**** BEGIN LOGGING AT Tue Jan 07 02:59:58 2014 Jan 07 08:40:39 What does it mean when a custom kernel module that I'm building is "not found in the base feed"? Jan 07 08:41:16 bitbake works fine, but when I add "kernel-module-" to my image recipe, I get that error message. Jan 07 08:41:48 i meant . Jan 07 08:58:08 ah. I know why it doesn't work. The kernel module name (recipe name is "Something" while the Makefile produces a "SomeOther.ko". Jan 07 08:58:25 Can I use FILES_${PN} += "SomeOther.ko" ? Jan 07 08:58:42 or doesn't that work for specifying the kernel object? Jan 07 09:08:21 does the recipe name _has_ to match the .ko file? I don't find any "options" for specifying another name. Jan 07 09:57:19 morning all Jan 07 09:59:22 good morning Jan 07 10:02:10 OutBackDingo: happy new year Jan 07 10:33:40 mckoan: hey happy new year to you also Jan 07 10:58:03 OutBackDingo: I noticed your messages in backlog, were you able to build for AT91 using Yocto? Jan 07 10:58:48 mckoan: no not as of yet when i ran into the errors i put it on hold after speaking of it in irc, but would really like to get a build done for this device Jan 07 11:00:02 OutBackDingo: you simply need to add meta-atmel Jan 07 11:01:06 OutBackDingo: but I have to admit thath I haven't tested it recently, I'm only creating custom layers for custom AT91 boards Jan 07 11:01:07 thats the thing after reading the site, it got a bit confusing so i forged through as best as i could....... ill go back into it today and see..... Jan 07 11:01:34 im thinking the meta atmel layer was in Jan 07 11:06:25 mckoan: ill go through the steps again fresh as listed on the site and see where it leads me Jan 07 11:11:59 bluelightning: hello, lucky man ... and all the best to you into this new year :) Jan 07 11:12:22 Marex: thanks :) happy new year to you! Jan 07 11:13:23 hi everyone Jan 07 11:18:55 rburton: ah, you are back! Jan 07 11:19:12 indeed i am! Jan 07 11:20:57 rburton: hny... :-) Jan 07 11:27:25 tmp/deploy/images/README_-_DO_NOT_DELETE_FILES_IN_THIS_DIRECTORY.txt Jan 07 11:27:26 why? Jan 07 11:27:36 I always do. :] Jan 07 11:27:39 what shall I expect? Jan 07 11:27:52 morning Jan 07 11:28:09 hi otavio, all Jan 07 11:28:14 morning Jan 07 11:28:16 :D Jan 07 11:28:25 morning Jan 07 11:30:29 lpapp: how about reading it ? Jan 07 11:31:09 bluelightning: I do not understand that. Jan 07 11:31:15 it _is_ recreated for me. Jan 07 11:31:37 or well, what do you mean by recreated? Jan 07 11:31:43 I use myimage, and everything is ok there. Jan 07 11:31:47 try deleting the kernel binary, that will not be re-created automatically Jan 07 11:31:52 well Jan 07 11:32:00 I use bitbake -c compile -f virtual/kernel either way Jan 07 11:32:08 so probably it is just fine for me to delete the rootfs, etc Jan 07 11:32:35 i have some troubles compiling alsa-tools with current master Jan 07 11:33:19 do_compile fails: ".../alsa-tools/1.0.26.1-r1/temp/run.do_compile.18343: 110: export: automake-1.14: bad variable name" Jan 07 11:33:41 bluelightning: the problem is that I rebuilt my kernel without modification, and it is now stuck at Uncompressing Linux... done, booting the kernel. Jan 07 11:34:03 run.do_compile:110 says " export AUTOMAKE_DIR=${STAGING_DATADIR_NATIVE}/$(ls ${STAGING_DATADIR_NATIVE} | grep automake)" Jan 07 11:34:15 which comes directly from the alsa-tools recipe Jan 07 11:34:53 first of all, i think there are some ' " 's missing Jan 07 11:35:15 lpapp: it's fine - the notice is there for new users; you're free to ignore it as long as you know how to deal with the situation Jan 07 11:35:19 so it should be 'export AUTOMAKE_DIR="${STAGING_DATADIR_NATIVE}/$(ls ${STAGING_DATADIR_NATIVE} | grep automake)"' Jan 07 11:35:38 lpapp: don't know why that would happen, sorry Jan 07 11:36:59 but that leads to some troubles with automake not finding some files, that are to be copied Jan 07 11:37:11 but that are present. Jan 07 11:40:00 and i just solved that. Jan 07 11:40:07 never mind. Jan 07 11:41:23 bluelightning: bitbake -c compile -f virtual/kernel and then bitbake myimage does not produce uImage there Jan 07 11:41:26 nor modules in there. Jan 07 11:41:42 bluelightning: last night before going home, I interrupted an ongoing build like that Jan 07 11:41:48 perhaps some sstates got corrupted? Jan 07 11:42:20 for the record: 'ls ${STAGING_DATADIR_NATIVE} | grep automake' returns 'automake-1.13 \n automake-1.14' since automake just got upgraded and rebuilt, but the sysroot/STAGING_DATADIR_NATIVE is not cleaned off the old version... manually deleting the folder for automake-1.13 solved it Jan 07 11:42:38 bluelightning: how can I force the re-stuff? Jan 07 11:42:51 simply add -f to bitbake myimage, too? Jan 07 11:42:54 or something else ? Jan 07 11:43:04 lpapp: I think you want -c deploy not -c compile... Jan 07 11:43:10 * lpapp is trying bitbake -f virtual/kernel Jan 07 11:46:34 dRp3pP3r: ouch. can you file a bug or send a mail about that? we need to patch in a "head -n1" . Jan 07 11:47:19 well maybe a tail -n1? Jan 07 11:47:58 bluelightning: does not work Jan 07 11:48:05 same issue, stuck. Jan 07 11:48:17 i sure can file a bug :) Jan 07 11:57:21 rburton: what about the ' " 's i thought were missing? Jan 07 11:57:40 it works without, but it seems strange to me Jan 07 12:02:49 rburton: #5706. filed. :) Jan 07 12:57:33 rp: have you had a chance to look at the git fetcher changes i sent in december, after your initial review? i noticed that Andrei Gherzan also sent a patch fixing that issue Jan 07 12:59:12 zibri: not yet, no. Its on the todo list for this week Jan 07 13:00:09 ok, thanks Jan 07 13:44:34 zibri: I've sent some questions via email Jan 07 13:53:37 rp: ack, will reply as soon as possible Jan 07 13:54:26 zibri: thanks Jan 07 15:54:31 I'm having the devils own time of getting a 640x480 18bit LVDS working with an i915 controller. Anyone have experience with these things? I've tried VGA16, VESA, and i915. All work in some fashion but not 100% Jan 07 16:00:22 The Yocto Project Tech Meeting Con-Call will be starting at the top of the hour Jan 07 16:00:22 Dial-in number: 1.972.995.7777 / Participant passcode: 42001078 Jan 07 16:00:23 This call is open to all and the channel remains open to discuss any topic Jan 07 16:00:48 YPTM: I'm here... Jan 07 16:00:54 happy new year all Jan 07 16:01:03 YPTM: joined Jan 07 16:01:12 YPTM: HNY All, Saul is On and will likely be leading Jan 07 16:01:22 YPTM: Scott Rifenbark joined the call Jan 07 16:01:27 YPTM: Bruce dialing. Jan 07 16:01:29 YPTM: IonutC on Jan 07 16:01:48 YPTM: polk is on Jan 07 16:01:58 YPTM: tom z here Jan 07 16:03:14 * RP is on here Jan 07 16:03:21 YPTM: Bruce is on Jan 07 16:03:47 YPTM: ross joined. Jan 07 16:04:11 YPTM: nitin is on the call Jan 07 16:04:37 YPTM: Paul Eggleton is on Jan 07 16:04:54 YPTM: belen is on Jan 07 16:05:02 YPTM: Michael is on. Jan 07 16:05:03 YPTM: jzhang on the call Jan 07 16:05:09 Beth Flanagan & Alex D are also in the room Jan 07 16:07:35 YPTM: Beth has an open - Yocto AB downtime & upgrade Jan 07 16:11:46 That's master on the autobuilder, not oecore master Jan 07 16:52:03 I am working with the meta-qt5 layer. I noticed that the qtimageformats package is not build because there are no files in it but the qtimageformats-dev packages is being build even though it doesn't have any files in it either. This is breaking my sdk build because it is trying to install qtimageformats-dev and qtimageformats package does not exist. Does anyone have suggestions for what I can do? Jan 07 16:52:36 gjohnson: fix it? Jan 07 16:52:46 JaMa is your man. Jan 07 16:53:52 lpapp: I would fix it but I don't know that the fix is. If the dev package shouldn't be build then I guess this must be a yocto bug but if the main package should be built even though there are no files then it is a meta-qt5 issue. Jan 07 16:55:02 gjohnson: qtimageformats should be available. Jan 07 16:55:11 it is necessary for loading image into QPixmap, QImage, etc. Jan 07 16:55:20 images* Jan 07 16:55:31 that is a fundamental part of the fw. Jan 07 16:55:57 lpapp: I understand that but the qtimageformats-plugins package is the one that contains the image plugins Jan 07 16:57:43 lpapp: so the problem I have is qtimageformats-dev is created even though it contains no files and qtimageformats is not created because it contains no files. When I try to populate an sdk it tries to install qtimageformats-dev which has a dependency on qtimageformats that does not exist. Jan 07 16:58:44 qt5-imageformats should be working Jan 07 16:58:52 if it is split into dev non-dev, both should work, I would say. Jan 07 17:00:53 gjohnson: set RDEPENDS_${PN}-dev to empty Jan 07 17:00:55 ALLOW_EMPTY is set for -dev packages in bitbake.conf Jan 07 17:01:23 that's why you end up with an empty -dev package, i guess Jan 07 17:01:41 JaMa: I will try that, thanks Jan 07 17:21:13 zibri: Do you know why by default dev packages are allowed to be empty? Jan 07 17:21:49 there was a discussion on the mailing list (which list? can't remember :-() recently. let me check Jan 07 17:24:48 gjohnson: http://lists.openembedded.org/pipermail/openembedded-core/2013-December/087879.html Jan 07 17:24:56 zibri: thanks for the info Jan 07 17:26:03 Ha, JaMa is the one that had the response :) Jan 07 17:30:10 but are there many instances where a staticdev is created but no -dev? Jan 07 17:30:41 and would it be neater to fix up those dependencies in the recipes, instead of forcing empty -dev packages? Jan 07 17:49:09 How Do I copy binay file from one ${WORKDIR} to oher ${WORKDIR} Jan 07 17:54:25 ramose: that's not how things are done. if one recipe needs to use files from another, thats what the sysroots are for. install files in recipe A's do_install, then look int he sysroot in recipe B to find them (assuming A is in its DEPENDS) Jan 07 18:04:59 ok Kergoth: I have one recipe file which is compiling and installing a bin file to one $WORKDIR but it is required by some other receipe file which use different $WORKDIR in /tmp/work Jan 07 18:05:21 Would you suggest how should I do it,any example you would like to point out Jan 07 18:05:22 again, any files installed to ${D} in do_install will automatically end up in the sysroots Jan 07 18:05:43 further, i hope that binary isn't executbale, as obviously you can't run a binary from a different arch on the host Jan 07 18:07:22 finr Kergoth: so I have to copy the requied file from sysroots dir into recipe file which required it? Jan 07 18:07:33 I don't understand the question Jan 07 18:08:02 the files will bein the ssyroots, and there are variables which point ot th esysroots, so you just need to make sure recipe B's buildsystem knows how to look in the sysroot for recipe A's files Jan 07 18:10:50 ok Kergoth ,would you like to point to some example of same kind? Jan 07 18:10:53 I got the point Jan 07 18:11:12 see the thousands of orecipes which reference STAGING_* variables Jan 07 18:11:15 grep :) Jan 07 18:12:54 OK kergoth ,will try searching with STAGING_* Jan 07 18:13:08 Am i right? Jan 07 18:25:53 kergoth: sooooon (re: dockbake) Jan 07 18:54:08 Kergoth : I just gone through the DEPENDS varibale and looks like i have to use DEPENDS in the required recipe fie(say B) ,which will take required bin file from sysroots dir Jan 07 18:54:24 that's what i told you earlier, yes Jan 07 18:55:07 but do i also need to add other than DEPENDS in required recie file(say B)? Jan 07 18:58:21 Kergoth :I mean how do I tell this particular bin file I required in recipe file B Jan 07 19:05:40 Thanks Kergoth for ur help and ideas . Jan 07 20:15:28 fray or anyone: mtahmed needs a build machine to continue with Clang & oe-core... does YP have any decent builders that are publicly accessible? Jan 07 20:27:42 Anyone here working on ellcc? http://ellcc.org/blog/?topic=adding-ellcc-as-a-yocto-layer Jan 07 20:29:20 mtahmed: also has a basic clang recipe that he is going to add to meta-tc-llvm or some repo. Jan 07 20:30:26 I suspect that ellcc is a more complete solution but I haven't and won't be able to take the time to investigate. mtahmed said he might do that when not in class. :) Jan 07 21:13:24 ellcc does sound promising for yocto, dont' know of anyone who's digging into it though **** ENDING LOGGING AT Wed Jan 08 02:59:59 2014