**** BEGIN LOGGING AT Thu Sep 20 02:59:57 2012 Sep 20 09:16:09 Hello everyone. I have a question concerning quickly rebuilding *one* kernel module. What is the "bitbake" command to just rebuild, say, the "rotary-encoder" driver? Sep 20 09:16:15 morning all Sep 20 09:16:31 erwin: unless it's an external module, you can't really do that Sep 20 09:17:39 erwin: you can force a build of the kernel tree after you have changed it behind the scenes, that will result in the module being rebuilt and packaged if its source has changed Sep 20 09:22:08 Thanks, bluelightning Sep 20 09:23:34 hi all Sep 20 09:25:08 If I do a "bitbake -c compile -f virtual/kernel" as I'm used to do (getting the kernel uImage in my arch/arm/boot directory), the ko modules are not rebuilt. It seems I have to do a full package generation (bz2), which takes time. Is there a shorter way? Sep 20 09:28:32 hi pb_ Sep 20 09:30:08 erwin: ah, I think you may need to do bitbake -c compile_kernelmodules -f virtual/kernel Sep 20 09:53:39 bluelightning, you made my day. It compiles all .ko files in the driver source directories. Takes time the first time, but after it just recompiles the module I'm working on. Exactly what I needed, thanks! Sep 20 09:54:29 erwin: no problem :) Sep 20 11:26:29 looks like amount of patches from me will require pull request rather then sending yet another single patch Sep 20 11:27:39 dont know Sep 20 11:29:07 will collect them (most are harmless for existing archs) and then send one patch series Sep 20 11:41:56 db, ncurses, slang, byacc, xinetd so far Sep 20 12:01:20 hum is there some way to make a bit of python run each time a recipe is actually going to be built?? Sep 20 12:05:13 ? Sep 20 12:06:11 for eg, an event handler to be called when the recipe is going to be built (regardless of what tasks from the recipe are going to be executed) Sep 20 12:07:37 what should that be usefull for? Sep 20 12:08:36 for setting a DEFAULT_PROVIDER depending on the recipe that is being worked on Sep 20 12:09:47 the BuildStarted event would probably be useful for that.. however, i only seem to get RecipePreFinalise, RecipeParsed and TaskSucceeded events, no other types of events Sep 20 12:09:49 ???? Sep 20 12:10:49 when i define a bitbake event handler in my recipe Sep 20 12:27:48 or let me rephrase: what i'm trying is to see what build target is given on the bitbake command line, and set PREFERRED_PROVIDER_virtual/kernel based on that Sep 20 12:27:54 from within the recipe Sep 20 12:39:15 gmc: I have to say it doesn't really sound like the right approach Sep 20 12:43:12 at all, or my specific idea of implementing it? Sep 20 12:54:24 gmc: I think your requirement is going to be tricky to implement, but if it is possible I don't think this is the way to do it Sep 20 12:55:44 actually, thinking about it, it won't work as i planned it.. since the event would be generated AFTER the recipes have been parsed and so setting the preferred provider would be too late Sep 20 12:58:05 right, that would be one issue Sep 20 12:58:50 after adding a module_autoload, must the kernel be rebuilt? Sep 20 12:59:36 i guess i am going to give up on this requirement... i've been digging through bitbake for hours already Sep 20 12:59:54 did gain a lot of understanding about how bitbake works, but alas didn't reall ytfind a way to do what i want Sep 20 13:00:42 any other ways of dealing with multi-machine setups besides having two oe-core trees ? Sep 20 13:01:59 gmc: why have two oe-core trees? can you not just have two separate machine config files? Sep 20 13:02:21 I know you were trying to avoid that but that's still a better solution than entirely separate metadata trees Sep 20 13:04:26 uhm yes, that.. :) although that still gives issues with PREFERRED_PROVIDER being cached in between changes to local.conf, or? Sep 20 13:06:47 hi guys Sep 20 13:06:58 need to clear one thing about @ operator Sep 20 13:07:07 I am trying TUNE_LDARGS += "${@'-m ${LDEMULATION}' if LDEMULATION else ''}" Sep 20 13:07:39 if LDEMULATION is not define what is the expected behaviour Sep 20 13:07:40 ? Sep 20 13:08:20 I am getting EOL while scanning string literal (TUNE_LDARGS, line 1) Sep 20 13:08:47 I thought if LDEMULATION will handle it as it is checking that LDEMULATION exists or not Sep 20 13:29:36 Noor: perhaps rewrite it with base_contains or base_conditional ? Sep 20 13:32:33 gmc: it wokrs fine when I define LDEMULATION == "" and LDEMULATION_mips64 = "some options" Sep 20 13:32:47 *LDEMULATION = "" Sep 20 13:33:17 previously LDEMULATION = "" was missing .... so just want to understand 2 things Sep 20 13:34:16 first when we define a variable with override it looks that variable is not defined without override by default ... is it correct Sep 20 13:35:00 secondly usage of LDEMULATION in if doesn't say that is it is defined .... does it has another meaning in this context Sep 20 13:38:24 Is there a way to build libx11 even though you removed x11 from DISTRO_FEATURES? Sep 20 13:39:56 fcooper: I think not coz some recipes checks that x11 is present in DISTRO_FEATURE or not and add things in depends Sep 20 14:55:53 yesterday when I compileed my hello world.c program a.out appeared on the desktop but i couldn't get it to run, then I was told to run ./a.out and it ran fine. but now I've deleted it, recompiled it and it will run from anywhere with just a.out. Why/ Sep 20 14:57:07 It's not like desktop shows up when I i run PATH Sep 20 14:57:10 the directory is included in $PATH Sep 20 14:59:02 check it bash: /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games: No such file or directory Sep 20 15:02:01 sorry wrong room, I was told yesterday, probabbly gonna get banned before I ever get here. Sep 20 15:15:55 has nothing todo with oe Sep 20 15:25:18 woglinde: hi Sep 20 16:02:23 he zecke Sep 20 16:31:27 JaMa: hmm, somehow the backtrace for 3133 has gone away :( **** ENDING LOGGING AT Fri Sep 21 02:59:58 2012