**** BEGIN LOGGING AT Mon Jan 18 02:59:56 2010 Jan 18 06:04:49 can anybody help me Jan 18 06:04:49 ./bitbake/lib/bb/COW.py:29: DeprecationWarning: the sets module is deprecated Jan 18 06:04:51 OE is not building Jan 18 06:25:47 bloody shit Jan 18 06:26:31 oe_imxdev.html from From iMXwiki Jan 18 06:26:49 wget http://download.berlios.de/bitbake/bitbake-1.8.12.tar.gz Jan 18 06:27:07 they have older version of bitbake bcos of that i m getting that error Jan 18 06:27:18 damn iMxwiki page guys. Jan 18 06:40:46 morning Jan 18 06:59:41 hi all Jan 18 06:59:44 ERROR: Nothing PROVIDES 'x11-gpe-image' Jan 18 07:00:07 when i run bitbake x11-gpe-image i am getting above error Jan 18 07:00:15 any idea guys Jan 18 07:06:10 got the mistake. Jan 18 07:37:40 good morning Jan 18 07:53:09 03Koen Kooi 07org.openembedded.dev * r1c14a32da4 10openembedded.git/recipes/linux/linux-omap-zoomsync/omapzoom2/defconfig: linux-omap-zoomsync 2.6.31: enable some more drivers for zoom2 Jan 18 07:53:10 03Koen Kooi 07org.openembedded.dev * rc36054c475 10openembedded.git/ (conf/checksums.ini recipes/gstreamer/gst-rtsp_0.10.5.bb): gst-rtsp: add 0.10.5 Jan 18 07:53:11 03Koen Kooi 07org.openembedded.dev * rdf43057ec0 10openembedded.git/ (conf/checksums.ini recipes/gnome/gnome-dvb-daemon_0.1.14.bb): gnome-dvb-daemon: add 0.1.14 Jan 18 07:53:12 03Koen Kooi 07org.openembedded.dev * r21b83f8aa0 10openembedded.git/recipes/ggz/ggz-client-libs_0.0.14.1.bb: ggz-client-libs: package more files Jan 18 07:53:15 03Koen Kooi 07org.openembedded.dev * r33c882b663 10openembedded.git/recipes/mplayer/mplayer_svn.bb: mplayer: bump SRCREV for some more ARM fixes Jan 18 07:53:18 03Koen Kooi 07org.openembedded.dev * rdfa0d36cac 10openembedded.git/contrib/angstrom/build-feeds.sh: angstrom feed builder: add some more stuff Jan 18 07:53:21 03Koen Kooi 07org.openembedded.dev * r234f947712 10openembedded.git/contrib/angstrom/sort.sh: angstrom feed sorter: add openrd-* Jan 18 07:53:24 03Koen Kooi 07org.openembedded.dev * r7d25400166 10openembedded.git/recipes/xbmc/libcdio_0.81.bb: libcdio: add 0.81 Jan 18 07:53:27 03Koen Kooi 07org.openembedded.dev * r7b7605a0d1 10openembedded.git/recipes/musicbrainz/libmusicbrainz.inc: musicbrainz: add -ldl tp LDFLAGS, needed for some archs Jan 18 07:53:30 03Koen Kooi 07org.openembedded.dev * r8b6d842214 10openembedded.git/recipes/shadow/ (files/pam.d/login files/securetty shadow_4.1.4.2.bb): shadow: try to fix serial-login issue Jan 18 07:53:35 03Koen Kooi 07org.openembedded.dev * r1d1e92c5a2 10openembedded.git/recipes/policykit/policykit_0.9.bb: policykit: try harder to package /var/run/PolicyKit Jan 18 07:53:38 03Koen Kooi 07org.openembedded.dev * r60508b8749 10openembedded.git/ (6 files in 4 dirs): Jan 18 07:53:41 empathy: add 2.28.2 Jan 18 07:53:43 gnome-doc-utils: add 0.19.2 Jan 18 07:53:45 telepathy-glib: add 0.7.37 Jan 18 07:55:10 hello Jan 18 07:57:09 i am trying to build an embedded linux image for s3c2440 arm processor but the help and documentation isn't sufficient Jan 18 07:57:22 may anyone please tell me how should i proceed Jan 18 07:57:53 actually my first question is why should i recompile my existing gcc Jan 18 07:59:56 anyone..... Jan 18 08:07:08 ummm........ Jan 18 08:07:09 would anyone please help me out.... Jan 18 09:25:09 ok, need to take care of other things. bye Jan 18 09:32:13 hi, what would be the best way to add building an additional recipe after gcc cross is made, but before any target recipe is build ? Jan 18 09:36:36 eFfeM: probably to add it to DEPENDS_prepend using similar logic to that in base_dep_prepend() Jan 18 09:38:11 pb__: hi, are you going alone at fosdem? Jan 18 09:38:47 mckoan: I'm not sure what you mean by "going alone" Jan 18 09:39:43 pb_ thanks, also noticed uclibc.inc does a similar thing (although commented out) Jan 18 09:39:43 #DEPENDS_append_bfin = " elf2flt " Jan 18 09:40:34 guess prepend will be better because then it has a higher chance of being done first Jan 18 09:41:04 pb__: you told maybe there is also your family? Jan 18 09:41:38 oh, right. yeah, my wife and daughter will be coming to Brussels with me (though not to FOSDEM itself) Jan 18 09:52:28 good morning Jan 18 09:53:04 good morning Jan 18 09:53:16 good morning Jan 18 10:22:32 florian: good morning Jan 18 11:24:31 morning all Jan 18 11:26:17 hi RP Jan 18 11:27:54 moin RP Jan 18 11:28:39 RP: please check my message about bitbake patches.. or I can resend if you don't have it in backlog Jan 18 11:30:57 hello, maybe someone can help me getting rid of some strange "perl cpan recipe" error? i was trying to add a recipe for XML-LibXML, that needs libxml2 installed; unfortunatelly someting with the Makefile.PL goes wrong and it tries to open a 32 bit conftest.pm with a x64 (staging) perl Jan 18 11:31:44 hi rp Jan 18 11:33:01 JaMa|Off: I saw your patch on the bitbake list Jan 18 11:33:23 JaMa|Off: The 1.10 branch is going to be entirely replaced with something from master Jan 18 11:33:30 JaMa|Off: please don't spend any time on it Jan 18 11:33:39 RP: it's for both 1.10 and master and one should be cherry-picked from master.. Jan 18 11:33:40 ah ok Jan 18 11:34:08 http://pastebin.com/d18981f95 (mb somegot a clue) Jan 18 11:43:55 Best policy for old/outdated distro files? git-rm, git-mv (where?)? Jan 18 11:45:18 kristoffer: depends who "owns" them I guess Jan 18 11:46:04 RP, me. Personally I would prefer having an obselete/archive foder where I could move them. Could be good for reference. Jan 18 11:46:19 foder=folder Jan 18 11:47:05 (`foder' in portuguese is `to f***', BTW) Jan 18 11:47:10 kristoffer: We don't keep obsolete files around in OE, this is why we have an SCM Jan 18 11:48:03 RP, I guess. git-rm it is then. Jan 18 11:48:21 mario-goulart, :P good to know Jan 18 12:39:48 doesn anyone here know why busybox httpd applet doenst support url parameters like http://test/cgi-bin/test.sh?blabla Jan 18 12:40:54 because it was not implemented ? Jan 18 12:44:53 hmm i doubt that Jan 18 12:50:25 check the src ? Jan 18 13:01:33 into it now Jan 18 13:33:01 re Jan 18 13:33:43 wb hrw Jan 18 13:49:09 hi Jan 18 13:49:55 some devs are going to fosdem(for instance the ones listed on the wiki) Jan 18 13:49:58 right? Jan 18 13:50:07 are the hotel already chosen? Jan 18 13:52:50 GNUtoo: astrid for me as usual Jan 18 13:53:15 ok Jan 18 13:53:23 so there is no oe hotel? Jan 18 13:54:00 in which way? Jan 18 13:54:22 in previous years we mostly gathered in astrid but some devs used other hotels too Jan 18 13:54:50 ok thanks Jan 18 13:55:42 GNUtoo|oeee: me and other italian (non OE dev) people are at Ibis Gran Place Jan 18 13:56:14 ok thanks Jan 18 13:56:28 GNUtoo|oeee: but gremlin is going at Astrid for example Jan 18 14:04:06 astrid seems full.so I'll go to the usual one Jan 18 14:08:12 yeah, the astrid is the traditional team o[pi]e hotel, but there are plenty of others. Jan 18 14:08:22 ok Jan 18 14:08:26 thanks Jan 18 14:09:07 I guess it wouldn't be a completely bad idea to add some hotel information to that wiki page. Jan 18 14:09:12 maybe you'd like to go ahead and do that. Jan 18 14:13:32 I've not many infos... Jan 18 14:13:48 but I can put me in the list and my hotel Jan 18 14:14:35 righto Jan 18 14:15:25 I think i found a bitbake bug. Jan 18 14:16:55 pix_: congratulations Jan 18 14:17:27 I have a recipe which use the git fetcher, this recipe use: SRCREV = "${AUTOREV}" and DEFAULT_PREFERENCE = "-1" the other recipe build using a tarball. When i bitbake ths os task, if the git server is down bitbake seems to try to connect to the git server and fail. Jan 18 14:20:18 indeed, that is the expected behaviour Jan 18 14:22:24 pb_: If git is online, bitbake will build the recipe using the tarball, but if git is offline the recipe fails <- this is "expected behavior" ? Jan 18 14:22:34 yes Jan 18 14:23:13 pb_: ok, sorry for the false alarm. Jan 18 14:23:45 there isn't really any sensible alternative. if the git server is offline, there is no way to expand SRCREV for the autorev recipe, and hence no way to determine what its PV would be. Jan 18 14:25:04 in theory it would be possible for bitbake to do a little bit better than it does at the moment, but there is no way to completely solve that problem in the general case. Jan 18 14:25:18 pb_: bit if another exists and this one has it's preference to -1, It should try the other no ? Jan 18 14:25:28 its Jan 18 14:29:15 yeah, that would be desirable. right now, though, I think pretty much any error during parsing is fatal. Jan 18 14:30:08 and expanding ${DEFAULT_PREFERENCE} is, obviously, not guaranteed to succeed if the metadata is only half there, since it might have been defined in terms of ${SRCREV}. Jan 18 14:34:39 pix_: which bitbake version do you have? Jan 18 14:35:26 phpmod seem to have some problems in the dev branch. "[notice] child pid 3438 exit signal Segmentation fault (11)" Jan 18 14:35:39 that was from apaches error log Jan 18 14:36:32 pb_: HEAD of bitbake 1:10 has BB_GIT_CLONE_FOR_SRCREV which changes that error a bit.. because it tries to do full git clone for that recipe Jan 18 14:36:53 I mean BB_GIT_CLONE_FOR_SRCREV enabled by default (debug left-over) Jan 18 14:38:10 this patch from master is needed to turn it off by default again 0159f66e0d95c45537de4ab77924ba6521022370 Jan 18 14:48:49 morning Jan 18 14:51:53 g'day kergoth Jan 18 15:28:29 I've copied the hal_0.5.12.bb into my overlay and put PREFERRED_VERSION_hal=0.5.12 in my machine.conf but it still tries to build 0.5.13. Any ideas why that might be? Jan 18 15:28:36 or how I can find out why :) Jan 18 15:29:24 sounds like you need to set your BBFILE_COLLECTIONS appropriately, along with the patterns & priorities Jan 18 15:29:31 this is how bitbake prioritizes amongst overlays Jan 18 15:29:39 kergoth, that's already working for a lot of other packages Jan 18 15:30:06 and even then 0.5.12 also exists in the OE repo, so it should pick that one up instead of 0.5.13? Jan 18 15:30:23 true. maybe the distro/machine is overriding your preference Jan 18 15:30:23 v Jan 18 15:30:26 erify with bitbkae -e Jan 18 15:30:27 blah Jan 18 15:30:33 bitbake -e | grep \^PREFERRED_VERSION_hal Jan 18 15:31:48 yeah it has picked up 0.5.13 from somewhere.. thanks :) Jan 18 15:32:15 np Jan 18 15:32:54 it has include/angstrom-2008-preferred-versions.inc:PREFERRED_VERSION_hal = "0.5.13" Jan 18 15:32:57 PREFERRED_VERSION_hal_local is your friend :) Jan 18 15:32:59 how can I override that? Jan 18 15:33:01 ahh Jan 18 15:33:03 ty :)) Jan 18 15:33:07 well, any override would do Jan 18 15:33:21 but this is what _local is for, its the most specific of the overrides, so you can override everything else Jan 18 15:33:54 well you just saved me a lot of trouble Jan 18 15:34:06 so thanks a lot :) Jan 18 15:34:25 np Jan 18 15:34:51 I guess the distro include should have ?= if it was nice Jan 18 15:34:54 bitbake -e is great for that sort of thing, whenever something is happening that just doesn't make any sense ā€” time to make sure things are what we think they are :) Jan 18 15:34:56 yeah, indeed Jan 18 15:40:53 btw Jan 18 15:41:20 if I have a "do_install_append()" in an .inc file and also one in the main bb file do they get appended to each other or does the second override the first? Jan 18 15:41:33 and if the latter then is there a recommended way of sorting that out? Jan 18 15:43:35 its the former Jan 18 15:43:40 they're both appended Jan 18 15:43:41 great Jan 18 15:43:48 that's what I hoped Jan 18 15:43:56 the latter could be implemented by creating a separate function, letting that be overridden, and appending that Jan 18 15:44:00 so that gives the most flexibility Jan 18 15:44:14 huh, http://distributions.freedesktop.org/wiki/DistributionLocations is a nice little resource Jan 18 15:45:17 useful enough to deserve a bookmark Jan 18 15:45:52 indeed Jan 18 15:46:09 there's also the distributions fd.o mailing list, though its pretty quiet at the moment Jan 18 16:03:43 * kergoth grumbles and tries to finish cleaning up the bluez crap Jan 18 16:36:54 I use pstaging, is there a command to figure out which staging package a specific file in staging dir comes from? Jan 18 16:37:53 it's just opkg, there's an opkg database in there Jan 18 16:38:09 * kergoth wonders if bt-configure needs the bluez4 utils or just the libs Jan 18 16:38:14 * kergoth guesses probablyt he former Jan 18 17:04:21 http://git.openembedded.org/cgit.cgi/openembedded/commit/?h=kergoth/bluez&id=a053da8743966445342b495307e532178a1294e9 - heh, needs more cleanup, but its a start Jan 18 17:07:10 nice Jan 18 17:07:31 is that the patch to get rid of all those bluez warnings? Jan 18 17:07:52 kergoth: gstreamer, libx11 are in utils or libs? Jan 18 17:08:22 separate binary packages from utils Jan 18 17:08:30 should probably use a distro feature for those Jan 18 17:08:54 great Jan 18 17:09:10 so if no utils but need bluez libs then build time will be much smaller Jan 18 17:09:16 yep Jan 18 17:09:48 it's mainly pb_'s bluez-apps 4.x recipe with some consolidation, and a rewritten patch for 4.5x Jan 18 17:27:04 * kergoth thinks we should think about deprecating bb.error/bb.msg.error, as it confuses people.. people think of warnings and errors, not so much errors that don't break the build and those that do (fatal) Jan 18 17:30:51 * florian agrees Jan 18 17:31:11 * kergoth just got bit by that at work Jan 18 17:33:12 * kergoth also thinks these horrible hacks where errors are generated by expansions of variables warning about the consequences of their use in certain circumstances need to die Jan 18 17:33:32 All these strange messages from Angström while parsing alr likely to confuse people too. Jan 18 17:38:26 * kergoth grumbles, really need to implement some sort of variable reference tracking.. it would kick ass if bitbake -e only showed the variables a recipe actually utilized or exported Jan 18 17:40:11 o yes Jan 18 17:42:49 bb.msg.error(bb.msg.domain.Fetcher, "SRCREV was used yet no valid SCM was found in SRC_URI") Jan 18 17:42:49 raise ParameterError Jan 18 17:42:50 ugh.. Jan 18 17:43:00 why is that message not included in the exception? Jan 18 17:51:06 * kergoth thinks we should ditch the catchall exception handler in bb.data_smart.DataSmart.expand() Jan 18 17:52:09 better to let it be hit by the toplevel exception handler as an uncaught exception, and hence, a bug, when someone uses a var they shouldn't be using, than to print a message and continue on our merry way.. Jan 18 17:53:21 ah, right, it does re-raise, but it dumps a message too.. hmm Jan 18 17:53:41 * kergoth downs more caffeine Jan 18 18:12:32 ugh, bb.data.data_smart.DataSmart.expand makes me sad Jan 18 18:20:30 i wonder how much would break if we made references to undefined variables return an empty string Jan 18 18:42:15 03Kristoffer Ericson 07org.openembedded.dev * r74a308c3d3 10openembedded.git/conf/distro/ (4 files): (log message trimmed) Jan 18 18:42:15 Remove obselete distro files: Jan 18 18:42:15 * We earlier had an need for different distro files Jan 18 18:42:15 due to differences in archs. Now we have managed to merge Jan 18 18:42:15 all platforms into a single distro file. Besides these old Jan 18 18:42:18 distro files havent been used in between 1-4 years so its Jan 18 18:42:20 time to get rid of them. Jan 18 18:42:22 03Kristoffer Ericson 07org.openembedded.dev * rb1a0492b21 10openembedded.git/recipes/linux/linux_2.6.32.bb: Jan 18 18:42:25 We should not use this kernel so add an -1 for all supported Jan 18 18:42:27 machines. Jan 18 18:42:29 Signed-off-by: Kristoffer Ericson Jan 18 18:42:31 03Kristoffer Ericson 07org.openembedded.dev * r1b292552e4 10openembedded.git/conf/distro/jlime-2009.1.conf: Jan 18 18:42:34 Rename jlime distro file. Stupid having 2009 for a file we Jan 18 18:42:36 are still perfecting, so bump it abit. Jan 18 18:42:38 Signed-off-by: Kristoffer Ericson Jan 18 18:42:40 03Kristoffer Ericson 07org.openembedded.dev * r1ae803339a 10openembedded.git/recipes/linux/linux-jlime-ben_2.6.32.bb: Jan 18 18:42:49 Lets use the linux-hpc v2.6.32-hpc branch to build for Jan 18 18:42:51 nanonote. Eventually we will merge all our 2.6.32 releases Jan 18 18:42:55 to linux-hpc, but for now this seperate nanonote .bb file Jan 18 18:42:58 kristoffer: if that's you, your first commit there has a terrible commit message. you don't actually say that the distro files you refer to are for jlime. Jan 18 18:42:59 will do. Jan 18 18:43:05 Signed-off-by:Kristoffer Ericson Jan 18 18:43:07 03Kristoffer Ericson 07org.openembedded.dev * ra212259c67 10openembedded.git/conf/distro/jlime-2009.1.conf: Add update-alternatives for distro file Jan 18 18:43:10 03Kristoffer Ericson 07org.openembedded.dev * r4a2927f4d5 10openembedded.git/recipes/linux/ (22 files in 2 dirs): Jan 18 18:43:13 Remove ben nanonote from the linux 2.6.31.bb file. We intend on using Jan 18 18:43:19 linux-hpc branches (and eventually vanilla). Jan 18 18:43:21 Signed-off-by: Kristoffer Ericson Jan 18 18:43:25 03Kristoffer Ericson 07org.openembedded.dev * r4e4b5a1626 10openembedded.git/conf/machine/ben-nanonote.conf: Jan 18 18:43:32 Add require tune-mips32 and also the currently best supported Jan 18 18:43:36 kernel for the nanonote. Jan 18 18:43:40 Signed-off-by:Kristoffer Ericson Jan 18 18:44:08 kergoth, those with jlime in their name. But agreed. Jan 18 18:44:23 kergoth, now we only got one (deleted 4) Jan 18 18:44:25 you shouldnt' be forced to look at the file list to see what happens Jan 18 18:44:25 t Jan 18 18:44:25 h Jan 18 18:44:26 a Jan 18 18:44:30 thats why our commit messages have a policy Jan 18 18:44:34 and you aren't following it Jan 18 18:45:01 kergoth, oki, will read up on it again and make sure to follow it. Jan 18 18:45:24 cool, thanks. not a big deal, just something we're trying to work on going forward Jan 18 18:45:31 makes it easier for users to see whats what looking through the logs Jan 18 18:45:53 Hmm Jan 18 18:46:02 Is there a way to have git have a template commit message? Jan 18 18:46:08 I swear BK for example could Jan 18 18:46:21 I'm pretty sure there is, but i think its via .git/, so you can't automatically get it via a clone Jan 18 18:46:24 iirc, anyway Jan 18 18:46:29 kergoth, roger understand. Been somewhat lazy with the commit messages its not my usual kernel-quality, but will do it better. Jan 18 18:46:38 kristoffer: cool :) Jan 18 18:49:36 kristoffer: the wiki has all you need Jan 18 18:49:41 http://wiki.openembedded.org/index.php/Category:Policy Jan 18 18:50:01 http://wiki.openembedded.org/index.php/Commit_log_example Jan 18 18:50:18 Laibsch, great thanks. Jan 18 18:53:09 jesus, my OE ideas and tasks omnioutliner docs are FULL of stuff.. how the hell to turn this into a cohesive list of tangible tasks sanely.. Jan 18 18:54:10 send pix Jan 18 18:55:08 anyone else having problems getting ncurses to build with bitbake? Jan 18 18:55:14 kergoth: you should send it weekly to OE list as 'list of ideas for developers' Jan 18 18:55:20 http://dl.dropbox.com/u/112715/Documents/OpenEmbedded%20Tasks.html/index.html is an old copy of one of them.. but some of those are resolved, and have added some others, and there are 4-5 other docs with random notes on it Jan 18 18:55:42 * kergoth sets aside some time this week to organize this mess Jan 18 19:02:08 ~curse rygel Jan 18 19:02:09 May you be reincarnated as a Windows XP administrator, rygel ! Jan 18 19:02:46 foo123: http://tinderbox.openembedded.org/packages/ncurses/ Jan 18 19:02:54 looks good Jan 18 19:03:15 even gulessoft.com who usually finds way to break all kinds of packages builds it successfully ;-) Jan 18 19:03:29 even gulessoft.com who usually finds all kinds of ways to break all kinds of packages builds it successfully ;-) Jan 18 19:03:40 great that means it's my fault Jan 18 19:04:37 it's probably somehow related to the TMPDIR or something with the profile Jan 18 19:09:30 kergoth: wow, that looks messy Jan 18 19:09:58 foo123: http://wiki.openembedded.net/index.php/How_do_I_send_automatic_success_and_failure_reports Jan 18 19:10:08 then we can have a look at what's failing for you Jan 18 19:11:10 i noticed on the wiki it might be possible to get access to a rsync distfile mirror... Does it exist? Who would I contact? Jan 18 19:11:45 rphillips: what page? Jan 18 19:12:06 the wiki has nothing about an rsync mirror Jan 18 19:14:07 oh your right Jan 18 19:14:09 http://wiki.openembedded.net/index.php/Sources_mirror Jan 18 19:14:16 i inferred it was a rsync mirror Jan 18 19:14:17 nm Jan 18 19:17:04 Laibsch: is there a script that generates the source mirror? Jan 18 19:19:55 for bb in `find /home/rolf/oe/openembedded.org/recipes/ -mindepth 2 -maxdepth 2 -name *[a-fh-mo-ru-z0-9].bb`; do echo Now fetching $bb ; bitbake -c fetch -b $bb ; done &> /tmp/log Jan 18 19:19:58 simple as that Jan 18 19:20:16 rphillips: why are you asking? Jan 18 19:21:59 thanks Jan 18 19:22:16 I'm working on getting an internal deployment of OE setup Jan 18 19:22:26 I'm restricted behind a firewall Jan 18 19:23:26 i've gotten my management to sign off on an OE 'prototype' Jan 18 19:25:19 svn problem? Jan 18 19:25:32 that was discussed on the mailing list just yesterday Jan 18 19:26:04 and as you can see above, I specifically try to exclude svn, cvs, hg and git recipes Jan 18 19:26:12 df Jan 18 19:26:48 I haven't checked the list within the past couple days Jan 18 19:26:52 itd be nice to have a bitbake argument like -s but which shows all versions, not just the selected version Jan 18 19:26:55 heh Jan 18 19:27:11 but no, I don't have the svn problem, at least not yet Jan 18 19:28:53 rphillips: then why do you worry about the mirror? Jan 18 19:29:09 several devs that should fetch sources from a local mirror? Jan 18 19:29:26 yes, it's going to be a team of people Jan 18 19:29:30 kergoth: bitbake -i has a "which" option Jan 18 19:30:11 rphillips: rgrep in OE for premirror Jan 18 19:30:17 PREMIRROR, even Jan 18 19:30:23 yes, but my thought was that you could use the argument to dump a list of available recipes, minus the masked out recipes, and use that to do your fetching, rather than trying to exclude things via filenames Jan 18 19:30:31 heh Jan 18 19:30:40 my recent recipe has UGLY "depends: automake-native-1.11.1" Jan 18 19:30:45 thanks Jan 18 19:31:20 Laibsch: better is to just use urls from checksums.ini ;D Jan 18 19:31:37 kergoth: the best approach (tested, but so far unsuccessful) would be to just disable git, hg, cvs and svn fetchers Jan 18 19:32:08 hrw: and where do you fetch what you haven't gotten, yet? Jan 18 19:32:24 plus, how are you sure that every file does indeed have an entry? Jan 18 19:32:28 could use an anonymous python function to rip out any urls from SRC_URI with that type. Jan 18 19:32:35 nah Jan 18 19:32:38 alternatively, could monkeypatch git/hg/cvs/svn fetch objects to do nothing Jan 18 19:32:44 I'm quite happy with what I've got, actually Jan 18 19:32:51 there may be more elegant solutions Jan 18 19:32:56 but it's working Jan 18 19:33:01 heh, can't argue with that :) Jan 18 19:33:07 I take the occasional look at apache logs Jan 18 19:33:10 and that's that Jan 18 19:33:31 if something important was missing, I'm sure it would cook up somewhere Jan 18 19:33:38 and it's only the last line of defense Jan 18 19:34:00 Laibsch: "for recipe in recipes/*/*.bb;do bitbake -cfetch $recipe;done" is even safer Jan 18 19:34:29 hrw: have you tried? Jan 18 19:34:31 ;-) Jan 18 19:34:44 * kergoth thinks about taking a walk, can't focus for some reason Jan 18 19:35:25 Laibsch: did that many times Jan 18 19:35:38 I don't think you did *that* Jan 18 19:35:42 it won't work Jan 18 19:35:45 too many files Jan 18 19:35:51 * Laibsch been there, done that Jan 18 19:35:51 Laibsch: forgot -b Jan 18 19:36:02 maybe that, too Jan 18 19:36:02 Laibsch: "for recipe in recipes/*/*.bb;do bitbake -cfetch -b $recipe;done" Jan 18 19:36:14 nope Jan 18 19:36:17 fails Jan 18 19:36:57 Laibsch: it is cleaner basically Jan 18 19:38:28 cleaner, but doesn't do the job Jan 18 19:38:32 trust me Jan 18 19:38:35 * Laibsch been there, done that Jan 18 19:38:55 kergoth: I even set FETCHCMD_cvs to /bin/true, but for some reason, even that did not work. That would have been the best solution, actually. Jan 18 19:49:43 03Marcin Juszkiewicz 07org.openembedded.dev * r9a3007faba 10openembedded.git/recipes/rygel/ (files/configure.ac.patch rygel_0.4.8.bb): (log message trimmed) Jan 18 19:49:43 rygel: added 0.4.8 (requires automake-native 1.11.1) Jan 18 19:49:43 Information from website (http://live.gnome.org/Rygel): Jan 18 19:49:43 Rygel is a collection of DLNA (UPnP AV) services (devices in UPnP Jan 18 19:49:45 speak), implemented through a plug-in mechanism. Jan 18 19:49:45 whateven Jan 18 19:49:47 Important facts and features Jan 18 19:49:49 * Specifically designed for GNOME (Mobile) and is therefore Jan 18 19:49:51 03Marcin Juszkiewicz 07org.openembedded.dev * rb33cef056c 10openembedded.git/conf/checksums.ini: checksums.ini: added rygel 0.4.8, automake 1.11.1 Jan 18 19:49:54 03Marcin Juszkiewicz 07org.openembedded.dev * r7bf770ae12 10openembedded.git/recipes/upnp/gupnp-vala_0.6.2.bb: gupnp-vala: added 0.6.2 Jan 18 19:49:57 03Marcin Juszkiewicz 07org.openembedded.dev * r331054c3af 10openembedded.git/recipes/upnp/ (gupnp-0.13.1/nodoc.patch gupnp_0.13.1.bb): gupnp: added 0.13.1 Jan 18 19:50:00 03Marcin Juszkiewicz 07org.openembedded.dev * rc82529d1b9 10openembedded.git/recipes/upnp/gssdp_0.7.1.bb: gssdp: added 0.7.1 Jan 18 19:50:03 03Marcin Juszkiewicz 07org.openembedded.dev * r3fb36bf300 10openembedded.git/recipes/upnp/ (gupnp-av-0.5.2/nodoc.patch gupnp-av_0.5.2.bb): gupnp-av: added 0.5.2 Jan 18 19:50:10 03Marcin Juszkiewicz 07org.openembedded.dev * rd9c5076be1 10openembedded.git/recipes/upnp/gupnp-tools_0.8.bb: gupnp-tools: added 0.8 Jan 18 19:51:49 Laibsch: there's FETCHCMD, but there's also FETCHCOMMAND, RESUMECOMMAND, and UPDATECOMMAND. probably needed the FETCH and UPDATE commands for cvs and svn.. but it doesn't look like the metadata can control certain scms at all (no git anywhere) :\ Jan 18 19:51:54 ah well Jan 18 20:13:07 Laibsch: I use the little snippet that hrw pointed and it works for me one difference I have is a space between -c and fetch Jan 18 20:13:57 gm all Jan 18 20:14:00 well, it did not for me Jan 18 20:14:08 Maybe something Ubuntu-specific Jan 18 20:14:21 Laibsch: I have ubuntu on all boxes :) Jan 18 20:14:30 I don't care Jan 18 20:14:38 it didn't work and that was the end of it Jan 18 20:14:43 melo is the way it is Jan 18 20:14:55 there's two things, still Jan 18 20:15:01 a) too many files Jan 18 20:15:10 b) svn recipes, etc. included Jan 18 20:15:16 actually 3 Jan 18 20:15:25 3) mine works and is in bash_history Jan 18 20:15:34 what payoff do I get by changing? Jan 18 20:15:46 none, I would think Jan 18 20:16:12 both approaches would still have the same limitations (even if both worked for me) Jan 18 20:16:20 I dont know what you have and why it was suggested .. why I use it is for checking if all recipes have correct entries in checksum.ini Jan 18 20:16:37 you fetch world? Jan 18 20:16:39 you might have different needs and what you are doing might be working well for you thats perfectly alright Jan 18 20:16:48 Laibsch: yes for the sake of OE Jan 18 20:16:52 I populate the sources mirror Jan 18 20:16:54 sometimes I do Jan 18 20:17:00 cool Jan 18 20:17:08 publically available somewhere? Jan 18 20:17:13 no Jan 18 20:17:17 bye all Jan 18 20:17:25 khem: you are aware of the official sources mirror? Jan 18 20:17:30 bye hrw Jan 18 20:17:32 good night Jan 18 20:17:35 yes Laibsch Jan 18 20:17:37 and thank you for the suggestion Jan 18 20:17:46 np Jan 18 20:17:59 khem: let me give you a list of sources I still don't have and that are not fetchable upstream Jan 18 20:18:28 Laibsch: OK Jan 18 20:18:39 http://paste.debian.net/57042/ Jan 18 20:18:52 however I might not be having much of world these days anymore because I reinstalled the boxed with 9.10 Jan 18 20:18:55 I know there are a couple in there not for publication or otherwise falsely positive Jan 18 20:19:39 just do a "cat $filelist|xargs locate" or something Jan 18 20:19:42