**** BEGIN LOGGING AT Sat Aug 08 02:59:57 2009 Aug 08 03:27:34 hmm Aug 08 03:33:08 03Kirtika Ruchandani  07org.openembedded.dev * redf9505ccc 10openembedded.git/recipes/wvstreams/ (3 files in 2 dirs): Aug 08 03:33:08 wvstreams_4.4.1.bb: Make it compilable with gcc 4.3 Aug 08 03:33:08 Signed-off-by: Khem Raj Aug 08 03:56:23 Yo ml Aug 08 03:56:31 Jom Aug 08 03:56:39 Yo Aug 08 04:01:36 Hi Aug 08 04:12:56 03Khem Raj  07org.openembedded.dev * r6cb210accf 10openembedded.git/ (3 files in 2 dirs): Aug 08 04:12:56 wvstreams_4.6.bb: Add new recipe. Aug 08 04:12:56 This is the latest release available as of now. Aug 08 04:12:56 it does not compile on uclibc based system for Aug 08 04:12:56 now. Aug 08 04:12:58 Signed-off-by: Khem Raj Aug 08 04:34:31 hrm, my newcollection.inc doesn't behave with upstream oe.. Aug 08 04:46:22 Why does bitbake create packages of the kind lib, when I want just the standard ${PN} and ${PN}-dev packages ? Aug 08 04:46:48 I have changed libs to misc/base in my SECTION var, but to no effect Aug 08 04:49:14 inherit debian does that Aug 08 04:49:59 Tartarus: but I dont have that Aug 08 04:50:37 Tartarus: oh, that could be in one of my conf files you mean ? (I am using ipk and not deb though) Aug 08 05:15:32 rkirti: its probably your DISTRO's .conf doing it. Aug 08 05:15:43 distro generally makes those kinds of policy decisions Aug 08 05:17:30 * rkirti surfs through the Angstrom config files Aug 08 05:21:22 03Khem Raj  07org.openembedded.dev * rd8839b3726 10openembedded.git/recipes/ccache/ (ccache_2.4.bb files/ccache-gcov.patch): Aug 08 05:21:22 ccache_2.4.bb: Cope with .gcno files generated for coverage analysis. Aug 08 05:21:22 Signed-off-by: Khem Raj Aug 08 06:21:03 * * OE Bug 5282 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Aug 08 06:21:05 * * uClibc 0.9.30.1 / gcc-cross-4.4.0 libstdc++v3 build failure (possibly locale related) Aug 08 06:21:07 * * http://bugs.openembedded.net/show_bug.cgi?id=5282 Aug 08 06:30:03 * * OE Bug 5281 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Aug 08 06:30:05 * * ffmpeg_git do_configure fails due to broken git URL Aug 08 06:30:07 * * http://bugs.openembedded.net/show_bug.cgi?id=5281 Aug 08 06:32:05 * * OE Bug 5280 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Aug 08 06:32:07 * * fakeroot-native_1.12.1.bb do_fetch failed Aug 08 06:32:09 * * http://bugs.openembedded.net/show_bug.cgi?id=5280 Aug 08 06:40:12 03Marc Poulhiès  07org.openembedded.dev * r02850e2af9 10openembedded.git/recipes/gpe-mixer/ (gpe-mixer/fix-install.patch gpe-mixer_0.42.bb): Aug 08 06:40:12 gpe-mixer_0.42.bb: Fix fix-install.patch to apply cleanly. Aug 08 06:40:12 Signed-off-by: Khem Raj Aug 08 06:53:04 * * OE Bug 5278 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Aug 08 06:53:06 * * wrong SRC_URI for ttf-sazanami Aug 08 06:53:08 * * http://bugs.openembedded.net/show_bug.cgi?id=5278 Aug 08 07:03:03 * * OE Bug 5271 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Aug 08 07:03:05 * * emotion_svn fails to build Aug 08 07:03:07 * * http://bugs.openembedded.net/show_bug.cgi?id=5271 Aug 08 07:32:03 * * OE Bug 5277 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Aug 08 07:32:05 * * wmiconfig_svn fails due to 'no GNU_HASH' bug Aug 08 07:32:07 * * http://bugs.openembedded.net/show_bug.cgi?id=5277 Aug 08 07:36:04 * * OE Bug 5089 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Aug 08 07:36:06 * * New Recipe for mumpot-0.5 Aug 08 07:36:08 * * http://bugs.openembedded.net/show_bug.cgi?id=5089 Aug 08 07:39:05 * * OE Bug 5130 has been RESOLVED (FIXED) by raj.khem(AT)gmail.com Aug 08 07:39:07 * * Coreutils-native_7.2 fails to compile Aug 08 07:39:09 * * http://bugs.openembedded.net/show_bug.cgi?id=5130 Aug 08 07:46:40 03Khem Raj  07org.openembedded.dev * r7fc8160b0e 10openembedded.git/ (2 files in 2 dirs): Aug 08 07:46:40 ttf-sazanami_20040629.bb: Update SRC_URI. Aug 08 07:46:40 Signed-off-by: Khem Raj Aug 08 07:46:41 03Christopher Friedt  07org.openembedded.dev * r2729982f02 10openembedded.git/recipes/efl1/emotion_svn.bb: Aug 08 07:46:41 emotion_svn.bb: Make it look for edje_cc in STAGING_DIR. Aug 08 07:46:45 Signed-off-by: Khem Raj Aug 08 07:46:47 03Khem Raj  07org.openembedded.dev * ra702c1ad05 10openembedded.git/recipes/wmiconfig/wmiconfig_svn.bb: Aug 08 07:46:50 wmiconfig_svn.bb: Fix build error due to choice of wrong --hash-style. Aug 08 07:46:52 Signed-off-by: Khem Raj Aug 08 08:36:50 ~seen khem Aug 08 08:36:51 khem is currently on #oe (3d 12h 18m 54s) #uclibc (3d 12h 18m 54s). Has said a total of 13 messages. Is idling for 2d 10h 17m 43s, last said: 'I think devshell can be powerful but its under developed'. Aug 08 08:39:05 g'morning woglinde Aug 08 08:59:29 hi rkirti Aug 08 08:59:31 hi jay7 Aug 08 08:59:40 woglinde: morning Aug 08 09:01:24 http://tinderbox.openembedded.net/builds/22631/ Aug 08 09:01:34 | grep: .config: No such file or directory Aug 08 09:01:56 * Jay7 have problem when building angstrom for tosa and akita Aug 08 09:04:08 hm.. according to stats ant have success with the same linux-rp version Aug 08 09:10:21 hehe.. direct bitbake -c rebuild linux-rp is working ok Aug 08 09:10:54 fakeroot was build before Aug 08 09:21:55 03Steve Sakoman  07org.openembedded.dev * r8e40f0f867 10openembedded.git/ (conf/checksums.ini recipes/alsa/alsa-plugins_1.0.20.bb): alsa-plugins: add 1.0.20 Aug 08 09:21:56 03Steve Sakoman  07org.openembedded.dev * ra40dd855b6 10openembedded.git/recipes/desktop-file-utils/desktop-file-utils_0.15.bb: Aug 08 09:21:56 desktop-file-utils: add 0.15 Aug 08 09:21:56 md5/sha256 already in checksums.ini Aug 08 09:21:59 03Steve Sakoman  07org.openembedded.dev * r1ebbc0245e 10openembedded.git/ (conf/checksums.ini recipes/devicekit/devicekit-power_009.bb): devicekit-power: add recipe for devicekit-power version 009 Aug 08 09:22:02 03Steve Sakoman  07org.openembedded.dev * rcfa8d5252e 10openembedded.git/ (conf/checksums.ini recipes/eel/eel_2.26.0.bb): eel: add 2.26.0 Aug 08 09:22:05 03Steve Sakoman  07org.openembedded.dev * r7ffde32d35 10openembedded.git/ (conf/checksums.ini recipes/glibmm/glibmm_2.21.2.bb): glibmm: add 2.21.2 Aug 08 09:22:08 03Steve Sakoman  07org.openembedded.dev * rb47a02e9ee 10openembedded.git/ (conf/checksums.ini recipes/gcalctool/gcalctool_5.27.5.bb): gcalctool: add 5.27.5 Aug 08 09:22:11 03Steve Sakoman  07org.openembedded.dev * rdb6f9ba936 10openembedded.git/recipes/gnome/gnome-power-manager_2.24.2.bb: gnome-power-manager: fix broken build - don't build man Aug 08 09:22:18 03Steve Sakoman  07org.openembedded.dev * rab91c4b0b5 10openembedded.git/ (conf/checksums.ini recipes/gnome/at-spi_1.26.0.bb): at-spi: add 1.26.0 Aug 08 09:22:21 03Steve Sakoman  07org.openembedded.dev * r2bc6b78b36 10openembedded.git/ (conf/checksums.ini recipes/gnome/libgail-gnome_1.20.1.bb): libgail-gnome: add 1.20.1 Aug 08 09:22:24 03Steve Sakoman  07org.openembedded.dev * r6e4f48a748 10openembedded.git/classes/gconf.bbclass: Aug 08 09:22:27 gconf.bbclass: adjust the ordering of things so that a package pkg_postinst_append is executed after the class postinst. Aug 08 09:22:30 This is important since recipes like gpanel need to to execute gconftool-2 --load after the gconf class code has processed the schema info Aug 08 09:22:33 03Steve Sakoman  07org.openembedded.dev * r30f6512165 10openembedded.git/ (conf/checksums.ini recipes/atk/atk_1.26.0.bb): atk: add 1.26.0 Aug 08 09:22:40 03Steve Sakoman  07org.openembedded.dev * raa5fa3a8bd 10openembedded.git/ (conf/checksums.ini recipes/gnome/gedit_2.26.2.bb): gedit: add 2.6.22 Aug 08 09:22:43 03Steve Sakoman  07org.openembedded.dev * r853caf13db 10openembedded.git/ (2 files in 2 dirs): gtksourceview: add 2.7.3 Aug 08 09:22:46 03Steve Sakoman  07org.openembedded.dev * r795240d1c1 10openembedded.git/ (conf/checksums.ini recipes/gnome/gnome-terminal_2.26.2.bb): gnome-terminal: add 2.26.2 Aug 08 09:22:51 03Steve Sakoman  07org.openembedded.dev * r7eed60cd78 10openembedded.git/ (2 files in 2 dirs): gnome-python-extras: add 2.25.3 Aug 08 09:22:54 03Steve Sakoman  07org.openembedded.dev * r6c47435bdb 10openembedded.git/classes/image.bbclass: image.bbclass: fix bug in lingua code (code wasn't spitting entries like fr-fr and thus locale-fr packages were bing omitted) Aug 08 09:22:58 03Steve Sakoman  07org.openembedded.dev * r691e173d19 10openembedded.git/ (4 files in 3 dirs): gconf: add 2.26.2 Aug 08 09:23:03 03Steve Sakoman  07org.openembedded.dev * rb41a97d3bb 10openembedded.git/ (conf/checksums.ini recipes/vte/vte_0.20.0.bb): vte: add 0.20.0 Aug 08 09:23:31 any idea why opkg(native) could hang forever? rootfs now taking 9h Aug 08 09:23:40 doesn't seem to do anything Aug 08 09:24:04 mickeyl: I've remember some talks about waiting for user input from some package Aug 08 09:24:18 hmm, that could be possible Aug 08 09:24:23 then again, it's taking 98% of CPU Aug 08 09:24:24 try to send -SIGCONT e.g. Aug 08 09:24:32 ah.. Aug 08 09:24:35 not sure then Aug 08 09:25:02 stefan_schmidt: can you build images atm.? my do_rootfs is hanging forever Aug 08 09:25:23 * Jay7 is building angstrom for akita right now Aug 08 09:25:25 mickeyl: hi Aug 08 09:25:30 mickeyl: from .dev you mean? Aug 08 09:25:33 yes Aug 08 09:25:44 mickeyl: I build an console image yesterday evening Aug 08 09:25:57 bummer, then it's something local here :/ Aug 08 09:26:05 mickeyl: zhone-image is failing with -lXext linkening problems in evas-native here Aug 08 09:26:21 mickeyl: I just removed my tmp for .dev on the buildhost Aug 08 09:26:29 Let me trigger a new build Aug 08 09:26:37 ok Aug 08 09:26:44 buildhost was full btw. Aug 08 09:26:48 that killed my build :) Aug 08 09:26:49 mickeyl: That is a known problem Aug 08 09:27:03 Laibsch1: really? since when and what's the fix? Aug 08 09:27:08 mickeyl: yeah, that's the reason I deleted my tmp :) Aug 08 09:27:14 most likely cause is a problem with bluez Aug 08 09:27:24 let me dig up the patch I posted to the list Aug 08 09:27:56 mickeyl: We need to talk to shoragan about space on the buildhost. What brings a 8 core machine if we run out of space :) Aug 08 09:28:45 !oebug 5133 Aug 08 09:28:47 * * Bug 5133, Status: CONFIRMED, Created: 2009-05-28 19:26 Aug 08 09:28:48 * * matt(AT)genesi-usa.com: do_rootfs hangs when opkg-cl asks for input Aug 08 09:28:49 * * http://bugs.openembedded.org/show_bug.cgi?id=5133 Aug 08 09:28:52 stefan_schmidt: right Aug 08 09:28:52 mickeyl: and http://patchwork.openembedded.org/patch/716/ Aug 08 09:29:07 Laibsch: thanks Aug 08 09:29:08 mickeyl: ok, triggered a new build for fso-console-image Aug 08 09:29:17 gern geschehen Aug 08 09:29:33 It's kind of sad that it's still unfixed Aug 08 09:29:50 although I agree that the patch I posted isn't the proper solution Aug 08 09:30:09 well Aug 08 09:30:18 by now i'm afraid we can call opkg unmaintained (again) Aug 08 09:30:31 mickeyl: What happened to tick? Aug 08 09:30:34 i doubt that tick and his friends are working on it Aug 08 09:30:40 nothing i know of Aug 08 09:30:49 but they seem to have moved to other things with 0xlab Aug 08 09:30:50 * stefan_schmidt is not on the opkg list Aug 08 09:30:53 like... android and stuff Aug 08 09:30:56 *cough* Aug 08 09:31:02 anyone asked tick? Aug 08 09:31:39 In this case, it's not really an opkg bug Aug 08 09:31:45 it's a bug in the metadata Aug 08 09:31:55 some bluez mess Aug 08 09:31:58 fun Aug 08 09:32:01 bluez3 vs bluez4 Aug 08 09:32:07 * mickeyl changes distro back to his own one Aug 08 09:32:14 I tried to untangle it and made a few commits to that end Aug 08 09:32:30 Eventually, I got fed up and just ripped out anything bluez-related Aug 08 09:32:37 I don't use bluez anyway ;-) Aug 08 09:32:49 mickeyl: minimal? Aug 08 09:33:09 yeah, that seemed to build things last time i used it Aug 08 09:33:21 it's been a long time for me without .dev. I'm back now. Aug 08 09:33:43 actually, it's more of a problem in minimal Aug 08 09:33:57 angstrom made a decision not to support bluez3 Aug 08 09:34:12 AFAIU, minimal tried to impose less such restrictions Aug 08 09:34:42 the problem is that bluez4 is not split up into the smaller -utils and other packages Aug 08 09:34:53 thus, both bluez3 and bluez4 packages get built Aug 08 09:35:05 and then they conflict upon installation in do_rootfs Aug 08 09:35:17 It should be fairly straight-forward to fix Aug 08 09:35:20 Please note that I don't have this problem here Aug 08 09:35:26 but the devil is in the details Aug 08 09:35:35 stefan_schmidt: are you building for angstrom? Aug 08 09:35:36 Build a lot images during tha last days and it never happened to me Aug 08 09:35:39 Laibsch: yes Aug 08 09:36:01 they took care of the problem by blocking bluez3 out Aug 08 09:36:14 that doesn't mean the metadata itself is fixed Aug 08 09:36:33 it's more or less just masking the problem out, which is OK for a distro to do Aug 08 09:36:38 could be, just stated that it does not happen to me Aug 08 09:36:46 yes, thanks Aug 08 09:36:58 I'm wondering as mickeyl will likely build the same images... Aug 08 09:37:22 yes, this time i tried angstrom, but i can't seem to get any images done w/ it Aug 08 09:37:40 *shrug* Aug 08 09:37:42 http://gitweb.openembedded.net/?p=openembedded.git&a=commit;h=20f4fa12b285b8f41b1a215415a5732d633a0a83 Aug 08 09:37:57 that is what fixed it for Angstrom Aug 08 09:38:13 supposedly Aug 08 09:38:26 mickeyl: what image do you build? fso-console-image for gta02? Aug 08 09:38:34 stefan_schmidt: correct, exactly that Aug 08 09:38:46 mickeyl: I build that yesterday. Aug 08 09:39:01 mickeyl: Let's wait until it build from scratch here again Aug 08 09:39:13 i wonder how this commit can get through Aug 08 09:39:18 it's megaugly Aug 08 09:39:29 *shrug* Aug 08 09:40:29 mickeyl: that commit is gone Aug 08 09:40:37 mickeyl: you guys are using mega old metadata Aug 08 09:41:02 doubtful Aug 08 09:41:06 * mickeyl is on org.oe.dev HEAD Aug 08 09:41:19 anyways, Jay7 was right Aug 08 09:41:25 mickeyl: then look at the actual git log, I reverted that commit and replaced it with something working Aug 08 09:41:28 http://pastebin.ca/1521706 Aug 08 09:42:35 opkg needs a -yes mode Aug 08 09:42:42 or we should call it with yes | opkg Aug 08 09:43:00 it's unacceptable that it just stalls with a question in image-build mode Aug 08 09:43:04 mickeyl: it does has a -yes mode, and look in OM git tree I enabled it for image building Aug 08 09:43:15 mickeyl: but the patch was rejected for OE Aug 08 09:43:40 why? Aug 08 09:43:55 on what terms, i mean Aug 08 09:43:57 because its not deterministic Aug 08 09:44:09 and I kinda agreed Aug 08 09:44:12 so never pushed it Aug 08 09:44:34 what opkg really needs is an -always-fail switch Aug 08 09:50:47 yes, that might help Aug 08 10:07:23 probably called -non-interactiv Aug 08 10:13:38 03Koen Kooi  07org.openembedded.dev * re96d3bd8b2 10openembedded.git/conf/distro/angstrom-2008.1.conf: angstrom: prefer fribidi 0.19.x over libfribidi 0.10.x Aug 08 10:13:39 03Koen Kooi  07org.openembedded.dev * r87ae19b3f6 10openembedded.git/recipes/gnome/gedit_2.26.2.bb: gedit: depend on iso-codes Aug 08 10:13:39 03Koen Kooi  07org.openembedded.dev * rf957642ad2 10openembedded.git/ (conf/checksums.ini recipes/libfribidi/fribidi_0.19.2.bb): fribidi: add 0.19.2 Aug 08 10:26:29 re Aug 08 10:29:43 micro micro distro Aug 08 10:30:35 hi mrmoku Aug 08 10:34:30 03Klaus Kurzmann  07shr/import * r23c0b8d881 10openembedded.git/recipes/callrec/callrec.bb: Aug 08 10:34:30 callrec: rename recipe / get scenario path from frameworkd.conf Aug 08 10:34:30 Signed-off-by: Klaus Kurzmann Aug 08 10:34:30 woglinde, hi Aug 08 10:34:40 03Klaus Kurzmann  07shr/import * rd79130b4d5 10openembedded.git/recipes/shr/initscripts-shr_git.bb: Aug 08 10:34:40 initscripts-shr: bump PR Aug 08 10:34:40 Signed-off-by: Klaus Kurzmann Aug 08 10:36:29 hehe.. something is wrong with kernels building Aug 08 10:36:40 now I have same trouble with linux-kexecboot Aug 08 10:37:35 bitbake x11-image failing but bitbake linux-kexecboot is working Aug 08 10:37:53 then doing again x11-image continuing to build Aug 08 12:07:26 0 Aug 08 12:48:53 03Stanislav Brabec  07org.openembedded.dev * r168db79585 10openembedded.git/recipes/pkgconfig/pkgconfig_0.23.bb: pkgconfig: Bump PR to keep incremental versioning. Aug 08 12:48:53 03Stanislav Brabec  07org.openembedded.dev * r55e26ebf17 10openembedded.git/recipes/gmp/gmp_4.2.4.bb: gmp: Bump PR to keep incremental versioning. Aug 08 12:48:54 03Stanislav Brabec  07org.openembedded.dev * r08df94776a 10openembedded.git/recipes/xorg-xserver/xserver-kdrive-1300_1.3.0.0.bb: xserver-kdrive-1300: Bump PR to keep incremental versioning. Aug 08 13:08:58 is there a way to force oe to redownload things even if DL_DIR is populated ? Aug 08 13:09:55 (only for particular packages) I have some obsolete url's in my tree (I dont know exactly which packages)...and removing DL_DIR altogether would be a real showstopper with a crappy bandwidth here. Aug 08 13:13:22 rkirti, delete only the stuff you want to redownload? Aug 08 13:15:12 Crofton|work: which is the issue. Since I have a huge number of recipes, and dont know which of them have SRC_URIs that are obsolete Aug 08 13:28:10 ah Aug 08 13:28:24 check in contrib and see if there is a script that can help Aug 08 13:28:27 he crofton Aug 08 13:29:14 sound like an idea for the site kergoth mentioned Aug 08 13:29:55 http://openembedded.uservoice.com/pages/6963-general Aug 08 13:30:38 Crofton : oh thanks for that link. I have been trying to get it for quite some time. Aug 08 13:31:08 * rkirti has quite a few rants about do_clean not removing ipks and arguments about why it should :-/ Aug 08 13:31:25 I like the voting idea Aug 08 13:31:42 it lets the people that hack at stuff know what people are really interested in Aug 08 13:32:01 I'll re-mention it on the list and ask the website guys to add a link Aug 08 13:32:06 maybe they have ... Aug 08 13:33:28 can't find be search Aug 08 13:45:09 hi mwester Aug 08 13:47:47 ~seen florian Aug 08 13:48:12 florian was last seen on IRC in channel #oe, 15h 17m 21s ago, saying: 're'. Aug 08 15:07:46 * kergoth makes a plan of attack for 1) revamp of staging, 2) revamp of service handling, 3) revamp of {pre,post}{inst,rm} handling Aug 08 15:09:56 Can anyone think of areas to cover for pre/post rm/inst other than 1) Permissions / Ownership, 2) New user(s) / group(s), 3) Service handling, 4) Alternatives, 5) Actual shell snippets to perform things like config file format changes, database schema upgrades, etc Aug 08 15:12:13 RP: any particular reason why bb.cache needs the cooker rather than just passing in the configuration data? that seems to be all it uses Aug 08 15:22:04 hrmph Aug 08 15:27:49 kergoth: morning Aug 08 15:28:10 kergoth: I have my own OE experience and use cases... maybe I can catch you tomorrow morning :} Aug 08 15:28:16 sure Aug 08 15:31:45 kergoth: I really saw dozens of people cross compiling gstreamer and deps by hand :( Aug 08 15:31:49 yikes Aug 08 15:37:26 hrmph Aug 08 15:38:22 taking an element from DEPENDS/RDEPENDS_* and mapping that to the actual recipe bitbake will build / has built to satisfy that dep, programmatically, from the metadata, is not easy Aug 08 15:38:51 kergoth: yeah, external toolchain + packaging everything into a tar or "rootfs stage area" would be better Aug 08 15:39:30 kergoth: and in general they go from "I need mp3 playback" and then they need to map that to gst-plugin-mad without knowing what gstreamer is :} Aug 08 15:40:00 what I'm toying with right now for staging is to capture the output of do_install (or packaging), store that by pn/pv/pr/pe quadruplet, then in a task that runs before configure, map our deps to those quadruplets and grab up the files this recipe will need, populating the private staging area as appropriate, and cleaning it up when compile is completed Aug 08 15:43:51 kergoth: do install? so like /usr/include /usr/lib/pkgconfig? Aug 08 15:44:12 kergoth: and untar with no overwrite? how will this be different from the -dev package of PN-PV-PR? Aug 08 15:44:31 -dev would work too, as i said, either do_install or packaging Aug 08 15:44:38 hmm Aug 08 15:45:05 -dev forces us to fix the -dev packages :) Aug 08 15:45:10 anyway, sleep first... Aug 08 15:46:01 night Aug 08 15:48:46 thats not the hard part, the hard part is figuring out how to map the items in RDEPENDS/DEPENDS into the pn-pv-pr-pe quadruplet :) Aug 08 15:51:04 * rkirti tries to recall what pe is Aug 08 15:51:14 epoch Aug 08 15:51:41 debian based distros use it in their binary packaging too, epoch is generally used to deal with versioning issues Aug 08 15:51:47 i.e. you accidentally made the version go backwards Aug 08 15:51:55 ok Aug 08 15:52:00 epoch is before the version, so if its bumped, its definately forward Aug 08 15:52:12 1:1.0 > 1.0rc1, but 1.0rc1 > 1.0 Aug 08 15:53:39 err...what does 1:1.0 mean ? Aug 08 15:54:00 thats how its used in debian. 1 is the epoch, 1.0 is the version Aug 08 15:54:41 good morning, fellow oe developers/users Aug 08 15:54:48 lets say you had a binary package with version 1.0rc1 instead of the correct 1.0+1.0rc1. if 1.0 then released, you'd be in trouble, because if you release a package with 1.0, it will be seen as older than 1.0rc1 Aug 08 15:55:20 so you bump epoch. if the epoch is newer, its newer, regardless of whether the version is Aug 08 15:57:14 03Klaus Kurzmann  07shr/import * rf539919b9f 10openembedded.git/recipes/images/shr-image.inc: Aug 08 15:57:14 shr-image: readd task-fso-compliance and bump PR Aug 08 15:57:14 Signed-off-by: Klaus Kurzmann Aug 08 15:58:08 has anyone had experience with the at91sam9263ek machine and OE? Aug 08 16:01:40 damnit damnit Aug 08 16:03:58 kergoth: that bad? Aug 08 16:15:41 aha, thats whats happening.. hmm Aug 08 16:17:33 can someone remind me what the MULTIMACH bits are for? Aug 08 16:18:19 kg4ysn: yes, lots of people use OE for that board. hrw|gone has been working on it recently. There used to be good support at linux4sam but its obsolete now. Maybe at91 forums can give you latets info Aug 08 16:18:25 err..latest Aug 08 16:20:03 rkirti: thanks for the info. i've scoured the linux4sam and at91 forums for information, but have run into a dead end. everything compiles fine. running is a different story. the at91bootstrap program prints its debug messages to the serial console, but then nothing else happens ... U-boot just seems to hang with no serial output. i was wondering if this was a known issue on the latest OE tree. Aug 08 16:20:14 rkirti: (or if i'm doing something wrong!) Aug 08 16:21:34 kg4ysn: ouch. I am sorry, I cant help here. I stopped working with that board some months back, so I have no clue whats up with the latest bits. Aug 08 16:21:54 kg4ysn: as I said, hrw|gone can give you a better idea when he is back Aug 08 16:22:04 rkirti: no worries. any tips/tricks from your time working with it that you feel might be helpful? Aug 08 16:22:48 rkirti: advice like "turn around three times and clap your hands to make the bootstrap work" is fine, too. ;-) Aug 08 16:26:11 hello! wich variable can I define to add a specific define (-D) during make? Aug 08 16:26:20 I'm using qt4e :-/ Aug 08 16:26:40 if it's not specific, tell me nontheless, it could be helpfull Aug 08 16:28:59 rkirti: just sent hrw a message. thanks for the advice! Aug 08 16:29:54 mm, maybe some exports in do_config_prepend() Aug 08 16:31:19 hmm Aug 08 16:33:18 kg4ysn-away-food: sorry for the lack of response. I was trying to check out things if I could make sense. Will ping you incase I do dig up anything Aug 08 16:47:36 EXTRA_QMAKEVARS_PRE was what I was looking for :) Aug 08 16:48:08 rkirti: sorry for disappearing. food was prepared, and food trumps computers in maslow's hierarchy. :-) i appreciate the help! Aug 08 16:51:53 * kergoth is tantalizingly close to populating private staging areas Aug 08 16:54:01 hmmm Aug 08 16:55:05 re Aug 08 16:55:08 hi kergoth Aug 08 16:55:57 hmm, itd be really, really easy to implement things to automatically rebuild dependent recipes when library ABIs change doing it this way Aug 08 16:56:04 * kergoth thinks Aug 08 17:03:51 * rkirti wonders if it is acceptable practice to place the recipes for packages A,B,C under dir. "foobar" if the name foobar is a generic description of the group of packages A,B,C belong to.. Aug 08 17:04:26 or do I need to have separate directories to place the recipes for each of A,B,C (ugghh) Aug 08 17:05:04 rkirti, that's commonly done in fact. Aug 08 17:05:24 look at the recipes for the X stuff, for example Aug 08 17:06:37 It also makes sense to take special care to see if the packages A,B,C share certain parts of their recipes in common, and break that out to an include file for ease of maintenance. Aug 08 17:07:13 mwester: hmm...thanks. being reminded of the recipes under python,I realise my query was stupid. Aug 08 17:08:08 http://bit.ly/OTvBa Aug 08 17:11:40 kergoth: wouldn't that be just one more (but large) extension to the autobuilders we have (for eg. narcissus/amethyst - I forget :-( ) ..the already existing builder allows picking packages, so we need one more dialog to pick files from packages Aug 08 17:12:51 (I definitely do not think its as straightforward to implement as I seem to be conveying; I am just checking if I got the idea right) Aug 08 17:13:00 Those are even more limited, from what I've seen. They don't show binary packages, even, they show lists of sets of apckages Aug 08 17:13:17 plus, they're web based, no? Aug 08 17:14:33 kergoth: true. A non-web gui for OE would rock Aug 08 17:14:49 guis are in development already in the master branch, there's a server/client split and all Aug 08 17:15:23 kergoth, have you seen http://amethyst.openembedded.net/~koen/narcissus/ Aug 08 17:15:28 ah Aug 08 17:15:29 ok Aug 08 17:15:33 yes, thats what i just said Aug 08 17:15:36 :) Aug 08 17:15:55 I added it as a comment to your idea, so people don't gert confused Aug 08 17:16:07 similar concept, but i want something a lot more powerful than that, and i want a native interface Aug 08 17:16:17 the web based one is very good for new people Aug 08 17:16:23 they do not have to set up anthing Aug 08 17:16:28 and i want the image recipe stuff to just go away, in favor of template configs for the builder Aug 08 17:16:34 but what you are talking about is also interesting Aug 08 17:16:36 oh yes, i dont want that to go away, that sort of thing is great Aug 08 17:16:49 they're just two separate things. one abstracts away the whole build process, the other doesn't Aug 08 17:16:55 right Aug 08 17:17:14 but that could leverage this, if we design it right Aug 08 17:19:42 why do i *always* misspell definite as definate? Aug 08 17:19:45 every fucking time Aug 08 17:20:23 heh Aug 08 17:20:35 the user voice people have made a suggestion Aug 08 17:20:50 indeed, the site starts up with those Aug 08 17:21:06 http://openembedded.uservoice.com/pages/6963-general/suggestions/75290-put-the-top-suggestions-or-feedback-tab-widget-on-your-site Aug 08 17:21:17 I gave it three votes :) Aug 08 17:21:28 it amuses me that this prototype private staging implementation is nearly functional, and is about 110 lines of code for the whole thing Aug 08 17:21:31 hopefully it is easy to install in mediawiki Aug 08 17:21:37 yeah, that woudl be nice Aug 08 17:21:39 cool Aug 08 17:22:04 i had to do evil though. i made bitbake put its Cooker object into the datastore, so i could get at provider info ;) Aug 08 17:23:17 let me know if any of the coreish folk want admin on the uservoice site, I'm the only one at the moment Aug 08 17:25:23 probably should ask Laibsch and cbrake_away Aug 08 17:25:25 for this staging thing, its really only a few steps. first, i made a task to capture the output of do_install into a tarball, whose name includes MULTIMACH_TARGET_SYS & PF. then I'm going to map recipe filename<->archive name in an sqlite database in TMPDIR. finally, there'll be a task that runs prior to configure that gets the filenames of its deps from bitbake, then looks up the archive names in the database and deploys those into its pri Aug 08 17:25:27 okay Aug 08 17:37:52 hmm Aug 08 17:38:10 * kergoth tries to figure out how to get all this happy but avoid the do_setscene stamp madness Aug 08 18:10:40 mmm....the oe git-native program seem broken in the shr branch Aug 08 18:15:20 anyone? Aug 08 18:15:27 here what I get: Aug 08 18:16:18 http://pastebin.com/f2a3394da Aug 08 18:24:38 damn, why isnt this working.. Aug 08 18:26:05 hmmmm Aug 08 18:27:42 anyone know what the 'selfstamp' flag is for? Aug 08 18:27:51 ah, i think i see Aug 08 18:30:11 hmm Aug 08 18:39:01 argh Aug 08 18:50:52 i wonder what would happen if a task created a bitbake runqueue and started it under the other runqueue Aug 08 18:50:56 * kergoth chuckles Aug 08 18:58:19 03Khem Raj  07org.openembedded.dev * rf2ee6f9937 10openembedded.git/ (3 files in 3 dirs): Aug 08 18:58:19 gcc-4.4.1.inc: Make cross compile for SH architecture. Aug 08 18:58:19 checksums.ini: Add checksum for lilo. Aug 08 18:58:19 Signed-off-by: Khem Raj Aug 08 18:59:48 Someone playing w/ SH? Fun Aug 08 19:00:11 I think I almost have everything I need for making my dreamcast usable Aug 08 19:01:54 * kergoth wonders how much work itd be to add a flag to let a task use something other than a stamp to determine whether it needs to be run Aug 08 19:39:59 03Khem Raj  07org.openembedded.dev * r8ccfc13fef 10openembedded.git/recipes/linux/linux-titan-sh4_2.6.21.bb: Aug 08 19:39:59 linux-titan-sh4_2.6.21.bb: Include limits.h in sumversion.c Aug 08 19:39:59 Signed-off-by: Khem Raj Aug 08 19:47:04 03Klaus Kurzmann  07shr/import * rc34840863c 10openembedded.git/ (16 files in 3 dirs): Aug 08 19:47:04 shr: add a bunch of illume keyboard layouts Aug 08 19:47:04 Signed-off-by: Klaus Kurzmann Aug 08 19:50:05 * * OE Bug 5221 has been RESOLVED (FIXED) by Aug 08 19:50:07 * * bugzilla increments bug number by 5 Aug 08 19:50:09 * * http://bugs.openembedded.net/show_bug.cgi?id=5221 Aug 08 19:52:04 * * OE Bug 5201 has been RESOLVED (WONTFIX) by Aug 08 19:52:06 * * make .net hosts CNAMES for .org Aug 08 19:52:08 * * http://bugs.openembedded.net/show_bug.cgi?id=5201 Aug 08 20:00:19 03Phil Blundell  07org.openembedded.dev * r5a728f256a 10openembedded.git/recipes/eglibc/ (eglibc-package.bbclass eglibc.inc): (log message trimmed) Aug 08 20:00:19 eglibc-package.bbclass: Replace hardcoded /lib references Aug 08 20:00:19 It turns out that eglibc contains a bunch of hard-coded Aug 08 20:00:19 references to /lib. This patch replaces them with the appropriate Aug 08 20:00:20 ${base_libdir}. Also, the globs in ${libc_baselibs} were a bit too Aug 08 20:00:24 permissive and would accept, for example, "libc.a" which isn't something Aug 08 20:00:26 that we want in the main library package. - Phil Aug 08 20:05:32 wow. even core packages have such nasty hard-codings. >.< Aug 08 20:09:06 03Elena Grandi  07org.openembedded.dev * r272c915e84 10openembedded.git/ (5 files in 4 dirs): Aug 08 20:09:06 gtksourceview: renamed the 1.x tree to gtksourceview1 Aug 08 20:09:06 GtkSourceView 1.x is an older stable version that is still required Aug 08 20:09:06 by some packages and is currently used in 2 recipes. Aug 08 20:09:06 Signed-off-by: Elena Grandi Aug 08 20:09:08 Signed-off-by: Khem Raj Aug 08 20:11:11 hm.. Aug 08 20:11:28 other funny error when building hal Aug 08 20:11:32 http://tinderbox.openembedded.net/builds/22716/ Aug 08 20:11:48 03Klaus Kurzmann  07shr/import * r6088ddf9dd 10openembedded.git/recipes/callrec/callrec_svn.bb: Aug 08 20:11:48 callrec: fix postinst script and bump PV Aug 08 20:11:48 Signed-off-by: Klaus Kurzmann Aug 08 20:11:58 probe-storage.c:47:19: error: blkid.h: No such file or directory Aug 08 20:16:35 this was a tosa build Aug 08 20:18:01 03Michal Hrusecky  07org.openembedded.dev * r6e0c17c141 10openembedded.git/recipes/justreader/ (files/missing-include.patch justreader_2.0k.bb): Aug 08 20:18:01 justreader: fixed build Aug 08 20:18:01 Adding missing include file so package can be built again. Aug 08 20:18:01 Signed-off-by: Michal Hrusecky Aug 08 20:18:01 Signed-off-by: Khem Raj Aug 08 20:29:18 03Ramax Lo  07org.openembedded.dev * rb1497a8e1d 10openembedded.git/recipes/u-boot/u-boot-mkimage-openmoko-native_oe.bb: Aug 08 20:29:18 u-boot-mkimage-openmoko: Fix SRC_URI Aug 08 20:29:18 Signed-off-by: Khem Raj Aug 08 20:33:45 03Florian Boor  07org.openembedded.dev * r14d7697f93 10openembedded.git/recipes/eglibc/eglibc-package.bbclass: (log message trimmed) Aug 08 20:33:45 RFC: eglibc libc6 and tzdata Aug 08 20:33:45 Hi, Aug 08 20:33:45 the eglibc libc6 package provides tzdata. The applied patch add this knowledge Aug 08 20:33:45 to the eglibc-package class. Aug 08 20:33:49 Greetings Aug 08 20:33:51 Florian Aug 09 02:46:32 whee, i have an almost complete private staging implementation Aug 09 02:46:42 does some crazy stuff, though.. Aug 09 02:49:55 kergoth: congrats Aug 09 02:50:25 I had to make bitbake expose the instance of Cooker in the datastore Aug 09 02:50:26 hehe Aug 09 02:53:45 kergoth: okay, what of cooker do you use? Aug 09 02:54:31 i needed the populated CacheData to use to construct a TaskData, which i used to get the filenames of the recipes we should be pulling in, based on our deps/rdeps and user preferences Aug 09 02:54:42 then i stored a fn->archivepath mapping with PersistData Aug 09 02:55:10 its actually pretty clean, easier to understand than the current twisty maze Aug 09 02:56:32 zecke: http://gist.github.com/164598 Aug 09 02:57:25 it does an rm_work too, figure i can make the packaging tasks use the captured install archive Aug 09 02:57:32 so the workdir is only needed to get that created **** ENDING LOGGING AT Sun Aug 09 02:59:57 2009