**** BEGIN LOGGING AT Thu Jan 03 02:59:56 2008 Jan 03 03:45:59 03xjqian 07org.oe.dev * ra9cf4a7f... 10/ (5 files in 3 dirs): Jan 03 03:45:59 gpsd: add 2.35 Jan 03 03:45:59 * gpsd_dbus.patch has been fixed in upstream svn Jan 03 09:10:25 morning Jan 03 09:22:35 gm Jan 03 11:16:52 03pfalcon 07org.oe.dev * r25fe74a5... 10/ (1 packages/xrestop/xrestop_0.4.bb): xrestop 0.4: Add DESCRIPTION & SECTION. Jan 03 11:16:57 03Paul 07org.oe.dev * r4a6e894f... 10/ (4 files in 3 dirs): opie-taskbar 1.2.3: Fix first-run wizard after install path revamp. Jan 03 11:17:05 03koen 07org.oe.dev * rb56480f5... 10/ (1 packages/ntpclient/ntpclient_2007_365.bb): ntpclient: update to 2007_365 at the authors request Jan 03 11:17:10 03koen 07org.oe.angstrom-2007.12-stable * r70648790... 10/ (1 contrib/angstrom/upload-packages.sh): Jan 03 11:17:10 contrib: upload-packages.sh: adapted superfeed.sh to upload packages for the 2007.12 release Jan 03 11:17:10 * for use with autobuilders with small bandwidth Jan 03 11:19:24 hi all Jan 03 11:19:28 ping mickeyl Jan 03 12:00:02 yo Jan 03 12:03:53 morninh Jan 03 12:03:58 *morning Jan 03 12:06:31 hi CoreDump Jan 03 12:19:40 re Jan 03 13:10:10 are there any good books on DBUS? Jan 03 13:17:35 Afternoon. Jan 03 13:18:06 I'm involved in a project that writes a web browser. Somebody has recently attempted to port it to a Zaurus, but it segfaults when starting up. Jan 03 13:18:29 They're not really in a positition to debug it. Does anybody here have the time and ability to give building it and running it a whirl? Jan 03 13:18:39 (especially through gdb, so we can see where it explodes) Jan 03 13:19:30 rjek: Is it worth testing in qemu? Jan 03 13:19:43 Well, the browser's native platform is ARM-based. Jan 03 13:19:47 So I doubt it's CPU-related. Jan 03 13:20:01 The native platform is ARM/RISC OS, and it certainly functions just dandy on x86/Linux. Jan 03 13:20:35 And Debian build it for all their platforms and there are no bugs filed against it of similar, so I was assuming it must be specific to the Zaurus's OSes and/or hardware Jan 03 13:20:46 And somebody suggested here might be a good place to ask :) Jan 03 13:21:07 * broonie would be that somebody Jan 03 13:22:30 :) Jan 03 13:22:34 Hello broonie. :) Jan 03 13:22:38 rjek: what browser? Jan 03 13:22:48 hrw: NetSurf. http://www.netsurf-browser.org/ Jan 03 13:23:13 If you want to try building it under Linux, svn co svn://svn.netsurf-browser.org/trunk/netsurf/ Jan 03 13:23:26 Make sure you have the dependancies installed that are mentioned in Docs/BUILDING-GTK Jan 03 13:23:31 And then do make -f Makefile.unix Jan 03 13:24:55 that guy who has built it for his Z, I assume he did so via OE? Jan 03 13:25:09 not necessary via oe Jan 03 13:25:48 * XorA hides from bad memories conversation Jan 03 13:27:18 CoreDump: I have no idea. I also have no idea how to ask, given I've never seen a Zaurus. :) Jan 03 13:28:04 ;) so it *might* just be a library mismatch or random toolchain uglyness Jan 03 13:28:27 or eabi problms Jan 03 13:28:42 as 'arm/risc os' is what from oabi came iirc Jan 03 13:28:52 I'll see if I can hack a .bb... Jan 03 13:28:55 RISC OS uses APCS, which even predates OABI. Jan 03 13:29:17 But one assumes when he rebuilt it with his Zaurus-targeting compiler, it used the right ABI. Jan 03 13:29:52 The program launches: it segfaults when it attempts to open a window and fetch a page. It initialises fine. Jan 03 13:30:32 Which points away from ABI issues. Jan 03 13:30:43 rjek: you could always use qemu to debug, that saves you hunting out a zaurus Jan 03 13:30:44 Unless he's copied some dependancy from his RISC OS box or something :) Jan 03 13:31:11 XorA: Is it easily obtained, and does it actually emulate a Zaurus, or does it just emulate similar ARM-based hardware? Jan 03 13:31:19 If the later, it won't tell us anything: it already works on ARM hardware. Jan 03 13:31:47 fast build (I already built it for x86-64) Jan 03 13:32:00 rjek: it is unlikely it is anything zaurus specific, more likely it is EABI or modern glibc specific, in which case qemu will help Jan 03 13:32:19 rjek: we are well ahead of arm debian in library versions AFAIK Jan 03 13:32:27 I suppose I could talk to Wookey to try NetSurf under armel. Jan 03 13:32:44 XorA: Not for unstable, the main thing Debian lacks at the minute is EABI. Jan 03 13:32:56 Which is what armel's for. :) Jan 03 13:33:15 We did have a bit of a hold up with some stuff while preparing the last release but we're still a way off proper freezes now. Jan 03 13:34:52 broonie: hurry up and merge armel then :-D Jan 03 13:35:00 rjek: qemu/arm is best with arm versatilepb emulation Jan 03 13:35:11 * rjek shudders versatile. Jan 03 13:35:22 I have real ARM hardware I could run Debian armel on if it came to that. Jan 03 13:36:01 rjek: I do not know how versatile HW looks and how good/bad it is. Jan 03 13:36:19 rjek: I know that I use versatilePB emulation to test ARMv6 software Jan 03 13:36:24 It's eyewateringly expesnive for no apparent reason. :) Jan 03 13:37:09 netsurf breaks on my website - it is displayed but css are not quite good parsed Jan 03 13:37:43 hrw: If your CSS is correct and NetSurf does not correctly parse it, file a bug. Jan 03 13:37:56 There are plenty of parts of CSS that NetSurf does not implement. Jan 03 13:38:03 Just remember that its x86 binary is less than 500kB. :) Jan 03 13:38:06 rjek: http://blog.haerwu.biz/ Jan 03 13:38:22 -rwxr-xr-x 1 hrw hrw 533K sty 3 14:38 nsgtk Jan 03 13:38:31 533K > 500K :D Jan 03 13:38:35 but its x86-64 Jan 03 13:38:36 Now strip it :-p Jan 03 13:38:41 rjek: stripped Jan 03 13:38:49 was over 2M unstripped Jan 03 13:38:57 Something on your machine bloats it :) Jan 03 13:39:36 8 byte pointers instead of 4 :-) Jan 03 13:39:51 03xjqian 07org.oe.dev * rfd9c39fa... 10/ (5 files in 3 dirs): gpsd: update to 2.36, drop 2.35. Jan 03 13:39:56 03xjqian 07org.oe.dev * r75652b38... 10/ (1 packages/gpsd/gpsd.inc): gpsd.inc: fix staging Jan 03 13:40:46 del.icio.us is unreadabl Jan 03 13:40:47 e Jan 03 13:41:17 It's unreadable in any browser :-p Jan 03 13:41:33 rjek: http://www.macchiato.com/unicode/Unicode_transcriptions.html renders quite good Jan 03 13:41:55 but thats due to pango rather then netsurf itself Jan 03 13:42:03 I'm not actually infront of a machine that can run NetSurf at the moment, so I'm unable to confirm or investigate any of these issues :) Jan 03 13:42:23 no problem for me Jan 03 13:42:39 Quite of a few of the CSS Zen Garden styles render nicely. There's a handful of features we need to implement that'll make most of them look perfect. Jan 03 13:44:24 and some more usability would be nice to have Jan 03 13:44:53 scrolling with space/pgup/pgdn, zooming Jan 03 13:46:45 tabs Jan 03 13:50:05 hrw: It does have zooming. Jan 03 13:50:10 Complete with proper scaling of the images. Jan 03 13:50:21 And yes, I know what would be nice to have. Jan 03 13:51:00 I'm waiting for your patches or your money such that I can afford to work on it full-time :) Jan 03 13:52:11 ;) Jan 03 13:52:15 common problem Jan 03 13:52:34 It's not bad given it's a hobby project of 4 or 5 people. Jan 03 13:55:04 indeed Jan 03 13:55:39 Plus, its native platform is a ballache to develop for. Jan 03 13:55:55 Fastest machine available talke half an hour to build the bloody thing, and there is no debugging tools. Jan 03 13:56:04 No gdb, no electricfense type stuff, etc. Jan 03 13:56:43 03xora 07org.oe.dev * r42ff173c... 10/ (1 conf/distro/include/preferred-om-2008-versions.inc): Jan 03 13:56:43 preferred-om-2008-versions.inc : OM is starting to deviate from angstrom Jan 03 13:56:43 2007.1 so add a file to record these deviations until we are ready to Jan 03 13:56:43 follow angstrom 2008 Jan 03 14:01:39 XorA, why doesn't OM follow the angstrom stable branch? Jan 03 14:02:01 Crofton|home: incompatible release schedules Jan 03 14:02:07 hrh Jan 03 14:03:22 Crofton|home: makes no sense to base a release on old Angstrom when it is more likely 2008 will match our schedules Jan 03 14:03:31 ok Jan 03 14:03:46 Crofton|home: and we already need newer versions of some core software Jan 03 14:03:55 no one can agree on what stable is :) Jan 03 14:04:43 Crofton|home: angstrom stable is a good start for someone bringing a board up from scratch, if GTA01/2 had run on time 2007.1 would probably be OM, as it is everything is later than originally planned Jan 03 14:04:53 2007.12 Jan 03 14:05:20 We will try building gumstix from stable, combined with sakoman gumstix specific collections Jan 03 14:05:44 where his collections contain stuff he didn't like about .dev :) Jan 03 14:29:11 morning Jan 03 14:52:51 03hrw 07org.oe.dev * re470e738... 10/ (8 files in 3 dirs): dbus: added 1.1.1 from Poky (non default) Jan 03 14:53:46 I have a .bb that creates multiple packages. A machine-specific file in installed so BB changes PACKAGE_ARCH to MACHINE for all packages (sigh). In which variable is the proper h/w ARCH hidden so I can change it back? TARGET_ARCH = arm, not armv4t Jan 03 14:57:09 PACKAGE_ARCH_${PN} = "${MACHINE_ARCH}" Jan 03 14:57:47 hrw: I need to set that back to the original ARCH Jan 03 14:57:54 ie: armvt4 Jan 03 14:58:12 it already has been set (wrongly) to MACHINE_ARCH Jan 03 14:58:44 morning Jan 03 14:58:56 mickeyl: hey Jan 03 14:58:58 hi mickeyl Jan 03 15:00:31 CoreDump: tried bitbake -e? Jan 03 15:02:17 hrm Jan 03 15:02:27 FEED_ARCH would be armv4 Jan 03 15:03:03 feed_arch is angstromizm Jan 03 15:03:04 CoreDump: Set SRC_URI_OVERRIDES_PACKAGE_ARCH=0 Jan 03 15:03:07 see base.bbclass Jan 03 15:03:22 RP: perfect, thanks! Jan 03 15:03:55 morning all Jan 03 15:05:12 morning all :) Jan 03 15:05:37 hi hvontres|work Jan 03 15:06:50 morning RP & hvontres|work Jan 03 15:07:01 hi Henry Jan 03 15:28:49 is the openembedded-issues list down? Jan 03 15:32:36 03xjqian 07org.oe.dev * r40bac439... 10/ (6 files in 4 dirs): Jan 03 15:32:36 qpegps: upgrade 0.9.3.1. bump angstrom-preferred Jan 03 15:32:36 * fix DEPENDS and libqte-mt linking close #oebug 3955 thanks to bluelightning Jan 03 15:32:36 * patch 0.9.3.1 Place.* file name bug Jan 03 15:32:36 * fix .desktop QA issue Jan 03 15:32:45 03hrw 07org.oe.dev * r0d7f35f4... 10/ (4 files in 3 dirs): base and sanity bbclasses + qemu: move gcc3 availability checking into base.bbclass (from Poky) Jan 03 15:35:42 would anyone like to be annoyed with #3593? :) Jan 03 15:36:45 !oebug 3593 Jan 03 15:36:47 * * Bug 3593, Status: NEW, Created: 2008-01-02 10:05 Jan 03 15:36:48 * * jin(AT)mediatomb.org: initscripts ignores IMAGE_DEVICE_TABLES and installs it's own table Jan 03 15:36:49 * * http://bugs.openembedded.org/show_bug.cgi?id=3593 Jan 03 15:37:09 hrw: how did the centos thing go yesterday? Jan 03 15:37:18 who won? :) Jan 03 15:37:38 Jin^eLD: I think it was debian :) Jan 03 15:38:05 hvontres|work: no, no, hrw was trying to install CentOS and it did not work repeatedly :) Jan 03 15:39:09 Jin^eLD: I won after midnight Jan 03 15:39:13 :) Jan 03 15:39:33 centos repo is just one year behind any other Jan 03 15:39:39 quite fresh stuff Jan 03 15:40:25 Greetings Jan 03 15:40:57 hi mwester Jan 03 15:41:16 hrv ... use fedora luke !! Jan 03 15:41:29 well, centos is good for long term support Jan 03 15:41:33 so it's ok for servers Jan 03 15:41:41 I would not use it as a desktop system Jan 03 15:41:49 ahhh for a server, so yes i use centos too ! Jan 03 15:42:19 hrw u'll come to fosdem ? Jan 03 15:42:27 Jin^eLD, u'll come to fosdem ? Jan 03 15:42:28 gremlin[it]: no Jan 03 15:42:35 gremlin[it]: no :( Jan 03 15:42:40 damn :( Jan 03 15:42:49 Jin^eLD: BitBake require python-sqlite2 which is not in CentOS Jan 03 15:43:08 gremlin[it]: well, have no budget for it :( and company won't send me Jan 03 15:43:21 hrw: it may indeed be missing a couple of things, yes Jan 03 15:43:28 hrw: but most fedora src rpm will compile on centos Jan 03 15:43:41 Jin^eLD: then it is not centos anymore Jan 03 15:43:43 re zecke Jan 03 15:44:00 hrw: true, but for special cases where you are missing one extra package it's probably still acceptable :P Jan 03 15:44:11 I never tried OE on centos anyway... Jan 03 15:44:24 Jin^eLD: I did not tried OE neither Jan 03 15:44:25 * hvontres|work wonders how many people here still have a "stock" distibution Jan 03 15:44:29 I use Poky on centos Jan 03 15:44:50 hvontres|work: my centos 5.1 install has only two packages from outside and nothing built from source Jan 03 15:45:09 hvontres|work: so it is 'stock' ;D Jan 03 15:45:35 gremlin[it]: my daughter will born in Feb Jan 03 15:46:24 ohhhh .... congratulation !!! Jan 03 15:46:29 thx ;D Jan 03 15:46:52 hrw: exciting times! Jan 03 15:47:25 we will put a pink ribbon on OE booth as soot she born ! Jan 03 15:47:36 soot -> soon Jan 03 15:49:28 hrw: you'll have fun...:) Jan 03 15:49:35 * hvontres|work has two daughters now Jan 03 15:50:12 hrw: better make sure you get all the toys you want now tho.... Jan 03 15:50:35 * gremlin[it] think is time to find a GF and stop have fun with young girls :P Jan 03 15:50:58 gremlin[it]: sounds like you are getting old ;) Jan 03 15:52:23 Jin^eLD, i'm 34 ... Jan 03 15:52:31 Shreyas Srinivasan Jan 03 15:52:38 shres: hey happy new year Jan 03 15:52:46 hi zecke Jan 03 15:53:04 zecke: hey hey, Happy new year to you too dude! Jan 03 15:53:53 zecke: hows it going? Jan 03 15:54:29 shres: fine, only my sleeping schedule is upside down :) Jan 03 15:55:08 zecke: well, bats are nocturnal AND they sleep upside down.. Jan 03 15:55:32 zecke: heh, tell me about it! I just got back from 6 intense days of partying and cant think straight! *hic *hic Jan 03 15:55:33 =D Jan 03 15:56:22 shres: hehe, we had our 24C3 (24th Chaos Communication Congress)... Jan 03 15:57:35 zecke: ah yes, i read harald's review about it. I so wanted to attend the camp which happened this year but unfortunately i just had gotten back from guadec so no leave :( Jan 03 15:58:37 shres: two more years and there is a big camp in the netherlands Jan 03 15:58:51 * shres marks his calendar! Jan 03 15:59:37 i wish the currency conversion didnt stink so bad, everything in europe is super expensive for me! Jan 03 16:00:24 :( Jan 03 16:10:18 shres: we felt the same for a long time ;) Jan 03 16:28:58 florian: heh Jan 03 16:39:33 hey all Jan 03 16:39:58 Should SRC_TARBALL_STASH be the first place tried for fetching of any sort of tarball? Jan 03 16:41:06 no Jan 03 16:41:11 OK Jan 03 16:41:36 Is there a way, aside from a PREMIRRORS_append bbclass? Jan 03 16:41:44 hey mdp Jan 03 16:41:45 ga Jan 03 16:42:38 PREMIRRORS is proper Jan 03 16:43:54 ok. But does it have to be in a bbclass? I tried doing it in the local.conf, but either had a stupid typo, or bad luck, or it just couldn't be done Jan 03 16:44:18 it has to ;( Jan 03 16:44:39 k, thanks Jan 03 16:55:23 bye all Jan 03 20:13:54 re Jan 03 20:28:19 hi florian Jan 03 20:59:47 flo_lap: Hi, it seems https is down on serenity, could you check? Jan 03 21:01:17 psokolovsky: yes sorry, I just started a major upgrade to fix some bad security issues Jan 03 21:02:41 flo_lap: ah, ok, np Jan 03 21:03:50 psokolovsky: its a little bit nasty that the upgrade stopped apache and now waits for mailman queues to become clear Jan 03 21:07:15 morning all Jan 03 21:08:34 hi Fique Jan 03 21:26:55 http://www.ohloh.net/projects/3928?p=OpenEmbedded Jan 03 21:27:13 "Only one single active developer" -LOL Jan 03 21:27:48 hee Jan 03 21:27:54 Mostly written in C/C++ Jan 03 21:27:56 ;) Jan 03 21:28:29 rotfl Jan 03 21:28:31 :) Jan 03 21:29:02 they lack the mtn conduit Jan 03 21:29:02 likewise: they're looking at http://repo.or.cz/r/org.openembedded.dev.git/ Jan 03 21:29:11 maybe we should aim at it the git mirror? Jan 03 21:29:27 I could point them at SlugOS svn stable release repos ... Jan 03 21:29:40 oh, yeah, the git mirror would be better Jan 03 21:29:47 will that make all committers visible? Jan 03 21:31:42 29 person years by one person :) Jan 03 21:34:18 :) Jan 03 21:36:15 03pfalcon 07org.oe.dev * ref2c3679... 10/ (1 contrib/feed-browser/includes/config.inc): contrib/feed-browser: Capure current config.inc from Angstrom site. Jan 03 21:36:21 03pfalcon 07org.oe.dev * re41398d2... 10/ (1 contrib/feed-browser/update.php): contrib/feed-browser/update.php: Allow to update individual feeds. Jan 03 21:36:26 03xjqian 07org.oe.dev * r128beb53... 10/ (1 packages/qpegps/qpegps_svn.bb): add qpegps_svn, close #oebug 2194 Jan 03 21:36:31 03likewise 07org.oe.dev * r0d2aa1a8... 10/ (1 packages/util-linux-ng/util-linux-ng.inc): util-linux-ng: Install binaries into same locations as util-linux does. Jan 03 21:36:38 03pfalcon 07org.oe.angstrom-2007.12-stable * r1512eaa0... 10/ (3 files in 3 dirs): Jan 03 21:36:38 applied changes from 67cc2c9842bf449626180fddd68ba2a4b7a0c519 Jan 03 21:36:38 through 97df34dbfe924823e092b0d579539be286af6cea Jan 03 21:36:38 gpe-login 0.90: Call "C" locale what it is. Jan 03 21:36:39 * Stop using confusing names like "English" for C locale, because people Jan 03 21:36:41 select that "English" and then bang head for hours trying to figure why Jan 03 21:36:43 nothing locale-related works. Jan 03 21:36:45 03pfalcon 07org.oe.angstrom-2007.12-stable * r1369baac... 10/ (1 BACKPORTS.txt): BACKPORTS.txt: Note 97df34dbfe924823e092b0d579539be286af6cea applied. Jan 03 21:39:34 At least he made $1,596,814 Jan 03 21:40:24 I guess we should start a fight who is the one committer then, so that she can cash before the $ drops any lower Jan 03 21:41:23 Hmm - I only made $1,260,370 for nslu2-linux: http://www.ohloh.net/projects/4571?p=NSLU2-Linux Jan 03 21:41:38 * mwester is will to split that sum three ways... anything less than that won't buy a cup of coffee in the US anymore. Jan 03 21:43:10 nslu2-linux is "mostly written in shell script"????!! Jan 03 21:47:05 rwhitby: why not up your salary a bit? :-) Jan 03 21:50:21 flo_lap: just fyi: PHP Warning: [eAccelerator] This build of "eAccelerator" was compiled for PHP version 5.2.3-1+b1. Rebuild it for your PHP version (5.2.4-2) or download precompiled binaries. Jan 03 21:50:21 in Unknown on line 0 Jan 03 21:50:21 PHP Fatal error: Unable to start eAccelerator module in Unknown on line 0 Jan 03 21:51:09 * flo_lap curses Jan 03 21:51:43 I still wonder why we had Debian unstable in sources.list Jan 03 21:52:02 flo_lap: that's for command-line php, web one seems to work Jan 03 21:52:52 psokolovsky: yeah, but mysql and mailman both failed to restart Jan 03 22:17:04 psokolovsky: ok, web is back Jan 03 22:20:18 re Jan 03 22:21:04 hrw: wb Jan 03 22:42:21 8GB for build is small size Jan 03 23:01:45 Khem, ping Jan 03 23:02:45 hi thesing Jan 03 23:02:54 thesing: hey there :) Jan 03 23:03:07 hi everybody Jan 03 23:04:01 thesing: how life? Jan 03 23:04:38 hrw: ? Jan 03 23:04:49 thesing: could you take a look at http://bugs.openembedded.org/show_bug.cgi?id=3599 ? I am not too familiar with SD errors Jan 03 23:06:31 hrw: I'm fine, thanks (if this was the question). Jan 03 23:06:48 hvontres|work: probably timeouts are wrong Jan 03 23:07:03 thesing: that was Jan 03 23:09:28 hvontres: I don't know how mmc cards are connected on poodle. But I think hrw is right. Some cards are quite particular with timing. Jan 03 23:10:33 hvontres|work: http://bugs.openembedded.org/show_bug.cgi?id=3582 - can you? Jan 03 23:10:46 thesing: poodle use pxamci Jan 03 23:11:44 I ported nearly all patches to 2.6.24rc. Jan 03 23:12:14 thesing: when you plan to send them to LAKML for review? Jan 03 23:12:27 * hrw wants Zaurus patches upstream... Jan 03 23:13:02 thesing: can you add descriptions in patches? From:/Signed-Off-By: lines? Jan 03 23:13:13 others too ;d Jan 03 23:14:00 hrw: Some patches are not from me and are quite ugly. Jan 03 23:14:32 thesing: but you use them anyway Jan 03 23:14:54 if we will not send patches upstream they will bitrot Jan 03 23:14:54 hrw: I don't know where they came from. I will send my locomo-spi to spi-devel when it works again. Jan 03 23:15:18 thesing: are they useful? if yes - send to lakml Jan 03 23:15:29 hrw: I included them because they were there. I didn't look at every patch in detail yet. Jan 03 23:15:57 ok Jan 03 23:16:40 hrw: from first view some might be dropped. Jan 03 23:17:09 I spent few hours in last week just on updating 'unknown' patches to apply to newer kernel Jan 03 23:18:01 good that at least on laptop I do not have to use non-vanilla kernel Jan 03 23:18:16 hrw: That work is no fun. It hard enough to port own patches to newer kernels ;) Jan 03 23:18:56 on monday I will start updating kernel for client board (which I do not have). this will be fun... Jan 03 23:19:14 2.6.16.x probably + ugly patchset or whole tree Jan 03 23:19:29 and now I go to bed ;) Jan 03 23:19:35 At least its already 2.6 Jan 03 23:19:46 Ok. Good night. Jan 03 23:19:50 thesing: with closed modules for half of cpu Jan 03 23:20:25 bye all Jan 03 23:29:47 Crofton|home: pong Jan 03 23:38:07 Khem, some guy is apparently trying to build omap2430sdp and having an issue with the kernel build Jan 03 23:38:36 maybe having to do with SRCREV ..., I am running a build to see if I can reproduce the problem Jan 03 23:39:13 Hvae you done any builds for that lately? Are there any tricks people should be aware off? Jan 03 23:42:03 hrw|gone: already fixed in 2.6.23. Jan 03 23:49:49 psokolovsky: ping Jan 03 23:50:00 thesing: pong Jan 03 23:51:06 I'm trying to build initramfs-module for collie. Jan 03 23:51:39 thesing: ok Jan 03 23:51:54 But do_rootfs fails because it can't satisfy the dependencies update-modules. Jan 03 23:52:29 I didn't find and update-modules package in oe. Jan 03 23:52:38 thesing: what exactly do you bitbake? Jan 03 23:52:48 initramfs-image Jan 03 23:54:28 but initramfs-minimal-image doesn't work also because Jan 03 23:55:51 thesing: last time I build initramfs-bootmenu-image mostly Jan 03 23:56:32 thesing: and update-modules is right in OE, and any kernel module has dep on it Jan 03 23:56:44 I want an image that only mount jffs2 partion and kexecs a new kernel from there. Jan 03 23:57:22 03tmbinc 07org.oe.dreambox * r6280190b... 10/ (3 files in 3 dirs): tuxbox-tuxtxt-32bpp: properly handle console container failure conditions Jan 03 23:57:31 03coredump2 07org.oe.dev * r0475d495... 10/ (4 files in 3 dirs): Jan 03 23:57:31 openmoko-gps: Update to latest version and add new icon. Jan 03 23:57:31 mtN: ---------------------------------------------------------------------- Jan 03 23:57:37 03coredump2 07org.oe.dev * ra925a41d... 10/ (3 files in 3 dirs): openmoko-apps: openmoko-gps: Fix icon name Jan 03 23:57:42 initramfs-minimal-image fails because there is no buildable provider for initramfs-nfsboot. Jan 03 23:59:07 thesing: I'll drop *-minimal-image Jan 03 23:59:41 thesing: otherwise, one reason you don't have update-modules in image is that you don't have any kernel module there. Not good ;-) Jan 04 00:00:30 or maybe not drop, it's sample for PACKAGE_INSTALL_NO_DEPS Jan 04 00:01:11 But I don't need any kernel module in the image. Everything this image will need is compiled statically into the kernel. Jan 04 00:02:29 thesing: well, then make it work for you. but initramfs is intended to be generic thing, and you cannot even think about generic with static kernel. Jan 04 00:05:58 thesing: changing "remove ${PACKAGE_REMOVE};" into "remove ${PACKAGE_REMOVE} || true;" would be quick workaround Jan 04 00:06:42 An other thing: is there already a way to include the initramfs-image into a kernel in oe? Jan 04 00:08:15 thesing: not that I know off. again, anti-generic. but would be nice too ;-). would require 2-stage kernel build process, or using external util altogether... Jan 04 00:12:15 with no modules include in initramfs it wouldn't even need a 2-stage build process. Jan 04 00:13:31 btw. I was able to build initramfs-image after I build module-init-tools and update-modules manually. Jan 04 00:13:33 thesing: well, good catch, but *unnnn*generic ;-) Jan 04 00:14:21 thesing: also, external util won't help it seems, image is linked into kernel by the linker. (well, so ld could be used by such utils, but that's too hacky ;-) ) Jan 04 00:15:16 Well the first stage bootloader-linuxkernel will be device specific anyway. Jan 04 00:15:29 thesing: hm, strange. well, as you didn't post error log, I just assumed that it fails trying to remove non-existent packages from *image* Jan 04 00:15:49 thesing: *can* be, only can ;-) Jan 04 00:16:21 thesing: if you have device-specific drivers compiled in, and generic driver as modules, it won't be ;-) Jan 04 00:16:31 that's semi-generically of course ;-) Jan 04 00:17:33 True. But the location where to load the real kernel from will be device specific. Jan 04 00:17:41 ~pastebin Jan 04 00:17:42 [~pastebin] A "pastebin" is a web-based service where you can paste anything over 3 lines without flooding the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://www.rafb.net/paste Jan 04 00:18:41 thesing: well, why? initramfs-bootmenu-image allows user to decide ;-) Jan 04 00:21:02 but its probably too big. initramfs + kernel < 1048576 KB for collie Jan 04 00:21:12 not KB but B Jan 04 00:23:00 psokolovsky: http://rafb.net/p/EfbTqV36.html is the important part of the errorlog. Jan 04 00:23:01 thesing: well, device limits are another matter. it would be big, because it uses shell, so need busybox. you also either use shell, or do adhoc C hacking. Jan 04 00:25:38 thesing: well, weird. but it always can be .bbclasses' missing dependencies... Jan 04 00:25:57 thesing: you make fresh build as I understood? Jan 04 00:27:15 .. or the fact that you didn't use "-c buildall" for bitbake ... Jan 04 00:28:20 nearly fresh. I build kernel and nano before.. Jan 04 00:29:32 should I use "-c buildall" ? Jan 04 00:30:29 thesing: well, I can't see sth wrong in initramfs-* recipes. e.g., *-nfs pulls update-modules, so you likely have kernel module for nfs after all. the matter, somehow update-modules *ipk* is not written when its needed already ;-( Jan 04 00:31:02 thesing: yes, for any issues that you don't have ipk's for the built packages, -c buildall should fix that Jan 04 00:34:36 03pfalcon 07org.oe.dev * r662671ac... 10/ (1 packages/images/initramfs-minimal-image.bb): initramfs-minimal-image: Revamp dependencies. Jan 04 00:35:40 Thanks for your help. Jan 04 00:38:30 thesing: I'm trying to utter it all in the mail to RP, maybe I'll be able to describe it, and he'll be able to suggest where to fix it ;-) Jan 04 00:40:33 running iptables gives me "getsockopt failed strangely: No such file or directory" Jan 04 00:40:43 any ideas what I might be missing? oe on gumstix Jan 04 00:41:26 maybe device files are missing? Jan 04 00:43:56 I'll try to find what dev files it needs Jan 04 00:46:25 if normal networking works you should have the needed device files. Jan 04 01:50:49 thesing: yes normal networking does work **** ENDING LOGGING AT Fri Jan 04 02:59:56 2008