**** BEGIN LOGGING AT Mon Apr 30 02:59:56 2007 Apr 30 04:20:36 03lenehan 07org.oe.dev * re1f2e353... 10/ (5 files in 4 dirs): (log message trimmed) Apr 30 04:20:36 perl 5.8.8: Improvements for cpan modules: Apr 30 04:20:36 * Modify perl to install a copy of it's configuration during staging. This Apr 30 04:20:36 will allow us to get at the perl settings for the target when building Apr 30 04:20:36 cpan modules. Apr 30 04:20:36 * Modify perl-native to allow selection of the host or target configuration Apr 30 04:20:38 based on an environment variable. This will allow the cpan class to select Apr 30 04:20:44 03lenehan 07org.oe.dev * r4ae75f1e... 10/ (1 packages/perl/libxml-parser-perl_2.34.bb): Apr 30 04:20:44 libxml-parser-perl: Add a dependency on expat-native and pass in the Apr 30 04:20:46 appropriate libarary and include directories for exapt to the Apr 30 04:20:48 makefile. Apr 30 04:20:50 NOTE: This still doesn't work since it tries to use the host headers Apr 30 04:20:53 when linking to perl. Fixing this properly will need "perl" to be Apr 30 04:20:54 installed in the staging area (currently we stage perl-native only). Apr 30 04:35:03 03lenehan 07org.oe.dev * r9fca9729... 10/ (1 packages/quagga/quagga_0.99.7.bb): Apr 30 04:35:03 quagga 0.99.7: Add latest release. This includes some important BPG DOS Apr 30 04:35:03 fixes and OSPF configuration releated fixes. Apr 30 04:42:44 03lenehan 07org.oe.dev * r2fa2932e... 10/ (1 packages/quagga/quagga.inc): Apr 30 04:42:44 quagga: Use DEBIAN_NOAUTONAME to stop the debian class renaming the Apr 30 04:42:44 packages instead of manually naming them with their original name to Apr 30 04:42:44 stop it. Apr 30 06:19:45 I'm using BBFILE_COLLECTIONS and the configured priority is working fine for .bb's but not for .conf files. Is that expected? Apr 30 07:48:44 03lenehan 07org.oe.dev * re01aa57a... 10/ (6 files in 3 dirs): (log message trimmed) Apr 30 07:48:44 rp-pppoe: Add version 3.8. Apr 30 07:48:44 * The source for version 3.5 hasn't been available for a long time. Apr 30 07:48:44 * Make sure pppoe is installed suid root. Apr 30 07:48:44 * Always try to reconnect, pppd handles stop/starting pppoe. This prevents Apr 30 07:48:44 lots of annoying problems with pppoe totally giving up (debian patch) Apr 30 07:48:46 * Accept PPPoE packets from multicast addresses (debian patch) Apr 30 07:56:53 morning all Apr 30 07:57:19 morning Apr 30 07:59:00 I'm using BBFILE_COLLECTIONS and the configured priority is working fine for .bb's but not for .conf files. Is that expected? It's driving me nuts here :\ Apr 30 08:01:00 morning Apr 30 08:01:26 hi XorA Apr 30 08:03:30 good morning all Apr 30 08:03:51 yo koen CoreDump|home Apr 30 08:14:43 for future reference: It appears that BBFILE_COLLECTIONS doesn't affect .conf files at all. BUILDIR must point at the propper configs Apr 30 08:31:38 03koen 07org.oe.dev * re29c9d52... 10/ (1 packages/loudmouth/loudmouth_1.2.2.bb): loudmouth: update to 1.2.2 Apr 30 08:45:04 03koen 07org.oe.dev * r3e584182... 10/ (1 packages/linux/linux-gta01_2.6.20.bb): linux-gta01: switch URI to 2.6.20 branch for patches Apr 30 09:06:03 03koen 07org.oe.dev * rfc08ef3a... 10/ (4 files in 3 dirs): Apr 30 09:06:03 linux-gta01: add 2.6.21.1 kernel Apr 30 09:06:03 * NOTE: this is a cleaned and fixed up version of the one in the openmoko svn repo Apr 30 09:26:19 hmmm minimum bootstrap-image + kernel is too big for a 4MB flash Apr 30 09:26:29 * Ifaistos wonders why.... Apr 30 09:26:58 morning Apr 30 09:27:06 Ifaistos: use task-boot instead of task-base Apr 30 09:27:09 hey hrw Apr 30 09:27:26 hi hrw Apr 30 09:35:10 03koen 07org.oe.dev * rf5dbd0eb... 10/ (1 packages/images/angstrom-minimal-image.bb): angstrom-minimal-image: add an image for people that needs something smaller as angstrom-console-image Apr 30 09:35:22 Ifaistos: try that one Apr 30 09:40:12 heh Apr 30 09:40:15 http://www.oesf.org/forums/index.php?showtopic=23865&view=findpost&p=160096 Apr 30 09:41:10 koen : my task-base is rather old r20, for task-boot to work does any of the classes need update also ? Apr 30 09:41:30 dunno Apr 30 09:42:41 let's copy and pray.... Apr 30 09:56:45 first try failed -> ERROR: Nothing provides dependency task-boot Apr 30 09:58:53 speaking of dependencies, anyone besides zecke doing bitbake and interested in those not-added-automatically deps? Apr 30 10:00:16 polyonymous: show it Apr 30 10:00:49 hrw, when building angstrom from scratch, it, for instance, failes with ipkg-cl: command not found and fakeroot: command not found. Apr 30 10:01:01 hrw: http://rafb.net/p/4Q4ObM33.html Apr 30 10:01:19 (that's the fix) Apr 30 10:01:29 talking of bitbake, is anybody else having problems w/ 1.8 just hanging on the first task? Apr 30 10:01:54 Hmm.. Well, I thought my problem is that I'm using 1.6 :) Apr 30 10:02:20 Well, it's a matter of choosing problems that you prefer, of course :) Apr 30 10:02:41 I think I'll have to /part /join - something's wrong with xchat... Apr 30 10:02:53 hmm.. Apr 30 10:03:12 well Apr 30 10:03:15 better now. Apr 30 10:03:23 koen : If i use MACHINE_TASK_PROVIDER = "task-base" then it builds ok -> http://rafb.net/p/nxVPNT28.html Apr 30 10:03:46 koen : but if i change that to task-boot it fails Apr 30 10:07:13 polyonymous: don't use bitbake 1.6.x Apr 30 10:07:21 OE in the press: http://www.notacloud.com/blog/?p=32 Apr 30 10:07:47 koen, I'm planning to move on to 1.8, but since I was there anyway and it's not officially deprecated, I thought I'd fix it first. Apr 30 10:10:08 here he is Apr 30 10:10:16 hey zecke Apr 30 10:10:18 zecke, http://rafb.net/p/4Q4ObM33.html Apr 30 10:11:20 polyonymous: bitbake 1.6.x is essentially dead, since 1.8.x features a rewrite of core parts Apr 30 10:11:50 polyonymous: that is a patch to 1.8? Apr 30 10:11:58 koen, I do not argue with that. Apr 30 10:12:03 zecke, no, to 1.6 Apr 30 10:12:34 Actually, the function in question in base.bbclass is something 'oldbitbake', so, I dunno of the problem applies to 1.8 at all Apr 30 10:13:06 I'll go for 1.8 now, just want to see if fresh build with 1.6 will work Apr 30 10:13:16 polyonymous: I will be back ah ~15h Apr 30 10:13:34 zecke, have fun. but by that time I may well be at 1.8 :)) Apr 30 10:13:55 polyonymous: the patch looks sane Apr 30 10:14:12 polyonymous: but it can only make a difference if we have some sort of inherit cycle (do we have it?) Apr 30 10:14:13 zecke, good, because I'm not familiar with bb and even with python :)) Apr 30 10:14:19 zecke, we do. Apr 30 10:14:28 that's about indirectly inherited rootfs_ipk Apr 30 10:14:42 ah okay, good catch then :) Apr 30 10:15:14 yup. dunno if it applies to 1.8, we'll see later today :) Apr 30 10:20:12 the fix seems to fix all build problem for what I can see so far. We'll see if the build completes without introducing ew problems :) Apr 30 10:24:59 koen : The smallest i could get is around 2.7MB -> http://rafb.net/p/uZlQMX98.html Apr 30 10:25:28 and that's using squashfs-lzma Apr 30 10:25:46 that's minimal-image? Apr 30 10:27:25 i used the one Clif mentions in his blog ->http://rafb.net/p/i7LDkZ58.html Apr 30 10:33:12 koen : but the size differences are big... will have to compare what it's included in both cases Apr 30 10:33:35 try the angstrom-minimal image Apr 30 10:33:58 does not "compute" on my image :( Apr 30 10:34:48 http://rafb.net/p/phjWgn64.html Apr 30 10:35:35 this is from my svn repo, which was laste synced with OE about a month ago (maybe more) Apr 30 10:35:35 Ifaistos: your OE is over a month old Apr 30 10:35:46 :) Apr 30 10:38:03 koen : I am afraid to sync it to the current OE... probably need to see what has changed Apr 30 10:38:28 Ifaistos, can't you revert to this revision later on? Apr 30 10:38:30 i synced classes,images,tasks and bitbake.conf Apr 30 10:42:19 polyonymous : yes i can... that's the idea behind source control ;) the problem is it takes time to find out why things don't work as they used to once you sync Apr 30 10:42:48 Ifaistos, that's also the point of distributed development :))) Apr 30 10:44:25 polyonymous : as i learned very early on my work with OE, once you find something that works, keep it :) You never know when the next time will be :) Apr 30 10:44:58 polyonymous : dev branch is a fast moving target Apr 30 10:45:28 Ifaistos, of course. You can keep the revision id, though :) Also, I think I should learn more about monotone, probably you can "keep" tag and branch for your modifications (I take it it's pretty much like git). Apr 30 10:45:36 03koen 07org.oe.dev * r7546d663... 10/ (1 packages/images/angstrom-minimal-image.bb): angstrom-minimal-image: fix up for situation where the ssh daemon isn't defined in local.conf Apr 30 10:45:52 But I don't follow this paradigm, I prefer bleeding edge and fix stuff if I can. Apr 30 10:47:59 polyonymous : oh well its like religion,politics, favourite distro etc. everyone has a different opinion... we are using svn here Apr 30 10:48:43 Ifaistos, no, it's not. You're totally right and I admit I'm not following it deliberately looking for troubles :) Apr 30 10:49:07 :) Apr 30 10:49:16 Or if you were talking about the choice of VCS, then it is, indeed, but I wasn't discussing it. Apr 30 11:11:06 koen : http://rafb.net/p/jlSddm16.html Apr 30 11:11:51 pretty close Apr 30 11:16:58 Hi all, how do I disable insane.bbclass? Apr 30 11:17:29 trying to do a fresh build.. Apr 30 11:17:45 you don't, you fix the bugs insane.bbclass shows you Apr 30 11:18:14 I think it's pretty hard to fix the gcc rpath issue... Apr 30 11:18:37 I went for http://rafb.net/p/YJi4tU90.html :-( Apr 30 11:18:57 polyonymous: the patch to do that is in the bugtracker Apr 30 11:19:11 the patch to insane? Apr 30 11:19:27 !oebug 2140 Apr 30 11:19:29 * * Bug 2140, Status: RESOLVED (FIXED), Created: 2007-04-25 14:48 Apr 30 11:19:30 * * likewise(AT)gmx.net: RPATH of shared libraries contains host path, causing overhead Apr 30 11:19:31 * * http://bugs.openembedded.org/show_bug.cgi?id=2140 Apr 30 11:20:24 aha.. Seen that, haven't applied yet. Apr 30 11:20:31 Thanks for pointing to it once more. Apr 30 11:21:40 I prefer having local patches that are really local, but on the other hand I'm not good at having strict policies :) Apr 30 11:26:14 hello all Apr 30 11:28:39 hey likewise Apr 30 11:29:13 hi koen Apr 30 11:30:32 likewise hi Apr 30 11:30:45 koen ; Anything above 2.9MB is too big Apr 30 11:31:48 My girlfriend thinks otherwise - excuse me, what was the question? :-) Apr 30 11:32:38 :)))))))))) Apr 30 11:34:10 likewise : heh regarding *flash* sizes.... not flesh :) Apr 30 11:34:52 flesh nor flash probably Apr 30 11:35:17 likewise : its another case of "size matters" Apr 30 11:35:44 Well, it's like with nose - not the size, but what's inside. Apr 30 11:39:44 * Ifaistos tries uclibc for size Apr 30 11:40:41 koen, I don't understand this bug. I mean I don't understand why is it closed and patch isn't in the tree. Apr 30 11:41:03 ask likewise :) Apr 30 11:41:30 likewise, I don't understand this bug. I mean I don't understand why is it closed and patch isn'tin the tree. ;-)))) Apr 30 11:41:44 polyonymous: which bug? (I just logged on here) Apr 30 11:41:49 ah, yes, 2140 Apr 30 11:43:13 polyonymous: ok, the rpath detection was faulty in insane.bbclass. that is now fixed. I suggest that further RPATH issues get one bug report per package. Apr 30 11:44:04 polyonymous: I'll add a bug reference in bugzilla: #2146, #2147, #2152, #2160 Apr 30 11:44:16 ah Apr 30 11:44:42 likewise, but the patch from the bug isn't in the tree yet? Apr 30 11:45:21 And 'resloved', basically, means "forget about this bug - it's history". But koen still have to point at the bug for resolution. That's what I mean. Apr 30 11:46:28 ok, I see what you mean. However, the patch will not work for most host/target combo's. chrpath does not respect targets != host architectures. Apr 30 11:46:56 Well, that would justify why it's not in the tree, but not why the bug is closed :) Apr 30 11:47:32 Anyway, I'm not trying to force you into anything, just stating my opinion. Apr 30 11:47:57 polyonymous: you are right, I'll add this information, I'ld rather see a neat solution (rpaths fixes in package, or a works-always workaround). Apr 30 11:48:58 work-always-around :) Well, I've tried to look into the fixing gcc rpath issue and thought it's a bit too complex for me. The .bb part seems to be more complicated than gcc itself :) Apr 30 11:52:39 polyonymous: I tried it as well: first run bitbake gcc-cross -c configure then grep if any bad (r)paths appear in the config/make files, etc etc, step by step, but I could not track them down. I need a few hours to sit on this sometime. Apr 30 11:56:10 03koen 07org.oe.dev * r09678265... 10/ (1 conf/checksums.ini): checksums: remove xorg checksum till we find out which is the right one Apr 30 11:56:14 03koen 07org.oe.dev * r18ca2479... 10/ (5 files in 4 dirs): gamin: add 0.1.8 Apr 30 11:56:19 03koen 07org.oe.dev * r7ffa5d74... 10/ (1 packages/gamin/gamin_0.1.8.bb): gamin: inherite pkgconfig Apr 30 11:56:23 03koen 07org.oe.dev * rca317b0a... 10/ (5 files in 4 dirs): tracker: add 0.5.4 Apr 30 12:00:17 tranks, koen for tracker Apr 30 12:00:24 np Apr 30 12:04:38 with fixed bitbake 1.6 build takes much longer, guess there were some dependencies missing that weren't fatal :) Apr 30 12:11:11 likewise, isn't that something to do with LDFLAGS_FOR_TARGET ? Apr 30 12:11:59 polyonymous: yes, I've fiddled with that patch as well. I've tried @TARGET LDFLAGS@ instead of @TARGET LDFLAGS@, no luck. Apr 30 12:12:09 polyonymous: yes, I've fiddled with that patch as well. I've tried @TARGET LDFLAGS@ instead of @LDFLAGS@, no luck. Apr 30 12:13:28 likewise, yeah, that's a bit tricky to track down, but I think that's the way to go. Maybe I'll have time to play with it, but for now I just want to make a semi-stable image before I go to Moscow for two weeks. Don't feel like digging deeper now... Apr 30 12:14:05 polyonymous: then simply do not inherit insane.bbclass would be my best bet. Apr 30 12:14:29 likewise, at this point I just commented out 'return False' in rpath check. Apr 30 12:15:09 polyonymous: yes, ok Apr 30 12:15:44 great, fixed bitbake-1.6 build is complete. Now I'll switch to 1.8 to face brand new problems (which is directly against my idea of getting things working at the least cost:)). Apr 30 12:17:32 2h10m build :) Apr 30 12:17:43 03koen 07org.oe.dev * rc543d011... 10/ (1 packages/linux/linux-gta01/defconfig-2.6.21.1-fic-gta01): linux-gta01 2.6.21.1: enhance defconfig a bit Apr 30 12:35:22 koen: http://www.nu.nl/news/1060599/22/Snelweg_San_Francisco_stort_in_door_ongeval_met_truck.html Apr 30 12:35:55 wow Apr 30 12:55:23 morning all Apr 30 12:55:38 hi RP Apr 30 13:03:39 hi all Apr 30 13:03:52 hey RP Apr 30 13:04:55 hey RP Apr 30 13:05:39 RP: fixed the compact flash serial port problem while you were AFK :-D Apr 30 13:12:33 XorA: Ah, how did you fix it? Apr 30 13:13:03 hey Apr 30 13:13:31 XorA: http://rafb.net/p/LP5nGM60.html Apr 30 13:13:35 zecke: wb Apr 30 13:13:58 RP: HWUART was taking the minor that we used for first 8250 ttySX device, so I moved them all up by one Apr 30 13:14:11 morning Apr 30 13:14:20 hey chouimat Apr 30 13:15:03 XorA: So you shifted the serial console? Apr 30 13:15:39 RP: no, the 8250 minor numbers Apr 30 13:15:48 XorA: ah, thats ok that :) Apr 30 13:15:50 RP: so it moved from ttyS3 ttyS4 Apr 30 13:16:12 RP: and I fixed the dmesg output to actually say the right ttySX when 8250 card is plugged in :-) Apr 30 13:16:19 XorA: cool :) Apr 30 13:17:52 RP: the dmesg for the device caps still prints the wrote tty but I didnt want to start messing with the tty layer to kludge tat Apr 30 13:17:53 s/tat/that/ Apr 30 13:17:57 RP: have you seen the patch from polynomous? Apr 30 13:20:26 XorA: I wonder if that YUV code would help with Ximageon as well Apr 30 13:20:53 koen: it should do Apr 30 13:43:17 zecke: I don't think so Apr 30 13:43:22 XorA: Fair enough Apr 30 14:05:25 When building from scratch with 'bitbake gpe-image' I get a failure 'X11/libX.h' not found on compilation of xdamage package. If I then 'bitbake -c clean virtual/libx11' and then restart build with 'bitbake gpe-image' then it reinstalls diet-x11 and continues happily. Any ideas how I can debug this problem? Apr 30 14:11:28 vervain: I wonder if its something to do with x11 being used instead of diet-x11? Apr 30 14:18:22 03likewise 07org.oe.dev * r548157e6... 10/ (3 files in 3 dirs): linux-efika-2.6.19-rc6.bb: Bring patch local, was no longer available online. Apr 30 14:19:15 likewise: what's the status of th efika in 2.6.21? Apr 30 14:19:29 koen: nice you ask, I was just browsing the web for it :-) Apr 30 14:19:44 koen: wanted to build a real-time 2.6.21 kernel for efika, actually. Apr 30 14:19:58 likewise, please do :) Apr 30 14:20:03 low latency? Apr 30 14:20:05 we don't want to cheap taiwanese time Apr 30 14:20:10 s/to/no/ Apr 30 14:20:27 RP: Sure - I dunno, like I say I'm trying to debug it... however to debug I have to start a clean build and use an other few cpu hours. Apr 30 14:20:29 Crofton|home: yes, the -rt patch. ~6us latencies on x86. Apr 30 14:20:31 * chouimat is listening to Fade To Black by Sonata Arctica on Takatalvi [Amarok] Apr 30 14:20:32 oops Apr 30 14:20:38 chouimat: :-) Apr 30 14:21:13 * chouimat remembers that today is the monthly fire alarm test :( Apr 30 14:21:14 the gnu radio guys report they can use ridiculously small buffers with that kernel Apr 30 14:21:28 vervain: That was just a hint as I think I've seen that problem too Apr 30 14:21:51 RP: I got my original clue from the DEPENDS from xdamage Apr 30 14:22:56 from the xdamage bb file I meant. Apr 30 14:25:52 Crofton|home: where do they talk on this? gnu radio ml? Apr 30 14:32:28 koen, Crofton|home: http://www.powerdeveloper.org/forums/viewtopic.php?p=7735#7735 - no big things missing, it *seems*. Trying a build. Apr 30 14:33:01 * koen reads Apr 30 14:35:20 * koen gets angry at mesa for try to feed x86 asm into a powerpc compiler Apr 30 14:35:41 heh -> 1) a package for every distro on x86 we care about - it contains the compiler/tools required Apr 30 14:36:03 ) a package for the Efika which turns on the cross-compiling and distribution client, and one which makes the Efika act as a server in case you have 40 Efikas laying on a desk on the network. Apr 30 14:36:20 3) a prebaked virtualpc and vmware image Apr 30 14:36:33 wow Apr 30 14:36:49 any takers ? :) Apr 30 14:37:15 Ifaistos: you can send me the 40 eifkas :) Apr 30 14:38:03 chouimate : NP. Send me an email with your address and i will send you the bank account to make the deposit :) Apr 30 14:38:22 hehe Apr 30 14:40:10 huhu Apr 30 14:43:52 heya Jin^eLD Apr 30 14:43:58 hey likewise Apr 30 14:44:02 CosmicPenguin: hi Apr 30 14:44:15 morning Apr 30 14:49:09 left uclibc without compiling for a while and now he/she/it had forgot how to compile correctly :( Apr 30 14:49:34 libm/powerpc/s_ceil.c:78: error: impossible constraint in 'asm' Apr 30 14:52:45 Ifaistos: bi section approach between last know good date and now? Apr 30 14:56:53 03koen 07org.oe.dev * r5613d967... 10/ (1 site/arm-common site/arm-linux): site/arm: updates from poky Apr 30 14:56:57 03koen 07org.oe.dev * r139f3317... 10/ (1 packages/glib-2.0/glib.inc): glib: merge printf fix from poky Apr 30 15:04:40 hi Apr 30 15:06:04 hi Bernardo Apr 30 15:07:22 hi hrw Apr 30 15:07:42 I've seen the emails on moving openzaurus to angstrom Apr 30 15:07:52 most stuff now builds fine in angstrom Apr 30 15:08:23 kdepimpi seems to have been broken by qte-mt, fails looking for a qmutex.h Apr 30 15:09:52 Bernardo: Can you fix it? Apr 30 15:10:11 provide a patch and attach it to the BTS Apr 30 15:10:23 That would be very nice of you Apr 30 15:10:28 no idea. I mostly have time to build and not much more Apr 30 15:10:41 hey Bernardo Apr 30 15:10:43 but I'll try Apr 30 15:10:43 Bernardo: kdepimpi bug is known. Apr 30 15:10:45 hi Jin^eLD Apr 30 15:11:04 Bernardo: it was never designed to work with qte-mt Apr 30 15:11:07 hrw: I just stumbled on that one Apr 30 15:11:33 hrw: and since it seems abandoned, I think I'll have to live without kopi Apr 30 15:11:40 Bernardo: it is done as 'if no threads and qt2 then build. if threads that assume qt3' Apr 30 15:14:11 hrw: thanks Apr 30 15:16:01 * chouimat is bored Apr 30 15:21:46 hrw, Bernardo: I have a patch for kdepimpi. I need to test if it compiles with recent tools, though. Apr 30 15:21:58 great Apr 30 15:22:19 polyonymous: great, post it and we'll test it Apr 30 15:22:38 a quick mtn diff: http://rafb.net/p/uC17iS16.html Apr 30 15:22:41 the only showstopper I've had until now with angstrom/opie is lirc-modules Apr 30 15:23:37 hmm.. I think there was a longer patch... Apr 30 15:23:55 well, anyway, I'm trying to build it later today, so I'll see. Apr 30 15:24:15 this patch I used like half a year ago for .dev-based OZ. Apr 30 15:25:47 It would also have been great if we could aply the new version of konq/E. I have buildt it and it compiles, but i havent tried it out yet. Apr 30 15:26:22 polyonymous: I'll update my rep and try in a few minutes Apr 30 15:27:02 ok. I'm building angstrom-x11-image now and opie will be the next to try, I'd estimate it as a couple of hours :) Apr 30 15:28:03 polyonymous If you are building wih Angstrom distro you have to set up some stuff in local.conf. There is a howto at angstrom wiki Apr 30 15:28:58 goxboxlive, I know Apr 30 15:29:06 good Apr 30 15:29:18 goxboxlive, actually, I did build it yesterday, but I started afresh with bb 1.8 now that I fixed 1.6 :) Apr 30 15:29:31 ok Apr 30 15:29:43 * Bernardo forgot he had just rebooted his server and got bitten by the md raid bug in ubuntu once more Apr 30 15:30:01 will take some time to get the build started... Apr 30 15:35:39 polyonymous: you fixed bb 1.6 upstream or locally? Apr 30 15:42:49 likewise: it is not yet upstream Apr 30 15:43:17 likewise: the patch looks sane and should be applied to 1.8,1.6 and trunk but I would like to understand the issue a bit better Apr 30 15:43:35 zecke|siesta: stop talking in your sleep ;) Apr 30 15:44:19 * Bernardo tries to do a 3 way merge in vim and fails miserabily Apr 30 15:45:49 how do I convince it to accept what is shown in the left pane? Apr 30 15:52:32 zecke|siesta: http://gtk-webcore.svn.sourceforge.net/viewvc/gtk-webcore?view=rev&revision=97 Apr 30 15:53:06 koen: seen it yesterday Apr 30 15:53:30 koen: I don't have any strong opinion yet. But he seems to have difficulties working with communities and upstream Apr 30 15:53:53 like all nokia employees ;) Apr 30 15:54:10 koen: I think I will currently ignore him (probably not his API draft) Apr 30 15:54:33 koen: the naming of files, structs is not gtk'ish :) Apr 30 15:55:45 RP: ping Apr 30 15:56:01 zecke|siesta: I'll wait for philippe to add the recipes for it :) Apr 30 15:56:31 Nokia gang... he should push him to do stuff upstream Apr 30 15:56:55 anyway back to windows :} Apr 30 16:01:41 hvontres|poodle: pong Apr 30 16:03:46 RP: Do you have any clues where I should look for problems in builting the modules.symbols file? It seems it isn't getting build correctly for me and causing th poodle sound modules not to load Apr 30 16:03:56 http://www.bloomberg.com/apps/news?pid=20601087&sid=aHigP6BkC8pE&refer=home Apr 30 16:04:34 hvontres|poodle: The KERNEL_VERSION used to cause problems with that. Is it your own kernel? Apr 30 16:05:34 zecke|siesta: ok, enjoy your siesta Apr 30 16:06:50 RP: the host is fedora 6 stock, I am trying to build linux-rp-2.6.20 Apr 30 16:07:03 hvontres|poodle: ok, I doubt thats the problem then Apr 30 16:07:16 03koen 07org.oe.dev * rbdbd61d0... 10/ (1 packages/linux/linux-gta01_2.6.21.1.bb): linux-gta01: lower default pref Apr 30 16:07:23 hvontres|poodle: In short I don't know. You can just work through the problem sysmtematically I guess... Apr 30 16:07:41 RP: I do get a strange warning about a missing path that includes the output of uname -r... Apr 30 16:08:07 what could be the reason for error message "syslogd: semget: No space left on device" after start syslogd -C 16 -m 20 Apr 30 16:08:23 hvontres|poodle: I think I've seen that - it could be the problem, I don't know. Worth checking Apr 30 16:08:52 RP: Thanks Apr 30 16:09:17 mrdata: disk full is the most likely cause Apr 30 16:09:45 koen: i have 128MB ram Apr 30 16:10:24 koen: MemFree: 73348 kB Apr 30 16:11:12 koen: and syslogd should use shared mem (circular buffer) Apr 30 16:11:50 mrdata: have you read the OE mailinglist? Apr 30 16:11:59 koen: no Apr 30 16:12:32 koen: you have post for changing syslog.conf right Apr 30 16:18:08 zecke|siesta, the patch (the logic of the patch) seems to be applicable to 1.8 as well, but I do not experience the same problems with 1.8. Probably because MY problem was showing up in *oldbitbake* something function. Apr 30 16:21:45 koen: you have wrote to use /var/log/messages thats okay, because /var/tmp is tmpfs, and thats work Apr 30 16:22:29 koen: but why could i not start syslogd -C 1 for example, became the same error message Apr 30 16:22:59 cu Apr 30 16:23:42 koen: ipc is enabled in kernel Apr 30 16:25:11 circular buffer should live in ram i think, and why can syslogd not using it Apr 30 16:39:49 Jin^eLD: http://www.angstrom-distribution.org/repo/?action=details&pnm=mediatomb Apr 30 16:55:24 like|away, gnu radio ml Apr 30 17:01:05 koen: cool! thanks! Apr 30 17:09:41 Jin^eLD: are you still using a fedora 6 box for building? Apr 30 17:12:10 hvontres|poodle: yes, you ask because of this qemu problem? Apr 30 17:12:46 Jin^eLD: no, Now I am trying to track down a wierd problem when building a 2.6.20 kernel Apr 30 17:13:07 what is happening? Apr 30 17:13:55 I did not try building 2.6.20 yet Apr 30 17:14:54 Jin^eLD: It seems that when generating the modules.symbols file, the output from uname -r on the host machine sneaks in and causes problems. My problem is I don't really know where to start looking.. Apr 30 17:16:23 I guess the question is, where uname -r is called from? but hm.. if the wrong uname is called, then other kernel builds should also show the problem? Apr 30 17:17:05 ..unless the bug is in the .bb file for this particular version Apr 30 17:17:09 but thats just a wild guess Apr 30 17:21:26 Jin^eLD: it seems something changed in the kenrnel build system between 2.6.17 and 2.6.20. I tried out the 2.6.20 kernel form the angstom unstable feed and it worked... So I guess it's digging time... Apr 30 17:23:45 I have to use some older vendor kernels because of the drivers, so I never ran into this problem, simply did not have to build the new one yet Apr 30 18:15:14 Good evening Apr 30 18:17:12 hey sirfred Apr 30 18:17:21 sirfred: http://rafb.net/p/LP5nGM60.html Apr 30 18:19:18 koen: Imageon 100 hardware handles directly planar yv12. Apr 30 18:19:49 what happens if you have yuv420? Apr 30 18:19:56 * koen knows nothing about YUV Apr 30 18:20:21 koen: It seems that they're talking about a non planar format, where the planes information is interleaved. Apr 30 18:20:54 koen: It seems that they need to convert to that non-planar format to output it to the nokia video hardware. Apr 30 18:21:00 yes Apr 30 18:21:18 and they now have some accelerated conversion code Apr 30 18:21:36 Yes, that's what I've understood. Apr 30 18:21:49 But the ati w100 is able to handle also non planar formats directly. Apr 30 18:22:26 even better :) Apr 30 18:22:28 And the xv overlay driver I've written is exposing those posibilities. But I never tried it with such formats. I'm mainly trying divx, and that one seems to use always YV12 Apr 30 18:22:37 koen: And also some RGB formats. Apr 30 18:22:47 sirfred: one question I have with the xv interface Apr 30 18:22:48 koen: But I never had to test them . Apr 30 18:23:02 siesta, how "spanish". ;-) Apr 30 18:23:20 sirfred: would it be possible to use lets say the PXAFB hw overlay/underlay inside kdrive for Xv? Apr 30 18:23:37 zecke|siesta: I suppose so, as we are using the Xw100 one inside kdrive. Apr 30 18:23:56 zecke|siesta: xserver-kdriver-imageon is using the hardware to implement the xv layer. Apr 30 18:24:50 koen: Is gtk-engines/gtk-sato-engine_svn.bb a good name/directory for the o-hand sato gtk engine and theme? Apr 30 18:25:50 zecke|siesta: The only needed thing is to implement PutImage for the exposed and supported video color formats. Apr 30 18:25:56 sirfred: I think so Apr 30 18:26:25 koen: Perhaps I should include also the matchbox theme in that recipe, or perhaps I should use another place? Apr 30 18:26:37 koen: I mean, different package, same recipe Apr 30 18:26:44 * koen has no real idea on that Apr 30 18:32:16 * XorA|gone wonders how to get debug info on serial ports Apr 30 18:40:12 XorA|gone: what device? Apr 30 18:40:35 8250 Apr 30 18:40:49 hvontres|poodle: my rtm8000 to be precise Apr 30 18:41:13 XorA|gone: ohh, you mean how to debug the serial port...I thought you were trying to get a sereial console to work ...:) Apr 30 18:41:39 * koen hands XorA|gone a logic analyzer Apr 30 18:42:39 * hvontres|poodle thinks debugging serial is almost as much fun as trying to diagnose hangs on resume... Apr 30 19:05:26 03pfalcon 07org.oe.dev * r3744a34d... 10/ (1 packages/linux/linux-handhelds-2.6/h5000/defconfig): Apr 30 19:05:26 linux-handhelds-2.6: Add defconfig for h5000. Apr 30 19:05:26 * Lart mtn's changing of add command semantics. Apr 30 19:05:26 * Closes #2175. Apr 30 19:11:39 gah rtm8000 problem is beyond my fixing capability Apr 30 19:22:25 like|away, ping Apr 30 19:25:30 weird. Yesterday, when I was building with bb 1.6 it went okay, today I'm struck with wireless kernel headers failures... Apr 30 19:33:06 koen: wanna go with me to Oslo? Apr 30 19:37:33 with you everyone wants to go anywhere! :) Apr 30 19:38:34 :) Apr 30 19:39:21 woglinde: hi Apr 30 19:39:39 hi mrdata :) Apr 30 19:44:47 zecke|siesta: sure, when? Apr 30 19:53:35 anyone have an y ideas on how we can rid packages of bad RPATHS Apr 30 19:53:42 hmm, what is rpath? Apr 30 19:54:18 Crofton|laptop: like RDEPENDS Apr 30 19:54:27 Crofton|laptop: runtime path to find shared libraries Apr 30 19:54:54 Crofton have you had time to build kernel packages for osk ? :P Apr 30 19:55:03 hmm Apr 30 19:55:06 I may have Apr 30 19:55:15 but I am kind of stuck on rpath Apr 30 19:55:17 koen: or should we ask to visit australia? Apr 30 19:55:28 I did push upates to build modules Apr 30 19:55:46 putumutukas, do you have something working atm? Apr 30 19:56:08 no Apr 30 19:56:16 ok Apr 30 19:56:27 i had to visit st petersburg because of the problems between estonia and russia :s Apr 30 19:56:29 Let me see if I hvae a kernel and the ipks for the modules Apr 30 19:56:39 over the statue? Apr 30 19:56:41 Crofton|laptop: if you currently don't look for security, disable the insane.bbclass Apr 30 19:56:45 yes Apr 30 19:57:18 I saw an article, how did that impact you? Apr 30 19:57:39 http://lists.debian.org/debian-devel/1999/01/msg02245.html Apr 30 19:58:03 it's more than over the statue ... it's a fight between proud and bullheaded estonians and russian propaganda machine :D Apr 30 19:58:37 I wish peope could try and get along better Apr 30 19:58:38 and at some part i'm responsible for the security in foreign embassies Apr 30 19:58:44 ah Apr 30 19:58:57 actually the people are getting along very good Apr 30 19:59:07 Crofton|laptop: mkdir classes/; touch classes/insane.bbclass in some path inside the BBPATH Apr 30 19:59:15 Crofton|laptop: this temp. will disable the insane.bbclass Apr 30 19:59:40 just the bullheaded ones looking for audience are casuning trouble? Apr 30 19:59:58 but the truth is served in different light in russian media and that makes russian people angry in russia, although most of the people here are getting along very good Apr 30 19:59:58 zecke, I will have to remember this Apr 30 20:01:21 i think the trouble started because the occupation in WWII is a sensible subject for estonians and the fallen troops in WWII are a sensible subject to russians Apr 30 20:02:17 I can see why it would be easy to get people upset over the whole thing Apr 30 20:02:34 you need a kernel and an ipk with the serial usb module? Apr 30 20:02:52 now both wounds are torn open and estonians pour salt by moving the statue out of the city centre and russian government is producing false news to heat the trouble Apr 30 20:03:11 yes Apr 30 20:03:22 and preferrably the scsi and usb-storage module too Apr 30 20:03:23 I hope cooler heads prevail Apr 30 20:03:28 ok Apr 30 20:03:30 me too Apr 30 20:03:49 but the worst is yet to come i think Apr 30 20:04:04 what is the name os the scis module? Apr 30 20:04:06 9. may is the historical end of WWII Apr 30 20:04:32 ah Apr 30 20:05:03 lets run ver to #ol Apr 30 20:26:26 zecke: do you have a minute or two? Apr 30 20:49:12 Crofton: pong (RTT ~3h) Apr 30 20:49:25 Crofton|home, Crofton|laptop: pong (RTT ~3h) Apr 30 20:49:43 heh Apr 30 20:50:03 I was googling rpath and came across a link from teh debian archives Apr 30 20:50:40 Crofton|laptop: be quick, I'm off to bed soon :-) Apr 30 20:50:59 Crofton|laptop: in an attempt to run away from rpaths :-) Apr 30 20:51:20 good night Apr 30 20:51:34 its sounds like maybe they come from libtool Apr 30 20:52:01 Crofton|laptop: yes, but libtool itself is often generated from something else. Apr 30 20:52:33 well let me know if you have any ideas on how we can work on the troublesome packages Apr 30 20:53:05 Crofton|laptop: it's the whole autoconf/libtool stuff which I'm not familiar with. I can hardly understand all the Make magic (spending too little time with them to remember all the details) Apr 30 20:53:14 heh Apr 30 20:53:21 I'll try looking some this week Apr 30 20:53:25 Maybe we should do this in a debug run someday Apr 30 20:53:37 http://lists.debian.org/debian-devel/1999/01/msg02245.html Apr 30 20:55:13 Crofton|laptop: 2.6.21 kernel resets on my EFIKA... Apr 30 20:55:27 ug Apr 30 20:55:41 The last thing the kernel says is "Calling quiesce ... returning from prom_init" Apr 30 20:57:20 hmm... kismet fails because it finds host's Magick-config, should it depend on oe imagemagick or should it have --disable-gpsmap ? Apr 30 20:57:37 I'm off now. Good nite to you for later. Apr 30 20:57:41 gn Apr 30 20:57:49 nightwise Apr 30 20:57:55 oops Apr 30 21:31:11 finally the office is empty Apr 30 21:32:02 zecke: and now it's time for fun :) Apr 30 21:32:51 hvontres|poodle: well, it is fun with people around as well Apr 30 21:33:04 crazy DVB-H folks :) Apr 30 22:26:03 zecke: Happy May Day :) Apr 30 22:27:37 thanks a lot, it will be a boring one Apr 30 22:28:35 well, boring = good in this case =) Apr 30 22:29:10 * hvontres|poodle remembers seeing the beat up police vans in Berlin back in 1987 or 1988 May 01 00:44:36 koen: angstrom-minimal-image works for nslu2le, and rootfs is only 56% full after boot :-) May 01 00:47:13 trying again now with normal bootstrap-image and MACHINE_TASK_PROVIDER set to "task-boot dropbear sysconf" so my network settings come up automatically. May 01 02:13:26 which package provides libstdc++.so.6?? May 01 02:14:25 odd. I see a libstdc++ entry in work directory... **** ENDING LOGGING AT Tue May 01 02:59:57 2007