**** BEGIN LOGGING AT Tue Nov 28 02:59:59 2006 **** ENDING LOGGING AT Tue Nov 28 05:44:59 2006 **** BEGIN LOGGING AT Tue Nov 28 05:46:13 2006 **** ENDING LOGGING AT Tue Nov 28 05:51:13 2006 **** BEGIN LOGGING AT Tue Nov 28 05:51:43 2006 Nov 28 07:57:23 yo Nov 28 08:17:43 hi Nov 28 08:29:28 morning Nov 28 08:30:03 morning Nov 28 08:36:13 * XorA begins to cry Nov 28 08:38:07 morning all Nov 28 08:41:25 yop RP Nov 28 08:43:36 good morning Nov 28 08:58:16 RP: do you have any idea why libc6 would end up missing from all my images Nov 28 08:59:13 ls /lib is quite simple :-( "libgcc_s.so.1 libproc-3.2.7.so modules udev" Nov 28 08:59:49 XorA: It the ipk being installed (is libc in the status file?_ Nov 28 09:00:06 * RP wishes he could type... Nov 28 09:00:29 re Nov 28 09:00:33 XorA: Were there errors in the image generation log? Nov 28 09:01:23 RP: no errors and no attempt was made to install libc6 Nov 28 09:02:06 Depends: update-rc.d, libapm1 (>= 3.2.2), libgcc1 (>= 4.1.1) Nov 28 09:02:13 XorA: Do your packages have depends on it? Nov 28 09:02:16 that suspiciously missies libc6 Nov 28 09:02:37 So the shlibs code is malfunctioning? :-/ Nov 28 09:02:45 it looks kind of like shlibs code is broken Nov 28 09:02:50 snap Nov 28 09:03:45 glibc doesnt seem to have generated any shlibs Nov 28 09:04:03 this happens on both Angstrom and Generic distro for me Nov 28 09:10:50 Good morning Nov 28 09:11:27 Can the 2.95 toolchain from Sharp build glibc 2.5 or is it too old? http://rafb.net/paste/results/IngoSW29.html Nov 28 09:11:49 old Nov 28 09:11:49 * cdbot2 is 7 hrs, 20 min old Nov 28 09:11:55 ~seen lrg Nov 28 09:12:03 Laibsch: build 2.2.5 with 2.95 Nov 28 09:12:03 Genesis: he is at FOSS.in Nov 28 09:12:04 Isn't he at FOSS.in? Nov 28 09:12:10 :x Nov 28 09:12:11 lrg was last seen on IRC in channel #oe, 3d 17h 54m 58s ago, saying: 'hehe'. Nov 28 09:12:30 oki thanks XorA & bots :) Nov 28 09:12:43 RP: shlibs code doesnt seem to dump anything in logs that looks like error Nov 28 09:13:14 hrw|work: Thanks for the info. Would sharprom-compatible.conf be the right place to fix that? Nov 28 09:14:48 Laibsch: yes Nov 28 09:15:08 Thank you. I will attach a patch to the BTS Nov 28 09:15:35 XorA: my gcc libgcc1 angstrom ipk ended up as i686 instead of arm. Maybe that is related to your problem? Nov 28 09:15:36 XorA: I'm sure I've seen shlibs do that before. I just can't remember why :-( Nov 28 09:16:03 rwhitby: did you find solution? Nov 28 09:16:05 koen pointed to http://www.openembedded.org/viewmtn/diff.psp?id1=2152f3464031a4bf3e4bc5c8203b46ebb0e44dff&id2=9a4fb914b7b1a293ffcc388e0301d3158ea42794 about 12 hours ago, but I haven't heard anything since Nov 28 09:16:47 and I can't work out why it's packaging as i686 - it's definitely building with the arm cross compiler Nov 28 09:18:33 isnt ${PACKAGE_ARCH} in cross i686 as cross runs on i686 Nov 28 09:19:03 well, that patch seems to have changed things from TARGET_ARCH to PACKAGE ARCH ... Nov 28 09:19:10 (through MY_ARCH) Nov 28 09:19:34 libgcc_s.so.1 is the one usefull lib that is in the right place for me though :-( Nov 28 09:19:57 ok, maybe unrelated then. Nov 28 09:20:29 (I had another error, which may have caused the problem I saw, and am currently rebuilding from scratch to test that) Nov 28 09:20:30 I heard that test is 43 Nov 28 09:20:56 thanks cdbot2 (huh?) Nov 28 09:21:20 CoreDump|afk: take this stupid imitation of ibot out of here please Nov 28 09:21:30 ~no-botsnack Nov 28 09:21:30 aw, gee, rwhitby Nov 28 09:22:07 What does PREFERRED_PROVIDER_libgpewidget ?= "libgpewidget" do? Nov 28 09:22:25 Laibsch: it select libgpewidget as preverred provider of libgpewidget Nov 28 09:22:37 Laibsch: because there are more then one of them Nov 28 09:22:52 But wouldn't just = do it instead of ?= Nov 28 09:23:03 Laibsch: no, ?= is weak Nov 28 09:23:17 Laibsch: someone can want to use libgpewidget-hildon instead Nov 28 09:23:40 I see. This preference is only set if it is not set elsewhere? Nov 28 09:23:50 yes Nov 28 09:23:57 ?= is 'set if not set' Nov 28 09:24:04 RP: NOTE: Couldn't find shared library provider for librt.so.1 (comes during do_package) Nov 28 09:24:09 RP: 4 of them Nov 28 09:24:16 for different libs Nov 28 09:24:29 thanks, hrw|work Nov 28 09:24:32 Laibsch: np Nov 28 09:26:29 hi Kooen Nov 28 09:27:09 morning all Nov 28 09:29:26 good morning all Nov 28 09:29:49 2 hours of linear math + cold != fun Nov 28 09:33:03 Re: libgcc problem: they come out as armv4t and armv5te for me, so I wonder why it comes out as i686 for some people Nov 28 09:33:37 and afaics cross.bbclass doesn't fille with *_ARCH Nov 28 09:33:56 greetings ! Nov 28 09:34:08 hey Ifaistos Nov 28 09:34:11 Crofton|oe: ossie/*_svn.bb are still broken ;( Nov 28 09:34:29 FETCHCOMMAND_svn = "echo r|/usr/bin/env svn co ${SVNCOOPTS} ${SVNROOT} ${SVNMODULE}" Nov 28 09:34:29 UPDATECOMMAND_svn = "echo r|/usr/bin/env svn update ${SVNCOOPTS}" Nov 28 09:34:41 koen: it seems bizarro land has landed here, I join rwhitby in wierd shit happening in builds Nov 28 09:34:53 nice solution for https ones Nov 28 09:37:51 if i build a toolchain ,, how do i control for which host system i build the chain ? Nov 28 09:44:42 rob_w: by default it is built for your host Nov 28 09:45:16 how can i change this as my host is amd64 but i want it for a simple intel one Nov 28 09:45:46 no idea - never tried it Nov 28 10:14:31 DEBUG: calculating shlib provides for glibc Nov 28 10:14:31 DEBUG: XorA sonames : [] Nov 28 10:14:49 Im guessing that that my root cause, now to find out why Nov 28 10:15:26 bb.debug(2, "XorA sonames : %s" % sonames) Nov 28 10:15:33 is that a valid way to print sonames? Nov 28 10:15:44 * XorA doesnt really grep this python thing Nov 28 10:16:25 good Nov 28 10:16:25 I bet good - the only problem at the Z currently is that I cant get ogg123 to run. Install terminates :-( Nov 28 10:16:41 ~curse coredump for cdbots Nov 28 10:16:50 May the fleas of a thousand camels infest your most sensitive regions, coredump for cdbots ! Nov 28 10:16:51 cdbot2: part Nov 28 10:16:52 I don't know what part is, hrw|work. Nov 28 10:23:12 morning Nov 28 10:31:37 koen_: are you on oz-dev ML? Nov 28 10:35:14 hrw|work: no Nov 28 10:35:27 koen: recently unsubscribed? Nov 28 10:36:10 I was subscribed via gmane Nov 28 10:36:14 ah. Nov 28 10:36:22 making it subscriber only broke that Nov 28 10:36:26 koen: then look at my 'Future of OpenZaurus' post Nov 28 10:36:31 ah. Nov 28 10:36:37 I can forward you mail Nov 28 10:36:53 and the 'OMG!!! No 0p3r4!!!' was plain annoying Nov 28 10:37:09 ;D Nov 28 10:37:52 RP: it seems that within glibc the objdump command isnt working on libs Nov 28 10:38:57 XorA: Strange :-/. Any idea why? Nov 28 10:40:01 RP: no, I just adding debug to print the variables at each stage, when doing shlibs for glibc it crerates output for sbin/ldconfig the nothig for the rest of files in the package Nov 28 10:41:07 hrw|work: your mail doesn't include a proposal, so most people won't respond Nov 28 10:41:14 XorA: try running the shlibs objdump command manually and see if you can see why glibc isn't being seen? Nov 28 10:42:02 koen: first line of mail say why Nov 28 10:42:02 * koen still wonders why binary localegen doesn't want to run on his machine Nov 28 10:42:03 RP: working on that, I have tried arm-linux-objdump -p on the contents of the package and that works fine Nov 28 10:47:44 sweetness Nov 28 10:47:51 * koen hugs his nokia 6021 Nov 28 10:48:00 syncing works, internet works Nov 28 10:48:16 that's 2 points more as the a780 Nov 28 10:48:32 * koen gets his soldering iron and a serial connector Nov 28 10:48:47 koen: likewise: Due this unexpected two tests next week, I will chicken-out :( Nov 28 10:50:14 NOTE: Removing failed runtime build target python-pygnome Nov 28 10:50:14 NOTE: Removing failed build target hal Nov 28 10:50:14 NOTE: Removing failed build target gnome-power Nov 28 10:50:25 hal+gnome-power -> removal.txt? Nov 28 10:50:46 hal works Nov 28 10:51:03 koen: not hal 0.5.7 then Nov 28 10:51:08 and this is default Nov 28 10:51:08 and this is default Nov 28 10:51:20 CoreDump|afk: I HATE YOU Nov 28 10:53:09 koen: so choices: 1. add python-(py)gnome, 2. drop hal-device-manager from hal Nov 28 10:53:31 3. drop pygnome from DEPENDS Nov 28 10:53:41 RP: it looks like BUILD_PREFIX isnt being set Nov 28 10:53:54 RP: so its calling amd64 objdump Nov 28 10:54:08 koen: hal-device-manager needs it Nov 28 10:55:23 * koen looks for the python-pygnome recipe Nov 28 10:55:46 I am looking at bug #68. My question is does it make sense to disable java for all uclibc based distro by issuing JAVA_linux-uclibc="" ? Nov 28 10:56:25 is anyone using it ? i see mips,and arm achs disable it, it also fails on x86 Nov 28 10:56:32 Anybody care to push bug 1551? It fixes sharprom-compatible.conf Nov 28 10:56:59 RP: any ideas how to fix? Nov 28 10:58:28 Laibsch: when I will get it loaded then will look Nov 28 10:59:24 hrw|work: Thank you. You mean, bugs.oe.org does not load? Neither for me, ATM. Shall I send you the patch? Nov 28 10:59:58 * koen looks at the topic Nov 28 11:00:18 koen: want 'tail -n1 /etc/hosts'? Nov 28 11:00:56 03koen 07org.oe.dev * r9571370c... 10/ (1 packages/hal/hal_0.5.7.bb): hal 0.5.7: disable hal-device-manager till someone adds python-pygnome Nov 28 11:03:38 XorA: BUILD_PREFIX doesn't sound right? Nov 28 11:04:04 RP: line 578 in package.bbclass Nov 28 11:04:34 just for info: bugs.treke.net also does not load Nov 28 11:09:46 koen: Same here. I can reach bugs.oe.org via /etc/hosts, but it seems to be generally down. Nov 28 11:12:42 XorA: I think it should be HOST_PREFIX Nov 28 11:13:12 RP: trying Nov 28 11:13:45 XorA: I'm tempted to recommend declaring an entry like OBJCOPY in bitbake.conf Nov 28 11:14:50 RP: that makes a lot of sense Nov 28 11:14:58 RP: as all the other command are in there Nov 28 11:15:23 XorA: I was the one who added OBJCOPY there when I added -dbg support Nov 28 11:15:55 XorA: I just wonder how BUILD_PREFIX worked for anyone? :-/ Nov 28 11:16:08 RP: I was wondering that as well Nov 28 11:18:19 NOTE: the following files were installed but not shipped in any package: Nov 28 11:18:19 NOTE: /a/home/hrw/devel/build/test/tmp/staging/arm-linux/include/Eet.h Nov 28 11:18:22 hm. Nov 28 11:20:27 RP: HOST_PREFIX certainly seems to be working looking at tail -f of the log Nov 28 11:20:58 tmp/staging/arm-linux/shlibs/glibc.list exists as well Nov 28 11:21:32 ~seen JustinP Nov 28 11:21:58 justinp is currently on #oe (5h 4m 49s) #openzaurus (5h 4m 49s), last said: 'once it's in a release branch I'll use it ;-)'. Nov 28 11:23:10 time to build an image Nov 28 11:24:17 XorA: On my system the standard objdump seems to work with ARM binaries Nov 28 11:24:28 XorA: I guess its a 64 bit issue? Nov 28 11:24:55 RP: oddly it seems to work at home, not at work, both are amd64 Nov 28 11:25:03 RP: and both use openSuSe Nov 28 11:25:24 XorA: you see different output with the plain objdump? Nov 28 11:25:41 RP: it says file format not recognised Nov 28 11:27:03 Ifaistos: disable java in uclibc is a good idea Nov 28 11:28:47 ade|desk : i'll do some more tests and then push it to the repository Nov 28 11:28:52 XorA: I wonder if its something about changes in the toolchain. Have you tried the same arm binary on both? Nov 28 11:30:19 RP: world builds will need few more progressbars ;( Nov 28 11:30:46 hrw|work: How do you mean? Nov 28 11:30:54 NOTE: consider defining PREFERRED_PROVIDER_libao-alsa Nov 28 11:30:55 ERROR: No providers of runtime build target libesmart-trans-x11-0 (for ['/home/hrw/devel/oe/org.openembedded.dev/packages/e17/entice_20060416.bb']) Nov 28 11:31:11 RP: between those both lines is lot of time with no info Nov 28 11:31:43 RP: thats the trick, mine recognises armv5te but not arm920t libs Nov 28 11:32:07 Ifaistos: its only gcj thats not working at least for gcc4 Nov 28 11:32:49 hrw|work: For the Preparing runqueue? Nov 28 11:32:58 aha Nov 28 11:33:04 RP: yes Nov 28 11:33:25 MY_ARCH only works for machines that explicitly set PACKAGE_ARCH Nov 28 11:33:48 RP: currently we do not have any information for user what is going on. and I do not want to think how long it take on <2G 32bit Nov 28 11:34:03 hrw|work: Its only really a problem on world builds and you kind of expect those dependencies to take a while Nov 28 11:34:58 ok then Nov 28 11:41:26 ...should'n the ipk packges that were build for eabi target be marked somehow? because they can not be used on non-eabi systems, or am I wrong? Nov 28 11:43:10 there are too much ABIs to make for arm Nov 28 11:43:45 hardfpa, softfpa, hardvfp, softvfp, eabi, etc Nov 28 11:44:10 but are the packages compatible between those ABIs? Nov 28 11:44:16 they are not Nov 28 11:44:40 ...so is it possible that the particular distro sets _forabi.ipkg Nov 28 11:44:52 aem, I mean, adds that to the package name or something? Nov 28 11:45:30 my problem right now is that I am building for eabi and non-eabi targets, and all packages land in deploy Nov 28 11:45:41 deploy/ipk that is, and I can not keep them apart anymore Nov 28 11:47:44 that's why I have ${DISTRO} in tmpdir :) Nov 28 11:49:19 hmm, but various distros can share stuff when using multimachine, no? like the i686 native stuff? Nov 28 11:50:18 no, DISTROs can't share stuff Nov 28 11:50:24 it's multimachine, not multidistro Nov 28 11:50:59 oh, ok I get it Nov 28 11:51:00 thanks Nov 28 11:51:01 you could put ${TARGET_PREFIX} in ${DEPLOY_DIR} Nov 28 11:51:03 hello fooks Nov 28 11:51:08 hey likewise Nov 28 11:51:34 hey koen Nov 28 11:51:47 * koen flips a coin Nov 28 11:52:06 decision making trouble? Nov 28 11:52:21 yeah, "stay home" or "go to werkcollege" Nov 28 11:52:37 in combination with fever Nov 28 11:53:20 I noticed this morning my attention span was 15minutes too short for the lecture Nov 28 11:58:16 k Nov 28 11:58:53 I wonder when last time JustinP built e17 from .dev Nov 28 12:14:52 XorA: what is an address to asoc info other then the one on RP page? Nov 28 12:15:11 http://opensource.wolfsonmicro.com/~lg/ Nov 28 12:15:14 http://opensource.wolfsonmicro.com/~lg/asoc/ Nov 28 12:16:25 thx Nov 28 12:29:52 koen: Can you explain why you needed MY_ARCH in the first place? Nov 28 12:30:09 RP: is your libvolumeid fix in .dev as well? Nov 28 12:30:14 koen: I don't think requiring PACKAGE_ARCH for every machine is a good idea Nov 28 12:30:17 RP: see bug #1265 Nov 28 12:30:24 koen: no, its not Nov 28 12:30:55 setting PACKAGE_ARCH to TARGET_ARCH is wrong Nov 28 12:31:19 koen: Why? Nov 28 12:31:43 if I build for armv5te libgcc will claim to be 'arm' (=armv4l) while containing armv5 isns Nov 28 12:32:08 if I build for athlon libgcc will claim to be 'i686' while containing mmx2 and 3dnow isns Nov 28 12:32:10 etc Nov 28 12:33:34 So I'd rather set PACKAGE_ARCH for each machine than let OE create bogus packages Nov 28 12:33:38 koen: ok, so gcc packaging is broken. Why does it touch PACKAGE_ARCH_* at all? Nov 28 12:33:51 RP: 'inherit cross' Nov 28 12:34:45 http://rafb.net/paste/results/WJOobp70.html Nov 28 12:36:18 RP: not only gcc packaging, see wpa-supplicant 0.5.5 as well Nov 28 12:37:07 koen: wpa-supplicant inherits cross? Nov 28 12:37:31 #we introduce MY_ARCH to get 'armv5te' as arch instead of the misleading 'arm' on armv5te builds Nov 28 12:37:31 MY_ARCH := "${PACKAGE_ARCH}" Nov 28 12:37:32 PACKAGE_ARCH = "${@base_contains('COMBINED_FEATURES', 'pci', '${MACHINE}', '${MY_ARCH}', d)}" Nov 28 12:38:00 koen: Have you tried MACHINE_ARCH in these places? Nov 28 12:39:05 wouldn't work Nov 28 12:39:07 MACHINE_ARCH = "${@[bb.data.getVar('HOST_ARCH', d, 1), bb.data.getVar('MACHINE', d, 1)][bool(bb.data.getVar('MACHINE', d, 1))]}" Nov 28 12:39:23 that will give you 'arm' or '$MACHINE' Nov 28 12:40:40 isn't linux fun? Nov 28 12:41:11 koen: The thing is, tune-xxx.inc sets PACKAGE_ARCH = "xyz" which should come after the entries in bitbake.conf? Nov 28 12:41:21 yes Nov 28 12:42:11 putting 'PACKAGE_ARCH := "${HOST_ARCH}"' as the first line in each machine.conf would 'solve' it Nov 28 12:42:17 but I don't like that Nov 28 12:42:34 which is why I sent the mail to the list Nov 28 12:43:57 koen: I don't understand the problem. PACKAGE_ARCH = armv5te when compiling something armv5te specific? PACKAGE_ARCH = arm when compiling something arm specific? Nov 28 12:44:40 libgcc is built by gcc-cross.bb, which is 'i686' for me Nov 28 12:44:52 but the resulting libgcc is armv5te Nov 28 12:44:55 So the issue is inherit cross? So should bitbake.conf say PACKAGE_ARCH := "${HOST_ARCH}"? Nov 28 12:45:16 libgcc shouldnt really come from gcc-cross, but from gcc_X.bb Nov 28 12:45:32 (and remove the misplaced TARGET_ARCH settings in gcc) Nov 28 12:46:10 I suspect the immediate expansion in bitbake.conf will break a lot of assumptions Nov 28 12:46:27 koen: You can't immediately expand that :-/ Nov 28 12:46:49 koen: ok, cross.bbclass needs to do something about it Nov 28 12:48:11 koen: Something like OLD_PACKAGE_ARCH := ${PACKAGE_ARCH} PACKAGE_ARCH = ${OLD_PACKAGE_ARCH} Nov 28 12:48:27 (before changing HOST_ARCH) Nov 28 12:49:12 wpa-supplicant should work with MACHINE_ARCH and is a different problem Nov 28 12:50:32 wpa supplicant is either machine specific, of architecture specific Nov 28 12:50:43 so 'c7x0' or 'armv5te' Nov 28 12:50:47 hence, MY_ARCH Nov 28 12:50:59 otherwise it's 'c7x0' or 'arm' Nov 28 12:53:40 * koen gets some more soup Nov 28 12:54:12 koen: ah, right, yes. different problem but your solution is correct there as we don't allow self referencing variables in OE :} Nov 28 12:54:47 I presume PACKAGE_ARCH = "${@base_contains('COMBINED_FEATURES', 'pci', '${MACHINE}', '${PACKAGE_ARCH}', d)}" gives parse errors? Nov 28 12:57:59 yes Nov 28 12:58:07 self referencing indeed Nov 28 12:58:39 I don't see why we couldn't support that :} Nov 28 12:58:46 but yeah, that should use MACHINE_ARCH to account for the kergoths in the world Nov 28 12:59:43 hey Nov 28 12:59:52 hey lrg Nov 28 13:00:04 * lrg has to survive lots of power cuts in his hotel :( Nov 28 13:00:17 hey Koen Nov 28 13:00:20 hey lrg Nov 28 13:00:32 hey Philipp Nov 28 13:00:44 hey pH5 Nov 28 13:00:53 pH5: i think the div is too big now Nov 28 13:01:04 hey koen Nov 28 13:01:30 pH5: try with 0 again Nov 28 13:02:03 03koen 07org.oe.dev * r45e7f7da... 10/ (1 packages/wpa-supplicant/wpa-supplicant_0.5.5.bb): wpa supplicant 0.5.5: $MACHINE} isn't valid as PACKAGE_ARCH, use ${MACHINE_ARCH}, makes kergoth happy when he's building without setting MACHINE Nov 28 13:02:42 * koen adds a missing { to the commit message Nov 28 13:03:09 koen: Fancy testing the proposed change to cross.bbclass? :) Nov 28 13:03:09 hello lrg Nov 28 13:03:17 hi Liam Nov 28 13:03:22 hey Richard, vlo Nov 28 13:03:52 RP: could you show the patch first? ;) Nov 28 13:03:55 lrg: so the get_audio_scr can not be used with the audio clock? Nov 28 13:04:15 pH5: i think so, it's not very clear in the ref manual Nov 28 13:04:34 'very clear', lol. to me everything is very unclear :) Nov 28 13:04:35 pH5: maybe some trial and error Nov 28 13:04:46 ok Nov 28 13:05:34 lrg: still no success in my mx31ads serial boot ... Nov 28 13:06:22 hey lrg Nov 28 13:06:23 vlo: I'll be working on mx31 next week Nov 28 13:06:29 hey Graeme Nov 28 13:06:56 koen: http://www.rpsys.net/openzaurus/temp/cross_package_arch.patch Nov 28 13:08:56 lrg: yeah, but it's strange that with the same kernel and hardware, you can boot your image and I cannot ... Nov 28 13:09:40 vlo: maybe we have different bootloaders Nov 28 13:10:29 vlo: btw my board is non RoHS (it has pb), and not for general sale, so maybe other things are also different Nov 28 13:11:16 RP: woowoo, an image with some libc6 goodness :-) Nov 28 13:11:33 hmm. did anyone have a problem in not being able to run any threads? I made a custom distro for a custom machine, tried both, eabi and non eabi, tested with a simple c program that does a simple pthread_create() and a printf inside a thread (so nothing memory intensive); pthread_create always returns with out of mem; same program works on the vendor distro on the same hardware. any hints or ideas on what it might be related to? Nov 28 13:13:12 INIT: version 2.86 booting finally Nov 28 13:14:21 Jin^eLD: I have had some intermittent problems with threading in Python in an older openslug distro, but never tracked down what was going on. Nov 28 13:14:33 Jin^eLD: we are running threads on recent builds (using glib) with no problems. Nov 28 13:15:16 Jin^eLD: with the Python issue, threads would just mysteriously get killed or disappear. Nov 28 13:15:27 hmm, well I can not even start them Nov 28 13:15:35 Jin^eLD: yeah, that sounds different. Nov 28 13:15:47 any idea on where to start looking? Nov 28 13:16:10 lrg: the ssp clock has to be disabled while changing the SSCR0 register? Nov 28 13:17:28 pH5: yes, according to the ref manual Nov 28 13:17:54 pH5: only for the clk though Nov 28 13:18:39 so I should put those pxa_set_cken around the call to ssp_config instead Nov 28 13:19:24 pH5: doh, of course Nov 28 13:19:33 * lrg needs more coffee Nov 28 13:20:00 RP: works, I shall commit it Nov 28 13:20:09 pH5: must be lack of sleep ;) Nov 28 13:20:27 cbrake: any idea on how to debug such a thing? I am not even sure how to get started with this Nov 28 13:20:46 hey lrg , do you know the status of usbhost/usbotg on the mx21 ? Nov 28 13:21:16 Genesis: sorry, no Nov 28 13:21:51 oki Nov 28 13:21:55 03rpurdie 07org.oe.dev * r9562dbac... 10/ (1 classes/cross.bbclass packages/gcc/gcc-package-cross.inc): cross.bbclass + gcc-package-cross.inc: put the immediate expansion trick in cross.bbclass Nov 28 13:22:27 RP: shall I push that change? Nov 28 13:24:38 lrg: yes, I think there must be a few differences between our iMX boards ... we should check what revisions we have Nov 28 13:24:44 XorA: Adding OBJDUMP to bitbake.conf? Nov 28 13:25:21 vlo: I'm in India until Friday, will check on return Nov 28 13:25:40 lrg: ok we will see then Nov 28 13:27:31 I have disabled binary locale generation lately. Now a fresh build errors out on "Cannot find package glibc-localedata-i18n." Nov 28 13:28:24 koen: For what its worth we can do PACKAGE_ARCH := ${PACKAGE_ARCH}, self references are allowed with immediate expansion :} Nov 28 13:28:29 RP: well I can make the BUILD_PREFIX to HOST_PREFIX change, but Im not confident enough in python to change the whole statement to use an OBJDUMP defined in bitbake.conf Nov 28 13:29:07 * XorA is scared of the hoards of angry devs if hy breaks package.bbclass Nov 28 13:31:14 Jin^eLD: sounds pretty difficult to debug. I would probably try some control builds and see if they do the same thing. Nov 28 13:36:03 cbrake: control builds like what? Nov 28 13:36:14 XorA: cmd = bb.data.getVar('OBJDUMP', d, 1) + " -p " + path + " 2>/dev/null" Nov 28 13:37:46 cbrake: the interesting thing is - when I first setup OE, added the machine and used the prebuilt toolchain - threads worked; but from the moment I switched to an environment that created a distro and everything from scratch it stopped working; I wonder what I messed up and how that was possible Nov 28 13:37:49 Jin^eLD: x86/generic, logicpd-pxa270/generic -- something that seems to be working for other people. Maybe your custom distro is not that much different from one of these though? Nov 28 13:37:50 RP: ah that simple Nov 28 13:38:26 XorA: Untested but it should work :} Nov 28 13:38:44 it should not be too different, I started off with generic and also lookged at angstrom.conf files, hrw and koen helped me out to get things going and it seemed fine.. apart from this problem Nov 28 13:38:50 RP: ok, Ill give it some testing and commit later Nov 28 13:39:02 XorA: ok, thanks Nov 28 13:39:03 hmm, ok I will try a generic distro then Nov 28 13:39:41 Jin^eLD: you can boot most bootstrap-image builds on most machines, so I would just try something else. Nov 28 13:39:50 ok, thanks Nov 28 13:39:54 I'll try that Nov 28 13:40:26 Jin^eLD: if you still get failures, I can send you one my builds to try as a control. Nov 28 13:43:45 ENABLE_BINARY_LOCALE_GENERATION="0" with LINGUAS_INSTALL non-empty seems to fail here. Is that a bug in my config? Nov 28 13:52:29 Hi! Nov 28 13:52:30 Now even bugs.treke.net doesn't respond? Nov 28 13:53:10 psokolovsky: their T3 doesn't like rain :( Nov 28 13:53:25 ;-( Nov 28 13:53:30 koen: when he said T3 he meant his Tungsten :-) Nov 28 13:54:14 yeah, bugzilla is running under cygpalm ;) Nov 28 13:54:46 .oO(that's the reason it doesn't work?) Nov 28 13:55:03 another set of bugdays coming up this weekend Nov 28 13:55:32 koen: well thats an easy one, currently we have no bugs Nov 28 13:56:48 bitbake task-bootstrap should result in a image.tar.bz2 right ?? i cant find it ?!? Nov 28 13:58:20 bitbake bootstrap-image Nov 28 13:58:43 ah thanks .. Nov 28 14:14:31 * RP wonders about the ethics of submitting all the bugs on his todo list to the bugtracker... Nov 28 14:16:35 RP use a pseudonym ? Nov 28 14:16:36 Hmmm. I think RP is your libvolumeid fix in .dev as well? Nov 28 14:16:44 RP: http://www.angstrom-distribution.org/history, rule 5 :) Nov 28 14:17:09 what do need to create my own task-myownbootstrab .. which i want to build per bitbake myownbootstrap-image Nov 28 14:17:15 ~prod cdbot2 Nov 28 14:17:24 * ibot_ zaps cdbot2 with a high voltage cattle prod Nov 28 14:17:33 i copied the task-bootstra.bb to task-myownbootstrap.bb but it doenst find it Nov 28 14:17:45 koen: poky != angstrom and most of my todo list is poky related although the bugs are mostly in OE too... Nov 28 14:18:03 RP: sounds like fair game to me then Nov 28 14:19:00 re Nov 28 14:19:06 I should add rootfs_deb.bbclass and its assicated bugs :) Nov 28 14:19:51 share and share alike Nov 28 14:20:00 rob_w: task-bootstrap is going to be removed soon Nov 28 14:22:31 hrw|work, hmm i like the idea of a minimum system Nov 28 14:26:10 rob_w: task-base is better way Nov 28 14:26:10 Well, task-base is better way Nov 28 14:26:20 god damnit Nov 28 14:26:24 cdbot2: die! Nov 28 14:26:24 koen: don't ask me, I don't know. Nov 28 14:26:56 can we not get a server op to boot that bot off? Nov 28 14:29:25 ibot_: kick cdbot2 Nov 28 14:37:33 koen: Have you had working bluetooth on your c7x0? Nov 28 14:37:45 RP: yes Nov 28 14:38:07 using a CF bt adaptor Nov 28 14:38:21 koen: Is there some trick needed with the serial ports for that? Nov 28 14:38:30 RP: got a biontoncard? Nov 28 14:38:34 billionton Nov 28 14:38:39 XorA: yes Nov 28 14:38:45 RP: you need to setserial Nov 28 14:38:48 RP: no, hciattach "just works(TM)" Nov 28 14:39:00 (after editing /etc/sysconfig/bluetooth) Nov 28 14:39:29 setserial /dev/ttyS3 baud_base 921600 Nov 28 14:39:43 hciattach -s 921600 /dev/ttyS3 bcsp Nov 28 14:39:53 is needed to initialise billionton Nov 28 14:40:18 XorA: Which modules do I need loaded? Nov 28 14:40:28 XorA: I don't have a /dev/ttyS3 yet :-/ Nov 28 14:40:55 RP: serial_cs I think Nov 28 14:41:22 XorA: Hmm. Thats what I thought too. 8250 tells me device or resource busy though :-/ Nov 28 14:41:24 RP: it may be that the whole arm serial port naming thing is screwing you, as if you look in dmesg kernel thinks just plugged in card is ttyS0 Nov 28 14:41:46 * RP wondered if it was the renaming Nov 28 14:41:52 RP: I havent got my card on me today Nov 28 14:42:16 This is with a 2.6.19-rc4 kernel Nov 28 14:42:27 dont think I tried since 2.6.17 Nov 28 14:42:37 I should try 2.6.17... Nov 28 14:43:12 but that setserial has always been neccessary Nov 28 14:44:22 ~seen sirfred Nov 28 14:44:31 sirfred was last seen on IRC in channel #oe, 3d 22h 9m 41s ago, saying: '56Mb of source code'. Nov 28 14:47:13 can each distro be used with the "multimachine" setting? I would like to add it to my site.conf so it is always used Nov 28 14:47:20 or would that be a bad idea? Nov 28 14:47:28 Jin^eLD: every distro can be used with that Nov 28 14:47:33 thanks Nov 28 14:47:51 koen: it won't hurt if INHERIT has multimachine twice in there or something? Nov 28 14:48:06 for the case that it is already defined in the somedistro.conf file Nov 28 14:48:40 03xora 07org.oe.dev * r1fb8812d... 10/ (1 classes/package.bbclass conf/bitbake.conf): Nov 28 14:48:40 bitbake.conf : introduce the OBJDUMP variable, by default set to Nov 28 14:48:40 ${HOST_PREFIX}objdump Nov 28 14:48:40 package.bbclass : make use of the OBJDUMP variable rather than Nov 28 14:48:40 calling ${BUILD_PREFIX}objdump inside do_shlibs. As the original Nov 28 14:48:40 usage was faulty and ended up calling host objdump which works Nov 28 14:48:42 Didn't you say which is sent the mail to the list? Nov 28 14:48:42 for some arm targets but not all. Nov 28 14:48:49 RP: do you know off-hand why udev 100 has default_pref -1? Nov 28 14:49:25 koen: Either someone was nervous of the new version or it has bugs Nov 28 14:49:30 I didn't add it... Nov 28 14:49:41 it has def_pref = "1" for slugos Nov 28 14:49:46 which makes me wonder Nov 28 14:49:46 I'm sure which is sent the mail to the list Nov 28 14:49:57 cdbot2: shut up Nov 28 14:49:57 I'm not sure what shut up is. Nov 28 14:50:16 CoreDump|afk: remove those &$*@#&$*# bots from #oe Nov 28 14:51:01 * RP was wondering what the cd stood for :) Nov 28 14:51:19 koen: Its probably ok to use and they just didn't want to break things for anyone else Nov 28 14:51:23 RP: 28 12:17 < hrw|work> CoreDump|afk: I HATE YOU Nov 28 14:51:37 rm -rf CoreDump|afk Nov 28 14:52:57 cross dressing bot? Nov 28 14:53:39 03koen 07org.oe.dev * r190c717f... 10/ (5 files in 2 dirs): udev: merge stuff from 097 and 100 into udev.inc, add do_stage from poky, bump PR Nov 28 14:53:59 RP: consider merging the udev.inc from .dev into poky :) Nov 28 14:54:13 libvolumeid-dev package Nov 28 14:56:17 koen: I'm happy enough for it to be merged... Nov 28 15:07:02 03koen 07org.oe.dev * rf62c6236... 10/ (1 packages/hal/hal_0.5.8.1.bb): hal: add 0.5.8.1 Nov 28 15:33:46 hrw|work: bugs.oe.org seems to be working again Nov 28 15:33:57 can you take a look at bug 1551? Nov 28 15:34:58 will push Nov 28 15:40:50 thanks a lot. Nov 28 15:41:39 03koen 07org.oe.dev * rbf833230... 10/ (3 files in 2 dirs): python-pygtk2: update to 2.10.3 Nov 28 15:48:52 03hrw 07org.oe.dev * r1a4cfa89... 10/ (1 conf/distro/sharprom-compatible.conf): sharprom-compatible: prefer glibc 2.2.5 - close #1551 Nov 28 15:49:02 is there a good reference for assmebler in gcc? Nov 28 15:50:19 XorA: Thanks for those commands, they're helping a lot :) Nov 28 15:50:41 RP: took me a lot of googling to find them so I spread them as much as I can Nov 28 15:51:04 XorA: I presume that's followed by a hciconfig hci0 up and then things like maemo-mapper should work? Nov 28 15:51:36 XorA: 2.6.17 works so it looks like 2.6.19-rc4 has issues with the serial interface numbering :-( Nov 28 15:52:27 RP: I assume so, all my stuff is rfcomm, so just some rfcomm commands Nov 28 15:53:02 koen: Any further hints? maemo-mapper segfaults on me after the above when I try and ebale the gps :-/ Nov 28 15:54:11 RP: the pxa vs 8250 stuff probably got altered Nov 28 15:54:36 XorA: It'll be our serial port number hack :-/ Nov 28 15:55:18 RP: I actually didn't try using the gps Nov 28 15:55:41 RP: as dbus is broken, so gpe-bluetooth can't enable the bt interface Nov 28 15:55:41 RP: with udev I dont see the need for that anymore, finding the serial ports in /dev is a doddle Nov 28 15:55:56 koen: dbus is borken? Nov 28 15:55:56 I'm sure dbus is borken? Nov 28 15:56:23 Crofton: prolly better off w nasm. Nov 28 15:56:40 XorA: We have a problem in that if we disable the hack, the standard 8250 ports aren't usable at all Nov 28 15:56:40 I have inlined asm in gcc code Nov 28 15:56:59 XorA: #1567 Nov 28 15:57:18 Crofton: trying to remember. done it before but it's a long time ago. Nov 28 15:57:35 koen: oh, that one, I had thought that was fixed :-( Nov 28 15:57:49 google is helpful, but I fugure someone around here would have a fav url for looking this up Nov 28 15:58:04 RP: should we not do what RMK keeps telling us to though and move pxa to its allocated ports, same as the other arm platforms? Nov 28 15:58:39 RMK isn't serial maintainer anymore Nov 28 15:58:49 koen: I am aware of that Nov 28 15:59:20 XorA: Perhaps. It needs a flag day to update all the references in software though :-( Nov 28 15:59:40 Crofton: http://www.tldp.org/HOWTO/html_single/Assembly-HOWTO/#P-GCC about 2 sections down Nov 28 15:59:56 RP: thats where udev come in and we link /dev/ttySXXX to whatever we like depending on machine Nov 28 16:00:18 XorA: Doesn't help the SERIAL_CONSOLE Nov 28 16:00:20 RP: although Angstrom should really do away with it, as its supposed to be a clean start Nov 28 16:00:39 he Nov 28 16:00:39 RP: you talking bootloaders? Nov 28 16:00:40 I bet he is at FOSS.in Nov 28 16:00:50 see linux kernel source for useful examples :) Nov 28 16:00:56 thanks a-atwood Nov 28 16:01:17 XorA: No, kernel commandline Nov 28 16:01:36 (and inittab) Nov 28 16:01:48 RP: for smdk2440 I am already not using /dev/ttyS0 Nov 28 16:02:05 RP: oe handles that Nov 28 16:02:18 S:2345:respawn:/sbin/getty 115200 ttySAC0 Nov 28 16:02:23 XorA: but the PXA has used ttyS0 for years. Hence the need for a flag day. We change OE, old kernels stop working Nov 28 16:02:59 RP: ok, I would suggest if OZ 3.5.5 release is near, choose a time 7 days after that Nov 28 16:03:34 XorA: We don't do it until/if mainline does it. I'm not maintaining that in OE Nov 28 16:04:05 XorA: We also have to wipe all kernel versions other than the most recent mainline with that change in Nov 28 16:04:43 so no .16 .17 or .18 as they will no longer be compatible with OE Nov 28 16:06:08 ~lart historic fsckups Nov 28 16:06:09 * ibot_ send killer squirrels to attack historic fsckups Nov 28 16:11:33 bye Nov 28 16:14:59 bye Nov 28 16:19:54 I'm going to mass report all cross compile badness entries Nov 28 16:20:55 grr fontconfig is the worst Nov 28 16:21:23 ade|desk: yup, fc-lang, fc-case, and fc-glyphname call native gcc with arm arguments :-( Nov 28 16:21:59 indeed Nov 28 16:22:45 but fontconfig builds Nov 28 16:23:35 hrw|work: on some systems, not on all Nov 28 16:23:44 i did a hack a while ago that i just slapped the corrected Makefile and ignored the configure.. don't know what i did with it .. quite anoyed that i didn't push it Nov 28 16:25:13 hrw|work: try fontconfig with uclibc on any arch. glibc arm was ok but by far the worst was uclibc i386 with a build sys of i686 glibc ... that is not a good combo Nov 28 16:25:25 ade|desk: I think a quitck sed -i "s/LDFLAGS =.*/LDFLAGS =/" fc-case/Makefile in do_configure_apend() Nov 28 16:26:09 hmm needed more than that iirc Nov 28 16:26:25 ade|desk: that fixes the problem I see, which is using glibc Nov 28 16:26:27 Didn't you say which is using glibc? Nov 28 16:26:45 cdbot2: piss off Nov 28 16:26:45 Yo momma knows about piss off! Nov 28 16:26:51 it moaned about isystem too Nov 28 16:27:29 ade|desk: probably a similar replacement with CFLAGS = then Nov 28 16:28:49 anyway time to head home Nov 28 16:28:52 hmm needs CFLAGS with correct path for host Nov 28 16:34:45 http://pastebin.ca/260702 - can you guys look at this patch? it should not do any harm for slugos builds but it fix world build a bit Nov 28 16:42:50 lrg: Thanks for patching up alsa :) I rebuilt the modules last night and setting the volume in opie works now :) Nov 28 16:43:18 hvontres|poodle: cool :) Nov 28 16:45:38 lrg: RP: CoreDump updated the 2.6.17 recipies in .dev and .oz354x, but you might want to double check them..:) Nov 28 17:03:13 hrw|work: Looks ok to me although the glibc change makes me nervous - does it work for uclibc? Nov 28 17:03:54 RP: skip that part and look at all those ?= "" Nov 28 17:05:05 hrw|work: They look safe to me Nov 28 17:07:14 1~3~3~1~1~1~1~1~1~1~1~1~1~1~1~1~1~1~1~ Nov 28 17:08:38 cu Nov 28 17:15:14 I have moved over to hh2 from hh1. But g_ether is still compiled in as a modules as hh0 and hh1. How do i remove the hh0 modules? I have cleaned linux-handelds-2.6 and task-base. Is there anything else i should have cleaned up? Nov 28 17:48:09 hrw|gone: Updating Ångström unstable Progear webpad: 129 packages Nov 28 18:17:12 Is bugs.treke.net down again? Nov 28 18:22:18 Why is OE always restarting from scratch the calculation of dependencies for the bb files instead of using the cache when I change the distro? Does not seem logical for me since there are separate caches being used for the separate distros (I have $DISTRO set for the tmp dir). Nov 28 18:28:59 any fpga people around ? Nov 28 18:32:37 03hrw 07org.oe.dev * r755167f1... 10/ (1 packages/libexif/libexif_0.6.13.bb): libexif: stage few more headers in 0.6.13 Nov 28 18:36:26 "This is intended to let users write their own "FNapps" for the card, and compile them using GCC-4.3.4. " Nov 28 18:36:33 linuxdevices.com at their best Nov 28 18:37:00 hi koen Nov 28 18:37:05 hey rob_w Nov 28 18:37:12 can hint me please ... Nov 28 18:37:37 so i need .dreambox to get ppc stuff .. but i was not able to locate a MACHINE which is actually ppc in htere .. Nov 28 18:38:01 you don't need .dreambox for ppc stuff Nov 28 18:38:24 ah , that would go along my logics Nov 28 18:38:39 so which MACHINE in oe is actually a ppc ? Nov 28 18:38:45 hrw|gone: why bump PR on libexif? The resulting package didn't change, right? Nov 28 18:38:51 rob_w: efika Nov 28 18:39:29 koen: but libgphoto2 2.2.1 (not yet pushed) will break without it Nov 28 18:39:44 rob_w: dreambox Nov 28 18:40:01 koen: so I bumped it and had in queue to push Nov 28 18:40:24 argos-wlan:~/Projects/OpenEmbedded/org.openembedded.dev/conf/machine koen$ grep powerpc * | awk -F":" '{print $1}' | sort | uniq Nov 28 18:40:24 dht-walnut.conf Nov 28 18:40:24 efika.conf Nov 28 18:40:24 lite5200.conf Nov 28 18:40:32 hrw|tv, file not found Nov 28 18:40:53 rob_w: ignore that.. .dev you was talking.. Nov 28 18:41:43 * rob_w ignores Nov 28 18:42:25 so today is the day that i first booted a ppc kernel , knowingly Nov 28 18:42:48 the proc/cpuinfo is 6 lines "huge" Nov 28 18:42:53 new day, new toolchain version.. Nov 28 18:46:31 time for some background music, might get a longer day Nov 28 18:50:10 ~lart toolchains rebuilding Nov 28 18:50:11 * ibot_ drops a humongous exploding nuke on toolchains rebuilding Nov 28 18:51:02 hrw|tv, but the learn effect is tremendous Nov 28 18:51:44 ;) Nov 28 18:51:51 hi all Nov 28 18:55:26 hi cyrilRomain Nov 28 18:55:41 questions me if gcc knows a mcpu=885 ?!? or such Nov 28 19:01:24 koen: ERROR: No providers of runtime build target hal-info (for ['/home/hrw/devel/oe/org.openembedded.dev/packages/hal/hal_0.5.8.1.bb']) Nov 28 19:08:25 ERROR: Task /home/hrw/devel/oe/org.openembedded/packages/update-alternatives/update-alternatives-dpkg-native_1.13.22.bb (do_package) has circular dependency on /home/hrw/devel/oe/org.openembedded/packages/dpkg/dpkg-native_1.13.22.bb (do_package) Nov 28 19:09:34 03koen 07org.oe.dev * rb9ba4abc... 10/ (1 packages/hal/hal_0.5.8.1.bb): hal 0.5.8.1: surpress dependencie on hal-info Nov 28 19:10:06 koen: why not add hal-info? Nov 28 19:10:38 when they release a tarball Nov 28 19:10:43 ok Nov 28 19:11:16 ~lart the one who added dpkg^wupdate-alternatives-dpkg-native_1.13.22.bb Nov 28 19:11:16 * ibot_ explains, ever so gently, that if the one who added dpkg^wupdate-alternatives-dpkg-native_1.13.22.bb doesn't give the channel more information, they can't help Nov 28 19:11:31 koen: angstrom is not buildable here Nov 28 19:11:39 it is over here Nov 28 19:11:51 and my crystal ball is in repairs Nov 28 19:11:54 ;) Nov 28 19:12:00 bitbake apt then Nov 28 19:12:04 or dpkg Nov 28 19:12:15 angstrom doesn't use that Nov 28 19:12:26 ok. then I will revert my words Nov 28 19:13:03 irc needs a revert operation Nov 28 19:14:19 NOTE: Applying patch 'autofoo.patch'Command Error: exit status: 1 Output: Nov 28 19:14:27 that's what happens to apt Nov 28 19:15:02 not here ;d Nov 28 19:35:34 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=398381 Nov 28 19:35:38 can be interesting Nov 28 19:40:56 03hrw 07org.oe.dev * rad9a12c3... 10/ (3 files in 2 dirs): libpcap: added 0.9.5 Nov 28 19:53:36 47 unknown files in metadata.. Nov 28 19:54:02 03hrw 07org.oe.dev * r15b6f453... 10/ (1 packages/netperf/netperf_2.4.2.bb): netperf: added 2.4.2 Nov 28 19:54:21 when I use the mtd update... does it also update the kernel stuff to the most current hh.org kernel source? Nov 28 19:54:38 hey Nov 28 19:54:52 boah I'm hopelessly overworked :( Nov 28 19:54:56 that's life ;) Nov 28 19:55:19 yes Nov 28 19:55:30 woglinde: hey Nov 28 19:55:42 <---- yes to me? or aggreance to statement about life? Nov 28 20:00:32 Laibsch: If you changed the path to tmp and the cache is within tmp, it will reparse ;-) Nov 28 20:00:43 hi zecke Nov 28 20:01:10 hey Nov 28 20:01:40 hi rp Nov 28 20:01:59 hi woglinde Nov 28 20:03:08 k... here is an interesting question... maybe someone can point me in a direction... but how do I specify "what" gets compiled into a kernel ... I'm guessing from within a machine .conf ....? Nov 28 20:03:25 usually from the defconfig file Nov 28 20:03:58 defconfig? = default config? or ??? Nov 28 20:04:28 hrw|tv: 25k calls to gettimeofday :) Nov 28 20:06:22 http://pastebin.ca/260940 Nov 28 20:07:01 I get 160 skipped Nov 28 20:07:57 but I removed something qmake related to stop qpe crap from messing up the build Nov 28 20:07:57 s/removed/mtn mv -e foo packages/nonworking/foo Nov 28 20:08:46 I use mtn drop Nov 28 20:08:46 RP http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=141&logfile=20061128072708.log fetch failure on setserial Nov 28 20:09:01 koen: notice 3990 packages Nov 28 20:09:32 koen: I'm still unable to fetch the ep93xx-kernel, is it fetchable for you? Nov 28 20:09:39 I might need to ask to open more ports :) Nov 28 20:09:46 http://pastebin.ca/260946 is my fetch.sh Nov 28 20:09:59 NOTE: Parsing finished. 3847 cached, 0 parsed, 160 skipped, 0 masked. Nov 28 20:10:24 hrw|tv: you miss a way to deprecate stuff :) Nov 28 20:10:39 zecke: ? Nov 28 20:10:46 ~lart 300K/s Nov 28 20:10:46 * ibot_ whips out a hot clue gun and makes sure that 300K/s is stuck to the floor Nov 28 20:11:08 hrw|tv: remove unused source packages Nov 28 20:11:31 zecke: that will be another script Nov 28 20:11:49 zecke: I got info 'we have 500G space, removing can be later' Nov 28 20:12:16 zecke: what part fails to fetch? Nov 28 20:12:37 koen: tinderclient.bbclass is too crappy to say that :) Nov 28 20:12:47 koen: I'm fetching now Nov 28 20:12:55 NOTE: fetch http://ftp.kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.18-rc1-git9.bz2;patch=1 Nov 28 20:12:57 RP: Can that be avoided by moving the cache out of tmp? Will that create problems if one changes distros or machines? Nov 28 20:12:59 12:36:23 ERROR 404: Not Found. Nov 28 20:13:12 zecke: *.gz, *.bz2, *.patch, *.diff is at http://www.angstrom-distribution.org/unstable/sources/ Nov 28 20:13:57 koen: I will check tomorrow :) Nov 28 20:14:02 koen: you need to add /old/ Nov 28 20:14:10 http://ftp.kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.18-rc1-git9.bz2 -> http://ftp.kernel.org/pub/linux/kernel/v2.6/snapshots/old/patch-2.6.18-rc1-git9.bz2 Nov 28 20:14:36 hrw|tv: your qemu needs to remove linux/compiler.h Nov 28 20:14:43 hrw|tv: true, but I'm more interested why it doesn't try the fallbacks in CVS_TARBALL_STASH Nov 28 20:15:04 koen: because CVS_TARBALL_STASH is not in MIRRORS Nov 28 20:15:04 koen: hmm, okay I will look into the machine now Nov 28 20:15:25 zecke: patches welcome Nov 28 20:15:48 koen: CVS_TARBALL_STASH is for SCM snapshots, not for files Nov 28 20:15:52 MIRRORS is for files Nov 28 20:16:02 it used to work for files as well Nov 28 20:16:06 when? Nov 28 20:16:22 ~3 months ago Nov 28 20:16:23 in .oz354x I just added ewi to MIRRORS Nov 28 20:17:34 03koen 07org.oe.dev * r22940140... 10/ (1 packages/linux/ep93xx-kernel_2.6.17+2.6.18-rc1.bb): ep93xx-kernel: fix src uri for kernelpatch Nov 28 20:21:16 Is there a good reason not to set CACHE = "${PATH2OE}" where ${PATH2OE} is the root OE directory? Will this put the cache file in ${PATH2OE}? Will this create problems? Nov 28 20:22:36 hrw|tv: so what's the best way to add a mirror without touching base.bbclass? Nov 28 20:23:13 koen: pcap fails badly as well :) Nov 28 20:23:22 hmm Nov 28 20:23:22 zecke: libpcap 0.9.5? Nov 28 20:23:46 yes Nov 28 20:24:37 which tree? Nov 28 20:24:38 koen: and for ep93xx kernel, see hrw's remark Nov 28 20:24:51 right Nov 28 20:25:08 so what's the best way to add a mirror to a distro without touching base.bbclass? Nov 28 20:27:11 zecke: which tree failed on libpcap? Nov 28 20:27:46 hrw|tv: angstrom Nov 28 20:28:15 hrw|tv: http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=147&logfile=20061128212715.log Nov 28 20:28:32 ---> TINDERBOX Task do_patch started Nov 28 20:29:32 hrw|tv: and then bitbake raised an exception and exited (crashed) :) Nov 28 20:30:59 zecke: libpcap 0.9.5 works for me Nov 28 20:31:47 ok. failed Nov 28 20:32:29 ~curse monotone Nov 28 20:32:32 May you be reincarnated as a Windows XP administrator, monotone ! Nov 28 20:32:50 ljp: instead of source, we get more varpoware? what is wrong with TT? Nov 28 20:33:22 * hrw|tv request mtn add -r Nov 28 20:33:47 in 0.30 and older 'mtn add .' added ALL recursive Nov 28 20:34:00 0.31 need 'mtn add `find .`' instead Nov 28 20:34:23 03hrw 07org.oe.dev * reaf6133c... 10/ (1 packages/libpcap/libpcap-0.9.5/shared.patch): libpcap: added forgotten patch Nov 28 20:34:38 zecke: will work now Nov 28 20:38:11 NOTE: package libpcap-0.9.5: completed Nov 28 20:47:08 hrw|tv: hhm, is he interested in Free Software? Nov 28 20:47:46 #opie guy? dont know Nov 28 20:47:56 he came few lines before you Nov 28 20:48:26 hrw|tv: yes, sounds like free, cheap support for Qtopia Nov 28 20:48:50 zecke: how many people does not differentiate opie<>qtopia? Nov 28 20:49:34 too many, this is why I stopped working on it Nov 28 20:53:40 GMT is 21:19 now? Nov 28 20:53:52 zecke: source code is overrated when you have a marketing dept. Nov 28 20:54:02 koen: or a Greensuite Nov 28 20:54:25 koen: so we have a green Phone, green Suite, so only green shoes and green shoes are missing? Nov 28 20:54:37 "Holger Freyther announces a complete frikandel stack" Nov 28 20:54:59 and Green Soylent Nov 28 20:55:00 sorry for the chicken out on t-dose, unexpected tests suck :( Nov 28 20:55:32 koen: Holger Freyther announces successfull enforcement on the frikandel patent :) Nov 28 20:56:30 soylent green is people Nov 28 20:57:27 IPKG_INSTALL governs what ipks are installed on image? Nov 28 20:57:40 will these packages RDEPENDS also get installed? Nov 28 20:57:53 "yes" and yes Nov 28 20:58:02 thanks Nov 28 20:58:16 I forget so fast Nov 28 20:58:47 re Nov 28 20:58:49 IPKG_INSTALL will be moved to PACKAGE_INSTALL iirc Nov 28 21:01:22 22:00 PST^A^K04:00 GMT will my first autofetch start... Nov 28 21:05:13 libass/ass.c ;) Nov 28 21:13:31 CosmicPenguin: interested in xfce 4.4-rc1? Nov 28 21:13:42 or rc2 even Nov 28 21:14:06 Not super interested Nov 28 21:14:09 at this point Nov 28 21:14:13 but i'm sure somebody would be Nov 28 21:14:20 I have them built but not tested yet Nov 28 21:14:27 ptxdist is kinda odd Nov 28 21:17:00 03koen 07org.oe.dev * r64b62e71... 10/ (1 packages/python/python-pygtk2_2.10.3.bb): python-pygtk2 2.10.3: stage files needed for codegen as well Nov 28 21:18:54 what monotone version is the OE.db nowadays? Nov 28 21:22:08 ~lart mplayer Nov 28 21:22:09 * ibot_ takes out a seltzer bottle and sprays mplayer in the face. You know, one of those old-school seltzer bottles clowns have? Yeah those. Anyway, consider yourself spritzed Nov 28 21:27:35 03koen 07org.oe.dev * r9e630059... 10/ (1 packages/python/python-pygtk2_2.10.3.bb): python-pygtk2 2.10.3: stage even more defs... Nov 28 21:28:23 ~lart python Nov 28 21:28:24 * ibot_ grabs a large, mis-shapened log, with squirrels, and beats python until only the nuts remain ... which the squirrels run off with Nov 28 21:28:29 yesss! Nov 28 21:29:22 grrr Nov 28 21:29:33 python-pygnome needs gnome-vfs >= 2.13.2 Nov 28 21:29:52 florian: do you know where the gnome-vfs-dbus 2.16 port source is located? Nov 28 21:30:22 koen: not offhand, movial svn maybe? Nov 28 21:30:37 it used to be in immendio svn Nov 28 21:32:00 florian, are u visiting SPS Drives in Nürnberg ? Nov 28 21:32:20 koen: it may have migrated to one of the numerous maemo.org repos.. Nov 28 21:32:45 koen: eh well... that's what i meant :-} Nov 28 21:32:50 http://cvs.gnome.org/viewcvs/gnome-vfs/configure.in?rev=1.432&view=markup instills hope Nov 28 21:32:53 god Nov 28 21:33:15 rob_w: unlikely... unless it is next year Nov 28 21:33:15 I got pissed of by the "the buildbot can only build from maemo svn" comment from carlos Nov 28 21:33:45 koen, have you ever considered a class in anger management? :) Nov 28 21:34:01 they took away my cluebat Nov 28 21:34:04 never again! Nov 28 21:34:07 :p Nov 28 21:34:24 luckily you can always make a new cluebat Nov 28 21:34:38 maybe a class in woodworking then? Nov 28 21:35:09 hehe Nov 28 21:35:53 koen: or maybe just lart france a little bit an make him remove your old homedir at hh.org ;) Nov 28 21:36:07 didn't he do that already? Nov 28 21:36:37 koen: the gpe wiki pulled the banner from there last week... Nov 28 21:37:40 http://cvs.gnome.org/viewcvs/gnome-vfs/?only_with_tag=vfs-dbus Nov 28 21:37:41 koen: someone closed it for a while when the lig wiki got it from the same location, but it re-appeared when i fixed the ltg wiki Nov 28 21:37:56 it looks like *something* vfs-dbus related is in gnome cvs Nov 28 21:41:27 anyone gotten parted 1.8.0 to work? compiles nicely, but gives error message when trying to run Nov 28 22:03:41 * v8jlene is back Nov 28 22:10:38 afternoon Nov 28 22:12:39 I just checked in to see what was going on with the bugzilla server Nov 28 22:13:28 Right now it's looking like a routing issue with the new ISP for the machine. I'm hoping they get it sorted out this afternoon. Nov 28 22:16:07 hi, all! Nov 28 22:16:26 is there a package where usb.ids and lsusb reside? Nov 28 22:19:59 usbutils? Nov 28 22:26:32 yeah, thanks! Nov 28 22:27:40 woglinde: added minix to OE yet? Nov 28 22:29:11 bye Nov 28 22:30:13 zecke: I'll check out setserial Nov 28 22:30:27 :) Nov 28 22:31:41 * RP celebrates working maemo-mapper-nohildon :) Nov 28 22:32:01 but I can't have wifi and bt in the zaurus at the same time :-( Nov 28 22:33:07 RP: yay, and you can on a spitz/akita series Nov 28 22:33:23 XorA|gone: how? Nov 28 22:34:22 RP: BT usb dongle, CF wifi Nov 28 22:34:42 * RP had forgotten about the usbhost :} Nov 28 22:34:57 I may even have a USB bt dongle somewhere... Nov 28 22:35:12 my safecom runs without a powered hub Nov 28 22:37:05 ~lart the terrier on/off switch location Nov 28 22:37:06 * ibot_ gives the terrier on/off switch location a good seeing to Nov 28 22:37:43 usbhci is easier than CF BT as well Nov 28 22:38:03 XorA|gone: Mine TDK dongle appears to run without a hub too :) Nov 28 22:41:29 the usb dongle works :) Nov 28 22:41:36 * RP is happy ;) Nov 28 22:42:10 RP: now go out and map the UK Nov 28 22:42:32 XorA|gone: On the bike, with pleasure ;-) Nov 28 22:43:37 RP: http://wiki.openstreetmap.org/index.php/Main_Page Nov 28 22:47:28 XorA|gone: I've seen it before. I might end up submitting traces, we'll see... Nov 28 22:49:06 anyway time for me to sleep so I can beat up samsung chips tomorrow Nov 28 22:49:14 rp: better yet, Nov 28 22:49:30 just figure out the length of the coastline of england Nov 28 22:50:16 'night XorA|gone (and thanks for the bt tips :) Nov 28 23:55:00 chalk this one up to famous last words Nov 28 23:55:03 but I think it's fixed Nov 29 01:03:01 hrw|gone: yeah, I know about those...haven't gotten around to fixing yet. I suspect it's the sed line in the bbclass **** ENDING LOGGING AT Wed Nov 29 02:59:58 2006