**** BEGIN LOGGING AT Sun Mar 07 02:59:58 2010 Mar 07 07:21:04 hi guys Mar 07 07:21:21 hi Mar 07 07:21:45 i am working on setting up a package in OE for the first time and I'm having a small issue Mar 07 07:22:48 so I made a .bb file and configure stage is completing without error. The package I am trying to build depends on gtk. However the makefile that comes out doesn't have the gtk folders in the CFLAGS Mar 07 07:23:22 gtk is a dependency in the .bb file. I'm guessing I might have to make a patch of some kind to solve this? Mar 07 07:23:54 TheRain: depends on the package i guess Mar 07 07:23:58 is gtk build before it Mar 07 07:24:10 is gtk found by configure ? Mar 07 07:24:41 yah, gtk is there and other packages that build against it are building fine Mar 07 07:25:07 i don't know if it's found by configure or not.... how would I tell? Mar 07 07:26:44 actually, one sec Mar 07 07:27:51 TheRain: go to the work dir Mar 07 07:27:55 there's no mention of gtk in the configure log, or in the configure file Mar 07 07:27:56 soemwhere below is the dir you build Mar 07 07:28:24 then my guess is that you need to explicitly enable it when calling configure Mar 07 07:28:34 ok hmm Mar 07 07:28:57 look at other recipes and check for lines like OECONF or OEEXTRACONF or so Mar 07 07:29:19 ok will do Mar 07 07:29:24 EXTRA_OECONF (thanks to my goodl friend grep) Mar 07 07:30:52 ok, giving that a shot Mar 07 07:33:11 awesome! --enable-gtk+ fixed it Mar 07 07:33:16 thank you very much Mar 07 07:35:35 yw Mar 07 07:35:40 and have a nice day :-) Mar 07 07:39:05 woops, I was mistaken... that didn't work, but i'm sure that's not the flag I should set to do it anyway. still looking for the correct one if there is one Mar 07 11:18:04 hi all, yesterday pb_ told me about using _virtual-native. I was wondering if there is something similar for target packages. Mar 07 11:19:17 that way if there is one file not needed for native, it can be added for target only (currently I end up overriding e.g. DEPENDS with DEPENDS_virtclass-native and copying most of DEPENDS in it, but I'd rather use something like DEPENDS_target += (which ofc is not executed for native) Mar 07 11:21:34 no, there isn't currently any override like that. Mar 07 11:22:22 to avoid duplicating the contents of DEPENDS you can do something along the lines of: Mar 07 11:22:39 effem_is_cool = "target-only-package" Mar 07 11:22:47 effem_is_cool_virtclass-native = "" Mar 07 11:22:58 DEPENDS = "${effem_is_cool} rest of depends" Mar 07 11:23:08 pb__: I get the point, thanks! Mar 07 11:23:46 (although maybe having a _target override would be cool too, not sure whether I am capable to add that though) Mar 07 11:42:15 03Martin Jansa  07org.openembedded.dev * r1ad4ccb8e6 10openembedded.git/recipes/util-linux-ng/util-linux-ng.inc: Mar 07 11:42:15 util-linux-ng: use u-a for /etc/default/mountall (collision with busybox-mountall) Mar 07 11:42:15 Signed-off-by: Martin Jansa Mar 07 11:42:26 03Valery Febvre  07org.openembedded.dev * r904b132c96 10openembedded.git/recipes/openmoko-3rdparty/podboy_svn.bb: (log message trimmed) Mar 07 11:42:26 podboy: update to 1.5.0 Mar 07 11:42:26 Changes since 1.3.4 Mar 07 11:42:26 * New feature: played episodes that are older than a specified amount of days (defined in Settings) can be auto-deleted on every startup. Mar 07 11:42:27 * New feature: episodes can now be tagged as "(un)deletable" in page "Episodes". All episodes with status "undeletable" will be ignored when auto-deletion of old episodes will be done on startup. Mar 07 11:42:27 * New feature: podcasts can be renamed. Mar 07 11:42:28 * New feature: downloads can be canceled. Mar 07 11:42:28 03Martin Jansa  07org.openembedded.dev * r289532f007 10openembedded.git/recipes/busybox/busybox.inc: Mar 07 11:42:29 busybox: use u-a for /etc/default/mountall (collision with util-linux-ng-mountall) Mar 07 11:42:29 Signed-off-by: Martin Jansa Mar 07 11:45:01 03Martin Jansa  07org.openembedded.dev * r7db0a57210 10openembedded.git/recipes/util-linux-ng/util-linux-ng.inc: util-linux-ng: missing PR bump after u-a change Mar 07 12:58:04 yo mickey Mar 07 12:58:32 sports, that does not really sound like computing, sounds more like something that is actually good for you :-) Mar 07 13:08:55 hi eFfeM1 Mar 07 13:08:58 indeed Mar 07 13:09:08 i enjoy sports a lot Mar 07 13:09:19 squash once a week Mar 07 13:09:23 * eFfeM1 did spent an hour in the forest with his ATB Mar 07 13:09:24 tabletennis (almost) twice a week Mar 07 13:09:36 atbß Mar 07 13:09:37 ? Mar 07 13:09:55 all terrain bike Mar 07 13:10:07 aka mountain bike, but that seems a pretty weird name for the netherlands :-) Mar 07 13:10:49 many packages using autotools seem to have only autotools_stage_all in do_stage() Mar 07 13:12:16 do_stage and later stages by default do not use ${S} any more do they? Mar 07 13:12:31 * eFfeM1 wonders about do_stage () { cp ${S}/clutter.pc ${S}/clutter-0.4.pc autotools_stage_all Mar 07 13:13:27 another beauty: "cmake/cmake-native_2.6.4.bb Mar 07 13:13:28 that's usually not necessary Mar 07 13:13:36 there's a lot of historic cruft in OE, no doubt Mar 07 13:14:04 well, let's say I feel it is time for spring cleanig :-) Mar 07 13:14:32 not just you Mar 07 13:14:35 many of us feel that Mar 07 13:14:42 and some are finally starting to do something about it Mar 07 13:15:19 i'm building console-image with distro minimal in my branch if that works out reasonably well I'll push the branch so others can have a go at it too Mar 07 13:15:51 but doing a lot of edits atm that touch many files so at this point it is better that I spend a little time before pushing Mar 07 13:15:52 awesome Mar 07 13:16:00 hope to be able to push later today Mar 07 13:16:04 cool stuff. Mar 07 13:16:06 * eFfeM1 keeps fingers crossed Mar 07 13:16:08 ty Mar 07 13:16:19 i see some momentum behind a no-legacy-branch Mar 07 13:16:28 i guess something will come along this month Mar 07 13:16:44 yes Mar 07 13:17:50 mickeyl: is the DESTDIR part here useful: oe_runmake 'DESTDIR=${D}' install Mar 07 13:18:20 yes Mar 07 13:18:39 compare w/ autotools.bbclass; it does something similar Mar 07 13:19:10 ah ok, will Mar 07 13:19:26 actually this package inherits autotools Mar 07 13:20:07 hmm Mar 07 13:20:15 in that case i wonder why it overrides do_stage at all Mar 07 13:20:19 or do_install Mar 07 13:26:47 actually i'm now removing do_stage functions that only call autotools_stage_all which seems pretty pointless since they inherit autotools, saw the other construct while editing Mar 07 13:28:06 eFfeM1: I think it should be possible for the likes of native.bbclass to check and filter out ${PN}-native from their own depends Mar 07 13:28:24 eFfeM1: IMO a cleaner option Mar 07 13:31:01 XorA: guess you're right, i am not an expert on the build system Mar 07 13:31:33 but it could also be useful at other places. E.g. I saw that some other dependencies are sometimes also not there for target Mar 07 13:31:45 or a target specific patch that is not needed for native Mar 07 13:31:50 there are more options Mar 07 13:32:04 eFfeM1: yeah I have a couple of recipes with the same issue Mar 07 13:32:21 I had to do BASE_SRC_URI messing around Mar 07 13:36:29 yeah, yuk Mar 07 13:36:37 recipes/eggdbus/eggdbus_0.6.bb Mar 07 13:36:43 is a prime example of messiness Mar 07 13:47:17 03Michael Lippautz  07org.openembedded.dev * rff2dd30ccc 10openembedded.git/recipes/python/python-wsgiref_0.1.2.bb: Mar 07 13:47:17 python-wsgiref: Add new recipe Mar 07 13:47:17 * Adds Web Server Gateway Interface (WSGI) Framework Mar 07 13:50:33 XorA: seen worse recibes than the eggdbus one Mar 07 13:50:40 afk coffee break Mar 07 13:52:53 * XorA is finished here, and its sunny in Scotland which is a 1 in a million so Im off out for a bit Mar 07 14:34:13 03Michael Lippautz  07org.openembedded.dev * ra0b042d1f0 10openembedded.git/recipes/libol/libol_0.3.18.bb: libol: Convert to new staging Mar 07 15:27:11 Romke: did you get a suitable answer yesterday, saw you replying whenthe system was shutting down, so could not answer Mar 07 15:30:57 yeah, i said something about git Mar 07 15:31:01 i'm new to git ;) Mar 07 15:31:19 i'm an svn user Mar 07 15:31:49 well if they are just new files you might also just send the files and a description, but if it is complicated git is definitely preferred Mar 07 15:32:02 and atm I cannot add them, so send to the mailing list Mar 07 15:32:34 ah, well, i will find out how git works Mar 07 15:32:54 knowing how git works won't kill me Mar 07 15:33:03 it is not that difficult, check out a tree, git add your new files Mar 07 15:33:24 then git commit, git format-patch; git send-email Mar 07 15:33:30 there are lots of tutorials out there Mar 07 15:33:35 yeah Mar 07 15:33:46 i just need to find the time and will to do it :P Mar 07 18:03:03 hi xora Mar 07 18:03:14 hey woglinde_ Mar 07 18:04:03 only flying visit from me before off out again :-) Mar 07 18:04:12 hm Mar 07 18:04:20 nothing happens Mar 07 18:47:40 anyone else having problems to connect to the gnu mirrors ? Mar 07 18:50:46 sunsite.ust.hk does not seem to exist any more and and ftp.cs.ubc.ca and ftp.ayamura.org do not respond any more Mar 07 18:56:08 morning Mar 07 18:58:19 hehe Mar 07 18:58:23 morning hrw Mar 07 19:00:45 ~curse koen for his gnutls overreaction Mar 07 19:00:46 May you be reincarnated as a Windows XP administrator, koen for his gnutls overreaction ! Mar 07 19:00:59 hm Mar 07 19:01:07 Is someone else getting problem with expat 2.0.2? after latest pull today I get crc error on the package Mar 07 19:01:10 so far only gnome-vfs failed here. and this is 2.24 so over year old? Mar 07 19:01:23 palisv try to delete it from archive Mar 07 19:01:34 your local archive Mar 07 19:02:14 yes, I deleted the tgz file from sources a couple of times but the downloaded package always get wrong crc, (it was 2.0.1 i ment) Mar 07 19:02:42 let me fetch it Mar 07 19:03:25 ee8b492592568805593f81f8cdf2a04c /home/hrw/devel/sources/dl_dir/expat-2.0.1.tar.gz Mar 07 19:03:33 fresh fetch Mar 07 19:04:18 same as in recipe Mar 07 19:04:32 NOTE: Running task 4506 of 15839 (ID: 8018, /home/hrw/devel/oe/openembedded/recipes/xorg-proto/xf86dgaproto_2.1.bb, do_package_stage) Mar 07 19:04:39 angstrom testing build Mar 07 19:09:36 yes, the md5 is correct on my file also, but if I do a "tar tvfz expat-2.0.1.tar.gz" i get gzip: stdin: invalid compressed data--crc error Mar 07 19:10:41 unpacked fine here Mar 07 19:12:20 strange, Mar 07 19:30:46 fixed it by download a fresh expat 2.0.1 from sourceforge and created a new md5 Mar 07 19:32:59 hrw, wrt the curse: what's new ? Mar 07 19:33:21 but XP administrator seems a very harsh punishment :-) Mar 07 19:33:34 Hi all. Running u-boot in qemu has a terrible networking performance. Ping command takes 10 seconds or more. TFTP transfer of kernel image dies with timeout. Any ideas? Mar 07 19:34:43 eFfeM1: gnutls 2.8.5 got uploaded, fixed security bugs. so far angstrom reverted to 2.4.2 because of 'build failures' which for me so far came only in gnome-vfs ;D Mar 07 19:34:47 eFfeM1, hi I'll push the xorg.conf very soon(need to double check everything before commit and I'll do it) Mar 07 19:35:03 anyway 10k tasks to finish my test builds Mar 07 19:35:14 * hrw -> off now Mar 07 19:36:40 GNUtoo: great, but don't think I have time for it today Mar 07 19:36:45 hrw, cya Mar 07 19:36:50 but in the todo list we have: Mar 07 19:37:01 ok Mar 07 19:37:15 np Mar 07 19:37:15 can I still describe the todo list? Mar 07 19:37:31 sure, i have time to chat, just not to build, my system is humming at soemthign else Mar 07 19:37:35 ok Mar 07 19:37:40 tkask 1986 of 2988 :-) Mar 07 19:38:08 *fix init symlink: init get symlinked to busybox instead of init.sysvinti in the sysvinit package postinstall Mar 07 19:38:17 *fix password handling Mar 07 19:38:25 (maybe that's a local bug for me) Mar 07 19:38:36 *xorg.conf that I'll push Mar 07 19:40:19 sounds ok Mar 07 19:42:41 hi all Mar 07 19:42:53 hi florian Mar 07 19:43:00 hi woglinde Mar 07 19:46:21 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * rfdd1165ca6 10openembedded.git/recipes/xorg-xserver/ (xserver-xorg-conf/eee701/xorg.conf xserver-xorg-conf_0.1.bb): (log message trimmed) Mar 07 19:46:21 xserver-xorg-conf : fix AllowEmptyInput problem for eee701 and bump PR Mar 07 19:46:21 In newer Xorg versions AllowEmptyInput is on by default, Mar 07 19:46:21 And that appear in Xorg.0.log of the eee701 machine: Mar 07 19:46:21 (WW) AllowEmptyInput is on, devices using drivers 'kbd', 'mouse' or 'vmmouse' will be disabled. Mar 07 19:46:22 (WW) Disabling Keyboard Mar 07 19:46:23 (WW) Disabling Mouse Mar 07 19:52:50 Hi all. Running u-boot in qemu has a terrible networking performance. Ping command takes 10 seconds or more. TFTP transfer of kernel image dies with timeout. Any ideas? Mar 07 19:55:46 drasko, did you read qemu FAQ...they said things like ping doesn't work inside qemu Mar 07 19:55:53 maybe it's not related but it could be Mar 07 19:56:30 I don't know well TFTP at the network level but if it uses icmp.... Mar 07 19:56:59 all that I know is that it uses udp with its own control protocol on top of it Mar 07 21:38:08 meh. Mar 07 21:49:14 indeed Mar 07 21:49:45 Blah Mar 07 21:50:06 pb__ in your branch does all of the gcc versions have FILESPATHPKG .= gcc-${PV} ? Mar 07 21:50:41 hm, dunno. let me have a quick look. Mar 07 21:51:34 no, they don't Mar 07 21:51:47 gcc-4.1.0.inc:FILESPATHPKG .= ":gcc-4.1.1" Mar 07 21:51:47 gcc-4.4.1.inc:FILESPATHPKG .= ":gcc-$PV" Mar 07 21:51:47 gcc-4.4.2.inc:FILESPATHPKG .= ":gcc-$PV" Mar 07 21:51:47 gcc-4.4.3.inc:FILESPATHPKG .= ":gcc-$PV" Mar 07 21:51:47 gcc-csl-arm-2007q3.inc:FILESPATH = "${FILE_DIRNAME}/gcc-csl-arm-2007q3:${FILE_DIRNAME}/gcc-csl-arm" Mar 07 21:51:48 gcc-svn.inc:FILESPATHPKG .= ":gcc-svn" Mar 07 21:51:53 that's all I have, probably the same as .dev Mar 07 21:52:40 it is also a bit sad that we have (for example) both gcc-4.4.2.inc and gcc-4.4.3.inc which are nearly identical Mar 08 00:30:16 03Chris Larson  07org.openembedded.dev * rf285e1e754 10openembedded.git/ (12 files in 2 dirs): Mar 08 00:30:16 m4: use autotools_stage, checksum in SRC_URI, switch to .inc, add 1.4.14 Mar 08 00:30:16 Signed-off-by: Chris Larson **** ENDING LOGGING AT Mon Mar 08 02:59:57 2010