**** BEGIN LOGGING AT Tue Feb 02 02:59:58 2010 Feb 02 05:40:14 こんにちは、iwamatsuさん Feb 02 06:25:21 Hi Laibsch Feb 02 06:27:32 That patch was sent mostly for mickey to review/ack, we talked about it before sleep().. You're right, I should say this in e-mail with patch.. Feb 02 06:28:39 Laibsch: and mickey has to fix something and then it can be aplied with updated srcrev with fix.. that's why I cannot push it directly Feb 02 06:29:08 I see Feb 02 06:29:16 that makes perfect sense, then Feb 02 06:31:34 * JaMa Off to work.. Feb 02 06:32:25 * grg home from work Feb 02 06:32:30 bye Feb 02 07:09:58 morning Feb 02 07:21:31 morning everyone Feb 02 08:00:15 RP, thanks for your reply, posted my reply with some more data, but guessing now the issue is that the page from hrw on which I started this does not mention moving DEPLOY_DIR out of tmp Feb 02 08:59:44 morning all Feb 02 09:00:16 hi gremlin Feb 02 09:00:33 hi ! Feb 02 09:26:38 jo rp Feb 02 09:27:30 eFfeM: Moving DEPLOY_DIR could well break it :/ Feb 02 09:28:10 hi woglinde Feb 02 09:28:46 args Feb 02 09:29:11 RP, well the problem I had was with DEPLOY_DIR inside tmp; btw i might have misphrased things, I meant to say that I put the DEPLOY_DIR outside tmp, and rebuilding, not issueing a mv command Feb 02 09:29:25 or should DEPLOY_DIR remain in tmp ? Feb 02 09:29:34 hi florian Feb 02 09:30:04 eFfeM: You meant DEPLOY_DIR or DEPLOY_DIR_PSTAGE ? Feb 02 09:30:10 florin can you send a mail to oe-ml which devices you will bring to fosdem? Feb 02 09:30:20 eFfeM: There is a big difference Feb 02 09:30:29 seems we are a bit unorganized Feb 02 09:30:59 DEPLOY_DIR_PSTAGE is already outside tmp, but that one does not have sysvinit-inittab (see my email) Feb 02 09:31:49 eFfeM: I don't understand why you need to change DEPLOY_DIR Feb 02 09:32:05 eFfeM: Changing DEPLOY_DIR_PSTAGE is understandable Feb 02 09:33:22 RP, this blog from denix suggested changing DEPLOY_DIR and your email also seemed to indicate that (although it was between the lines) Feb 02 09:33:49 eFfeM: Nothing of mine would mean that, only DEPLOY_DIR_PSTAGE Feb 02 09:34:13 RP what happens is that if I build with only DEPLOY_DIR_PSTAGE outside TMP_DIR, I get the errors as posted on the ML Feb 02 09:34:45 eFfeM: and as I said, this seems to be that the ipk is missing Feb 02 09:34:53 RP, not saying you said so, it was just something I concluded from it (perhaps faulty based upon bad understanding or english not being my native language) Feb 02 09:34:55 what you don't say is whether its in the staging package or not Feb 02 09:35:07 If its not in the staging package, we need to work out why Feb 02 09:36:27 good morning Feb 02 09:36:41 hi florian Feb 02 09:36:44 hi hrw Feb 02 09:36:45 It is there angstrom-dev/deploy/glibc/ipk/beagleboard/sysvinit-iniitab* Feb 02 09:36:45 the staging package is there but only in the tmp dir, not in the dir given by DEPLOY_DIR_PSTAGE Feb 02 09:37:12 *sigh* inittab with sysv and busybox is a mess Feb 02 09:37:19 eFfeM: I'm not talking about the staging package. The problem is the missing sysvinit-inittab ipk package in DEPLOY_DIR_IPK Feb 02 09:38:07 eFfeM: if it was finding that you wouldn't see the rootfs error Feb 02 09:38:31 eFfeM: It looks like its there as I mention above so the question is then why rootfs generation isn't seeing in. Feb 02 09:38:35 * RP -> back later Feb 02 09:38:36 eFfeM which image are you building? Feb 02 09:38:47 RP, yes but angstrom-dev == TMPDIR. I use the default angstrom local.conf which says: TMPDIR = "${OETREE}/${DISTRO}-dev" Feb 02 09:39:03 woglinde: angstrom head, beagleboard, packaged staging Feb 02 09:39:15 image!!!!!!!!!!!!!!!! Feb 02 09:39:22 you mentioned Feb 02 09:39:25 distro Feb 02 09:39:26 woglinde: oops sorry, console-image Feb 02 09:39:27 machine Feb 02 09:41:52 RP, maybe I misunderstand things, but it was my impression that after building once with DEPLOY_DIR_PSTAGE outside TMPDIR, I could remove TMPDIR and it would retrieve all from the PSTAGE dir. The sysviniit-inittab file lives within TMPDIR (as does DEPLOY_DIR_IPK I assume) Feb 02 09:42:52 RP your staging-pkg dir is where DEPLOY_DIR_PSTAGE points to ? Feb 02 09:46:15 other Q: can I set a preference for a specific version of a lib inside another recipe. Feb 02 09:46:27 eg something like PREFERRED_VERSION_xyz_lib = "1.0+cvs20100102" Feb 02 09:57:29 eFfeM: The staging .ipk should contain that other sysvinit-inittab .ipk Feb 02 09:57:47 eFfeM: So it should be recreated in DEPLOY_DIR_IPK Feb 02 09:58:37 eFfeM seems we need the long outstanding rev dep in bitbake Feb 02 10:04:55 I'm trying to create a tcl8.5 package, my first major package Feb 02 10:05:24 tcl8.5, in a Makefile, is calling "./tclsh blahblah" Feb 02 10:05:36 i.e. its calling a binary built during its compilation Feb 02 10:05:52 I should say, in an installation hook in the makefile Feb 02 10:06:03 But we need to use tclsh provided by tcl-native Feb 02 10:06:12 Since tclsh we just compiled is ARM Feb 02 10:06:20 Is there a general approach to doing this? Feb 02 10:07:23 RP, nuked the staging ipk this morning when trying with DEPLOY_DIR outside the tree, will rebuild and check what is in it (actually I don't recall that that there was an .ipk in it, but will check and report back Feb 02 10:12:39 woglinde: rev dep would be nice; tagging and pinning in machine files can help too Feb 02 10:14:57 eFfeM sane toolchain and revs Feb 02 10:15:01 are onky vehicles Feb 02 10:15:16 * XorA wonders when \s* broke in grep Feb 02 10:16:49 woglinde: will dig into sane Feb 02 10:35:09 Laibsch: btw, is there any reason for holding 1515 on patchwork? I'm using it there and also had it with rebuild and no problems of course.. Feb 02 10:35:57 Laibsch: and with python upgrade it would be nice to have it included Feb 02 10:36:11 Laibsch: which I should ask before.. Feb 02 10:39:53 I'm all for getting that committed Feb 02 10:40:00 but that is core and needs review Feb 02 10:40:10 JaMa: reviewed it :-) Feb 02 10:40:11 Haven't had a single ACK for it Feb 02 10:40:19 still no ACK Feb 02 10:41:04 The patch itself is sorely needed, I think. Only problem is that to do this really properly, one would need to bump PR for all affected packages Feb 02 10:41:28 link? Feb 02 10:41:53 http://patchwork.openembedded.org/patch/1515/ Feb 02 10:42:00 cheers Feb 02 10:42:49 that one gets my ACK Feb 02 10:43:01 Acked-by: Graeme Gregory Feb 02 10:43:01 mine too.. Feb 02 10:43:19 Acked-by: Martin Jansa Feb 02 10:43:25 lol Feb 02 10:43:48 if I could get tin to set my correct email address Id ack original email Feb 02 10:44:16 Laibsch: agreed that PR should be bumped but as soon as it's included even without PR bump we will build better packages.. Feb 02 10:45:39 mickeyl: Hehe. I had asked you for an ACK earlier and you kind of declined, being unsure of what that affected. Feb 02 10:46:01 I'll push it in a minute Feb 02 10:46:10 must have imagined a different thing Feb 02 10:46:20 or maybe I gave you an incorrect URL Feb 02 10:46:22 i guess i didn't see that it only affects -dbg packages Feb 02 10:46:34 or I just plain hallucinate (that happens) Feb 02 10:46:36 ;-) Feb 02 10:46:38 FILES_${PN}-dbg is another of the variables that shouldnt be overridden unless really really sure its right Feb 02 10:46:39 nah ;) Feb 02 10:47:58 we need an rureallyreallysure.bbclass Feb 02 10:48:54 kergoth could add that to recipe_sanity Feb 02 10:49:23 well unlike PACKAGES I can see that in some cases a .bb file might be right Feb 02 10:49:31 yes Feb 02 10:49:46 But a warning a la lintian would be OK Feb 02 10:49:55 I guess Feb 02 10:50:19 lintian checks for errors and emits warning based on severity and certainty Feb 02 10:50:24 -doc -static in the same class Feb 02 10:50:30 this one would be certainty: possible Feb 02 10:51:23 hmm Feb 02 10:51:24 hi rkirti Feb 02 10:51:31 NOTE: Unpacking ../meta-bug/packages/bug-apps/com.buglabs.app.bugdash.bb to tmp/work/all-angstrom-linux-gnueabi/com.buglabs.app.bugdash-1_20-r15.1/ Feb 02 10:51:37 ERROR: Error, lockfile path does not exist!: /home/hrw/devel/buglabs/trunk/build/tmp/work/all-angstrom-linux-gnueabi/com.buglabs.app.bugdash-1_20-r0/install Feb 02 10:51:42 wtf? Feb 02 10:51:57 why r0 instead of r15.1? Feb 02 10:52:32 ideas? Feb 02 10:56:26 could be a build that was interrupted Feb 02 10:56:32 then something happened in the source tree Feb 02 10:56:37 and then the build was restartet Feb 02 10:56:41 and now the revisions don't match Feb 02 10:56:52 try cleaning all stamps for that and rebuild that package Feb 02 10:58:14 ok Feb 02 10:58:56 same thing Feb 02 10:59:03 removed all stamps Feb 02 10:59:13 DEBUG: Executing task do_package_write_ipk Feb 02 10:59:13 DEBUG: update_data() Feb 02 10:59:13 DEBUG: mkdirhier(/home/hrw/devel/buglabs/trunk/build/tmp/work/all-angstrom-linux-gnueabi/com.buglabs.app.bugdash-1_20-r15.1/image) Feb 02 10:59:16 DEBUG: mkdirhier(/home/hrw/devel/buglabs/trunk/build/tmp/work/all-angstrom-linux-gnueabi/com.buglabs.app.bugdash-1_20-r0/image) Feb 02 10:59:19 ERROR: Error, lockfile path does not exist!: /home/hrw/devel/buglabs/trunk/build/tmp/work/all-angstrom-linux-gnueabi/com.buglabs.app.bugdash-1_20-r0/install Feb 02 11:00:41 hmm, very strange Feb 02 11:01:14 XorA, mickeyl: http://bugs.openembedded.org/show_bug.cgi?id=5395 Feb 02 11:02:17 dvar = bb.data.getVar('D', d, 1) in package_ipk.bbclass is wrong Feb 02 11:02:43 bitbake -e gives proper value Feb 02 11:09:16 shit Feb 02 11:11:07 Laibsch: agree with that bug report Feb 02 11:21:39 Laibsch: ack from me Feb 02 11:29:23 hrw: I always got this when that recipe changes arch.. Feb 02 11:30:08 hrw: like something what was build as armv4t and now is all then it always use old armv4t path in IIRC do_stage Feb 02 11:30:27 hrw: and workaround is to remove all traces of armv4t version from tmp/pkgdata Feb 02 11:34:39 so 18k tasks to rebuild Feb 02 11:35:00 or maybe ~13k if all will go fine Feb 02 11:51:45 any hint about "Runtime target 'qt4-x11-free-gles' is unbuildable, removing..." when building normal qt4-x11-free? Details here: http://www.mail-archive.com/openembedded-devel@lists.openembedded.org/msg02703.html Feb 02 11:53:30 JaMa sure Feb 02 11:53:46 hm Feb 02 11:53:49 I think you can build only one Feb 02 11:53:56 or you dont have the gles libs Feb 02 11:54:31 woglinde: I want to build only qt4-x11-free.. Feb 02 11:54:50 there is only inherit for qt4x11 in recipe Feb 02 11:55:24 and that message is always shown even with provider set (in compaitibility-providers.conf) to non-gles version Feb 02 11:57:02 here is that recipe http://git.openembedded.org/cgit.cgi/openembedded/tree/recipes/wpa-supplicant/wpa-gui_0.6.9.bb Feb 02 11:57:30 but any recipe with "inherit qt4x11" shows that error too (ie stopwatch, qgpslog) Feb 02 11:58:55 that error message by now is harmless (and pretty much unavoidable AFAIK) Feb 02 11:59:31 unless you mask out the gles one Feb 02 12:00:35 jama ask koen and zecke about it Feb 02 12:00:43 they had a thread on the ml Feb 02 12:01:57 woglinde: that's reply to that thread.. :) Feb 02 12:02:57 Laibsch: I guessed that it's only annoying and already masked it locally, but I was hoping there should be better fix.. Feb 02 12:03:58 Laibsch: why is it "tested" by bitbake as qt4-x11-free provider when preferred_provider is set in environment to qt4-x11-free? Feb 02 12:06:13 PREFERRED_PROVIDER_qt4-x11-free ?= "qt4-x11-free" is in compaitibility-providers.conf Feb 02 12:13:25 JaMa: I would expect something similar to bug 4230 at work here Feb 02 12:13:30 IOW, blame bitbake Feb 02 12:22:22 Laibsch: ok, fair enough, thanks for confirmation Feb 02 12:24:15 btw: why it's not named virtual/qt4-x11-free iirc someone said that virtual/ providers are handled better as long as they are used only in DEPENDS not RDEPENDS Feb 02 13:28:48 ok, booked my train ticket Feb 02 13:28:59 can only stay one night this year Feb 02 13:29:05 (fosdem, that is) Feb 02 13:29:31 the ICE speed outside germany is incredible Feb 02 13:29:51 it's taking me the same amount of time to Munich as it's taking me to Brussels Feb 02 13:29:56 ~lart german rails Feb 02 13:29:57 * ibot slaps a compatible dib on german rails's head Feb 02 13:30:11 mickey|sofa: come to the UK, we have the worlds slowest trains :-) Feb 02 13:30:19 hehe Feb 02 13:38:41 re Feb 02 13:38:45 wb hrw Feb 02 13:39:29 XorA: at least in UK you can pay with credit card during buying ticket in train Feb 02 13:39:57 hi mickey|sofa Feb 02 13:39:59 * pb_ stabs freenode Feb 02 13:40:01 hey pb_ Feb 02 13:40:34 which night are you staying in brussels, is that saturday? Feb 02 13:41:25 hrw: thats because our stations dont have working ticket machines Feb 02 13:41:42 * XorA got bored of having to get the guys to open it up to reclaim his stolen card Feb 02 13:41:52 pb_: yes, saturday to sunday. I'll arrive @ 10:30 sat morning Feb 02 13:41:55 pb_: which hotel are you in this year? Feb 02 13:42:07 XorA: I was buying tickets only on airport stations Feb 02 13:42:08 as for the german rails, I think Brussels is actually about the same distance from Frankfurt as Munich is, so the belgian rail network is maybe not as impressive as you might have thought :-} Feb 02 13:42:16 pb_: oops, right .) Feb 02 13:42:36 XorA: and was surprised that one of languages available was Polish Feb 02 13:42:38 remember that germany is about ten times the size of brussels :-) Feb 02 13:42:44 er, of belgium Feb 02 13:42:45 heh Feb 02 13:43:03 mickey|sofa: um, it is a new (to me) hotel and I have forgotten the name. let me check. Feb 02 13:43:49 ah, here we go, Le Meridien Brussels. Feb 02 13:44:24 ah, relatively near to the grote markt Feb 02 13:44:33 yeah, it's near centraal station I think Feb 02 13:44:46 so, the other side of the grand place from the Astrid etc, but not too far Feb 02 13:44:49 where are you staying this year? Feb 02 13:44:54 * hrw started building nano for maemo5-compat Feb 02 13:45:09 * hrw Astrid as usual Feb 02 13:46:41 i got a good deal at the NH Atlanta Feb 02 13:47:12 pb_ i guess you are compiling gnu radio for the beagleboard, not on the beagleboard :-) Feb 02 13:47:20 eFfeM: neither Feb 02 13:47:30 I think you have mixed me up with crofton (again?) Feb 02 13:47:30 heh Feb 02 13:47:36 yes Feb 02 13:47:48 mickey|sofa: very good Feb 02 13:47:56 why do you share the same initials ;-) Feb 02 13:47:58 yes he is :) Feb 02 13:49:29 to or on ? Feb 02 13:50:21 hrw: my old local council has all forms available in polish due to that being the language of 10% of inhabitants Feb 02 13:50:45 XorA: does not surprise me Feb 02 13:51:12 hrw: means the food and alcohol has improved :-) Feb 02 13:52:17 ~summon Robert Schuster Feb 02 13:52:18 apt takes out 20 clean, identical-looking phones, some extra hands, and pretends to be a telemarketer for a large corporation, so he gets delivered a phonelist containing Robert Schuster's coordinates Feb 02 13:56:41 2.5.1-1eglibc20+0m5 Feb 02 13:57:02 ~curse nOkia for glibc 2.5.1-1eglibc20+0m5 with Debian maintainers inside Feb 02 13:57:03 May you be reincarnated as a Windows XP administrator, nOkia for glibc 2.5.1-1eglibc20+0m5 with Debian maintainers inside ! Feb 02 13:57:18 hehe Feb 02 13:57:25 pb_: such version was never present in Debian - right? Feb 02 13:57:28 gosh, they put the debian maintainers in there? that is a bit harsh Feb 02 13:57:53 hrw: don't think so. iirc, debian only started using eglibc recently, after about 2.9 Feb 02 13:57:57 pb_: Version: 2.5.1-1eglibc20+0m5 Maintainer: GNU Libc Maintainers Feb 02 13:58:32 nokia will never learn how to cooperate Feb 02 14:00:08 haha... nokia rocks Feb 02 14:02:32 sigh, rebuilding glibc, system is running 15 instances of make, 5 instances of python, some 10 sh's 3 cc1's and a few other tings Feb 02 14:02:46 no wonder why the system feels like crawling Feb 02 14:02:59 sounds like you have set PARALLEL_MAKE and/or BB_N_T too high Feb 02 14:03:12 the glibc build system itself is not inherently parallel like that Feb 02 14:03:39 * eFfeM renames his system "Turtle" Feb 02 14:03:52 pb_ I use the default settings Feb 02 14:04:13 and this is a quad core Xeon @ 2.5 Ghz Feb 02 14:05:52 eFfeM: by default P_M and BB_N_T are set so only 1 process runs Feb 02 14:07:13 the default for P_M for angstrom is -j4 Feb 02 14:07:20 no idea what BB_N_T stands for Feb 02 14:07:22 ah, crazy angstrom Feb 02 14:07:43 yeah, maybe i should move to minimal Feb 02 14:08:03 BB_N_T = BB_NUMBER_THREADS Feb 02 14:08:12 guess koen has all systems at his uni hooked up as compilation slaves Feb 02 14:08:14 * XorA doesnt think that P_M default sounds right Feb 02 14:08:15 might be worth checking that angstrom is not also setting that to some high value Feb 02 14:08:38 eFfeM: and half of TI machines Feb 02 14:09:54 eFfeM: I think you are talking bogusness about Angstrom Feb 02 14:11:37 pb_: its not crazy Angsyrom dudes, but crazy eFfeM dudes it seems Feb 02 14:11:46 ah Feb 02 14:12:01 at least my grep shows no distro changing those cars Feb 02 14:12:03 vars Feb 02 14:12:04 fair enough Feb 02 14:12:24 anyone know of a library that contains functions useful for init processes? such as mounting a filesystem, bringing a network device up, etc.? Feb 02 14:12:46 mickey|bbl: libbusybox :-) Feb 02 14:12:58 and mickey|sofa as well Feb 02 14:13:00 i was afraid of that answer ;) Feb 02 14:13:21 * mickey|sofa checks out the latest busybox Feb 02 14:13:49 mickey|sofa: you propsed program actually sounds like it would be cool inside busybox anyway Feb 02 14:14:05 mickey|sofa: busybox fancierinit :-) Feb 02 14:14:15 XorA: true, that might be suitable Feb 02 14:15:38 nokia toolchain for maemo5 is 616M shell script Feb 02 14:15:54 hrw: heh, insane Feb 02 14:16:18 XorA it wasn't me who said crazy Angstrom, Feb 02 14:16:29 mickey|sofa: I noticed your blog about the startup Feb 02 14:16:44 hi RP Feb 02 14:16:50 i had to build on a new machine so as test I decided to jjust follow the instructions Feb 02 14:16:50 mickey|sofa: Its certainly a good way to go, using dbus activation is how a lot of moblin works Feb 02 14:16:51 mickey|sofa: I expect random code drops inside as toolchain Feb 02 14:17:11 mickey|sofa: Poky's init scripts are still shell based but we have them down to about 5 forks so the overhead is minimal though Feb 02 14:17:29 RP__: yeah, i really want to try to see how fast i can get by just doing the minimum amount of things necessary to bring the system up. thanks to devtmpfs and dbus system activation i don't need to do a lot Feb 02 14:17:41 eFfeM: recomended config is different from default settings Feb 02 14:18:07 mickey|sofa: I understand, I just think in practise you won't end up much different from whats in poky :) Feb 02 14:18:32 mickey|sofa: We did teach that to defer a number of things until after the desktop starts too (ssh server and so on) Feb 02 14:18:39 btw BB_N_T is also 4 and these are the values that originate from contrib/angstrom/local.conf Feb 02 14:18:53 There is a "dbus-wait" util that waits for a signal from the desktop Feb 02 14:19:15 eFfeM: but they are not in conf/distro/angsrom* Feb 02 14:20:22 RP__: true, i will probably not get substantially faster than your solution, however i find the idea nice to be somewhat self-contained. I always wanted to write something that can act as init process :) Feb 02 14:20:45 yeah, I think the idea of a self-contained init is a good one. Feb 02 14:20:53 mickey|sofa: fair enough :) Feb 02 14:20:56 I would be interested to see the results of that too. Feb 02 14:21:08 cool, i will keep you posted Feb 02 14:21:49 mickey|sofa: its that task I always meant to do at OM, but spent my time fannying around with qtopia-phone instead Feb 02 14:22:15 XorA: true, but as I said I just faithfully followed the instructions on http://www.angstrom-distribution.org/building-angstrom Feb 02 14:22:19 which say "cp ${OETREE}/openembedded/contrib/angstrom/local.conf ${OETREE}/build/conf/" Feb 02 14:22:29 XorA: yeah, i would have loved to see you doing that instead :/ Feb 02 14:22:45 XorA: the same as raster is doing now the things i wanted him to do back then Feb 02 14:22:55 but alas, the powers overruled me Feb 02 14:22:55 *sigh* Feb 02 14:23:11 mickey|sofa: its what happens when people who hve artistic dreams get in charge Feb 02 14:23:48 XorA: hmm yeah. these days, without extreme polish and design you have a hard time to do products, but you should still not make the designer the CEO Feb 02 14:24:17 mickey|sofa: being able to make a phone call should come before the bubbles Feb 02 14:24:35 yep Feb 02 14:24:43 mickey|sofa: but I got blanked by certain people when I said going with FSO made more sense than qtopia Feb 02 14:24:48 fun Feb 02 14:25:09 I think certain people were also pissed at me when my estimate for August for qtopia-phone turned out accurate Feb 02 14:25:24 maemo5 uses two toolchains now: scratchbox and madde. first uses gcc4.2.1, binutils 2.18.50.something, second gcc4.3.3, binutils 2.19.51.something Feb 02 14:25:35 hrw: Whats madde? Feb 02 14:25:40 XorA: *nod* i don't want to estimate how much money we burned for that E/X11/Qtopia monster :/ Feb 02 14:25:58 mickey|sofa: too much it seems Feb 02 14:26:02 yep Feb 02 14:26:15 RP__: other nokia stuff Feb 02 14:26:49 "MADDE stands for Maemo Application Development and Debugging Environment " - http://wiki.maemo.org/MADDE Feb 02 14:27:22 Yet another build system :( Feb 02 14:27:30 Just what the world needs... Feb 02 14:27:55 looks sboxish for me anyway Feb 02 14:36:44 Making phone calls is so last century Feb 02 14:38:10 Crofton: well personally the first app I like to see is Jabber Feb 02 14:39:20 but I suppose that does depend on a working gsm modem :) Feb 02 14:40:04 if only I had an sms->jabber gateway Feb 02 14:43:52 * kgilmer wonders what it is about build systems that makes everyone want to roll their own... Feb 02 14:45:15 * mickey|sofa discovers mount(2) Feb 02 14:46:00 kgilmer: NIH Feb 02 14:46:10 kgilmer: It starts with "this one will be more simple" Feb 02 14:47:00 yeah, exactly. I don't think it's strictly NIH, it's just that you always end up with this feeling that the existing build systems are either too complex or don't do quite what you want, and it'd be much simpler if you just write a new one which... Feb 02 14:47:02 makes sense mickey|sofa, zecke. in fact a dev at bug started writing one in haskell before we decided on poky. Feb 02 14:47:26 *nod* Feb 02 14:47:58 pb_: ... ultimately turns out to require the same complexity once you understand enough of the domain Feb 02 14:48:09 mickey|sofa: right Feb 02 14:49:51 oe isn't supported on windows though... os x looks pretty raw based on the wiki as well. maybe that was it. Feb 02 14:50:11 "Unfortunately there are various issues building on OS X that will most likely prevent the toolchain from building. Feb 02 14:50:11 Unfinished - tbd " Feb 02 14:51:27 hms Feb 02 14:51:56 it would seem from reading the blog post that win/os x was a core requirement. Feb 02 14:52:58 yeah, lack of windows support is a bit of an issue. I'm not actually sure how hard it would be to get oe to run on a modern windows host. I suspect you would probably run into some of the longstanding suckages around the way that bits of oe handle filenames, which might be annoying. not sure what else would actually be a showstopper. Feb 02 14:53:05 not that I have any particular desire to find out, either :-} Feb 02 14:53:20 i hear that pb_ Feb 02 14:56:28 kgilmer I am right now try to compile java zero with thumbee support Feb 02 14:56:55 nice woglinde :) Feb 02 14:57:14 i am stuck now w/ openjdk build in stable. Feb 02 14:57:46 an upstream dep, i think jamvm-initial is failing on me. Feb 02 14:58:34 I pushed new icedtead6-native into jalimo Feb 02 14:59:51 nice, i'll have to take a look woglinde :) Feb 02 15:00:25 so now a working one Feb 02 15:00:29 is ine Feb 02 15:00:31 args in Feb 02 15:00:44 I might happen that the compile stuckz Feb 02 15:00:50 but I saw this only one time Feb 02 15:01:17 kglimer I pushed a sane-toolchain-jave.inc into distro/include Feb 02 15:01:26 try to include it and build from sratch Feb 02 15:01:27 i should try against dev or stable woglinde ? Feb 02 15:01:31 shoud work Feb 02 15:01:40 I mainly work against dev Feb 02 15:01:50 I used md5 stuff inside recipe Feb 02 15:02:20 include sane-toolchain-java.inc where woglinde ? Feb 02 15:03:08 local.conf Feb 02 15:03:47 include conf/distro/include/sane-toolchain-java.inc Feb 02 15:03:50 should work Feb 02 15:04:46 ah ok woglinde thanks, i will try that now! Feb 02 15:05:24 it worked here Feb 02 15:05:30 plain 686 build Feb 02 15:05:36 do you have amd64? Feb 02 15:05:44 my host is amd64, yes Feb 02 15:05:49 okay Feb 02 15:05:57 lets see if it works for you too Feb 02 15:06:09 and pull the latest dev Feb 02 15:06:13 yep Feb 02 15:06:20 I checked in some temporary fixes yesterday Feb 02 15:09:22 * kgilmer waits. bitbake is now handling some files. Feb 02 15:13:50 hi all, I'm trying to setup my local.conf for ngw100, I have MACHINE="atngw100" etc, but bitbake don't want to build anything. It's printing lots of lines like ERROR: Exception: Message:'NoneType' object has no attribute 'append' Feb 02 15:13:56 any idea why ? Feb 02 15:14:53 soltys: which DISTRO? Feb 02 15:15:31 angstrom-2008.1 Feb 02 15:15:40 ANGSTROMMODE="uclibc" Feb 02 15:15:55 bye! need to catch the flight Feb 02 15:16:12 ANGSTROMLIBC = "uclibc" I meant Feb 02 15:16:14 sorry Feb 02 15:16:38 zecke: have a safe trip Feb 02 15:16:45 zecke: later Feb 02 15:16:54 bye zecke Feb 02 15:16:56 snowing now Feb 02 15:17:06 zecke: later Feb 02 15:17:09 oops Feb 02 15:17:28 its snowing for several hours now... Feb 02 15:17:47 still don't have that "focus on the window I am looking at" window manager that I needed Feb 02 15:18:28 :-) Feb 02 15:18:40 XorA: thanks ;) Feb 02 15:18:52 hmm... eye tracker for window manager control Feb 02 15:19:06 one day debian will get all the avr32 patches into glibc and we can rock on :-) Feb 02 15:19:52 lol Feb 02 15:19:54 into eglibc please :) Feb 02 15:20:08 make ulf an debian maintainer Feb 02 15:20:37 hm now it's time for bitbaking image and going to buy SD card ;) Feb 02 15:28:06 soltys: hej here Feb 02 15:28:33 woglinde: no, please - no Ulf Feb 02 15:28:47 hrw *g* Feb 02 15:28:51 woglinde: Ulf is more sales man then software Feb 02 15:29:07 woglinde: at least according to atmel hw guys which I spoke with at ELC-E Feb 02 15:30:02 hrw: that was impression I got at FOSDEM Feb 02 15:30:13 hrw I know Feb 02 15:30:17 you already told me that Feb 02 15:31:08 ok Feb 02 15:31:20 RP, regenerated my system including sysvinit, but there is no sysvinit-inittab*.ipk in the ipk in the staging dir Feb 02 15:32:49 the toc of my package in staging (staging-sysvinit-beagleboard-angstrom-linux-gnueabi_2.86-r57_i686-linux.ipk ) is here: http://pastebin.ca/1775709 Feb 02 15:34:28 hrw: hello ;) Feb 02 15:36:58 eFfeM: Is there a sysvinit-inittab ipk file *within* taging-sysvinit-beagleboard-angstrom-linux-gnueabi_2.86-r57_i686-linux.ipk ? Feb 02 15:37:41 eFfeM: Or is that a listing of the contents of that .ipk? Feb 02 15:38:11 eFfeM: Assuming it is (it looks like it), then that is your problem, its missing Feb 02 15:38:24 that listing is from the data.tar file from the staging ipk Feb 02 15:38:36 so the answer is no that ipk is not there Feb 02 15:39:23 only ipk in the staging one is sysvinit-2.86-r57.do_package_write_ipk Feb 02 15:39:48 RP__: guess indeed that is my problem, but no idea how to fix it Feb 02 15:40:25 (but meanwhile eFfeM got a somewhat better understanding on how packaged staging works :-) ) Feb 02 15:40:54 eFfeM: Well, it gives you somewhere to start looking... Feb 02 15:41:12 https://bugs.maemo.org/show_bug.cgi?id=8788 - Maemo5 packages have Debian maintainers listed but are not from Debian Feb 02 15:41:13 RP__: any idea where this is added ? Feb 02 15:41:24 Are there any .ipk files in the sysvinit build directory in staging-pkg/* ? Feb 02 15:41:35 That staging .ipk is built from there Feb 02 15:42:14 checking Feb 02 15:46:23 RP__: hm, the sysvinit build dir only contains the tmp file. only ipk for sysvinit are in deploy Feb 02 15:46:39 ./deploy/glibc/ipk/beagleboard/sysvinit-inittab_2.86-r57.5_beagleboard.ipk Feb 02 15:47:30 need to leave now, will test some more tonight Feb 02 15:47:46 eFfeM: There should be a copy of that placed into staging-pkg Feb 02 15:48:08 hm Feb 02 15:48:38 eFfeM: The code for that is somewhere in packaged-staging.bbclass Feb 02 15:49:12 ok will have a look at it, guess the issue will be trivial, just did a find on my TMPDIR and there is no staging-pkg dir :-( Feb 02 15:49:43 btw pstage/pstaging_lists is also empty Feb 02 15:50:20 need to go now, but will get back to this, RP__ thanks for your help Feb 02 16:18:48 GNUtoo, ok pull Feb 02 16:19:53 GNUtoo, but I lied it's not totally working. it needs some headers at least include/msm_mdp.h and I think it's most likely to work well with the _latest_ codeaurora branch and not the .29 as I said earlier Feb 02 16:20:09 dcordes, ok Feb 02 16:20:14 need to get back to organic chemistry Feb 02 16:20:19 bbl Feb 02 16:20:20 dcordes, ok thanks a lot Feb 02 16:20:54 GNUtoo: you plan to take BUGs to fosdem? Feb 02 16:21:06 hrw, maybe Feb 02 16:21:39 what should I bring? Feb 02 16:21:54 I did not planned to take any devices this year to travel really light Feb 02 16:22:30 I travel by plane...my laptop is this eeepc 701 so it does not takes much place...but I fear that I will maybe buy some stuff at fosdem Feb 02 16:22:37 such as t-shirt or devices Feb 02 16:22:55 I take only phone (n900) + charger for it Feb 02 16:22:56 t-shirt is shure but device is unshure Feb 02 16:22:59 ok Feb 02 16:23:04 I saw you pushed a kenrel Feb 02 16:23:14 but not tested it yet on device Feb 02 16:23:34 no need to take charger as we do not have support for that in oe(just joking) Feb 02 16:23:56 because I bet no-one will reverse-engineer the battery charging... Feb 02 16:24:00 ;D Feb 02 16:24:47 hmm arm32 sux ;P i can't build toolchain, glib and gcc fail;/ Feb 02 16:25:31 soltys: o rly? I did avr32 builds and used on that ngw100 which you have on desk Feb 02 16:25:31 arm 32bit? or avr32? Feb 02 16:25:45 avr Feb 02 16:26:03 soltys: you need to have uclibc build and thats all Feb 02 16:26:03 ok Feb 02 16:26:27 how far is the hardware based java implementation on avr32? Feb 02 16:26:46 no idea Feb 02 16:26:49 ok Feb 02 16:26:51 hrw: I tryed bitbake vim but it fails on do_configure of glib and do_compile of gcc-cross-initial Feb 02 16:26:56 no idea why ;) Feb 02 16:28:03 * XorA kicks off avr32 build Feb 02 16:31:05 * woglinde started hopefully last thumbee compile Feb 02 16:34:15 hrw, I'll try to bring a bug device if it fits into my bag Feb 02 16:34:57 but I can't bring 2 bugs because I've only one transport camera bag Feb 02 16:36:41 * XorA is going to miss the rum this year :-( Feb 02 16:38:07 damn Feb 02 16:38:18 seems we dont get the toucbook Feb 02 16:39:16 touchbook? Feb 02 16:42:08 yes Feb 02 16:42:14 compay bought one Feb 02 16:49:30 * XorA burgles woglinde company Feb 02 16:49:55 someone could ask koen if he is taking one Feb 02 16:51:02 I'ld like to see one before buying it Feb 02 16:51:46 I want one to try and make it an EEE replacement Feb 02 16:51:56 unfornatly parcel get stucked somewhere at sending it to our department Feb 02 16:52:13 jo lergoth Feb 02 16:52:19 someone beat me to burglarising :-) Feb 02 16:52:19 ups kergoth Feb 02 16:52:29 hey Feb 02 16:53:44 XorA: 256MB ram will not hurt too much? Feb 02 16:54:19 hrw hm no 512 yet? Feb 02 16:54:47 I thought omap3550 256 mb are out of stock Feb 02 16:54:49 right, 512MB now Feb 02 16:54:56 :) Feb 02 16:56:00 are there some people comming with gumstix Feb 02 16:57:31 woglinde, build is still running! good sign Feb 02 16:59:10 kgilmer, hi I bet you don't come to fosdem this year....too bad Feb 02 16:59:24 heh... I thought that touchbook has wifi and bt onboard not in 2 usb slots Feb 02 17:01:52 yeah GNUtoo not this year :< thanks for bringing your BUGs though! Saw that on the ml. Feb 02 17:02:15 kgilmer, I'll try to bring one Feb 02 17:02:28 kgilmer, + some modules Feb 02 17:02:37 that would be great GNUtoo Feb 02 17:04:43 I can't bring the 2 because I've no case for the second one Feb 02 17:05:18 by case you mean bugbundle box? Feb 02 17:05:56 or you have device without plastics? Feb 02 17:08:36 GNUtoo: when yo arrive and which hotel? Feb 02 17:09:12 hrw, friday in a hotel which is at brugman Feb 02 17:09:26 hrw, no I've a camera bag Feb 02 17:10:33 hmm.. I can take bug with some modules Feb 02 17:10:40 if I bring the bug I'll try to take 4 modules Feb 02 17:10:41 or bugs even Feb 02 17:11:01 but I don't know where to put the very precious debug board Feb 02 17:11:15 bugdock? Feb 02 17:11:24 I have 2 bugs, 8 modules Feb 02 17:11:48 I've 2 bugs and 6 modules + 1 serial cable sent by you and one dock Feb 02 17:12:01 yeah Feb 02 17:12:13 build nearly finished Feb 02 17:12:18 2 screens,bug-sound,von-hippel,motion-detector and gps Feb 02 17:12:45 2xlcd, sound, vh, motion, gps, camera, wifi Feb 02 17:13:01 ok Feb 02 17:13:33 kgilmer|lunch: too bad that you will not come - my collection would grow a bit ;D Feb 02 17:13:34 avr32 build seems to be progressing as normal Feb 02 17:13:41 also we should closely monitor the devices Feb 02 17:13:55 so far we been lucky and not lost any devices Feb 02 17:14:00 ok Feb 02 17:14:05 xora or stolem Feb 02 17:14:08 ups tolen Feb 02 17:14:10 args Feb 02 17:14:12 lol Feb 02 17:14:17 woglinde: by lost I meant stolen Feb 02 17:14:34 xoa okay :) Feb 02 17:14:46 so lets boot beaglboard Feb 02 17:14:58 also I would need at least one phone for using it as a phone...maybe I should take my old phone then Feb 02 17:15:06 normal procedure is make sure stand is well watched anyway Feb 02 17:15:34 xora its not my first event Feb 02 17:15:35 also how we diferentiate between 2 equals devices that have different owners? Feb 02 17:15:41 first was back in 2000 Feb 02 17:15:56 woglinde: I know Feb 02 17:16:01 gaffertape with name on Feb 02 17:16:06 ok Feb 02 17:16:10 woglinde: just reassuring any newbs Feb 02 17:16:17 will be a little hard on the bug Feb 02 17:16:33 cebit is far worser Feb 02 17:16:33 ha yes maybe where there is written hiro-p edition Feb 02 17:16:40 cant be worse than the overo last year :-) Feb 02 17:16:51 that was so bloody tiny Feb 02 17:16:56 what was that? Feb 02 17:17:03 some overo were stolen? Feb 02 17:17:04 gumstix board Feb 02 17:17:07 I know Feb 02 17:17:14 no, it was just bloody tiny on the stand Feb 02 17:17:15 no Cebit the german it fair Feb 02 17:17:17 there is overo fire,earth etc... Feb 02 17:17:22 ok Feb 02 17:17:25 I kept worrying I would sneez and overo would fly Feb 02 17:17:51 too bad that the fire consume more than the other Feb 02 17:17:54 static discharge is probably the worst danger of the stand :-) Feb 02 17:17:58 lol Feb 02 17:18:06 everyone pokes the bare boards Feb 02 17:18:35 so bringing dock(= debug board) is not a good idea Feb 02 17:18:54 GNUtoo: if its delicate dont have it out Feb 02 17:18:58 ok Feb 02 17:18:59 it is Feb 02 17:19:00 GNUtoo: especially when there will be no BUGgers rather Feb 02 17:19:13 XorA: it is pcb without any case Feb 02 17:19:27 and it's also pre-production Feb 02 17:19:29 hrw: some like OM debug board is not delicate though Feb 02 17:19:33 hrw: or precious Feb 02 17:19:42 XorA: same stuff Feb 02 17:20:04 in past years I pnly bring devices I can live without Feb 02 17:20:12 serial+usb+ethernet+powe+reset-button+ some other unknown things Feb 02 17:20:21 GNUtoo: o yes, preproduction or even rather internal only device Feb 02 17:20:23 power means plug Feb 02 17:20:52 GNUtoo: powerled is also present Feb 02 17:21:00 yes Feb 02 17:21:28 and "the sound of the machine"(elephant dream) like audio is also present Feb 02 17:22:01 this year I had hoped to show zoom2 but Im unfortuneately working Feb 02 17:22:20 and elphel camera Feb 02 17:22:36 ok Feb 02 17:22:52 anyway Im off home Feb 02 17:22:56 too bad that zoom 2 doesn't have a modem that you can easily find Feb 02 17:23:08 hm wha arm core was bug? Feb 02 17:23:22 btw I don't know well elphel camera apaart of the fact that it runs GNU/Linux Feb 02 17:23:28 woglinde, armv6 with vfp Feb 02 17:23:35 woglinde: imx31 so arm1136 Feb 02 17:23:45 128M of ram Feb 02 17:24:17 32MB nor flash for u-boot Feb 02 17:24:26 the cpu is like 533 Mhz or something like that Feb 02 17:24:37 woglinde: ask Robert for bug - he has one Feb 02 17:24:42 533 yes Feb 02 17:24:42 hm with jazell= Feb 02 17:24:53 woglinde, he has jazelle working???!!! Feb 02 17:25:05 Features : swp half thumb fastmult vfp edsp java Feb 02 17:25:05 gnutoo I am asking Feb 02 17:25:12 no not robert Feb 02 17:25:20 but ed nevielle Feb 02 17:25:24 args neville Feb 02 17:25:35 I am going testing now on beagleboard Feb 02 17:25:39 hrw, I know but jazelle has not been reverse enginerred last time I looked(last year) Feb 02 17:25:44 yep Feb 02 17:25:58 GNUtoo no need Feb 02 17:26:07 ah? Feb 02 17:26:07 ed was lead designer of jazelle Feb 02 17:26:11 lol Feb 02 17:26:21 now he contributes to openjdk/icedtea Feb 02 17:26:29 wow Feb 02 17:26:44 so...we now have jazelle-enabled openjdk/icedtea ? Feb 02 17:26:48 here is latest change set Feb 02 17:26:50 http://icedtea.classpath.org/hg/icedtea6/rev/ba5d4644582d Feb 02 17:26:58 gnutoo I said I am testing it Feb 02 17:27:17 ok Feb 02 17:28:15 woglinde: so arm926, 1136, 1176, cortex-A{5,8,9} will get java stuff to be nicer Feb 02 17:28:16 kergoth: Can you please fix the chat group configuration? I think somebody mentioned to you what needs to be done after server move. If not, let me know and I can dig it out. Feb 02 17:28:16 ?? Feb 02 17:28:28 kergoth: BTW, why the hostile tone in the BTS? Feb 02 17:30:04 why? because you didn't take 10 minutes to bother to look at the thing you're requesting features in Feb 02 17:31:00 *sigh* Feb 02 17:31:15 3700 taks for console-image Feb 02 17:31:26 beagel-board angstroem Feb 02 17:31:29 this suckz Feb 02 17:39:01 * RP__ pushed new autoconf and automake into poky Feb 02 17:40:10 rp *g* Feb 02 17:43:57 rp oh you have co-workers which send patches Feb 02 17:45:40 re Feb 02 17:47:05 re effem Feb 02 18:05:56 anyone know whether UML is now fully mainline? Feb 02 18:06:00 (the kernel, not the modelling language) Feb 02 18:06:25 did not it died? Feb 02 18:06:31 did it? Feb 02 18:07:18 oh well, i just grab a .32 and see for myself Feb 02 18:08:17 it did not - uml-devel speaks about .32 Feb 02 18:08:25 excellent Feb 02 18:08:32 need it for my little init project Feb 02 18:08:58 always booting one of the phones is a bit too much of turnaround time ;) Feb 02 18:11:03 * hrw -> off Feb 02 18:11:17 mickeyl: qemu? Feb 02 18:15:34 too much overhead Feb 02 18:15:41 uml is faster Feb 02 18:15:53 just a small binary that i can feed with a rootfs directly on my host Feb 02 18:16:04 i don't need system emulation Feb 02 18:27:02 mickeyl, why not using the kernel thing... Feb 02 18:27:27 lguest Feb 02 18:27:31 or something like that Feb 02 18:27:56 what's that? Feb 02 18:28:24 ah, simple hypervisor Feb 02 18:29:33 yes but it permit you to run a linux kernel only Feb 02 18:29:45 and it's in mainline Feb 02 18:32:35 will consider this Feb 02 18:32:37 for now uml works fine Feb 02 18:32:38 ./linux rootfstype=hostfs rw init=/bin/bash Feb 02 18:32:41 ... Feb 02 18:32:47 root@(none):/# uname -a Feb 02 18:32:48 Linux (none) 2.6.32.7 #1 Tue Feb 2 19:16:30 CET 2010 x86_64 GNU/Linux Feb 02 18:32:49 :) Feb 02 18:34:44 RP, did a bitbake -c clean sysvinit; bitbake sysvinit, this time on my system at home, I get no ipk's in staging-pkg. Feb 02 18:34:45 http://www.pastie.org/806255 contains the full listing of the work tree. There are no .ipk files in the dir and also not in the package in DEPLOY_DIR_PSTAGE Feb 02 18:37:05 btw there is an error in log_packagedstaging_fastpath statig that it cannot stat $TMPDIR/cross/armv7a Feb 02 18:37:55 log.staging_helper and log.staging_package_installer are empty Feb 02 18:39:09 log.do_package_stage_all and log.do_package_write are also empty as is log.sysroot_stage_all Feb 02 18:43:51 (last addendum: just verified that TMPDIR/cross/armv7a exists and contains things; misread the dir it is $TMPDIR/work/beagleboard-angstrom-linux-gnueabi/sysvinit-2.86-r57/sysroot-destdir///home/frans/oe/tmp_angstrom/cross/armv7a/* : apparently the /// got me) Feb 02 18:44:09 :q Feb 02 19:52:34 eFfeM: The code which is meant to handle this is python do_package_stage () in packaged-stage.bbclass Feb 02 19:52:52 eFfeM: The bb.copyfile(srcfile, destpath + srcname) under if bb.data.inherits_class('package_ipk', d): Feb 02 19:55:37 RP__: ah thanks, was just peeking at that func but a little bit before (the if not packaged line) Feb 02 19:56:29 but i am not really a python wiz, not really an idea how to debug this Feb 02 19:56:59 eFfeM Add some "print srcname" and "print srcfile" statements into that if, see where its looking for what... Feb 02 19:57:11 then you can work out why it doesn't find it... Feb 02 19:57:23 RP__: will do , thanks Feb 02 19:59:39 i also did a diff with the poky version but did not find changes that (in my humble opinion) could explain this Feb 02 20:01:46 eFfeM: My guess is angstrom plays with some paths that break something Feb 02 20:02:06 eFfeM: but its just a guess Feb 02 20:02:33 RP__ well i was a litlte bit surprised with the /// in the path, but that is harmless Feb 02 20:02:34 eFfeM: See if packaged-staging is looking in the right place for a file of the correct name. If its not, we can learn from that Feb 02 20:03:00 * RP__ -> back in a bit Feb 02 20:03:29 RP__: catch you later, will definitely keep you posted Feb 02 20:06:08 mmm I found that: XINERAMA = "${@['--disable-xinerama','--enable-xinerama'][bb.data.getVar('MACHINE',d) in ['ion']]}" in xserver-xorg_1.7.4.bb without putting the package in the machine arch.... Feb 02 20:06:14 is it normal? Feb 02 20:09:26 RP__: we've got him :-) Feb 02 20:10:39 RP__: angstrom adds DISTRO_PR to the package name Feb 02 20:16:30 re Feb 02 20:23:46 RP__: changed Feb 02 20:23:46 srcname = bb.data.expand(pkgname + "_${PV}-" + pr + "_" + arch + ".ipk", d) Feb 02 20:23:46 into Feb 02 20:23:46 srcname = bb.data.expand(pkgname + "_${PV}-" + pr + "${DISTRO_PR}" + "_" + arch + ".ipk", d) Feb 02 20:23:46 tested it and now the ipk files end up in the .ipk in pstage, so I consider it resolved Feb 02 20:23:47 Only question left is: is this change also needed in the package_deb and package_rpm parts Feb 02 20:24:29 RP__ btw: DISTRO_PR is ?= assigned to "" so guess it is always there Feb 02 20:39:43 yeah Feb 02 20:39:45 root@beagleboard:~# java -version Feb 02 20:39:45 java version "1.6.0_17" Feb 02 20:39:45 OpenJDK Runtime Environment (IcedTea6 1.8pre) (Debian build 1.6.0_17-b17) Feb 02 20:39:46 OpenJDK Zero VM (build 14.0-b16, interpreted mode) Feb 02 20:42:25 hms Feb 02 20:42:30 not functional Feb 02 20:53:45 uh Feb 02 20:53:49 its functional Feb 02 20:53:50 but slow Feb 02 21:05:45 ok Feb 02 21:05:50 slower than without/ Feb 02 21:05:51 ? Feb 02 21:07:16 woglinde, btw do you still have your openoffice recipe? Feb 02 21:08:43 :ls Feb 02 21:08:50 dah :-) Feb 02 21:57:23 back tomorrow mornig :-) Feb 02 22:16:26 gnutoo hm yes saw it yesterday in the office :( Feb 02 22:16:50 woglinde, ok and you didn't took it with you I bet Feb 02 22:17:20 unfornatly ;( Feb 02 22:17:23 not Feb 02 22:17:25 ok Feb 02 22:17:36 and unfornatly I will be not in the office tomorrow Feb 02 22:17:41 only with luck Feb 02 22:17:47 because before buying touchbook I'd like to try to cross-compile openoffice Feb 02 22:18:07 it's the only free software I know which can make presentations Feb 02 22:18:15 you would like to do that? Feb 02 22:18:22 that doesn't sound like fun Feb 02 22:18:39 I've a computer that is on the night Feb 02 22:18:45 so I've spase cycles Feb 02 22:19:07 haha. I'm sure the build failures will waste many a cycle Feb 02 22:19:31 grg it isnt fun Feb 02 22:19:43 you have to restart and restart ad idl running Feb 02 22:20:04 ad idl ? Feb 02 22:20:28 at idl Feb 02 22:20:39 I may also have access to a dual-quad-core machine with 4GB of ram Feb 02 22:23:38 Does anyone else get messages from their chinese manufacturer like this: "the LCD driver , show erro ." but no actual error message in the email? Feb 02 22:23:45 gnutoo next week I will send it to you Feb 02 22:23:50 promised Feb 02 22:24:20 ok thanks Feb 02 22:29:20 grg: uh Feb 02 22:31:39 .... and since the software has been working fine in pre production units, shouldn't they realise that its probably a hardware issue with those samples? Feb 02 22:32:08 and perhaps not ask a software guy to remotely debug their hardware problems? Feb 02 22:32:22 * grg is just a little frustrated Feb 02 22:45:19 grg: that's a quite well known effect... Feb 02 22:54:22 hi tmbinc Feb 02 22:56:30 hi steliosk Feb 02 23:08:54 good nite Feb 02 23:17:26 I need bitbake to use the kernel headers that are specific to configured machine in order to build this recipe http://cgit.openembedded.org/cgit.cgi/openembedded/tree/recipes/xorg-driver/xf86-video-msm_git.bb . a known working hack is to replace SRC_URI in recipes/linux-libc-headers/linux-libc-headers_2.6.31.bb . how to do this correctly? Feb 02 23:37:55 If you build an appropriate MSM kernel then you can use the staging headers from that Feb 02 23:38:59 EXTRA_OECONF="--with-kernel-headers=${STAGING_KERNEL_DIR}/include" **** ENDING LOGGING AT Wed Feb 03 02:59:56 2010