**** BEGIN LOGGING AT Fri Oct 29 02:59:57 2010 Oct 29 04:05:47 i most definitely downloaded the 2.6.35-rc3 kernel source, but when i run "uname -r" in my built rootfs (on my pandora), it says 2.6.27 (which is the standard kernel version for pandora) Oct 29 04:05:55 anywhere that might be getting changed? Oct 29 05:09:17 vorporeal: no you have the same kernel booting. Oct 29 05:09:31 vorporeal: are there different kernels on various partitions Oct 29 05:10:28 khem: i figured it out - the boot line given to me booted my rootfs with the kernel off the internal filesystem Oct 29 05:11:09 usually computers are seldom wrong they usually do what humans ask them to do Oct 29 05:12:07 khem: very true :) Oct 29 05:12:28 although with my ill-informed hacks of bitbake recipes to get the newer source, i assumed i screwed something up somewhere Oct 29 05:12:36 and was wondering if there was something i might have forgotten to change Oct 29 05:18:57 03Khem Raj 07master * r3b44d9f3e8 10openembedded.git/recipes/xerces-c/xerces-c_2.8.0.bb: Oct 29 05:18:57 xerces-c_2.8.0.bb: Call gnu-configize in do_configure Oct 29 05:18:57 * This gets correct config.guess and config.sub generated Oct 29 05:18:57 which have OE changes ro note uclibc based OS Oct 29 05:18:57 * Remove redundant do_stage Oct 29 05:54:24 Hi, I have written a couple of new recipes for the latest versions of a couple of applications . How do I submit this patch back to OE ? Oct 29 06:09:10 screwgoth, http://wiki.openembedded.org/index.php/How_to_submit_a_patch_to_OpenEmbedded Oct 29 06:14:45 grg: Thanks. Oct 29 06:16:22 grg: do u have logs for your latest bitbake world somewhere Oct 29 06:16:50 khem, just a moment. ill put them somewhere Oct 29 06:16:55 grg: if you can send to me or upload somewhere accessible would be great Oct 29 06:17:04 * khem goes to bed Oct 29 06:17:07 gn all Oct 29 06:18:25 night Oct 29 07:31:33 good morning Oct 29 07:35:33 moin Oct 29 07:35:40 JaMa|Off: was it kexec-eglibc? Oct 29 07:38:21 ant_work: ? Oct 29 07:38:58 ant_work: are you talking about my attempts to kexec from maemo to meego|SHR? Oct 29 07:40:56 y Oct 29 07:41:14 iirc they are using eglibc Oct 29 07:41:29 and kexec-eglibc was not tested (by me) Oct 29 07:41:38 fwiw kexec-dietlibc was failing Oct 29 07:42:24 atm kexec-glibc and kexec-klibc are working, at least on armv5te / armv4 Oct 29 07:42:46 and on x86 it seems Oct 29 07:43:17 I tried both shared and static versions Oct 29 07:43:37 ...but... Oct 29 07:43:56 it looks like on armv7 / beagle still fails miserably Oct 29 07:44:17 kernel issues iirc (registers not unset or like) Oct 29 07:53:10 ant_work: looks like glibc GNU C Library stable release version 2.5, by Roland McGrath et al. Oct 29 07:53:40 ant_work: at least for new eglibc I got "GNU C Library (EGLIBC) stable release version 2.12.1, by Roland McGrath et al." in SHR Oct 29 07:54:40 ant_work: here is list of kernel issues with meamo http://wiki.meego.com/ARM/N900/Install/kexec but those are supposed to be fixed in pr-1.3 I have now Oct 29 07:55:10 good morning Oct 29 08:38:42 likewise, ping Oct 29 09:18:46 I don't get it .. I just submitted a patch for a new recipe and I received a mail saying I can't post to the mailing list... even though I have subscribed to it months go Oct 29 09:18:47 ago Oct 29 10:03:30 screwgoth, are you sure you posted from the address you are subscribed with Oct 29 10:16:26 hi Oct 29 10:16:48 is there a place to follow oedem from remote ? Oct 29 10:42:50 ericben: I don't think so :] Oct 29 10:43:09 ynezz: hi ! Oct 29 10:43:16 hi Oct 29 10:44:35 ericben: just was looking at barebox sources, I'm thinking about trying it on ts72xx Oct 29 10:44:44 ynezz: that's a cool bootloader Oct 29 10:44:52 I have recipes to push for it Oct 29 10:45:06 to update the one I already put in oe Oct 29 10:45:07 and have seen some of your commits there Oct 29 10:45:17 so would like to ask you about your experience Oct 29 10:45:26 ts72xx it's a cirrus EP93xx CPU ? Oct 29 10:45:30 yep Oct 29 10:45:37 there's already some stuff at barebox Oct 29 10:45:45 so this may be quite direct Oct 29 10:45:47 so it shouldn't be that hard to add support for it Oct 29 10:46:16 and the EP93xx has a bootrom so you can play with the bootloader without many risk to brick your board :-) Oct 29 10:46:36 yep Oct 29 10:46:50 actually bricked it few times already :) Oct 29 10:47:31 we did a design on these CPU a few year ago. this was not really a pleasure because of all the bugs in the chips Oct 29 10:47:54 true if you need FPU, but otherwise it's not that bad Oct 29 10:48:01 or that crunch Oct 29 10:48:18 well we started with silicon rev C, then there was D1 D2 E0 D1 E2 I think Oct 29 10:48:25 yep :) Oct 29 10:48:34 so there were more problems than only the crunch ;-) Oct 29 10:48:41 which never got fixed btw Oct 29 10:48:48 yep Oct 29 10:49:55 on the ts, isn't there a fpga to handle nand ? Oct 29 10:50:19 I don't remember this board Oct 29 10:51:51 don't know offhand too Oct 29 10:51:54 but it seems so Oct 29 10:52:12 the platform nand driver is working fine, so I've never touched that part Oct 29 10:52:19 this may be the more problematic thing for the bootloader Oct 29 10:53:12 but barebox is well structured so you can try to import linux drivers in it as a basis for your work Oct 29 10:59:16 yep, it shoudln't be that hard Oct 29 11:16:26 Hi All, To disabling of option CONFIG_SYSLOGD in busybox defconfig, standalone package sysklogd is creating binary file sysklogd.sysklogd in /etc/init.d/ folder. But it is not soft linking to /etc/rc3.d folder as sysklogd. please let me know some pointers. Oct 29 11:23:55 you need to use an appropriate update-alternatives line in your postinst Oct 29 11:24:08 you could look at the sysklogd recipe in poky Oct 29 11:25:10 http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-extended/sysklogd Oct 29 11:36:49 Crofton: pong Oct 29 11:44:40 heh Oct 29 11:44:46 I forgot what I was asking Oct 29 11:46:26 Crofton: how goes oedem? Oct 29 11:46:34 good Oct 29 11:46:41 we are going to lunch Oct 29 11:46:53 smacznego then Oct 29 11:47:01 if it is good, we need to but more Intel processors :) Oct 29 11:47:07 ;D Oct 29 11:50:09 incandescant: thanks :) , I will verify. Oct 29 12:49:08 back from lunch... Oct 29 12:49:41 denix|oedem: how was it? Oct 29 12:50:26 pb__: weren't you there? Oct 29 12:50:54 denix|oedem: no, I had to do some errants for my wife Oct 29 12:50:57 errands, even Oct 29 13:07:02 hi Oct 29 13:12:15 pb__: it was Ok, better than what was at ELC... Oct 29 13:13:54 denix|oedem: hi, ELC was not that interesting ? Oct 29 13:15:25 ericben: ELC was interesting, we were talking about lunch... :) Oct 29 13:15:48 denix|oedem: ah ok ;-) Oct 29 13:16:12 did they say where will be next ELC ? Oct 29 13:18:02 San Francisco and probably Prague Oct 29 13:18:31 ok thanks Oct 29 13:45:07 Any change to the class files has to be deliberated upon in the mailing lists? Oct 29 13:45:16 yes Oct 29 13:45:25 classes/ is base of oe Oct 29 13:45:41 OK Oct 29 13:46:44 Just to ask if it's worth it, I added a base-conditional to rootfs_ipk allowing me to define PACKAGE_FORCE_OVERWRITE and have opkg ignore any duplicate files when creating an image Oct 29 13:47:30 The reason behind this is that I have a package, jlime-extras which adds some files, overwrites some others for a specific image Oct 29 13:47:47 The overwritten files are mainly icons, .desktop files and the like. Oct 29 13:48:28 Mostly desktop customization stuff. Oct 29 13:48:59 Um, distro customization stuff. Oct 29 13:49:20 Is this completely stupid or does it have a chance? Oct 29 13:49:28 I suspect that's what deliberation is for :-/ Oct 29 13:50:00 B_ Oct 29 13:50:17 B_Lizzard: you are using opkg options as discussed yesterday ? Oct 29 13:50:40 Where? Oct 29 13:50:56 in your change to rootfs_ipk ? Oct 29 13:50:59 Defining IPKG_ARGS in the image recipe doesn't do a thing Oct 29 13:51:23 yes that's why you need to hack the class Oct 29 13:51:36 Yes. Oct 29 13:51:55 Which "hack" I want to send upstream, if it is accepted. Oct 29 13:52:29 distro maintainers may not like this but for creating one image for one target this seems interesting Oct 29 13:53:00 Well, PACKAGE_INSTALL_NO_DEPS is only used in two images so ehm. Oct 29 13:53:15 And it doesn't really change anything. Oct 29 13:54:53 We'll see. Oct 29 14:00:51 Is there some switch I have to turn on to get OE to build and install the development libraries, header files, etc for packages? Oct 29 14:01:50 RobotGuy: look for the -dev packages in your deploy dir Oct 29 14:04:20 Ahhh, OK, I will look for that. I must have missed them somehow. Oct 29 14:06:47 RobotGuy: check PACKAGE_GROUP_dbg and PACKAGE_GROUP_dev in image.bbclass I think this what you are looking for Oct 29 14:09:06 Wouldn't there be something I need to set in local.conf to make OE give me the dev packages? I don't see them in deploy. Oct 29 14:13:09 RobotGuy: and here is the email presenting this : http://www.mail-archive.com/openembedded-devel@lists.openembedded.org/msg08705.html Oct 29 14:27:47 ericben, mzs Thanks. Oct 29 14:35:59 03Stefan Schmidt 07master * r0b7b712dc5 10openembedded.git/recipes/cacao/ (4 files): Oct 29 14:35:59 cacaoh-*native: Don't rely on host libtool Oct 29 14:35:59 Reported-by: Martin Jansa Oct 29 14:35:59 Signed-off-by: Stefan Schmidt Oct 29 14:37:54 * RobotGuy returns Oct 29 14:38:38 There seems to be a problem.. I added IMAGE_FEATURES += "dev" to my build recipe, but am now getting ERROR: '['/xdev/Devel/Beagle/sources/openembedded/recipes/images/walter-console-dev-image.bb']' RDEPENDS/RRECOMMENDS or otherwise requires the runtime entity 'task-sdk-native-dev' but it wasn't found in any PACKAGE or RPROVIDES variables Oct 29 14:43:16 Am I using the right branch of OE to get the dev packages? Oct 29 14:49:39 I am using "git://git.openembedded.org/openembedded" Oct 29 14:49:53 For beagleBoard Oct 29 15:09:41 RobotGuy: yes Oct 29 15:38:52 I switched to using "IMAGE_FEATURES_dev = " Is this correct to give me the dev stuff? Oct 29 15:44:33 RobotGuy: I just tried this and have the same problem : ERROR: Nothing RPROVIDES 'base-files-dev' Oct 29 15:45:00 RobotGuy: can you post a mail to the mailing list concerning this problem with PACKAGE_GROUP_ ? Oct 29 16:09:42 ericben: To which group? Oct 29 16:10:34 openembedded-devel@lists.openembedded.org Oct 29 16:10:45 bitbake is crashing for me: http://dpaste.org/SxLw/ Oct 29 16:11:01 I'm not on that one right now. Where do I sign up? Oct 29 16:11:48 http://wiki.openembedded.org/index.php/Mailing_lists Oct 29 16:12:28 this after following http://wiki.openembedded.org/index.php/Getting_started - my local.conf looks like this: http://dpaste.org/fgit/ Oct 29 16:13:33 the host system is an x86_64 Fedora 14, which also provided the bitbake Oct 29 16:15:55 muep: which ver of bitbake you have? Oct 29 16:16:05 muep: muep http://permalink.gmane.org/gmane.comp.handhelds.openembedded/38983 Oct 29 16:16:27 hrw|uds: 1.8.18 Oct 29 16:16:39 ericben: I am joining the OE lists now. Oct 29 16:17:04 RobotGuy: that's where you have some chances to get an answer Oct 29 16:17:13 hmm, let's get a more recent bitbake... Oct 29 16:18:06 muep: grab 1.10 one Oct 29 16:24:27 hrw|uds, ericben: 1.10 works better, thanks Oct 29 16:25:07 03Klaus Kurzmann 07org.openembedded.dev * rf503cc21ff 10openembedded.git/recipes/udev/ (files/nokia900/udev-rules-nokia-n900-hacks.rules udev_151.bb): Oct 29 16:25:07 udev_151: path fix in udev-rules-nokia-n900-hacks.rules and bump PR Oct 29 16:25:07 Signed-off-by: Klaus Kurzmann Oct 29 16:32:59 ericben: I just created a text file with the recipe and the error I am getting, so I can post it. Oct 29 17:19:47 is there any way to override PREFERRED_VERSION_cairo in Angstrom without modifying conf/distro/include/angstrom-2008-preferred-versions.inc? Oct 29 17:21:56 PREFERRED_VERSION_cairo_local ? Oct 29 17:24:20 effeM: so is _local something special? Oct 29 17:25:00 yeah seems to have to do with ordering, forgot details Oct 29 17:25:11 oedem/friday is closed so heading for food Oct 29 17:25:15 trying ... Oct 29 17:29:13 eFfeM_work: yes, that work, thanks! Oct 29 17:41:41 likewise: hey Oct 29 17:41:48 likewise: how was OEDEM Oct 29 17:42:14 busy; lots of Intel / TI ppl talking. I have to run now though for the informal part which is dinner. Oct 29 17:42:26 cya Oct 29 17:43:06 ok Oct 29 17:44:03 bon appetit Oct 29 17:52:28 how do i determine why a recipe is bringing in another as a dependency? Oct 29 17:52:40 i tried graphviz but the graphs are WAY too big to render Oct 29 17:53:10 vorporeal: well you can look into the .dot files and follow the trail Oct 29 17:53:11 nevermind, got it Oct 29 17:53:21 that's exactly what i just did Oct 29 17:53:23 silly me Oct 29 17:53:26 vorporeal: or you can just bitbake -g Oct 29 17:53:44 some are direct dependencies thru DEPENDS Oct 29 17:53:53 others are indirectly pulled in thru the DEPENDS Oct 29 17:55:27 got it Oct 29 17:58:27 I don't think anyone parses the .dot files Oct 29 17:58:33 I kinda wonder if they're broken, honestly Oct 29 17:58:43 but haven't tried making a tiny graph of something and getting that parsed Oct 29 17:59:12 (s/anyone parses/anyone makes pictures out of them/) Oct 29 17:59:46 i missed the beginning, what dot files are you using? Oct 29 17:59:55 the bitbake ones? Oct 29 18:00:33 the ones i've tried have borked my graphviz because they're too big. haven't cared enough to try one of the aftermarket visualizers that purports to deal with larger graphs Oct 29 18:01:46 He's using graphviz, yeah Oct 29 18:08:16 to parse a .dot file from bitbake on a recipe w/ lots of dependencies? Oct 29 18:08:25 what's the failure mode/symptom? Oct 29 18:15:48 Tartarus: are you at ELCE too Oct 29 18:17:08 Nope, took last week and this week off for http://www.flickr.com/photos/arizonarinis/5105371728/in/set-72157625218056098/ :) Oct 29 18:17:52 Tartarus: congrats :) Oct 29 18:18:35 thanks :) Oct 29 18:19:34 Tartarus: congrats Oct 29 18:19:53 Tartarus: its your second baby IIRC Oct 29 18:20:29 yeah Oct 29 18:20:35 Tartarus: when was the baby born 10/20 Oct 29 18:21:02 yeap, a nice easy to remember date, 10/20 2010, or 20/10 2010 if you prefer :) Oct 29 18:21:04 in indian way of writing the date it would be 2010 2010 :) Oct 29 18:21:25 hm.. Prague is good target to next ELC-E :) Oct 29 18:21:29 alright so you have your hands full now Oct 29 18:25:12 Tartarus: 1 boy and 1 girl now ? Oct 29 18:25:31 yeap Oct 29 18:25:44 he's going to be 2 in a few more days Oct 29 18:26:08 Tartarus: cool thats a good combo and I believe right difference I am 2 years ahead of you :) Oct 29 18:27:00 heh, fun :) Oct 29 18:27:48 now they can play with each other for few moments Oct 29 18:27:59 all you have to be good at is conflict resolution :) Oct 29 18:29:18 ;) Oct 29 18:30:09 hrw|uds: how is Orlando treating you Oct 29 18:30:25 did you visit Disneyland or not Oct 29 20:05:25 hi concerning the dot files, I think they are broken : http://bugs.openembedded.org/show_bug.cgi?id=5471 Oct 29 20:06:52 congratulation Tartarus , I was in the same configuration 2.5 months ago ;-) Oct 29 20:07:22 thanks :) Oct 29 20:07:36 and .... good nights :-D Oct 29 20:08:26 ericben: fyi the link to the fixed version is broken, heh Oct 29 20:08:50 ericben: oh, yeah, i remember that. the reason graphviz screwed up was because the graph was too large, but the graph was only too large because of that stupid bug where each character got its own vertex :P Oct 29 20:09:09 Tartarus: which fixed version ? Oct 29 20:09:43 "This problem was first discussed and fixed on 2009-04 here : http://bugs.openembedded.net/0 Oct 29 20:09:43 " Oct 29 20:09:49 This problem was first discussed and fixed on 2009-04 here : Oct 29 20:09:50 yeah Oct 29 20:09:58 ah ok Oct 29 20:10:57 I've reworked busybox recipes to factorize them as it was a big mess, will post patches to the list but I think several vversions could also be removed Oct 29 20:12:06 probably, yes Oct 29 20:12:23 I think busybox hasn't been put to the same "remove non-pinned versions" stuff others have Oct 29 20:12:35 And that's probably unwise Oct 29 20:14:04 i might try to just hack a fix for the dotfile stuff, it looks straightforward enough (though i don't know python :P ) Oct 29 20:14:20 ok original link for the bug was : http://bugs.openembedded.org/show_bug.cgi?id=5002 Oct 29 20:15:47 mrj10: I mark the bug as confirmed as you also meet it ? Oct 29 20:16:44 mmm it seems buzilla don't like the url Oct 29 20:20:01 yeah Oct 29 20:23:20 I'm testing this on the host http://www.tenshu.net/terminator/ it's quite interesting when we need several terminals and don't want to use the mouse to much Oct 29 20:36:10 * Tartarus is a screen guy but opens that up Oct 29 20:36:56 hmm, maybe that + screen could cover some good cases, heh Oct 29 20:38:42 yes, i'm using this + byobu over screen Oct 29 20:39:11 * mrj10 feels it's his duty to bring up tmux, as someone does whenever screen is mentioned http://tmux.sourceforge.net/ Oct 29 20:41:49 interesting that's terminator in screen instead of having screen in terminator ;-) Oct 29 20:44:13 mrj10, intersting, but I'm only allowed one big switch per decade and I've already moved from bitchx to other things for IRC this decade :) Oct 29 20:44:34 Tartarus: irssi? Oct 29 20:44:53 * hrw|uds is using screen since 1996 Oct 29 20:45:10 I learnt screen second day of using Unix Oct 29 20:45:39 pidgin for work (since it does other protocols), python client a buddy writes for freenode/linuxnet/efnet Oct 29 20:50:07 have a nice weekend Oct 29 20:50:17 for me it will be mostly travelling back home Oct 29 20:50:22 have a nice trip :-) Oct 29 20:50:25 thx Oct 29 20:50:34 and have fun with the mx51 :-D Oct 29 20:50:46 I am not at all sure that package grouping is going to allow me to get dev packages - with all headers and dev libs - installed. Oct 29 20:51:58 RobotGuy, it should be able to Oct 29 20:52:19 Since -dev and -doc are rules based things, get it done right in one spot and all pkgs will ahve the right magic Oct 29 20:53:08 Tartarus: In theory, yes, that would be ideal. However, I'm still not sure it works like that. Oct 29 20:53:40 There doesn't seem to be any -dev packages. Oct 29 20:54:01 RobotGuy: post the error you get Oct 29 20:55:04 ericben: I am going to do that right now. Oct 29 20:56:01 I love how being a developer keeps sucking me into different things, just so I can do what I do. ;) Oct 29 20:58:58 why should one use tmux instead of screen except screen is GNUism Oct 29 21:01:45 khem, looks like different features in some cases Oct 29 21:02:43 i am by no means an expert on either, so i defer to http://tmux.cvs.sourceforge.net/viewvc/tmux/tmux/FAQ http://blogs.techrepublic.com.com/opensource/?p=1901&tag=nl.e102 Oct 29 21:02:55 both sources are biased, i'd say, but some of the feature differences are, i think, objective Oct 29 21:03:12 i'm not a power user so i like screen just fine Oct 29 21:03:23 but people over at http://news.ycombinator.com seem to prefer tmux by and large Oct 29 21:03:26 for me screen serves my purpose Oct 29 21:03:37 screen mutt irssi Oct 29 21:03:38 yeah, i wouldn't switch if it works fine Oct 29 21:03:40 all set Oct 29 21:05:06 ericben: Posted it just now. Oct 29 21:10:06 Tartarus: gcc kung-fu here http://pastebin.com/bCb3ps5P do u see any issues ? Oct 29 21:16:01 * saa doesn't see in tmux the equivalent of: screen /dev/ttyS0 115200,-crtscts,-ixon,-ixoff Oct 29 21:17:18 ericben: you around? Oct 29 21:17:27 i unborked the bitbake graphviz files Oct 29 21:17:38 actually, i forget who originally had problems. tartarus? Oct 29 21:17:56 mrj10: if you fixed it thats nice Oct 29 21:18:10 send the patch Oct 29 21:20:52 mrj10: cool ! Oct 29 21:22:33 it was the same approach as the previous fix, applied different places in lib/bb/cooker.py Oct 29 21:23:29 most of the rdepends structure is lists of strings, but rdepends and rrecs are lists of characters, for whatever reason, so you have to ''.join() them before doing the regex search Oct 29 21:24:00 i think the proper fix is to fix whoever is inserting that stuff the wrong way, but i dont want to deal with making sure that all the other consumers of that data, outside the graphviz printer, still function properly Oct 29 21:24:51 diff -uNr lib/bb/cooker.py.busted lib/bb/cooker.py Oct 29 21:24:51 --- lib/bb/cooker.py.busted 2010-10-29 16:15:01.157515847 -0500 Oct 29 21:24:51 +++ lib/bb/cooker.py 2010-10-29 16:17:12.367546106 -0500 Oct 29 21:24:51 @@ -435,11 +435,11 @@ Oct 29 21:24:51 for depend in depgraph["depends"][pn]: Oct 29 21:24:51 print >> depends_file, '"%s" -> "%s"' % (package, depend) Oct 29 21:24:51 for package in depgraph["rdepends-pkg"]: Oct 29 21:24:52 - for rdepend in depgraph["rdepends-pkg"][package]: Oct 29 21:24:52 - print >> depends_file, '"%s" -> "%s" [style=dashed]' % (package, rdepend) Oct 29 21:24:53 + for rdepend in re.findall("([\w.-]+)(\ $.+$)?", ''.join(depgraph["rdepends-pkg"][package])): Oct 29 21:24:53 + print >> depends_file, '"%s" -> "%s%s" [style=dashed]' % (package, rdepend[0], rdepend[1]) Oct 29 21:24:54 for package in depgraph["rrecs-pkg"]: Oct 29 21:24:54 - for rdepend in depgraph["rrecs-pkg"][package]: Oct 29 21:24:55 - print >> depends_file, '"%s" -> "%s" [style=dashed]' % (package, rdepend) Oct 29 21:24:55 ouch Oct 29 21:25:11 (sorry about that): khem: should i send to the openembedded-devel list, or is there a separate bitbake one? Oct 29 21:25:41 mrj10: yes oe list Oct 29 21:25:48 mrj10: bitbake-dev@lists.berlios.de Oct 29 21:26:03 mrj10: next time dont paste so much here Oct 29 21:26:10 yeah, sorry about that Oct 29 21:26:37 do i have to subscribe to bitbake-dev? Oct 29 21:29:30 mrj10: I don''t think so Oct 29 21:29:40 but send it also to oe please (so that I get it and test it ;-) Oct 29 21:33:58 sent Oct 29 21:34:01 to both Oct 29 21:35:39 mrj10: patch looks sane from quick look but it needs to be tested Oct 29 21:35:44 khem: re http://pastebin.com/bCb3ps5P what's the multilib bit? Oct 29 21:36:18 saa: yeah, screen does make a handy terminal program for serial stuff too Oct 29 21:36:26 I think i picked that one up from you :) Oct 29 21:37:18 Tartarus: thats for SH4 Oct 29 21:37:34 it need compiler to have nofpu because kernel uses it Oct 29 21:37:54 Tartarus: otherewise the default is always m4 Oct 29 21:38:06 saa: are you hacking on OE these days ? Oct 29 21:38:20 I would say using it. Oct 29 21:38:36 So far, so good. Oct 29 21:38:43 for work ? or personal use Oct 29 21:38:48 Work Oct 29 21:38:54 cool Oct 29 21:39:12 Yup. Much better than the bad old days of embedded. Oct 29 21:39:25 heh yes we all have seen them :) Oct 29 21:39:49 saa: are you at Google ? Oct 29 21:39:53 Tartarus: congratulations on your girl! Oct 29 21:40:06 khem: yup, I'm contracting with them. Oct 29 21:40:12 saa: cool. Oct 29 21:40:22 Tartarus: I forwarded the link on to Kristin Oct 29 21:40:41 saa: say my hello to Kristin as well Oct 29 21:41:02 Will do. Oct 29 21:41:03 saa, thanks Oct 29 21:41:51 mrj10: your patch need few tuning for bitbake master Oct 29 21:42:18 khem, m4 bit looks fine then, just sep from the libgcc fixup? Oct 29 21:42:23 but the that works Oct 29 21:42:36 Tartarus: yes these are two different commits Oct 29 21:42:51 Tartarus: I am knee deep in libtool shit Oct 29 21:43:06 so I need to comeout clean as quick as I can Oct 29 21:43:50 mrj10: your kernel patch needs rework too Oct 29 21:44:00 mrj10: basically just bump the right pr bits Oct 29 21:45:11 Tartarus: your opinion on nuking those all unused/broken recipes from toolchain ? Oct 29 21:47:41 what recipes? Oct 29 21:47:47 * Tartarus isn't reading oe-devel until monday :) Oct 29 21:47:52 oh ok Oct 29 21:48:18 basically I am proposing to remove some old binutils/gcc/glibc/uclibc recipe removal from master Oct 29 21:48:33 its PITA to fix anything in those Oct 29 21:48:45 khem: trrue Oct 29 21:48:53 and we hardly use 3 compiler versions and 2 binutils versions Oct 29 21:49:00 khem, i'm a little torn, but maybe it's OK depending on the versions / types left Oct 29 21:49:10 thats the request Oct 29 21:49:18 identify what you need Oct 29 21:49:22 rest will go away Oct 29 21:49:26 I'll reply monday, ok? :) Oct 29 21:49:35 yes I have given 1 weeks time Oct 29 21:49:40 ericben: thanks for the heads-up, i don't think i'll have time to tune up the patch in the next couple days, so if you'd like to clean it up it'd be greatly appreciated, otherwise i can do it next week Oct 29 21:53:00 time to go to daughter's halloween parade Oct 29 22:10:18 * Jay7 wish to have some script to check that current system have all needed tools to build something with OE Oct 29 22:11:07 that's because I'm installing pre-requisites in FreeBSD VM Oct 29 22:22:12 are there any recipes that create /etc/ld.so.conf.d and /etc/ld.so.conf? seems like perhaps something base-files would do Oct 29 23:30:54 when looking for pinned verisons of a packages : is there somewhere to check except from conf directory ? Oct 29 23:48:39 ericben: DEFAULT_PREFERENCE var in the recipes also should be checked Oct 29 23:48:59 khem: OK I missed this one Oct 29 23:49:41 ericben: too late in EU sleep ! Oct 29 23:50:20 gotta reboot this system Oct 29 23:50:22 brb Oct 30 00:28:21 khem: as busybox 1.17.3 is considered as the latest stable version : what about removing DEFUALt_PREFRENCE=-1 to this version ? Oct 30 00:28:56 this way 1.11.3 and 1.14.3 can go out quite safely Oct 30 00:35:28 03Eric Bénard 07org.openembedded.dev * r291693ea81 10openembedded.git/recipes/busybox/files/find-touchscreen.sh: Oct 30 00:35:28 busybox/find-touchscreen: handle input/device case Oct 30 00:35:28 On a new configuration using 2.6.36 kernel + busybox 1.17.1 , Oct 30 00:35:28 find-toucscreen.sh fails because $MDEV is input/event0 and not event0. Oct 30 00:35:28 This patch fixes this problem by removing input/ if present in $MDEV. Oct 30 00:35:35 03Eric Bénard 07org.openembedded.dev * rab9224ff76 10openembedded.git/recipes/busybox/ (10 files in 2 dirs): Oct 30 00:35:35 busybox: add 1.17.3 Oct 30 00:35:35 this patch adds the 1.17.3 release of busybox and its up to date Oct 30 00:35:35 patches. Oct 30 00:35:36 Signed-off-by: Eric Bénard Oct 30 00:35:40 03Eric Bénard 07org.openembedded.dev * rcd7c9542ba 10openembedded.git/recipes/busybox/ (busybox_1.1x.inc files/mdev.conf): Oct 30 00:35:40 busybox's mdev: add script to populate /dev/bus/usb Oct 30 00:35:40 this patch allows libusb to work on kernel configured without Oct 30 00:35:40 CONFIG_USB_DEVICEFS Oct 30 00:35:53 oops forget to add khem's signoff Oct 30 00:36:06 or ack Oct 30 00:36:28 ok bye, now it's really time to sleep :-) Oct 30 00:36:33 its fine as long as patch is looked at Oct 30 00:36:36 gn Oct 30 00:36:55 ericben: yes please make 1.17.1 as default Oct 30 00:37:10 usually when someone adds a new version it should become default Oct 30 00:37:13