12 16:44:28 and linking to it from buglabs wiki Aug 12 16:45:25 03Tom Rini  07org.openembedded.dev * r157bb114df 10openembedded.git/recipes/binutils/binutils.inc: Aug 12 16:45:25 binutils.inc: Drop spurious CC definition. Aug 12 16:45:25 There's always been a definition of CC here that gets out of sync Aug 12 16:45:25 with bitbake.conf. Drop. Aug 12 16:45:25 Signed-off-by: Tom Rini Aug 12 16:47:14 git send-email --compose --to "openembedded-devel@lists.openembedded.org" --subject "[PATCH] update bug2.0 confi " --smtp-server [...] --smtp-server-port=[...] --smtp-pass [...] --smtp-user patch.patch Aug 12 16:47:45 subject is taken from patch file Aug 12 16:48:02 smtp things is better to set in ~/.git/config Aug 12 16:48:12 ah ok Aug 12 16:48:19 mailing list set in repository/.git/config Aug 12 16:48:28 ok thanks a lot Aug 12 16:48:30 nice to know Aug 12 16:48:43 then it is just 'git send-email --compose PATCHFILE' Aug 12 16:48:49 nice Aug 12 16:48:53 and compose is needed only if you want to add something Aug 12 16:48:59 I'll do that next time that I'll have to send a patch Aug 12 16:49:13 I use it only when sending few patches and want to give description for series Aug 12 16:49:26 otherwise commit message has to be enough Aug 12 16:49:45 not using --compose teach to write good commit descriptions Aug 12 16:59:43 ok Aug 12 17:00:09 another hole in gsm coverage Aug 12 17:01:38 03Tom Rini  07org.openembedded.dev * r7245c191e2 10openembedded.git/conf/bitbake.conf: Aug 12 17:01:38 bitbake.conf: Add ext[23].bz2 images Aug 12 17:01:38 Boards with small flash and large ram need all the compression they can get. Aug 12 17:01:38 Signed-off-by: Tom Rini Aug 12 17:04:05 03Koen Kooi  07org.openembedded.dev * r60df02db6f 10openembedded.git/recipes/gnome/ (gnome-panel_2.30.0.bb metacity_2.30.0.bb): Aug 12 17:04:05 metacity: bump PR Aug 12 17:04:05 gnome-panel: bump PR Aug 12 17:12:56 03Koen Kooi  07org.openembedded.dev * r066d2f7b28 10openembedded.git/recipes/octave/octave_3.2.4.bb: octave: check in WIP recipe for 3.2.4 Aug 12 17:16:08 RP__: I have change do_distribute_sources as you suggested (with exit 1), the odd thing is that even though it fails, do_patch is still started. Aug 12 17:16:11 from the log: Aug 12 17:16:12 ERROR: Build of /home/frans/oe/openembedded/recipes/cdparanoia/cdparanoia_svn.bb do_distribute_sources failed Aug 12 17:16:12 NOTE: Running task 527 of 539 (ID: 1, /home/frans/oe/openembedded/recipes/cdparanoia/cdparanoia_svn.bb, do_patch) Aug 12 17:16:13 ERROR: Task 13 (/home/frans/oe/openembedded/recipes/cdparanoia/cdparanoia_svn.bb, do_distribute_sources) failed Aug 12 17:16:24 is this a bug? feature? Aug 12 17:17:02 I would expect do_patch to depend on successful completion of do_distribute_sources Aug 12 17:20:50 http://pastebin.com/UEna3hgB so how do I resolve the conflict between these two conf files? this was output from do_rootfs on a clean angstrom-gnome-image-big build... Aug 12 17:21:43 andyj: you need to remove ipks Aug 12 17:21:46 from deploy Aug 12 17:22:00 oh its a clean build Aug 12 17:22:01 hmm **** ENDING LOGGING AT Thu Aug 12 17:22:21 2010 **** BEGIN LOGGING AT Thu Aug 12 17:34:15 2010 Aug 12 17:36:25 hi Ken Aug 12 17:46:08 WarGames watched Aug 12 17:46:33 8" disks, acoustic modem - nice things Aug 12 17:47:08 now something more modern - Hackers with Angelina Jolie in 720p Aug 12 17:47:15 yes, I am bored ;D Aug 12 18:01:49 hrw|gprs: I think that some day I come and visit you with portable hdd ;) Aug 12 18:02:19 soltys: take sata one - will be faster ;D Aug 12 18:02:43 soltys: 17GB of DL_DIR want too? Aug 12 18:03:18 soltys: but I thought that you are studying so you should not have a problems with movies availability Aug 12 18:06:49 hrw|gprs: with new ones there is no problem, but with the older.. Aug 12 18:07:10 hrw|gprs: I had only usb2.0 portable hdd none sata.. Aug 12 18:08:13 soltys: fine. if its 2.5" and more then 200GB then sata is inside but no worry - I have ~15 usb ports available in desktop Aug 12 18:08:50 hrw|gprs: 2.5" and 500GB :) Aug 12 18:09:23 soltys: 1.5TB 3.5" as /media/storage/ Aug 12 18:09:57 hrw|gprs: If I had 200 pln more I'd buy external 1.5TB ;) Aug 12 18:12:43 I planned to buy 1TB but price difference was too small Aug 12 18:12:43 1.5TB 5400rpm was 370pln and 1TB 7200 was ~300 Aug 12 18:14:17 storage is cheap Aug 12 18:14:19 It was the same when I bought my one, but I lacked of money Aug 12 18:16:26 happens Aug 12 18:16:40 ETA: one hour to Warsaw Aug 12 18:18:18 hrw|gprs: are you ever travelling through austria? :) I'm looking for a special paint that is made in Poland :) Aug 12 18:19:22 Jin^eLD: never was in Austria Aug 12 18:19:38 well, if you ever plan a visit... ;) Aug 12 18:19:57 ;) Aug 12 18:20:14 Jin^eLD: ever heard of internet shops? Aug 12 18:20:37 ;D Aug 12 18:20:52 hrw|gprs: shipment from germany costs more than the can of paint itself, which is kind of stupid Aug 12 18:21:05 and I was not able to figure out a shop in poland Aug 12 18:22:55 Jin^eLD: mail me vendor, name etc - will check in september Aug 12 18:23:05 remind me if I will forget Aug 12 18:23:19 thanks! btw what is in september? planning a visit? Aug 12 18:23:30 no, I will be after vacations Aug 12 18:23:39 aah Aug 12 18:24:01 what is a vendor btw? Aug 12 18:24:29 APP Aug 12 18:24:55 auto plast produkt I think Aug 12 18:24:57 or whatever the spelling is Aug 12 18:26:11 never heard Aug 12 18:26:53 http://www.app.com.pl/products/?grid=11&gridn=3&prid=52 Aug 12 18:27:17 what kind of paint it is? Aug 12 18:27:17 ok, send me vendor, website, name of paint (url if exists) etc Aug 12 18:27:17 * hrw|gprs back to Hackers Aug 12 18:27:47 will do Aug 12 18:31:32 ugh... Hackers should be rated in US - there are naked tits there Aug 12 18:39:40 :) Aug 12 18:39:47 l8r **** BEGIN LOGGING AT Thu Aug 12 19:12:16 2010 Aug 12 19:34:09 03Khem Raj  07org.openembedded.dev * rc9b7719013 10openembedded.git/recipes/klibc/klibc_1.5.18.bb: Aug 12 19:34:09 klibc: Fix staging and PR Aug 12 19:34:09 Signed-off-by: Khem Raj **** BEGIN LOGGING AT Thu Aug 12 20:08:15 2010 Aug 12 20:20:29 http://wiki.openembedded.org/index.php/OpenEmbeddedJanitors now talks about FILESDIR stuff, if anyone is looking for something :) Aug 12 20:24:27 Tartarus: my preference is to delete Aug 12 20:24:58 FILESPATHPKG is an error prone mechanism Aug 12 20:25:36 if you point it to another dir someone might at some point remove the older version and leave a non functioning build without being aware of it Aug 12 20:25:40 biab, reboot Aug 12 20:27:19 huh? Aug 12 20:34:14 03Martin Jansa  07org.openembedded.dev * r37df60115d 10openembedded.git/recipes/nano/nano.inc: Aug 12 20:34:14 nano: RRECOMMENDS ncurses-terminfo Aug 12 20:34:14 Signed-off-by: Martin Jansa Aug 12 20:34:16 03Martin Jansa  07org.openembedded.dev * rb3c61f19e8 10openembedded.git/recipes/htop/htop_0.7.bb: Aug 12 20:34:16 htop: RRECOMMENDS ncurses-terminfo Aug 12 20:34:16 Signed-off-by: Martin Jansa Aug 12 20:34:17 03Martin Jansa  07org.openembedded.dev * r925829fdf7 10openembedded.git/recipes/procps/procps.inc: Aug 12 20:34:17 procps: RRECOMMENDS ncurses-terminfo Aug 12 20:34:17 Signed-off-by: Martin Jansa Aug 12 20:34:22 03Martin Jansa  07org.openembedded.dev * rb3cd464790 10openembedded.git/recipes/joe/joe_3.1.bb: Aug 12 20:34:22 joe: RRECOMMENDS ncurses-terminfo Aug 12 20:34:22 Signed-off-by: Martin Jansa Aug 12 20:34:24 03Martin Jansa  07org.openembedded.dev * rad112ee1ce 10openembedded.git/recipes/tasks/task-shr-feed.bb: Aug 12 20:34:24 task-shr-feed: add supertux-qvga wesnoth Aug 12 20:34:24 Signed-off-by: Martin Jansa Aug 12 20:45:05 I need to change some files in an existing rootfs image build. I put my changed files in a package and added the package to IMAGE_INSTALL but the bitbaking fails due to two sources for the same file. Aug 12 20:45:27 There must be a simple way to overlay files ... Aug 12 20:45:28 yep, that will happen Aug 12 20:45:47 if the recipe uses file:// to get the files (i.e. config files), you can leverage FILESPATHBASE to override them in the existing packages Aug 12 20:46:00 otherwise you can manipulate the fs directly with ROOTFS_POSTPROCESS_COMMAND (it's a shell snippet) Aug 12 20:47:01 otherwise you might need to deal with update-alternatives Aug 12 20:47:49 or more dirty, put the files in a different dir and whack init to copy them to the proper location (really ugly imho) Aug 12 20:47:52 yeah, i was thinking about mentioning that too, but for that one you'd really have to modify the recipes Aug 12 20:48:07 or make a patch for your recipe and put it in an overlay Aug 12 20:48:09 kergoth: yes Aug 12 20:48:10 though, with .bbappend and bitbake master, or with amend.bbclass, you could adjust the recipe that way Aug 12 20:48:10 OK, I understand how to do the rootfs post processing and can do manual patching from a dir, but is there an easy way to get at the guts of files in a package Aug 12 20:48:36 cdhm, you can't modify what goes into a package without modifying the recipe in some way. Aug 12 20:48:37 cdhm: i don't understand you Aug 12 20:48:53 effeM: he wants to inject his changes in the packaging process, as far as i can tell Aug 12 20:48:55 ah ok Aug 12 20:49:30 if you are only into updating the rootfs, there is no point in change the packaging, that is only useful if you are going to distribute the package Aug 12 20:49:37 I did make a recipe that builds my overlay patch then added it to IMAGE_INSTALL but that causes problems becaue the original and new files conflict. Aug 12 20:49:43 yes, it does Aug 12 20:50:04 you can't just throw files into a filesystem that another package also installs, how would it know which should win? Aug 12 20:50:22 cdhm: copy the original recipe, put it in an overlay, patch your files Aug 12 20:50:30 Perhaps by order... I was hoping... Aug 12 20:50:48 cdhm: what files do you want to toss in (and what package) Aug 12 20:50:57 most package managers have special mechanisms for this. update-alternatives, debian's .. shit, what was that feature called.. debian has a thing where a later package can force its own file to be used instead of an existing one, without modifying the original package, but i can't remember what it's called Aug 12 20:51:06 but opkg doesn't have that capability Aug 12 20:51:43 you could also do something like.. require conf/foo.inc in local.conf, then in conf/foo.inc, define do_install_append_pn- () {} Aug 12 20:51:55 and have that install your files from whereever on your filesystem, replacing the originals Aug 12 20:52:25 It is my own package. Files from svn that I want to override (eg. init scripts). If I have these in a directory I can do it with rootfs post processing. It seemed ugly to have rootfs post processing create a dir, extract svn files too it, then copy yhem in. Aug 12 20:52:25 install -m 0644 ${TOPDIR}/files/${PN}/thefileiwant ${D}${sysconfdir}/ Aug 12 20:52:27 or whatever Aug 12 20:52:37 why are you failing to grasp this? Aug 12 20:52:39 it's not that complex Aug 12 20:52:58 you can't have the files in your own package replace those of another, without installing with --force, which our rootfs stuff does not do Aug 12 20:53:07 so once again, to modify what goes into ap ackage, you have to modify the recipe Aug 12 20:53:11 (not yours, the other) Aug 12 20:53:44 Modifying other packages to turn stuff off sucks badly. Aug 12 20:53:49 either by copying the recipe to an overlay, as effeM suggests, modifying it in place in the oe tree, using amend.inc, or using a pn-${PN} override, which i just suggested Aug 12 20:54:00 well, welcome to the world of package management Aug 12 20:54:09 no package manager worth a damn lets you just replace packages in another package from yours Aug 12 20:54:11 kergoth: ;-) Aug 12 20:54:14 not rpm, not dpkg, not pkg Aug 12 20:54:17 s/pkg$/opkg/ Aug 12 20:54:20 I'll just do a manual hacky thing in rootfs Aug 12 20:54:30 we've told you like 7 different ways to do this Aug 12 20:55:45 cdhm: have you looked at RREPLACES_${PN} Aug 12 20:55:45 " install -m 0644 ${TOPDIR}/files/${PN}/thefileiwant ${D}${sysconfdir}/" where do I do this? In my overlay package or in the rootfs post processing. Aug 12 20:56:08 I'll look at RREPLACE. Aug 12 20:56:13 cdhm: my understanding is this tells opkg its OK to replace files in another package Aug 12 20:56:14 no, i just told you Aug 12 20:56:22 the two lines i just said above the one with the install command Aug 12 20:56:24 tell you where to put it Aug 12 20:56:29 you should try reading everything, not just snippets Aug 12 20:57:35 bedtime, here cya all tomorrow & stay well! Aug 12 20:57:42 night Aug 12 20:58:04 Ah ok... so do_install_append does stuff after the IMAGE_INSTALL stuff is processed and will overwrite existing files Aug 12 20:58:08 no Aug 12 20:58:16 do_install_append runs after the do_install task Aug 12 20:58:30 it modifies what files go into the package for the original pckage Aug 12 20:58:38 it has nothing to do with IMAGE_INSTALL, which is image creation / rootfs population Aug 12 20:59:51 you know what would be useful, instead of making the user use ROOTFS_POSTPROCESS_COMMAND to inject files, add a ROOTFS_EXTRA variable, which is a path, and everything in that path gets cpoied into the rootfs after the packages are installed, automatically Aug 12 21:00:04 then you could just put a file in ${ROOTFS_EXTRA}/etc/ to override the config file Aug 12 21:00:12 and set ROOTFS_EXTRA to something, of course Aug 12 21:00:27 * kergoth thinks he'll do a quick prototype of that Aug 12 21:01:47 so do_install_append just overrides the end part of a package build? That's fine for just overwriting one file in one package but is unwieldy if I need to do this for quite a few packages. Aug 12 21:02:31 are these all configuration files you're dealing with? Aug 12 21:02:47 No, some data files etc too Aug 12 21:04:54 I'm replacing the init sequence with my own scripts (coz the default start up is too slow), then changing some config files (which come from a few packages) + adding some default web pages/cgi etc. Aug 12 21:05:15 Some of the stuff does not conflict and some stuff does. Aug 12 21:08:02 I don't think RREPLACES would work. IIRC It replaces a whole package. I just want to replace a file or two. Aug 12 21:10:23 This is surely something that many embedded folk do, which is why I was expecting a one-liner answer of the form IMAGE_OVERRIDE = "package I want to drop on top" Aug 12 21:14:10 cdhm: I've not tried RREPLACES yet, but I plan to at some point. A quick hack is: http://pastebin.com/rh22XFaA Aug 12 21:14:16 cdhm: that is my current solution Aug 12 21:16:27 Thanks cbrake. While I'm sure force-overwrite does work, it is dangerous if there are unintended conflicts. Aug 12 21:17:04 Also don't like forking classes to make stuff work.... Aug 12 21:17:07 cdhm: agreed Aug 12 21:21:32 oh.. bitbake interactive mode is really needed.. Aug 12 21:22:58 kergoth: ping Aug 12 21:29:01 How do I find the package file name that gets generated? ie. I'm thinking of building foo-overlay which will make something like tmp/.../foooverlay-0.10-r0/, then just getting at those files and doing a copy from rootfs_post_process Aug 12 21:41:22 ROOTFS_POSTPROCESS_COMMAND += "opkg ${IPKG_ARGS} install --force thepackage" Aug 12 21:41:26 or whatever the opkg argument is to force the install, i don't recall offhand Aug 12 21:41:28 Thanks Aug 12 21:41:30 er, opkg-cl, not opkg Aug 12 21:41:32 but you get the idea Aug 12 21:41:36 you install it the way the rootfs population does, just forcing it in Aug 12 21:41:36 let me know how that goes Aug 12 21:41:38 you're right, this is probably a fairly common situation, i expect everyone solves it differently, on their own :) Aug 12 21:41:38 Yup, I'll try that out. Shall I post on oe user mailing list? Aug 12 21:41:44 yeah, or better yet, write it up on the wiki and email the link to the list Aug 12 21:41:49 thanks. been a big help Aug 12 21:42:06 sorry i didn't think of that sooner Aug 12 21:42:11 in retrospect its really obvious Aug 12 21:42:13 heh Aug 12 21:44:27 khem, by the way, thanks for taking the time to implement the srctree fixes we discussed, for some reason i never seemed to get around to fixing it Aug 12 21:44:28 I'm an OE newbie... For the-package in the above do I just put foo.overlay or do I need the full file name foo-overlay-xxx-yyy? Aug 12 21:44:43 just the package name, no tthe version, etc Aug 12 21:44:52 sweet. thanks Aug 12 21:45:57 ant_: hello Aug 12 21:46:55 hi Aug 12 21:47:36 kergoth: about bitbake interactive mode.. is there any IPC method ready for this? Aug 12 21:47:49 or even.. how much code is needed? :) Aug 12 21:48:02 it should be pretty trivial to implement Aug 12 21:48:06 what's missing is a UI for it Aug 12 21:48:14 I will be happy with unix socket even Aug 12 21:48:16 (ui in bitbake parlance, the new ui stuff) Aug 12 21:48:25 it's really straightforward Aug 12 21:48:35 a UI is just a python function that acts as a main loop, sending commands and receiving events Aug 12 21:48:44 see bitbake/lib/bb/ui/ Aug 12 21:48:58 knotty is the main one, the default commandline interface, but there's also ncurses and others Aug 12 21:50:03 * Jay7 wishes 48 hours per day.. Aug 12 21:50:22 indeed Aug 12 21:50:36 too long todo :) Aug 12 21:50:47 no doubt Aug 12 21:50:54 plus need some downtime to relax or you go insane Aug 12 21:51:02 yep Aug 12 21:51:49 ant_: any news about initramfs /var overbloat? Aug 12 21:53:28 ah, opkg... Aug 12 21:53:35 grg seem s not around Aug 12 21:54:38 there is a 'fix' on the ML but needs review and 2 acks Aug 12 21:54:53 and iirc Aug 12 21:54:54 is it working properly? Aug 12 21:55:16 ..JaMa had side-effects with the feeds Aug 12 21:55:28 atm I don't care, I applied it Aug 12 21:55:41 will it fix initramfs? Aug 12 21:55:46 yes Aug 12 21:55:52 ok, good.. Aug 12 21:56:09 but il will remove /var/lib/opkg in every image... Aug 12 21:56:19 not only initramfs Aug 12 21:56:31 hm.. a bit insane Aug 12 21:56:41 opkg is insane ;) Aug 12 21:57:06 but we have nothing better :) Aug 12 21:57:11 grg is doing grat battle against the monster Aug 12 21:58:14 is there way to detect initramfs image to remove /var/lib/opkg from there? Aug 12 21:58:33 I thought ONLINE_PACKAGE_MANAGEMENT was the key Aug 12 21:58:41 but isn't respected anymore Aug 12 21:59:26 other bad thing - qemu-system-arm is failing Aug 12 21:59:37 at least with -M akita Aug 12 21:59:39 hm..report to khem :) Aug 12 22:00:05 both my host's qemu and OE's one Aug 12 22:10:52 ant_, it's always been respected but iirc it's never done all that much Aug 12 22:10:56 maybe git log -p shows me wrong tho Aug 12 22:12:33 Tartarus: a patch broke that Aug 12 22:13:02 package_ipk.bbclass: add lists_dir to sysroots' /etc/opkg. conf Aug 12 22:13:12 maybe both :) Aug 12 22:21:58 khem: thx for the new efforts wrt klibc. now libs are staged (even /lib/klibc.so which seems cruft here) Aug 12 22:22:07 klcc still missing in action... Aug 12 22:29:00 ant_: hmmm klcc is missing ? **** ENDING LOGGING AT Thu Aug 12 22:30:20 2010 **** BEGIN LOGGING AT Thu Aug 12 22:56:14 2010 Aug 12 22:56:48 Angelox_123: did you build a full image on the same box ? Aug 12 22:56:53 Angelox_123: sqlite3.OperationalError: attempt to write a readonly database Aug 12 22:57:18 that means some file permission issue on your tmpdir Aug 12 22:57:18 i tried bitbake nano and i get the same error Aug 12 22:57:29 hum Aug 12 22:57:41 is this local disk Aug 12 22:57:49 i am using another partition Aug 12 22:58:31 ant_: let me see how packaged staging packages are generated Aug 12 22:58:43 how to fix this? Aug 12 22:58:46 ideally it should just use the ipk's from deploy Aug 12 22:59:04 Angelox_123: partition on same disk ? Aug 12 22:59:11 or is it some nfs mounted Aug 12 22:59:13 yes Aug 12 22:59:16 same disk Aug 12 22:59:23 that should be ok Aug 12 22:59:54 i need change permissions of any directory? Aug 12 23:00:54 Angelox_123: I would suggest a fresh git clone Aug 12 23:01:06 of oe and bitbake Aug 12 23:01:09 and go from there Aug 12 23:01:12 i cloned 2 minutes back Aug 12 23:01:18 ok Aug 12 23:01:37 what was first thing you did with bitbake Aug 12 23:01:57 trying to run bitbake -b /mnt/hd/stuff/openembedded/recipes/opie-tictac/opie-tictac_1.2.2.bb Aug 12 23:02:02 only for test Aug 12 23:02:04 heh Aug 12 23:02:21 well you need to parse the recipes Aug 12 23:02:36 I would suggest read about OE and bitbake Aug 12 23:02:57 bitbake -b wont parse stuff it will expect that its already done Aug 12 23:03:09 i aready had this problem but i dont remember how i fixed! Aug 12 23:03:19 bitbake opie-tictac Aug 12 23:03:22 will fix it Aug 12 23:03:38 same error Aug 12 23:03:44 ant_: do you have old staged tree somewhere Aug 12 23:04:11 yes Aug 12 23:04:13 mom Aug 12 23:05:03 Angelox_123: you are not running bitbake as root on the box Aug 12 23:05:15 no Aug 12 23:05:39 i am running as oe user ( i created ) Aug 12 23:05:56 03Tom Rini  07org.openembedded.dev * rc5a4936846 10openembedded.git/recipes/squashfs-tools/ (5 files): Aug 12 23:05:56 squashfs-tools: Switch to BBCLASSEXTEND, drop legacy staging, bump PR Aug 12 23:05:56 Signed-off-by: Tom Rini Aug 12 23:05:58 ok Aug 12 23:06:14 Angelox_123: check file permissions on tmp/cache/bb_persist_data.sqlite3 Aug 12 23:06:59 do u want to view my local.conf ? i dont have this file.. Aug 12 23:07:05 ops Aug 12 23:07:07 i found Aug 12 23:07:09 sorry Aug 12 23:07:23 -rw-r--r-- Aug 12 23:07:33 i do : chmod 777 bb_persist_data.sqlite3 ?????? Aug 12 23:07:43 and who is the owner Aug 12 23:07:50 root Aug 12 23:07:51 root Aug 12 23:08:03 well then at some point you ran bb as root Aug 12 23:08:21 yes Aug 12 23:08:27 by accident Aug 12 23:08:29 =( Aug 12 23:08:36 khem, http://filebin.ca/czvvac Aug 12 23:08:39 well I asked you this before Aug 12 23:08:44 and you said no Aug 12 23:08:48 heh Aug 12 23:08:50 sorry..my english is bad Aug 12 23:09:06 now delete your tmp Aug 12 23:09:20 i will try NOW !! thanks by time! Aug 12 23:09:24 and do bitbake opie-tictac Aug 12 23:09:38 i cannot do -b ? Aug 12 23:09:51 btw you will need root to delete your tmp Aug 12 23:10:05 -b does not build dependencies Aug 12 23:10:17 whats the use unless you have a populated build Aug 12 23:10:22 my pc is too old and slow.... Aug 12 23:10:28 khem, the link was renamed: original is http://filebin.ca/czvvac/staging-klibc-c7x0-angstrom-linux-gnueabi_1.5.18-r17.2_c7x0-linux-gnueabi.ipk Aug 12 23:10:36 Pentium 4 1GB Memory Aug 12 23:10:37 then buy a new one Aug 12 23:10:48 I have similar PC Aug 12 23:11:00 i cannot use a existing rootfs to OE install ? Aug 12 23:11:06 and I know its limitations and I know by compiling OE what I am asking for Aug 12 23:11:42 i cannot use a existing rootfs to OE install ? Aug 12 23:12:03 we heard you the first time Aug 12 23:13:10 now i get another error Aug 12 23:13:15 http://pastebin.com/76LDRKB4 Aug 12 23:14:10 trying chmod -R 777 stuff Aug 12 23:14:26 now compiling Aug 12 23:15:09 Angelox_123: You'll need to do sudo rm -rf /mnt/hd/stuff/tmp first. Aug 12 23:15:18 i aready Aug 12 23:15:19 khem: just reminding me of something... :) guys at the company did not have a linux PC, but needed an image till monday, so they let it run on an asus netbook ;) hehe Aug 12 23:15:21 now compiling Aug 12 23:15:24 khem: see the klibc_xyz.so and klibc.so symlink libs were in /lib/klibc/lib and now are in /lib of armv5te sysroot too Aug 12 23:15:28 thanks all help Aug 12 23:15:29 now twice Aug 12 23:15:31 took well, friday till monday or so :) but worked Aug 12 23:16:01 Angelox_123: sudo chown -R oe /mnt/hd/stuff Aug 12 23:16:28 done Aug 12 23:16:33 btw chmod 777 is a bad idea Aug 12 23:16:48 why? Aug 12 23:16:53 ant_: OK Aug 12 23:17:18 Angelox_123: security risk Aug 12 23:17:28 im compiling now Aug 12 23:17:39 thanks khem for all help Aug 12 23:18:03 Jin^eLD: heh did it compile in 2 days Aug 12 23:18:08 or needed another day Aug 12 23:18:27 in computing world we have to follow moore's law Aug 12 23:18:31 khem: i can pause the compile,turn off computer,and re-start ?? Aug 12 23:18:47 we can not keep churning software in 2010 on machines build in 2006 Aug 12 23:18:53 its will be awfully slow Aug 12 23:19:09 Angelox_123: you can do whatever you want Aug 12 23:19:20 make sure you Ctrl+c bitbake Aug 12 23:19:21 khem: well, on monday there was some minor issue, but it was at about 90% of all tasks, so it needed a couple of more hours on top but was ok :) Aug 12 23:19:33 Jin^eLD: heh ok Aug 12 23:19:40 and no smoke or anything :) Aug 12 23:19:58 yeah, the white color of the casing went grey ;) Aug 12 23:19:58 do u know by my machine how time it take to compile? Aug 12 23:20:11 Angelox_123: it depends what you are compiling Aug 12 23:20:13 no just kidding, it survived, I think they used it a couple of times more, until they got the buildserver Aug 12 23:20:27 it could take 24 hours to 3 days Aug 12 23:20:36 Here's a new one to me: Somehow mtd-utils-dev is getting included in my image, and apparently both mtd-utils-dev and linux-libc-headers-dev provide e.g. /usr/include/mtd/inftl-user.h. How can I either (1) figure out where in the chain mtd-utils-dev is getting included and remove it, or (2) remove linux-libc-headers-dev? Aug 12 23:21:06 Xobs: bitbake -g Aug 12 23:21:16 will spit out some .dot files Aug 12 23:21:24 which will list the depchain Aug 12 23:21:25 i am using to compile uclibc...the time takes down ????? Aug 12 23:21:36 Angelox_123: yes a bit Aug 12 23:21:43 about 1.5hrs Aug 12 23:22:09 there are machines where people have 16G RAM Aug 12 23:22:30 and they use rm_work and ramfs to build the beast in RAM Aug 12 23:22:41 takes about 1:30 hrs for 6000 tasks Aug 12 23:22:56 OE uses around 8GB at peak Aug 12 23:23:12 nice.......i m using Slackware 13...i can get any error ? Aug 12 23:23:30 see wiki Aug 12 23:23:39 someone might have some rants about slackware Aug 12 23:23:43 I dont use it here Aug 12 23:23:52 what u use? Aug 12 23:23:59 do you care > Aug 12 23:24:21 I use debian/ubuntu/suse sometimes fedora Aug 12 23:24:22 only to know... Aug 12 23:24:56 personally what is the best(fast) ? Aug 12 23:25:25 for OE ubuntu is easier to get it going Aug 12 23:25:33 it may not be fastest Aug 12 23:25:39 in executing Aug 12 23:25:44 people use gentoo Aug 12 23:25:56 to tune it for machine for best perf Aug 12 23:26:11 but on your machine you will need like 2 days to install gentoo first Aug 12 23:26:14 so dont do that Aug 12 23:26:32 hum.ok..khem this page http://wiki.openembedded.net/wiki/OeFaq dont exists? Aug 12 23:27:50 lol 2551 tasks for my build Aug 12 23:27:57 http://wiki.openembedded.net/index.php/Category_talk:FAQ Aug 12 23:28:29 ant_: I think it should be ok to have klibc in /lib Aug 12 23:29:35 ok Aug 12 23:29:36 Angelox_123, i'm use slackware/x86_64 13.1 Aug 12 23:29:40 works fine Aug 12 23:30:01 i used slackware/x86 13 before that Aug 12 23:30:06 worked fine too Aug 12 23:30:16 grg..what is configuration of u machine? Aug 12 23:30:31 i7, 6gb ram Aug 12 23:30:45 how time takes? Aug 12 23:31:05 ~45 minutes for minimal/minimal-image Aug 12 23:31:24 2-3hours for something with a gui and gui apps Aug 12 23:31:32 like angstrom/gpe-image Aug 12 23:31:33 03Khem Raj  07org.openembedded.dev * r04a9722f7d 10openembedded.git/classes/srctree.bbclass: Aug 12 23:31:34 srctree.bbclass: Also account fot deptask Aug 12 23:31:34 * Use exec_func instead of exec_task Aug 12 23:31:34 * Use a new name for meta task Aug 12 23:31:34 Signed-off-by: Khem Raj Aug 12 23:31:47 forgot to add the ACKS args Aug 12 23:31:50 anyway Aug 12 23:31:53 lol...on my machine it will take days.... Aug 12 23:32:11 =( Aug 12 23:32:35 grg: whats your box's config Aug 12 23:32:55 khem, what do you mean? Aug 12 23:33:29 i thing hes mean console ? Aug 12 23:34:17 i doubt it Aug 12 23:34:40 local.conf ? Aug 12 23:35:44 grg: configuration of the build machine Aug 12 23:36:12 after i run bitbake opie-tictac i run bitbake opie-image i need the bitbake will re-do all tasks ???? Aug 12 23:36:15 after i run bitbake opie-tictac i run bitbake opie-image i need the bitbake will re-do all tasks ???? Aug 12 23:37:25 khem, P6T deluxe motherboard, i7/2.67GHz, 6gb ram Aug 12 23:37:29 Angelox_123: repeating yourself isn't helping anything, and it is annoying people. stop it. Aug 12 23:37:49 sorry..i pressed two times for accident ! Aug 12 23:38:11 grg: hi, any news about /var/lib/opkg ? Can be safely removed? Aug 12 23:38:45 ant_, yep, i don't see why it needs to be present after do_rootfs Aug 12 23:39:06 JaMa once said needs further tests Aug 12 23:39:14 he got broken feeds Aug 12 23:39:21 hmm.. ok Aug 12 23:39:26 grg: cool box is it i7 870 or something ? Aug 12 23:39:46 khem, uhhh.. dunno... i forgot 5 seconds after it was purchased Aug 12 23:39:52 there is new 970 with 6 cores releases recently Aug 12 23:40:11 but 899 is little too much for a CPU Aug 12 23:40:32 /proc/cpuinfo lists 8 procs, model name : Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz Aug 12 23:42:38 yikes, 970 has a *lot* of power Aug 12 23:43:10 a lot of cache Aug 12 23:44:26 hm..khem Aug 12 23:44:28 configure:2607: arm-angstrom-linux-gnueabi-klcc conftest.c >&5 Aug 12 23:44:44 ->/oe/build/tmp/sysroots/i686-linux/usr/armv5te/bin/arm-angstrom-linux-gnueabi-ld: /oe/build/tmp/work/c7x0-angstrom-linux-gnueabi/klibc-1.5.18-r18.0/image/lib/klibc/lib/crt0.o: No such file: No such file or directory Aug 12 23:50:48 ant_: which recipe is giving that issue Aug 12 23:51:05 that's config.log of kexec-tools-.. Aug 12 23:51:47 I seem to have image/lib/klibc/lib/crt0.o Aug 12 23:52:11 I have rm_work Aug 12 23:52:23 ok actually it should have been staged Aug 12 23:52:27 and seems wrong that kexec-tools is peeking there Aug 12 23:53:00 let me rebuild from scratch once more Aug 12 23:53:15 we'll continue tomorrow ;) Aug 12 23:53:36 thanks again for your time Aug 12 23:55:58 ok Aug 12 23:56:04 after i run bitbake opie-tictac i run bitbake opie-image i need the bitbake will re-do all tasks ???? Aug 12 23:56:08 klibc seems like a bad recipe Aug 12 23:56:22 Angelox_123: no Aug 12 23:56:33 you know when one says 'a can of worms' Aug 12 23:56:33 it will only do the tasks which are needed Aug 12 23:56:41 oh Aug 12 23:56:45 thanks again Aug 12 23:56:46 ant_: yeah I think it should be done properly Aug 12 23:56:54 ant_: what recipes use klibc Aug 12 23:57:13 just kexecboot (kexec klibc static) Aug 12 23:57:41 ah, and a couple of initramfs-* images Aug 12 23:58:05 I wonder if we can use it as yet another system C library Aug 12 23:58:08 some moved to uclibc + busybox Aug 12 23:58:26 for initramfs, yes, definitely Aug 12 23:58:28 like uclibc/eglibc/glibc/klibc Aug 12 23:58:43 which means you can only select one of them Aug 12 23:58:56 is that how it is used now Aug 12 23:59:01 iirc there are a lot of missing headers Aug 12 23:59:05 or do you have some recipes use it Aug 12 23:59:21 well if there are images which only include specific recipes + klibc Aug 12 23:59:23 then we are ok Aug 12 23:59:26 I don't think most recipes would build Aug 12 23:59:32 they dont have to Aug 12 23:59:48 but we have images based on klibc Aug 12 23:59:56 true Aug 12 23:59:59 like maybe kexecboot-image Aug 13 00:00:05 or initramfs-image Aug 13 00:00:07 exactly Aug 13 00:00:23 initramfs-kexecboot-image precisely Aug 13 00:00:40 but if klibc needs to live with glibc or uclibc then its a separate story Aug 13 00:00:40 does the cpio for linux-kexecboot kernel Aug 13 00:00:54 oh may be we wont be able to bootstrap a toolchain with klibc Aug 13 00:01:10 hmm Aug 13 00:01:40 thats why one uses uclibc or glibc to bootstrap the toolchain and klcc to wrap around it Aug 13 00:01:42 until now, was a replacement for busybox Aug 13 00:02:15 when one just need simple tools and a shell Aug 13 00:02:18 me needs some snack Aug 13 00:02:29 I'll have to sleep... Aug 13 00:02:34 gnight Aug 13 00:02:42 grg: is 920 good ? Aug 13 00:02:45 thx again and see you tomorrow Aug 13 00:02:50 grg: I am planning on buying one Aug 13 00:02:51 good night Aug 13 00:03:12 grg: how much time does it take for you for a x11-image from scratch Aug 13 00:03:24 khem, i have no complaints. Aug 13 00:03:38 khem, i haven't done an x11-image.. Aug 13 00:03:52 i can kick one off though Aug 13 00:04:51 grg: and is it too noisy Aug 13 00:05:08 the price/performance is really good. Sans monitor/keyboard/etc, the box cost $1200 AUD (~800-900USD) Aug 13 00:05:17 not noisy (yet) Aug 13 00:05:28 the air conditioning in my office is much worse Aug 13 00:05:33 anyone knows americas army?? i can play it running bitbake on my machine ?? Aug 13 00:05:34 heh ok Aug 13 00:05:55 as the fans all get older, i'm sure it will get noisy Aug 13 00:06:04 grg: does it have SATA3 hdd Aug 13 00:06:14 i have it in a "3 hundred" gamer case... Aug 13 00:06:21 i forget what hdd Aug 13 00:06:50 grg: yeah price/perf ratio is key Aug 13 00:07:33 i'll see if i can dig up the invoice to see what was bought... Aug 13 00:12:23 khem, my root dir is on a ST3500418AS, which is 3gb/s sata. But i have my tmpdir mounted on an old ST3500418AS, which is ata/100. Aug 13 00:12:48 oe git repo + sources are on the new drive Aug 13 00:13:07 (source tarballs i mean) Aug 13 00:16:23 Hmm... what are the preferred programs for reading the .dot files? Both Graphviz for OS X and dotty on my Ubuntu VM have been sitting frozen for a half hour trying to open the generated dependency graphs. Aug 13 00:17:19 grg: same split here (2 ST3500418AS) Aug 13 00:17:55 oh. oops. i think i pasted the wrong thing. :) Aug 13 00:18:08 the old ata/100 drive is an ST3160215A Aug 13 00:18:42 it was pulled from my last workstation when it gave up Aug 13 00:18:45 build launched, now zZZ Aug 13 00:19:07 Xobs, the .dot files need to be processed a little first Aug 13 00:19:18 i get error Aug 13 00:19:54 http://pastebin.com/60ryhiJe Aug 13 00:19:56 Xobs, openembedded/contrib/dependsgraph/dependsgraph.sh Aug 13 00:20:32 grg: That makes sense. Wonderful! I really need to remember to check contrib/ more often. Aug 13 00:20:49 paste bin off error here...to any help .... http://pastebin.com/60ryhiJe Aug 13 00:21:34 Angelox_123, that is nowhere near enough information. You just pasted that there is an error Aug 13 00:22:10 Angelox_123, but i think it is related to ccache on your slackware/x86 13.0 system Aug 13 00:22:30 ccache ? Aug 13 00:22:38 Angelox_123, disable ccache in your local.conf Aug 13 00:23:13 actually, from memory that may not be enough. I think i had to move the ccache binary away from /usr/bin/cache Aug 13 00:23:24 otherwise oe still tried to use it Aug 13 00:24:26 http://pastebin.com/UVfW2L9R my local.conf | i think ccache aready disactived Aug 13 00:26:02 i need remove /usr/bin/cache and re-try ?? Aug 13 00:26:33 Angelox_123, mv /usr/bin/cache /usr/bin/cache-is-built-with-bad-options Aug 13 00:26:51 or without the right options. i forget Aug 13 00:27:00 mv: cannot stat `/usr/bin/cache': No such file or directory Aug 13 00:27:04 is ccache ?? Aug 13 00:27:10 yeah, sorry Aug 13 00:27:15 ok Aug 13 00:27:26 Note that we don't use ccache by default now Aug 13 00:27:40 * grg shrugs Aug 13 00:27:51 after this i re try bitbake opie-tictac ? Aug 13 00:28:05 yep Aug 13 00:28:10 tnx Aug 13 00:28:13 trying Aug 13 00:28:55 hmm... maybe it was shasum-native which had the ccache problem Aug 13 00:31:30 grg: Hmm... still need to massage it some more. "dit: width (82517 >= 32768) is too large." Aug 13 00:33:16 Xobs, you could also try bitbake -g -u depexp your-package-here Aug 13 00:34:21 but reading the reduced.dot file manually instead of using dot or the depexp works better for me... your milage may vary Aug 13 00:35:12 grg: bitbake doesn't appear to have a -u option. But looking at the reduced.dot does give me the answer I'm looking for. Aug 13 00:42:45 * kergoth_ chuckles at http://stopwritingramblingcommitmessages.com/ Aug 13 00:43:52 more one error Aug 13 00:44:02 http://pastebin.com/YLW6pVgY Aug 13 00:46:29 Angelox_123, this just says *what* failed. you haven't provided the logs from the failure Aug 13 00:47:34 where i found log ? ( i see linux-libc-headers_2.6.31.bb and dont have do_unpack ) Aug 13 00:50:01 there might be more information if you scrolled up. For instance, something else failing might be causing linux-libx-headers to fail Aug 13 00:50:37 identify the first failing item and look for its log file in its work dir Aug 13 00:50:56 see this: Aug 13 00:50:57 http://pastebin.com/caSDB6iQ Aug 13 00:51:22 dont have command do_unpack into linux-libc-headers_2.6.31.bb... normal ? Aug 13 00:52:41 this doens't look good: Aug 13 00:52:43 # Aug 13 00:52:43 /usr/bin/opkg-cl: /usr/lib/libcrypto.so.0.9.8: no version information available (required by /usr/local/angstrom/arm/lib/libopkg.so.0) Aug 13 00:52:43 # Aug 13 00:52:43 /usr/bin/opkg-cl: symbol lookup error: /usr/bin/opkg-cl: undefined symbol: conf Aug 13 00:53:53 but compiled successfully correct and its for binutils? Aug 13 00:54:22 i give up Aug 13 00:54:56 huh ? Aug 13 00:55:43 i tried run ( bitbake opie-tictac ) and apparently skipped Aug 13 01:08:17 grg: Woot. Image finished. Thanks for your help. Aug 13 01:08:46 no probs Aug 13 01:10:00 i tried run ( bitbake opie-tictac ) and apparently skipped...is normal ? Aug 13 01:13:06 Angelox_123, what skipped? Aug 13 01:14:21 the error of linux kernel Aug 13 01:14:25 http://pastebin.com/caSDB6iQ Aug 13 01:16:33 but ok...now is going fast =) i i am so happy :-) Aug 13 01:18:12 on question: I'm compiling opie-tictac. If after that I compile opie-image so it will build the rest right? Aug 13 01:18:16 one* Aug 13 02:42:29 grg: it's usually better to inherit gettext-native rather than just adding the dep, unless there's a reason to do otherwise.. is there? :) Aug 13 02:58:15 Shine on your crazy diamond.... **** ENDING LOGGING AT Fri Aug 13 02:59:57 2010