**** BEGIN LOGGING AT Sat Oct 30 02:59:57 2010 Oct 30 07:15:52 03Roman I Khimov  07org.openembedded.dev * r349668f166 10openembedded.git/recipes/httperf/httperf_0.9.0.bb: Oct 30 07:15:52 httperf: new recipe, version 0.9.0 Oct 30 07:15:52 Httperf is a tool for measuring web server performance. It provides a Oct 30 07:15:52 flexible facility for generating various HTTP workloads and for Oct 30 07:15:52 measuring server performance. Oct 30 07:16:00 paxtest: new recipe, version 0.9.9 Oct 30 07:16:00 PaX is a Linux kernel patch which adds much stricter control on how memory is Oct 30 07:16:00 being used by applications. A normal Linux kernel leaves the control to the Oct 30 07:16:00 application and does not implement any enforcement. Especially buffer overflow Oct 30 07:16:01 attacks benefit from the absence of kernel enforced memory control. PaX tries Oct 30 07:16:01 to do its best to enforce this control of memory used by applications, thereby Oct 30 07:16:02 03Roman I Khimov  07org.openembedded.dev * rf8579580b4 10openembedded.git/recipes/suricata/suricata_1.0.1.bb: Oct 30 07:16:56 03Vladimir Sorokin  07org.openembedded.dev * r79deba4e95 10openembedded.git/recipes/sshpass/ (sshpass.inc sshpass_1.04.bb): Oct 30 07:16:56 sshpass: initial commit. Version 1.04. Oct 30 07:16:56 Sshpass is a tool for non-interactivly performing password authentication with Oct 30 07:16:56 SSH's so called "interactive keyboard password authentication". Oct 30 07:16:56 Signed-off-by: Vladimir Sorokin Oct 30 08:06:05 gm Oct 30 08:12:50 gm Oct 30 11:38:24 How does bitbake choose the order which to extract two tar gzs in SRC_URI? Oct 30 11:39:19 I have two tar.gz's, one which is the normal source code and one that is in SRC_URI_append and which is supposed to extract over the previous one, replacing some icons Oct 30 11:39:50 bitbake always extracts the one in _append before the normal one Oct 30 11:43:14 How can I get bitbake to extract the second one after the first one? Oct 30 11:50:05 ho Oct 30 11:52:06 hrw, could you help me by answering me how bitbake chooses which archive to extract first when there are more than one in an SRC_URI? Oct 30 11:52:54 Um, I have two, one is the normal source code and the other is supposed to extract over it and contains some icons Oct 30 11:52:56 B_Lizzard: try _prepend instead ;) Oct 30 11:53:16 but order is strange.. Oct 30 11:53:32 should be _prepend, normal value then _append Oct 30 11:53:33 Shouldn Oct 30 11:53:39 't prepend mean before? Oct 30 11:53:44 And append mean after? Oct 30 11:53:48 :/ Oct 30 11:53:54 yeah, it should Oct 30 11:54:10 better to ask more bitbake-familiar people Oct 30 11:54:14 B_Lizzard: I would expect 'in order' Oct 30 11:56:56 The first archive is in SRC_URI, the second one is in SRC_URI_append but they still extract in reverse Oct 30 11:57:03 Lemme see with prepend Oct 30 11:57:31 B_Lizzard: iirc there was a possibility to tell to which dir you want it to unpack Oct 30 11:58:00 so unpack source, then addons to other dir and in extra task after unpack copy files Oct 30 11:58:07 thats best way Oct 30 11:58:50 I wouldn't want to litter the recipe with distro-specific crap, so I made the addons archive mirror the directory structure of the source archive Oct 30 11:59:02 I know your way makes order irrelevant Oct 30 12:00:12 It just seems strange that an archive that follows the source archive in every way (in the file, in the tasks order, in the alphabet) is still extracted first Oct 30 12:00:37 Does bitbake consider file size when extracting? Oct 30 12:00:38 :/ Oct 30 12:00:57 no diea Oct 30 12:01:10 Could this be considered a bug? Oct 30 12:01:17 also no idea Oct 30 12:01:22 its 7:56 here Oct 30 12:01:25 :D Oct 30 12:01:28 OK, sorry. Oct 30 12:01:41 my brain does not got warmed up yet Oct 30 12:01:50 I'll ask in a couple of hours :) Oct 30 12:01:52 base_do_unpack() looking sane Oct 30 12:02:21 for filename in glob(oe.path.join(srcdir, os.path.basename(urldata.path))): Oct 30 12:02:38 may be this.. Oct 30 12:02:46 B_Lizzard: do you using file:// url? Oct 30 12:02:51 Yeap Oct 30 12:03:01 Makes a difference? Oct 30 12:03:09 can you check with http:// Oct 30 12:03:14 just to check :) Oct 30 12:03:16 OK, lemme see. Oct 30 12:03:51 if urldata.type == "file" and "*" in urldata.path: Oct 30 12:04:13 hm.. that glob is only used when file:// used with * Oct 30 12:19:13 Does the same with http:// Oct 30 12:22:36 :/ Oct 30 12:30:36 03Martin Jansa  07master * ra0830a5e12 10openembedded.git/ (23 files in 3 dirs): Oct 30 12:30:37 xorg: new versions for 20101029 Oct 30 12:30:37 Signed-off-by: Martin Jansa Oct 30 12:30:47 03Martin Jansa  07master * rbb9bd7379e 10openembedded.git/recipes/freesmartphone/frameworkd_git.bb: frameworkd: bump SRCREV Oct 30 12:30:48 03Martin Jansa  07master * r4c1b1944d1 10openembedded.git/recipes/supertux/ (supertux-qvga_0.1.3.bb supertux_0.1.3.bb): Oct 30 12:30:48 supertux(-qvga): include supertux recipe to supertux-qvga Oct 30 12:30:48 * if do_configure in supertux-qvga fails for you, then it's because gp2x.patch adds libmikmod to configure.ac and Oct 30 12:30:48 libmikmod-config returns -L/usr/lib, libmikmod should be fixed Oct 30 12:30:49 Signed-off-by: Martin Jansa Oct 30 15:11:33 03Martin Jansa  07master * r931fb39f49 10openembedded.git/recipes/util-linux-ng/util-linux-ng.inc: Oct 30 15:11:33 util-linux-ng: Make cal, script, flock use u-a, drop D_P = -1 Oct 30 15:11:33 * cal and script were added to usrbinprogs_a in af3f52db17181108a45a0484a57e84495e865393, Oct 30 15:11:33 but u-a calls weren't added to postinst/prerm Oct 30 15:11:33 * drop D_P = -1, doesn't make much sense in .inc file Oct 30 15:11:41 03Martin Jansa  07master * rb3094bc2b4 10openembedded.git/recipes/coreutils/coreutils-target.inc: Oct 30 15:11:41 coreutils: install df and nice to /bin as busybox does Oct 30 15:11:42 * otherwise u-a will fail Oct 30 15:11:42 Signed-off-by: Martin Jansa Oct 30 15:11:44 03Martin Jansa  07master * r05794e79b4 10openembedded.git/recipes/coreutils/coreutils-target.inc: Oct 30 15:11:44 coreutils: use u-a for sha256sum, sha512sum, timeout (conflict with busybox-1.17.3) Oct 30 15:11:44 * surround "u-a --install" and "mv" calls with "if" to be sure it won't fail for old coreutils version without some ${bindir}/$i Oct 30 15:12:57 03Martin Jansa  07master * r090669b2ff 10openembedded.git/conf/distro/include/preferred-shr-versions.inc: Oct 30 15:12:57 SHR: prefer latest busybox Oct 30 15:12:57 Signed-off-by: Martin Jansa Oct 30 16:42:20 * Jay7 is started first attempt to build OE under FreeBSD Oct 30 16:43:45 bad thing is that my current qemu-kvm can't use more than one cpu Oct 30 16:43:52 I don't know why atm Oct 30 16:44:05 so.. it will take a lot of time Oct 30 16:51:00 and there is first bug - bitbake wish to have md5sum tool Oct 30 16:53:10 we should emulate this around bundled md5 tool or make port of real md5sum Oct 30 16:54:51 ah.. there is sysutils/coreutils port with gmd5sum Oct 30 17:20:02 03Sebastian Krzyszkowiak  07master * rf24863c45f 10openembedded.git/recipes/xserver-nodm-init/xserver-nodm-init_2.0.bb: Oct 30 17:20:02 xserver-nodm-init: update INITSCRIPT_PARAM to stop it in runlevel 0 Oct 30 17:20:02 Signed-off-by: Martin Jansa Oct 30 17:20:12 03Sebastian Krzyszkowiak  07master * ra505f567b1 10openembedded.git/recipes/mokomaze/ (2 files in 2 dirs): Oct 30 17:20:12 mokomaze: fix accelerometers on gta02 with 2.6.34 kernel Oct 30 17:20:12 Signed-off-by: Martin Jansa Oct 30 17:23:17 03Martin Jansa  07master * r658b9a4ff2 10openembedded.git/ (30 files in 7 dirs): Oct 30 17:23:17 xorg: new versions for 20101030 Oct 30 17:23:17 Signed-off-by: Martin Jansa Oct 30 19:34:34 well.. bitbake 1.10 run on FreeBSD producing following errors Oct 30 19:34:37 http://pastebin.ca/1977499 Oct 30 19:34:48 on every task seems Oct 30 19:35:42 we need compatibility layer Oct 30 19:36:20 bsd tar have no --owner option Oct 30 19:56:21 seems openembedded built me an ufifs image Oct 30 19:56:37 are these suitable for writing directly on an mtd device with nandwrite? Oct 30 19:57:45 how big is a full OE (gui) build if I do not want to erase the work files? Oct 30 20:04:45 ffthe documentation says about 10G without rm_work but I seem to remember it being quite a bit bigger Oct 30 20:08:29 03Martin Jansa  07master * rf0629ca6b0 10openembedded.git/recipes/tasks/task-shr-feed.bb: Oct 30 20:08:29 task-shr-feed: add efm recipes Oct 30 20:08:29 Signed-off-by: Martin Jansa Oct 30 20:08:59 mine is 25 GB Oct 30 20:09:10 but I only built minimal-image so far Oct 30 20:10:25 yeah that's kind of what I thought... that 10GB size is out of date :-) Oct 30 20:15:04 heh... now quilt-native fails on sed checking Oct 30 20:15:27 well.. I'll give up on this for now Oct 30 20:25:18 huge Oct 30 20:25:20 let me chuck Oct 30 20:25:25 *check Oct 30 20:25:27 what distro? Oct 30 20:26:36 i want to say 130-150gb Oct 30 20:26:47 but let me du -shc for sure. it takes a while if you're not on an ssd Oct 30 20:36:22 a scant 96GB Oct 30 20:53:06 so it looks like the demo image actually shipped with the beagleboard xm isn't available in the OE repos? Oct 30 20:53:35 i found a patch from jason kridner http://patchwork.openembedded.org/patch/2737/ that looks like it might be what i want Oct 30 20:54:25 i can't get a desktop environment working with what's in the repos, i get GPE to come up with some kind of on-screen keyboard with a bunch of tiny text boxes above it, but all the text is illegible :P Oct 30 20:54:41 i thought it might be something i did to my tree so i blew it all away and started over, but same thing happened :( Oct 30 21:08:22 mrj10: well its lot of debug info that contributes to size Oct 30 21:08:44 mrj10: you can add INHERIT += "rm_work" to your local.conf if size is bothering you Oct 30 21:10:23 tzanger asked how big it was if you didn't want to remove work files. does INHERIT += "rm_work" not remove work files? Oct 30 21:10:35 yes it does Oct 30 21:10:44 you can also change the level of debugging Oct 30 21:10:54 right now it uses -g3 Oct 30 21:11:01 you can change that to -g Oct 30 21:11:20 yeah, the downloads are only 5-6gb for me, so i'd imagine you can get under 10gb if you rm_work Oct 30 21:11:35 debug output out of bitbake? Oct 30 21:15:42 angstrom distro Oct 30 21:18:02 03Martin Jansa  07master * r7634d114f0 10openembedded.git/recipes/busybox/ (busybox.inc busybox_1.17.1.bb busybox_1.17.3.bb): Oct 30 21:18:03 busybox: use INC_PR in consistent way Oct 30 21:18:03 Signed-off-by: Martin Jansa Oct 30 21:18:04 03Martin Jansa  07master * re361253742 10openembedded.git/recipes/busybox/ (6 files): Oct 30 21:18:04 busybox: don't require busybox.inc twice Oct 30 21:18:04 * busybox_1.1x.inc includes it too resulting in Oct 30 21:18:04 ERROR: busybox-linuxrc is listed in PACKAGES multiple times, this leads to packaging errors. Oct 30 21:19:22 sorry for being a bit dense khem, what tool takes -g3 as an argument? Oct 30 21:20:42 holy cow Oct 30 21:20:45 i didn't know gcc did that! Oct 30 21:20:51 * mrj10 is ashamed Oct 30 21:21:49 ok, so g3 includes all your macro definitions and things like that on top of -g2 (which == -g) Oct 30 21:29:58 good evening. Oct 30 21:31:33 does anyone in here happen to know if dream multimedia still works on porting a recent kernel to dm800 and that generation of devices... Oct 30 21:31:58 or anyone else, for that matter Oct 30 21:52:59 mrj10: perfect, thanks guys :-) Oct 30 21:54:10 np Oct 30 22:56:26 03Khem Raj  07master * r279419af0e 10openembedded.git/recipes/gcc/ (gcc-4.5.inc gcc-package-cross.inc gcc-package-target.inc): Oct 30 22:56:26 gcc-package-cross, gcc-package-target: Stage libgcc_s.so linker stub Oct 30 22:56:26 * This stub is not staged in /usr/lib which is Oct 30 22:56:26 a problem for many python recipes as it looks for libgcc in Oct 30 22:56:26 that directory and fail to find it. Oct 30 23:25:56 03Martin Jansa  07master * ref297b9feb 10openembedded.git/recipes/accelges/accelges_svn.bb: Oct 30 23:25:56 accelges: don't put libraries to RDEPENDS Oct 30 23:25:56 * they are automatically added by .so providers Oct 30 23:25:56 * in this case libcurl was pulling curl-nativesdk to DEPENDS which is really wrong Oct 30 23:25:56 Signed-off-by: Martin Jansa Oct 31 00:23:15 aloha Oct 31 00:26:00 morning Oct 31 00:32:21 * hrw stuck at MCO Oct 31 00:33:04 Thunderstorms? Oct 31 00:33:26 or did Mickey Mouse cause a backup? Oct 31 00:34:39 no idea Oct 31 00:34:51 boarding had to start 25 minuts ago Oct 31 00:35:02 whoops... Oct 31 00:35:04 so I think one hour delay at least Oct 31 00:35:24 wow...hopefully you don't have another connection to make. Oct 31 00:35:45 I have Oct 31 00:36:08 MCO -> FRA -> TXL and then 2.5h bus home Oct 31 00:37:50 which means sometime tommorrow morning Oct 31 00:37:59 yes Oct 31 00:38:14 I hate redeyes Oct 31 00:38:20 11:30 at fra, then iirc 14:00 to txl and 17:25 bus Oct 31 00:38:42 ok, changed time.. now plane at 21:40 Oct 31 00:38:56 so an hour... **** ENDING LOGGING AT Sun Oct 31 02:59:58 2010