**** BEGIN LOGGING AT Mon Feb 11 02:59:58 2013 Feb 11 08:51:44 morning! Feb 11 09:07:45 morning Feb 11 09:11:34 night Feb 11 09:38:04 Hi all Feb 11 09:47:43 all: Nine years ago I bought Sharp Zaurus SL-5500 as my first Linux PDA. And due to this I am where I am. Feb 11 09:47:46 http://marcin.juszkiewicz.com.pl/2013/02/11/nine-years-of-embedded-linux/ Feb 11 09:47:48 ;D Feb 11 09:49:20 Happy birth-of-a-vocation day ! Feb 11 09:49:43 thanks Feb 11 09:51:29 good morning Feb 11 09:52:01 hi florian Feb 11 09:52:54 hrw: Nice blog entry... I should take a look back as well :) Feb 11 09:53:18 florian: I was planning to write that for over month Feb 11 09:53:34 and finally sat today and wrote 90% of it. Feb 11 09:53:58 florian: note lack of links inside of entry ;) Feb 11 09:54:36 hrw: heh Feb 11 09:59:52 hrw: at least this one, for the youngest : http://en.wikipedia.org/wiki/Sharp_Zaurus Feb 11 10:00:26 ascor: normally I link to my previous posts Feb 11 10:13:47 morning all Feb 11 10:15:03 hi bluelightning Feb 11 10:27:38 morning all, quick question Feb 11 10:28:04 is it right that VIRTUAL-RUNTIME variables are only taken into account when used in the distro.conf file? Feb 11 10:28:17 and that they cannot be used to set a provider in image recipes Feb 11 11:13:23 jackmitchell: that is correct Feb 11 11:13:31 hey florian Feb 11 11:28:43 I am trying to use LTO with codesourery (sourcery code bench actually) but I am getting this `_kill' referenced in section `.text._kill_r' of /opt/arm-2012.09/bin/../lib/gcc/arm-none-eabi/4.7.2/../../../../arm-none-eabi/lib/thumb2/libc.a(lib_a-signalr.o): defined in discarded section `.text' of common.c.obj (symbol from plugin) Feb 11 11:29:19 any idea why is that? I am using function sections and data sections .. but I have disabled gc sections for this test and it does not help. Feb 11 12:40:29 what's the trick to be able to update busybox with opkg ? when doing an update it first delete busybox and all the links, then opkg is unable to use basic utils like wget and fails at updating busybox Feb 11 13:06:59 I have libsqlite3 in my image.bb, now when I do -c populate_sdk, sqlite3.h doesn't make it into the sysroot Feb 11 13:07:33 do I have to specify libsqlite3-dev in my image, I would have though -c populate_sdk would automatically resolve to -dev packages.... Feb 11 13:08:31 jackmitchell: IMAGE_FEATURES += "dev-pkgs staticdev-pkgs" needs to be in image Feb 11 13:08:52 hrw: thanks! Feb 11 13:11:14 hrw: btw, I enjoyed your blog post by the way, similar to the career path I hope to follow ;) Feb 11 13:11:22 ;) Feb 11 13:36:25 Despite I removed bluetooth from DISTRO_FEATURES and IMAGE_FEATURES (confirmed with bitbake -e), it is still enabled in pulseaudio. Is it possible that when pulseaudio recipe is parsed, the _FEATURE variable do not have bluetooth ? Feb 11 13:37:10 oups, please read "the _FEATURE still have bluetooth". Feb 11 13:38:07 ascor: pulseaudio.inc has the following line Feb 11 13:38:08 ${@base_contains('DISTRO_FEATURES', 'bluetooth', '--enable-bluez', '--disable-bluez', d)} \ Feb 11 13:38:18 can you check to see if yours is the same Feb 11 13:38:40 what version of oe-core are you running? Feb 11 13:39:09 my oe-core is master, HEAD from the git Feb 11 13:39:27 I do have this line in pulseaudio.inc Feb 11 13:39:29 it should be the same then Feb 11 13:40:07 But I also have this line in the recipe i build : DISTRO_FEATURES := "${@oe_filter_out('bluetooth', bb.data.getVar('DISTRO_FEATURES', d, 1), d)}" Feb 11 13:40:11 I don't know why it is still enabled, as it is explicity disabled in the recipe, maybe a configure bug upstream Feb 11 13:40:34 I think it does the same thing in danny Feb 11 13:40:45 hmm, you are running latest master Feb 11 13:41:13 as in master, master, not danny, master Feb 11 13:42:47 danny is a tag in the master branch or a branch itself ? Feb 11 13:43:04 it's a branch Feb 11 13:43:33 Ok, so I'll check again with danny. Feb 11 13:43:57 no, I was saying that danny may not have the bluetooth changes Feb 11 13:43:58 However, don't you think it's strange ? Feb 11 13:44:06 https://github.com/openembedded/oe-core/blob/master/meta/recipes-multimedia/pulseaudio/pulseaudio.inc Feb 11 13:44:11 Ah .. Feb 11 13:44:12 that is current pulseaudio.inc Feb 11 13:44:34 and it doesn't have your oe_filter statement, so you may be out of sync somewhere Feb 11 13:45:31 if the recipe as it is in master, still builds with bluetooth support - I would say that it is a bug upstream Feb 11 13:45:42 oe_filter works, because it disappears from bitbake -e output Feb 11 13:46:05 https://github.com/openembedded/oe-core/blob/danny/meta/recipes-multimedia/pulseaudio/pulseaudio.inc is the same as in master branch Feb 11 13:46:28 ok, that means that no big changes have happened to put you out of sync, or break the support Feb 11 13:46:52 without the oe_filter statement do you still have bluetooth in your distro_features Feb 11 13:47:01 yes Feb 11 13:47:05 if so, I would find out what is causing that to get in there Feb 11 13:47:21 my filterout filters everything that I tell, except pulse audio. Feb 11 13:48:27 that's strange, i'm not sure how good an idea filter_out is though - I'm no bitbake expert though, so it may be ok Feb 11 13:48:45 why do you use filter and not just remove bluetooth from distro features in the distro conf? Feb 11 13:48:49 That reminds me something hrw told me about pulse audio ... it's included though a special mechaDISTRO_FEATURES_BACKFILnism : Feb 11 13:49:13 exactallly, when you start backfilling and doing odd things like that, it can get messy Feb 11 13:49:48 so pulseaudio may backfill, before oe_filter, filters.. does that make sense? Feb 11 13:50:14 because my distro uses the defaultsetup So I have to modify oe .conf file. I thing it will be no good Feb 11 13:50:43 ascor: you can set up your own distro conf file which sets DISTRO_FEATURES Feb 11 13:51:02 That makes sense to me. I think I'll ask hrw, he seems to understand well the backfilling mechanism Feb 11 13:51:06 hmm, there are two other ways, one is have another copy of DISTRO_FEATURES, or keep a pulseaudio.inc patch which forces bluetooth off Feb 11 13:52:42 bluelightning : I think that I will do that, but I am a the starting point of my learning curve, so I would also like to understant why this does not work. Feb 11 13:53:15 patching is a little bit dirty, no ? Feb 11 13:53:20 ascor: check bitbake -e pulseaudio | grep ^EXTRA_OECONF Feb 11 13:53:48 then you can determine whether the problem lies with OE or pulseaudio's configure step Feb 11 13:54:04 patching is a bit dirty, but trying to manipulate the default config to a custom one is also a bit dirty Feb 11 13:54:16 output is : --enable-bluez Feb 11 13:54:40 well the output contains --enable-bluez Feb 11 13:54:53 and it shouldn't Feb 11 13:55:14 I solved a lot of problems for myself when I created my own distro configuration, which is pretty much a copy of the oe-core distro with a couple of things stripped Feb 11 13:55:55 ascor: ok, then the oe_filter_out is not working Feb 11 13:55:57 ok, so I have IMAGE_FEATURES = "dev-pkgs staticdev-pkgs" and I also have libsqlite3 in my image.bb but when I do -c populate_sdk sqlite.h never makes it to my sysroot Feb 11 13:55:59 Frustrating... Feb 11 13:56:11 all the other library include files make it into the sysroot Feb 11 13:56:36 bluelightning : it does function for wifi, IPV6, pci ... There is something with pulseaudio. Feb 11 13:56:45 however, sqlite3 is the only recipe where the -dev is not named the same as the package name Feb 11 13:57:22 so is it possible that the parser isn't clever enough to notice that sqlite3 is the package, but libsqlite3-dev is the dev package that needs including? Feb 11 13:57:23 ascor: hang on a second, you have this oe_filter_out statement in a recipe? Feb 11 13:57:53 ascor: which recipe did you add it to? Feb 11 13:58:26 bluelightning : yes, in a .bb file, its my main image bb file. The one I use on bitbake command line Feb 11 14:00:23 ascor: that is not going to work at all Feb 11 14:00:32 :'( Feb 11 14:00:52 ascor: recipe cannot influence variable values outside its own context Feb 11 14:01:08 s/recipe/a recipe/ Feb 11 14:01:34 you really need your own distro config to do what you're trying to do I'm afraid Feb 11 14:01:48 ok Feb 11 14:02:16 sounds logic Feb 11 14:03:43 But, if so, how do you explain that bitbake -e does show bluetooth if I remove it ? Feb 11 14:04:01 Sorry Feb 11 14:04:07 one more time Feb 11 14:04:19 But, if so, how do you explain that bitbake -e does NOT show bluetooth if I remove it ? Feb 11 14:04:44 :q Feb 11 14:04:53 whoops :) Feb 11 14:08:00 ascor: what is the exact bitbake -e command line you are using? Feb 11 14:09:17 bitbake -e my-recipe-with-filter-out-bluetooth Feb 11 14:10:11 ascor: right, so you are telling bitbake to print out the context of the image Feb 11 14:10:21 ascor: which does not affect pulseaudio or indeed any other recipe Feb 11 14:11:36 I'm going to test if all this is related to features backfill Feb 11 14:12:10 it isn't Feb 11 14:12:20 only the addition of pulseaudio to DISTRO_FEATURES is Feb 11 14:18:13 It's only added to DISTRO_FEATURES_BACKFILL. Feb 11 14:18:42 which is never added to DISTRO_FEATURES. It looks like there is a special path .. Feb 11 14:20:00 ascor: git grep features_backfill Feb 11 14:29:37 hmm. I'm having troubles making RDEPENDS conditional on eglibc. I first did RDEPENDS_${PN}-ptest_libc-eglibc += "foo" but as RP pointed out in the ml that gets overwritten. however his suggestion of RDEPENDS_${PN}-ptest_append_libc-eglibc = "foo" doesn't work either. any ideas? Feb 11 14:44:08 Zagor: I think you need a leading space in the value there Feb 11 14:44:25 it's probably just mangling the previous item in the list instead of adding an additional item Feb 11 14:44:46 bluelightning: I actually have a leading newline in both examples Feb 11 14:45:29 the previous list is unharmed, but the new items are not added Feb 11 14:48:43 Zagor: try glibc Feb 11 14:48:48 Zagor: not eglibc Feb 11 14:49:05 ok Feb 11 14:50:38 that worked in my tests here Feb 11 14:53:44 yes, that works for me too. thanks. Feb 11 15:28:00 woglinde, ping Feb 11 15:30:52 afternoon all Feb 11 15:31:36 hi pb_ Feb 11 15:32:09 hi bluelightning Feb 11 15:45:57 hi folks Feb 11 15:46:05 * eren long time, no see :) Feb 11 15:49:25 hi eren Feb 11 16:59:21 Good evening Feb 11 20:47:41 JaMa: have you seen any fallouts from systemd class merge into oe-core Feb 11 20:48:18 yes, e.g. avahi now fails to parse Feb 11 20:48:36 I deleted the class from meta-systemd Feb 11 20:48:44 just dropping systemd from meta-systemd does not resolve that here Feb 11 20:48:50 and the class in OE-Core does not offer all services Feb 11 20:49:03 yes Ross said that in pull request Feb 11 20:49:24 did he also say how to "port" the recipes which already use the logic from old class Feb 11 20:50:09 I don't think so and I'm too busy to compare what exactly changed in "new" implementation Feb 11 21:28:18 hello all Feb 11 21:55:11 hello Feb 11 21:55:13 how are you Feb 11 21:55:37 bluelightning: hello long time no see Feb 11 21:55:50 hi Zoidburg Feb 11 21:55:59 how are you? Feb 11 21:56:39 bluelightning: fine fine, having a lot of stuff todo Feb 11 21:57:07 bluelightning: but i have now a new workstation, where i can compile stuff relatively fast Feb 11 21:57:10 ^ Feb 11 21:57:17 bluelightning: and you? Feb 11 21:57:27 Zoidburg: not too bad... also been pretty busy Feb 11 21:57:33 bluelightning: you havent been to the 29c3 have you? Feb 11 21:57:42 Zoidburg: no... I did go to FOSDEM though Feb 11 21:58:12 bluelightning: ok i havent been to fosdem :-) 29c3 was awesome this year in hamburg... Feb 11 22:03:17 can someone suggest some good RT benchmarks ? Feb 11 22:11:31 bluelightning: how was fosdem? Feb 11 22:12:03 Zoidburg: didn't get to see many talks, I was running the OE stand this year Feb 11 22:12:16 Zoidburg: but it seemed like a good conference Feb 11 22:15:38 hey guys can anyone help if i build angstrom i'm encountering some errors... http://pastebin.com/Nbn9apqq Feb 11 22:17:14 can this be an error connected to the missing makeinfo in ubuntu? Feb 11 22:22:09 Zoidburg: there was a binutils patch for it Feb 11 22:22:46 khem: how can i patch this? Feb 11 22:25:39 Zoidburg: may be use 2012.12 release Feb 11 22:26:40 where can i get this? **** ENDING LOGGING AT Tue Feb 12 02:59:58 2013