**** BEGIN LOGGING AT Fri Mar 25 02:59:57 2011 Mar 25 05:37:14 I have a build that says my do_configure() depends on a do_populate_sysroot() for the same package. How do I find out why? I've looked through the files that make up the recipe, but can see nothing that would do this. **** ENDING LOGGING AT Fri Mar 25 07:03:19 2011 **** BEGIN LOGGING AT Fri Mar 25 07:04:14 2011 **** ENDING LOGGING AT Fri Mar 25 07:16:35 2011 **** BEGIN LOGGING AT Fri Mar 25 07:18:14 2011 Mar 25 07:36:49 03Siddharth Heroor  07org.openembedded.dev * r8d10a72e59 10openembedded.git/recipes/ti/ti-xdctools_3.20.08.88.bb: Mar 25 07:36:49 TI XDCTools: Add version 3.20.08.88 Mar 25 07:36:49 Signed-off-by: Siddharth Heroor Mar 25 07:36:49 Signed-off-by: Koen Kooi Mar 25 07:52:20 git over git:// seems down, ka6sox-away ? Mar 25 07:55:12 JaMa|Wrk, okay will look. Mar 25 07:56:54 thanks Mar 25 08:17:22 JaMa|Wrk, whats it not doing? Mar 25 08:17:28 no checkout or pull or what? Mar 25 08:17:45 its like really busy Mar 25 08:20:57 there are 75 uploads happening now. Mar 25 08:21:07 all being packed up. Mar 25 08:23:11 03Koen Kooi  07org.openembedded.dev * rec37fbd570 10openembedded.git/recipes/x-load/x-load_git.bb: Mar 25 08:23:11 x-load git: update to 1.5.0 Mar 25 08:23:11 Tested on beagle-xm/angstrom and pandaboard/angstrom Mar 25 08:23:11 Signed-off-by: Koen Kooi Mar 25 08:28:34 ka6sox-away: fatal: read error: Connection reset by peer, while git+ssh works fine Mar 25 08:28:43 during pull Mar 25 08:29:12 okay jas...thanks Mar 25 08:33:54 mornin Mar 25 08:37:50 ka6sox-away: works now, thanks Mar 25 08:38:51 yw Mar 25 08:39:11 its just overloaded right now...all of .eu is banging pretty hard. Mar 25 08:39:55 ka6sox-away: ok, right, it doesn't work again (I was just lucky once :)) Mar 25 08:40:41 ya,its like 39 connections and thrashing Mar 25 08:40:57 so there are a lot of changesets updating Mar 25 08:41:31 koen's was the first one to get thru Mar 25 08:53:07 I mean I can restart the server but basically there are 64 folks using the git server and its taking a long time to process all this. Mar 25 08:53:38 34 are uploading changes Mar 25 08:54:03 oh my god they used reiserfs on our internal oebuilder Mar 25 08:54:09 * woglinde grumbles Mar 25 08:57:25 hi obi Mar 25 08:57:38 hi woglinde Mar 25 09:22:59 03Martin Jansa  07master * r181c4ed14b 10openembedded.git/recipes/mc/mc.inc: Mar 25 09:22:59 mc: add RRECOMMENDed mc-skins package and more files to FILES_PN Mar 25 09:22:59 * fixes SHR bug http://trac.shr-project.org/trac/ticket/1348 Mar 25 09:22:59 * and not packaged files Mar 25 09:22:59 Signed-off-by: Martin Jansa Mar 25 09:23:11 03Martin Jansa  07master * r76e1134237 10openembedded.git/recipes/tasks/task-shr-feed.bb: Mar 25 09:23:11 task-shr-feed: add subversion, cvs Mar 25 09:23:11 * fixes SHR bug http://trac.shr-project.org/trac/ticket/1349 Mar 25 09:23:11 Signed-off-by: Martin Jansa Mar 25 09:23:13 03Martin Jansa  07master * r61f9ea4223 10openembedded.git/recipes/freesmartphone/fsoaudiod_git.bb: Mar 25 09:23:13 fsoaudiod: add asound.conf to CONFFILES and RCONFLICTS with alsa-state Mar 25 09:23:14 * fixes SHR bug http://trac.shr-project.org/trac/ticket/1350 Mar 25 09:23:14 Signed-off-by: Martin Jansa Mar 25 09:32:26 morning all Mar 25 09:37:16 kill -9 git-daemon Mar 25 09:58:44 03Roger Monk  07org.openembedded.dev * r93548f06c1 10openembedded.git/conf/machine/omap4430-panda.conf: Mar 25 09:58:44 omap4430-panda: Prefer u-boot-git and remove extra custom x-load Mar 25 09:58:44 * Switch to using u-boot-git (instead of u-boot-sakoman) Mar 25 09:58:44 * Remove reference to custom x-load Mar 25 09:58:44 * Remove extra image depends (which will come from common include) Mar 25 09:58:45 Signed-off-by: Roger Monk Mar 25 09:58:45 Signed-off-by: Koen Kooi Mar 25 09:58:54 03Roger Monk  07org.openembedded.dev * r8f1b80d918 10openembedded.git/recipes/u-boot/ (2 files in 2 dirs): Mar 25 09:58:54 u-boot-git: Add omap4430-panda to u-boot git recipe Mar 25 09:58:54 * Enable u-boot git to build for pandaboard Mar 25 09:58:54 Signed-off-by: Roger Monk Mar 25 09:58:54 Signed-off-by: Koen Kooi Mar 25 10:01:30 ka6sox-away: ping Mar 25 10:02:56 hi xora Mar 25 10:03:03 hey woglinde Mar 25 10:04:59 morning XorA Mar 25 10:05:11 (middle of the night here) Mar 25 10:05:15 ka6sox-away: noticed you were still talking at insane hours :-) Mar 25 10:05:46 trying to see if your git server would process all the changesets without falling over. Mar 25 10:06:29 down from 64 leechers|34 uploaders to 43/11 Mar 25 10:07:40 there were 34 peoples doing git push? Mar 25 10:07:59 I didnt think we had that many active developers Mar 25 10:08:02 34 upload|pakcks Mar 25 10:08:19 er 34 upload|packs simultaneous Mar 25 10:08:32 thats a lot of work to sort that out. Mar 25 10:08:37 yeah Mar 25 10:09:10 git isnt the most efficient of SCMs brings my VM to a halt quite regular and there are only 4 pushers Mar 25 10:09:25 right...and we have a LOT More than that. Mar 25 10:09:39 we *need* to get some RO mirrors Mar 25 10:09:49 we are dying. Mar 25 10:10:14 waer cool wenn du es mergest, ich acke dann einfach Mar 25 10:10:15 there was discussion on a official github mirror, I cant remember if that happened Mar 25 10:10:17 ooops wrong window Mar 25 10:10:32 if it did we should change the wiki to point to github for all newbs Mar 25 10:10:35 if it wasn't seeing 30+ pulls at the same time all from .eu Mar 25 10:11:01 I have accounts for the project for both github and gitoriuos Mar 25 10:11:23 we need hooks to push changes out to them after commits on git.oe.org Mar 25 10:12:01 git hooks are still a mystery to me Mar 25 10:12:01 similar to the CIA hooks Mar 25 10:12:08 me too Mar 25 10:12:21 * XorA should write some for the angstrom mirrors Mar 25 10:12:39 ya, thats another "interesting" one. Mar 25 10:12:47 82GB of mirror Mar 25 10:13:15 thats the used Angstrom diskspace? Mar 25 10:13:29 yes, for the local mirror @ osuosl Mar 25 10:13:47 its so big I can't put it on the OSUOSL mirrors :( Mar 25 10:14:02 ok, Ill talk to koen about that, what do we need to get it down to? Mar 25 10:14:10 we can't Mar 25 10:14:17 he is using it for narcissus too Mar 25 10:14:55 I need to push it over to the package server and maintain the backend links Mar 25 10:15:03 apache sucks for serving up this stuff Mar 25 10:15:25 so I have a package server with nginx and varnish Mar 25 10:15:51 it already serves up 120+GB already so whats another 82GB Mar 25 10:16:53 I also need to move the sources off of melo(again apache) and move them over to the package server Mar 25 10:17:13 thats another 30GB or so. Mar 25 10:19:51 03Roger Monk  07org.openembedded.dev * r6902570b88 10openembedded.git/recipes/gstreamer/gst-plugins-good_0.10.28.bb: Mar 25 10:19:51 gst-plugins-good: update to 0.10.28 Mar 25 10:19:51 Signed-off-by: Roger Monk Mar 25 10:19:51 Signed-off-by: Koen Kooi Mar 25 10:22:33 03Koen Kooi  07org.openembedded.dev * r0d9d17c9e8 10openembedded.git/recipes/u-boot/u-boot_git.bb: Mar 25 10:22:34 u-boot git: remove duplicate filesdir line Mar 25 10:22:34 Signed-off-by: Koen Kooi Mar 25 10:23:41 hm Mar 25 10:23:46 bitbake git dead Mar 25 10:23:48 *g* Mar 25 11:17:05 hm wtf here populate_sysroots freezes now Mar 25 12:30:28 hey. i have a git repository for bitbake pointing at git://git.openembedded.org/bitbake.git. Was the bitbake git repo moved? Mar 25 12:30:54 i am trying to fetch and i get fatal: read error: Connection reset by peer Mar 25 12:31:30 ka6sox did something Mar 25 12:33:47 well, more i wanted to ask if it's my connection problem or the git repo moved Mar 25 12:34:10 connection Mar 25 12:34:21 or rather server problme Mar 25 12:34:27 server problem Mar 25 12:35:32 woglinde: but if your connection get's us mirror you can be more lucky ;) Mar 25 12:36:11 gm crofton Mar 25 12:36:33 okay; so i guess it will get back online in a while Mar 25 12:36:45 thanks for the reply Mar 25 12:38:36 later than sooner Mar 25 12:51:32 hello Mar 25 12:52:20 I m build a file for mini2440.... but I m getting an error during wvstreams task .... http://pastebin.mozilla.org/1191770 Mar 25 12:52:25 http://pastebin.mozilla.org/1191770 Mar 25 12:52:33 plz someone help me... Mar 25 12:53:43 | ipstreams/wvunixdgsocket.cc:50:32: error: 'chmod' was not declared in this scope Mar 25 12:53:46 fix it Mar 25 12:54:00 aloril: is there #include in that file? Mar 25 12:54:09 JaMa its c++ Mar 25 12:54:35 so please no c header until the c++ suckz Mar 25 12:55:58 git version has it in #ifdef MACOS already Mar 25 12:56:12 *g* Mar 25 12:57:27 how to I fix this error ipstreams/wvunixdgsocket.cc:50:32: error: 'chmod' was not declared in this scope Mar 25 12:57:35 so it's missing in wvstreams / include / wvunixdgsocket.h Mar 25 12:58:03 jama ? Mar 25 12:58:16 woglinde: https://github.com/wlach/wvstreams/blob/master/include/wvunixdgsocket.h Mar 25 12:59:18 but it is present .... in wvstreams-4.6/include/wvunixdgsocket.h Mar 25 13:00:46 #include "wvunixdgsocket.h" is wrong too Mar 25 13:00:58 in my opinion Mar 25 13:01:39 so should I try ... #include Mar 25 13:02:15 sanket no Mar 25 13:02:19 that wont help Mar 25 13:02:30 you need the right header inclusion Mar 25 13:02:33 as in git Mar 25 13:02:44 version Mar 25 13:02:54 so how should I solve this error Mar 25 13:03:15 make recipe for git version Mar 25 13:03:23 or apply the fix to wvunixdgsocket.h Mar 25 13:05:34 hm Mar 25 13:05:40 its not fixed there Mar 25 13:06:01 so #include in wvunixdgsocket.h Mar 25 13:06:06 as jama suggested Mar 25 13:10:20 03Matt Rice  07org.openembedded.dev * r6a40f720fc 10openembedded.git/contrib/angstrom/omap3-mkcard.sh: Mar 25 13:10:21 angstrom omap3 mkcard: improve loop mount handling Mar 25 13:10:21 Signed-off-by: Koen Kooi Mar 25 13:13:08 now I had got other errors.... http://pastebin.mozilla.org/1191787 Mar 25 13:14:11 # Mar 25 13:14:12 streams/wvatomicfile.cc: In member function 'bool WvAtomicFile::open(const WvFastString&, int, mode_t)': Mar 25 13:14:12 # Mar 25 13:14:12 | streams/wvatomicfile.cc:40:31: error: 'lstat' was not declared in this scope Mar 25 13:14:12 # Mar 25 13:14:12 | streams/wvatomicfile.cc:40:60: error: 'S_ISREG' was not declared in this scope Mar 25 13:14:45 same Mar 25 13:15:37 fix Mar 25 13:19:57 how to fix this error? can anyone help me... Mar 25 13:20:24 sanket, as woglinde said, add the include Mar 25 13:21:18 wvunixdgsocket.h file is already present in wvstreams-4.6/include/wvunixdgsocket.h Mar 25 13:21:31 o.O Mar 25 13:21:50 -> 14:02 < woglinde> so #include in wvunixdgsocket.h Mar 25 13:21:58 TGIF Mar 25 13:22:36 eFfeM_work, I hv added #include in wvunixdgsocket.h .... after that I had got ythis error http://pastebin.mozilla.org/1191787 Mar 25 13:23:23 sanket, the answer has been given already twice: Mar 25 13:23:29 (02:20:49 PM) woglinde: -> 14:02 < woglinde> so #include in wvunixdgsocket.h Mar 25 13:23:36 sanket man 2 chmod Mar 25 13:23:42 or in the .cc file Mar 25 13:24:18 effem ah right in the .cc is cleaner Mar 25 13:24:19 sorry Mar 25 13:25:14 you might also consider adding for lstat etc Mar 25 13:25:23 and maybe Mar 25 13:31:19 eFfeM_work, I hv added #include in wvunixdgsocket.h as well as #include ..... with same error Mar 25 13:32:01 oh my god Mar 25 13:32:12 http://pastebin.mozilla.org/1191789 Mar 25 13:32:35 okay Mar 25 13:32:47 add #include in wvunixdgsocket.c Mar 25 13:33:02 woglinde: no Mar 25 13:33:04 now you should be able to fix the other stuff too Mar 25 13:33:16 sanket the file you need to modify is wxatomicfile.cc Mar 25 13:33:33 effem I dont know why he cannt abstract it to the other files Mar 25 13:33:45 woglinde: true Mar 25 13:33:58 the error is not in wvunxidgsocket.cc Mar 25 13:34:18 ./CXX -c streams/wvatomicfile Mar 25 13:34:26 thsi one will probalby not even include the .h file Mar 25 13:34:32 sanket: anyway I hope you get the idea Mar 25 13:41:43 eFfeM_work, thanx I had done with it.... Mar 25 13:41:46 thanx to all Mar 25 13:51:38 yw Mar 25 15:00:27 03Chris Larson  07master * r416d24912f 10bitbake.git/lib/bb/build.py: Mar 25 15:00:27 build: in exec_func, mkdirhier ${T} Mar 25 15:00:27 This should fix the -c clean traceback people are seeing. Mar 25 15:00:27 Signed-off-by: Chris Larson Mar 25 15:05:21 03Roger Monk  07org.openembedded.dev * raa34ecbbaa 10openembedded.git/recipes/ti/gst-ducati_git.bb: (log message trimmed) Mar 25 15:05:21 ti-gst-ducati: Update to latest SRCREV, package libs Mar 25 15:05:21 * Update to latest commit Mar 25 15:05:21 * Switch to use latest gstreamer included depends Mar 25 15:05:21 * Fix lib packaging Mar 25 15:05:22 * Bump PR Mar 25 15:05:22 Signed-off-by: Roger Monk Mar 25 15:08:17 03Roger Monk  07org.openembedded.dev * r460d4afbcb 10openembedded.git/recipes/ti/libdce_git.bb: Mar 25 15:08:17 ti-libdce: Update to latest SRCREV Mar 25 15:08:17 * Update to latest version Mar 25 15:08:17 * Bump PR Mar 25 15:08:17 Signed-off-by: Roger Monk Mar 25 15:08:17 Signed-off-by: Koen Kooi Mar 25 15:51:00 could it be that tzdata-2011b is not only stale but also unavailable nowadays? Mar 25 15:53:46 meh Mar 25 15:54:14 heh, it's pretty common for tzdata to be unfetchable from what i've seen, wouldn't surprise me, though that one doesn't seem particularly old. *shrug* Mar 25 15:54:41 kergoth, sounds like d is current but whatever. Mar 25 15:54:48 ah. Mar 25 15:55:06 * blindvt waves Mar 25 15:55:07 * blindvt & Mar 25 16:03:12 03Chris Larson  07master * r7a29ab5343 10bitbake.git/lib/bb/build.py: Mar 25 16:03:12 build: fix dir removal traceback Mar 25 16:03:12 This one is to cover the case where the current directory vanishes out from Mar 25 16:03:12 under us, so os.getcwd() raises an OSError. Mar 25 16:03:12 Signed-off-by: Chris Larson Mar 25 16:15:35 03Chris Larson  07master * r278fb41793 10bitbake.git/lib/bb/build.py: Mar 25 16:15:35 build: don't choke if olddir no longer exists Mar 25 16:15:35 Signed-off-by: Chris Larson Mar 25 16:17:58 03Roger Monk  07org.openembedded.dev * rd719b8bb5b 10openembedded.git/recipes/ffmpeg/ (files/omapfbplay-errorhandling.patch omapfbplay.inc): (log message trimmed) Mar 25 16:17:58 omapfbplay: Add patch for dce codec error handling Mar 25 16:17:58 * Add patch to handle return codes from dce codec library Mar 25 16:17:58 * Check for non-fatal errors Mar 25 16:17:58 * Update to latest SRCREV Mar 25 16:17:58 * Bump PR Mar 25 16:17:58 Signed-off-by: Roger Monk Mar 25 16:55:35 is git down ? Mar 25 16:56:21 khem depends Mar 25 16:56:29 somtimes ti works sometimes not Mar 25 16:56:32 git:// is down Mar 25 16:56:35 ssh:// is ok Mar 25 16:56:38 ah Mar 25 16:56:40 seems like Mar 25 16:56:40 right Mar 25 16:57:07 woglinde: ti is guilty? ;) Mar 25 16:57:15 florian? Mar 25 16:57:23 no khem is right Mar 25 16:57:29 git:// is broken Mar 25 16:58:11 khem ka6sox did soemthing Mar 25 17:00:28 woglinde: it was a bad joke about your typo :-) Mar 25 17:00:43 args Mar 25 17:01:03 florian btw. your friendly arm is still at my home Mar 25 17:01:16 do you want it back as fast? Mar 25 17:01:48 woglinde: it doesn't have to be fast but to get it back some time would be nice :) Mar 25 17:02:25 hm or maybee linuxtag? Mar 25 17:02:35 good point Mar 25 17:02:57 I should ask robert if got some news about a booth Mar 25 17:03:09 I need to send a mail - I got a mail that we are accepted at LinuxTag. Mar 25 17:03:17 uhm? Mar 25 17:03:19 you Mar 25 17:03:21 lol Mar 25 17:03:22 yes Mar 25 17:03:31 do we have two now? ;) Mar 25 17:03:32 thanks for applying Mar 25 17:03:45 I thought robert made the entries Mar 25 17:03:50 or seems you were faster Mar 25 17:03:55 heh Mar 25 17:04:08 I told robert at the cebit to do it Mar 25 17:04:15 where the deadline was over Mar 25 17:04:24 but we have some special connections Mar 25 17:04:26 woglinde: I had the vcc account and the description from the past few years :-) Mar 25 17:04:32 okay Mar 25 17:04:36 superb Mar 25 17:05:02 * florian needs more time Mar 25 17:05:21 florian for fairs? Mar 25 17:06:23 qmake is making autotools look pretty nice ... Mar 25 17:06:25 woglinde: For everything... lots of work, renovations at home, baby at home and now i have to find a new car. Mar 25 17:06:38 ugh Mar 25 17:06:48 cbrake o.O Mar 25 17:07:29 cbrake, amusing Mar 25 17:07:45 I am fighting with PyQwt building on an omap3 Mar 25 17:08:04 crofton got it worked now? Mar 25 17:08:15 IO built sip and PyQt native Mar 25 17:08:31 I am trying to get some stuff running, then I will fight the recipes some Mar 25 17:08:38 after I understnd how it works native better Mar 25 17:08:45 Crofton|work: I'm in the process of adding the latest qwt beta to OE Mar 25 17:08:46 sip was updated since you looked at it Mar 25 17:08:51 Crofton|work: what version you using? Mar 25 17:09:02 whatever is in dev Mar 25 17:09:48 be back in a bit ... Mar 25 17:09:58 good idea Mar 25 17:12:30 03Henning Heinold  07org.openembedded.dev * rb8d3dc71df 10openembedded.git/recipes/ecj/ (4 files in 2 dirs): ecj: for ecj-initial 512 mb heapsize is okay for ecj-bootstrap we need 1024 Mar 25 17:38:03 hi stefan Mar 25 17:38:12 hi woglinde Mar 25 17:38:23 you puhsed a patch for ecj-initial Mar 25 17:38:39 woglinde: hit some of our problems as well now? Mar 25 17:38:44 yes Mar 25 17:38:59 cool :) Mar 25 17:39:20 but now It builds fine for over an hour Mar 25 17:39:33 great Mar 25 17:40:02 I will cherry-pick the commit in our branch and put it in my next pull request for 2011.03 then Mar 25 18:18:21 Is the git server down, or is my box acting up again? Mar 25 18:18:29 Grr Mar 25 18:18:48 ssh only works again Mar 25 18:22:38 Is it possible to use bash with the devshell. That is not to spawn another shell through xterm or screen, but use bash right in line. I tried setting the dev shell term variable to bash, but when I run -c devshell it returns without loading a bash session. Mar 25 18:27:36 woglinde: retry now Mar 25 18:27:39 it should be fixed Mar 25 18:36:27 cbrake, ping Mar 25 18:36:50 Crofton|work: yes Mar 25 18:37:40 Could not find mkspecs for your QMAKESPEC(linux-g++) after trying: Mar 25 18:37:40 /home/balister/oe/tmp/sysroots/x86_64-linux/qt4/mkspecs Mar 25 18:37:56 looks like qmake installed on target has a bad path compiled in Mar 25 18:38:08 qmake-qt4 Mar 25 18:42:18 good morning Mar 25 18:42:44 re ka6sox Mar 25 18:43:11 ty woglinde Mar 25 18:43:28 hate being up till 5am working on servers...I'm too OLD for this. Mar 25 18:43:51 hm Mar 25 18:43:54 take a break Mar 25 18:43:59 make some holidays Mar 25 18:45:01 woglinde, I'd like that...someday.. Mar 25 18:45:04 when I'm older. Mar 25 18:45:10 (or dead) Mar 25 18:45:27 Germans are much better at this holiday thing Mar 25 18:45:52 he Mar 25 18:45:59 Crofton|work, in General Europeans have it down...its only us crazy colonials. Mar 25 18:46:08 come over Mar 25 18:46:43 and you dont need a car here Mar 25 18:46:51 saves a lot of money Mar 25 18:47:11 woglinde, them's are fighting words for a Californian....they will take our cars out of our cold, dead, hands. Mar 25 18:47:25 oh hehe Mar 25 18:47:59 crofton damn I missed the games yesterday Mar 25 18:48:37 heh Mar 25 18:48:48 my friend is no longer number one in the bracket thing Mar 25 18:48:58 *g* Mar 25 18:49:08 yes san diego is out Mar 25 18:49:16 and butler is suprising Mar 25 18:49:30 duke out too Mar 25 18:50:06 * ka6sox-work didn't think .eu cared about basketball...was wrong Mar 25 18:50:37 I get all my NCAA news from woglinde Mar 25 18:51:35 what? Mar 25 18:52:10 I do not pay attention to it :) Mar 25 18:52:12 didnt you watch the last world champion chip? Mar 25 18:52:17 crofton I know Mar 25 18:52:23 PyQwt buidling Mar 25 18:53:32 native :( Mar 25 18:54:34 tconant: no, devshell and such needs to spawn *something*. What's wrong with screen? Mar 25 19:00:04 Tartarus: what about a new sub-shell in the same tty? Mar 25 19:01:27 I suspect there's some limitation or another and kergoth would know the details Mar 25 19:14:05 03Simon Busch  07org.openembedded.dev * ra57ee950ce 10openembedded.git/recipes/e17/elfe_svn.bb: Mar 25 19:14:06 elfe: bump SRCREV to latest Mar 25 19:14:06 Signed-off-by: Simon Busch Mar 25 19:14:20 03Lukas Märdian  07org.openembedded.dev * re7a96874f5 10openembedded.git/recipes/shr/e-wm-config-illume2-shr_git.bb: (log message trimmed) Mar 25 19:14:20 e-wm-config-illume2-shr: use palmpre config option and bump SRCREV Mar 25 19:14:20 * bump SRCREV and PR Mar 25 19:14:20 * set PACKAGE_ARCH_palmpre = "${MACHINE_ARCH}" Mar 25 19:14:21 * add --enable-machine-palmpre to EXTRA_OECONF_palmpre Mar 25 19:14:21 * now a bigger illume-indicator (top-bar) is used for the palmpre Mar 25 19:14:22 Signed-off-by: Lukas Märdian Mar 25 19:27:59 evening Mar 25 19:28:03 hi jay7 Mar 25 19:30:49 woglinde: what was your outcome of kexec[boot] experiments on ac100? Mar 25 19:33:25 jay7 dont works Mar 25 19:33:32 kernel problem Mar 25 19:34:03 woglinde: ah Mar 25 19:34:39 woglinde: try to build chained kernel with builtin cmdline Mar 25 19:34:52 and as minimal as possible Mar 25 19:35:01 just to check Mar 25 19:35:12 having serial console is good idea as well Mar 25 19:35:55 We have found problem on nanonote only with serial console help. There was no add_memory_region() call Mar 25 19:42:32 How do I use a keyboard with a serial console? Mar 25 19:43:31 mwester-laptop enable kernel support Mar 25 19:43:41 and then what? Mar 25 19:43:49 connect it? Mar 25 19:43:53 :) Mar 25 19:44:33 Ok, done that. And I get events on a node in /dev for each key press and each key release... Mar 25 19:45:10 But I cannot figure out how to make it so that the keys are decoded and sent as input to the console session, which is on /dev/ttyS0 Mar 25 19:45:41 ugh Mar 25 19:45:43 hm Mar 25 19:45:49 looks like hardware 'feature' Mar 25 19:45:55 (My use case is that I want an NSLU2, which has no framebuffer or graphics, to be able to accept input from a USB keyboard during boot. Mar 25 19:45:56 ) Mar 25 19:46:08 ah thats diffrent Mar 25 19:46:12 serial and keyboard may use same gpio or something else Mar 25 19:46:24 than connecting a keyboard to the serial port Mar 25 19:46:43 Depending on what the user types at certain points in the boot, I can enable diagnostics, etc. I was hoping not to have to detect and decode key events, but perhaps that's the only way to do it. Mar 25 19:47:47 mwester-laptop: you should listen /dev/input/eventX device for input Mar 25 19:47:57 and do appropriate things Mar 25 19:48:11 Ok. I had hoped for an easier way, but that will have to do. :) Mar 25 19:48:26 mwester-laptop: look into kexecboot code for example :) Mar 25 19:48:33 or I may write something for you Mar 25 19:48:56 kexecboot code is in the OE repo, or elsewhere? Mar 25 19:49:05 OE have only recipe Mar 25 19:49:15 which points to kexecboot git repo ;) Mar 25 19:49:25 Ah, that'll work. :) I'll go look. Mar 25 19:49:32 http://git.linuxtogo.org/?p=groups/kexecboot/kexecboot.git;a=summary Mar 25 19:50:16 mwester-laptop: look into evdevs.c for process_events() function Mar 25 19:50:37 btw, I've unpushed and untested changes for that code.. Mar 25 19:52:00 Thanks! Mar 25 19:53:09 np Mar 25 19:55:59 mwester-laptop: you just need some 'daemon' code around this code :) Mar 25 19:56:15 but better to wait some time Mar 25 19:56:24 I'll push updated code Mar 25 19:56:35 it easier to read and smaller and faster :) Mar 25 19:56:43 I'll have no time until Tuesday, so that will be fine. Mar 25 19:57:06 I see it supports a mouse as well -- that's even better; simpler and more common. Mar 25 19:57:18 and even have placeholders for dealing with sockets Mar 25 19:57:27 yes, it does Mar 25 19:57:38 ah, no :) Mar 25 19:57:44 it's commented Mar 25 19:58:29 I've done some attempts to deal with touchscreen but this need some changes and synchronization Mar 25 19:59:01 I'll be making something self-contained; a daemon that will run for a while at boot, look for events, and based on the event it will do something with the beeper and the LEDs on the device to signal the results of diagnostics, or success in resetting the IP address to defaults, etc. Mar 25 19:59:35 mwester-laptop: good idea to use dbus.. but it's impossible at early boot Mar 25 20:00:02 may be behave like zaurusd then :) Mar 25 20:00:10 or acpid Mar 25 20:00:29 Ideally I'd like to be able to use this with only /proc and /sys mounted (in addition to the rootfs) -- if possible, I'd like to make it work before even udev runs. Mar 25 20:00:56 kexecboot is usable w/o udev :) Mar 25 20:01:04 Perfect. Mar 25 20:01:10 but you need /dev/input/eventX or /dev/eventX pre-created then Mar 25 20:01:38 That's ok -- we already do create similar devices for the beeper device. Mar 25 20:01:48 Adding a few extra is no problem. Mar 25 20:04:16 The boot sequence for SlugOS boots to a script (linuxrc), which figures out the real rootfs, then switches to that rootfs and runs init. The daemon would run during the delay time in the linuxrc boot script while we wait to ensure that the rootfs is ready. Mar 25 20:04:37 mwester-laptop: good idea to compile it with klibc then Mar 25 20:05:05 I have the full eglibc available from the flash fs -- SlugOS can run entirely from flash as well. Mar 25 20:05:21 ah, ok then Mar 25 20:06:01 I just need to make sure that we have no daemons that remain running as we switch rootfs, so that the user can have a different libc on the rootfs and have it actually work. Mar 25 20:07:39 This will also be useful on the Sheevaplug -- although it has the serial console more accessible than the NSLU2, neither one has external general-purpose buttons that one can use to affect the boot process. Mar 25 20:08:16 attaching keyboard or mouse may be useful :) Mar 25 20:08:43 mouse have at least 2 buttons that may be used to control boot process :) Mar 25 20:09:46 or you may use usb storage with files :) Mar 25 20:10:11 * Jay7 have lot of crazy ideas in his head Mar 25 20:11:46 I used usb storage with special files for the "Unslung" firmware for the NSLU2. That worked too. Mar 25 20:14:24 But I would like a way with SlugOS that would not require altering the external storage configurations, for example when /dev/sda becomes /dev/sdb which sometimes happens when one adds USB storage devices... Mar 25 20:15:00 With Unslung, the device name was locked to the USB port, so the device ID never changed. Mar 25 21:33:48 hi, can anyone help with how to "install" something into the sysroots (staging) area? I have some recipes using cmake that need to be able to find a library (built by another recipe) in a consistent place Mar 25 21:39:31 hbeck: are you inherting the cmake class? Mar 25 21:39:42 and have you looked at other cmake using recipes as examples? Mar 25 21:40:01 (git grep -l cmake recipes/*/*.{inc,bb}, browse a few of them) Mar 25 21:41:07 hm some should update vlc to 1.1.8 Mar 25 21:41:17 Tartarus: yep, cmake is working happy, but currently I have a Find.cmake Module (pulled by the toolchain file) that is going into the /oe/build/work/... directory to find the library needed Mar 25 21:42:27 I want to be able to direct the library in question to staging so that I don't have to edit the path info in Find.cmake when we change version/revision numbers on the recipe that builds it Mar 25 21:56:21 Tartarus: regarding the devshell with an in-line shell. The scroll back in screen is not very convenient compared to a native shell. Mar 25 21:57:06 tconant: I've been using screen so long I have problems with scrollback outside of it :) Mar 25 21:58:17 In theory it'd be possible to set aside the original file descriptors from the starting terminal and spawn a shell there, but it wouldn't work for all bitbake UIs, since the UI may very well want to read from the tty's stdin as well Mar 25 21:58:45 in addition, the ui may well output messages to that tty as well, particularly if other tasks run in parallel with devshell (theoretically possible, if unlikely) Mar 25 21:59:09 and remember that bitbake uses a client/server model now, the server may well be daemonized at some point in the past Mar 25 21:59:13 er, future Mar 25 21:59:32 a daemon can't go ahead and spawn a shell on a tty given it's properly daemonized and had those file descriptors closed Mar 25 21:59:34 Tartarus: I got mac where shells have really nice kinetic scrolling so, I've gotten spoiled. Mar 25 22:01:35 one possibility would be to add direct shell support to bitbake, and the UI could spawn it in whatever way is most appropriate, so knotty could spawn one on the original tty, while the gui might spawn an integrated shell window, or whatever Mar 25 22:01:49 which seems like it may be the best route in the very long term Mar 25 22:02:43 devshell started spawning off a new process the way it does when bitbake got multithreading support and all. when everything was simple and linear, it was easier Mar 25 22:02:46 (afaik) Mar 25 22:04:52 if anyone is curious about alternative ways of setting up oe-core, https://github.com/kergoth/reloc-tests is the sort of thing i do, rather than using either the poky scripts or the angstrom scripts Mar 25 22:05:23 03Denys Dmytriyenko  07master * ree44889f30 10openembedded.git/recipes/meta/external-toolchain-csl.bb: Mar 25 22:05:23 external-toolchain-csl: update licensing info, document exceptions Mar 25 22:05:23 * Split GCC and RLE licenses and base them off different gcc versions Mar 25 22:05:23 * Document all the details about the licenses and exceptions, so they Mar 25 22:05:23 won't get changed by accident again Mar 25 22:05:24 Signed-off-by: Denys Dmytriyenko Mar 25 22:05:30 03Denys Dmytriyenko  07master * radd7a3d31a 10openembedded.git/recipes/meta/external-toolchain-angstrom.bb: Mar 25 22:05:30 external-toolchain-angstrom: update licensing info, document exceptions Mar 25 22:05:30 * Split GCC and RLE licenses and base them off different gcc versions Mar 25 22:05:30 * Document all the details about the licenses and exceptions, so they Mar 25 22:05:30 won't get changed by accident again Mar 25 22:05:31 Signed-off-by: Denys Dmytriyenko Mar 25 22:07:19 tconant: if you run bitbake master, try using bitbake -u goggle sometime :) (won't help you here, but may give you an idea of the sort of interfaces which are available, and why spawning a shell on the tty isnt' as trivial as it seems) Mar 25 22:08:59 03Denys Dmytriyenko  07master * r23ed108cb7 10openembedded.git/recipes/meta/external-toolchain-angstrom.bb: Mar 25 22:08:59 external-toolchain-angstrom: variable name for gdb license got changed Mar 25 22:08:59 Signed-off-by: Denys Dmytriyenko Mar 25 22:22:27 do someone use xephyr with evdev? Mar 25 22:26:01 kergoth_: bb server may be run on other machine Mar 25 22:26:14 03Jeff Lance  07master * r3b907f7e4e 10openembedded.git/recipes/ti/ (3 files): (log message trimmed) Mar 25 22:26:14 matrix-gui: Bump up SRCREV Mar 25 22:26:14 * Bump up SRCREV to bring in mods on am180x and other platforms Mar 25 22:26:14 * am180x: Automated Wifi Mar 25 22:26:14 * am181x, am180x, omap-l138, dm368: Removing TI SVG icons as they were found to slow down update performance Mar 25 22:26:15 * am180x: Fixing PRU EDMA example Mar 25 22:26:15 * pulling in addtional mods from other platforms Mar 25 22:26:16 Jay7: yeah, good example, better than the ones i was thinking of :) Mar 25 22:26:28 definitely need bitbake knowledge / UI knowledge to implement it well Mar 25 22:26:35 even then it'd be tough Mar 25 22:26:36 so, bb client then should have possibility to pass commands to server and pass output back to client Mar 25 22:26:41 probably better off .. yeah Mar 25 22:26:44 more like bitbake interactive mode Mar 25 22:26:54 we could always make it act shell-like Mar 25 22:27:00 yep Mar 25 22:27:05 ant__: evening :) Mar 25 22:27:06 pass off the shell command to the server, wait for results, whatever Mar 25 22:27:07 * kergoth_ nods Mar 25 22:27:27 kergoth_: around ? if I include python function in the .bb itself it works fine. If I add python function to an inc file and do require include.inc in the .bb it does not work and if I do .bb -> .inc -> another .inc (contains the python func) it works again Mar 25 22:27:42 kergoth_: its an interesting behavior Mar 25 22:27:47 hehe Mar 25 22:27:57 khem: mystical Mar 25 22:28:15 otavio: it seems it like 1,3,5,7 and so on Mar 25 22:28:21 when it comes to file nesting level Mar 25 22:28:25 khem: weird. Mar 25 22:29:06 kergoth I kept moving that python function down the nesting levels Mar 25 22:29:14 khem: the def'd python function code in bitbake is really pretty simple. compiles/execs in a special context which is shared with our python tasks, stores in a dictionary associated with the file name to implement the warning where multiple files define it Mar 25 22:29:22 and stores the python function text in the metadata for reference Mar 25 22:29:35 maybe its something in that methodpool code Mar 25 22:29:49 kergoth_: its reproducable now Mar 25 22:30:10 its nothing to do with python version, oe or oe-core Mar 25 22:30:31 that's good to hear at least Mar 25 22:30:46 would you be interested to debug it ? Mar 25 22:31:11 do you know enough python to wade into the bitbake code on this, or do you want me to look at it? it won't be for a few days most likely, and ideally if you could just zip up a test case that'd make it quicker for the reproduction Mar 25 22:31:26 ok Mar 25 22:31:35 do u want it against oe or oe-core? Mar 25 22:32:07 heh, recently i'm preferring to do development on oe-core and port back to oe if necessary :) Mar 25 22:32:11 a bit more forward-looking that way Mar 25 22:32:13 I will give you two testcases one will define pythong func at nesting level 0 and another at nesting level 1 Mar 25 22:32:28 cool, thanks Mar 25 22:32:38 and if you like I can give you another at nesting level 2 Mar 25 22:32:45 so you can have 2 sandboxes Mar 25 22:32:49 or 3 Mar 25 22:33:29 I am interested on start moving to oe-core as our development base ... is it usable as is now? Mar 25 22:33:34 whatever you think will be most useful in debugging :) Mar 25 22:33:42 otavio: its stable enough sure Mar 25 22:33:44 otavio: yep. see the blog post someone did about building with it Mar 25 22:33:48 i do builds on it every day just about Mar 25 22:33:52 both poky and angstrom are buildable Mar 25 22:33:52 kergoth_: it was me :) Mar 25 22:33:56 oh right :) Mar 25 22:34:12 khem: for comparison, setup wise: https://github.com/kergoth/reloc-tests Mar 25 22:34:14 khem: is minimal being maintained there? Mar 25 22:34:20 http://sakrah.homelinux.org/blog/2011/03/using-openembedded-core-to-build-angstrom-for-qemu/ Mar 25 22:34:29 otavio: no minimal is not there Mar 25 22:34:34 otavio: don't think so, think its just poky and angstrom right now (and of course angstrom uses its own layers on top of oe-core) Mar 25 22:34:35 otavio: I have no intention Mar 25 22:34:47 I think angstrom is good Mar 25 22:35:07 otavio: any reason you would want minimal ? Mar 25 22:35:14 khem: minimal used to be smaller then angstrom Mar 25 22:35:21 khem: but didn't compare it latelly Mar 25 22:35:29 i think between poky and angstrom we have decent coverage, i think poky is a bit more basic than angstrom is Mar 25 22:35:38 otavio: yeah Mar 25 22:35:48 kergoth_: by basic you mean small? Mar 25 22:35:50 the whole micro collapsed filesystem is really intreguing to me, so i might have to try playing with that with it one of these days.. Mar 25 22:36:01 otavio: yeah, less full featured Mar 25 22:36:05 otavio: I think for now if you want benefits of oe then angstrom is something you can use Mar 25 22:36:10 but i haven't looked at their actual config files Mar 25 22:36:14 so that's just an impression Mar 25 22:36:14 we have oe gcc in meta-oe Mar 25 22:37:00 khem: currently it is not a big deal for me to use one specific gcc ... for me what I need is i586 or compatible support Mar 25 22:37:09 ok Mar 25 22:37:18 otavio: then you can choose poky or angstrom Mar 25 22:37:52 otavio: eventually you can have your own distro with set of layers Mar 25 22:37:59 just like angstrom Mar 25 22:38:14 khem: the idea for oe-core is to have only few machines right? Mar 25 22:38:20 otavio: yes Mar 25 22:38:27 it will have reference machines Mar 25 22:38:34 and so far its only qemu machines Mar 25 22:38:38 and it will remain so Mar 25 22:39:03 distro or bsp or machine layers should add more Mar 25 22:39:23 khem: right Mar 25 22:39:41 khem: i like the idea since it makes it more manageable Mar 25 22:39:48 yep Mar 25 22:39:55 khem: is any x86 qemu support on it? Mar 25 22:39:58 and you wont get zillion versions of managed recipes Mar 25 22:39:59 hey folks, found (http://docs.openembedded.org/usermanual/html/directories_staging.html) in ref to my earlier question about installing to staging. Question now is what would determine an "automatic" install to staging vs needing a direct reference? Mar 25 22:40:21 yes arm ppc mips x86 x86-64 qemus are supported Mar 25 22:40:28 nice Mar 25 22:40:44 otavio: they all work well Mar 25 22:41:05 so I believe the foundation is layed out and it strong Mar 25 22:41:10 khem: any reason why oe gcc didn't get merged into oe-core? Mar 25 22:41:13 now we need to populate layers on top Mar 25 22:41:28 otavio: I have not proposed it yet Mar 25 22:41:34 it needs discussion Mar 25 22:41:57 khem: afaik it has some advantage, no? Mar 25 22:42:02 it does Mar 25 22:42:06 newer linaro patches and so Mar 25 22:42:16 basic question is if we want to remain vanilla Mar 25 22:42:23 or provide patches Mar 25 22:42:27 so gcc is better Mar 25 22:42:49 there is another way where one could create layers for toolchain Mar 25 22:42:50 humm; for long term it would be indeed better to have it vanilla Mar 25 22:43:07 since it could "push" for us to try to merge the stuff into upstream Mar 25 22:43:12 but toolchain is so fundamental that its very difficult for someone to qualify a toolchain Mar 25 22:43:37 otavio: yes intention is always that Mar 25 22:43:42 specially because some weird bugs happen from time to time Mar 25 22:43:48 but gcc needs a lot of soak time Mar 25 22:44:13 now people have started to use gcc 4.5 and 4.6 is around the corner for release Mar 25 22:44:20 it takes about 6 - 12 months Mar 25 22:47:41 [19:40] < khem> but gcc needs a lot of soak time Mar 25 22:47:49 this is the last I saw Mar 25 22:49:07 ~logs Mar 25 22:49:07 All conversations are logged to http://ibot.rikers.org/%23oe/ Lines starting with spaces are not logged. Logs are updated daily. Mar 25 22:49:19 !logs Mar 25 22:49:20 Channel logs for #oe are archived at: Mar 25 22:49:21 http://hentges.net/tmp/logs/irc/%23oe Mar 25 22:49:22 Live-logs are available at Mar 25 22:49:23 http://hentges.net/tmp/logs/irc/livelogs/%23oe.livelog Mar 25 22:49:25 See ?? help-logs for usage instructions Mar 25 22:49:30 otavio_: ^ :) Mar 25 22:49:36 Jay7, gm Mar 25 22:50:04 ka6sox-work: morning :) Mar 25 22:50:21 long night last night...moving slow today. Mar 25 22:50:49 hard times Mar 25 22:50:57 indeed. Mar 25 22:52:39 otavio_: 15:32 < khem> now people have started to use gcc 4.5 and 4.6 is around the corner for release Mar 25 22:52:42 15:33 < khem> it takes about 6 - 12 months Mar 25 23:02:38 kergoth_: yes; maybe it might be good to start putting 4.6 into meta-oe Mar 25 23:03:41 otavio_: yes I will start with that soon Mar 25 23:03:43 god, this relocation testing script had damn well better be worth it, this is a gigantic pain in my ass Mar 25 23:03:45 there is more to it Mar 25 23:05:23 otavio_: about the compression in bitbake.conf, if you add lzo that would be perfect Mar 25 23:05:52 ant__: I can't look at it Mar 25 23:06:06 ant__: in fact I am at hospital ... will do a surgery tomorrow Mar 25 23:06:08 ant__: ;-) Mar 25 23:06:25 best wishes Mar 25 23:06:33 hope it goes well Mar 25 23:06:34 ant__: I can look at it later but can't promise when Mar 25 23:06:39 * otavio_ too Mar 25 23:07:40 kergoth_: there you go Mar 25 23:07:45 I have sent 2 testcases Mar 25 23:07:48 otavio: I ack the patch as it is for now, another ack and we'll push Mar 25 23:07:49 okay, thanks Mar 25 23:08:27 otavio: have a speedy recovery! Mar 25 23:17:15 otavio: just 3 recipes grepping for lzma-native: linux.inc, linux.kexecboot.inc, linux-wrt_2.4.30 Mar 25 23:17:38 linux-wrt_2.4.30 heh Mar 25 23:17:49 it that even relevant Mar 25 23:17:52 in current context Mar 25 23:17:54 oh, yes, I've seen some dinosauria recently Mar 25 23:19:43 khem: somehow task-boot pulls modutils-initscripts Mar 25 23:19:58 and I end up with 2.4/2.6 mixed modules stuff Mar 25 23:20:16 task-base separates that Mar 25 23:25:41 this is odd. gettext-native doesn't show up in -g for gmp-native or pseudo-native, yet building gmp-native shows gettext-native being built Mar 25 23:25:44 hmm Mar 25 23:26:05 is it one of early deps Mar 25 23:26:15 kergoth_ because its filtered out Mar 25 23:26:32 if its filtered out then it should not show in -g output Mar 25 23:26:36 woglinde: what do you mean? Mar 25 23:26:47 khem -> gettext-native doesn't show up in -g for gmp-native Mar 25 23:26:50 if something is being built, it should be in the -g Mar 25 23:26:51 doesnt!!!!!!!!!! Mar 25 23:27:01 hm ups Mar 25 23:27:26 * kergoth_ does a -g, then a -n to compare Mar 25 23:27:26 o.O Mar 25 23:28:03 for classpath the behavoiur changed Mar 25 23:28:12 inherit gettext wasnt enough Mar 25 23:28:13 actually two rogue deps, both tar-replacement-native and gettext-native seem to show up in the build but not in the -g.. that can't be right, i must be doing something wrong Mar 25 23:28:16 * kergoth_ investigates Mar 25 23:31:16 * Jay7 -> sleep Mar 25 23:31:49 nite jay7 Mar 25 23:42:14 good nite Mar 25 23:52:38 this is definitely happening. -n shows things that aren't in the -g Mar 25 23:52:42 * kergoth_ grumbles Mar 26 00:01:30 hey Mar 26 00:02:33 ERROR: Could not inherit file classes/rootfs_${IMAGE_PKGTYPE}.bbclass while parsing /.../openembedded/recipes/images/e-image.bb Mar 26 00:02:45 with bitbake 1.10.2 and latest oe git Mar 26 00:02:51 is this worrying? Mar 26 00:11:07 CMoH-notebook: misconfiguration. your DISTRO isn't set Mar 26 00:23:43 seems i must give up angstrom-2008.1 for angstrom-2010.x Mar 26 00:24:57 hmm... strange Mar 26 00:25:13 it says here i should be using angstrom-2008.1 for a stable angstrom Mar 26 00:25:38 is the config file still conf/local.conf? Mar 26 00:27:18 yup, it is... Mar 26 00:28:25 CMoH-notebook: if you want stable (and angstrom-2008.1), look at 2011.03-maintenance branch Mar 26 00:28:58 so angstrom-2008.1.conf disappeared from org.openembedded.dev ? Mar 26 00:29:03 the branch i mean Mar 26 00:30:35 ok, i found the commit **** ENDING LOGGING AT Sat Mar 26 02:59:58 2011