**** BEGIN LOGGING AT Sun Jan 23 02:59:57 2011 Jan 23 11:53:51 03Koen Kooi  07org.openembedded.dev * r3b6561e026 10openembedded.git/recipes/linux/ (18 files in 4 dirs): Jan 23 11:53:51 linux-omap 2.6.37: fix mmc regulators when using wl1271 expansion board Jan 23 11:53:51 linux-omap 2.6.37: cherry-pick a few wl1271 patches Jan 23 11:53:51 * tested on a beagleboard xM with wl1271 expansionboard Jan 23 11:53:52 linux-omap 2.6.37: update wl127xx patchset to include BT and FM drivers Jan 23 11:53:52 Signed-off-by: Koen Kooi Jan 23 11:53:57 03Koen Kooi  07org.openembedded.dev * r1db2deca67 10openembedded.git/recipes/ti/ti-wifi-utils_git.bb: Jan 23 11:53:57 ti-wifi-utils: add git version Jan 23 11:53:57 Signed-off-by: Koen Kooi Jan 23 12:25:35 <_alx_> hi GNUtoo|laptop :) Jan 23 12:25:42 hi Jan 23 12:26:35 <_alx_> i'm getting Xorg segfault :( Jan 23 12:26:43 that's bad Jan 23 12:26:49 do you know how to gdb it? Jan 23 12:27:06 <_alx_> no :( Jan 23 12:27:41 <_alx_> reading information for Xorg.log I discover that the sig11 came from fb discovery Jan 23 12:28:02 http://www.elinux.org/GDB Jan 23 13:39:09 <_alx_> GNUtoo|laptop, do you know why virtual/kernel create an image for uboot (uImage) ? Jan 23 13:44:26 _alx_, yes Jan 23 13:44:32 look at the machine configs Jan 23 13:44:40 grep for uImage Jan 23 13:45:11 <_alx_> no result Jan 23 14:39:37 <_alx_> GNUtoo|laptop, how come that I don't have any /dev/fb0 ? Jan 23 14:58:41 _alx_, ah? Jan 23 14:59:01 _alx_, do you have the framebuffer driver for your board in the kernel? Jan 23 15:02:49 <_alx_> GNUtoo|laptop, I suggest so: it's an x86 based board Jan 23 15:02:58 <_alx_> GNUtoo|laptop, how to check if I'm doing it right ? Jan 23 15:03:11 add VESA framebuffer then, if it has a BIOS Jan 23 15:04:47 <_alx_> how to check if it's present ? Jan 23 17:02:22 good morning (or afternoon) Jan 23 17:07:58 <_alx_> or evening Jan 23 17:08:00 <_alx_> who knows Jan 23 17:10:38 I'm using a local overlay an always when I make a receipt with a "require" of a file that exists in the same directory in the upstream tree it cannot find it. Jan 23 17:11:18 I seem to have some problem with the include/search path. Jan 23 17:11:18 <_alx_> Andy_, you should copy needed files in the overlay recipe Jan 23 17:11:47 <_alx_> anyone using with success qt4-x11-base-image ? Jan 23 17:12:10 Why? Is it not the idea to re-use included and not changed files from their original place? Jan 23 17:15:10 <_alx_> Andy_, somewhere, somehow I read that included files must be copied :( Jan 23 17:21:08 I am having trouble building MACHINE="mpc8323e-rdb" where an error occurs somewhere during the building of glibc. Is MACHINE="mpc8323e-rdb" known to work? Jan 23 17:35:26 03Chris Larson  07master * rc03c179bb1 10openembedded.git/recipes/netperf/ (files/vfork.patch netperf_2.4.4.bb): Jan 23 17:35:26 netperf: fix patch Jan 23 17:35:26 Signed-off-by: Chris Larson Jan 23 17:53:11 _alx_, hi any luck with framebuffer, what is your video card, and what xf86-video-something do you use? Jan 23 17:54:18 have we mirrored that ncurses patch yet? Jan 23 17:54:40 <_alx_> on qemu don't know, on real machine it is ... who know either. I first need it to be working on qemu Jan 23 17:55:33 what's not working on qemu? Jan 23 17:55:58 I fired up x11-image shortly after I changed over to xserver-xorg for qemu stuff, on qemux86+DISTRO=minimal Jan 23 18:02:03 <_alx_> Tartarus, well, I need python and qt so i fired qt4-x11-base-image and qt4-x11-demo-image, the latter has a script to launch Xorg and then qtdemo. Simply they don't find any fb device. Nevertheless I don't talk that qemux86 machine prefer Xvesa on Xorg so the script provided by qtdemo-init.bb is wrong. Jan 23 18:02:36 <_alx_> tried to start Xorg from console, cannot find FB device, while I get sig11 trying Xorg -configure. Jan 23 18:03:16 Well, start with just x11-image and confirm it works Jan 23 18:03:24 Then it should be a simple'ish matter to see what the others do wrong Jan 23 18:04:20 <_alx_> x11-image or xorg-image ? Jan 23 18:04:47 x11-image Jan 23 18:08:22 <_alx_> what are the differences ? Jan 23 18:08:34 Well, for qemux86 nothing Jan 23 18:09:18 <_alx_> the real board is using an atom Jan 23 18:09:23 heh Jan 23 18:09:44 <_alx_> it's not so far different from qemu :) Jan 23 18:19:45 _alx_, regarding copying include files. This means that also some .bb files has to be copied to local overlay when setting up my own local overlay .bb files for toolchains. Jan 23 18:24:49 Anybody out there how knows about why files (in upstream OE tree) required from receipt in local overlay is not found? Jan 23 18:26:15 that depends on what files you're talking about Jan 23 18:28:01 I've just set up toolchain .bb files in local overlay to build SDK (as described in the OE user manual). These receipts require som other .bb files that are not in the local overlay but in the upstram OE tree. Jan 23 18:29:18 kergoth, one of the files not found is task-sdk-host.bb Jan 23 18:30:49 include/require are found via BBPATH Jan 23 18:31:37 I'll check that. Jan 23 18:33:43 How deep down do BBPATH have to be? Now I have it set to ...build: ...local: ...openembedded. Meaning that both the top of my local overlay and the top of the upstream OE is in BBPATH. Jan 23 18:33:50 there is no "deep" Jan 23 18:34:01 if i include 'foo.bb', it looks for foo.bb in teh root of each dir in bbpath Jan 23 18:34:17 it also happens that hte dir where a recipe resides gets temporarily added to bbpath Jan 23 18:34:41 if, on the other hand, i include 'recipes/foo/foo.inc' or similar... Jan 23 18:36:56 So what, the best thing is to specify the .inc file or .bb file with require and the full path from the top? Like recipes/foo/foo.bb? Jan 23 18:37:13 if you want it to be able to find that included file in other collections/layers, yes Jan 23 18:37:28 it really depends on what you're trying to accomplish Jan 23 18:39:16 I've been making my own adapted versions of some receipts like the toolchain/SDK receipts which are referring to other .inc or .bb files from the standard upstream collection Jan 23 18:39:40 It would be nice to just keep on using the untouched standard files without copying them. Jan 23 18:40:35 right, in that case you'd want to include the path from the root of the collection Jan 23 18:42:50 Well, that is a ok solution. Nicer would be to have a default dynamich search path set in some way to search for included files in both the local overlay and in the same place in the upstream tree. Jan 23 18:43:36 that isn't how bitbake works, and that would be prone to problems. bitbake doesn't prevent you from having multiple recipes with the exact same filename in a single collection Jan 23 18:46:41 How do you mean that this prevents from searching for files in two or more directory trees? If the include file does not exist in the same dir as the receipt that is parsed it could find the inlcude file in another dir. Jan 23 18:46:52 Couldnt that work? Jan 23 18:47:17 Just like the search for the .bb files in general? Jan 23 18:48:40 .bb files aren't searched in the common case, for one Jan 23 18:48:51 and once again, there's nothing stopping you from having 73 foo.inc's in your tree Jan 23 18:48:56 what magic do you think its going to use to pick one? Jan 23 18:51:53 advanced magic Jan 23 18:51:59 Yes, i can have 73 foo.inc. But the include of one now from my receipt is the one in the same directory as the receipt itself (in the local overlay). If not found there a potential good place too look would be in the same directory in the upstream tree. Jan 23 18:52:20 Sorry if the comments/questions are stupid. :-) Jan 23 18:54:27 I am wondering why I cannot "touch .config" in tmp/work/...linux-2.6.34/" and then "bitbake virtual/kernel" and get a new uImage? Jan 23 18:55:20 ckrinke, because the kernel tree is already built Jan 23 18:55:51 ckrinke, the status about comiled packages are kept in some status files. Jan 23 18:56:16 Ok, but, .... what I need to do is modify a kernel tree for a new board based on an existing reference board to get a new kernel. Obviously I dont "grok" the OE way yet. Jan 23 18:56:31 ckrinke, try to do a "bitbake -f virtual/kernel" to force a build Jan 23 18:57:06 ckrinke, maybe you have to be more specific like "bitbake -f virtual/kernel -c compile" Jan 23 18:57:38 Thanks. good hints. I am coming up on the OE curve so need to get my mind around the work flow. Jan 23 18:58:50 crinke, If you want to change something in the kernel konfiguration more than just touching the .config open a devshell. This is really neat I think. Jan 23 18:59:22 Let me try a bit more specific question. In the PPC architecture a .dtb file (DeviceTreeBinary) is usually compiled. Is the .dtb a part of the virtual/kernel recipe or a seperate recipe? Jan 23 19:00:02 ah, devshell? From that, would I do "make menuconfig" or "make" in the kernel tree? Jan 23 19:00:17 crinke, Try "bitbake virtual/kernel -c devshell". You should get a shell window (within the linux kernel dir) with all environment variable set up and can do things like "make menuconfig". Jan 23 19:01:28 Coolness, Andy. That starts getting me back to more familiar territory. Jan 23 19:01:39 crinke, OH your working with a PPC. So am I from time to time. Which PPC? Jan 23 19:02:20 I have been working with 8313 and recently a P2020 and now started a project with an 8323. Jan 23 19:03:01 My work to date has been with ltib, so I am coming up to speed on OE Jan 23 19:03:01 Oh, so you are the one posting in the users mailing list about glibc? Jan 23 19:03:26 yeah, Khem is very helpful. I am recompiling right now looking for the glibc issue Jan 23 19:03:39 Takes a "looooong time" to compile Jan 23 19:04:10 I'll try a build myself... Jan 23 19:04:13 So, my mind is trying to get ahead a bit and work through the issues I see coming up in the next week or so. Jan 23 19:04:41 would you be Andy as in "Andy Fleming"? Jan 23 19:05:14 Have you been using LTIB in earlier PPC projects and now switch to OE? Jan 23 19:05:30 Not Andy Fleming. Anders Törnqvist Jan 23 19:06:15 just guessing on that one. But good to meet you (virtually). Jan 23 19:07:47 Regarding the DTB build. From what I have seen so far it is managed in the kernel receipts. Jan 23 19:11:37 That would make sense. I have used the "dtc", DeviceTreeCompiler from a bash in an ltib shell in the past and it works fine, so I would assume a similar thing should be possible in OE Jan 23 19:12:42 In morphing u-boot and uImage to a new board, one finds a lot of time spent on register initialization, interrupt settings and memory maps and that leads to lots of interest in u-boot, linux kernel initialization and making sure the .dtb is sane. Jan 23 19:15:04 Andy_: no, bitbake has zero knowledge of the "same place in upstream tree". it has no idea what "upstream" is. also, you could do that yourself entirely in the metadata, by manipulating BBPATH based upon FILE. further, there's no guarantee all your layers have the same layout, bitbake doesn't require that. its up to you what you include for a reason Jan 23 19:15:59 cranke, I just fired up a build of mpc8223e-rdb with angstrom-2008.1 and minimal imag like you describe in the mailing list. I'll see if I get the same problem. Jan 23 19:16:51 Thanks Andy. BTW, I'm usually not "cranky", I usually try to be "ckrinke", as in "Charles Krinke" Jan 23 19:17:10 most of the time, anyway Jan 23 19:17:58 kergoth, Ok, I follow you. Sounds like it is a better idea then to be specific about the path in the "require" line if I want get the result I'm after. Jan 23 19:18:03 I'm at step 300 of 1400 on the compile. I havent added in the psyco.net compiler or multiple tasks in local.conf yet. Jan 23 19:18:27 crinke, sorrry for the misspelling. Jan 23 19:18:57 crinke, I thinkyou have to get a fater computer, Mine just passed line 552. :-) Jan 23 19:20:25 crinke, unfortunately we will probably not have the exact same result. I'm using release-2010.12 and the the main git. Jan 23 19:21:21 Jan 23 19:23:51 Is that bitbake release-2010.12 and OE main git? Jan 23 19:24:32 In the build tree for the moment it is bitbake 1.10.2 and OE release 2010.12 Jan 23 19:25:03 Ok, I've got bitbake 1.20.2 and I did a git pull on OE last night Jan 23 19:25:09 er, 1.10.2 Jan 23 19:26:17 Although, I am having an issue with "gettext" in the gcc-cross-initial task right now, even though I did an apt-get install gettext with my Debian installation Jan 23 19:33:34 crinke, What is the error message? I just sa a "Missing inherit gettext" message pass by. But it didn't stop. Jan 23 19:34:51 Hmm, mine stopped (or at least paused for a long time), but, .... I dont think the gettext is the issue I was seeing with glibc last night. I think the error is further along in the build. Although, if you dont get an error, that helps a great deal as it indicates a setup issue on my end and not a problem with OE. Jan 23 19:34:52 crinke, the message iwas for gcc-cross-initial-4.3.3 Jan 23 19:35:08 right, but that is not the error, that is a Jan 23 19:35:12 nag" Jan 23 19:35:24 hopefully you will *not* get an error Jan 23 19:35:33 this was added when bitkeeper was in use Jan 23 19:35:44 hmm Jan 23 19:36:08 bitkeeper? Jan 23 19:36:21 was meant for some other channel Jan 23 19:38:14 crinke, my build also paused for a while. gcc is a really BIG package to build. Even for a fast computer. Jan 23 19:38:38 crinke, Now I got a build error! Jan 23 19:38:52 glibc_2.9 fails to build Jan 23 19:59:26 Ok, Andy. Now the question becomes "why". The good news is you get a similar error. The bad news is there is an error. Jan 23 19:59:45 I wonder if other PPC MACHINE definitions might build? Jan 23 20:05:06 I restarted my compilation again to see if the newly installed gettext gives me no "nag" and am kerchunking my way past task 278 of 1411 but with a "bitbake minimal-image | tee log.txt" so I can see if I can intuit where it goes awry a bit better. Jan 23 20:15:40 03Khem Raj  07master * rcb23dffdbe 10openembedded.git/recipes/vdr/vdr_1.7.16.bb: Jan 23 20:15:40 vdr_1.7.16.bb: Specify the libc-uclibc override correctly Jan 23 20:15:40 Signed-off-by: Khem Raj Jan 23 20:16:27 Interesting. I rebooted my laptop and now went by the gcc-cross-initial with no nag about gettext. There is still hope that I can get through this compilation of mpc8323e-rdb with a sane glibc Jan 23 20:17:06 oops. Spoke too soon, still "missing inherit gettext", but the pause is short. Jan 23 20:59:10 03Klaus Kurzmann  07master * r072779ca8c 10openembedded.git/recipes/shr/ (2 files in 2 dirs): Jan 23 20:59:10 initscripts-shr: add script to populate /dev/cmt on nokia900 Jan 23 20:59:10 Signed-off-by: Klaus Kurzmann Jan 23 21:03:49 anyone here using gtk-sharp? Jan 23 21:04:06 or mono winforms for that matter Jan 23 21:45:12 crinke, Have you tried another version of glibc than the failing one for 8323? Jan 23 22:04:45 Dear Andy: "Another version of glibc" ? Can you help me understand how to do that? I am a tiny-bit glibc challenged as to how to change to another. Jan 23 22:25:57 he khem Jan 23 22:27:15 hm is chicago comming back now Jan 23 22:28:54 I am somewhat concerned that woglinde watch US sports Jan 23 22:29:03 touchdown Jan 23 22:29:06 crofton why? Jan 23 22:29:10 :) Jan 23 22:29:15 you know more than me Jan 23 22:29:24 I am watching biathlon too Jan 23 22:29:28 ok Jan 23 22:29:30 and other wintersports Jan 23 22:29:33 that is fun Jan 23 22:29:44 I would think more American would like biathlon Jan 23 22:29:50 because it involves guns Jan 23 22:30:03 but i think they are put off by the physical activity bit :0 Jan 23 22:30:11 hm there es one worldcup race at least Jan 23 22:30:14 in usa Jan 23 22:30:34 they should try it Jan 23 22:30:42 yeah Jan 23 22:30:45 it is interesting Jan 23 22:30:47 and tries to fire other the skiing Jan 23 22:31:22 crofton are you watching too? Jan 23 22:31:27 no Jan 23 22:31:29 at work Jan 23 22:31:38 Work bah Jan 23 22:31:43 not fun Jan 23 22:31:47 learning permenantly change screen resolution on over tide Jan 23 22:32:20 ka6sox? Jan 23 22:33:46 I need a better connection some day Jan 23 22:52:51 thats it Jan 23 22:57:16 wow Jan 23 23:18:51 good nite Jan 24 00:26:36 03Mike Westerhof  07org.openembedded.dev * r3f96355dd1 10openembedded.git/recipes/madwifi/madwifi-ng_r3878-20081204.bb: SlugOS: madwifi-ng - remove unnecessary pinning of an old version for SlugOS Jan 24 00:40:21 good nite, see you all tomorrow Jan 24 01:03:21 03Mike Westerhof  07org.openembedded.dev * rf90b785681 10openembedded.git/recipes/cryptsetup/cryptsetup_1.1.3.bb: Jan 24 01:03:21 Updates cryptsetup recipe to version 1.1.3, the most recent version that builds Jan 24 01:03:21 with the current version of automake in OE. Note that we need to keep the old Jan 24 01:03:21 1.0.5 recipe, even though it is not fetchable any longer, because the license Jan 24 01:03:21 has changed. Jan 24 01:03:21 Signed-off-by: Mike Westerhof Jan 24 01:03:22 Acked-by: Khem Raj Jan 24 01:49:36 03Khem Raj  07master * r4140085f8a 10openembedded.git/recipes/opensp/opensp_1.5.bb: (log message trimmed) Jan 24 01:49:36 opensp_1.5.bb: Fix compilation failure seen with -Os/gcc-4.5, Bugzilla/5494 Jan 24 01:49:36 This has been reported many times on mailing list and also in bugzilla/5494 Jan 24 01:49:36 The problem seems to be that gcc is not inlining the functions Jan 24 01:49:36 marked inline in class definitions and emits calls to them but then Jan 24 01:49:37 there is no definitions of these functions elsewhere. I dont know the Jan 24 01:49:38 real cause yet could be a gcc problem. Jan 24 02:36:38 03Khem Raj  07master * rffc1260a51 10openembedded.git/recipes/v8/ (4 files in 2 dirs): Jan 24 02:36:38 v8: Add recipes for v8/google's opensource JavaScript Engine Jan 24 02:36:38 * Fixed Bugzilla/5486 Jan 24 02:36:38 Signed-off-by: Khem Raj Jan 24 02:36:48 03Khem Raj  07master * re5274a6cce 10openembedded.git/conf/distro/ (2 files in 2 dirs): Jan 24 02:36:49 angstrom-2010.x.conf: Prefer binutils 2.21 and uclibc-git Jan 24 02:36:49 * no need to pin older udev with uclibc/git either Jan 24 02:36:49 Signed-off-by: Khem Raj **** ENDING LOGGING AT Mon Jan 24 02:59:58 2011