**** BEGIN LOGGING AT Tue Feb 08 02:59:57 2011 Feb 08 07:25:47 hello anyone can help me? Feb 08 07:31:57 help me Feb 08 07:36:08 whoami_: I pray for you ;) Feb 08 07:36:20 thanks Feb 08 07:36:35 but i cant get this package gconf-dbus Feb 08 07:36:40 Hi, I have created a new recipe: openembedded/recipes/visstrim_m10-utils/visstrim_m10-utils_0.1.bb Feb 08 07:36:49 but when I run bitbake -p it is not parsed Feb 08 07:37:50 delete the cache Feb 08 07:37:55 Helio_2000, Feb 08 07:38:04 it's in build/tmp/cahce Feb 08 07:38:13 *cache Feb 08 07:38:21 ok, I'll give it a try Feb 08 07:38:32 Helio_2000, additionally supply me the package gconf-dbus Feb 08 07:38:53 the svn repo is not valid Feb 08 07:40:20 you want sources or ipkg? Feb 08 07:41:36 source Feb 08 07:43:13 Helio_2000, ?? Feb 08 07:43:34 also anyone has any idea why wfica segfaults Feb 08 07:43:36 :( Feb 08 07:43:46 whoami_: it seems I don't have it available in my GIT version of openembedded Feb 08 07:45:41 nayone with ica client issue solved it Feb 08 08:01:32 good morning Feb 08 08:03:16 Jay7: thx, looks good to me, I'm going to test it Feb 08 11:53:14 hi ericben Feb 08 11:53:19 you didn't go to fosdem? Feb 08 11:57:33 fosdem was very quiet on OE people side Feb 08 11:57:40 setenv bootargs 'console=ttyS2,115200n8 ubi.mtd=filesystem root=ubi0:hawkboard-rootfs rootfstype=ubifs mem=34M vpif_capture.ch0_bufsize=1658880' Feb 08 11:57:48 I looked for ericben and I didn't found him Feb 08 11:58:08 how to write this entire bootargs in one line is my predicament ! Feb 08 11:58:14 Any idea ! Feb 08 11:58:39 I am not able to wrte this correctly for the u-boot args Feb 08 11:58:55 GNUtoo|laptop: did I found you? Feb 08 11:59:03 not sure Feb 08 11:59:16 GNUtoo|laptop: were you at beer event? Feb 08 11:59:20 yes Feb 08 11:59:28 but after I went away with tuxbrain people Feb 08 11:59:42 then I was for quite some time to the tuxbrain/oe booth Feb 08 11:59:57 I also attended to conferences Feb 08 12:00:04 including some in the embedded room Feb 08 12:01:24 I was at beer event at back with OE people, then at FOSDEM only sat morning Feb 08 12:08:05 ah you only came for the beer :-) Feb 08 12:08:14 * eFfeM_work was there only on sunday Feb 08 12:12:58 eFfeM_work: my girlfriend has familly in Brussels so we went there on Sunday Feb 08 12:13:10 ah nice Feb 08 12:13:26 take the best of both worlds! Feb 08 12:23:28 03Koen Kooi  07org.openembedded.dev * r8da4059223 10openembedded.git/recipes/u-boot/ (51 files in 2 dirs): Feb 08 12:23:28 u-boot git: beagleboard: add more expansionboards and rediff patches Feb 08 12:23:28 Signed-off-by: Koen Kooi Feb 08 12:26:54 03Koen Kooi  07org.openembedded.dev * r69ad11570c 10openembedded.git/recipes/linux/linux-omap-2.6.37/base/0013-omap3-beagleboard-add-WIP-support-for-beagleboardtoy.patch: Feb 08 12:26:54 linux-omap 2.6.37: catch up with beagleboardtoys wifi rename Feb 08 12:26:54 Signed-off-by: Koen Kooi Feb 08 13:14:56 * Crofton|work is home Feb 08 13:15:19 * XorA cheers at Crofton|work's good luck Feb 08 13:17:04 yeah, no late planes and a seat in first class from PHL to CLT Feb 08 13:17:23 jo crofton Feb 08 13:17:53 Crofton|work: now get on with email :-) Feb 08 13:23:08 working on it, Feb 08 13:23:51 re stefan Feb 08 13:24:23 * stefan_schmidt shakes the fist and people around him turning offf the swicth Feb 08 13:24:26 switch Feb 08 13:25:09 haha Feb 08 13:25:12 where are you? Feb 08 13:27:33 woglinde: university lab Feb 08 13:28:27 woglinde: someone just kicked the switch under the table and it lost power Feb 08 14:10:10 Jay7: hi, how about pulling my stuff? Feb 08 14:18:58 hi filip Feb 08 14:19:03 are your patches still outstanding? Feb 08 14:20:53 * Tartarus sees yes Feb 08 14:21:54 03Filip Zyzniewski  07org.openembedded.dev * rd2e75021ae 10openembedded.git/recipes/apt/ (apt-package.inc apt_0.7.20.2.bb): Feb 08 14:21:54 apt: remove references to not-built man pages. Feb 08 14:21:54 Manual pages are not being built because doxygen is currently not Feb 08 14:21:54 supported by openembedded. Feb 08 14:21:54 Signed-off-by: Filip Zyzniewski Feb 08 14:21:54 filip, thanks for fixing that stuff Feb 08 14:21:54 Signed-off-by: Tom Rini Feb 08 14:21:58 03Filip Zyzniewski  07org.openembedded.dev * r46a91960cc 10openembedded.git/classes/rootfs_deb.bbclass: Feb 08 14:21:58 rootfs_deb.bbclass: removed $(( ... )) operator usage. Feb 08 14:21:58 Bitbake's pyshlex.py cannot parse it raising NotImplementedError in Feb 08 14:21:58 line 295 (bitbake @ c8d8388ce9662574a390be3b849c07d1b044275d). Feb 08 14:21:58 Signed-off-by: Filip Zyzniewski Feb 08 14:21:58 Signed-off-by: Tom Rini Feb 08 14:22:00 03Filip Zyzniewski  07org.openembedded.dev * r086eb0df2d 10openembedded.git/recipes/apt/apt_0.7.20.2.bb: Feb 08 14:22:01 apt_0.7.20.2: fixed compilation errors with gcc-4.4. Feb 08 14:22:01 Just like in apt-native_0.7.20.2.bb. Feb 08 14:22:01 Related bugzilla entry: http://bugs.openembedded.org/show_bug.cgi?id=5511 Feb 08 14:22:01 Signed-off-by: Filip Zyzniewski Feb 08 14:22:13 example). Feb 08 14:23:51 *g* Feb 08 14:31:16 damnit, i want the chrome search bar dealy to search github automagically so i don't have to manually bookmark all the projects i watch Feb 08 14:41:28 mickey|office: is there already a recipe for libgisi in OE? Feb 08 14:46:25 * Tartarus finds only a little, so far, cruft in MAINTAINERS Feb 08 14:47:27 should ts_calibrate update the values in the lib or only save them to /etc/calibrate.conf ? Feb 08 14:48:14 not sure what you mean by that. it won't magically tell all your running processes about it, no Feb 08 14:48:27 hmm right thats what i meant Feb 08 14:48:45 my setup is tslib servinng Xorg server Feb 08 14:48:59 so what would be a way to gain on-the-fly calibration ? Feb 08 14:50:28 * kergoth_ doesn't know how the X support of tslib works Feb 08 14:50:39 ok thanks Feb 08 14:51:45 on the fly? Feb 08 14:51:58 so not to restart Xorg Feb 08 14:52:14 wasnt there an xorg-calibration rpogramm? Feb 08 14:52:21 yep there is one Feb 08 14:52:21 may it work with tslib? Feb 08 14:52:26 i thought so yeah, that would be the way to go, if it can handle it Feb 08 14:52:34 kergoth are you still maintainer of tslib? Feb 08 14:52:36 i think its works on the xinput layer Feb 08 14:52:54 maintainer mia Feb 08 14:53:48 okay, regarding autotools, does anyone have any thoughts on how to name the split out pieces? i'm separating the ./configure execution into a different class from autoreconf execution -- though i guess we could use a variable too i think this is a bit cleaner. currently its named autotools_base, the ./conigure part.. i was thinking maybe autotools/base, as a step toward creating class subdirs to organize Feb 08 14:54:14 woglinde: i am, but only cursory, i pretty much just apply bugfixes people send, not actively working on it beyond that -- if someone else wants to step up, they're welcome to take over Feb 08 14:54:24 i don't even have a device running linux that has a touchscreen right now to test it on ;) Feb 08 14:54:52 kergoth hm sounds good Feb 08 14:55:24 maybe a goal would be to get libxcalibrate support in Xorg Feb 08 14:55:31 could be autotools/configure or something, and also an autotools/autoreconf, but have autotools.bbclass inherit the latter automatically Feb 08 14:55:33 then xtscal would be the tool Feb 08 14:55:49 like in kdrive Feb 08 14:56:05 hmmmm Feb 08 14:57:01 or, i could call the autoreconf part autotools/bootstrap, that would sort of make sense, given what it does Feb 08 14:57:38 rob_w send patches to kergoth Feb 08 14:57:53 he's talking about a non-tslib approach Feb 08 14:57:57 so no, i don't want those patches ;) Feb 08 14:58:46 :q Feb 08 14:58:53 no i love tslib .. Feb 08 14:59:04 ah Feb 08 14:59:11 i just dont love the fact that Xorg is not prepared to take new values Feb 08 15:00:09 rob_w: is xinput set-int-prop not working? Feb 08 15:00:19 rob_w ah hm Feb 08 15:00:42 Tartarus: it's practically WIP, still needs fixing Feb 08 15:00:45 Tartarus: the dpkg stuff Feb 08 15:01:13 filip, heh, for stuff you don't want pushed just please prefix the email with RFC :) Feb 08 15:01:33 Tartarus: I do want it pushed, it does fix things Feb 08 15:01:43 Tartarus: but not all of them :) Feb 08 15:03:35 rob_w, why can't you change the driver to report evedv touch events instead of tslib touch events Feb 08 15:03:35 k Feb 08 15:03:45 you could use x-input-calibrator Feb 08 15:04:04 well i am actually confused Feb 08 15:04:13 GNUtoo|laptop: not sure what you mean by that.. tslib supports the event interface Feb 08 15:04:26 the question is what does the calibration on top of that :) Feb 08 15:04:34 (and filtering, though that's done in kernel in some cases) Feb 08 15:04:37 let me find a code example Feb 08 15:04:48 as the kernel driver sorts himself as evdev driver , then tslib hooks onto it with its evedev module and generates tslib events .. then xorg tslib driver takes it all Feb 08 15:05:05 i don't need a code sample, i know how input layer drivers work, i'm saying use of tslib doesn't imply that hte driver isn't using it Feb 08 15:05:26 * kergoth_ nods at rob_w Feb 08 15:05:27 code example was just to explain better Feb 08 15:06:03 someone updates freetype? Feb 08 15:06:16 so kergoth could we leave out tslib actually ? and hook the kernel evdev directly into xorg ?? Feb 08 15:06:29 yes, that's almost certainly an option Feb 08 15:06:38 i don't know anything about how that's done, though Feb 08 15:06:43 we would loose the debounce and stuff from tslib Feb 08 15:06:46 right Feb 08 15:06:51 hmm Feb 08 15:07:26 hmm let see if xinput-calibrator made it into oe already Feb 08 15:08:02 oh hell Feb 08 15:08:10 using subdirs in classes completely fucks EXPORT_FUNCTIONS Feb 08 15:08:17 * kergoth_ mutters Feb 08 15:11:10 gm rick Feb 08 15:15:31 xinput_calibrator does work .. but yet .. its night mare to restore it to a new xorg session .. udev rules and what not Feb 08 15:17:36 rob_w hm Feb 08 15:18:00 re stefan Feb 08 15:18:28 woglinde: :D Feb 08 15:18:36 woglinde: I should switch to wireless Feb 08 15:18:47 woglinde: but then my 1Gbs internet link is gone Feb 08 15:19:11 and? Feb 08 15:19:26 dont watch kino.to at univeristy Feb 08 15:19:33 woglinde: whats that? Feb 08 15:19:46 woglinde: do_fetch is enough :) Feb 08 15:19:52 ? Feb 08 15:20:05 what the hell you need to fetch? Feb 08 15:20:08 new computer? Feb 08 15:20:18 rofl Feb 08 15:20:54 woglinde: removed source dir Feb 08 15:20:59 oha Feb 08 15:21:07 woglinde: you kow, you have to clean up from time to time :) Feb 08 15:21:30 I am sure its not for research Feb 08 15:21:33 instead of filtering what is still needed I just get it again :) Feb 08 15:21:37 so you maybee violating the dfn rules Feb 08 15:21:45 * woglinde hides Feb 08 15:21:53 woglinde: sure it is. Need to build an image for my board I use for my study thesis :) Feb 08 15:22:03 ah Feb 08 15:22:05 okay Feb 08 15:22:06 woglinde: I band the rules to my needs Feb 08 15:22:15 so whats your thesis about? Feb 08 15:23:05 ongoing from my study thesis. Delay tolerant networking over 802.15.4 networks Feb 08 15:23:19 and my trafiic is hiden in all this torent traffic anyway... Feb 08 15:24:33 hm thats zippy? Feb 08 15:24:48 ach zigbee Feb 08 15:24:56 args Feb 08 15:25:07 write a zigbee implementation for the usrp e100 :) Feb 08 15:25:15 I am getting old Feb 08 15:25:18 old as crofton Feb 08 15:25:20 * woglinde runs Feb 08 15:25:46 woglinde: not really. Its the two lower layers specified somewhere else then the zigbee forum Feb 08 15:25:47 young whipper snappers Feb 08 15:25:55 means you can use it without the license hassle Feb 08 15:26:15 * stefan_schmidt stays away from zigbee Feb 08 15:26:45 we drive dtn directly on top of the mac layer and addressing scheme of 802.15.4 Feb 08 15:26:58 just going with the IEEE spec Feb 08 15:27:12 * woglinde is ieee member now Feb 08 15:27:51 woglinde: woaho Feb 08 15:27:55 woglinde: why that? Feb 08 15:27:59 fun Feb 08 15:28:05 cool Feb 08 15:28:09 and leaving university this year Feb 08 15:28:18 same plan for me :) Feb 08 15:28:27 so I want to have access to the digital library Feb 08 15:28:35 ah Feb 08 15:28:41 maybe I should do the same :) Feb 08 15:29:02 my univeristy has only regisrted for computer assoc Feb 08 15:29:11 so we dont get the paper from the other Feb 08 15:29:16 I would need to check here Feb 08 15:29:31 or you get a hint to check the old rotten paper Feb 08 15:29:36 802 specs are public anyway after some months Feb 08 15:29:48 its not about the specs Feb 08 15:29:54 its about the papers Feb 08 15:30:31 woglinde: so you are planing to read them after leaving university? Feb 08 15:30:31 so i recompiled xserver-xorg with -enable-xcalibrate yet it says XCALIBRATE extension missing when i call xtscal ? Feb 08 15:30:33 just kidding Feb 08 15:30:42 anyway, have to go. cu later Feb 08 15:30:58 stefan and its really fun to read the original papers Feb 08 15:31:01 oh Feb 08 15:31:03 okay Feb 08 15:31:06 bye Feb 08 15:31:19 cu Feb 08 15:31:31 rob_w did you check if really compiled? Feb 08 15:32:46 hmm Feb 08 15:35:25 woglinde gm yourself. Feb 08 15:35:47 Did kergoth and Jay7 get anything more done on the "update only on finish or fail" patch to bb for oestats? Feb 08 15:50:10 woglinde, how do i tell ? Feb 08 15:54:05 rob_w checking the do_compile Feb 08 15:54:13 and do_configure Feb 08 16:55:40 03Chris Larson  07master * r16e296b918 10bitbake.git/lib/bb/build.py: Feb 08 16:55:40 build: improve set -x handling Feb 08 16:55:40 - set -x always, not just when debugging Feb 08 16:55:40 - set -x immediately before executing the function, rather than at the top of Feb 08 16:55:40 the script, so the output is no longer cluttered with all the variable and Feb 08 16:55:41 function definitions Feb 08 16:55:41 Signed-off-by: Chris Larson Feb 08 17:02:25 how can I do an equivalent of exec &> somefile in perl? Feb 08 17:04:42 ok I found it Feb 08 17:04:47 lazy me Feb 08 17:05:44 ? Feb 08 17:06:19 set the right fd's? Feb 08 17:06:35 woglinde: I wanted open(STDOUT, "> somefile"); open(STDERR, ">somefile"); Feb 08 17:07:33 yeah Feb 08 17:07:35 as I said Feb 08 17:07:51 how can I force building git version of package ? Feb 08 17:09:39 only once Feb 08 17:09:46 or in local.conf? Feb 08 17:09:50 local.conf Feb 08 17:09:52 bibtake foo-git Feb 08 17:09:57 should work Feb 08 17:09:59 hm Feb 08 17:10:06 local.conf was trickier Feb 08 17:10:20 PREFERRED_VERSION, same as anything else, substituting % for the git hash if its in there Feb 08 17:10:40 iirc, anyway Feb 08 17:12:29 hm was it Feb 08 17:12:30 git push org.openembedded.dev:origin/org.openembedded.dev Feb 08 17:12:39 or git push git push org.openembedded.dev:org.openembedded.dev Feb 08 17:13:25 woglinde: why would you bother with that nowadays? Feb 08 17:13:33 just use master Feb 08 17:13:45 but neither of those will work, you need to specify the remote when using taht syntax Feb 08 17:13:55 git push origin localbranch:master Feb 08 17:13:58 or whatever Feb 08 17:14:27 hm yeah right Feb 08 17:14:35 git is awesome, so simple yet so powerful Feb 08 17:14:50 indeed Feb 08 17:16:06 03Henning Heinold  07master * rf26754d6bc 10openembedded.git/recipes/directfb/ (6 files in 3 dirs): Feb 08 17:16:07 directfb: update directfb 1.4 to directfb 1.4.11 and switch to INC_PR support Feb 08 17:16:07 * remove directfb < 1.2 stuff from .inc file Feb 08 17:16:07 * bump PR through INC_PR support Feb 08 17:16:42 hm I need to talk with jama to solve gl/gles Feb 08 17:16:45 for mesa Feb 08 17:20:36 khem: hmm, good point about the source packages, will look into it :) Feb 08 17:21:04 god, tab and space indentations mixed. AUGH Feb 08 17:23:02 filip yeah we have some of them Feb 08 17:23:13 :( Feb 08 17:23:27 i try to at least make classes more consistent when i touch them, but i hate to do commits just for whitespace Feb 08 17:23:29 fixm them Feb 08 17:23:40 I am not Feb 08 17:23:50 I like sexy style fixes Feb 08 17:25:16 woglinde: nah, these are debian folks (update-alternatives) Feb 08 17:27:54 hm the template is spaces and tabs? Feb 08 17:28:23 woglinde: no, the update-alternatives.pl perl script Feb 08 17:28:43 ah hahahahaa Feb 08 17:30:18 hmm, there is a C update-alternatives somewhere Feb 08 17:30:23 never had a chance to look at it though Feb 08 17:30:32 c-worth? Feb 08 17:31:29 damn this X gl, gles, cairo, bla stuff is so confusing Feb 08 17:32:48 hmm, http://lists.debian.org/debian-dpkg/2010/05/msg00086.html Feb 08 17:43:48 03Chris Larson  07master * rd93fcbd64a 10bitbake.git/lib/bb/persist_data.py: Feb 08 17:43:48 persist_data: loop on database lock for table creation Feb 08 17:43:48 Signed-off-by: Chris Larson Feb 08 17:44:25 RP__: this may well have been the random locked issue seen often in poky.. Feb 08 17:44:45 damn, maybe not Feb 08 17:44:47 still just hit one Feb 08 17:44:49 * kergoth_ mutters Feb 08 17:46:07 lock issue? Feb 08 17:46:10 kergoth_: I suspect the python bindings are broken but I've never dug into it :/ Feb 08 17:46:34 woglinde: randomly see a parse failure with 'database is locked', even with a fresh tmpdir Feb 08 17:46:49 i just hit it with stock oe, it just seems to happen more often with my poky based builds for some odd reason Feb 08 17:46:59 uh Feb 08 17:47:18 hm is it hided from us user? Feb 08 17:47:27 because I never saw the error Feb 08 17:48:07 again, random :P Feb 08 17:48:19 today's the first time i ever hit it with regular oe Feb 08 17:48:20 * kergoth_ shrugs Feb 08 17:51:14 Higher fraction of recipes using SRCPV? Feb 08 17:51:24 yeah that could be Feb 08 17:51:44 hm so its an sqlite-problem? Feb 08 17:51:51 * RP__ remembers the order on disk of file affecting the preferred provider selection bug Feb 08 17:51:59 woglinde: yep Feb 08 17:52:12 woglinde: yes, the sqlite database is locked and the timeout doesn't seem to apply Feb 08 17:52:14 that's the exact error message sqlite comes back with when there's a lock Feb 08 17:52:25 kergoth whats the sqlite version on the hostsystem? Feb 08 17:52:50 woglinde: its done this for a couple of years Feb 08 17:52:53 not a new problem Feb 08 17:53:04 rp 3.7 has more features Feb 08 17:53:12 solving this kind of problem Feb 08 17:53:25 if I remember correctly Feb 08 18:06:44 | /home/gnutoo/embedded/oe/oetmps/htcdream/sysroots/armv6-novfp-oe-linux-gnueabi/usr/include/integers.h:50:67: fatal error: srtp_config.h: No such file or directory Feb 08 18:07:14 from libpciaccess-0.12.1-r9.0 with scanpci.o Feb 08 18:07:15 hmmm Feb 08 18:07:26 intressting Feb 08 18:08:20 where should I start Feb 08 18:08:24 I've no pci on htcream Feb 08 18:08:27 *htcdream Feb 08 18:09:05 blacklist? Feb 08 18:09:33 okay Feb 08 18:09:36 time to go home Feb 08 18:09:38 till later Feb 08 18:13:26 add debug message for database is locked, watch it stop happening Feb 08 18:13:27 * kergoth_ glares Feb 08 18:13:29 * kergoth_ hates those Feb 08 18:49:01 hi folks, question about the binutils recipe that is using cvs - it looks like the repo for that one is down? cvs [update aborted]: unrecognized auth response from sources.redhat.com: Service temporarily unavailable, please try again later Feb 08 18:49:22 * kergoth_ kicks bitbake Feb 08 18:50:33 that a recipe problem or a source repo problem (the latter of which should resolve itself hopefully soon?) Feb 08 18:59:24 * kergoth_ sighs, database is locked isn't happening now Feb 08 19:07:15 RP: okay, i just added traceback to debug the database is locked with a poky based build, all of the ones i've seen so far were line 51 of persist_data, which is the bug/fix i just pushed to master, so that may very well have fixed it Feb 08 19:51:41 evening Feb 08 19:51:56 re jay7 Feb 08 20:03:44 03Koen Kooi  07org.openembedded.dev * r273d760f6f 10openembedded.git/recipes/u-boot/u-boot-git/beagleboard/0051-add-support-for-beagleboardtoys-expansionboards.patch: Feb 08 20:03:44 u-boot git: add missing patch to beagle. Thanks to Steve K for spotting that Feb 08 20:03:44 Signed-off-by: Koen Kooi Feb 08 20:12:24 well... Feb 08 20:12:56 I see preparations for release Feb 08 20:13:14 hay7 march is comming Feb 08 20:13:22 march next release Feb 08 20:13:26 after that yocto Feb 08 20:13:29 We should consider about using automated testing report collecting Feb 08 20:14:14 Q1) will we change oestats-client to post all data after build (I can't create it from scratch) Feb 08 20:14:34 well.. no other Q now :) Feb 08 20:15:22 after I'll try to change client/server sides to send/accept more parameters and create report that looks like our scary Testing table Feb 08 20:18:20 btw, I should upgrade my debian in buildbox Feb 08 20:51:01 03Chris Larson  07master * r3539c9474a 10bitbake.git/lib/bb/ (__init__.py event.py): Feb 08 20:51:01 Shift traceback pre-formatting into LogHandler Feb 08 20:51:01 It's cleaner to leave it behaving as usual, passing the exception data in the Feb 08 20:51:01 exc_info attribute of the LogRecord where it normally lies, and then let Feb 08 20:51:02 LogHandler make it pickleable so it can be sent to the UI. Feb 08 20:51:02 Signed-off-by: Chris Larson Feb 08 20:51:03 03Chris Larson  07master * r7486b38603 10bitbake.git/lib/bb/persist_data.py: Feb 08 20:51:04 persist_data: kill unreachable break line Feb 08 20:51:04 Signed-off-by: Chris Larson Feb 08 20:52:56 03Chris Larson  07master * rc0f3ff3eb5 10bitbake.git/lib/bb/event.py: Feb 08 20:52:56 event: limit the number of traceback entries to 5 Feb 08 20:52:56 This is slightly more sane. Feb 08 20:52:56 Signed-off-by: Chris Larson Feb 08 21:01:15 Hi all Feb 08 21:02:21 is it possible to apply a patch to the source of a package only if you're running a certain version of a different package? Feb 08 21:03:18 I revisited the issue discussed on the mailing list here ( http://www.mail-archive.com/angstrom-distro-devel@linuxtogo.org/msg04090.html ) and came up with a minimal patch for ti-xdctools Feb 08 21:03:21 SRC_URI += inside the recipe Feb 08 21:03:43 but it should only be applied if you're using xdctools < 3.21 and binutils >= 2.19 Feb 08 21:04:15 ok, so i should have some conditionals that look at the binutils version, and conditionally add the .patch file to SRC_URI? Feb 08 21:04:46 mrj10 thats not possible Feb 08 21:04:58 we dont have version depends in bitbake Feb 08 21:05:04 a long missing feature Feb 08 21:05:10 but I know the patch Feb 08 21:05:36 I was the first one which found a solution for Feb 08 21:05:37 it Feb 08 21:05:49 while compiling it with uclibc Feb 08 21:06:14 the ti-xdctools thing? or you came up with a patch to add version depends to bitbake? Feb 08 21:06:57 no Feb 08 21:07:08 ti-xdctools linker script fixing Feb 08 21:07:16 for binutils > 2.20 Feb 08 21:07:49 gotcha. i found the workaround on the eclipse forum for xdctools Feb 08 21:08:20 ? Feb 08 21:08:35 its since some months in oe Feb 08 21:09:21 wait, the fix already exists in OE? my tree must be messed up in some way then Feb 08 21:10:15 there was a thread on the OE mailing list a while ago that didn't come to any satisfactory resolution on the matter http://www.mail-archive.com/angstrom-distro-devel@linuxtogo.org/msg03995.html Feb 08 21:11:17 it branched into 2 subthreads; in one, Denys just said that you shouldn't be trying to build it with binutils > 2.18. in the other, i pointed out the workaround i had found on the eclipse forums, but didn't have time to turn it into a proper patch submission to OE Feb 08 21:11:36 hm I though koen checked it already in Feb 08 21:12:46 hmm. i'm on master and current up to yesterday afternoon Feb 08 21:13:05 maybe it's on a different branch? Feb 08 21:13:10 ah hm Feb 08 21:13:16 okay see the reason now Feb 08 21:13:17 # This fixes c6accel, but breaks codec-engine and lpm Feb 08 21:13:17 #SRC_URI += "file://arm-linker-hack.diff" Feb 08 21:14:41 but I have no time right now to look deeper into it Feb 08 21:27:34 what way to find system 'bintess' is better - file /sbin/init or file `which file` ? Feb 08 21:27:52 I mean userspace arch Feb 08 21:30:31 or may be any better way Feb 08 21:32:59 e.g. objdump -a $file Feb 08 21:33:28 readelf Feb 08 21:33:31 is the best Feb 08 21:33:44 objdump dont get if its eabi Feb 08 21:34:10 ah, forgot it Feb 08 21:34:18 hmm Feb 08 21:34:27 kergoth si Feb 08 21:35:21 woglinde: thanks for finding that for me, i'll look at it a bit more. that's a different workaround than the one suggested on the ti-xdctools forum. it looks cleaner than what I did, but iirc my ti-codec-engine and lpm didn't break as a result of the one i used Feb 08 21:35:50 mrj10 ah could you pastebin it? Feb 08 21:36:19 ti-codec-engine didnt break either If I remember correctly Feb 08 21:37:20 hm ah no Feb 08 21:37:24 thats not the patch Feb 08 21:37:29 well.. readelf -h `which readelf` should be ok Feb 08 21:37:34 hrms Feb 08 21:38:11 is there a new maillist for oe-core development? Feb 08 21:38:27 * woglinde dont knows Feb 08 21:45:57 cbrake, hasn't been one Feb 08 21:46:01 Was that part of the plan or not? Feb 08 21:47:22 Tartarus: I thoughts someone mentioned it once, but can't recall exactly Feb 08 22:29:24 03Henning Heinold  07org.openembedded.dev * rd6e5809d08 10openembedded.git/recipes/disko/ (7 files in 3 dirs): disko: update git version, use INC_PR, re work disko.inc usage Feb 08 22:29:27 03Henning Heinold  07org.openembedded.dev * r48b894463d 10openembedded.git/recipes/disko/disko.inc: disko: fix legacy staging Feb 08 22:30:21 woglinde: sorry, stepped away for a bit, i can pastebin my patch Feb 08 22:30:58 http://pastebin.com/Mgjx9cKq Feb 08 22:32:18 the arm-linker-hack.diff commented out in OE attempts to be smart by further gating the _utils.genSections() call using the binutils version. while this is a theoretically superior approach, the comment in OE implies that it doesn't actually work as implemented. Feb 08 22:33:33 to my eyes, the patch i pastebinned (which just blows away the end of the file) doesn't break ti-codec-engine or lpm, but if anyone has suggestions for specific tests I should run (I have a beagleboard xm rev a2), i can confirm this further Feb 08 22:36:25 03Chris Larson  07master * re4db09fd12 10bitbake.git/lib/bb/event.py: Feb 08 22:36:25 event: fix exception format Feb 08 22:36:25 Forgot to concatenate the lines returned by format_exception. Feb 08 22:36:25 Signed-off-by: Chris Larson Feb 08 22:36:58 03Henning Heinold  07org.openembedded.dev * r024e010e93 10openembedded.git/recipes/libsdl/ (3 files): Feb 08 22:36:58 libsdl: use new .inc for 2.1.14 version Feb 08 22:36:58 * make autotools-usage and libtool2.4 from directfb generic, Feb 08 22:36:58 this fixes libsdl-x11 too Feb 08 22:36:58 * switch to INC_PR Feb 08 22:39:24 mrj10 best talk to koen in #beagle what the status is Feb 08 22:39:28 I am off Feb 08 22:39:37 disko und sdl rework is in Feb 08 22:42:16 RP, ping Feb 08 22:42:40 03Graeme Gregory  07org.openembedded.dev * r42f8d0260d 10openembedded.git/recipes/u-boot/u-boot_git.bb: Feb 08 22:42:41 u-boot : for omapzoom2/3 and gcc 4.5 the -Os optimisation produces a Feb 08 22:42:41 u-boot which doesnt boot or launch kernel. This makes u-boot a lot Feb 08 22:42:41 slower but as we only see it once on boot this is unimportant in long Feb 08 22:42:41 view so remove the -Os from build. Feb 08 22:42:41 Signed-off-by: Graeme Gregory Feb 08 22:43:34 maybe too late, heh Feb 08 23:26:54 khem: ping Feb 08 23:29:53 sh$t :( 32Gb is not enough to build both x11-image and opie-image Feb 08 23:30:08 at least with speed scheduler Feb 08 23:30:15 now switched to completion Feb 08 23:32:36 damn you, 'database is locked' Feb 08 23:35:30 kergoth: still unresolved? Feb 08 23:36:09 i thought i had it, but apparently not all the issues Feb 08 23:36:19 btw, I can run set of builds with bb threads from 1 to 8 e.g. Feb 08 23:37:02 not sure about useful outcome :) Feb 08 23:40:38 anyway, time to go to bed Feb 08 23:40:41 * Jay7 -> sleep **** ENDING LOGGING AT Wed Feb 09 02:59:57 2011