**** BEGIN LOGGING AT Sun Aug 29 02:59:57 2010 Aug 29 03:10:23 How do I define a preferred provider for virtual/bootloader? "bitbake virtual/bootloader" keeps building the wrong package and says I should define PREFERRED_PROVIDER_virtual/bootloader. Aug 29 03:10:57 My machine's .conf file has the line 'PREFERRED_PROVIDER_virtual/bootloader = "u-boot-chumby-silvermoon"', and I can build u-boot-chumby-silvermoon without a problem. Aug 29 03:14:07 xobs, look how other machines do it Aug 29 03:16:28 GNUtoo: Most seem to use the same syntax. But for some reason it's not getting picked up in my conf. Aug 29 03:26:03 Running bitbake -D -D -D -v doesn't seem to help, either. I'm guessing the variable is getting overwritten or lost somewhere. Any idea on how I can figure out where? Aug 29 03:31:05 Okay, that's embarrassing. Running "getvar PREFERRED_PROVIDER_virtual/bootloader" shows that it's defined, just defined wrong. Which is caused by having the wrong MACHINE set. Aug 29 04:03:21 xobs: for future reference, there's also bitbake -e Aug 29 04:13:13 kergoth: Perfect! That looks very handy. Aug 29 04:19:40 I love it when things come together. Aug 29 04:19:46 And devshell is so very handy. Aug 29 05:00:20 hmmm Aug 29 05:00:32 hi khem Aug 29 05:01:20 I've a big issue with immodule Aug 29 05:01:34 basically it's a dynamic packages things Aug 29 05:01:50 and so I know that immodule is present only after beeing dynamically-packaged Aug 29 05:01:58 and so the code after that has no effect Aug 29 05:02:10 it added the file to FILES_gtk-immodule-xim Aug 29 05:02:22 and was putting it as machine arhc Aug 29 05:02:24 *arch Aug 29 05:02:28 but it's already too late Aug 29 05:02:30 no? Aug 29 07:00:02 gm Aug 29 07:00:40 lol, hope i didn't scare you away XorA|gone :-) Aug 29 07:01:13 eFfeM: nah, just forgot to set nick last night Aug 29 07:02:10 is there an easy way to override this in a makefile: Aug 29 07:02:12 KREV := $(shell echo $(KVER) | \ Aug 29 07:02:12 sed -e 's/^[0-9][0-9]*\.[0-9][0-9]*\.\([0-9][0-9]*\).*/\1/') Aug 29 07:02:39 aims to get the last number of the rev takes the host string Aug 29 07:02:53 kver is: Aug 29 07:02:57 KVER := $(shell $(CC) $(CFLAGS) -E -dM $(VERSION_FILE) | \ Aug 29 07:02:58 grep UTS_RELEASE | awk '{ print $$3 }' | sed 's/\"//g') Aug 29 07:03:08 and defining KVER on the cmd line didn't help Aug 29 07:03:26 anybody an idea on this (or is it still too early :-) ) Aug 29 08:55:19 kergoth: thanks for your words in the dependency checking thread Aug 29 10:24:18 03Michael 'Mickey' Lauer  07org.openembedded.dev * r4123611d60 10openembedded.git/recipes/vala/vala_0.9.6.1.bb: Aug 29 10:24:19 vala: add 0.9.6.1 Aug 29 10:24:19 NOTE: Unfortunately upstream decided to change the PC version, so this Aug 29 10:24:19 and later versions claim to be vala-0.10 instead of vala-1.0. Please Aug 29 10:24:19 adjust your configure.ac, if necessary. To ease the transition, we Aug 29 10:24:19 install vala-1.0.pc as well in the meantime. Aug 29 10:40:49 03Martin Jansa  07org.openembedded.dev * r54ee43bfb6 10openembedded.git/recipes/tasks/task-shr-feed.bb: task-shr-feed: add gnuplot Aug 29 11:18:42 03Martin Jansa  07org.openembedded.dev * r768611b6c8 10openembedded.git/recipes/tasks/task-shr-feed.bb: Aug 29 11:18:43 task-shr-feed: add galculator Aug 29 11:18:43 Signed-off-by: Martin Jansa Aug 29 11:18:53 03Martin Jansa  07org.openembedded.dev * r533c749c16 10openembedded.git/recipes/galculator/galculator_1.3.4.bb: Aug 29 11:18:53 galculator: add version 1.3.4 Aug 29 11:18:53 Signed-off-by: Martin Jansa Aug 29 11:47:44 (Why) Is the GIT server dead slow? Anyone has access to look into this? Aug 29 11:47:59 It is in Mongolia Aug 29 11:48:05 ~20 to 30 kB/s Aug 29 11:48:22 Crofton_|work: you kidding us? Aug 29 11:48:30 yes Aug 29 11:48:58 that makes it worse then :-) Aug 29 11:49:25 who has access to our servers anyway? Aug 29 11:50:11 The OE infrastructure is totally unknown to me. I thought we once went to the nslu2 server park, or was this cancelled? Aug 29 11:58:29 the git server is there Aug 29 11:58:43 I don't have access Aug 29 11:59:12 oe.org reads: "We currently have a need for funds to upgrade server hardware at our OSUOSL hosting location." Aug 29 11:59:41 yeah Aug 29 11:59:49 they had to buy some stuff recently Aug 29 12:01:35 I offered last year to sponsor for a dedicated server like these ones : http://www.kimsufi.com/ks/ (the one @ 49.99/month seems great) if there is interest this could be done this year Aug 29 12:05:46 still fighting with ncurses : now sdk build fine but it seems the split in several packages (lib, terminfo & co) breaks some dependencies : example : launching alsamiser returns : "Error opening terminal: linux." because ncurses-terminfo is not installed (only libncurses5 and libncursesw). What is the best way to solve this ? Aug 29 12:06:56 ericben: never worked with the sdk... I should try that one day. Aug 29 12:07:34 hi, Aug 29 12:07:40 should I add RDEPENDS_alsa-utils-alsamixer = "ncurses-terminfo" to alsa-utils' bb ? Aug 29 12:07:50 the gtk-immodule-xim is created by the gtk+ recipe Aug 29 12:08:31 it uses package split Aug 29 12:08:47 and package-dynamic Aug 29 12:09:07 at package split time you know that the immodule-xim module has been created Aug 29 12:09:46 I worte some python code that altered the pakcages's FILES and ARCH but I made it to be executed just after splitting the packages Aug 29 12:09:53 and it didn't wor Aug 29 12:09:55 *work Aug 29 12:10:04 what should I do? Aug 29 12:28:02 ericben, that server would be nice for testing purposes Aug 29 12:28:48 likewise: there seem to be some strange network things. If I am correct the angstrom builder is also hosted there, but it is impossible to even get a small image from it. Aug 29 12:29:16 ka6sox4: do you have an idea on the slowness? are you involved in the operational things? Aug 29 12:31:40 eFfeM, I suspect ka6sox-work is still in bed Aug 29 12:36:19 me too, but he'll see the Q when he shows up Aug 29 12:36:38 there seem to be some kind of issue Aug 29 12:36:47 hi eFfeM maybe I should look at the 2 msm orphaned patches Aug 29 12:47:30 bbl, bye all Aug 29 13:14:21 03Koen Kooi  07org.openembedded.dev * r26b7d65141 10openembedded.git/recipes/openssl/openssl.inc: openssl: move cnf file to -misc making different libcrypto versions parallel installable Aug 29 13:21:55 bye likewise Aug 29 13:22:11 GNUtoo|laptop: great, be my gest Aug 29 13:23:38 guest Aug 29 13:26:13 gone too, cya Aug 29 13:27:01 hmmm Aug 29 13:27:13 maybe we should delete linux-replicant Aug 29 13:27:14 ? Aug 29 13:29:27 but there are other phones there Aug 29 13:47:25 hmmm Aug 29 13:47:30 is there someone here? Aug 29 13:47:52 who knows package split and its implications of some FILES_ += executed after it Aug 29 13:48:02 same for ARCH=${MACHINE_ARCH} Aug 29 14:02:12 hmmm Aug 29 14:02:18 maybe I should do something else then Aug 29 14:28:43 last Sunday in August, even Christmas isn't this dead :) Aug 29 15:16:43 hi ericben Aug 29 17:01:46 eFfeM_work, you here? Aug 29 17:21:45 ka6sox-work, people have been saying git is slow Aug 29 17:49:40 Crofton_|work, okay I'll look @ it... Aug 29 17:52:04 can someone tell me whether its the git portion or CIA's reporting? Aug 29 17:52:16 the server is bored...and quite responsive. Aug 29 17:52:38 I think git pulls to .eu Aug 29 17:53:52 oh...I"m in .us so suggests pipes to .eu Aug 29 18:35:24 03Martin Jansa  07org.openembedded.dev * rfae1bd6463 10openembedded.git/recipes/galculator/galculator_1.3.4.bb: Aug 29 18:35:24 galculator: add link galculator_48x48.png, otherwise it segfaults Aug 29 18:35:24 Signed-off-by: Martin Jansa Aug 29 18:58:06 Hej! I'm trying to convert htmldoc to BBCLASSEXTEND = "native" but this always creates dependency loops. And I find no example how to use it's on -native recipe from inside a recipe. Aug 29 18:58:56 htmldoc_x.y.bb contains DEPENDS += "htmldoc-native" and BBCLASSEXTEND = "native". This fails Aug 29 18:59:25 Is there a DEPENDS_htmldoc-nonative += "htmldoc-native" syntax? Aug 29 19:03:15 Silly me, it's DEPENDS_htmldoc += "htmldoc-native" Aug 29 19:57:09 03Enrico Scholz  07org.openembedded.dev * rc73d107c2f 10openembedded.git/recipes/ncurses/ncurses_5.7.bb: Aug 29 19:57:09 ncurses: set --with-ticlib for widec libs too Aug 29 19:57:09 It was used for narrow lib and there are no reasons why it should not be Aug 29 19:57:09 set for widec tool. Noticed by Eric Benard. Aug 29 19:57:09 Signed-off-by: Enrico Scholz Aug 29 19:57:17 03Enrico Scholz  07org.openembedded.dev * rcaa254da31 10openembedded.git/recipes/ncurses/ncurses_5.7.bb: Aug 29 19:57:17 ncurses: use ${PN} instead of hardcoded 'ncurses' in package names Aug 29 19:57:17 Thanks to Eric Benard for noticing this. Aug 29 19:57:17 Signed-off-by: Enrico Scholz Aug 29 19:57:18 03Enrico Scholz  07org.openembedded.dev * r6f81dd32a1 10openembedded.git/recipes/ncurses/ncurses_5.7.bb: Aug 29 19:57:20 ncurses: fixed widec builds for native targets Aug 29 19:57:21 -D_GNU_SOURCE was not propagated when building native helper tools and Aug 29 19:57:21 must be added to CPPFLAGS hence. Aug 29 19:57:21 By fixing this, widec build can be enabled for native and sdk. Aug 29 19:57:21 Signed-off-by: Enrico Scholz Aug 29 19:57:21 03Enrico Scholz  07org.openembedded.dev * r39294e54ba 10openembedded.git/recipes/ncurses/ (ncurses-5.7/config.cache ncurses_5.7.bb): Aug 29 19:57:22 ncurses: set some configure cache variables Aug 29 19:57:22 Corresponding tests returned either wrong results or were generally broken. Test Aug 29 19:57:23 for GPM was not prepared for crosscompiling and GPM support has been disabled hence. Aug 29 19:57:23 Signed-off-by: Enrico Scholz Aug 29 19:57:23 03Enrico Scholz  07org.openembedded.dev * r447e7396ad 10openembedded.git/recipes/ncurses/ncurses_5.7.bb: Aug 29 19:58:01 ncurses: preserve default PACKAGES list by using +=, not = Aug 29 19:58:01 Signed-off-by: Enrico Scholz Aug 29 19:58:01 03Enrico Scholz  07org.openembedded.dev * r3aa1ef6caf 10openembedded.git/recipes/ncurses/ncurses_5.7.bb: Aug 29 19:58:01 ncurses: bumped PR Aug 29 19:58:02 Signed-off-by: Enrico Scholz Aug 29 19:58:02 03Enrico Scholz  07org.openembedded.dev * rce59023d4e 10openembedded.git/recipes/ncurses/ncurses_5.7.bb: Aug 29 19:58:02 ncurses: splitted common terminal definitions in own package and depend on it Aug 29 19:58:03 Moved /etc/terminfo from base 'ncurses' packages into -terminfo-base. Aug 29 19:58:04 Recommend this package in -libtinfo and suggest the large -terminfo Aug 29 19:58:04 package. Aug 29 19:58:05 Signed-off-by: Enrico Scholz Aug 29 19:58:05 03Enrico Scholz  07org.openembedded.dev * rd6ee8215a0 10openembedded.git/recipes/ncurses/ncurses_5.7.bb: Aug 29 19:59:01 ncurses: made -tools depend on base package Aug 29 19:59:01 There is a symlink in -tools to a file in the base package. Aug 29 19:59:01 Signed-off-by: Enrico Scholz Aug 29 20:01:38 hi all Aug 29 20:18:33 eFfeM, ka6sox-work thinks the git server should be ok Aug 29 20:19:11 Crofton_|work: ah ok, thanks for the info Aug 29 20:20:04 actually I did not really notice git issues, it was likewise who thought things were ok. i noticed poor performance from the angstrom server and actually I have no idea if that is the same system Aug 29 20:20:44 here (Bordeaux, south west of France), I have ~200-500KiB/s on the git sever Aug 29 20:21:33 cool ncurses-sdk seems fixed Aug 29 20:21:53 they aren't Aug 29 20:22:00 I am pretty sure angstrom is on ltg Aug 29 20:22:05 ericben, hi Aug 29 20:22:19 hi GNUtoo|laptop Aug 29 20:22:33 btw is it possible to do that: Aug 29 20:22:45 FILES_gtk-immodule-xim += "somthing" Aug 29 20:22:53 when the packages are already splited Aug 29 20:23:17 personally i've or 3g which is limited to 2GB Aug 29 20:23:52 or a vpn over open wifi with some port blockeds like port 80,443,6667 etc... Aug 29 20:23:57 vpn is 50k max Aug 29 20:24:54 ericben, I'll PM you Aug 29 20:24:55 ah ok, well i pull regularly so I am not too keen on the performance of the git server, but I needed a quick image the other day and pulled it from the angstrom autobuilder, and it stalled after a while, had to restart it a few times before i got the image, if I recall correctly koen stated on #beagle that this was a known isuue Aug 29 20:24:58 issue Aug 29 20:25:11 i was not sure if we were on the same server Aug 29 20:25:40 maybe if many peoples are using the image builder the bandwidht & the server gets loaded Aug 29 20:26:38 i then would expect the bandwidth to lower, now it seems like the connection is completely lost Aug 29 20:27:50 anyway, wrt the angstrom builder, i only use it rarely, but frankly i have no idea what is where, so if it were the same server as git, that would explain things, but apparently it isn't so that does not explain what likewise saw Aug 29 20:28:36 hmm, if the image buyilder is getting hammered I'll have to talk with koen and see what optios we have Aug 29 20:28:44 that is an important service Aug 29 20:31:06 koen said something about it when i asked about it, guess it was in #beagle, forgot when we discussed it Aug 29 20:31:51 but noticed the download issue again about a week ago when trying to download an image for eeepc, had to restart the download several times (firefox) Aug 29 21:50:58 hm... Aug 29 21:51:03 /home/users/builder/openembedded/recipes/dialog/dialog.inc:DEPENDS = "ncurses" Aug 29 21:51:14 | configure: error: no curses library found Aug 29 21:51:14 NOTE: package dialog-1.1-20080819-r1.0: task do_configure: Failed Aug 29 21:51:16 wtf... Aug 29 21:51:37 | checking for ncurses version... 5.7 Aug 29 21:51:39 lol Aug 29 22:08:38 /lib/libgcc_s.so.1: file not recognized: File format not recognized Aug 29 22:08:43 um... Aug 29 22:13:20 should -L/lib be present in arm-oe-linux-gnueabi-gcc's flags when crosscompiling? Aug 29 22:40:51 03Florian Boor  07org.openembedded.dev * r9cd92740e6 10openembedded.git/conf/machine/smartqv7.conf: smartqv7.conf: Device description for Smartdevices V7 Aug 29 22:40:53 03Florian Boor  07org.openembedded.dev * r274a3e5941 10openembedded.git/recipes/xserver-common/xserver-common_1.34.bb: Aug 29 22:40:53 xserver-common: Add new release with several bugfixes and merged patches from OE. Aug 29 22:40:53 tested with Angstrom 2008.1 Aug 29 22:40:54 03Florian Boor  07org.openembedded.dev * r4824de3bbf 10openembedded.git/recipes/linux/ (3 files in 2 dirs): linux: 2.6.28 patch and workarounds for smartqv7. Aug 29 22:46:42 03Florian Boor  07org.openembedded.dev * rbf8ba265bb 10openembedded.git/recipes/psplash/ (files/smartqv7/psplash-default psplash.inc): psplash: Add configuration for smartqv7 Aug 29 22:46:51 03Florian Boor  07org.openembedded.dev * r8b071c0fa5 10openembedded.git/recipes/tslib/ (tslib.inc tslib/smartqv7/tslib.sh tslib_1.0.bb): tslib: Configuration for smartqv7 Aug 29 23:46:26 03Stanislav Brabec  07org.openembedded.dev * r0e59321a3f 10openembedded.git/recipes/udev/ (4 files in 2 dirs): udev: udev-compat-wrapper-patch: Use KERNELMICROVER evaluation that does not fail on rc and four-part versions. Aug 29 23:51:01 03Stanislav Brabec  07org.openembedded.dev * r49da1aba1a 10openembedded.git/recipes/crda/crda_1.1.1.bb: crda: Skip verification of regulatory.bin of the build host. Aug 29 23:59:36 03Stanislav Brabec  07org.openembedded.dev * r8100653c22 10openembedded.git/recipes/ksymoops/ (3 files in 3 dirs): ksymoops: Updated to version 2.4.11, fixed for current libbfd, adjusted /proc/kallsyms. Aug 30 01:22:14 who created this branch remotes/origin/oe-write/org.openembedded.dev ? **** ENDING LOGGING AT Mon Aug 30 02:59:57 2010