**** BEGIN LOGGING AT Sun Jun 01 02:59:56 2008 Jun 01 03:50:16 03otavio 07org.oe.dev * raf52c5bd... 10/ (3 files in 3 dirs): Jun 01 03:50:16 libvncserver: add possibility to use a logger by client. partly closes 4340. Jun 01 03:50:16 * commit by Laibsch Jun 01 05:50:38 hello Jun 01 05:50:43 anybody awake? Jun 01 05:50:50 I want to investigate my QA erors Jun 01 05:51:13 QA_LOG="1" Jun 01 05:51:13 did not have any effect Jun 01 05:51:23 or I did look in the wrong places, of course Jun 01 05:56:38 03Laibsch 07org.oe.dev * r61ead9be... 10/ (1 packages/xfce/xfwm4.inc): Jun 01 05:56:38 xfwm4: don't mangle the include path Jun 01 05:56:38 * Currently all related libraries properly install at /usr/include/xfce4/ Jun 01 05:56:38 and the mangle is then breaking the compilation process. Jun 01 05:56:38 * commit by Laibsch Jun 01 05:56:43 03Laibsch 07org.oe.dev * rca2ecd71... 10/ (3 files in 3 dirs): acpid: 1.0.6 and later don't need the gcc40.patch anymore Jun 01 06:03:09 btw: how could I reduce memory consumption on the h2200 even further? (except stopping stuff I don't use) Jun 01 06:18:58 !logs Jun 01 06:18:59 Channel logs for #oe are archived at: Jun 01 06:19:00 http://hentges.net/tmp/logs/irc/%23oe Jun 01 06:19:01 Live-logs are available at Jun 01 06:19:04 http://hentges.net/tmp/logs/irc/livelogs/%23oe.livelog Jun 01 06:19:06 See ?? help-logs for usage instructions Jun 01 06:37:37 I now get: Jun 01 06:37:38 NOTE: QA checking staging Jun 01 06:37:38 ERROR: QA Issue: libgmodule-2.0.la failed sanity test (workdir) in path /usr/src/ipaq/angstrom-tmp//staging/arm-angstrom-linux-gnueabi/lib Jun 01 06:37:38 ERROR: QA Issue: libpangocairo-1.0.la failed sanity test (workdir) in path /usr/src/ipaq/angstrom-tmp//staging/arm-angstrom-linux-gnueabi/lib Jun 01 06:37:49 how do I make the QA stuff even more verbose? Jun 01 06:38:02 I do have QA_LOG enabled, so where is the log? Jun 01 07:49:02 hi Jun 01 07:49:30 is there a big difference in performance between the C7xx and a C1000? Jun 01 07:55:29 Check the specs, I would say Jun 01 08:02:36 udovdh, i'm wondering what the difference between the PXA270 and 255 Jun 01 08:02:48 since I heard the 25x series are relatively slow Jun 01 08:02:59 higher number is basically newer, bigger, better, faster Jun 01 08:03:31 http://en.wikipedia.org/wiki/XScale#PXA27x Jun 01 08:04:45 if you scroll up there is info on the 25x as well Jun 01 08:05:03 I have a 250 in my H2210 Jun 01 08:05:26 runs OK, can be overclocked for a bit more speed (in wince at least) Jun 01 08:12:54 i see Jun 01 08:14:57 akita should run on angstrom Jun 01 08:19:15 http://bugs.openembedded.net/show_bug.cgi?id=4344 also happens on stable. **** BEGIN LOGGING AT Sun Jun 01 09:27:33 2008 Jun 01 09:33:40 morning Jun 01 09:42:43 03koen 07org.oe.dev * rf89fc42c... 10/ (1 conf/distro/include/angstrom-2008-preferred-versions.inc): angstrom 2008: prefer pixman 0.10.0 Jun 01 10:09:28 cyrilRomain, good morning Jun 01 10:09:57 cyrilRomain, http://bugs.openembedded.net/show_bug.cgi?id=4344 Jun 01 10:17:02 <_gints___> does env variable CACHE should be set? .. bitbake complains that it's should .. but manuals do not say so Jun 01 10:20:00 udovdh: good, please also include the patch for linux-handhelds-2.6_2.6.21-hh20.bb (with PR = "r21" and your patch added in SRC_URI) as well (tip: use mtn to create the patch) Jun 01 10:20:18 will look into that Jun 01 10:20:26 udovdh: if a core developer find your patch OK it will be commited :) Jun 01 10:21:15 udovdh: you will have better chance the patch to be commited if it can be applied without much effort ;) Jun 01 10:23:01 I cna understand Jun 01 10:23:08 will have a look into mtn now Jun 01 10:24:07 * * OE Bug 4349 has been created by udovdh(AT)xs4all.nl Jun 01 10:24:09 * * h2200: high cpu usage Jun 01 10:24:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4349 Jun 01 10:25:17 udovdh: go into packages/linux, 'mtn add pathToYourPatch', 'mtn diff .' Jun 01 10:25:38 thanks cyrilRomain ! Jun 01 10:29:26 03koen 07org.oe.dev * ra66740dc... 10/ (3 files in 3 dirs): linux-omap2 git: bump SRCREV and tweak defconfig to get a more stable bootprocess Jun 01 10:30:19 cyrilRomain, appears to work ok Jun 01 10:30:29 I only now get all my changes that I did Jun 01 10:30:41 I only want this specific patch for the kernel compile Jun 01 10:35:27 cyrilRomain, refresh :-) Jun 01 10:35:44 oe Jun 01 10:39:27 someone knows workaround for svn bitbake get work at ubuntu? Jun 01 10:39:41 oe Jun 01 10:41:49 workaround for what Jun 01 10:41:50 ? Jun 01 10:41:53 oe Jun 01 10:42:33 it does not work at my ubuntu box Jun 01 10:42:39 oe Jun 01 10:43:00 moment I try put messages at pastebin Jun 01 10:43:01 in what way? Jun 01 10:43:05 oe Jun 01 10:43:06 ah, please do Jun 01 10:43:09 oe Jun 01 10:44:36 http://pastebin.ca/1035760 Jun 01 10:45:15 oe Jun 01 10:45:38 * oxo silences ortanc Jun 01 10:45:49 RuntimeError: dictionary changed size during iteration ? Jun 01 10:45:54 oe Jun 01 10:46:33 looks like a python issue Jun 01 10:46:37 oe Jun 01 10:46:39 what version do you have? Jun 01 10:46:41 oe Jun 01 10:46:42 moment Jun 01 10:46:44 oe Jun 01 10:46:54 how did you install it? Jun 01 10:47:06 oe Jun 01 10:47:27 I used the info from http://www.angstrom-distribution.org/building-angstrom and from http://oe.linuxtogo.org/wiki/GettingStarted Jun 01 10:47:29 with Synaptic :) Jun 01 10:47:33 and did not have your problem Jun 01 10:47:51 yet. Jun 01 10:47:56 actualy I have unstable ubuntu Jun 01 10:48:13 i am on Fedora 9 x86_64 Jun 01 10:48:15 python 2.5 I think Jun 01 10:48:27 maybe not the puthon version but the bitbake version Jun 01 10:48:35 s/put/pyt Jun 01 10:48:37 s/put/pyt/ Jun 01 10:48:59 ok.. I try get bitbake from install Jun 01 10:49:03 not svn one Jun 01 10:51:11 ok ... distribution one does not fail Jun 01 10:52:25 with the oe setup instructions I built navit (after some otehr issues that were fixed) and now I am approaching a full build of x11-image Jun 01 10:56:07 * * OE Bug 4350 has been created by udovdh(AT)xs4all.nl Jun 01 10:56:09 * * x11-image errors (QA and others) Jun 01 10:56:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4350 Jun 01 10:56:14 oe Jun 01 10:56:16 that was the one I just hit upon Jun 01 10:57:10 oe Jun 01 11:06:06 is there way to tell bitbake which version of python to use? Jun 01 11:06:50 PATH? Jun 01 11:07:28 PATH=/usr/src/ipaq/bitbake/bin:$PATH Jun 01 11:07:44 export PATH... Jun 01 11:08:30 not exactly what I mean.. but seems like system decide that.. so I have to look for my ubuntu Jun 01 11:10:43 gints, you have multiple python versions on the system? Jun 01 11:11:31 gints: if it's the case, you might need to set PYTHONPATH Jun 01 11:13:15 I had 2.5 only.. but this gives me many errors while parsing .bb files Jun 01 11:13:23 so I decided to try put 2.4 on work Jun 01 11:13:52 gints: python 2.5.x is ok to run bitbake Jun 01 11:14:16 ok .. then I have to deal with this issue - Message:'NoneType' object has no attribute 'split' Jun 01 11:14:26 I get many this kind of errors Jun 01 11:15:19 gints: sounds like a missing configuration variable .. Jun 01 11:16:42 gints: feel free to pastebin your errors Jun 01 11:17:38 there is many.. but I try put some first lines Jun 01 11:22:28 DEBUG: update_data() Jun 01 11:22:28 ERROR: Error in executing: Jun 01 11:22:28 ERROR: Exception: Message:argument of type 'NoneType' is not iterable Jun 01 11:22:28 ERROR: Printing the environment of the function Jun 01 11:22:28 DEBUG: Exception when executing anonymous function: argument of type 'NoneType' is not iterable Jun 01 11:23:14 damn it .. sory Jun 01 11:23:16 gints: which version of bitbake do you use ? Jun 01 11:23:37 http://pastebin.ca/1035787 Jun 01 11:23:57 1.8.10 Jun 01 11:25:22 1.8.11 worked ok for me so far... Jun 01 11:26:02 I try search for 1.8.11 Jun 01 11:29:40 hmm .. is not 1.8.11 like svn version? Jun 01 11:29:48 then I have tryed it Jun 01 11:29:52 gints: hmm wait. Do you have a /oe/build/conf/bitbake.conf file ? Jun 01 11:30:00 yes Jun 01 11:30:01 :) Jun 01 11:30:03 gints: you shouldn't Jun 01 11:30:10 ups Jun 01 11:30:56 gints: please read carefully the GettingStarted page (or the angstrom one) :) Jun 01 11:36:41 gah.. ok it works fine now.. for some reason I started to mess around with bitbake.conf instead of local.conf Jun 01 11:42:53 cyrilRomain, bitbake x11-image also gives me a lot of QA errors Jun 01 11:43:14 so is it my system? Jun 01 11:43:20 or what is causing them? Jun 01 11:43:38 udovdh: well some packages have QA issues, but you should'nt have so much, especially with x11-image :-/ Jun 01 11:44:15 udovdh: hmm did you deleted the tmp dir when you switched back and forth from .stable to .dev ? Jun 01 11:44:25 I filed a bug so you can see http://bugs.openembedded.org/show_bug.cgi?id=4350 Jun 01 11:44:37 I did not switch recently Jun 01 11:44:46 and yes I did delete temp from time to time. Jun 01 11:44:53 either disk full or other changes Jun 01 11:45:02 at this moment there is plenty of space Jun 01 11:45:43 besides the QA I see some other error happening at the bottom of the log Jun 01 11:45:50 is that correct? Jun 01 11:50:46 now I have to create my machine info :) from existing ones... In feed with opensourced libs of that device Jun 01 11:53:48 udovdh: I never seen that error Jun 01 11:55:18 then it must be interesting ;-) Jun 01 11:55:40 udovdh: yeah, but I don't know ipkg stuff enough to understand what's wrong :-/ Jun 01 11:55:40 a retry give the same Jun 01 11:55:53 maybe I must step back a bit more and then retry Jun 01 11:58:40 udovdh: you could find what's wrong for yourself by uncommenting line 158 of /usr/src/ipaq/angstrom-tmp/staging/x86_64-linux/bin/ipkg.py Jun 01 11:58:55 will do Jun 01 11:58:59 udovdh: it's as if a content the file were empty Jun 01 11:59:12 udovdh: at least from what the code tells me .. Jun 01 12:00:01 what job do I restart to recreate the file? Jun 01 12:00:17 no need to recreate the file Jun 01 12:00:37 just uncomment the line in-place Jun 01 12:00:50 and re-run bitbake x11-image Jun 01 12:01:15 cyrilRomain, http://rafb.net/p/s4qQMf37.html Jun 01 12:01:24 looks like one of the pkgs is not OK? Jun 01 12:01:31 or missing somehow Jun 01 12:03:44 udovdh: hmm try running 'bitbake package-index' and rerun bitbake x11-image Jun 01 12:04:49 same sort of error for the package index Jun 01 12:05:49 http://rafb.net/p/rHmwLn93.html Jun 01 12:11:14 udovdh: something is wrong with suspend-desktop_1.0-r1_all.ipk Jun 01 12:11:26 does /usr/src/ipaq/angstrom-tmp//deploy/glibc/ipk/all/suspend-desktop_1.0-r1_all.ipk exists ? Jun 01 12:12:01 $ ls -l /usr/src/ipaq/angstrom-tmp//deploy/glibc/ipk/all/suspend-desktop_1.0-r1_all.ipk Jun 01 12:12:01 -rw-rw-r-- 1 udo udo 1014 2008-06-01 12:45 /usr/src/ipaq/angstrom-tmp//deploy/glibc/ipk/all/suspend-desktop_1.0-r1_all.ipk Jun 01 12:12:03 pretty small Jun 01 12:12:09 maybe clean it out and rebuild? Jun 01 12:12:14 aha Jun 01 12:12:47 udovdh: yeah give it a try Jun 01 12:13:03 bitbake -c clean suspend-desktop Jun 01 12:13:09 bitbake x11-image. Jun 01 12:13:47 udovdh: also remove the suspend-desktop_1.0-r1_all.ipk file Jun 01 12:14:00 udovdh: because cleaning will not remove the ipkg iirc Jun 01 12:14:09 s/ipkg/ipk file/ Jun 01 12:14:18 we progress... error remains but now log shows different file Jun 01 12:22:42 now I am stuck on /usr/src/ipaq/angstrom-tmp//deploy/glibc/ipk/all/udev-hostap-cs-vcc-workaround-dev_1.0-r1_all.ipk Jun 01 12:22:46 small file Jun 01 12:22:49 and recreated as such Jun 01 12:26:32 udovdh: hmm I suspect a mistake in your configuration ending in QA issues and empty packages :-/ Jun 01 12:26:48 build.conf? Jun 01 12:27:03 local.conf I mean Jun 01 12:27:04 03mickeyl 07org.oe.dev * r26b9fba4... 10/ (1 packages/efl1/epsilon_cvs.bb): epsilon cvs as of 20080520, epsilon no longer needs imlib2 nor epeg Jun 01 12:27:07 03mickeyl 07org.oe.dev * r1ba40601... 10/ (1 conf/distro/include/sane-srcdates.inc): EFL cvs bump SRCDATE to 20080524 Jun 01 12:28:35 udovdh: or in your environment variables. Jun 01 12:29:53 will have a look Jun 01 12:31:07 flo_lap: Big thank you for representing OE at Linuxtag! Jun 01 12:32:45 cyrilRomain, some stuff: http://rafb.net/p/LSDlWY16.html Jun 01 12:32:52 looks Ok to me... Jun 01 12:35:40 udovdh: to me too Jun 01 12:36:00 maybe just rm -f the ipkgs? Jun 01 12:36:04 and try again? Jun 01 12:37:33 udovdh: yeah, it would probably solve your empty ipk issue. But you'll not get rid of QA issues though Jun 01 12:37:43 we'll see Jun 01 12:37:46 progerss step by step Jun 01 12:38:55 udovdh: your QA issue comes from files in tmp/staging directory pointing to files in tmp/work directory. Jun 01 12:39:08 how to avoid? Jun 01 12:39:45 udovdh: by writing correct do_package() function in .bb files Jun 01 12:39:59 so bug in package? Jun 01 12:40:18 udovdh: It might be. But I'm wondering why you get QA issue on packages other do not have .. Jun 01 12:40:35 still wondering here Jun 01 12:40:53 is it my host system? Jun 01 12:40:56 fedora 9 Jun 01 12:40:59 x86_64? Jun 01 12:41:05 the target? Jun 01 12:41:07 h2200? Jun 01 12:44:34 udovdh: hmm do you have a /usr/src/ipaq/angstrom-2007.1-tmp/ directory ? Jun 01 12:44:46 not anymore Jun 01 12:45:02 I have an angstrom-tmp Jun 01 12:45:35 udovdh: ok, be carefull then in your local.conf, you use ${DISTRO} for TMPDIR Jun 01 12:46:01 I have TMPDIR = "/usr/src/ipaq/${DISTRO}-tmp/" Jun 01 12:47:27 udovdh: yeah but FWIW it will not use DISTRO set in your local.conf, but from a DISTRO variable set in your environment Jun 01 12:48:14 no such variable here Jun 01 12:49:06 now I have one Jun 01 12:50:09 no difference for the TMPDIR though Jun 01 12:50:11 udovdh: no no. avoid another variable ;) Jun 01 12:53:06 udovdh: ah ok: DISTRO is set to angstrom by default, hence your angstrom-tmp. No pb here Jun 01 12:57:07 * * OE Bug 3599 has been RESOLVED (FIXED) by deller(AT)gmx.de Jun 01 12:57:09 * * MMC/SD card compatibility on Poodle Jun 01 12:57:11 * * http://bugs.openembedded.net/show_bug.cgi?id=3599 Jun 01 13:19:59 03mickeyl 07org.oe.dev * r10c1bb58... 10/ (3 files in 2 dirs): fso-image|moko-underground-image: add some tools Jun 01 13:32:48 ok... Jun 01 14:26:25 03rpurdie 07org.oe.dev * rc381734b... 10/ (8 files in 3 dirs): libtool: Add 2.2.4 (all varients) from Poky. Its not default since it will need work to get it working with all packages. Jun 01 14:43:36 cyrilRomain, building from cratch again to see if that helps Jun 01 14:43:56 maybe a mtn diff of all my changes can help see the issue in my problems? Jun 01 14:47:52 03koen 07org.oe.dev * r7dd27046... 10/ (3 files in 2 dirs): angstrom zeroconf audio: fix RDEPENDS Jun 01 14:57:17 yay! I opened a bug and fixed it at the same time! :-) Jun 01 14:58:06 * * OE Bug 4351 has been created by udovdh(AT)xs4all.nl Jun 01 14:58:08 * * navit cvs migrated to svn; edit navit_cvs.bb Jun 01 14:58:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4351 Jun 01 15:12:23 Hi, what are all these -native things bitbake is building? native to the phone (I wanna try some openmoko stuff) or native to my PC? Jun 01 15:12:48 task 330 of 4358 sounds scary :-( Jun 01 15:13:08 LionKMP: native to your PC, and required to build things for your phone Jun 01 15:13:29 cyrilRomain: but my Mandriva had most of these things already. "native" to my pc. Jun 01 15:13:36 LionKMP: many task because you're building a complete flashable Linux distribution from scratch! Jun 01 15:13:53 flashable? then native to the moko, isnt it? Jun 01 15:14:40 LionKMP: then you could speed-up the build process by using ASSUME_PROVIDED in your local.conf file Jun 01 15:15:26 LionKMP: but I would avoid that since your Mandriva may have version of software not compatible with the one required by OE Jun 01 15:15:53 ok. will I have to do this 4500 only once for all? Jun 01 15:17:10 does it build moko-native when it does not say "-native"? e.g. glib-2.0.2.16.1 Jun 01 15:18:07 LionKMP: no it does not build all -native packages per target packages, only those required Jun 01 15:19:26 what if I want to make a hello-world only? I guess a binary "openmoko os" would be enough to download, isn't it? (The how-tos are so confusing, I tried following many of them!) Jun 01 15:20:31 LionKMP: there should be openmoko SDK around. See in the openmoko wiki Jun 01 15:20:40 LionKMP: or better ask in #openmoko for that Jun 01 15:21:53 ok. I asked #openmoko but nobody is there at the moment. I'll try finding more in the Wiki. (but I tried MANY of the development howtos already, no result :) ) Jun 01 15:25:20 LionKMP: FWIW, all you need is in openmoko wiki or openmoko mailing list Jun 01 15:28:35 yes, just there are way too many options (with no "do THIS if unsure") for someone who never worked with embedded stuff or phones or so. Jun 01 15:28:56 hehe, "do THIS if unsure" would be cool. :) Jun 01 15:30:31 OEandYourDistro is spammed in OE wiki. Jun 01 15:30:54 with spam links Jun 01 15:34:44 LionKMP: we know that. There are plans to create a new wiki with protection against spams Jun 01 15:35:49 I see Jun 01 15:44:31 in which directory is the output growing for this 'bitbake openmoko-devel-image'? Where are all these glibc, perl, libxml going? Jun 01 15:45:18 LionKMP: build/tmp/ Jun 01 15:45:48 hehe, yes that's the only dir near here, but I hoped 5400 files won't go to a "tmp". ;-) Jun 01 15:46:47 LionKMP: output ends in tmp/deploy directory. It's up to you then to make backup of the deploy dir Jun 01 15:47:25 thank you Jun 01 15:48:28 it has tons of hm... luch is ready :( tons of architecture dirs in for example uild/tmp/work/x86_64-linux/qemu-native-0.9.1-r7/qemu-0.9.1 Jun 01 15:48:36 lunch even :-p Jun 01 15:48:58 I'll be back soon. thanks so far! I wonder why I have sparc32 and such dirs and a QEmu in each. Jun 01 15:49:10 I'll read manual more Jun 01 15:50:22 ;) And it's about 510,000 files in tmp for a full build of OM... Jun 01 15:50:34 MACHINE = "om-gta01" Jun 01 15:50:34 DISTRO = "openmoko" Jun 01 15:50:55 I think I didn't set a "TARGET_ARCH" :( Jun 01 15:51:57 BUILD_ARCH = "i686" Jun 01 15:52:08 x86_64 I have Jun 01 15:52:12 Did you use the Makefile to set up the environment? Jun 01 15:52:26 no, I just followed "Getting started": Jun 01 15:52:44 http://oe.linuxtogo.org/wiki/GettingStarted Jun 01 15:53:04 and adjusted the conf file according to Getting Started on openmoko site. Jun 01 15:53:38 Hopefully the docs are up to date; but I know that the Makefile is up to date! Jun 01 15:54:00 It would set up all the magic variables for you, and ensure that they are sensible. Jun 01 15:54:18 (The makefile is, however, very much frowned upon on this channel) Jun 01 15:54:40 CTRL-F Makefile on GettingStarted has no result :-/. Anyway, I'll keep on reading the whole docs first. Jun 01 15:55:08 so far it only said, how "make" has been obsolted by much better tools :-) Jun 01 15:56:17 LionKMP: again, for application development you might use the openmoko toolchain available. See http://wiki.openmoko.org/wiki/Toolchain Jun 01 15:57:08 cyrilRomain: I did that, but that creates files what I must download to the Moko (and I don't have one yet) Jun 01 15:57:22 I built their sample-app2 or what, and that's an arm app. :) Jun 01 15:57:36 using "Toolchain" binary package. Jun 01 15:57:42 LionKMP: then run it under QEmu. See http://wiki.openmoko.org/wiki/Openmoko_under_QEMU Jun 01 15:59:06 ok, thanks, I'll try that again (I tried yesterday that too, I don't remember how did it fail). Jun 01 15:59:31 ~lart scons Jun 01 15:59:32 * ibot blasts scons to oblivion with a kamehameha wave Jun 01 16:01:46 * mwester agrees with woglinde's sentiments Jun 01 16:02:04 mwester it sets -Werror somewhere Jun 01 16:02:14 and I dont know how to disable it Jun 01 16:03:39 :( It seems that whenever I run into scons at one of my customers, they too do not know how it works -- scons knowlege seems to be the domain of a very small set of people... Jun 01 16:04:40 unfornatly ok-labs use it for their l4 kernel Jun 01 16:05:08 it is the same with others tools: cmake, autotools, .. Jun 01 16:05:33 hi all Jun 01 16:05:37 he florian Jun 01 16:05:40 hi florian Jun 01 16:06:00 florian got back to home? Jun 01 16:06:06 woglinde: are you hacking with OKL4 ? ;) Jun 01 16:06:36 woglinde: with the FreeRunner ? ;) Jun 01 16:07:36 cyril hehe I want to look at it Jun 01 16:07:51 not for the freerunner Jun 01 16:07:56 only neo Jun 01 16:08:22 I dont know why the l4 people are so crazy Jun 01 16:08:36 their are they from australia Jun 01 16:08:43 the from dresden/germany Jun 01 16:08:57 and they dont work together Jun 01 16:09:37 but "stealing" code from opposite Jun 01 16:10:54 yes, that's very sad Jun 01 16:12:54 woglinde: yep... finally, arrived around 4am today Jun 01 16:13:20 flo_lap, so linuxTag went well? Jun 01 16:14:25 Crofton|work: Yes, I'm quite happy with it. We got quite some feedback and talked to several people who are using it or intend to use it. Jun 01 16:14:38 he mickeyl Jun 01 16:15:16 I should post a short summary with information... Jun 01 16:15:58 yes :) Jun 01 16:16:36 mickeyl: I just read you asked about it already :) Jun 01 16:18:06 yeah Jun 01 16:20:30 03rpurdie 07org.oe.dev * r60707f52... 10/ (1 packages/lame/lame_3.96.1.bb): lame: Run gnu-configize in do_configure to work with libtool 2.2.4 (from poky) Jun 01 16:20:35 03rpurdie 07org.oe.dev * r15ec998d... 10/ (4 files in 3 dirs): libpng12: Add fix for libtool 2.2.4 from poky Jun 01 16:20:40 03rpurdie 07org.oe.dev * r7c773905... 10/ (3 files in 3 dirs): apmd: Use libtool --tag options to avoid problems with libtool 2.2.4 (from poky) Jun 01 16:22:17 uh Jun 01 16:22:21 libtool2 transition Jun 01 16:22:58 I said I'd add the libtool 2.x .bb files and push the fixes which shouldn't break anything... Jun 01 16:23:05 well some of them Jun 01 16:23:14 rp hehe okay Jun 01 16:24:05 it starts the ball rolling but I can say for 100% certainty I won't be pushing it all the way to be made default myself ;-) Jun 01 16:26:04 I will test it too Jun 01 16:30:40 A quick survey of poky says that networkmanager glib-2.0 libusb cairo bluez-utils web-webkit hal gstreamer* claws* all need patches Jun 01 16:31:32 RP, ar the libtool2 patches work with libtool1? Jun 01 16:32:02 if not, is there a way to select patches that applied based on libtool version? Jun 01 16:34:46 Crofton|work: yes, the ones I'm applying are backwards compatible Jun 01 16:35:14 There are some in poky which aren't and they will need more thought but are thankfully a minority Jun 01 17:01:43 hmm, thats a relief. 2.6.26-rc4 boots on spitz :) Jun 01 17:02:18 I'm having terrible problems with .25 and the OSK Jun 01 17:02:46 Crofton|work: What kind of problems? Jun 01 17:04:52 i2c_omap i2c_omap.1: Transmit overflow Jun 01 17:04:52 tps65010: power off button Jun 01 17:04:59 with tmpfs on Jun 01 17:05:25 I'm sending an email to the linux-omap list Jun 01 17:06:08 I have no idea about that stuff tbh... Jun 01 17:06:19 me niether Jun 01 17:06:23 its realy frustrating Jun 01 17:06:41 hopefully someone on the omap list haas some ideas Jun 01 17:06:44 bbl Jun 01 17:36:27 03mickeyl 07org.oe.dev * re6a459c5... 10/ (1 packages/images/fso-image.bb): fso-image: noauto for nfs filesystem Jun 01 20:17:20 re Jun 01 21:52:50 03rpurdie 07org.oe.dev * rf63f2b5f... 10/ (1 packages/linux/linux-rp_2.6.24.bb): linux-rp-2.6.24: Update locomo keyboard patch for collie/poodle (#2367) Jun 01 21:52:54 03rpurdie 07org.oe.dev * r7e12dc3b... 10/ (1 packages/linux/linux-rp.inc): linux-rp.inc: Remove INITRAMFS_IMAGE setting. This is a personal (local.conf) or distro configuration option since its policy. Please discuss this on the mailing list or with me if you disagree. Jun 01 21:54:23 RP: are you around ? Jun 01 21:54:34 ant_: yes Jun 01 21:54:39 hi Jun 01 21:54:50 I'm just playing with initramfs Jun 01 21:54:54 eh Jun 01 21:56:06 the good news is thesing's bootmenu-image is very small (130 kb) Jun 01 21:56:15 cool Jun 01 21:56:29 but he somewhat forgot to commit the changes Jun 01 21:56:43 feels perhaps unfinished ? Jun 01 21:56:53 I've not used it Jun 01 21:57:15 I could post these (hope thesing would agree) Jun 01 21:57:24 My commit above probably won't make me any friends on the initramfs front :/ Jun 01 21:57:37 Putting policy into the kernel reipes is not good though... Jun 01 21:57:48 hmm Jun 01 21:57:52 RP: the actual state was already very bad Jun 01 21:57:56 was/is Jun 01 21:58:00 as long as it still can be made to work without patching the metadata Jun 01 21:58:02 handbags at dawn :-) Jun 01 21:58:02 RP I think it is okay Jun 01 21:58:39 mickeyl: IMO INITRAMFS_IMAGE should be set in the distro config or local.conf, not in the kernel recipe Jun 01 21:58:54 RP: andrea@mizar /oe/org.openembedded.dev/conf $ grep -R INITRAMFS . Jun 01 21:59:04 is the old way.. Jun 01 21:59:13 RP: in general i agree, in practice, there are some models, where it is a necessity Jun 01 21:59:31 mickeyl: The zaurus and linux-rp is not one of them ;-) Jun 01 21:59:39 collie is Jun 01 21:59:44 which is a zaurus Jun 01 21:59:45 the problem I see is the fork linux-hh linux-rp Jun 01 21:59:54 pfalcon <> thesing Jun 01 21:59:54 collie may be an exception Jun 01 22:00:22 I was forgetting collie. In that case collie.conf would be a good place to set this Jun 01 22:00:30 collie only needs it if you stick a large DISTRO on it Jun 01 22:00:52 anyhow, my point is conf files are more appropriate... Jun 01 22:01:21 ant_: yes, pflacon's approach and thesings look to be hitting head on :/ Jun 01 22:01:34 ant_: I've lost track of how they're both trying to do it too... Jun 01 22:01:35 RP: this initramfs kernel needs a separate defconfig, remember linux-rp-lite ? Jun 01 22:02:26 ant_: I remember something about it Jun 01 22:02:56 I could really use a recap on the mailing list of the current positions, what they're aiming for and what the problems are Jun 01 22:09:03 RP: http://www.pastebin.ca/1036258 Jun 01 22:09:38 and http://www.pastebin.ca/1036260 Jun 01 22:10:29 and http://www.pastebin.ca/1036262 Jun 01 22:10:37 nite Jun 01 22:12:11 ant_: Eeek, you can't do that with PV Jun 01 22:12:32 RP: I'm building now, I integrated some ideas by hand Jun 01 22:12:33 There is a bug open with a description of why somewhere Jun 01 22:12:51 RP: that's thesing work Jun 01 22:12:55 ant_: There is discussion needed on a lot of those patches Jun 01 22:13:03 ^^^ Jun 01 22:13:18 I agree with you Jun 01 22:14:38 but thesing done an awesome job porting all to klibc Jun 01 22:14:57 I'm not saying otherwise! :) Jun 01 22:15:00 *has done Jun 01 22:18:05 RP: I'd propose a linux-rp-lite and a linux-rp-lite-inc for the purpose of initramfs (needs special defconfig) Jun 01 22:18:06 * * OE Bug 2367 has been RESOLVED (FIXED) by rpurdie(AT)rpsys.net Jun 01 22:18:08 * * angstrom poodle: no suspend support with "Cancel" button Jun 01 22:18:10 * * http://bugs.openembedded.net/show_bug.cgi?id=2367 Jun 01 22:18:26 ant_: I'm ok with that in principle Jun 01 22:20:00 RP: andrea@mizar ~/devel/thesing/test $ ls -al initramfs_data.cpio.gz Jun 01 22:20:01 -rw-r--r-- 1 andrea users 109385 Jun 1 01:35 initramfs_data.cpio.gz Jun 01 22:20:33 the kernel doesn't need special diet even Jun 01 22:21:04 this is the bootmenu Jun 01 22:28:17 RP: the second step/issue will be the CONSOLE_CMDLINE (for the kexec'ed kernel) and the fact the kernel must be in /boot Jun 01 22:28:40 ant_: If there are separate recipes that shouldn't be an issue Jun 01 22:29:01 this would apply to standard images I think Jun 01 22:29:20 if we "change" boot mechanism Jun 01 22:29:27 We can change the commandline from kexec now Jun 01 22:29:33 * ant_ thinks about zaurus Jun 01 22:30:17 the first kernel must ignore the bogus angelboot parameters Jun 01 22:30:41 and must provide sane ATAGS and no cmdline to the second one Jun 01 22:30:48 is it so ? Jun 01 22:31:24 the second one (bootmenu) sees there is no cmdline and shows the menu Jun 01 22:31:42 if finds /root in cmdline boots from there Jun 01 22:32:05 I think this was the last situation / agreement Jun 01 22:32:07 * * OE Bug 3568 has been RESOLVED (FIXED) by deller(AT)gmx.de Jun 01 22:32:09 * * Poodle - WPA wifi hangs system when unplugging power cable Jun 01 22:32:11 * * http://bugs.openembedded.net/show_bug.cgi?id=3568 Jun 01 22:36:11 ant_: I'm afraid I'm not following closely enough to comment Jun 01 22:36:33 RP: throw a stone on the ML, comments will follow Jun 01 22:36:37 ;-) Jun 01 22:46:08 may I have a noob question: /usr/bin/ld: cannot find -lc Jun 01 22:46:29 in bitbake openmoko-devel-image, in the task: module-init-tools-cross-3.2.2-r3 Jun 01 22:54:25 03rpurdie 07org.oe.dev * rf49a938c... 10/ (1 packages/linux/linux-rp-2.6.24/hrw-hostapcard.patch): linux-rp-2.6.24: Uncondionally apply a number of patches, add missing patch from poky Jun 01 22:54:29 03rpurdie 07org.oe.dev * r8a01d6a6... 10/ (1 packages/linux/linux-rp_2.6.24.bb): linux-rp-2.6.24: Uncondionally apply a number of patches, add missing patch from poky Jun 01 22:54:35 03rpurdie 07org.oe.dev * r771be5e7... 10/ (13 files in 3 dirs): linux-rp-2.6.24: Sync defconfigs with Poky. Making the crypto modules builtin is wrong and compounds an existing space problem, the should be marked as autoloaded. Jun 01 22:54:44 03rpurdie 07org.oe.dev * r6ebd050b... 10/ (18 files in 3 dirs): linux-rp-2.6.23: Sync with Poky in the same way as 2.6.24 Jun 01 22:55:02 ^^^ clap clap Jun 01 23:10:38 good night **** BEGIN LOGGING AT Sun Jun 01 23:12:49 2008 **** ENDING LOGGING AT Mon Jun 02 02:59:56 2008