**** BEGIN LOGGING AT Tue Apr 09 02:59:58 2013 Apr 09 07:08:59 hey! I have a problem, creating a recipe for the genivi DLT daemon. I'm getting a "ERROR: dlt: md5 checksum is not specified for file://LICENSE.txt;" Apr 09 07:09:35 But I have specified: LIC_FILES_CHKSUM = "file://LICENSE.txt; md5=99ba60c3fad7eaf8c56bca6dd75cba09" Apr 09 07:10:33 And the file path seems to be correct because if I change the file name bitbake complains about the missing file. Apr 09 07:13:36 dang, found the problem. the white space between "LICENSE.txt; md5" was the culprit. Apr 09 07:44:49 good morning Apr 09 07:46:03 john76: great ;-) Apr 09 08:30:44 morning all Apr 09 08:30:59 bluelightning: hi Apr 09 08:31:35 bluelightning: Hi, that patch for buildhistory works for me Apr 09 08:31:47 JaMa: great, thanks, I'll post it Apr 09 08:31:50 hi mckoan Apr 09 08:32:16 bluelightning: maybe add comment with warning that with SSTATE_MIRRORS you won't get complete list of SRCREVs Apr 09 08:32:46 JaMa: well, you won't get a complete buildhistory contents at all with sstate Apr 09 08:33:09 something I hope to fix for 1.5 Apr 09 08:33:19 great Apr 09 10:49:53 does anyone have yocto running on a nitrogen6x? Apr 09 11:10:11 what is that? Apr 09 11:18:31 Q: how to get target specific packages built and installed to image? For example firmware-imx, imx-audio and gst-fsl-plugin packages from meta-fsl-arm. They are needed by HW independent gstreamer, and are in the MACHINE_EXTRA_RRECOMMENDS but are not getting built, or installed to image. Apr 09 11:29:51 mcfrisk: which image you're building? Apr 09 11:30:26 we have a custom image for pandaboard, I'm adding sabresd too but image is missing those packages. Apr 09 11:30:47 mcfrisk: right Apr 09 11:30:59 mcfrisk: those are add by the tasks (packagegroups) Apr 09 11:31:46 ...ing/fsl-community-bsp-mx6/sources/poky% git grep MACHINE_EXTRA_RRECOMMENDS|grep packagegroup Apr 09 11:31:49 meta/recipes-core/packagegroups/packagegroup-base.bb:RRECOMMENDS_packagegroup-machine-base = "${MACHINE_EXTRA_RRECOMMENDS}" Apr 09 11:32:09 mcfrisk: so your image should install this one Apr 09 11:32:28 is this enough to get the packages built as well? Apr 09 11:32:58 mcfrisk: sure Apr 09 11:33:45 anything that's rdepended or rrecommended will be built Apr 09 11:35:25 sigh, and base system is denzil, meta-fsl-arm from danny with backported BSP 1.1... Too many hacks, sorry. Apr 09 12:31:34 mcfrisk: arrrg! Apr 09 12:59:37 On latest danny I get "NOTE: Your conf/bblayers.conf has been automatically updated. Please re-run bitbake." when I try to build something. Then I get it again. And again :) Apr 09 13:01:35 Hello, how do i get a crosscompiler for u-boot out of yocto? Because I think adt is only for applikations an do not run bare metal right? Apr 09 13:04:53 the adt compiler should build u-boot fine Apr 09 13:05:03 mattnie: you can use that compiler Apr 09 13:05:07 u-boot is careful not to link against libraries Apr 09 13:06:07 i got error using adt => arm-poky-linux-gnueabi-ld.bfd -Wl, -01 -Wl,--hash-style=gnu -Wl,--as-needed Apr 09 13:06:10 and some more Apr 09 13:06:41 mattnie: which error? Apr 09 13:07:09 unrecognized option '-Wl,-01' Apr 09 13:08:54 right, that's an option that gcc would handle Apr 09 13:09:01 -Wl means "pass to linker" Apr 09 13:09:13 I've seen fixes for those kind of issues.. Apr 09 13:09:23 so if you're invoking ld directly, just pass -01 Apr 09 13:09:34 (-O1?) Apr 09 13:12:21 http://pastebin.com/FcD1aqCQ Apr 09 13:12:24 this is the whole output Apr 09 13:17:20 tasslehoff: did you remove meta-yocto? if so, just decrease the value of LCONF_VERSION in your bblayers.conf by one Apr 09 13:17:37 mattnie: LDFLAGS too equal to CFLAGS? Apr 09 13:19:26 bluelightning: meta-yocto is still there, and bumping LCONF_VERSION didn't help Apr 09 13:20:45 tasslehoff: this situation can only happen if LAYER_CONF_VERSION != CONF_VERSION Apr 09 13:21:32 er, I mean LCONF_VERSION not CONF_VERSION Apr 09 13:21:40 mcfrisk: I compilied the same u-boot configuration with linaro-arm-linux-gnueabi-2012.04-20120426 but now i want to move completly to yocto!# Apr 09 13:24:25 mattnie: then I would update the yocto u-boot recipe, they for example unset LDFLAGS, CFLAGS and CPPFLAGS (on denzil anyway) before calling make. Apr 09 13:24:28 mcfrisk:: so i think it mus be a configuration failure with adt? I am not sure i "installed" the adt right.. and also do not know how to test.. :( I am still new to the embedded world and have a lot of questions... Apr 09 13:25:49 bluelightning: where is LAYER_CONF_VERSION set? Apr 09 13:25:56 * tasslehoff greps and asks at the same time Apr 09 13:26:37 mattnie: check meta/recipes-bsp/u-boot/u-boot.inc in your yocto for the compilation tricks for u-boot, they do unset for LDFLAGS etc. Do the same in adt. Apr 09 13:28:48 bluelightning: hm. it seems my LCONF_VERSION is actually higher Apr 09 13:30:22 Thank you very much. You are very helpfull to me. I will try it and come back later.:) Apr 09 13:30:22 I had to decrease from 6 to 5, then run bitbake so it could bump it to 6 itself.. Apr 09 13:32:43 gmacario: welcome Apr 09 13:33:40 * tasslehoff must run Apr 09 14:01:20 rburton: I included xinput-calibrator in Yocto master and tested it successfully Apr 09 14:01:57 rburton: I did my patched under meta Apr 09 14:01:59 mckoan: sweet. too late for 1.4 but we can merge it asap in 1.5. Apr 09 14:03:09 rburton: my doubt is, can I send the patches to oe-core? (I used them in yocto's meta) Apr 09 14:03:37 mckoan: yes, poky's meta/ is identical to oe-core's meta/ Apr 09 14:03:40 rburton: I usually used git send-email --to=openembedded-core@lists.openembedded.org patchfilename Apr 09 14:03:57 rburton: yep, just a confirm, thx Apr 09 14:04:20 mckoan: that's fine. just be careful when doing a series that you don't mix up changes to meta/ and meta-yocto/ as those need to go to oe-core@ and poky@ Apr 09 14:05:15 rburton: yes of course Apr 09 14:19:49 rburton: patches sent Apr 09 14:22:06 rburton: I'm afraid I wrote a wrong [meta-oe] in the subject :-( Apr 09 14:22:43 heh Apr 09 14:22:50 libxaw?! Apr 09 14:23:18 are you sure about that Apr 09 14:23:23 or was that just copied from the old recipe Apr 09 14:24:39 rburton: dependency chain was: ['xinput-calibrator', 'xterm', 'libxaw', 'xmlto-native'] Apr 09 14:25:10 ffs Apr 09 14:32:20 ugh libxaw Apr 09 14:39:20 mckoan: "just say no" to xterm Apr 09 14:41:55 mcfrisk: Now I have read a bit into what happens in u-boot.inc but I think furthermore it is a problem of adt because with linaro cross compiler I do not set up more ld and c flags ... Apr 09 14:42:31 rburton: any news about xinput-calibrator in oe-core? Apr 09 14:42:42 ant_work: for 1.5 Apr 09 14:42:59 great Apr 09 14:43:08 patches on the list, i'm not entirely happy about pulling in a third terminal into oe-core Apr 09 14:43:13 but patches on the list! :) Apr 09 14:43:13 mattnie: depends how the cross compile has been setup. btw I Apr 09 14:43:26 btw I've newer used adt.. Apr 09 14:44:00 rburton: iirc if it finds the pointercal fil eit doesn't spawn Apr 09 14:45:07 oh, I see now, I was fooled by the meta-oe prefix ;) Apr 09 14:45:25 ant_work: my fault, sorry Apr 09 14:45:28 mattnie: I see adt as a 'application development tool'. I provides easy access to default compiler, default compiler settings etc from eclipse. u-boot does not use those as u-boot.inc shows. I doubt adt remebers how each yocto package needs to be comfigured and compiled. Apr 09 14:45:52 mckoan: on the opposite, congrats for having revived the issue Apr 09 14:46:14 ant_work: unfortunately I needed it ;-) Apr 09 14:46:30 I'm still unpacking the ipk on device atm :) Apr 09 14:58:28 YPTM: Beth joined Apr 09 14:58:59 TI's UK conf number is still broken Apr 09 14:59:56 denix: are you able to ping someone about that? (it just goes through what sounds like the standard TI PBX) Apr 09 15:00:41 YPTM: Welcome to the technical team meeting, please let me know who's on the bridge. Apr 09 15:00:54 YPTM: Tom on the call Apr 09 15:00:55 Song_liu: Richard is on the call Apr 09 15:01:07 YPTM: hear, here Apr 09 15:01:10 YPTM: Michael on the call. Apr 09 15:01:12 YPTM: Bruce Ashfield is only available via IRC. I have a conflict. Apr 09 15:01:16 RP_: did you dial the US? Apr 09 15:01:21 YPTM: Kevin Strasser is dialing in Apr 09 15:01:23 YPTM: Jeff is here Apr 09 15:01:31 rburton: I found a UK number Apr 09 15:01:39 RP_: ooh, what was it? Apr 09 15:01:46 YPTM: Yo estoy aqui, creo que puedo ayudarlos ahora Apr 09 15:02:02 Björn Stenberg joined Apr 09 15:02:10 RP_: I would also like it :) Apr 09 15:02:18 YPTM: Fahad just joined from Mentor Graphics Apr 09 15:02:29 YPTM: Cristian is present Apr 09 15:02:31 YPTM: Corneliu joined Apr 09 15:02:32 YPTM: jzhang's on the call Apr 09 15:02:38 YPTM: ross joined Apr 09 15:02:45 YPTM: Scott Rifenbark joined the call Apr 09 15:02:57 YTPM: Saul is in Apr 09 15:03:40 Ramana joined YPTM Apr 09 15:03:41 oh the memories Apr 09 15:04:12 Hi msm, lurking are we? Apr 09 15:04:20 YPTM: Laurentiu Palcu joined Apr 09 15:04:21 YPTM: Paul joined, on the US number... Apr 09 15:04:35 alex d joined Apr 09 15:04:37 YPTM: Nitin joined the bridge Apr 09 15:04:46 use the US number, not hte toll free one Apr 09 15:04:52 YPTM: here Apr 09 15:05:00 bluelightning: let me find out Apr 09 15:05:30 sgw1: i never left ;) Apr 09 15:05:37 YPTM: Any opens? Apr 09 15:05:44 denix: we first noticed the issue last week - the number we were using was 01604 663003 Apr 09 15:06:28 post build configuration Apr 09 15:06:56 YPTM: I'm here.. sorry for being late Apr 09 15:08:04 Ramana: what's the issue number for your proposed feature? Apr 09 15:08:27 Ramana_: do you have a feature request in bugzilla for that? Apr 09 15:08:39 an open from me: status of ptest for 1.4. there are still outstanding bug fix patches on the list that have yet to be reviewed/merged Apr 09 15:08:49 yes Apr 09 15:09:13 Ramana_: please link ;) Apr 09 15:09:18 Corneliu: I think bug #4106 is it Apr 09 15:09:19 Bug https://bugzilla.yoctoproject.org/show_bug.cgi?id=4106 enhancement, Medium+, 1.5, richard.purdie, NEW , Deployment user experience enhancement Apr 09 15:09:21 Zagor: I've heard sgw1 was having problems with it Apr 09 15:09:34 Zagor: I'm also concerned about the debug package business :/ Apr 09 15:09:42 ok I got it Apr 09 15:09:48 Zagor: btw we would really appreciate some material for the docs for ptest ;) Apr 09 15:09:51 Zagor: I have general issues with the -dbg packages but that is 1.5 time period work Apr 09 15:10:07 Zagor: I was going to get back to you, there was the build history issue I sent you, I guess I want to make sure I have the setup correctly Apr 09 15:11:18 bluelightning: ah, I've forgotten that. where would that fit in? Apr 09 15:11:28 https://bugzilla.yoctoproject.org/show_bug.cgi?id=3252 Apr 09 15:11:29 Bug 3252: enhancement, Low, Future, jessica.zhang, NEW , Binary level package selection and configuration tool. Apr 09 15:12:10 Zagor: I'm guessing we would have a section within the development manual on the general topic of recipe-specific tests Apr 09 15:12:21 Zagor: I'd recommend opening up a dialogue with scottrif Apr 09 15:12:39 RP_, sgw1: I can't repeat saul's issue. the dbus recipe does not inherit ptest and therefore should not create a dbus-ptest package to conflict with the dbus-ptest recipe Apr 09 15:14:18 https://lists.yoctoproject.org/pipermail/yocto/2013-April/015297.html is RFC mentioning the post build configuration Apr 09 15:28:20 Zagor: I need to go back and look at the issue and see if I can figure out what is going on Apr 09 15:28:50 RP_: it seems sgw_ and I figured it out off-channel. looks like a config issue. Apr 09 15:33:04 I have a v4 of the ptest patch btw, I'll send it first thing tomorrow. Apr 09 15:43:42 YPTM: thank you all for joining the meeting. Have a nice day/evening Apr 09 17:40:22 I'm noticing something I don't yet understand: it seems like if I have a recipe which DEPENDS on another that 'inherit meta's--say, an image recipe--then the image isn't necessarily put into deploy/images when I build the depending recipe. Why is that? Apr 09 17:41:58 It would make sense to me were it not for the 'inherit meta'. I would have thought the do_build[recrdeptask] = "do_build" that corresponds to would be enough to force the image to be generated when something DEPENDS on it...what am I missing? Apr 09 17:45:24 evanp: look at how initramfs is done Apr 09 17:45:44 evanp: IIRC look at do_bootimg task Apr 09 17:45:52 and go from there Apr 09 18:59:42 Can yocto install the initramfs and bootloader to the ext3 rootfs instead of the hddimg? Apr 09 19:14:23 RP_: ping? Apr 09 19:32:59 hi, I have a quick question, I'm using latest yocto release Apr 09 19:33:17 is there anyway to make one particular recipe to not generate sstates Apr 09 19:35:24 in particular I have my own kernel recipe (based on the oe kernel class), and my developers need to keep debugging the kernel, requiring the full source Apr 09 19:35:55 we use sstate server to speed builds, but they are getting the kernel sstate, and therefore lacking the kernel sources to debug or develop Apr 09 19:36:24 I have a workaround to avoid uploading the sstate to the central server, but want to avoid the recipe from generating sstate at all Apr 09 20:21:51 zecke: pong Apr 09 20:32:45 RP_: Hi, nobody responded to icecc RFC, is it OK to go in (I've tested it with internal and external toolchain - but without access to ice scheduler) Apr 09 20:32:59 JaMa: I guess so... Apr 09 20:33:05 RP_: at least it doesn't make things worse and fixes our use case with external TC Apr 09 20:44:01 is DEPENDS = "foo" equivalent to do_configure[depends] = "foo:do_populate_sysroot"? some other task? Apr 09 20:46:48 evanp: roughly. it's handled through deptask, see base.bbclass, there's do_configure[deptask] = "do_populate_sysroot", iirc Apr 09 20:53:33 RP_: one quick question about debug packages.. Apr 09 20:54:11 RP_: First perf (from linux 3.6) only supports one style of debug packages (the one that is not default in OE-core).. (but then it is still broken) Apr 09 20:54:32 RP_: second. there doesn't appear to be a way to have debug packages without /usr/src/ Apr 09 20:55:01 so the effect for my hardware with a small flash partition is that I can not use perf report on the device.. and not on my host. :} Apr 09 20:55:19 but I will be offline now Apr 09 20:55:43 zecke: with oprofile we used to be able to run the trace, save it, then analyse it on a different machine using a full set of debug symbols Apr 09 20:56:07 zecke: It sounds like perf hasn't got to that level of sophistication yet :( Apr 09 20:56:11 RP_: yes, but perf only supports the other (non default) debug style. Apr 09 20:56:24 RP_: and even then it is somehow broken Apr 09 20:56:39 zecke: Its a shame it doesn't use standard libs like libbfd since those support the debug style we use Apr 09 20:57:09 okay, i will take another look. I thought it is using libbfd as well Apr 09 20:57:15 zecke: As for turning off /usr/src, I'd take an option to add that. I thought there was even a patch recently but I don't remember what happened to it Apr 09 20:57:25 zecke: Certainly this all worked in oprofile and the code was in libbfd Apr 09 20:57:44 zecke: I know as I used to use it and helped write some of it :) Apr 09 20:58:21 zecke: tomz1 is our perf expert btw Apr 09 20:58:36 i used it too.:) Apr 09 20:58:54 now I am forced to make journald useful and it feels like three steps back Apr 09 20:59:27 - i failed to create a ubi/ubifs big enough for my -dbg packages, i didn't find the flag for the flash size. :} Apr 09 20:59:42 - the huge -dbg packages are a bit of a problem Apr 09 20:59:58 - and perf has an option for another root but 'cross perf' appears to be bogus Apr 09 21:00:08 anyway. i will start with the /usr/src Apr 09 21:00:25 zecke: er... didn't you add the option to disable source in debug packages? Apr 09 21:00:51 :} Apr 09 21:01:05 no, my bad memory, it was JaMa Apr 09 21:01:28 * zecke wishes celf would have paid him to add prologue scanning for callbacks to MIPS/ARM. :} Apr 09 21:01:42 zecke: http://cgit.openembedded.org/openembedded-core/commit/?id=3c8452c3abae74a42989c0fbd5ba303788528750 Apr 09 21:02:14 JaMa++ Apr 09 21:02:24 reminds me, yet another var needing docs Apr 09 21:02:28 * bluelightning adds to todo list Apr 09 21:02:53 bluelightning: you're doing a good job, I wsh I could find more time to help Apr 09 21:03:22 RP_: no worries, you have more important things to do right now Apr 09 21:03:49 bluelightning: like destress now the sanity issue might finally be solved! :) Apr 09 21:04:13 RP_: :) let's hope that's confirmed by not seeing the failure ever again :) Apr 09 21:04:54 bluelightning: all green autobuilder at least :) Apr 09 21:04:57 :) Apr 09 21:05:00 good night! Apr 09 21:05:05 zecke: night Apr 09 21:05:09 'night zecke Apr 09 21:05:15 \o/ Apr 09 22:31:53 I have a package that's failing to build because Python is not accessible in a sysroot or staging dir. How do I make sure it's available? Apr 09 23:06:56 can someone quickly point me to some doc which says how do I add new machine ? Apr 09 23:06:58 to yp Apr 09 23:09:19 khem: http://www.yoctoproject.org/docs/current/mega-manual/mega-manual.html#platdev-newmachine Apr 10 00:04:49 thanks wmat **** ENDING LOGGING AT Wed Apr 10 02:59:58 2013