**** BEGIN LOGGING AT Sun Sep 17 02:59:57 2006 Sep 17 09:46:15 morning all Sep 17 09:46:22 hey lrg Sep 17 09:46:49 hi all Sep 17 09:47:09 hey koen, marcin Sep 17 09:47:27 hi Zero_Chaos Sep 17 09:47:30 zecke: hi Sep 17 09:47:40 hey zecke Sep 17 09:48:51 building oz 3541/collie progressing.. 3rd day Sep 17 10:00:13 hi, I'm trying to bitbake angstrom but I getting into troubles when building glibc-2.4: I'd be grateful if somebody could help me Sep 17 10:00:53 *hex* *hex* and your troubles should be away :) Sep 17 10:01:51 hi Sep 17 10:02:05 I get the folowing error: Sep 17 10:02:26 Inconsistency detected by ld.so: rtld.c: 306: _dl_start_final: Assertion `info->l.l_tls_modid == 0' failed! Sep 17 10:02:26 NOTE: Task failed: localedef returned an error. Sep 17 10:02:26 NOTE: package glibc-2.4-r10: task do_package: failed Sep 17 10:02:26 ERROR: TaskFailed event exception, aborting Sep 17 10:02:26 NOTE: package glibc-2.4: failed Sep 17 10:02:43 but at first: Sep 17 10:02:53 NOTE: preferred version 2.4 of glibc not available Sep 17 10:03:37 cristiano-p: I assume, as you do not specify it, that ld.so is your host linker? Sep 17 10:03:44 cristiano-p: are you using Gentoo? Sep 17 10:03:51 mandriva Sep 17 10:04:36 cristiano-p: well, from the limited input I would say your linker (ld.so) of your system is broken Sep 17 10:11:27 we have 2006 year and there are still companies which create own one cross-platform UI libraries.. Sep 17 10:11:49 I just started one of such ones.. nightmare is not strong enough word.. Sep 17 10:13:22 zecke: I'm quite a newbie: which information do you need? Sep 17 10:18:03 was there some other recent core change which means that the image directory does not get created automatically any more? Sep 17 10:18:28 (nslu2-linksys-libs, which has not been changed in over a year, suddenly fails in the last couple of days sometime) Sep 17 10:18:52 morning Sep 17 10:19:00 perhaps another thing like the bad change which caused do_install to be called twice? Sep 17 10:19:12 morning all Sep 17 10:19:56 hey RP Sep 17 10:20:06 hi RP Sep 17 10:20:08 actually, it looks like ${D} is not being created Sep 17 10:20:09 hi CoreDump|home Sep 17 10:20:41 hey Richard Sep 17 10:20:46 the do_install for nslu2-linksys-libs simply does a tar c from ${S} piped to a tar x in ${D}, and it now fails. Sep 17 10:21:04 So there is definitely some major bug in core OE now causing this. Sep 17 10:21:28 rwhitby: I think it relies on directories that are not being there Sep 17 10:21:32 (nslu2-linksys-libs does not inherit any other bb files) Sep 17 10:21:40 koen: ${D} has been there for two years now Sep 17 10:22:10 suddenly now it is not, when do_intall is called Sep 17 10:22:27 so a install -d ${D}/whatever should solve it Sep 17 10:22:35 age is not a deciding factor is it's correct or not, mind you Sep 17 10:22:38 lrg: I have reports that cat ing a file to /dev/dsp used to work on spitz but doesn't now - any ideas if anything has changed there? Sep 17 10:22:57 it should, yes. The question is whether it's an intentional change or not? Sep 17 10:23:42 RP: I'll check later today on Mainstone. I've not managed to build poodle yet. Sep 17 10:23:55 If it's intentional, then it will break quite a few packages probably. That's fine - I can deal with that, as long as I know it's an intentional incompatible change. Sep 17 10:24:10 RP: oss emulation is rather horrible :( Sep 17 10:24:28 lrg: ALSA sound/core/oss/pcm_oss.c:574: snd_pcm_oss_prepare: SNDRV_PCM_IOCTL_PREPARE failed Sep 17 10:24:28 ALSA sound/core/oss/pcm_oss.c:574: snd_pcm_oss_prepare: SNDRV_PCM_IOCTL_PREPARE failed Sep 17 10:24:36 koen: can you confirm it's an intentional incompatible change to the preconditions that do_install expects? Sep 17 10:24:36 lrg: I know :-( Sep 17 10:25:02 lrg: Could it be 32 bit related? Sep 17 10:25:37 lrg: Is poodle still failing on dbus? Sep 17 10:25:55 RP: not sure. It could be realted to channels, srate, bit size, etc Sep 17 10:26:19 lrg: I can probably find a poodle image to get you going if that would help Sep 17 10:26:27 RP: yes, stiil at dbus. Will try again on Monday. My time is a bit limted atm until Graeme starts Sep 17 10:26:29 rwhitby: I have no idea on that, I just object to the "it's been there for a long time, therefore it's correct" assertion Sep 17 10:26:29 rwhitby: and the recent re-addition of the double do_install proves that people don't really test their stuff Sep 17 10:26:38 RP: thanks :) Sep 17 10:26:40 lrg: np, I understand Sep 17 10:27:16 koen: I said, it's been there a long time, so it's a change to core which has happened, not a change to the package. You're right, I should not have characterised it as a "bug" in OE, just a change. Sep 17 10:27:26 RP: could you post the image on rpsys and I'll download on Monday (poodle is at work atm) Sep 17 10:28:13 I'm happy to accept incompatible changes to preconditions in the name of correcting past bugs. I just want to know whether it's intentional or not before I go adding "install -d ${D}" to lots and lots of packages ... Sep 17 10:28:29 rwhitby: I have no idea if the change was intentional or not, since I spent 2 days figuring out why do_install started blowing up again Sep 17 10:29:04 koen: ok, thanks. In that case, I will assume it's not intentional, and wait for it to be fixed, cause it's bound to affect lots of people other than me. Sep 17 10:29:23 rwhitby: your best bet would be to send a note to the mailinglist, and see if the 'guilty' person can explain his actions Sep 17 10:30:05 Hmm. My working poodle images are from April! Sep 17 10:30:19 koen: thanks for your answers, and sorry to trigger your response about age vs bugs :-) Sep 17 10:30:40 rwhitby: anf FWIW, IMO ${D} should be in do_install[dirs] Sep 17 10:30:49 a removal of that would indeed be a bug Sep 17 10:30:53 s/anf/and/ Sep 17 10:31:14 lrg: http://www.rpsys.net/openzaurus/poodle/ - the image will need the modules updating, I'll try and run off a newer build at some point as that one is rather old (but works) Sep 17 10:31:25 rwhitby: np, I have been fighting a bitrotted uclibc.bb this week :) Sep 17 10:31:29 koen: yes, I agree. Where would I look in the core to start debugging that? package.bb or somewhere else? Sep 17 10:31:31 RP: thanks Sep 17 10:31:50 rwhitby: classes/base.bbclass I think Sep 17 10:32:15 do_install[dirs] = "${S} ${B}" Sep 17 10:32:21 ${B} ? what's that? Sep 17 10:32:49 * rwhitby has never seen ${B} anywhere in OE before ... Sep 17 10:33:19 bitbake.conf:B = "${S}" Sep 17 10:33:22 * koen wonders.... Sep 17 10:33:58 and no mention of ${D} in base.bbclass except in a couple of comments Sep 17 10:34:11 what's the monotone command to show the past history of a single file? Sep 17 10:34:23 mtn log Sep 17 10:34:23 lrg: use OZ 3.5.4.2-rc2 images for poodle Sep 17 10:34:39 rwhitby, koen: Would that change have shown up due to install now becoming before staging? Sep 17 10:34:59 RP: probably Sep 17 10:35:01 lrg: they have older asoc but you can exchange kernel and modules with one ones Sep 17 10:35:13 RP: nothing else in base.bbclass creates ${D} as far as I can see. Sep 17 10:35:16 hrw: ok, thanks Sep 17 10:35:57 rwhitby, koen: do_package[dirs] = "${D}" in package.bbclass Sep 17 10:36:11 Hi Sep 17 10:36:13 shouldn't that be += ? Sep 17 10:36:17 hey sirfred Sep 17 10:36:23 ~lart wine for low speed Sep 17 10:36:24 * ibot calls wine on the phone ... the lights are on but nobody's home for low speed Sep 17 10:36:30 hi sirfred Sep 17 10:36:51 ok, so adding ${D} to do_install[dirs] should be the fix? Sep 17 10:36:58 RP, koen : Hello. Sep 17 10:37:33 rwhitby: I'd like to work out what changed to cause that as its a bit of a worry :-/ Sep 17 10:37:44 RP: Do you know the differences between the IMAGEON 100 and the IMAGEON 3200 Sep 17 10:37:46 ? Sep 17 10:37:49 RP: ok, I'll hold off on a quick fix. Sep 17 10:38:06 sirfred: w100fb probably has the best "documentation" on that Sep 17 10:38:10 * rwhitby is glad to know he's got the heavies debugging it now :-) Sep 17 10:38:14 RP: :( Sep 17 10:38:16 sirfred: spyro did the 3200 work Sep 17 10:39:16 koen: do_install used to be a package function and you uprated it to a task of its own? Sep 17 10:39:24 RP: It's because I've found a windows CE player that is supporting IDCT using the AHI API. Sep 17 10:39:54 sirfred: with sources/ Sep 17 10:39:55 ? Sep 17 10:39:58 hrw: Yes. Sep 17 10:40:01 sirfred: I was always told 3200 supported IDCT and the 100 didn't but I've no idea how true that is Sep 17 10:40:01 wow Sep 17 10:40:29 sirfred: It'll be nice to see an implementation :) Sep 17 10:40:29 RP: I think that the 3200 ATI provided API is supporting IDCT. Sep 17 10:40:47 Let me find the link again... Sep 17 10:42:41 http://corecodec.org/frs/?group_id=53&release_id=277#r277 Sep 17 10:43:05 That source code has a file called ati3200/ati3200.c Sep 17 10:43:28 Where they're making use of the AHI API to perform IDCT + MC on a IMAGEON 3200 Sep 17 10:43:49 sirfred: I have a w100 and a w3200 if you need some testers Sep 17 10:44:20 sirfred: interesting. I wish I had some time to fiddle :) Sep 17 10:44:23 I want to think that perhaps the two chips has a similar set of registers, and having in mind that the first AtiCore versions for IMAGEON 100 didn't support IDCT, perhaps it was a marketing decision. Sep 17 10:44:46 koen: Why was do_install removed from PACKAGE_FUNCS? Sep 17 10:44:49 They just dropped IMAGEON 100 support. THen, released the new AHI API, and the Imageon 100 IDCT support ... Sep 17 10:45:37 I'm afraid that with the current knowledge of the IMAGEON 100, we cannot think about speed up video playing too much. Sep 17 10:45:46 sirfred: :-/ Sep 17 10:46:15 I think that the key for that is reducing bandwidth, and we're not reducing bandwidth just overlaying YUV surfaces. Sep 17 10:46:29 AT least YUV422, they are also 16bpp, afaik. Sep 17 10:47:13 sirfred: not just bandwidth but processor load. If the processor has to convert YUV -> RGB and we remove the need to that, things will run faster Sep 17 10:47:35 RP: The problem with that is that scaling is very limited. Sep 17 10:47:42 bye all Sep 17 10:47:58 RP: 1.-I haven't found a way to rotate an YUV surface. It seems to work only for RGB surfaces. Sep 17 10:48:06 So, the video must be prerotated. Sep 17 10:49:03 RP: 2.-YUV surfaces scaling, using the VIDEO_CTRL fields video_hor_exp and video_ver_exp are limited to integer scaling( x2, x3 and x4) Sep 17 10:49:15 sirfred: I see the problem :-( Sep 17 10:49:20 RP: I think that it's pretty useless. Sep 17 10:50:00 afk Sep 17 10:50:36 We cannot scale video to fullscreen, for example. The only implementations I've seen for the IMAGEON100 are just converting to rgb when they need to scale properly, or to an YUV planar, where U and V surfaces are totally independent. Sep 17 10:50:50 sirfred: There are two ways to look at it. We can make certain cases faster so things can be improved, but we can't speed up the most useful cases :-( Sep 17 10:51:18 RP: Yes, it's that sad. Sep 17 10:52:10 RP: The only hope I have now is that perhaps in ACE_DDI.DLL we could "learn" new things, that could apply to the IMAGEON 100. Sep 17 10:52:46 It should be able to perform IDCTs, I'm pretty sure, because the last Aticore DLL we have, is implementing it. Sep 17 10:52:55 But, with that poor scaling. It can not be true. Sep 17 10:55:53 sirfred: I suspect they realised the flaws with the 100 and then made the 3200 to correct them :-/ Sep 17 10:56:12 That's what I don't want to think. :) Sep 17 10:56:20 But it's probably true. Sep 17 10:57:48 rwhitby: I agree with adding the do_install[dirs] = "${D}" although I can't see why we suddenly need to do it... Sep 17 10:58:20 RP: in was removed from PACKAGE_FUNCS since it was being called twice Sep 17 10:58:29 RP: nobody objected to that change Sep 17 10:58:41 RP: till kergoth added it back and stuff started breaking Sep 17 10:59:16 koen: Thanks for your offer for testing. Sep 17 10:59:16 koen: How did it end up in the two locations? Sep 17 10:59:27 koen: I don't object to the change, I'm just wondering why it happened Sep 17 10:59:56 RP: I think it always was in two locations and the do_populate_staging change made it a problem Sep 17 11:00:18 the staging change + zecke's PACKAGE_FUNC change combined Sep 17 11:00:19 koen: Before the do_populate_staging change, did it run twice? Sep 17 11:00:48 What was zeckes PACKAGE_FUNC change? :-/ Sep 17 11:00:49 I'm not sure Sep 17 11:01:38 it was a while ago that he made that change Sep 17 11:02:35 Interestingly, do_install just says to run after do_compile, so it could have run after do_package and we'd not see an issue Sep 17 11:03:05 in fact do_install probably didn't run at all Sep 17 11:03:14 (by default) Sep 17 11:05:59 Ah, and adding addtask package before do_build after do_install was the killer change. Before that, do_install wasn't part of do_build Sep 17 11:06:36 if back then you'd s/do_install/do_compile/, we'd not have this problem now Sep 17 11:07:01 Not that I'm blaming anyone, just trying to undertstand it. The tasks are a mess Sep 17 11:07:13 so we should remove it from base.bbclass and add it to PACKAGE_FUNCS? Sep 17 11:08:03 We can have two choices - leave as a separate task, or put it back as a PACKAGE_FUNCS Sep 17 11:09:14 Adding the do_install[dirs] = "${D}" will fix things with no ill effects and I'm probably in favour of leaving it as a separate task. Providing, nothing in PACKAGE_FUNCS needs to run before do_install Sep 17 11:10:15 Since it was the first PACKAGEFUNCS, that shouldn't be a problem Sep 17 11:10:45 someone wants to 'finish' the slides for me? Sep 17 11:12:50 * RP -> back shortly Sep 17 11:12:59 RP: http://rafb.net/paste/results/UQ2Fad73.html ? Sep 17 11:16:54 rwhitby: mtn pull :) Sep 17 11:17:18 03koen 07org.oe.dev * re422dfdd... 10/ (1 classes/base.bbclass): base.bbclass: add ${D} to do_install[dirs] to fix some issues introduced by the task reordering Sep 17 11:36:10 koen: I created a patch that removes five occurences of ASSA:ERROR in the source of granule (cf. http://oz.leggewie.org/granule/files/remove-ASSA_ERROR.patch), still it will not compile. My current work is again at http://oz.leggewie.org/granule/ Sep 17 11:37:34 Laibsch: deckview and topmenubar require two lines to be commented Sep 17 11:37:51 Laibsch: the ASSA_ERROR statement spans 2 lines in those files Sep 17 11:38:19 koen: thx Sep 17 11:38:56 koen: Looks good. Sep 17 11:39:03 * RP ponders ${B} Sep 17 11:39:43 RP: check glibc* Sep 17 11:40:10 I think it stands for ${B}uilddir Sep 17 11:40:20 where builddir != sourcedir Sep 17 11:40:37 koen: I was just looking at glibc having fired up grep. I think you're right :) Sep 17 11:40:39 RP: hmm, somehow glibc2.4 really got used for the poky x86 build :( Sep 17 11:40:49 why would do_install create builddir? surely that should exist well before do_install? Sep 17 11:40:54 RP: now sqlite (from OE) fails to link, and I have no clue at all... Sep 17 11:41:10 rwhitby: I have no idea Sep 17 11:41:36 zecke: For me, it built glibc-initial-2.4 but then build glibc-2.3.5+cvs... Sep 17 11:42:02 Laibsch: btw, did you see http://anderssonj.com/jards already ? Sep 17 11:42:03 RP: yes but my staging/lib/libc.so is 2.4 :( Sep 17 11:42:17 zecke: Did it build both? Sep 17 11:42:57 RP: it looks like it Sep 17 11:43:28 glibc-2.3.5+cvs20050627-r3/ glibc-initial-2.4-r10/ glibc-intermediate-2.4-r10/ Sep 17 11:43:30 hmm Sep 17 11:43:33 koen: Yes, that is where I started the work on jards from. Johnny promised to install OE and continue my work on getting it into OE. He has also uploaded his stuff to http://gakusei.sf.net Sep 17 11:44:10 the source into CVS. Sep 17 11:44:26 zecke: Strange :-/ Sep 17 11:46:16 yes Sep 17 11:50:00 RP: I have removed glibc2.4 and will rebuild Sep 17 11:51:03 ok, next problem: ipkg-0.99.154-r4 fails in do_install Sep 17 11:52:10 re Sep 17 11:53:29 NOTE: make DESTDIR=/home/slug/unslung/tmp/work/ipkg-0.99.154-r4/image install Sep 17 11:53:29 make[3]: Entering directory `/home/slug/unslung/tmp/work/ipkg-0.99.154-r4/image' Sep 17 11:53:33 make[3]: *** No rule to make target `install'. Stop. Sep 17 11:53:41 I wonder why do_install is trying to run from inside the image dir? Sep 17 11:54:36 maybe it's time to finally update nslu2-linux to the latest ipkg version ... Sep 17 11:54:49 argh.. someone remember how to burn CD for msdos? Sep 17 11:58:01 I see files but trying to access any of them fails 'file not found' Sep 17 11:59:07 hrw: disable joliet? Sep 17 11:59:32 has the rearrangement of tasks meant that the current directory has changed when do_install is called? Sep 17 11:59:47 koen: joliet and RR disabled Sep 17 11:59:50 rwhitby: it shouldn't Sep 17 12:00:04 koen: tried with RR/without joliet and it was same ;( Sep 17 12:00:22 hrw: and don't have too many files in \ Sep 17 12:00:25 I want to install win98 under qemu Sep 17 12:00:27 Hmm - the last two lines of run.do_install are: Sep 17 12:00:34 cd /home/slug/unslung/tmp/work/ipkg-0.99.154-r4/image Sep 17 12:00:34 do_install Sep 17 12:00:36 koen: 64 limit iirc Sep 17 12:00:44 other than that I hae no idea how get CDs working in msdos Sep 17 12:01:21 recreating again Sep 17 12:01:27 Four days ago, the last two lines were: Sep 17 12:01:34 ~lart gpe-beam for not building Sep 17 12:01:34 * ibot pries gpe-beam's back open with a screwdriver and flashes a new bootldr to gpe-beam for not building Sep 17 12:01:34 cd /home/slug/debianslug/tmp/work/ipkg-0.99.154-r4/ipkg/C Sep 17 12:01:34 do_install Sep 17 12:01:48 (which makes a lot more sense) Sep 17 12:02:45 * rwhitby goes to bed, and hopes that do_install is run from the build directory again tomorrow after an mtn pull. Sep 17 12:10:18 rwhitby: I think that is the result of koen's change - dirs probably does a cd as well as creates the dir :-/ Sep 17 12:10:39 Perhaps ${D} ${S} will work better... :-/ Sep 17 12:10:56 03koen 07org.oe.dev * rc4b82dde... 10/ (1 classes/base.bbclass): base.bbclass: attempt to fix the case where do_install wants to run from inside ${D} Sep 17 12:11:09 like that? Sep 17 12:11:46 koen: yes :) Sep 17 12:16:51 * koen adds some points to http://www.openembedded.org/wiki/OEDEM Sep 17 12:23:05 koen: you disabled diff module? Sep 17 12:23:25 hrw: I didn't Sep 17 12:24:00 grrr Sep 17 12:24:04 css is broken again Sep 17 12:24:41 koen: I do not see diff 'tab' Sep 17 12:25:30 hrw: your cookie expired? Sep 17 12:25:36 maybe Sep 17 12:26:15 koen: after relogin its same Sep 17 12:28:33 hi :) Sep 17 12:29:05 hey Marex Sep 17 12:29:28 morning Sep 17 12:29:42 hey chouimat Sep 17 12:29:52 uf.. finally win98 install in qemu works Sep 17 12:56:10 koen: It comiles now! Sep 17 12:56:27 Let's see if it works. Sep 17 12:56:57 it = granule Sep 17 12:57:31 BTW, the granule guy was not interested in OE when I contacted him and said that maemo was rather the way to go. Maybe he will reconsider now. Sep 17 12:57:44 I think you can also compile for maemo on OE, can't you? Sep 17 12:58:59 sort of Sep 17 12:59:20 'no' is probably closer to the truth Sep 17 12:59:44 I see. Sep 17 13:01:31 Speaking about this. How can I switch to compile for another machine (or another distro such as Sharp ROM)? I know the changes need to be made in conf/local.conf but would I need to clean out tmp and use a different deploy directory? Or can I just change the conf file, compile and publish everything at the same place? How do the feed maintainers do it? Sep 17 13:06:15 we mostly use multimachine.bbclass Sep 17 13:06:52 but contrary to popular belief, you can run 99% of the ipks for 'collie' on a clamshell Sep 17 13:19:41 koen: How do you use multimachine.bbclass? Do I need to inherit it in conf/local.conf or what? Sep 17 13:20:12 most likely your distro already inherits it Sep 17 13:20:25 NOTE: package gpe-image-1.0-r24: task do_build: completed Sep 17 13:20:27 uf Sep 17 13:20:59 hrw: with a recent dbus? Sep 17 13:21:21 yes Sep 17 13:21:29 ~lart thunderbird Sep 17 13:21:30 * ibot whacks thunderbird with the cluebat Sep 17 13:21:32 cool Sep 17 13:21:37 ~kill thunderbird Sep 17 13:21:54 * ibot shoots a hyper-charged fluxpositrino gun at thunderbird Sep 17 13:21:56 this shit does not allow to import VCards Sep 17 13:22:09 instead it import csv with broken design Sep 17 13:25:17 koen: So I can just change the machine type in conf/local.conf and worry no further, right? I use openzaurus.unstable. Sep 17 13:25:50 yes Sep 17 13:28:41 Where will the machine-specific stuff end up? I think I already compiled machine specific stuff such as nunome. But http://oz.leggewie.org/ipk is all in a single directory except for morgue. Or will I actually need to compile for a second machine type before the magic is going to happen? Sep 17 13:29:15 why is nunome machine specific? Sep 17 13:29:58 http://vbone.net/oz/akita/ Sep 17 13:30:05 arm5vte is not possible for collie. Sep 17 13:30:12 That is why I thought it is machine specific. Sep 17 13:30:38 AlexTe had to recompile before I could use his packages way back when. Sep 17 13:30:40 only the libqte2 in that dir is machine specific Sep 17 13:31:07 there rest is only specific to armv5 cpu's (every zaurus except collie) Sep 17 13:32:03 I also have http://oz.leggewie.org/ipk/libqte-mt2_2.3.10-r3_collie.ipk which is in the main dir. Sep 17 13:32:23 the word before '.ipk' is the architecture Sep 17 13:32:56 OK, I guess everything should be alright then. Sep 17 13:33:28 multimachine.bbclass seperates your toolchains in OE Sep 17 13:33:37 check work/, cross/ and staging/ Sep 17 13:36:06 Ah, yes, I recognized that. collie-linux is the magic word, I guess. Where will this stuff end up? I see there altboot for example. Sep 17 13:36:53 packages that are *really* machine specific will get built there Sep 17 13:37:14 but since ipkg has an 'architecture' field they all end up in the same deploy/ipk/ dir Sep 17 14:08:19 'mtn pluck' now gives an error message if the requested operation would have no effect. Sep 17 14:08:22 YES:) Sep 17 14:09:09 mickey|city: pdaX does not have release policy at all Sep 17 14:11:54 mickey|city: there were rcXY releases, then beta1/beta2/beta3 (each one support other machines + rebuilds done by other people), then beta4 with u-boot and 2.6 kernel (too many users failed on u-boot instalation, system totally not ready for 2.6 in many places). now work on beta5 is going but hard to get when it will be released. And they do not support spitz (as c3000) in any release. c3x00 use flash still so users complain about lack of easy way to install s Sep 17 14:14:29 whenever i try to compile gdb as part of the sdk-image, i get multiple errors. how do i address these? Sep 17 14:14:29 | iwmmxt.c: In function 'WMAC': Sep 17 14:14:29 | iwmmxt.c:2117: error: invalid lvalue in assignment Sep 17 14:15:20 which gcc? Sep 17 14:15:52 koen: I have a problem. Sep 17 14:16:07 koen: we load local.conf, machine.conf, distro.conf Sep 17 14:16:33 koen: distro cannot tell which kernel provider to use for machine because machine already set it up... Sep 17 14:16:54 hrw: i'm using: gcc (GCC) 4.0.3 20051222 (prerelease) Sep 17 14:17:19 hrw: there are other examples that fail with the include order Sep 17 14:17:30 hrw: what version is used by whoever successfuly compile gdb for the OZ feed? Sep 17 14:17:47 Interrobang: iirc 4.0.3 doesn't have good iwmmxt support Sep 17 14:17:49 Interrobang: googled already? Sep 17 14:17:56 Interrobang: http://sourceware.org/ml/crossgcc/2005-11/msg00155.html ?? Sep 17 14:18:28 zecke: sry, would know which keywords are really key ont his subject Sep 17 14:19:22 Interrobang: I just pasted your iwmmxt.c:2117... line and google for that Sep 17 14:19:46 Interrobang: it doesn't look like a patch was posted Sep 17 14:20:02 Interrobang: http://ewi546.ewi.utwente.nl/tmp/hrw/fb/?action=details&pnm=gdb Sep 17 14:20:07 Interrobang: it is already in the oz feeds Sep 17 14:20:19 Interrobang: either patch yourself, search a patch and don't forget to file a bug report at bugs.openembedded.org and append the patch you have used Sep 17 14:20:21 zecke: i find that exact output searches rarely work since they often incldue version and system specific info Sep 17 14:21:01 Interrobang: "iwmmxt.c:2117: error: invalid lvalue in assignment" was my search string it worked this time... Sep 17 14:21:04 koen: i'm trying to build a sdk-image for Z, i'd manually install the packages myself but i don't which packages i need for a complete compilation environment Sep 17 14:21:32 zecke: didn't work the 100's of other times Sep 17 14:21:40 Interrobang: try http://dominion.kabel.utwente.nl/koen/cms/building-ooo-natively Sep 17 14:23:36 koen: thanks! shold i be using the same version of gcc as OZ? Sep 17 14:24:41 Interrobang: ehm, it will install the stuff from the feeds Sep 17 14:24:53 Interrobang: so I don't get your question Sep 17 14:25:22 koen: i was it might fix my problem on my desktop if i use the same version of gcc as OZ. would it? Sep 17 14:25:34 s/was/was thinking Sep 17 14:42:20 koen: out of curiosity, do you have a twin brother in the US? Sep 17 14:43:00 Interrobang: nope, but I do have family living in CA Sep 17 14:43:16 hi all Sep 17 14:43:16 koen: San Diego, CA? Sep 17 14:43:59 Interrobang: could be, last time I visited they lived in Los Altos, maybe their kids moved there Sep 17 14:44:20 where 'last time' = '1994' Sep 17 14:44:28 koen: you got a twin cousin or something? i once had a class with your doppleganger Sep 17 14:45:37 that would be a few degrees (?) further, since it was my grandfather brother living there Sep 17 14:45:58 Hi! Sep 17 14:45:58 hey psokolovsky Sep 17 14:46:19 i guess human biology just isn't diverse enough for all 6 billion+ of us to have our own unique face Sep 17 14:46:25 :) Sep 17 14:47:05 koen: Did some tweaking to http://linuxtogo.org/gowiki/Angstrom , pls have a look, fix if anything wrong ;-) Sep 17 14:47:16 users complain of empty wiki ;-) Sep 17 14:48:26 it is indeed not the official logo Sep 17 14:48:36 (not that we have an official logo) Sep 17 14:48:55 koen: just took what was on old hh.org page. Sep 17 14:49:06 koen: where's official one? Sep 17 14:49:45 we don't have one yet Sep 17 14:50:24 that sucks ;-) Sep 17 14:50:33 indeed Sep 17 14:51:02 I though that A with the pyramid in background might be it, but I saw it only on screenshots Sep 17 14:51:33 that's what mickeyl made with xara, but we couldn't vectorize it, so we're still searching Sep 17 14:52:31 03hrw 07org.oe.oz354x * rb2098b22... 10/ (1 conf/distro/openzaurus-3.5.4.1.conf): oz 3.5.4.1: enable new DBUS Sep 17 14:52:55 http://en.wikipedia.org/wiki/%C3%85 Sep 17 14:53:06 psokolovsky: we're planning to make an small installer based on uclibc Sep 17 14:53:30 koen: I can make svg img in inscape for example ... *only think that i maybe can :) * Sep 17 14:53:40 koen: That's nice. Sep 17 14:54:07 s/inscape/inkscape Sep 17 14:57:20 koen: Installation of granule fails. It depends on libx11-6 among other things. I guess that means, currently it will only work on a GPE image of OZ, right? If at all. Does OE determine these dependencies and create the ipk accordingly? I have not specified any RDEPENDS. Sep 17 14:57:28 I use an OPIE image. Sep 17 14:57:40 psokolovsky: can you post link to prototype img? Sep 17 14:57:57 Laibsch: RDEPENDS involves magic Sep 17 14:58:11 Laibsch: it checks which libraries you link to and adds them to RDEPENDS Sep 17 14:58:16 Laibsch: yes, OE does that for you Sep 17 14:58:28 den-ros: see gallery on http://www.angstrom-distribution.org/ Sep 17 14:58:31 Laibsch: we aren't like pdaX Sep 17 14:59:01 den-ros, but you really should coordinate with koen before starting yet another logo which may just go to /dev/null Sep 17 14:59:24 psokolovsky: ok :) Sep 17 14:59:29 or even better, send a note to the mailinglist(s) Sep 17 15:01:36 "koen: Laibsch: we aren't like pdaX" -> which is a good thing. I appreciate that OE was seemingly slower to start but things the right way which makes dealing with complexity a lot easier now. Or should I say it really makes it possible? OE is scalable easily which I think is great! Sep 17 15:02:00 s/but things/but doing things/ Sep 17 15:02:10 s/but things/but does things/ Sep 17 15:02:26 I love win98 messages when libraries are missing ;) Sep 17 15:02:38 'device connected to computer does not work' Sep 17 15:02:48 OE was designed to scale, even if it was annoying at the start Sep 17 15:03:14 annoying??? I had to double memory to be able to use it Sep 17 15:04:11 can see some vector logo images, that i make for ros-linux and uOS: http://www.teleology.ru/files/images/ros-linux_logo/lemon11.png http://www.teleology.ru/files/images/ros-linux_logo/lemon10.jpg http://teleology.ru/files/images/ros_logo-64x64.png http://www.teleology.ru/files/images/uos_logo/uos-emblem05.png http://www.teleology.ru/files/teleology_ru-1_logo.png Sep 17 15:05:37 it is convertions from svg Sep 17 15:10:27 * den-ros read http://en.wikipedia.org/wiki/Å Sep 17 15:10:39 I was only able to barely use it after upgrading my complete machine which coincided with the improvements in RAM usage. Sep 17 15:11:48 well ROAD gave me another 1GB of RAM and allowed me to decrease RAM usage Sep 17 15:12:06 The X stuff will not work on my OPIE image, will it? Sep 17 15:12:27 unlikely Sep 17 15:13:19 OK, then I will postpone testing it. If anybody wants to test granule (seems like a smart flashcard soft/vocab trainer), it is available from http://oz.leggewie.org/ Sep 17 15:13:29 I'd appreciate some feedback in that case. Sep 17 15:19:11 hrw: my former boss at the uni said it took a week to finish parsing with 512MB + kde running in vnc Sep 17 15:22:04 den-ros: your interest is much appeciated. it would be nice if subscribed to angstrom ML and be around. Mntioning logo now was mostly random, as devlopers are mostly busy with code so far. But a bit sooner help with logos, images, docs, wikis will be much required. Sep 17 15:22:19 I let the "granule guy" know that his stuff now compiles on OE. koen, if you are intereste: http://sourceforge.net/tracker/index.php?func=detail&aid=1560198&group_id=111131&atid=658403 Sep 17 15:22:47 psokolovsky_: i will try Sep 17 15:22:59 this is driving me crazy, I keep hitting this wall: "checking for GCONF_IPC... Requested 'dbus-glib-1 >= 0.60' but version of dbus-glib is 0.23.4" Sep 17 15:23:30 T0mW: build dbus-glib 0.71 ? Sep 17 15:24:48 Laibsch, it goes info OE.dev too? Sep 17 15:25:26 I'm using generic.conf as the distro, I keep saying in local.conf "PREFERRED_VERSION_dbus = "0.60" Sep 17 15:25:26 " but it ignores me Sep 17 15:26:11 T0mW: http://www.openembedded.org/viewmtn/revision.psp?id=539d974b2217fb30c8b1d400aa8a8c1820a9c2e9 Sep 17 15:26:16 I tried saying PREFERRED_VERSION_dbus-glib = "0.60", but that is ignored Sep 17 15:26:34 yes. Once I got it cleaned up in line with http://www.openembedded.org/wiki/StyleGuide I will submit a bug report to the tracker. Sep 17 15:26:41 psokolovsky_: yes. Once I got it cleaned up in line with http://www.openembedded.org/wiki/StyleGuide I will submit a bug report to the tracker. Sep 17 15:27:13 koen: i did a pull & update yesterday Sep 17 15:27:47 Laibsch, nice! Sep 17 15:28:03 koen: it looks like i've got to hunt down where ever PREFERRED_VERSION_dbus = "0.23" Sep 17 15:28:04 is being stated and comment it out? Sep 17 15:29:10 maybe it is my perception, I keep thinking that the local.conf is "GOD", that whatever I say in there will override anything? Sep 17 15:29:30 T0mW: I think PREFERRED_PROVIDER_dbus-glib = "dbus-glib" would have a better effect Sep 17 15:29:35 or, is local.conf merely a suggestion? Sep 17 15:30:09 koen: ok, what does that mean? PREFERRED_PROVIDER_dbus-glib = "dbus-glib" Sep 17 15:30:14 isn't it redundant? Sep 17 15:30:36 no, since dbus_0.23.4 says PROVIDES = "dbus-glib" Sep 17 15:30:45 ok Sep 17 15:30:47 ... Sep 17 15:31:11 the dbus people split the package in two parts in dbus 0.71 Sep 17 15:31:14 that statement seems to say the same as "#define 5 5" Sep 17 15:31:38 that statement seems to say the same as "#define _dbus-glib _dbus-glib" Sep 17 15:32:04 hey. Sep 17 15:33:27 so, what is: PREFERRED_PROVIDER_dbus-glib = "dbus-glib" saying? I cannot seem to find any docs on this type of statement and yet these are sprinkled throughout bitbake system. To my way of thinking, that statement is a NOP Sep 17 15:33:39 psokolovsky_: Are you interested in the software? Sep 17 15:34:53 Laibsch, yes, but unfortunately neck-deep in build engineering stuff. But now I will know that OE has good flashcard app ;-) Sep 17 15:35:13 We are also working on jards. Sep 17 15:35:35 http://www.openembedded.org/wiki/StyleGuide Sep 17 15:35:39 Sorry, wrong linke Sep 17 15:35:41 http://www.anderssonj.com/jards Sep 17 15:35:56 T0mW: like I said, dbus_0.23.4.bb says "PROVIDES = 'dbus-glib'" Sep 17 15:36:01 psokolovsky_: What machine do you have? what image? Sep 17 15:36:23 I can build the stuff for you, if you want. Sep 17 15:36:28 Laibsch: I also trying to do some ports on leisure time, but I already got an idea that many people did something before, but now it just dust somewhere. ;-( Sep 17 15:36:35 koen: so.. Ok, doesn't dbus_0.71 say the same thing? Sep 17 15:36:40 Laibsch, so it's good to know it will be in mainline Sep 17 15:36:50 Laibsch, I'm with ipaq h4000 ;-) Sep 17 15:37:15 T0mW: dbus_0.92.bb doesn't say that, since it doesn't provide it anymore, but dbus-glib provides it Sep 17 15:37:24 (a package always provides itself) Sep 17 15:37:33 hey, folks. is there anybody familiar with liblbxutil package ? Sep 17 15:37:43 psokolovsky_: Are you using an X based image? Or opie? Sep 17 15:37:56 so if you want to use dbus 0.23.x on your system, dbus is all you need to build app Sep 17 15:38:07 if you want to use 0.9x, you need dbus and dbus-glib Sep 17 15:38:11 hey koen . do you know something about liblbxutil ? Sep 17 15:38:25 leoncamel: it's a recipe in OE Sep 17 15:38:36 Laibsch, I'm leaning towards GPE. I'm dual boot so far ;-) Sep 17 15:38:41 what's the difference between gdb and gdb-cross Sep 17 15:38:43 koen so the statement 'PREFERRED_PROVIDER_dbus-glib = "dbus-glib"' is essentially saying 'ignore anyone that said HEY! I provide that functionality!' ?? Sep 17 15:38:58 i tried google and the .bb files but i didnt' find much Sep 17 15:39:04 koen, hmm, what do you mean "recipe" ? Sep 17 15:39:33 T0mW: "ignore all packages that aren't called 'dbus-glib' when you need dbus-glib" Sep 17 15:39:48 leoncamel: a recipe for bitbake Sep 17 15:40:00 koen: ok, makes little sense right now, I'll try what you suggested though. Sep 17 15:40:02 psokolovsky_: OK, I have to go now but I will see if I cannot build the stuff for you later. That way we can test if things work. It would also entice the "granule guy" to stop dismissing OE, I guess if we can tell him that two devices have been confirmed working shortly after compiling things. Sep 17 15:40:29 koen: Maybe you are also interested in testing this out? You already have your own package ;-) Sep 17 15:40:48 Laibsch, if you want me to test sth, let me know too, I'd be glad to help as time permits Sep 17 15:40:56 great Sep 17 15:41:29 koen: why not fix the broken bitbake file 'dbus_0.23.4.bb' ? Sep 17 15:41:44 T0mW: what's broken about it? Sep 17 15:42:08 koen: isn't it falsely stating the functionality? Sep 17 15:42:16 http://teleology.ru/files/joke01.png - first try Sep 17 15:42:17 T0mW: it isn't Sep 17 15:42:24 RP: hmmmmmm Sep 17 15:42:35 T0mW: it provides dbus-glib 0.23.4 Sep 17 15:42:39 koen, and do you mean the package liblbxutil is not part of xorg ? Sep 17 15:42:41 koen: I almost got it, stay with me, ok? Sep 17 15:42:45 T0mW: your configure script wants > 0.60 Sep 17 15:43:13 koen: so, before 0.71, both dbus + glib were in the same package, then at 0.71 they broke them apart? Sep 17 15:43:46 RP: ah well, the symlinks point to libc.so.2.3.90 Sep 17 15:43:50 koen: ok, I think what you mean, wierd though Sep 17 15:43:56 leoncamel: no, I meant OE has a recipe for it: http://www.openembedded.org/repo/org.openembedded.dev/packages/xorg-lib/liblbxutil_X11R7.1-1.0.1.bb Sep 17 15:44:02 koen: ok, I think what I see you mean, wierd though Sep 17 15:44:07 T0mW: yes, that's correct Sep 17 15:44:09 bah Sep 17 15:44:10 heh Sep 17 15:44:16 koen: thanks Sep 17 15:44:18 bah^2 Sep 17 15:44:28 ~lart chouimat Sep 17 15:44:28 * ibot takes out chouimat with the trash Sep 17 15:44:46 chouimat: heh Sep 17 15:45:03 chouimat: old dog vs. new tricks Sep 17 15:45:03 T0mW: see http://www.freedesktop.org/wiki/Software_2fDbusReleaseArchive Sep 17 15:45:15 * T0mW looks Sep 17 15:45:27 0.90, second point Sep 17 15:45:33 koen, here is my problem. it was block at http://rafb.net/paste/results/qnztuy45.html while building gpe-image. Sep 17 15:45:52 koen, and i found the mkg3states is a arm binary. Sep 17 15:46:05 we don't have any QT users in OE, otherwise 0.23.4 would've had PROVIDES += "dbus-qt" Sep 17 15:46:15 (proper QT, that is) Sep 17 15:46:19 T0mW: hehe Sep 17 15:46:35 koen, and I have already setup a environment of scratchbox. so it report the error above . Sep 17 15:46:44 koen, could you give me some clue ? Sep 17 15:47:07 no idea on that Sep 17 15:48:14 koen, so, the programe named mkg3states should compiled *native* or *cross compiling* ? Sep 17 15:50:00 i was thinking of changing the compiler choice fort he OE build process to 3.x.x to fix my iwmmxt problems with gdb-cross. how bad an idea is that? Sep 17 15:50:47 leoncamel: natively Sep 17 15:51:40 koen, hmm, so is there anybody else had already built this package ? Sep 17 15:53:08 koen, and here is another problem while compiling opie-image. Sep 17 15:53:11 koen, http://rafb.net/paste/results/a4LZo428.html Sep 17 15:53:20 I don't care about opie-image Sep 17 15:53:26 koen, could you mind have a look at it ? Sep 17 15:53:29 people should stop using qt/e crap Sep 17 15:53:49 koen, but why ? Sep 17 15:54:12 hrw, could you mind have a look at this ? http://rafb.net/paste/results/a4LZo428.html Sep 17 15:55:32 koen I use qt/e ... ok qtopia-core :) Sep 17 15:55:35 koen, so. let's talk about the liblbxutil package . Sep 17 15:56:00 koen, so. where is wrong with this package ? Sep 17 15:56:35 leoncame`: one reason to stop using qt/e, instead others, that is QtopiaCore is much upgraded. Sep 17 15:57:48 den-ros, but, the qt/e is stable and there are a lot of project using it, right ? Sep 17 15:57:49 koen, to compile stuff for psokolovsky's ipaq 4000 would it be sufficient to just change DISTRO=familiar and machine=h4000 or would I need to clean out tmp/? Sep 17 15:58:26 clean out /tmp Sep 17 15:58:31 Laibsch, familiar won't work likely ;-) Sep 17 15:58:36 leoncamel: you are Leon Zhang, right? Sep 17 15:58:50 koen, yes. Sep 17 15:58:53 koen: OK, then this will have to wait a little. Sep 17 15:59:00 psokolovsky_: What do you need? Sep 17 15:59:23 Thanks, koen. And later. Sep 17 15:59:31 leoncame`: when i was use Qt/E and after start compile QtopiaCore - i understand for me - that did not whant to return or still work with Qt/E. Sep 17 15:59:38 03koen 07org.oe.dev * rddfe4434... 10/ (1 packages/xorg-lib/liblbxutil_X11R7.1-1.0.1.bb): liblbxutilo: fix mkg3states generation, reported by Leon Zhang Sep 17 15:59:46 leoncamel: there you go :) Sep 17 16:00:00 koen, haha . great. :) Sep 17 16:00:06 Laibsch, what you build for normally? OZ? in this case, angstrom would be ok. Sep 17 16:00:40 psokolovsky_: http://www.angstrom-distribution.org/repo/?action=details&pnm=granule :) Sep 17 16:01:39 psokolovsky_: There you go. The master has already done it for you. Thanks, koen. Sep 17 16:01:48 cool ;-) Sep 17 16:01:48 psokolovsky_: Let me know if you can start it. Sep 17 16:02:03 Will test when I rebuild gpe-image for angstrom Sep 17 16:02:15 may take couple of days ;-) Sep 17 16:03:07 psokolovsky_: I am impatient, so maybe I will compile for your gpe-image. What DISTRO would that be, then? Sep 17 16:03:34 koen, Ok. I am move forword to my gpe-image .. :) Sep 17 16:03:53 hi all Sep 17 16:04:11 hey cyrilRomain Sep 17 16:04:16 hi koen Sep 17 16:04:27 Laibsch, lol. issue is to find time to boot it, mostly ;-). DISTRO is "angstrom" Sep 17 16:05:33 cyrilRomain: hey Sep 17 16:07:09 Laibsch, btw, did you try to run it under qemu? that's good test, too. Sep 17 16:07:24 I have never used qemu. Sep 17 16:07:33 I hardly know a thing about it. Sep 17 16:07:41 Can I run that on my Desktop? Sep 17 16:08:25 koen, who is maintaining the opie-image now ? Sep 17 16:08:35 no idea Sep 17 16:08:43 Laibsch, yes, pretty neat for testing, so can use some love to be fully nice to use. OE.dev and OZ have "qemuarm" MACHINE to build for it Sep 17 16:08:49 koen, nobody care about it ? Sep 17 16:09:18 Laibsch, but actually, you just need a kernel for it, and gpe-image for any other machine kinda works, too Sep 17 16:09:25 psokolovsky_: OK, I will take a look at it later. Maybe I will get back to you with request for support ;-) Sep 17 16:10:43 hrw|afk, around ? Sep 17 16:11:04 hrw|afk, could you mind have a look at this ? http://rafb.net/paste/results/a4LZo428.html Sep 17 16:39:22 re Sep 17 16:39:32 leoncamel: looks like fscked distro config? Sep 17 16:40:34 hrw, what do you mean fscked ? Sep 17 16:41:33 leoncamel: you lack many pref_prov entries Sep 17 16:42:44 hrw, so . how many pref_prov should I set ? Sep 17 16:42:57 hrw, is there any list for it ? Sep 17 16:43:13 leoncamel: look at openzaurus-unstable Sep 17 16:50:12 hrw, http://rafb.net/paste/results/OTFjZH24.html Sep 17 16:50:45 leoncamel: switch to binutils 2.16 Sep 17 16:51:24 hrw, why ? Sep 17 16:51:29 should work.. I did not build from .dev recently Sep 17 16:51:39 leoncamel: less problems with native development Sep 17 16:52:15 hrw, OK. but which build is *stable* ? Sep 17 16:52:32 hrw, I want checkout a *stable* version of OPIE . Sep 17 16:52:56 leoncamel: any version of opie would be stable probably - it is not developed too much Sep 17 16:53:05 leoncamel, look at which distros were recently released Sep 17 16:53:42 hrw, I mean , the version which I could build with OE. Sep 17 16:53:57 emte, hmm, you mean oz354x ? Sep 17 16:54:06 that would be a good place to look Sep 17 16:54:17 release usually means stable Sep 17 16:54:56 emte, so, what you mean "release" is something like tag or something else ? Sep 17 16:55:20 leoncamel: .oz354x branch is a bit old when compared to .dev but it is stable Sep 17 16:55:47 leoncamel: it support only openzaurus distros and only zaurus machines (but anyone can build oz for own platform too) Sep 17 16:55:57 distro images released for public/supported use Sep 17 16:56:05 hrw, OK. i got it. I did sucessfully build opie a few days ago. Sep 17 16:56:42 hrw, but it bind to openzaurus too much. and what do you think of it ? Sep 17 16:57:17 leoncamel: basically any glibc based distro compared to OZ can be build from it Sep 17 16:58:33 hrw, OK. I got a other question about oz354x. the packages between oz354x and .dev is all the same, right ? I mean the same version. Sep 17 16:58:54 leoncamel: that depends Sep 17 16:58:54 hrw, or very close to each other ? Sep 17 16:58:57 leoncamel: not quite Sep 17 16:59:12 hrw, so , what is the difference ? Sep 17 16:59:13 leoncamel: many apps are old (as this branch has ~6 months now) Sep 17 16:59:28 hrw, so, which is older ? Sep 17 16:59:39 few devs add new software into it - but such things need to be discussed with me before Sep 17 16:59:56 leoncamel: no x11r7 iirc Sep 17 16:59:59 opie 1.2.1 Sep 17 17:00:51 hrw, i am not clear. which version is newer, the .dev one or the oz354x one ? Sep 17 17:01:10 dev is newer and possibly unbuildable Sep 17 17:01:20 hence being in dev Sep 17 17:01:33 emte, OK. Sep 17 17:02:05 leoncamel: it is more like Debian/stable (.oz354x) and Debian/unstable (.dev) Sep 17 17:02:49 leoncamel: oz354x is old, lack many nice things from .dev (like RDEPENDS->DEPENDS generation) Sep 17 17:02:52 hrw, and the oz354x based on opie-1.2.1 ? Sep 17 17:02:56 yes Sep 17 17:03:13 hrw, OK. I got it. :) Sep 17 17:03:31 maybe one day I will update it to 1.2.2 Sep 17 17:03:56 planning on releasing oz354.5 ? Sep 17 17:04:30 hrw, OK, the something (RDEPENDS -> DEPENDS generation) is added recently ? so there are a lot of "dependency bugs" booming ? :) Sep 17 17:05:08 emte: now 3.5.4.2 is on table Sep 17 17:05:54 hrw, so, I have a SUGGESTION. maybe you need a daily build on OPIE. it would be very useful for you controlling the software quality . :) Sep 17 17:06:18 really? i would have thought it might have shifted to 3.6.x Sep 17 17:06:54 emte, maybe the "oz 2007" or "oz XP" ? ;) Sep 17 17:07:05 leoncamel: why would you want to rebuild something that does not change at all? Sep 17 17:07:18 * emte witholds comments on naming things by year Sep 17 17:07:34 leoncamel: and building is only one aspect of the quality of the product Sep 17 17:07:57 there are probably 50-ish devices that opie is built for Sep 17 17:08:16 zecke, hmm, it is something like regression test. to *force* or *monitor* your software quality. is that clear ? Sep 17 17:08:36 leoncamel: opie builds without problems - problems were with recent gpe changes rather Sep 17 17:09:02 leoncamel: sure. but if nothing changes, why would one need to rebuild? Sep 17 17:09:23 emte: I will keep 3.5.4.x line until Angstrom will get released. Then I will dump all 2.6-powered-angstrom-supported zaurus models from it Sep 17 17:09:27 zecke, yes, it is. but, it is a *basic* or *base*. if it built failed, anything would be *blocked* , right ? Sep 17 17:09:32 leoncamel: and besides that there are daily builds... but I wonder about the relation between buildable and working Sep 17 17:10:06 zecke, the *working* is BASED on *building* ? right ? Sep 17 17:10:06 leoncamel: you still have no clue about QA, right? Sep 17 17:10:52 zecke, I mean, if the building failed, anything would be block. right ? Sep 17 17:11:14 leoncamel: why would you want to check pulse of a dead horse? Sep 17 17:11:26 leoncamel: why would you want to check pulse after a week again? Sep 17 17:11:39 zecke, so, it would keep each commit *must* be buildable. right ? Sep 17 17:11:44 leoncamel: even if it build - what sense if it does not work? or if it is not correctly packaged? Sep 17 17:12:14 leoncamel: .oz354x is not .dev - commits are usually not-breaking ones Sep 17 17:12:19 the package could build but all the db are broken Sep 17 17:12:36 leoncamel: first you still have no clue what you are taling about, then we allow and have daily builds and then Opie does not change at all as it is dead Sep 17 17:14:23 leoncamel: in general I agree with you on regular builds Sep 17 17:14:43 OK. I still keep my option. I think the buildable on .dev branches is also very important. Sep 17 17:15:43 zecke, :). so , I am the "daily building" now, right ? :) Sep 17 17:16:03 leoncamel: have you looked at the tinderbox last time? Sep 17 17:16:12 leoncamel: will you look at it this time? Sep 17 17:17:06 zecke, :). NO. you are very kindly . :) Sep 17 17:18:00 leoncamel: I think we have talked about the same stuff already Sep 17 17:18:02 zecke, So. what do you want me to have a look at ? Sep 17 17:18:25 leoncamel: and as I hate to repeat myself, I might sound a bit harsh Sep 17 17:19:33 zecke, OK. so, you are working on opie, right ? or maintaining ? Sep 17 17:19:55 leoncamel: no, I'm not working on Opie anymore. I used to work on it Sep 17 17:20:04 leoncamel: Opie is dead, at least this is what I believe Sep 17 17:20:21 zecke, and which is still alive ? gpe ? Sep 17 17:20:51 leoncamel: maemo, GPE, GNOME in general keeps on rocking but that is a different topic Sep 17 17:21:21 leoncamel: I have the feeling, I could be wrong, you regulary come here say "QA... regular builds" and leave Sep 17 17:22:09 leoncamel: and I tend to repeat. We have many many different configurations. Setup autobuilding for the configuration you are interested in, report to our tinderbox... Sep 17 17:23:14 but autobuilding is only a small part, things have to be tested on real devices Sep 17 17:23:36 zecke, so , what do you mean "tinderbox" ? Sep 17 17:23:37 emte: yes :) Sep 17 17:23:44 leoncamel: the same I meant last time Sep 17 17:24:08 you can read all about tinderbox at mozilla's site Sep 17 17:24:08 http://ewi546.ewi.utwente.nl/tinderbox/showbuilds.pl?tree=OpenEmbeddedBuilds Sep 17 17:24:24 leoncamel: a crappy perl script from the Mozilla Foundation, you send your data to a tinderbox and it represents the data Sep 17 17:25:50 leoncamel: again don't get me wrong. OE is in need for better QA Sep 17 17:27:35 zecke, yes, right . daily testing. Sep 17 17:28:09 leoncamel: you are so funny :) Sep 17 17:28:34 zecke, :). what do you mean funny exactly . Sep 17 17:28:46 zecke, I said too much "daily" ? ;) Sep 17 17:29:39 leoncamel: you just don't have a clue. And you keep coming here repeating the word daily. It is just plain funny Sep 17 17:29:42 we could buy him a book one QA definitions Sep 17 17:29:45 on* Sep 17 17:29:56 leoncamel: What is quality, what is the quality OE wants Sep 17 17:30:27 leoncamel: then we have all different aspects. As we bundle software we depend on the quality of all the projects we build Sep 17 17:30:54 leoncamel: we have native and non-native packages requiring different kinds of tests Sep 17 17:31:09 leoncamel: we have things one can test statically, only when running the application remotely Sep 17 17:31:14 zecke, who test it ? Sep 17 17:31:32 leoncamel: a lot of stuff is emergent as it requires that different projects work well together Sep 17 17:31:49 leoncamel: and I do not worry about the daily compiling as we can do that for a long time already Sep 17 17:33:06 zecke, come on. but it can't be built recently. it is time to fix it, right ? Sep 17 17:33:51 leoncamel, how many developers and users do you think build something everyday? Sep 17 17:34:16 leoncamel: you are so funny. I think OE has a infinite number of possible configurations Sep 17 17:34:23 pretty much everything gets built everyday with exception to rare devices Sep 17 17:34:28 leoncamel: and I only have finite amount of computing on my build farm Sep 17 17:34:45 leoncamel: so how likely is it, I'm building your configuration? Sep 17 17:36:16 zecke, but, do you thinks the building failed here is *normal* ? or *right* ? Sep 17 17:36:33 leoncamel: I have not looked at your failure at all Sep 17 17:37:45 zecke, now, could you mind have a look at it ? http://rafb.net/paste/results/a4LZo428.html Sep 17 17:38:18 leoncamel: why should I? is paraphone-gen a supported distribution by OpenEmbedded? Sep 17 17:39:13 leoncamel: there is not even a praphone-gen in OE, how should it get build daily Sep 17 17:39:20 leoncamel: it is even getting more funny :) Sep 17 17:39:44 It sure would be nice if people actually learned something about OE before complaining about it. Sep 17 17:40:19 leoncamel: I agree that things should always be buildable, no doubt about that Sep 17 17:40:51 leoncamel: but if you believe you can always use the latest version of OpenEmbedded, benefit from improvements without doing adjustments I'm sorry this is nothing we can promise Sep 17 17:41:01 OK . I will test the openzaurus-unstable.conf tonight. Sep 17 17:41:21 leoncamel: it would be nice to always be able to use latest, greatest but this is unrealistic Sep 17 17:43:41 zecke, Ok. I agree you. it is unrealistic. I come here just for help. you know, I am newbie on opie. so I need your help about this simple problem. Sep 17 17:43:55 leoncamel: There is another issue. paraphone is not part of OE, so it does not receive the automatic updates we are normally doing when we upgrade stuff Sep 17 17:44:02 speaking about issues - does anybody else has the problem that while glibc staging, the process loops and never ends? We have this on *some* machines, but not all. At least one of the working and one of the non-working are both running an up-to-date debian unstable, and so far nobody was able to find a difference. Sep 17 17:44:24 tmbinc: GNU make, koen had it as well and I might be responsible Sep 17 17:44:37 tmbinc: try changing bitbake.conf and replace -isystem with the old and plain -I Sep 17 17:46:04 oh Sep 17 17:46:08 tmbinc: I remember koen having had this issue after I replace -I with -isystem in bitbake.conf Sep 17 17:46:15 ok Sep 17 17:46:25 tmbinc: I do not remember the resolution and his conclusion Sep 17 17:46:30 do you know if the problem is a too new or too old 'make'? Sep 17 17:46:35 zecke, the disto here is modified from generic-unstable.conf. Sep 17 17:46:48 tmbinc: I think it could have been a GNU make (newer GNU make) issue as well, but I sadly can't remember Sep 17 17:46:50 s/disto/distro. Sep 17 17:47:09 zecke: i let one of the people with the problem test that :) Sep 17 17:47:30 "unfortunately" on all my machines it's working Sep 17 17:48:22 tmbinc: I would compare the versions of GNU make, and then try replacing -isystem with -I Sep 17 17:48:44 leoncamel: okay. Free training, you will not be charged :) Sep 17 17:49:03 leoncamel: It fails on installing itsy packages to the rootfs. So opie itself was built Sep 17 17:49:47 leoncamel: and lines 382-400 you see "ERROR: ..." Sep 17 17:50:35 zecke, yes. I got it. but I do "bitbake meta-opie". Sep 17 17:50:41 leoncamel: it fails installing a lot of opie packages (and now I do not know the versions of the stuff you are using, the date of your last upgrade) Sep 17 17:50:58 leoncamel: so what you could do is. Pick any of the missing packages Sep 17 17:51:23 leoncamel: e.g. opie-language, go to your TMPDIR and check deploy/ipk if you have a opie-language*.ipk package Sep 17 17:52:25 zecke, OK . here is my distro : http://rafb.net/paste/results/OTFjZH24.html Sep 17 17:52:49 leoncamel: no, listen. I ask you to check something and you have two options Sep 17 17:52:54 leoncamel: do it, or don't do it Sep 17 17:54:40 zecke, I can't find tmp/deploy/opie-freetype*.ipk. Sep 17 17:54:51 tmbinc: stupid question. Do you know if the dvbsnoop author is in irc? Sep 17 17:55:10 leoncamel: now try to find opie-freeytep in tmp/work/ Sep 17 17:55:38 leoncamel: when was the last time you have upgraded? Sep 17 17:56:09 leoncamel: have you seen http://www.openembedded.org/first-construction-phase-over-2nd-in-sight this? Sep 17 17:56:35 zecke, yes. I got tmp/work/armv5te-linux/opie-freetype-1.2.2-r0/ Sep 17 17:57:50 zecke, bitbake 1.6, and i do pulling and updating just now. Sep 17 17:58:19 leoncamel: no, listen again. Randomly sticking your finger somewhere will only make it more dirty... Sep 17 17:59:01 leoncamel: so you have opie-freetype-1.2.2-r0 but no package you could have done: Sep 17 17:59:09 leoncamel: randomly deleted ipackages (likely) Sep 17 17:59:22 leoncamel: do not inherit package_ipk (unlikely as the tasks are proper packages) Sep 17 18:00:13 leoncamel: even more unlikely. YOu lied have a opie-freetype but it has an architecture set that is not in your IPKG_ARCH list Sep 17 18:00:54 zecke: rasc? not at the moment, but he is usually in #dbox2dev in ircnet Sep 17 18:01:15 do we allow '+' in PN? Sep 17 18:02:14 zecke, I found "inherit opie" from opie-freetype.inc. Sep 17 18:02:29 zecke, I am not clear. what do you mean exactly ? Sep 17 18:02:32 tmbinc: I woke up and dvbsnoop tried to analyse itself and crashed when parsing its environment Sep 17 18:02:50 leoncamel: lol, why do you think randomly tossing names around helps? Sep 17 18:03:02 leoncamel: do you know what itsy packages are? Sep 17 18:03:45 zecke, what do you mean "itsy" ? Sep 17 18:03:57 leoncamel: I'm sorry for you Sep 17 18:04:10 leoncamel: ipk == Itsy Packages, Itsy == CRL research device Sep 17 18:04:22 leoncamel: CRL == Cambridge Research Lab, an OLD DEC lab Sep 17 18:04:34 leoncamel: DEC == Digital Equipment which was bought by Compaq Sep 17 18:04:42 zecke, OK. I got it. Sep 17 18:04:47 leoncamel: compaq was bought by HP, HP closed CRL last year Sep 17 18:05:30 zecke, so , let's move forward about the problem. Sep 17 18:06:39 leoncamel: now use your mind. Have you randomly deletd files from tmp? e.g. to reclaim disk space? Is this a clean rebuild? Sep 17 18:07:16 zecke, so, the package is build successfully, and it wasn't packaged sucessfully ? Sep 17 18:07:53 zecke, I do rebuild cleanly. I do "rm tmp/ -rf && bitbake opie-image" last time. Sep 17 18:08:18 leoncamel: okay one step at a time Sep 17 18:08:42 leoncamel: then what do we have now: We know opie-freetype exists in tmp/work Sep 17 18:08:57 leoncamel: could you check if the task-*.ipk exists in tmp/deploy/ipk? Sep 17 18:10:30 zecke, yes. there are a lot of task-opie*.ipk. Sep 17 18:10:58 leoncamel: okay then you don't tell the truth ;) Sep 17 18:11:21 leoncamel: It would mean it created packages for task-opie*.bb but not for the other bb files Sep 17 18:11:24 leoncamel: which can't be Sep 17 18:12:04 leoncamel: now it is up to you to find out why no package for opie-freetype is created Sep 17 18:14:58 zecke, http://rafb.net/paste/results/k0wggE72.html Sep 17 18:15:50 zecke, the file size of task-*.ipk is too small ? is that normal ? Sep 17 18:16:05 leoncamel: well, do ls -l deploy/ipk/* and paste that Sep 17 18:16:17 leoncamel: but what about using your mind? Sep 17 18:17:52 zecke, It is a long list. Sep 17 18:18:18 leoncamel: yes, paste it Sep 17 18:18:41 hmm.. something is fscked on my system: Sep 17 18:18:47 -rw-r--r-- root/root 0 2006-09-17 20:34 ./usr/bin/struOE2M Sep 17 18:18:47 -rwxr-xr-x root/root 85697 2006-09-17 20:34 ./usr/bin/interbench Sep 17 18:19:05 in many packages I got those usr/bin/st* tempfiles Sep 17 18:20:31 ideas? Sep 17 18:21:42 no Sep 17 18:24:47 leoncamel: around? Sep 17 18:25:02 zecke, yes. Sep 17 18:25:25 the list is too long. I can't paste it into rafb.net/paste Sep 17 18:25:40 leoncamel: split it in two? Sep 17 18:25:45 leoncamel: paste it Sep 17 18:27:32 zecke, http://pastebin.ca/174201 Sep 17 18:29:32 leoncamel: well okay Sep 17 18:29:45 zecke, so what ? Sep 17 18:29:57 leoncamel: then use grep to see who would DEPEND/RDEPEND on opie-freetype Sep 17 18:31:50 zecke, tasks/task-opie.bb . only this. Sep 17 18:32:09 zecke, tasks/task-opie.bb:32: opie-pics opie-sounds opie-freetype ttf-dejavu-sans \ Sep 17 18:33:42 leoncamel: now do you have packages like opie-alarm, opie-pics, opie-sounds? Sep 17 18:34:46 leoncamel: okay, opie-freetype seems missing Sep 17 18:34:46 zecke, yes . it got it : tmp/deploy/ipk/opie-alarm_1.2.2-r0_armv5te.ipk Sep 17 18:35:02 leoncamel: bitbake opie-freetype -cclean ; bitbake opie-freetype and watch the output Sep 17 18:35:16 and there is no file under the directory tmp/work/armv5te-linux/opie-freetype-1.2.2-r0/install/opie-freetype Sep 17 18:35:35 leoncamel: the difficulty is. You use some version of OpenEmbedded with configurations and modifications nobody besides you knows Sep 17 18:35:42 zecke, and tmp/work/armv5te-linux/opie-freetype-1.2.2-r0/install/opie-freetype-{dbg|dev|doc}. Sep 17 18:35:54 leoncamel: again, do what you are told or leave it Sep 17 18:36:25 leoncamel: I'm happy if you would take a look yourself, as it is something specific to your configuration/system Sep 17 18:37:19 zecke, hmm, it is almost the same to openzaurus-unstable.conf. Sep 17 18:37:50 leoncamel: so after you have rebuild opie-freetype Sep 17 18:38:01 leoncamel: did it say something like not creating empty package for...? Sep 17 18:38:31 03mickeyl 07org.oe.dev * r3b6b2904... 10/ (1 conf/distro/include/preferred-x-versions-pre-X11R7.inc): preferred-x-versions-pre-X11R7.inc: fix wrong xau entry, it's 0.1.1, not 1.1.0 Sep 17 18:38:39 03mickeyl 07org.oe.dev * r1904423d... 10/ (1 packages/git/git-native.bb): git-native: fix DEPENDS Sep 17 18:38:47 03mickeyl 07org.oe.dev * r315878d6... 10/ (3 files in 2 dirs): openssl: add openssl-native (needed for git-native) Sep 17 18:38:59 03mickeyl 07org.oe.dev * rdaa7392e... 10/ (1 packages/ltrace/ltrace_0.3.36.bb): ltrace: fix building for ARCH=i686 and ARCH=i586 Sep 17 18:39:02 03mickeyl 07org.oe.dev * r04cd7b14... 10/ (1 packages/xorg-lib/libxcalibrate_git.bb): libxcalibrate_git.bb: fix typo in PV Sep 17 18:39:11 evening Sep 17 18:39:18 mickeyl: ouch, git needs OpenSSL? Sep 17 18:39:27 mickeyl: did you try to build it with GNU TLS? Sep 17 18:39:34 no, i didn't Sep 17 18:39:37 btw. Sep 17 18:39:41 zecke, maybe there are something wrong while "do_install" ? Sep 17 18:39:43 we have to do something about the fetcher check Sep 17 18:39:52 there is a 'git' in debian which has nothing to do with 'linus-git' Sep 17 18:40:20 leoncamel: maybe, analyse it :) Sep 17 18:40:41 mickeyl: update-alternatives is your friend Sep 17 18:40:59 zecke, Ok. I have to fall sleep now. it is too late . :). and I will check it tomorrow. Sep 17 18:41:14 leoncamel: take care Sep 17 18:41:18 * chouimat is bored ... damned sunday afternoon ... Sep 17 18:41:48 hi mickeyl Sep 17 18:42:08 hi hrw, how are things? Sep 17 18:42:19 mickeyl: good Sep 17 18:42:43 mickeyl: blame Linus for choosing already used name for git Sep 17 18:42:53 ~hail Debian Sep 17 18:42:59 * ibot bows down to Debian and chants, "I'M NOT WORTHY!!" Sep 17 18:43:07 ~hail gentoo Sep 17 18:43:09 * ibot bows down to gentoo and chants, "I'M NOT WORTHY!!" Sep 17 18:43:14 hrw: :) Sep 17 18:43:20 mickeyl: Hi :) Does You drow Angstrom logo? Sep 17 18:43:35 I just took 4 Debian patches to get one app built Sep 17 18:43:51 den-ros: hey. yes, the ones you see in the installer / bootmanager screenshots Sep 17 18:43:59 hrw: nice :) Sep 17 18:44:49 * chouimat goes back to convince OE to build a wrtsl54gs image for him .... Sep 17 18:45:30 hey, when are you OEDEM guys scheduling to leave @ sunday? Sep 17 18:46:28 mickeyl: hmm, good question :) Sep 17 18:46:36 mickeyl: 12:52 or 16:52 from Ost station Sep 17 18:46:38 chouimat: yeah! i have a WRT54GS that's sitting unused... i want OE on it. Sep 17 18:46:52 mickeyl: Today we talk about Angstrom logo... sorry, but i draw joke picture :) ... i plan show it to You, but seems to delete it already. Sep 17 18:47:01 hrw: righto, that sounds good. i wanted to take a train ~ 1200 Sep 17 18:47:03 I have wrt54gs too - but used Sep 17 18:47:11 mickeyl: will take time ... I'm kind of rusted using OE ... Sep 17 18:47:15 hrw: that means we could have a nice breakfast together and then leave Sep 17 18:47:55 OEDEM? Sep 17 18:48:05 openembedded.org/oedem Sep 17 18:48:23 sure Sep 17 18:50:17 why everything cool and fun is in europe ... here we only have DDC, OLS and GNOME :( Sep 17 18:52:32 zecke: which hour are you arriving in Munich for DevDays? Sep 17 18:52:51 zecke: i could be at central station @ 13:33. would that work? Sep 17 18:53:26 guys: how do we mark recipes which can generate GPL and non-free subpackages? Sep 17 18:53:53 hmm Sep 17 18:53:58 hi [cc]smart Sep 17 18:54:13 LICENSE_subpackage ? Sep 17 18:54:39 LICENSE is more for source distribution than anything else... Sep 17 18:54:51 true, but it gets added into the ipk Sep 17 18:54:53 if the contents of SRC_URI contains that which we cannot distribute, LICENSE shouldnt be open Sep 17 18:55:01 When this celebrate DevDays start? Sep 17 18:55:05 then we need to split it into two license fields Sep 17 18:55:10 source and binary Sep 17 18:55:22 den-ros: TrollTech DevDays is 11.10/12.10 in Munich Sep 17 18:55:30 which would make sense, given either or both can have their distribution restricted Sep 17 18:55:42 mickeyl: Thank You. Sep 17 18:56:00 kergoth: currently I'm talking about lmbench which has gpl source but its results are non-free Sep 17 18:56:08 results == lmbench results data Sep 17 18:59:38 bbl Sep 17 18:59:44 have to go Sep 17 18:59:52 cu hrw Sep 17 18:59:53 hmmmm 1 hours 45 minutes before the beer store close ... Sep 17 18:59:56 cu hrw|gone Sep 17 19:00:59 hrw|gone: bye Sep 17 19:01:41 hrw|gone: ah, which is the reverse of what we usually have to deal with Sep 17 19:01:44 heh Sep 17 19:01:45 hm Sep 17 19:21:04 03freyther 07org.oe.dev * r6388d23a... 10/ (1 classes/test.bbclass packages/quilt/quilt-native.inc): Sep 17 19:21:04 classes/test.bbclass: Add a new task called test which we can run after compile Sep 17 19:21:04 packages/quilt/quilt-native.inc: Run make check for quilt native Sep 17 19:21:04 Run make check for quilt-native from the do_test task. This Sep 17 19:21:04 will run the first native test. If the test fails quilt will Sep 17 19:21:06 fail. Sep 17 19:21:08 The tests will only be run if you do INHERIT+=test Sep 17 19:22:58 ~praise zecke Sep 17 19:23:00 All hail zecke! Sep 17 19:23:59 kergoth: i heared you solved some glibc 2.4 for i386 problems locally? could you gimme a hint? Sep 17 19:24:46 mickeyl: next will be installing tests into a -test package Sep 17 19:24:55 mickeyl: and a start script to launch and run the test suite :} Sep 17 19:25:10 mickeyl: but doing this sounds like a diploma thesis to me :( Sep 17 19:25:16 *nod* Sep 17 19:25:31 at my current pace this will be next century Sep 17 19:25:52 crap, I should schedule some time to merge packaged-stagin Sep 17 19:25:53 g Sep 17 19:26:35 luckily my term starts soon Sep 17 19:27:07 hi kergoth Sep 17 19:27:16 mickeyl: I think I will take a Urlaubsemester in the summer to get rid of the mountain of to be taken tests :( Sep 17 19:27:34 zecke: cool stuff with tests. and yes, it sounds like thesis or at least term thing ;-) Sep 17 19:28:37 psokolovsky_: it will take ages until we are through with it :( Sep 17 19:28:40 zecke: sounds good. Sep 17 19:29:18 bbiab Sep 17 19:31:48 03koen 07org.oe.dev * r6639a14a... 10/ (1 packages/lcms/lcms_1.15.bb): lcms: update to 1.15 Sep 17 19:32:02 NOTE: package inkscape-0.44-r0: task do_configure: started Sep 17 19:32:10 * koen|away closes laptop and turns on tv Sep 17 19:39:18 hmm wird, it stopped working Sep 17 20:02:53 koen|away: thx for the latest do_install regression fix, ipkg now builds fine. Sep 17 20:07:25 good nite guys Sep 17 20:12:20 psokolovsky_: Are you still there? How do you actually use qemu? Sep 17 20:12:57 I called qemu-launcher and am presented with a bunch of choices. Maybe we can discuss in a private chat if you have time? Sep 17 20:13:44 Laibsch, http://handhelds.org/moin/moin.cgi/FamiliarUnderQemu is my collection of links, RP's page most important one. Sep 17 20:13:56 Laibsch, sorry, too late here ;-) Sep 17 20:14:37 I will study the page and see where this gets me. Sep 17 20:16:44 Too complicated for now. Sep 17 20:16:48 Maybe later. Sep 17 20:24:25 * T0mW watches bitbake build dbus-0.23.4 again despite PREFERRED_PROVIDER Sep 17 20:24:51 ~nuke bitbake for being so stupid Sep 17 20:24:56 * ibot prepares 100 missle silos, and targets them at bitbake for being so stupid ... B☢☢M! Sep 17 20:43:06 * RP returns Sep 17 21:06:04 is monotone.openembedded.org down or just sluggish? Sep 17 21:14:36 03kristoffer 07org.oe.dev * red935e15... 10/ (2 files in 2 dirs): Sep 17 21:14:36 packages/linux/linux-jlime-jornada6xx-2.6.17/io.h-fix.patch: Fix io bug Sep 17 21:14:36 * Fix IO code Sep 17 21:14:36 * Fixes strange pcmcia wifi behavior (= makes wifi work!) Sep 17 21:17:43 03tmbinc 07org.oe.dreambox * r662c2349... 10/ (1 conf/bitbake.conf): conf/bitbake.conf: replace -isystem with -I to work around unknown host toolchain problem Sep 17 21:39:38 03kristoffer 07org.oe.dev * rebc85345... 10/ (1 packages/linux/linux-jlime-jornada6xx_2.6.17.bb): Sep 17 21:39:38 packages/linux/linux-jlime-jornada6xx_2.6.17.bb: Add io.h patch Sep 17 21:39:38 * Addition of io.h-fix.patch to kernel .bb file Sep 17 21:51:18 koen|away, there? Sep 17 21:51:20 mickey|tv, ? Sep 17 21:52:06 Kristoffer, i am here though ;) Sep 17 21:53:01 TheCan, :D can you answer some OE questions then Sep 17 21:53:53 nevermind just found it :) Sep 17 21:56:09 heh Sep 17 22:38:48 03kristoffer 07org.oe.dev * r8f533ba4... 10/ (4 files in 3 dirs): Sep 17 22:38:48 gtk-webcore/osb-jscore-0.5.0: Addition of patches for gcc4 & superh Sep 17 22:38:48 * Addition of superh-aclocal.patch that adds sh* to host_cpu Sep 17 22:38:48 * Addition of gcc4-fno-threads-statics patch for sh3 Sep 17 22:38:48 * Change .bb file to reflect patches (used append_sh3) Sep 17 22:53:49 hi Sep 17 23:27:39 koen|away psokolovsky_: It runs. Check http://oz.leggewie.org/granule/granule-rotated.png and http://oz.leggewie.org/granule/granule.png for proof. Sep 17 23:30:29 03kristoffer 07org.oe.dev * r6f824aaf... 10/ (3 files in 3 dirs): Sep 17 23:30:29 gtk-webcore/osb-jscore-0.5.0: Addition of Patch Sep 17 23:30:29 * Addition of patch to modify const declarations (style KJS::) Sep 17 23:30:29 So that gcc likes them abit better (only for superh) Sep 17 23:30:29 * Change to osb-jscore_0.5.0.bb to reflect patch addition. Sep 17 23:30:31 03kristoffer 07org.oe.dev * r1e26a244... 10/ (1 conf/distro/jlime-donkey.conf): Sep 17 23:30:34 conf/distro/jlime-donkey.conf: Add preferred_versions of osb-jscore Sep 17 23:30:36 * We have a good osb-jscore version now (that finally builds) Sep 17 23:30:38 lets use it. Sep 17 23:39:47 anyone else using python with oe and got "messed up paths" in the compiled py files? the host-build-path (where the package was built) always ends up in error messages. everything else is fine, though. Sep 17 23:54:28 This may sound like a really stupid question, but would it be possible to compile windows binaries with OE? Sep 18 00:02:00 wasn't there somebody working on mingw support? Sep 18 01:54:23 morning ~! **** BEGIN LOGGING AT Mon Sep 18 02:35:45 2006 **** ENDING LOGGING AT Mon Sep 18 02:59:57 2006