**** BEGIN LOGGING AT Tue Jul 17 02:59:58 2012 Jul 17 03:28:31 Hi everyone, I have been having difficulty building oe (mini2440-angstrom) The fakeroot-native_1.12.4.bb package is no longer available Jul 17 07:40:06 good morning Jul 17 08:26:02 morning all Jul 17 08:27:37 bluelightning: hi Jul 17 08:32:26 hi mckoan Jul 17 08:53:49 hi Jul 17 09:19:11 how can I in a recipe check which driver version I'm using? I need to make a conditional in a driver recipe, which depends on a kernel version Jul 17 09:21:00 basically there are some patches I only wish to add if the kernel version is less than 3.3 Jul 17 09:28:39 whats the -c command to do the rm_work step ? Jul 17 09:32:27 you need to have rm_work inherited first Jul 17 09:32:39 then -c build will also run it for you Jul 17 09:34:58 ah right .. thx Jul 17 10:03:23 build is the default task of course, so you don't actually need to specify it unless you've changed it :) Jul 17 10:03:34 rather, the default command Jul 17 10:03:56 mertsas: I'm not sure if there's a clean way, your only choice might be to have separate recipes Jul 17 10:19:34 morning all Jul 17 10:19:59 hey pb_ Jul 17 10:20:06 hi pb_, florian Jul 17 10:21:25 wow, we have udev 182 patch for oe-core. meta-oe is 'just' 186+git Jul 17 10:24:21 hopefully they'll be soon unified Jul 17 10:32:14 bluelightning: ok, so theres no clean way to get the PV from another recipe? Not that weird, but damn:P Guess I'll have to think a bit on this then. Jul 17 10:39:09 mertsas: it's better you do the check in the patch Jul 17 10:40:33 we did smthg like #include #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,31) Jul 17 10:40:43 in the 2.x epoch Jul 17 10:49:21 khem: meta-shr ready for blacklist.bbclass move.. Jul 17 11:32:09 hi guys Jul 17 11:32:32 how can i checkout a git repo into a specific directory with the SRC_URI line? Jul 17 11:33:10 e.g. if by default it extracts into a "git" directory but i want to name it differently Jul 17 11:45:46 okay i'll probably just inherit git-project instead Jul 17 11:48:27 ant_work: yeah, might have to have a look at something like that Jul 17 12:21:40 hi. with angstrom I have: Jul 17 12:21:42 STAGING_KERNEL_DIR=${STAGING_DIR_HOST}/usr/src/kernel Jul 17 12:21:45 is it normal? Jul 17 12:21:55 because my kernel is at: Jul 17 12:22:02 STAGING_KERNEL_DIR=${STAGING_DIR_HOST}/kernel Jul 17 12:22:06 oops Jul 17 12:22:15 I meant ${STAGING_DIR_HOST}/kernel Jul 17 12:23:39 oe-core's kernel.bbclass changed that lately.. so if you're using meta-oe in other build then it's different Jul 17 12:24:39 http://git.openembedded.org/openembedded-core/commit/meta/classes/kernel.bbclass?id=6125ea40d4483965f793bd847b3ce14b668a5b1e Jul 17 12:28:03 yes I know that Jul 17 12:28:07 I looked with git blame Jul 17 12:28:22 specially on bitbake.conf Jul 17 12:28:44 I rebuilt kernel and linux-libc-headers Jul 17 12:28:51 and they're still at the wrong place Jul 17 12:29:07 I'll look at kernel.bbclass and the commit has you have just given me too Jul 17 12:29:52 ah ok it seem very close to the commit I saw Jul 17 12:30:48 what should I do to get my kernel at ${STAGING_DIR_HOST}/usr/src/kernel ? Jul 17 12:35:17 the point is that there are 2 kernel.bbclass one in oe-core and one in meta-oe Jul 17 12:36:26 ahhh ok Jul 17 12:36:43 I use meta-oe Jul 17 12:37:07 just sending patches to sync them again a bit Jul 17 12:37:40 ok thanks a lot Jul 17 12:38:10 basically I need a router for this summer, and I'm converting my bug 2.0 in router (it only lack compat-wireless and hostapd-daemon) Jul 17 12:53:28 fraxinas: there is an option for that Jul 17 12:54:30 bluelightning: in fact i've already done it the other way but thx :) Jul 17 12:54:42 but i've got a different question Jul 17 12:55:09 can i easily include a file from openembedded-core from within a meta layer recipe? Jul 17 12:55:11 fraxinas: well, for reference the option for the git fetcher is destsuffix Jul 17 12:55:20 aaah okay thanks bluelightning Jul 17 12:55:30 fraxinas: yep, just specify the path starting from the base directory Jul 17 12:55:48 okay Jul 17 12:56:07 e.g: require recipes-kernel/linux-libc-headers/linux-libc-headers.inc Jul 17 12:56:13 e.g. require openembedded-core/meta/recipes-multimedia/gstreamer/gst-plugins.inc ? Jul 17 12:56:23 oops okay i get it :) Jul 17 12:56:31 fraxinas: strip off openembedded-core/meta/ from that yes Jul 17 13:02:23 thanks bluelightning Jul 17 13:13:41 am i right in the optinion that i can delete tmp/work totally but the chain keeps working as tmp/sysroots etc are still in tact ? Jul 17 13:16:58 rob_w: I wouldn't advise deleting things behind the build system's back, but in theory it should work provided nothing is in the middle of being built Jul 17 13:17:16 true Jul 17 13:17:42 i am trying to find the smallest working way so i can give away my toolchain inside a virtualbox image Jul 17 13:37:46 how large is the gcc git repository which is downloaded by recent recipe? do_fetch started 3 hours and downloaded 1.3GB so far... Jul 17 13:43:14 rob_w: you could use a pre-populated shared state cache for that... Jul 17 13:43:42 uhh is that something i got or do i need to build such ? Jul 17 13:43:59 i though that pstage and sysroots is bascicly it Jul 17 13:44:13 and stamps of course Jul 17 13:47:47 rob_w: with OE-Core it's built-in Jul 17 13:48:15 rob_w: you mention pstage though, sounds like you are still using OE-Classic ? Jul 17 13:48:22 iam using classic so to speak Jul 17 13:48:33 ah ok Jul 17 13:48:36 old school , eh ? Jul 17 13:48:56 kind of... I can understand there are people out there who can't make the switch yet Jul 17 13:49:30 yeah but i need to get into it someday .. maybe thats the chance now to achieve this Jul 17 13:49:36 if there is anything other than inertia and working on an existing product that's stopping you please let us know however Jul 17 13:50:25 no, all cool .. its just the existing projecests where you dont want to trhough away your good old chain Jul 17 13:50:53 its even from the glibc days , rather then a eglibc ;-) Jul 17 13:51:28 sure, of course Jul 17 13:51:30 :) Jul 17 13:51:57 but i works like a charm ;-) Jul 17 13:52:08 kthxbye Jul 17 13:59:28 ensc|w: du -sh tells me it should only be about 300MB uncompressed... compressed it's only about 23MB Jul 17 14:00:12 ensc|w: not sure how efficient svn's protocol is but 1.3GB sounds like quite an inflation :( Jul 17 14:07:50 bluelightning: is using git now; svn was used in older -core Jul 17 14:12:13 ah, my mistake... I was looking at eglibc Jul 17 14:12:17 sorry Jul 17 14:20:14 is there a GettingStarted wiki for that oe-core n friends ? Jul 17 14:25:33 nvm Jul 17 14:25:39 http://www.openembedded.org/wiki/Getting_Started_with_OE-Core Jul 17 14:26:37 got it Jul 17 14:26:38 thx Jul 17 15:06:06 JaMa|Off: cool Jul 17 16:02:54 obi ich hab das hingekriegt (geht mit relativem pfad ausgehend von openembedded-core/meta/. Jul 17 16:31:54 JaMa|Off, hi again Jul 17 16:32:09 I've that now: Jul 17 16:32:20 ERROR: QA Issue: Architecture did not match (40 to 62) on /work/bug20-angstrom-linux-gnueabi/linux-bug20-linaro-2.6.35-r123+gitr37efe9e3ef0e12a207a92da21f28fe7d295c4aba/packages-split/kernel-dev/usr/src/kernel/scripts/kconfig/kxgettext.o Jul 17 16:32:34 still I imported the part that rmed the script files Jul 17 16:33:28 http://pastie.org/private/q7qexuj6mzzeog4o9ka8ww is what I did Jul 17 18:48:17 is there a way to remove automatically the temporary files of old packages? 'rm_work' does not remove 'package' + 'packages_split' which fill the disk very fast when building some large packages Jul 17 18:49:54 ensc|w, yes there is Jul 17 18:50:07 ensc|w, in SHR we have rm_old_work Jul 17 18:50:17 ensc|w, so look in meta-smartphone for the class Jul 17 18:50:27 and send a patch for adding it in meta-oe or oe-core Jul 17 18:54:30 GNUtoo-desktop: thx; but this class seems to assume another directory layout (work/.../${PN}/${PF} instead of default work/.../${PF}) Jul 17 18:54:41 change that then.... Jul 17 19:17:18 it doesn't assume different layout, it sets diferent layout so it should work.. but inherit it only after finishing build Jul 17 19:18:30 JaMa, hi Jul 17 19:18:34 hi Jul 17 19:19:01 JaMa, did you see my message where the merge of both kernel.bbclass files makes kernel build fails for me? Jul 17 19:19:08 GNUtoo-desktop: that file shouldn't be packaged with http://patchwork.openembedded.org/patch/32253/ Jul 17 19:19:09 IRC message Jul 17 19:19:31 ah ok Jul 17 19:19:50 thanks a lot Jul 17 19:20:02 did you apply all patches for meta-oe's kernel.bbclass? Jul 17 19:20:09 no Jul 17 19:20:13 I just merged both files Jul 17 19:20:16 with meld Jul 17 19:20:52 do you should have this change too Jul 17 19:21:01 yes I'll add this change Jul 17 19:21:25 let me look if I've got it Jul 17 19:21:39 it was in oe-core's version already Jul 17 19:24:17 ah ok Jul 17 19:26:08 thanks a lot!!! Jul 17 20:26:16 | make: *** .../sysroots/bug20/usr/src/kernel: No such file or directory. Stop. Jul 17 20:26:18 hmmm Jul 17 20:26:26 strange.... Jul 17 20:27:31 now the kernel moved who knows where Jul 17 20:30:04 ok it's in kernel-dev Jul 17 20:30:15 but I wonder if it's staged.... Jul 17 20:32:47 ahhh Jul 17 20:32:49 kernel-misc Jul 17 20:32:58 that's where staging stuff goes Jul 17 20:33:02 in meta-oe Jul 17 20:33:20 GNUtoo-desktop: JaMa posted few patches for meta-oe Jul 17 20:33:29 for kernel.bbclass Jul 17 20:33:31 cherry pick those Jul 17 20:35:26 ok thanks Jul 17 20:40:38 ~seen woglinde Jul 17 20:40:52 woglinde <~heinold@f052066209.adsl.alicedsl.de> was last seen on IRC in channel #oe, 2d 21h 41m 31s ago, saying: 'good nite'. Jul 17 20:57:32 howdy Jul 17 20:57:41 woglinde: hey Jul 17 20:57:49 I need your help Jul 17 20:58:11 khem whats up? Jul 17 20:58:19 I am having troubles with meta-java Jul 17 20:58:33 giflib mail? Jul 17 20:58:40 does it need bash to be your default shell ? Jul 17 20:58:45 woglinde: that is sorted Jul 17 20:59:03 woglinde: btw. what other deps does it need from meta-oe besided giflib Jul 17 20:59:39 woglinde: my default shell it dash and I am having all sort of issues Jul 17 21:00:15 oh we have to much bashims? Jul 17 21:00:34 angstroem forces you to use other shell than dash Jul 17 21:01:52 yeah Jul 17 21:02:11 woglinde: so now my problem is that virtual:native:/yocto/poky/meta-java/recipes-core/xerces-j/xerces-j_2.9.1.bb is failing Jul 17 21:02:18 woglinde: I will post the compile log Jul 17 21:02:38 okay Jul 17 21:02:53 with which distribution you are using oe-meta? Jul 17 21:03:09 no angstrom no meta-oe Jul 17 21:03:27 http://paste.ubuntu.com/1097314/ Jul 17 21:03:38 mine is poky Jul 17 21:04:17 NOTE: multiple providers are available for runtime java2-runtime (openjdk-6-jre, jamvm) Jul 17 21:04:18 uhm what did you choose as javac-native? Jul 17 21:04:20 NOTE: consider defining a PREFERRED_PROVIDER entry to match java2-runtime Jul 17 21:04:37 sure Jul 17 21:04:39 PREFERRED_PROVIDER_virtual/java-native = "jamvm-native" Jul 17 21:04:40 PREFERRED_PROVIDER_virtual/javac-native = "ecj-bootstrap-native" Jul 17 21:04:40 PREFERRED_PROVIDER_java2-runtime = "openjdk-6-jre" Jul 17 21:04:40 PREFERRED_VERSION_openjdk-6-jre = "6b24-1.11.1" Jul 17 21:04:40 PREFERRED_VERSION_icedtea6-native = "1.8.11" Jul 17 21:04:41 distro has to choose Jul 17 21:05:03 Do I need to set something more ? Jul 17 21:05:11 yes Jul 17 21:05:15 ah Jul 17 21:05:21 I just followed readme Jul 17 21:05:23 preffered version of jamvm to git Jul 17 21:05:30 hmmm Jul 17 21:05:41 yes I need to add something more to the readme Jul 17 21:06:20 PREFERRED_VERSION_jamvm = "git" Jul 17 21:06:20 PREFERRED_VERSION_jamvm-native = "git" Jul 17 21:06:24 is that ok ? Jul 17 21:07:20 does this work? Jul 17 21:07:30 I thought you need the complete git revision Jul 17 21:07:46 I havent dug down deep into bashism Jul 17 21:07:49 1.5.5+1.6.0-devel+git% Jul 17 21:07:53 prolly Jul 17 21:11:53 woglinde: hmm now it seeems to be progressing lets see Jul 17 21:12:08 woglinde: wait no Jul 17 21:12:50 NOTE: preferred version 1.5.5+1.6.0-devel+git of jamvm-native not available (for item virtual/java-native) Jul 17 21:12:53 NOTE: versions of jamvm-native available: 1.5.5+1.6.0-devel+git1+4617da717ecb05654ea5bb9572338061106a414d Jul 17 21:14:56 khem sorry I had to sing to the baby Jul 17 21:15:04 otherwise it would not sleep Jul 17 21:15:13 sure Jul 17 21:15:17 by all means Jul 17 21:15:23 I know those things very well Jul 17 21:15:28 :) Jul 17 21:17:02 so whats the problem now? Jul 17 21:17:09 woglinde: OK now I set it to 1.5.5+1.6.0-devel+git% it works better Jul 17 21:17:09 and whats the ecj version? Jul 17 21:17:34 ah okay Jul 17 21:17:36 ecj-bootstrap-native-1.0 Jul 17 21:17:38 its only one Jul 17 21:18:09 btw. libecj-bootstrap_3.6.2.bb doesnt work Jul 17 21:18:15 libecj-bootstrap_3.8.bb Jul 17 21:18:19 is what I changed it to Jul 17 21:18:30 upstream uri is gone Jul 17 21:18:52 plus the new libecj-bootstrap_3.8 is in jar form not in zip as libecj-bootstrap_3.6.2 was Jul 17 21:18:58 I hope thats not an issue Jul 17 21:19:18 uh? Jul 17 21:19:28 only changed download Jul 17 21:19:38 3.8 is maybee not working Jul 17 21:19:44 I renamed the recipe Jul 17 21:19:52 and changed the SRC_URI Jul 17 21:19:55 but thanks for the hint Jul 17 21:20:00 I need to update it Jul 17 21:20:29 Hi guys, a (quick) patching question: I've got a patch that applies cleanly in isolation, but when I try applying it after all the other patches specified by the kernel recipe it fails because the lines are too far off. I could just edit the context but I suspect there is a proper way to do this. Jul 17 21:20:55 git mergetool Jul 17 21:21:01 woglinde: here is patch for update Jul 17 21:21:27 khem hm yes I missed that I still have ibiblio in Jul 17 21:21:37 instead of archive.eclipse Jul 17 21:21:42 I suspect that if I added it to the recipe the bitbake machinery would handle it, but I'm not sure the code will compile with it phase yet. Jul 17 21:21:57 spacecolonyone o.O Jul 17 21:22:03 what you are talking about Jul 17 21:22:31 Trying to get LI-5M03 support on 3.2.18 Jul 17 21:23:04 woglinde: can you update the src uri and push ? Jul 17 21:23:08 so I can get the change Jul 17 21:23:22 Max Galemin has a great blog about it, but he is using buildroot so I'm trying to incorporate some of his patches from 3.2.8 and see how far I get Jul 17 21:23:22 woglinde: is it an issue to use 3.8 instead Jul 17 21:24:53 khem 3.7 did not work properly to compile openjdk Jul 17 21:24:54 woglinde: xerces-j-native is still puking on me Jul 17 21:25:01 3.8 I did not test Jul 17 21:25:04 woglinde: ok Jul 17 21:25:05 you see Jul 17 21:25:07 I'm trying to work off http://blog.elphel.com/2009/12/openembeddedangstrom-kernel-workflow/ but the patch wont apply after bitbake virtual/kernel -c configure because of the number of other patches already patched in, or so I gather Jul 17 21:25:09 ecj problem Jul 17 21:25:13 try 3.6.2 Jul 17 21:25:22 so lets update the src uri to fetch 3.6.2 from archive Jul 17 21:25:30 whats the archive address Jul 17 21:25:38 spacecolonyone yes so fix your patch Jul 17 21:25:54 khem mom I will push it after testing in a minute Jul 17 21:26:23 k Jul 17 21:28:19 woglinde is it "proper" to just muck with the line number context in this situation. I figured there is a proper way to do this with aptch or git, essentially saying I want to apply patch A to file B, be advised patches C,D,E have been applied and take that into accout when merging Jul 17 21:28:35 s/aptch/patch Jul 17 21:28:53 spacecolonyone git am Jul 17 21:28:59 than git mergetool Jul 17 21:29:08 bitte Jul 17 21:32:00 woglinde: we also need http://paste.ubuntu.com/1097385/ Jul 17 21:37:00 khem uhm why? Jul 17 21:37:17 otherwise in dash its copying the file to an non existing dir Jul 17 21:37:34 may be its sorting order Jul 17 21:37:36 normaly javac creates the dir Jul 17 21:37:37 yes Jul 17 21:37:43 oh Jul 17 21:37:49 so may be I have bigger issue Jul 17 21:37:52 but sorting order yes Jul 17 21:38:07 we had a problem with it in other package Jul 17 21:39:24 khem libecj is fixed Jul 17 21:39:28 k Jul 17 21:39:32 let me update Jul 17 21:39:48 I will search for the sort fix Jul 17 21:41:27 multiple providers are available for runtime java2-runtime (jamvm, openjdk-6-jre) Jul 17 21:41:34 I dont know why this si happening Jul 17 21:42:18 PREFERRED_PROVIDER_virtual/java-native = "jamvm-native" Jul 17 21:42:18 PREFERRED_PROVIDER_virtual/javac-native = "ecj-bootstrap-native" Jul 17 21:42:18 PREFERRED_PROVIDER_java2-runtime = "openjdk-6-jre" Jul 17 21:42:52 you need PREFERRED_PROVIDER_openjdk-6-jre = "openjdk-6-jre" Jul 17 21:43:17 than it should work Jul 17 21:44:03 khem problem with xml-resolver is that you might had a javac which didnot honored -d Jul 17 21:44:05 right Jul 17 21:44:19 which creates all class directories normaly Jul 17 21:45:18 I see Jul 17 21:45:35 after your commit things are progressing lets see where it beak Jul 17 21:45:49 sure Jul 17 21:46:01 did you made another systemd test? Jul 17 21:51:05 woglinde: next in line Jul 17 21:55:27 khem? Jul 17 21:58:58 woglinde: I mean systemd Jul 17 21:59:07 woglinde: it looks good Jul 17 21:59:13 progressing well thus far Jul 17 21:59:19 so I guess key is to use bash for now Jul 17 21:59:26 until we figure out the real problem Jul 17 21:59:38 other issues seems to work ok now Jul 17 22:18:33 okay Jul 17 22:22:43 woglinde: here is another problem Jul 17 22:23:11 http://paste.ubuntu.com/1097457/ Jul 17 22:24:32 here a new issue..after git pull, curl-native is failing (no libtasn). I see libtasn-native is building on another BB_THREAD... Jul 17 22:25:59 solved rebuilding it Jul 17 22:30:59 khem hm classpath is not automake 1.12 compatible? Jul 17 22:31:31 hmm Jul 17 22:31:46 there is a 0.99 release Jul 17 22:32:06 woglinde: should it be safe to try that ? Jul 17 22:32:15 should be Jul 17 22:32:19 look at the patches Jul 17 22:33:12 hm I could try it here too Jul 17 22:34:17 woglinde: http://gumstix.8.n6.nabble.com/mkdir-p-Command-not-found-td647494.html Jul 17 22:36:02 khem yes I know normaly of mkdir_p Jul 17 22:39:02 woglinde, I looked into git-am and merge tool and I'm not really seeing how they fit in to what I was asking earlier. Jul 17 22:39:33 Following that elphel post for I can just use quit and make the changes to the code manually following Max's work, but then the patches aren't passed off of a git commit (the 3.2.18 commit to be precise), right? Rather they are patches to a heavily patched 3.2.18 kernel based off the meta-ti recipe. Jul 17 22:39:42 woglinde: we need something like http://permalink.gmane.org/gmane.comp.handhelds.openembedded.core/23413 Jul 17 22:40:13 decalre your last port of entry, space wanderer... Jul 17 22:40:20 *declare even Jul 17 22:40:23 I'd like to get this right and submit it eventually, which is why I'm asking. If #beagle or #angstrom would be more appropriate let me know Jul 17 22:40:36 khem anyway it should notbuild classpath-target Jul 17 22:40:50 unless you specified it Jul 17 22:41:24 woglinde: I am doing bitbake openjdk-6-jre Jul 17 22:41:37 yes Jul 17 22:41:40 so if its getting pulled then its some implicit dep Jul 17 22:41:49 please look at the build path Jul 17 22:41:59 why it needs classpath-target Jul 17 22:42:06 spacecolonyone: what's the problem? Jul 17 22:42:10 no idea, I thought you would know :) Jul 17 22:42:48 bitbake -g Jul 17 22:43:01 okay good night Jul 17 22:43:24 woglinde: yes I am doing that Jul 17 22:45:10 denix, brb, ravioli and boiling water on stove Jul 17 22:46:27 that is definitely a problem for another channel... :) Jul 17 22:48:34 ugh random aside), are any of you on comcast and have(had) an issue where you can't maintain a steady irc connection Jul 17 22:49:31 ah jamvm is needed classpath Jul 17 22:50:57 ok denix, so I'm trying to patch in support for the patina sensor on the leopard imaging module for the BBxM I've got some patches that Max Galemin has used to get it working in 3.2.8 with buildroot and they apply cleanly to a checkout of mainline 3.2.18 Jul 17 22:52:19 but fail after bitbake virtual/kernel -c configure has run. As best I can tell this is since the file line numbers have changed too much from all the patches specified in the meta-ti recipe being applied automatically. Jul 17 22:53:14 I see a number of ways to proceed from here, but I've got no clue what would be considered best practices Jul 17 22:53:43 spacecolonyone: you can try disabling BB-specific patches temporarily - that will give you pristine 3.2.18 Jul 17 22:54:17 or if you already have a patch against 3.2.18, try inserting it in the patch list before BB-specific patches Jul 17 22:54:23 wouldn't that open a whole nother can of worms? Jul 17 22:55:01 Yea, but since this is my development, shouldn't I do that in a custom layer? Jul 17 22:55:05 could be, but worth a try Jul 17 22:55:07 http://git.yoctoproject.org/cgit/cgit.cgi/meta-ti/tree/recipes-kernel/linux/linux-ti33x-psp_3.2.bb Jul 17 22:55:19 around line 1179 Jul 17 22:56:00 yea? Jul 17 22:56:26 that's where you can try inserting your patch temporarily to see if it indeed conflicts with BB-specific patches Jul 17 22:56:55 if it still doesn't work, then it conflicts with TI PSP 3.2 kernel, unfortunately :( Jul 17 22:57:13 you'd have to massage it yourself, then Jul 17 22:57:32 I think the recipe I'm building against was linux-mainline_3.2.18.bb Jul 17 22:58:06 rather, I think that is what angstrom defaults too Jul 17 22:58:08 to* Jul 17 22:58:32 http://git.yoctoproject.org/cgit/cgit.cgi/meta-ti/tree/recipes-kernel/linux/linux-mainline_3.2.bb Jul 17 22:59:49 ok, so I don't think they conflict so much as it is simply that the line number for insertion of the code is moved too much by the other patches Jul 17 23:00:58 spacecolonyone: ah, BBxM, I thought about the BeagleBone... Jul 17 23:02:27 dangit, ok, denis I'm going to give that a whirl, then if it doesn't work I'll try what the elphel post said about quilt Jul 17 23:03:15 oh, what is the deal with the psp3.2 kernel Jul 17 23:03:22 i take it this is new? Jul 17 23:11:19 spacecolonyone: that's for the Bone, not XM Jul 17 23:11:36 ok, thought it like a bit different Jul 17 23:12:12 you are correct, BeagleBoard-xM uses linux-mainline recipe, but there are still BB-specific patches Jul 17 23:12:54 yea, the thing that still confuses me a bit is that it seems like the way bit bake apples them doesn't lead to a local commit Jul 17 23:13:39 by default, it doesn't use git-am, but rather quilt to apply patches Jul 17 23:14:13 so they are unstaged changes, which makes me a little confused about how I should generate a patch for submission when I finally get stuff ready, but I'll cross that bridge when I come to it Jul 17 23:14:41 easy, git-commit them and then git-format-patch Jul 17 23:15:20 but that commit will include all of those patches that were applied by bit bake with quilt + my work Jul 17 23:16:14 doesn't the patch I submit need to be based off of the 3.2.18 commit that the recipe uses as a base? Jul 17 23:16:15 do that 2 times Jul 17 23:17:22 bitbake -c configure, git-commit = baseline, then apply your patch, git-commit, git-format-patch Jul 17 23:17:44 ok, was just typing that out, lol Jul 17 23:18:05 then I'd submit that last patch? Jul 17 23:18:19 and could also add it at the end of the list in the recipe Jul 17 23:18:50 right Jul 17 23:19:13 ok, great, now to go see if I can actually get it to work! Jul 17 23:20:08 good luck **** ENDING LOGGING AT Wed Jul 18 02:59:58 2012