**** BEGIN LOGGING AT Fri Mar 07 02:59:58 2014 Mar 07 08:48:52 good morning Mar 07 09:44:24 morning all Mar 07 09:45:21 morning Mar 07 10:14:51 hello is there a chan for raspberry pupose with yocto ? Mar 07 10:15:17 i would like to know if there is a way to make gstreamer 1.0 decode h264 video Mar 07 10:15:30 in yocto Mar 07 10:15:42 i mean with yocto image Mar 07 10:36:29 otavio: ping Mar 07 12:32:15 jbrianceau: Hi Mar 07 12:50:19 otavio: hi, I see your name in qt5 recipes, and I'm just wondering if there is a plan to move to Qt 5.2 Mar 07 12:50:36 jbrianceau: master has 5.2.1 Mar 07 12:51:40 otavio: oh that's great, I missed that Mar 07 12:51:44 otavio: thanks Mar 07 12:58:07 hmm, I have to say that the way how /var/volatile settings are done is not really ideal, this all is very much cluttered around causing confusion, took me a while to figure out that fs_perms.txt was actually overwriting things that my base-failes recipe was doing Mar 07 13:00:13 jbrianceau: yw Mar 07 13:27:47 #join kernel Mar 07 13:28:09 hmm ... Mar 07 14:59:54 Hi everybody. Does anybody know a way to force an image into the recovery shell on boot without there actually being a problem? Is this even the correct channel for this question? Mar 07 15:34:08 Hi, how would I change the hardware capabilities of a target device? My goal is to build yocto for an overo gumstix, and make it use exclusively ARM instructions, no SIMD or THUMB[2] Mar 07 15:41:34 Hello. I need to add mercurial to the SDK but I cannot make it works. I made a recipe with : "require recipes-core/meta/meta-toolchain.bb" and "TOOLCHAIN_TARGET_TASK += "mercurial"". I got the following error message : E: Unable to locate package mercurial. Do you have any idea ? Mar 07 15:43:44 rbuchmann: and you have a recipe that provides mercurial? Mar 07 15:44:40 Garibaldi: yes. I have mercurial-native_1.9.bb from the dylan meta-oe Mar 07 15:45:07 rbuchmann: that's mercurial-native though i.e. for the host system, not the SDK or target Mar 07 15:45:33 rbuchmann: I think what you probably want is TOOLCHAIN_HOST_TASK += "nativesdk-mercurial" as well Mar 07 15:46:26 and then you'd need to provide a recipe that provides a nativesdk-mercurial target (either you write a nativesdk-mercurial recipe, or a mercurial target recipe that has BBCLASSEXTEND = "nativesdk" in it) Mar 07 15:46:50 you could use the mercurial-native recipe as a starting point of course Mar 07 15:50:37 I wrote a recipe with BBCLASSEXTEND += "nativesdk". Is it wrong to have finally "native nativesdk" ? Mar 07 15:51:36 rbuchmann: well, the native part would overlap the mercurial-native recipe (if you still have that in your configuration) Mar 07 15:51:43 but otherwise it wouldn't be wrong, no Mar 07 15:51:48 another question... TOOLCHAIN_HOST_TASK implicitely add the prefix "nativesdk-" right ? Mar 07 15:52:09 rbuchmann: no, you must use the nativesdk- prefix when adding items to the value Mar 07 15:52:42 see how it's set in existing examples Mar 07 15:52:56 ha ok. Thx. The nativesdk version fails while making. I guess I need to fix something. Mar 07 15:53:04 where are the existing examples ? Mar 07 15:53:17 git grep TOOLCHAIN_HOST_TASK Mar 07 15:53:18 ;) Mar 07 15:53:37 hi, can I trust Yocto and opkg to update my custom linux kernel to linux-yocto on the board properly? Mar 07 15:53:58 Does nobody know how to force the boot into the recovery shell without an actual error? Mar 07 15:55:26 hsychla: what error are you facing Mar 07 15:59:58 none currently. the problem is that we had a FS error which put us into recovery shell on reboot but we could not log into it. we suppose that is because there was no root password set. we then used a knoppix cd to fix the FS. now we have set a root password and want to test if we can now actually get into the recovery shell... Mar 07 16:39:01 JaMa: MODULE_IMAGE_BASE_NAME does not seem to be documented Mar 07 16:42:52 I wrote a recipe for mercurial. When i run bitbake nativesdk-mercurial, it works. When I build the metatoolchain, i got : "Unable to locate package nativesdk-mercurial". What is wrong ? :'( Mar 07 16:43:37 rbuchmann: it might have built it but did it package anything? Mar 07 16:44:06 how is it possible to check that ? Mar 07 16:45:17 if i check the package subdirectory in tmp/work, does it mean something ? Mar 07 16:47:16 rbuchmann: check the package split folder. Mar 07 16:47:55 there is no package folder. Mar 07 16:49:04 package-split Mar 07 16:49:15 no package-split too Mar 07 16:49:41 I have : image/ mercurial-2.0.2/ temp/ Mar 07 16:49:41 license-destdir/ sysroot-destdir/ Mar 07 16:49:55 find ./ -name package-split Mar 07 16:52:02 find ../build_yocto/tmp/work/x86_64-nativesdk-pokysdk-linux/ -name package-split => no dir Mar 07 16:53:51 check the image dir, thats what you installed into ${D} in do_install Mar 07 16:56:14 image dir ? You mean tmp/work/x86_64-nativesdk-pokysdk-linux/nativesdk-mercurial/2.0.2-r2/image/ ? Mar 07 16:57:04 yes Mar 07 16:57:09 If yes, there are several binaries into this directory Mar 07 16:57:21 for example : tmp/work/x86_64-nativesdk-pokysdk-linux/nativesdk-mercurial/2.0.2-r2/image/opt/poky/1.4.2/sysroots/x86_64-pokysdk-linux/usr/bin/hg Mar 07 16:57:36 if you have files there, but not in package, then either your FILES vars are wrong, or you installe yoru files to the wrong location, generally Mar 07 17:00:22 except a fix on staging dirs, I let autotools do its work. I got the files into image dir but not in package. I did not redefine FILES_${PN}. Do I need to do that ? Mar 07 17:02:47 I have NO run.do_package or log... I guess I missed a huge thing but i have no clue about it Mar 07 17:08:57 rbuchmann: your recipe doesn't still have "inherit native" in it does it? Mar 07 17:09:53 the original recipe has "inherit native". My append has BBCLASSEXTEND += "nativesdk" Mar 07 17:11:12 rbuchmann: I don't know if by append you mean a bbappend or just what you put on the end... Mar 07 17:11:29 ? Mar 07 17:11:30 bbappend Mar 07 17:11:43 ok, you can't do this as a bbappend Mar 07 17:11:49 oups Mar 07 17:12:07 you need to copy and modify the mercurial-native recipe to mercurial_.bb and then modify it Mar 07 17:12:34 and the first modification to make to the copy is to drop "inherit native" Mar 07 17:12:47 I suspect you'll find it will do the packaging stuff properly after that Mar 07 17:13:25 ok. I will try Mar 07 17:13:54 so to understand Mar 07 17:14:13 i can't "convert" a native into a "nativesdk" Mar 07 17:14:42 but if there is any other inherit that does NOT include native, I still can use BBCLASSEXTEND ? Mar 07 17:14:49 sure Mar 07 17:15:13 really, having separate -native recipes is an old convention; we much prefer to use BBCLASSEXTEND for those as well Mar 07 17:15:54 in fact once you have this working it might be nice to send a patch for meta-oe that drops the -native recipe and adds your new one with "native" in BBCLASSEXTEND as well Mar 07 17:17:20 I would be happy to send a patch to the community but at this moment I'm still a noob and I'm not sure to understand what I'm doing Mar 07 17:19:49 rbuchmann: sure, no worries Mar 07 18:05:05 I still have "Unable to locate package nativesdk-mercurial" but I have : tmp/work/x86_64-nativesdk-pokysdk-linux/nativesdk-mercurial/2.0.2-r2/packages-split/nativesdk-mercurial Mar 07 18:06:32 and tmp/work/x86_64-nativesdk-pokysdk-linux/nativesdk-mercurial/2.0.2-r2/deploy-debs/x86_64-nativesdk/nativesdk-mercurial/nativesdk-mercurial_2.0.2-r2_x86-64.deb Mar 07 18:11:45 rbuchmann: you are adding it to TOOLCHAIN_HOST_TASK (not TOOLCHAIN_TARGET_TASK) right? Mar 07 18:56:43 sigh, and hit another bitbake hang, and when trying to ^c it, got timeouts communicating with the bitbake server Mar 07 18:57:03 seems like 10 times a day i end up resorting to pkill -f bitbake, or even pkill -9 -f bitbake, to shut down jobs Mar 07 18:57:08 hmm Mar 07 18:58:45 there has to be something unusual going on in your configuration Mar 07 18:59:21 well, maybe not unusual, but some kind of mode that's not entered when I use bitbake, at any rate Mar 07 19:01:21 I did Ctrl+C in the middle of searching for available sstate packages, and a server timeout happened; but that's the only time I have seen that recently Mar 07 19:01:34 and I regularly Ctrl+C out of builds, forcibly Mar 07 19:15:51 hmm, SRC_URI_mymachine_append = "blah" did not append but did override the actual SRC_URI... what am I missing? Mar 07 19:59:32 Jin^eLD: you got it backwards Mar 07 19:59:43 that appends to SRC_URI_mymachine, which then blows away SRC_URI Mar 07 19:59:47 you want SRC_URI_append_mymachine Mar 07 20:07:32 thanks kergoth, i needed that a couple nights ago... Mar 07 20:13:47 kergoth: arg... thank you!! Mar 07 20:15:38 np Mar 07 20:16:20 its a common source of confusion, but there are valid cases for both (though admittedly FOO_machine += or .= is better in the _machine_append case) Mar 07 20:17:02 so you suggest I change it to SRC_URI_mymachine += Mar 07 20:17:03 ? Mar 07 20:17:17 I was just used to the _append stuff but have not used it in some time so got it all wrong :) Mar 07 20:17:26 no, in your case you should use _append_machine Mar 07 20:17:39 _append_machine is a conditional append, which is what you want Mar 07 20:17:45 as opposed to an append to a conditional value Mar 07 20:18:08 generally _append/_prepend should be avoided unless necessary, and this is a case where it's necessary :) Mar 07 20:18:41 :) Mar 07 20:18:44 there's no other syntax for an append/prepend based on an override. using SRC_URI_ymachine += would have the same problem you just hit, since it appends to an override which then blows away the original Mar 07 20:18:53 I have a specific layer that should support several machines that we are using Mar 07 20:19:22 * kergoth nods Mar 07 20:41:52 kergoth: why is the volatiles configuration spread in three different places? is it just "historical" or are there any technical reasons? Mar 07 20:44:54 and one more question... would things like pkg_postinst_${PN}_mymachine work? Mar 07 20:45:12 I somehow could not find a machine-condition example for the opkg hooks Mar 07 20:59:11 halstead: Hi, is there any known reason for git.yocto clones to be very slow now? linux-yocto-3.10 clone estimate shows 6-7 hours now Mar 07 20:59:39 JaMa, Checking. There is a backup in progress. Mar 07 21:00:21 JaMa, That doesn't explain it though. So no known reason. One moment. Mar 07 21:01:38 JaMa, We have some SYN flooding on port 80 right now. Are you cloning using http:// or git:// ? Mar 07 21:04:57 let me check with person which reported this to me Mar 07 21:05:49 halstead: git:// Mar 07 21:09:17 JaMa, I'm getting ~450KB/s from San Francisco, and 600KB/s from New York. Were are you at? And if possible, your ISP? Mar 07 21:10:23 I'm in San Francisco, but the report is from Ukraine (but saying that other bandwidth seems ok) Mar 07 21:10:42 halstead: so please don't worry about it Mar 07 21:10:54 halstead: JaMa: fwiw, I saw that intermittently last week, but nothing reliably enough to report as an issue. Mar 07 21:11:27 how do I have to change the gumstix layer to prevent it from using vfp instructions? Mar 07 21:12:28 JaMa, darknighte_ I appreciate the report. I'll keep an eye on things. We are always planning to increase bandwidth by the way. :) Mar 07 21:23:06 Has anyone here gone through setting up the SDK through yocto? Mar 07 21:24:00 I have this generated, but i'm a bit confused by how eclipse is reacting to it, it doesn't seem to offer the autotools chain, and it also doesn't seem to be using the includes from the sysrootfs Mar 07 21:24:15 which doesnt make sense, because what is the point if you dont use those includes... Mar 07 21:26:11 halstead: darknighte_: thanks Mar 07 21:44:34 hello everyone Mar 07 21:45:26 anyone know if pulseaudio 4.0, present on the yocto package, is compatible with blueZ5 ? Mar 07 21:52:51 no idea Mar 07 21:52:59 try it and see... Mar 07 22:08:09 hmm I can't seem to find that info, is it possible to use machine specific pkg_* hooks? (i.e. postist, preinst, etc.) Mar 07 22:29:20 seems the trick is to somply add an empty pkg_postinst_${PN} and then a pkg_postinst_${PN}_append_machine, I guess I figured it out now :> Mar 07 22:51:07 hi everyone. I have a quesion about a package..is it possibile to get the last version of pulseaudio, that'is pulseaudio 5.0? Mar 07 23:00:59 br1_21: if you update the recipe, sure. Mar 07 23:01:18 so, right now is not avaible? Mar 07 23:01:32 andhe_: ^ Mar 07 23:01:47 br1_21: I don't think anyone else has updated the pulseaudio recipe yet. Mar 07 23:02:07 and may I do a request for that? Mar 07 23:04:12 br1_21: we all like to request that someone else does our work for us, but it usually doesn't work like that. ;) Mar 07 23:04:38 br1_21: if you wait around, someone else will likely update pulseaudio sooner or later. Mar 07 23:04:39 yeah I know! It was just to understand xD Mar 07 23:05:15 andhe_: anyway, thanks! Mar 07 23:31:23 Does anyone know if the eclipse plugin no longer has autotools support? I only have cmake support. Mar 07 23:31:36 Has that been deprecated or did I mess up on my setup? Mar 08 02:31:52 what's tom zanussi's irc nick? **** ENDING LOGGING AT Sat Mar 08 02:59:59 2014