**** BEGIN LOGGING AT Wed Jun 04 02:59:57 2008 Jun 04 06:26:08 Yo Jun 04 07:17:29 heh, hey Speedy2, been a while Jun 04 07:17:35 hey kergoth Jun 04 08:25:50 03woglinde2 07org.oe.dev * r0d4e37bd... 10/ (3 files in 2 dirs): Jun 04 08:25:50 sablevm: fix download dir, now sourceforge Jun 04 08:25:50 * patch was provided by Elena "of Valhalla" Jun 04 08:25:50 * bug 4320 can not be closed, because it was filed against stable Jun 04 08:31:52 morning Jun 04 08:34:17 hallo Jun 04 08:40:25 03woglinde2 07org.oe.dev * r53e608ae... 10/ (3 files in 2 dirs): Jun 04 08:40:25 gdb-cross-sdk: include patches from the gdb-cross package for version 6.6 Jun 04 08:40:25 * bump PR Jun 04 08:40:25 * rework SRC_URI for gdb-cross 6.6 too, no need to bump the PR Jun 04 08:40:25 * fixes partly bug 4216 Jun 04 09:39:15 good morning Jun 04 09:42:37 yo florian Jun 04 09:43:02 hey florian Jun 04 09:43:46 03koen 07org.oe.dev * rd2747ed6... 10/ (3 files in 3 dirs): libsdl-x11: add (disabled) 1.2.11 for people with libtool 2.2.4 Jun 04 10:10:49 03koen 07org.oe.dev * r2bb52edf... 10/ (1 packages/kerneloops packages/kerneloops/kerneloops_0.11.bb): kerneloops: add 0.11 Jun 04 10:10:54 03koen 07org.oe.dev * r6bc539b8... 10/ (1 packages/libtool/libtool-native_2.2.4.bb): libtool-native: also stage libltdl header so -native packages that build against libltdl work (e.g. guile-native) Jun 04 10:21:34 03xjqian 07org.oe.dev * r8a346ac0... 10/ (6 files in 3 dirs): Jun 04 10:21:34 navit: upstream SRC_URL change from cvs to svn Jun 04 10:21:34 * tag svn srcrev 1096 Jun 04 10:21:34 * remove navit.xml-so.patch from _svn.bb Jun 04 10:21:34 * partly close oebug 4351, need to bp to stable. Jun 04 10:30:00 03woglinde2 07org.oe.dev * r9dee7a87... 10/ (5 files in 2 dirs): Jun 04 10:30:00 m4: update to version 1.4.11 Jun 04 10:30:00 * reasons are some security fixes Jun 04 10:30:00 * seems to build fine and runs fine so no DEFAULT_PREFERENCES Jun 04 10:30:00 * keep older recipes Jun 04 10:47:16 03koen 07org.oe.dev * r906860b5... 10/ (3 files in 3 dirs): guile-native: update to 1.8.5 and sanitize configure.in Jun 04 11:55:28 03koen 07org.oe.dev * r335a4917... 10/ (1 packages/gnome/goffice_0.6.3.bb): goffice: update to 0.6.3 Jun 04 12:44:26 :( uclibc_0.9.29.bb config can't find someone can help? Jun 04 12:45:32 gints: aeh? Jun 04 12:45:56 gints: that is native curses? Jun 04 12:46:03 no Jun 04 12:46:08 hi zecke Jun 04 12:46:24 gints: well, you have to be more precise ;) Jun 04 12:46:41 I try put output of log .. moment Jun 04 12:46:46 gints: unlikely that a compile of uclibc needs curses.h as curses needs a libc :) Jun 04 12:47:34 it's like gcc-cross-initial step Jun 04 12:49:54 http://pastebin.ca/1038587 Jun 04 12:53:41 ~curse ext3 for slowness Jun 04 12:53:44 May you be reincarnated as a Windows XP administrator, ext3 for slowness ! Jun 04 12:56:51 ~hi hrw :) Jun 04 12:56:52 Many greetings, hrw :), most strange traveller, to this IRCdom of plenty. Jun 04 13:00:41 hello everyone Jun 04 13:01:41 total n00b question: how do I build an individual module for my kernel? I need to build a network module so I can get my gumstix on the network Jun 04 13:04:42 Crofton|work: is the democrat result good or bad? Jun 04 13:33:09 ping cbrake about mediawiki Jun 04 13:33:29 dito hrw and mickey_away Jun 04 13:34:04 Laibsch: +1 from me as always Jun 04 13:34:17 hrw: good Jun 04 13:34:17 +1 Jun 04 13:34:38 the question was, are you willing to act as backup admins for the wiki ;-) Jun 04 13:34:48 so, you both agreed ;-) Jun 04 13:35:16 eeks Jun 04 13:35:18 ;) Jun 04 13:35:22 cbrake had committed publically, I'll do the same, so we should have enough redundancy Jun 04 13:36:10 I think there were some volunteers during previous discussions about mediawiki. I'll dig up those contacts and ask them, too, if it makes sense. Jun 04 13:37:11 did we elect a wiki team? Jun 04 13:37:33 one person team: Laibsch ;) Jun 04 13:37:40 reminds me, I really should resign from angstrom core team Jun 04 13:38:02 hrw: as long as we make it formal that he is Wiki team leader, then Im happy Jun 04 13:38:18 cbrake volunteered, too Jun 04 13:38:22 so Jun 04 13:38:29 Currently, we have cbrake and me as main admins, hrw and mickey for standby in case the mysql server needs a kick or some such. Jun 04 13:39:33 Laibsch: I am not admin in OE Jun 04 13:39:56 OK, that can be helped Jun 04 13:39:58 XorA: you are not they only one Jun 04 13:40:16 hrw: Depending on what makes sense Jun 04 13:40:24 either admin role in wiki soft Jun 04 13:40:28 sudo powers Jun 04 13:40:31 or full root Jun 04 13:40:48 I want some redundancy Jun 04 13:41:19 Not because I think the work will be overwhelming, but becuase redundancy is the right thing to have[tm] Jun 04 13:42:23 XorA: ya, i should, too Jun 04 13:44:06 not because I hate Angstrom, but I have so little free time Jun 04 13:44:08 mickey_away: How about some security updates while we are at it? They are at least three weeks overdue Jun 04 13:44:21 and I must ebay my zaurus Jun 04 13:44:32 * Laibsch proposes unattended-updates for security updates Jun 04 13:44:41 mickey_away: OK with you? Jun 04 13:45:01 sure Jun 04 13:45:22 OK Jun 04 13:45:36 wow, debian stable has 1.7 mediawiki Jun 04 13:45:44 Let's see if that is sufficient Jun 04 14:01:23 wrt to bug 3938, is there a problem to unconditionally rmmod a module even if not present in most circumstances? Why does Yan seem to need rmmod in addition to "pccardctl eject"? Jun 04 14:03:52 hello Jun 04 14:05:07 where can QA logging normally be found? Jun 04 14:20:42 udovdh: Try "find $OETEMP|grep -i qa|grep log" to find out Jun 04 14:21:11 mickeyl: can you please point wiki.openembedded.net to amethyst? Jun 04 14:21:44 Laibsch, running find... Jun 04 14:22:29 found something! Jun 04 14:22:32 thanks Jun 04 14:24:35 03koen 07org.oe.dev * r86141812... 10/ (1 packages/gnumeric/gnumeric_1.8.3.bb): gnumeric: add 1.8.3 Jun 04 14:24:39 added the findings to http://bugs.openembedded.net/show_bug.cgi?id=4348#c2 Jun 04 14:24:40 03koen 07org.oe.dev * r7dc73088... 10/ (1 packages/abiword/abiword_2.6.3.bb): abiword 2.6.3: tweak packaging Jun 04 14:24:52 any idea about fixing? (if at all) Jun 04 14:25:50 is the QA erroring because of the host system? (Fedora 9 on x86_64) Jun 04 14:25:54 or the recipe? Jun 04 14:26:03 or the sources? Jun 04 14:26:31 'evil hides inside the .la, staging, /staging/arm-angstrom-linux-gnueabi/lib/libmudflap.la' ???? Jun 04 14:31:10 bye Jun 04 14:33:39 Laibsch, got any idea about the evil I found? Jun 04 14:35:01 can anyone point me in the right direction on how to build a single module using bitbake? Jun 04 14:35:33 .la contains installed=yes ? Jun 04 14:35:39 what is that for a problem? Jun 04 14:38:28 so it is my .la files Jun 04 14:38:35 how could I fix them? Jun 04 14:52:32 Laibsch: I'd be glad to help in any way I can Jun 04 14:52:45 great Jun 04 14:52:56 I am currently installing mediawiki from backports Jun 04 14:53:07 once mickeyl has updated dns we can start rolling Jun 04 14:53:25 cbrake: are you aware of any moin2mediawiki converter? Jun 04 14:53:37 I think there are some, but I'd need to check it out Jun 04 14:53:49 Laibsch: I have not looked recently Jun 04 14:54:04 well, maybe it would be better to start from scratch anyway Jun 04 14:54:07 Laibsch: are we going to try to convert the old-old OE wiki? Jun 04 14:54:14 and hand-pick the most important pages to copy over Jun 04 14:54:15 Laibsch: I'd vote starting from scratch Jun 04 14:54:20 Laibsch: agreed Jun 04 14:54:21 it's OK Jun 04 14:54:52 I'd probably copy the stuff under policy that I kept in shape, plus the last know-good FAQ maybe Jun 04 14:55:03 we should have a brief look through what is applicable Jun 04 14:56:03 Laibsch: sounds good Jun 04 14:56:31 Laibsch: oe wiki is not mail imho Jun 04 14:57:00 03koen 07org.oe.dev * r0853a10f... 10/ (9 files in 4 dirs): angstrom led config: initscript that applies led config from /etc/default/leds Jun 04 14:57:46 florian: ? Jun 04 14:57:49 mail? Jun 04 14:58:24 eh sorry - "moin" Jun 04 14:58:41 ah, yes right Jun 04 14:58:56 drupal plugin with mediawiki syntax, right? Jun 04 14:59:06 I am always mistaken about that Jun 04 14:59:17 Should make it easier to copy over that odd page or two Jun 04 14:59:58 * florian votes for taking care about not loosing content Jun 04 15:01:31 Laibsch: looks like the OeFaq page is gone ... ohh, well it was getting pretty crusty anyway Jun 04 15:07:43 Laibsch: the new page is gonna be mediawiki based? on the list you discussed mediawiki+git, is that happening? Jun 04 15:08:04 mediawiki+git? Jun 04 15:08:11 Don't remember Jun 04 15:08:16 very unlikely Jun 04 15:08:35 if there is an addon, we might support that *later* in case people want it Jun 04 15:11:48 there was discussion about allowing for offline editing/control revisions Jun 04 15:13:06 it was ikiwiki not mediawiki iirc Jun 04 15:13:30 right, and then potential ways to make mediawiki have that functionality Jun 04 15:13:33 * hrw|gone -> gone now Jun 04 15:13:41 it doesn't look like it's definitive, I was just curious Jun 04 15:14:09 hi rschuster Jun 04 15:14:30 hi florian Jun 04 15:51:53 03koen 07org.oe.dev * r2697c4d7... 10/ (6 files in 3 dirs): webkit: update makefiles for x86_64 and add acinclude.m4 to get rid of infinite libtool loop Jun 04 18:05:16 hi ph5 Jun 04 18:05:52 'evil hides inside the .la, staging, /staging/arm-angstrom-linux-gnueabi/lib/libmudflap.la' Jun 04 18:06:00 anyone knows a fix or solution for that? Jun 04 18:06:24 udovdh whats wrong with libmudflap.la? Jun 04 18:06:42 that is the QA log entry Jun 04 18:06:48 after a navit bitbake Jun 04 18:06:58 part of the QA log, ther's a lot more Jun 04 18:07:07 I need to understand that 'evil' part Jun 04 18:07:11 and know how to fix it Jun 04 18:07:28 bottom of http://bugs.openembedded.net/show_bug.cgi?id=4348 Jun 04 18:07:32 could you pastebin it Jun 04 18:07:33 ah Jun 04 18:07:34 okay Jun 04 18:08:01 I found the qa logging and that is what it says (also one other .la issue) for multiple files Jun 04 18:08:23 is that stable or dev? Jun 04 18:08:38 angstrom is dev Jun 04 18:08:41 navit is cvs Jun 04 18:08:56 no Jun 04 18:08:59 angstrom is stable. Jun 04 18:09:05 navit is dev Jun 04 18:09:16 because navit needs to run on 2007.12 Jun 04 18:09:20 what? Jun 04 18:09:22 so I pull navit from their svn Jun 04 18:09:44 the angstrom repo only has older versions of navit pre-built Jun 04 18:09:47 hm seems no one backports some stage thinks from dev which take for the .la files Jun 04 18:09:52 and I want to be able to use the svn of navit Jun 04 18:10:06 what do la files do? Jun 04 18:10:11 I know so and ar etc Jun 04 18:10:23 so it is an oe/angstrom issue? Jun 04 18:10:34 I also have a number of these when I make x11-image Jun 04 18:10:38 and mtd-utils Jun 04 18:10:43 .la files is for libtool it tolds libtool how to link against librarys Jun 04 18:10:44 see (my) bugs Jun 04 18:10:50 yes Jun 04 18:10:55 ask koen to fix it Jun 04 18:10:59 for stable Jun 04 18:11:11 via email? maillist? is he in irc? Jun 04 18:11:41 mailinglist and point tothe bug Jun 04 18:11:52 ok, weill have a look at oe maillist then. Jun 04 18:11:53 tahnks Jun 04 18:11:55 thanks Jun 04 18:13:06 cbrake, summatusmentis: please create yourselves an account at http://wiki.openembedded.net Jun 04 18:13:39 Let me know about suggestions you may have Jun 04 18:14:14 hi, all! Jun 04 18:14:28 anybody tried to nuild ctorrent lately? Jun 04 18:14:49 s/nuild/build Jun 04 18:14:53 hi slapin long time not seen Jun 04 18:21:15 hi crofton Jun 04 18:21:43 gm Jun 04 18:25:07 * * OE Bug 4355 has been created by  Jun 04 18:25:09 * * spam protection Jun 04 18:25:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4355 Jun 04 18:26:27 slapin: I think I did Jun 04 18:26:32 I also did some work on it Jun 04 18:26:42 do you have any problems with it? Jun 04 18:28:07 * * OE Bug 4356 has been created by  Jun 04 18:28:09 * * decide on some wiki policies Jun 04 18:28:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4356 Jun 04 18:31:50 hi zecke, mickeyl Jun 04 18:32:08 pb__: master. Jun 04 18:37:52 hi pb Jun 04 19:09:42 hi woglinde Jun 04 19:10:31 rp ping Jun 04 19:18:32 hi all, it seems that whenever I update the OE dev branch, any attempts to re-build my image (x11-gpe-image) always fail at some package (in my current case, uclibc). However, like the last time this happened, I bet if I wiped out my build/tmp directory and let it rebuild everything from scratch, the image would complete successfully. Is there something I can do to avoid having to start from square one every time I update? Jun 04 19:21:34 loudawg use stable Jun 04 19:21:42 dev can brake often Jun 04 19:21:48 thats why its dev Jun 04 19:22:49 woglinde, I don't think that's the problem. I think it's a bitbake issue or something where it doesn't clean and rebuild all of the appropriate dependencies. That's my guess anyway. The fact that I can wipe everything out and THEN it will build successfully is what makes me think this Jun 04 19:24:39 the funny thing is, the last time I tried using stable it was a bigger headache than dev was, so I ditched it Jun 04 19:25:24 is there a way to tell bitbake to rebuild not only the specified package, but all of its dependencies as well? Jun 04 19:26:37 liudawg??? glibc an gcc changes are much often Jun 04 19:26:51 and it runs not always clean when you recompile gcc Jun 04 19:27:04 because a patch was applied Jun 04 19:31:15 hello, what is the easiest way to build an extra kernel module using bitbake? Jun 04 19:32:01 jadam look at the madwifi.bb Jun 04 19:32:08 or hostap.bb Jun 04 19:40:21 03woglinde2 07org.oe.dev * rd81f21e7... 10/ (1 packages/gdb/gdb-cross-sdk_6.6.bb): Jun 04 19:40:21 gdb-cross-sdk: remove sim-install-6.6.patch Jun 04 19:40:21 * this patch does nothing reasonable Jun 04 19:47:44 Laibsch: great -- will start working on it tomorrow morning :-) Jun 04 20:09:30 Laibsch, I had problems yesterday it did not build (failed to do_install, IIRC), have ever been any changes since then? ver 3.x is chosen in my case. Jun 04 20:23:15 hello Jun 04 20:34:12 hi there Jun 04 20:35:41 HopsNBarley: hi Jun 04 20:35:43 hernan_martinez: hi Jun 04 20:36:11 hi like2wise Jun 04 20:36:25 had coffee with Khem the other day. great guy. Jun 04 20:36:32 hi hopsnbarley Jun 04 20:36:55 HopsNBarley: shhh, he is listening in :-) Jun 04 20:37:10 oh crap! Jun 04 20:37:48 hi Jun 04 20:38:15 the lib cheaders version .. do you know where i set its version? Jun 04 20:38:40 i didnt to anything and it starts saying that my 2.6.25 prefered version is not found, and it fallbacks to 2.6.23 Jun 04 20:38:44 HopsNBarley: good to see the OE community grow, not only online Jun 04 20:39:02 libc headers* Jun 04 20:39:45 hernan_martinez: check if the kernel is there for your platform. AFAIK, linux_2.6.25.bb only is "unlocked" for some powerpc board and maybe something else. Jun 04 20:40:04 hernan_martinez: ah, the headers might not even be in OE... don't know. Jun 04 20:40:17 its okay , i dont needthe .2.6.25 i am okay with 2.6.23 Jun 04 20:40:25 but my kernel version is defaulting to 2.6.21 Jun 04 20:40:51 i tried changing my local.conf but bitbake -s still returns linux - 2.6.21 Jun 04 20:41:00 and libc-headers - 2.6.23 Jun 04 20:41:21 hernan_martinez: same thing, read packages/linux/linux-...bb and check if DEFAULT_PREFERENCE != -1 for your platform. Jun 04 20:41:35 ok Jun 04 20:41:35 tks Jun 04 20:45:25 -1 is to lock it? Jun 04 20:46:07 hernan_martinez: DEFAULT_PREFERENCE = "-1" prevent bitbake from choosing it in the dependency tree Jun 04 20:46:20 ok Jun 04 20:46:30 yes, my 2.6.23 was with -1 Jun 04 20:47:40 hernan_martinez: it probably was for a good reason. Looks at the log maybe you will find why Jun 04 20:47:45 hi everybody Jun 04 20:47:48 hernan_martinez: then you are the "first" to use that kernel for your platform. It probably does not have a "defconfig" yet in a sub-dir. Jun 04 20:48:03 it was -1 for def_pref.. but it was 1 for a couple of other machines Jun 04 20:48:21 its interesting that deselecting SMP on new kernels i get a FATAL: kernel too old from glibc I guess, since init says that the kernel is too old Jun 04 20:48:23 I'd like to override some files from base-files with my own custom version, is there any recommended method of achieving this? Jun 04 20:49:09 I'd like not to duplicate the entire base-files package, I just want to extend it to use some of my files instead. Jun 04 20:49:10 Fidget: of course. Just look at the basefiles dir. There are many examples. Jun 04 20:49:21 thesing: what examples? Jun 04 20:49:48 hi thesing Jun 04 20:50:02 have a look at the files dir. Jun 04 20:50:17 thesing: :) Jun 04 20:52:17 03woglinde2 07org.oe.dev * r3d5d2241... 10/ (3 files in 2 dirs): Jun 04 20:52:17 gdb-cross-sdk: add dependency on ncurses-cross-sdk Jun 04 20:52:17 * stolen from poky, courtesy to Richard Purdie Jun 04 20:52:21 03woglinde2 07org.oe.dev * rbcb31dc3... 10/ (6 files in 4 dirs): Jun 04 20:52:21 gdb-cross: fix some stuff for version 6.6 6.8 Jun 04 20:52:21 * remove sim-install-6.6.patch, it is the same patch as Jun 04 20:52:21 sim-install-makefile.patch and does nothing useful on 6.6 and 6.8 Jun 04 20:52:23 * drop early_debug_in_nptl.patch for 6.8 because it is long time applied Jun 04 20:52:25 upstream Jun 04 20:52:27 * bump PR for both Jun 04 20:52:55 Oooh Jun 04 20:53:07 I think that means my gdb-cross bug can be closed Jun 04 20:53:10 * Tartarus makes a note to check later Jun 04 20:55:41 hi florian Jun 04 20:56:01 re Jun 04 21:03:19 re Jun 04 21:14:07 * * OE Bug 4357 has been created by mmviii(AT)eclipticon.net Jun 04 21:14:09 * * ipkg update/upgrade renders HX4700 unusable Jun 04 21:14:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4357 Jun 04 21:24:07 * * OE Bug 4216 has been RESOLVED (FIXED) by heinold(AT)inf.fu-berlin.de Jun 04 21:24:09 * * differ Jun 04 21:24:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4216 **** BEGIN LOGGING AT Wed Jun 04 21:28:31 2008 Jun 04 21:36:46 Package ncurses wants to install file /usr/lib/libncurses.so.5 But that file is already provided by package libncurses5. any ideas? Jun 04 21:54:22 like2wise: hey Jun 04 21:54:52 hi khem Jun 04 21:55:07 woglinde: hi Jun 04 21:55:12 khem could please the bugs in the bugtracker when you apply your patches self Jun 04 21:55:17 +close Jun 04 21:55:24 woglinde: will do Jun 04 21:55:40 woglinde: I look once a week in the bugs concerning me. Jun 04 21:56:39 woglinde: I am ploughing through gcc 4.3 still not able to build a console-image Jun 04 21:57:13 khem http://bugs.openembedded.net/show_bug.cgi?id=4331 for instance Jun 04 21:58:27 woglinde: there was someone else who also said he had some patches for gcc 4.3 Jun 04 21:58:34 so I kept it open Jun 04 21:58:56 woglinde: I wish commits mails were also added to bugs as comments automatically Jun 04 21:59:58 khem better close the bug and open a new one Jun 04 22:00:04 or open a meta bug Jun 04 22:01:02 woglinde: ok will do Jun 04 22:01:37 hmmm with gcc 4.3 I think we need to fix packages which have misused inline before Jun 04 22:01:56 gcc 4.3 is complying to c99 now Jun 04 22:02:04 yes Jun 04 22:02:18 avahi build failure Jun 04 22:02:23 I am seeing demos this Jun 04 22:02:28 we should steal from distributions debian/redhat/gentoo/whatever Jun 04 22:02:34 :) Jun 04 22:02:47 I feel better if I reinvent the wheel Jun 04 22:03:22 leaverage is the mangerial word for stealing :) Jun 04 22:04:47 it saves a lot of time Jun 04 22:05:09 not inventing the wheel again and again Jun 04 22:05:16 yeah true Jun 04 22:06:30 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31389 Jun 04 22:08:20 over one year old Jun 04 22:11:17 khem but I didnt have a package yet which failed with gcc-4.3 koen fixed them all before me Jun 04 22:14:24 * Khem stealing Jun 04 22:14:42 http://bugs.gentoo.org/attachment.cgi?id=136465 Jun 04 22:15:11 woglinde: what image are you building with gcc 4.3 Jun 04 22:15:22 there are few problems that I am seeing Jun 04 22:15:31 oabi Jun 04 22:15:39 eabi Jun 04 22:15:40 armv4 Jun 04 22:15:48 angstrom-2008.1 Jun 04 22:15:58 but I cannt test all Jun 04 22:16:23 right Jun 04 22:17:49 ping cbrake Jun 04 22:23:06 * * OE Bug 4331 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Jun 04 22:23:08 * * GCC 4.3.0 needs patches Jun 04 22:23:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4331 Jun 04 22:24:21 woglinde: why do u still use oabi ? Jun 04 22:26:43 woglinde: hmm because you have armv4 ok got it. May be you should consider upgrading. Jun 04 22:26:50 HopsNBarley: hey :) Jun 04 22:27:43 khem upgrade my simpad? Jun 04 22:27:50 replace the strongarm with what? Jun 04 22:28:15 * Laibsch points out for Khem that he uses oabi too (for the collie) Jun 04 22:28:47 woglinde: :) Jun 04 22:29:00 woglinde: is it armv4 in there or armv4t Jun 04 22:29:25 strongarm sa-1100 Jun 04 22:29:32 no thumb mode Jun 04 22:30:37 ah th brutus not assabet Jun 04 22:31:22 ahno Jun 04 22:31:25 sa1110 Jun 04 22:31:50 damnit it, I always exchange this Jun 04 22:32:11 http://opensimpad.org/index.php/Hardware_Details Jun 04 22:32:35 ok so its assabet based Jun 04 22:35:55 still armv4 Jun 04 22:36:25 *g* Jun 04 22:36:43 its from 1995 Jun 04 22:36:45 If I want to override some machine specific base-files in my user.collection, is it possible to do that without duplicating the entire base-files package? Jun 04 22:37:49 fidget .bb or bbclass? Jun 04 22:38:52 woglinde: I have user.collection/packages/base-files/base-files//fstab that I want to override org.oe/packages/base-files/base-files/ Jun 04 22:39:11 I want to use the .bb in org.oe/packages/base-files/base-files_XYZ.bb Jun 04 22:41:10 fidget hm Jun 04 22:41:52 fidget sorry dont know exactly how the overlay is working Jun 04 22:41:58 Fidget: so you have a package in your collection that you do not want ? Jun 04 22:42:11 isnt is working as you suggest it Jun 04 22:42:25 khem I think the other way Jun 04 22:42:32 Khem: I have a package in org.oe and I want a few files from user.collection to override the files in the org.oe package Jun 04 22:42:37 he wants to overide the fstab from oe Jun 04 22:43:12 I'd like not to copy the entire base-files into my user collection Jun 04 22:43:24 I just want to override some specific files Jun 04 22:45:53 give it a higher version Jun 04 22:46:04 or PREFERENCE Jun 04 22:48:59 * cyrilRomain wonders between using a collection, or managing patchset over oe with quilt/stgit (for now I always did the later) Jun 04 22:49:29 My advice is to go with the collection Jun 04 22:51:19 Khem: I really want overlays to be overlays, not just package overlays Jun 04 22:51:47 Khem: I hear overlay, and I think unionfs, which is not at all what oe attempts to do, its a first come first serve package search Jun 04 22:53:34 actually, that gives me an evil idea for a build system that relies on a unionfs system for merging a bunch of trees into a single namespace Jun 04 22:54:16 Fidget: I have not used overlays so much Jun 04 22:54:29 Fidget: if you use git, you'd better create a branch with your packages and modifications over oe files, that you rebase to keep in sync with oe Jun 04 22:54:42 But I think if your collection appear first in search path then it should be searched first Jun 04 22:54:53 I have my own collection overlay Jun 04 22:55:18 the problem is that I have to create a full package, where all I want is to overlay a single file here and there Jun 04 22:55:33 so really I want the seperate collections Jun 04 22:55:55 but rather than searching each collection via the normal overlay process, I want them unionfs'ed together so that there's only one search repository Jun 04 22:57:34 Fidget: I like the unionfs approach, except I'm wondering how difficult it is to keep in sync with oe, (e.g. if a file you 'overrided' is modified in oe) Jun 04 22:58:16 cyrilRomain: well seperate collections are good so you'd have my-oe/{org.oe,user.collection} Jun 04 22:58:29 then create a my-oe/union-overlay Jun 04 22:58:46 and then you can have as many collections as you want, and use unionfs to unite them in union-overlay Jun 04 22:58:55 then there's only one search path Jun 04 22:59:19 would have to find rules to specify that user.collection overrides org.oe inside of union-overlay Jun 04 22:59:57 but I think you get the idea, the org.oe collection is independent branch that users would only keep in sync, they would put modifications inside of user.collection in a parallel structure to org.oe (as normal overlays do) Jun 04 23:00:20 Fidget: yes. I'm far from being against collection :) I'm just wondering how do you cope with modification of file that you overrides Jun 04 23:00:21 but there'd be more redundancy elilmination using unionfs in that if you wanted to override a single file, you could achieve that Jun 04 23:00:56 well if my user.collection/file override org.oe/file Jun 04 23:01:19 and I sync and org.oe/file has a newer version, I would assume that user.collection/file might still win out Jun 04 23:01:57 there needs to be a way to prune directory trees and specify logical overlay ordering / timestamp overlay ordering Jun 04 23:02:08 Fidget: but how can you actually *see* that org.oe/file has a newer version? Jun 04 23:02:09 that way everything can be super confusing Jun 04 23:02:19 cd org.oe; svn status ? Jun 04 23:02:24 when pulling many file has potentially changed Jun 04 23:03:36 well, I'm afraid this approach can lead to 'missing fix' in your overlay file (I think you see on what I'm trying to point out) Jun 04 23:03:44 treat org.oe as some remote tree (what it is) that you do not maintain (unless you do, in which case use your VC tools for updates) Jun 04 23:04:17 cyrilRomain: That I update org.oe and a security fix or update occurs that should override my user.collection file? Jun 04 23:04:54 Fidget: for instance, let's say you override libsdl-mixer.bb file Jun 04 23:04:57 cyrilRomain: Perhaps just use timestamp comparison to choose which file shows up in the overlay-union Jun 04 23:05:10 and someone fix e.g. the do_package() of libsdl-mixer.bb Jun 04 23:05:34 When I do an svn status, it should tell me that libsdl-mixer.bb was updated Jun 04 23:06:16 *or* because when I do an svn update, my local repo gets syncrhonized and the timestamps will update to something newer than my override of libsdl-mixer.bb Jun 04 23:06:27 wohoo gcc 4.3 rfs boots on my osk Jun 04 23:06:35 osk? Jun 04 23:06:50 omap5912 board Jun 04 23:07:02 Fidget: well, IHMO it's ok only if you have a few .bb files to manage :-/ Jun 04 23:07:39 Fidget: with many .bb and/or many changes in those files, I'm afraid it might becomes a nightmare to sync without a proper tool :-/ Jun 04 23:09:53 woglinde: libsdl-mixer compiled fine here (from oe repo) Jun 04 23:10:05 cyrilRomain: If you have many .bb files, you are probably maintaing your own local copy, might as well just do a recursive touch in your user.collection/packages Jun 04 23:10:06 * cyrilRomain diffing with OM repo Jun 04 23:10:23 find user.collection/packages -type f -exec touch {} \; Jun 04 23:11:40 Fidget: but how do you merge or 'cherry-pick' changes from oe files ? Jun 04 23:12:20 cyrilRomain: diff, patch, and manual inspection? Jun 04 23:12:27 03khem 07org.oe.dev * r52d1f4c1... 10/ (3 files in 3 dirs): Make inline patch also applicable to gcc 4.3 now that gcc 4.3 implements C99 conformant inline behaviour. Jun 04 23:13:10 Fidget: well this is rather tedious IHMO, when git can do that for you Jun 04 23:13:40 cyrilRomain: an overlay is a place for local packages. My suggestion was for increasing the flexibility of overlays so that you can cherry pick from an personal overlay to override a single file in a standard overlay like org.oe Jun 04 23:13:46 cyrilRomain: not familiar with git Jun 04 23:13:57 Fidget: a good solution would be to have both the advantages of a unionfs (file in a separate directory), and a SCM tools for diff/merge/patch Jun 04 23:15:05 cyrilRomain: if you intend to work on the org.oe tree, then yes, but if you have a custom set of packages, or are overlaying a specific file here and there in the org.oe tree, its overkill doing any VC stuff. Jun 04 23:15:48 Fidget: right. It depends on the number of file you want to 'override' Jun 04 23:16:18 cyrilRomain: depends on use case too Jun 04 23:16:28 Fidget: for one or two files, an VC can be overkill. For more files ... Jun 04 23:17:49 cyrilRomain: My current case is I need to modify fstab, so I had to copy the entire base-files tree into my local overlay (user.collection). Then modify fstab. Now how do you recommend I keep my local copy of base-files in sync with org.oe when all I want is a personal fstab. Jun 04 23:18:05 cyrilRomain: I'm really asking Jun 04 23:20:31 Fidget: well I would keep a personal patch with git, so that I can rebase the patch to keep in sync if the oe dudes change it Jun 04 23:20:35 Fidget: but you could keep an complete personal fstab, well in a unionfs is a rather sensible solution IMHO Jun 04 23:21:42 Fidget: in fact one can see an overkill using git/quilt, one can see an overkill to use unionfs ;) Jun 04 23:21:44 cyrilRomain: we should discuss the whole SCM tools some other time, I'm interested in exploring that idea Jun 04 23:22:00 cyrilRomain: Isn't using oe overkill compared to ubuntu? Jun 04 23:22:15 Fidget: ? Jun 04 23:23:21 Fidget: to build a desktop distro like ubuntu maybe, otherwise I don't understand your question Jun 04 23:23:36 cyrilRomain: I was making a joke Jun 04 23:24:00 Fidget: :) Jun 04 23:25:08 nite Jun 04 23:26:07 nite all Jun 04 23:36:51 Khem, ping Jun 04 23:37:13 Tartarus: pong Jun 04 23:37:24 Khem, gcc testsuite question for ya Jun 04 23:37:36 finally got things setup so 'make check' will run on a target Jun 04 23:37:46 Tartarus: k Jun 04 23:37:56 Now, in the context of make check Jun 04 23:38:01 what does unresolved mean? Jun 04 23:38:11 for a few days ago armv4t-gnueabi I get: Jun 04 23:38:20 * Tartarus pulls up pastebin Jun 04 23:39:05 Tartarus: for cross-testing gcc I usually invoke runtest manually and not do make check Jun 04 23:39:06 http://pastebin.com/m4db4cd47 Jun 04 23:39:54 I see Jun 04 23:40:21 So what's unresolved mean in the dejagnu context? :) Jun 04 23:40:22 unresolved means it was not able to run the tests due to some env issue Jun 04 23:40:40 or it could not determine the right way to run it Jun 04 23:41:00 generally this number is not that high Jun 04 23:41:08 so something is surely not alright there Jun 04 23:41:13 ok Jun 04 23:41:31 how did u set it up for cross testing Jun 04 23:41:49 env Jun 04 23:42:05 sec Jun 04 23:42:21 RUNTESTFLAGS="--target_board=myboard" make check Jun 04 23:42:22 and... Jun 04 23:42:26 load_generic_config "unix"; Jun 04 23:42:30 set_board_info gcc,timeout 5400 Jun 04 23:42:30 set_board_info g++,timeout 5400 Jun 04 23:42:30 set_board_info gfortran,timeout 5400 Jun 04 23:42:31 ok Jun 04 23:42:37 and then set_board_info for telnet/ftp Jun 04 23:44:08 ok Jun 04 23:44:15 do u know if it is talking to target Jun 04 23:44:33 I generally use ssh/scp Jun 04 23:45:03 set_board_info rsh_prog /usr/bin/ssh Jun 04 23:45:03 set_board_info rcp_prog /usr/bin/scp Jun 04 23:45:41 and I think you also need to set the compiler which you want to test Jun 04 23:45:43 Yes, I had inetd -d going at one point, I know we're ftp'ing in Jun 04 23:45:53 and telnet'ing Jun 04 23:46:05 set_board_info compiler Jun 04 23:46:07 can you pastebin a board config you have that works? Jun 04 23:46:20 ok Jun 04 23:47:06 * * OE Bug 4281 has been RESOLVED (FIXED) by trini(AT)embeddedalley.com Jun 04 23:47:08 * * mtd-utils-dev is missing headers Jun 04 23:47:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4281 Jun 04 23:47:18 * * OE Bug 4280 has been RESOLVED (FIXED) by trini(AT)embeddedalley.com Jun 04 23:47:20 * * The /usr/include/scsi headers are missing in some cases Jun 04 23:47:22 * * http://bugs.openembedded.net/show_bug.cgi?id=4280 Jun 04 23:50:10 Tartarus: http://pastebin.com/m255da717 Jun 04 23:51:09 Khem, OK, thanks Jun 04 23:51:16 I'll update my setup and-rerun Jun 04 23:51:24 which arch is it ? Jun 04 23:51:44 armv4 Jun 04 23:51:54 (ARM920T core) Jun 04 23:52:02 ah ok Jun 04 23:52:12 dont you something fast Jun 04 23:52:14 and nww Jun 04 23:52:17 new Jun 04 23:52:21 this will take ages Jun 04 23:52:28 Well, it's actually "new" and somewhat fast Jun 04 23:52:28 to run gcc tests on this boarf Jun 04 23:52:40 And it's got 128M ram :) Jun 04 23:52:49 oh thats better Jun 04 23:52:56 Mine is 32M Jun 04 23:53:03 so I am doomed Jun 04 23:53:25 sometimes gcc itself will need that much :) Jun 04 23:54:05 but on cross it wont matter much Jun 04 23:54:13 yeah Jun 04 23:54:19 Tartarus: are you using EABI ? Jun 04 23:54:22 yeah Jun 04 23:54:33 so its v4t Jun 04 23:54:39 yup Jun 04 23:54:46 thought i said that sorry :) Jun 04 23:55:08 I'm having problems with svn using an https server, how can I make it auto accept the cert from within OE? Jun 04 23:56:50 ERROR: Certificate verification error for svn.int.sigovs.com: unable to get local issuer certificate Jun 04 23:56:50 To connect to svn.int.sigovs.com insecurely, use `--no-check-certificate'. Jun 04 23:57:46 thats my internal test server Jun 04 23:57:56 and it won't use my use environment Jun 04 23:58:33 anyone? Jun 04 23:59:31 Hmm Jun 04 23:59:35 In local.conf, try... Jun 05 00:00:13 SVNCOOPTS=--no-check-certificate Jun 05 00:00:14 if you can do svn co --no-check-certificate .. Jun 05 00:00:23 thanks Jun 05 00:00:25 Otherwise, steal {FETCH,UPDATE}COMMAND_svn from bitbake.conf Jun 05 00:00:29 and update to have that in Jun 05 00:00:33 Tartarus: what gcc version are you using Jun 05 00:00:39 Khem, 4.2.4 Jun 05 00:01:02 OE.dev from a few days ago Jun 05 00:01:09 (Just did a pull and saw some changes go in) Jun 05 00:02:31 http://gcc.gnu.org/ml/gcc-testresults/2008-06/msg00261.html Jun 05 00:02:38 Next question, how do I get OE to use only my internal test server to pull from rather than trying to go to a distro server for a package it doesn't have? Jun 05 00:02:39 Tartarus: you can search for 4.2 Jun 05 00:02:46 thanks Jun 05 00:02:53 I dont see anyone reporting 4.2 though Jun 05 00:02:59 only 4.3 and 4.4 Jun 05 00:03:09 Fidget, "only" is hard Jun 05 00:03:16 Fidget, but see the angstrom-mirrors class Jun 05 00:03:29 If its found in your mirror first, you're good Jun 05 00:05:44 Tartarus: SVNCOOPTS=--no-check-certificate, doesn't seem to help Jun 05 00:06:08 Fidget, OK, time to override FETCHCOMMAND_svn and such Jun 05 00:06:10 see bitbake.conf Jun 05 00:06:15 thanks Jun 05 00:06:32 where is that? I'd rather use you than find Jun 05 00:07:26 openembedded/conf/bitbake.conf Jun 05 00:07:58 Tartarus: should I be using local.conf and not site.conf? Jun 05 00:08:22 policy question, i'm not good with that Jun 05 00:08:30 but, iirc, local.conf is the final place for overrides Jun 05 00:08:54 (ie if you put X=1 in bitbake.conf, X=2 in site.conf and X=3 in local.conf, X=3 will be what your builds get) Jun 05 00:09:35 so I need to setup SVNCOOPTS in my regular shell environment so that bitbake.conf will use that Jun 05 00:09:51 put it in local.conf and it'll be seen Jun 05 00:10:25 hmm Jun 05 00:10:26 ok Jun 05 00:20:54 Tartarus: nothing seems to be working, thanks a lot **** ENDING LOGGING AT Thu Jun 05 02:59:57 2008