**** BEGIN LOGGING AT Mon Oct 18 02:59:57 2010 Oct 18 03:00:17 says Special files such as character or block devices and pipes cannot be copied with this function Oct 18 03:01:59 well thats for copyfile Oct 18 03:02:11 I dont know if thats same for copytree as well Oct 18 03:02:24 I have to look into pythons source code how they implemented copytree Oct 18 03:03:01 presumably with the reused the copyfile func Oct 18 03:03:21 uggg... miscommunication between brain and keyboard :) Oct 18 03:05:20 yep, this python line hangs for me: shutil.copyfile('/tmp/fifo', '/tmp/fifo2') Oct 18 03:05:59 ok cool Oct 18 03:06:10 until i 'echo asdf > /tmp/fifo' Oct 18 03:06:11 what options do u need with cp to copy Oct 18 03:06:26 -R should work Oct 18 03:06:31 cp -R ? Oct 18 03:06:34 yup Oct 18 03:06:39 copyfile imitates cp -p Oct 18 03:07:08 can u try cp -R and see if it works ? Oct 18 03:09:05 khem, works fine Oct 18 03:09:16 grg: see this http://bugs.python.org/issue3002 Oct 18 03:09:19 :) Oct 18 03:09:35 that bug is biting our ass Oct 18 03:10:43 khem, nice find! Oct 18 03:11:27 grg: the fix really wont work for us Oct 18 03:11:35 because the fix it to make python barf Oct 18 03:11:51 and not fix the copying Oct 18 03:12:08 i think that fix should be applied Oct 18 03:12:22 and the fifo should be made using a postinst script Oct 18 03:12:43 and any other recipes that create fifos should be changed similarly Oct 18 03:13:17 ah git show 5fc5e69f026bccc43cb5e5a46da63e3dd148be3d Oct 18 03:13:22 this commit caused this Oct 18 03:13:27 I guess we need to revert it Oct 18 03:13:42 someone was wise enought to use cp Oct 18 03:13:45 for a reason Oct 18 03:16:07 and adding, e.g. "os.system('cp -pPR %s/.??* %s/' % (dest, dvar))" would fix the bug with the .dot files not being copied Oct 18 04:00:59 03Khem Raj  07master * r856478e665 10openembedded.git/recipes/lirc/ (3 files): Oct 18 04:01:00 lirc: Add recipes for 0.8.7 release Oct 18 04:01:00 Delete lirc-modules_0.8.6+cvs20100613.bb and lirc_0.8.6+cvs20100613.bb Oct 18 04:01:00 Signed-off-by: Khem Raj Oct 18 04:01:04 03Khem Raj  07master * r35881fcc6f 10openembedded.git/recipes/gcc/gcc-cross4.inc: Oct 18 04:01:04 gcc-cross4.inc: We dont support multilib so empty out the with-libdir-list for SH4 Oct 18 04:01:04 Signed-off-by: Khem Raj Oct 18 05:26:21 hi all, i m trying to build bitbake omap3-consol-image and getting this erro r: http://pastebin.com/qppPqZgU ple help me Oct 18 05:26:52 i m building image for gumstix overo board Oct 18 06:38:11 gm Oct 18 06:39:01 khem, still there? Oct 18 06:39:19 did the old lirc build for you? had some issues with them Oct 18 07:11:46 good morning Oct 18 07:22:19 eFfeM_work: khem reported that hanging lirc builds are because cp was replaced with shutil, isn't it the proplem you're seeing? Oct 18 07:22:58 JaMa|Zzz: yes, i only noticed that message after asking my question here Oct 18 07:23:23 I asked because I saw the cia msg for 0.8.7 Oct 18 07:23:36 I wonder if it's possible to replace cp, if it would be possible to replace sed, awk & co. with python also Oct 18 07:24:26 technically it is, but the question is how much work it is and how much python code you need to change Oct 18 07:24:43 and what it gives as things like sed are not executed as often as cp Oct 18 07:25:02 but you then don't need to spawn another process, it's more portable (the main reason for me) Oct 18 07:25:22 JaMa|Zzz: and i did not see the ml message on shlibs Oct 18 07:25:35 until after I asked Oct 18 07:26:23 ynezz: you want to port to osX? windows? any linux system will have those functions Oct 18 07:27:49 osx/bsd, whatever Oct 18 07:31:50 i have 0 knowledge on osx, bsd and other *nix systems won't have problems with awk/sed/whatever unless maybe when you use some weird extension/feature Oct 18 07:33:26 didn't tried it on freebsd yet Oct 18 07:36:43 ehlo # Oct 18 07:36:48 hi hrw Oct 18 07:37:24 ynezz: i think if you do not have a decent sed and awk lots of pacakges won't build anyway as they might use it in their build process Oct 18 07:52:54 JaMa|W: wrt having P and PN twice in FILESPATHPKG: I don;t think this is in local.conf; see here http://www.pastebin.ca/1965572 Oct 18 08:00:54 eFfeM_work: please try to grep bitbake -e output for FILESPATHPKG Oct 18 08:01:23 will try Oct 18 08:03:01 JaMa|W: FILESPATHPKG="xserver-xorg-conf-0.1-r34:xserver-xorg-conf-0.1:xserver-xorg-conf:xserver-xorg-conf-0.1:xserver-xorg-conf:files:. Oct 18 08:05:26 ah that's from BP and BPN.. I have it too, sorry for noise Oct 18 08:05:59 np Oct 18 08:06:25 should we get rid of it? Oct 18 08:06:36 but kergoth's uniq fix should remove those.. if you're using HEAD Oct 18 08:06:38 and/or change e.g. recipe_sanity to detect dup paths Oct 18 08:06:47 bb head or oe head? Oct 18 08:20:45 eFfeM_work: oe Oct 18 08:22:04 eFfeM_work: see a241106ca4a32215ecb6c079c4e6221fd5838f6f Oct 18 08:24:19 ah ok, didn't have that one yet, my last pull was from thu or fri Oct 18 08:42:54 ping ynezz Oct 18 09:21:05 anyone is using a small web browser with OE? could you please suggest any recipe name? Oct 18 09:21:47 dillo? Oct 18 09:21:50 w3m? Oct 18 09:21:53 grafical? Oct 18 09:21:57 flash? Oct 18 09:22:15 woglinde: thx, maybe dillo is the best known Oct 18 10:03:46 woglinde: pong Oct 18 10:04:22 ynezz please apply for rw Oct 18 10:06:08 sorry, for what? Oct 18 10:06:18 rw for oe-master Oct 18 10:06:31 ahh :D Oct 18 10:06:37 ok, thanks, will do Oct 18 10:11:05 mckoan: you might also want to peek at http://wiki.openmoko.org/wiki/Browser_review Oct 18 10:11:20 and you could also consider midori Oct 18 10:13:10 afaik dillo does not do javascript Oct 18 10:17:06 eFfeM_work: thx Oct 18 10:17:37 yw Oct 18 10:44:52 03Koen Kooi  07org.openembedded.dev * r91283c9724 10openembedded.git/recipes/ti/ti-staging.inc: ti-staging: expand staging hack to pkgconfig as well Oct 18 10:59:11 re Oct 18 11:26:13 re zecke Oct 18 11:41:48 huhu Oct 18 11:41:55 hi jineld Oct 18 11:42:01 hi woglinde Oct 18 11:44:39 if a package supplies a findXXX.cmake script, what is the proper way to stage it? Oct 18 12:02:25 In the Angstrom feeds there's a package called qt4-embedded-plugin-gfxdriver-gfxpvregl. Anyone got a clue what I must build to create this one myself? Oct 18 12:03:06 tasslehoff ask koen on beagle Oct 18 12:03:48 woglinde: ok Oct 18 12:12:19 03Michael Smith  07master * rcdb6be238a 10openembedded.git/classes/package.bbclass: (log message trimmed) Oct 18 12:12:19 Revert "package.bbclass: copy dotfiles in root D to PKGD" Oct 18 12:12:19 This reverts commit 5fc5e69f026bccc43cb5e5a46da63e3dd148be3d. Oct 18 12:12:19 lirc do_package was hanging because shutil.copytree() doesn't handle Oct 18 12:12:19 named pipes. Oct 18 12:12:20 Thread: Oct 18 12:12:20 http://news.gmane.org/find-root.php?message_id=%3cAANLkTikYuii4mEWxWh7HahgFVzYgaUz%2dDADyUyibqa4d%40mail.gmail.com%3e Oct 18 12:17:48 chitz: nein, habe ich nicht - wo bleibt dein MR? :) Oct 18 12:17:50 oops wrong window Oct 18 12:19:51 03Koen Kooi  07org.openembedded.dev * r244c787976 10openembedded.git/recipes/ti/ (ti-c6run.inc ti-c6run_0.94.04.05.bb): ti-c6run: add 0.94.04.05 Oct 18 12:19:52 03Koen Kooi  07org.openembedded.dev * r419fe63d99 10openembedded.git/recipes/gnome/gnome-icon-theme_2.31.0.bb: gnome-icon-theme: add 2.31.0 Oct 18 12:19:53 03Koen Kooi  07org.openembedded.dev * ree17160ca7 10openembedded.git/recipes/ti/ (4 files): recipes/ti: fix permissions **** BEGIN LOGGING AT Mon Oct 18 12:30:01 2010 Oct 18 12:49:52 libsmaperate is biting me Oct 18 12:49:59 sample Oct 18 12:50:03 why? Oct 18 12:50:13 checksum Oct 18 12:53:04 there was something on the list over the weekend Oct 18 12:54:20 yup there is patch updating checksums for new archive, should I push it or will you? Oct 18 12:55:21 if you have it ready to go, can you push it Oct 18 12:55:31 otherwise, I will Oct 18 12:55:47 better check whats diffrent Oct 18 12:55:51 it looks like what was on the list id not applyabel via git am Oct 18 12:55:55 not that they got hacked Oct 18 12:55:56 that patch talks about that Oct 18 12:56:04 and you now have a backdoor Oct 18 12:56:22 03Koen Kooi  07org.openembedded.dev * rc7e3dc8673 10openembedded.git/recipes/gtk+/gtk+_2.20.1.bb: gtk+ 2.20.1: disable introspection Oct 18 12:56:34 woglinde: diff was included in commit message and it's ok Oct 18 12:56:49 Crofton|work: patchwork picked that diff wrong.. I'll push it Oct 18 12:56:56 yeah Oct 18 12:57:04 it lok slike all we need is top change checksum Oct 18 12:57:17 yeah, they include the package diffs Oct 18 12:57:25 hm okay Oct 18 12:57:45 woglinde, there is proper docs on the list :) Oct 18 13:00:12 03Sean Cross  07master * r40e5976b29 10openembedded.git/recipes/libsamplerate/libsamplerate0_0.1.7.bb: (log message trimmed) Oct 18 13:00:12 libsamplerate: Fix checksums to match downloaded file Oct 18 13:00:12 * libsamplerate-0.1.7.tar.gz has changed upstream. The Angstrom mirror has an old version of the file. Pulling it from the source, I see that the following two changes have been made without bumping the version: Oct 18 13:00:12 * Oct 18 13:00:12 * diff -ur old-libsamplerate-0.1.7/ChangeLog libsamplerate-0.1.7/ChangeLog Oct 18 13:00:13 * --- old-libsamplerate-0.1.7/ChangeLog 2009-02-14 11:37:27.000000000 +0000 Oct 18 13:00:14 * +++ libsamplerate-0.1.7/ChangeLog 2009-02-15 00:17:35.000000000 +0000 Oct 18 13:05:40 we have master branch? Oct 18 13:05:55 yep Oct 18 13:06:41 thats probably in that ~100 emails from oe ml which waits Oct 18 13:11:28 03Martin Jansa  07master * rc1c581c503 10openembedded.git/recipes/xorg-lib/ (pixman_0.18.4.bb pixman_0.19.4.bb): Oct 18 13:11:28 pixman: drop D_P -1 for 0.18.4 and 0.19.4 versions Oct 18 13:11:28 * new cairo (which doesn't have negative D_P) needs it Oct 18 13:11:28 * most distros have pinned pixman from preferred-xorg-versions anyways.. Oct 18 13:11:28 Acked-by: Koen Kooi Oct 18 13:11:29 Signed-off-by: Martin Jansa Oct 18 13:27:35 hrw since some weeks Oct 18 13:32:31 hi Oct 18 13:32:49 I was wondering if OE supports the beagleboard-XM Oct 18 13:32:59 aditya_1010: yes, it does Oct 18 13:33:10 aditya_1010: it is same as beagleboard Oct 18 13:33:14 dsplink and stuff should be the same right? Oct 18 13:33:20 cause its still a TI dsp? Oct 18 13:35:32 actually same processor too Oct 18 13:35:34 nifty Oct 18 13:36:23 cpu is same just higher clocks Oct 18 13:43:29 03Koen Kooi  07org.openembedded.dev * r4988dd790d 10openembedded.git/recipes/gnutls/gnutls_2.10.1.bb: Oct 18 13:43:29 gnutls: don't fail on missing files we try to rm Oct 18 13:43:29 This is also needed when doing -c configure -f twice in a row Oct 18 13:44:51 hm right Oct 18 13:45:09 seems we need to touch all libtool2.4 fixes to touch Oct 18 15:41:29 03Chase Maupin  07org.openembedded.dev * r121b15fdfa 10openembedded.git/recipes/ti/ (6 files): Oct 18 15:41:29 matrix-gui: remove unneeded PV lines Oct 18 15:41:29 * Remove setting PV since the default value is correct. Oct 18 15:41:29 * This is one less thing to change when adding new versions. Oct 18 15:41:29 Signed-off-by: Chase Maupin Oct 18 15:41:29 Signed-off-by: Koen Kooi Oct 18 15:41:40 03Chase Maupin  07org.openembedded.dev * r382d644d45 10openembedded.git/recipes/ti/matrix-gui-common_1.3.bb: Oct 18 15:41:40 matrix-gui-common: use am-sysinfo Oct 18 15:41:40 * Created a new version of the matrix-gui-common recipe that Oct 18 15:41:40 RRECOMMENDS the am-sysinfo package to provide mem_util Oct 18 15:41:41 rather than distributing a pre-built version. Oct 18 15:41:41 Signed-off-by: Chase Maupin Oct 18 15:41:42 Signed-off-by: Koen Kooi Oct 18 16:45:15 gm Oct 18 18:20:25 anyone build x11-image lately? I'm getting gpe-conf do_configure failed with No package 'libxsettings-client' found (note that libxsettings-client is built) - pkgconfig issue? - http://pastebin.mozilla.org/820747 Oct 18 18:22:21 tharvey: No package 'libxsettings-client' found Oct 18 18:22:43 right, but bitbake libxsettings-client succeeds Oct 18 18:22:59 so its a pkgconfig issue? you build x11-image often right? Oct 18 18:23:22 tharvey: is right .pc file staged by libxsettings-client recipe? Oct 18 18:25:38 should see tmp/sysrootfs/.../usr/lib/pkgconfig/libxsettings-client.pc right? no its not, so failure of libxsettings-client recipe to not stage properly? Oct 18 18:26:23 tharvey: check pc file Oct 18 18:26:35 I see there was a patch 7 days ago for libxsettings-client: convert to new-style staging Oct 18 18:26:42 betting there is an issue there Oct 18 18:27:06 find tmp/sysroots -name libxsettings* turns up no *.pc file Oct 18 18:27:25 here was patch http://cgit.openembedded.net/cgit.cgi/openembedded/commit/?id=3abb865117555b5e4f2dc6e02d94c496c6a168a1 Oct 18 18:28:01 tharvey: do_install should be probably do_install_append there Oct 18 18:29:01 testing Oct 18 18:31:00 bitbake -cclean -f libxsettings-client && bitbake libxsettings-client - now find tmp/sysrootfs -name libxsettings*.pc is found - fix Oct 18 18:31:16 JaMa, thanks - seems like we need some sort of testing policy before commit Oct 18 18:31:19 ok, I'll push it Oct 18 18:31:24 appreciate it Oct 18 18:32:22 you dont need the append Oct 18 18:32:31 remove do_install completely Oct 18 18:32:56 and just restore old do_install_append Oct 18 18:34:05 khem`: ? Oct 18 18:34:26 khem`: http://paste.pocoo.org/show/277072/ is what I have Oct 18 18:34:55 JaMa: yes Oct 18 18:35:03 JaMa: I said dont have that Oct 18 18:35:24 in exisitng recipe delete do_install completely Oct 18 18:35:36 then restrore do_install_append from old commit Oct 18 18:36:13 do_install_append () { rm ${D}/${includedir}/xsettings-common.h } Oct 18 18:36:22 thats all you should have there Oct 18 18:36:33 tested that as well, it also works Oct 18 18:37:13 ah ok I see Oct 18 18:37:38 autotools install will take care of everything Oct 18 18:37:53 we just should do stuff thats out of ordinary Oct 18 18:38:10 tharvey: can you look into the buildir and see if it staged the headers Oct 18 18:38:51 you can see that in workdir of libxsettings-client Oct 18 18:39:45 tmp/work/armv7a-angstrom-linux-gnueabi/libxsettings-client-0.17-r4.1/image/usr/include/xsettings-client.h exists... no other headers Oct 18 18:40:27 or is that not what you mean for staging? Oct 18 18:40:35 s' Oct 18 18:40:55 staging is $WORKDIR/sysroot-destdir? Oct 18 18:44:46 yeah Oct 18 18:44:50 thats ok Oct 18 18:46:09 so what I suggested should be fine Oct 18 18:47:19 already checked and pushed it a minute ago.. Oct 18 18:48:33 so this is what I was talking about with regards to clean builds - according to tinderbox there have been successful builds of x11-image following that patch doing the breakage - I can only imaging they were not 'clean' builds and thus xsettings-client was already staged from before the bad patch Oct 18 18:48:54 03Martin Jansa  07master * r11690abe68 10openembedded.git/recipes/libxsettings-client/ (libxsettings-client.inc libxsettings-client_svn.bb): Oct 18 18:48:54 libxsettings-client: don't overwrite do_install from autotools, append to it Oct 18 18:48:54 * also use INC_PR in _svn.bb Oct 18 18:48:54 * remove unneeded install calls from old do_stage (it's already installed from Oct 18 18:48:54 make install) Oct 18 18:48:55 Signed-off-by: Martin Jansa Oct 18 18:49:19 understandable, just very misleading Oct 18 18:49:50 perhaps there should be a distinction between package builds in tinderbox and 'clean' builds - or perhaps thats what Testing group/tag is for Oct 18 18:52:11 tharvey: INC_PR was changed with that commit, so libxsettings-client was rebuilt ok.. but packages depending on its .pc cannot be.. Oct 18 18:52:32 tharvey: and that wouldn't be detected without you, so thanks for spotting it Oct 18 18:53:01 thx for pushing :) Oct 18 18:54:03 so I'm hoping to add a config to http://wiki.openembedded.org/index.php/Testing - should I be seeing a post to the openembedded-devel maillist when someone does the next tag? seems like the schedule for the tagging is a bit loose Oct 18 18:56:03 tharvey: schedule was updated last week Oct 18 18:59:00 i;m trying to create a recipe for xmlrpc-c but with not too much success; I get Oct 18 18:59:18 mv -f .libs/xmlrpc_base64.lo xmlrpc_base64.lo mv: cannot stat `.libs/xmlrpc_base64.lo': No such file or directory Oct 18 18:59:33 ok, I'll keep my eye out for a new tag. And its ok to add a row to the top of the table if I have results on a unique combination to add for latest testing_2010-10-08? Oct 18 18:59:47 can someone perhaps lend me a hand? recipe is at http://www.pastie.org/1230581 Oct 18 19:00:57 tharvey, does position matter? I'd say the ones at the bottom are equally important than the top ones Oct 18 19:01:19 and the schedule is thu evening EST Oct 18 19:01:40 dunno about position -thats why I was asking. I'll just follow whats there Oct 18 19:02:36 generally people append at the end Oct 18 19:03:17 would think chronological order makes sense but the table is all over the place - hmm I see posts about tags that I don't see either like testing_2010-10-14? http://cgit.openembedded.net/cgit.cgi/openembedded/ shows me latest tag is testing_2010-10-08? Oct 18 19:05:10 actually that as 9pm GMT Oct 18 19:05:38 but I'm not even seeing that tag in cgit? perhaps I have no clue what I'm doing heh Oct 18 19:05:43 the testing tag is the last accepted test; testers pull from the testing-next branch Oct 18 19:07:43 maybe we should make it the tested tag Oct 18 19:08:45 Crofton|work: I feel no; if the testing is not successful we do not tag Oct 18 19:09:04 that's why there is no tag between sep 27 and oct 8 Oct 18 19:09:55 but we test testing-next, so a successful result is tagged tested :) Oct 18 19:10:26 I think it is a subtle English present versus past tense thing :) Oct 18 19:10:41 tested suggest we have tested it Oct 18 19:10:49 testing implies we are testing it Oct 18 19:11:03 Crofton|work: ah ok, yes (as non native speaker I didn't immediately understood waht you meant; thought you wanted to tag the version that is being tested Oct 18 19:11:14 but yeah tested would be better Oct 18 19:11:42 btw wrt my recipe question before issues is fixed, apparently was a parallel make thingie Oct 18 19:11:57 although then you would need to think about tested-next :) Oct 18 19:12:02 to be consistent Oct 18 19:12:18 or just testing Oct 18 19:12:29 tharvey: you usually add to bottom of row Oct 18 19:13:20 tharvey: branch is created every thursday PST Oct 18 19:13:28 watchout email from Cliff Oct 18 19:16:24 03Frans Meulenbroeks  07org.openembedded.dev * r78f4e57aa4 10openembedded.git/recipes/mythtv/mythplugins_0.23+fixes.bb: Oct 18 19:16:24 mythplugins: added sed script to enable lighty modules Oct 18 19:16:24 and therefore add a DEPENDS on sed Oct 18 19:16:24 Signed-off-by: Frans Meulenbroeks Oct 18 19:16:35 03Frans Meulenbroeks  07org.openembedded.dev * r28ba661107 10openembedded.git/recipes/xmlrpc-c/xmlrpc-c_1.06.41.bb: Oct 18 19:16:35 xmlrpc-c_1.06.41.bb: new recipe Oct 18 19:16:35 Signed-off-by: Frans Meulenbroeks Oct 18 19:21:32 eFfeM: hmm xmlrpc why cant it be reautoconf'ed Oct 18 19:21:41 I am sure its going to break with libtool 2.4 Oct 18 19:21:50 ok, i'm about to create a package with web stuff (html and php) but as I had severe problems with apache+php this will be lighty only, should the package ow be named xyz-lighttpd ? Oct 18 19:22:04 khem what do you mean? Oct 18 19:22:24 this is build with libtool 2.4 (but wihtout sysroot enabled) Oct 18 19:22:43 khem` I'm still an autoconf n00b Oct 18 19:22:48 I mean you override autotools Oct 18 19:22:52 do_configure Oct 18 19:23:00 then ask for reviews :) Oct 18 19:23:31 hm, when I didn't do that it complained about Makefile.am not being present (if I recall correctly) Oct 18 19:23:37 had some trouble with it Oct 18 19:24:08 ok that means the package does not use autotools in its right sense Oct 18 19:24:22 but still we in OE need to take care of others lameness Oct 18 19:24:48 we need to atleast recreate libtool macros Oct 18 19:25:37 btw also peeked at e.g. vlc/vlc-davinci_0.8.6h.bb Oct 18 19:25:48 I am not saying its the only recipe Oct 18 19:25:56 btw did a quick grep: grep autotools `grep -l gnu-configize */*` |wc Oct 18 19:25:59 99 Oct 18 19:26:31 I know that we have to fix recipes to get going with libtool 2.4 Oct 18 19:27:00 but take care when we add new recipes we can not do anything about existing recipes and they need fixing Oct 18 19:27:28 khem` agree. Can you tell me how to fix ? Oct 18 19:27:45 :1 Oct 18 19:27:57 03Martin Jansa  07master * rd90969b82e 10openembedded.git/recipes/xorg-xserver/ (2 files in 2 dirs): Oct 18 19:27:57 xserver-xorg-conf: use tslib instead of evdev for om-gta02 Oct 18 19:27:57 * kernel driver included in 2.6.34 doesn't filter events and clicks are often interpreted as short drag Oct 18 19:27:57 Signed-off-by: Martin Jansa Oct 18 19:28:08 03Martin Jansa  07master * rd98685b02b 10openembedded.git/recipes/linux/ (2 files in 2 dirs): Oct 18 19:28:08 linux-openmoko-2.6.34: add tslib patch from qtmoko Oct 18 19:28:08 Signed-off-by: Martin Jansa Oct 18 19:29:03 eFfeM: http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=980a1ea8d5bbe2c0c6e4b8f79883b0f294de979b Oct 18 19:29:07 something on those lines Oct 18 19:29:46 khem` ah ok, and with or without autotools ? Oct 18 19:29:47 * khem` has spent enough time on fixing breakges lately Oct 18 19:30:35 with autotools ofcourse Oct 18 19:32:16 ok, just wanted to be sure Oct 18 19:32:32 btw is this not something to go in autotools so we only need to set a flag in the recipe ? Oct 18 19:39:19 jo Oct 18 19:41:37 khem` what to do with this: Oct 18 19:41:38 Makefile.common:LIBTOOL = $(SRCDIR)/libtool Oct 18 19:42:36 eFfeM: no it cant be because every package has its own source structure Oct 18 19:42:48 and it has to be dealt with case by case bases Oct 18 19:43:07 ideally all of them should be re-autoconfable Oct 18 19:57:50 florian: ping Oct 18 19:58:03 gm all Oct 18 19:59:35 hi likewise Oct 18 20:00:35 * florian didn't break it! ;) Oct 18 20:02:29 03Martin Jansa  07master * rd6153d998b 10openembedded.git/conf/distro/include/angstrom-jalimo.conf: angstrom-jalimo: update preferred version of icedtea6-native as 1.7.3 is no longer available in tree Oct 18 20:06:53 florian: :-) Oct 18 20:07:14 florian: I bought the TK71 but am looking for something like a kernel source tree to compile my driver against. Oct 18 20:07:32 florian: know of any? Oct 18 20:07:59 khem, any suggestion on how to fix that LIBTOOL line ? Oct 18 20:10:09 likewise: I don't think they have set up the pubic Git repo so the most up to date might be the downloads section at warpcomm.org Oct 18 20:10:15 jama thanks Oct 18 20:10:45 effem whats the libtool-exec ist called Oct 18 20:10:50 in the src-dir Oct 18 20:10:56 other running the autofoo Oct 18 20:11:12 likewise: but i set up a public git repo on a faster server last week... maybe they want to use it for this project too. i'll ask :-) Oct 18 20:12:29 woglinde: well I modeled my patch after the one khem mentioned above but it does not change this line, should I change and remove the SRCDIR part so it picks up the libtool in staging? Oct 18 20:12:55 hm where is Makefile.common comes from? Oct 18 20:12:58 it was called libtool, removed it and copied the ltmain thingie as in the patch Oct 18 20:13:06 surely not from autotools Oct 18 20:13:21 which recipe? Oct 18 20:14:04 xmlrpc-c Oct 18 20:14:18 just added & khem commented that it was not properly done Oct 18 20:14:43 it has .in files but not .am; INSTALL files just says configure;make;make install Oct 18 20:14:51 which btw also works Oct 18 20:16:27 hm let me see Oct 18 20:18:15 hm args Oct 18 20:18:30 I remeber koen changed the distro OR Oct 18 20:18:33 PR Oct 18 20:18:39 so no easy testing Oct 18 20:19:01 will try my checkout Oct 18 20:19:06 which is a bit behind Oct 18 20:19:57 trying a changed version with the patch from khem (adapted ltmain.sh dest path) Oct 18 20:20:13 maybe that's what was wrong Oct 18 20:20:32 normaly you dont need to alter ltmain.sh Oct 18 20:20:54 hm Oct 18 20:20:55 Nothing PROVIDES 'xmlrpc-c' Oct 18 20:21:21 the recipe is just pushed Oct 18 20:21:27 iehhks Oct 18 20:22:29 woglinde: http://www.pastie.org/1230822 Oct 18 20:23:07 i pushed it with a do_configure that only had gnu-configize and oe_runconf, then khem said i did it wrong now trying to gix it Oct 18 20:23:30 see discussion from an hour ago (21.18 my time) Oct 18 20:30:21 woglinde: khem, the alternative I see is not to use autotools at all and just do configure/make/make install; works too (although I had to do make DESTDIR=${D} install as oe_runmake DESTDIR=${D} install did not properly pass DESTDIR) Oct 18 20:33:27 effem no download raw or I am blind? Oct 18 20:33:45 SRC_URI = "${SOURCEFORGE_MIRROR}/${PN}/${P}.tgz" Oct 18 20:33:56 probably should move it to before do_configure Oct 18 20:34:02 shall run oe_stylize Oct 18 20:34:17 ah left Oct 18 20:35:33 effem how is the file named? Oct 18 20:36:08 xmlrpc-c_1.06.41.bb Oct 18 20:36:57 i was hoping to get wtorrent running, this is a web based frontent for rtorrent and needs xmlrpc-c Oct 18 20:37:52 effem hm why didnt you try to remove the libtool m4 macros only Oct 18 20:38:34 because I have hardly any knowledge about libtool and friends Oct 18 20:38:56 hm but you saw it on the other patches Oct 18 20:39:13 but I do have a version that just does what the INSTALL file says (configure/make/make install); thought the preferred way was to use the auto stuff Oct 18 20:39:20 checksum is wrong Oct 18 20:40:10 ??? will verify Oct 18 20:42:25 hm? Oct 18 20:42:27 Expected MD5: '02c6b89b8ff911341b6b6d4a6c621ea9' and Got: '02c6b89b8ff911341b6b6d4a6c621ea9' Oct 18 20:42:30 Expected SHA256: '"a13f5ef40ed75d89543b9cd086185121dead9fc7de850ca4d2edd9a0cec1220' and Got: 'a13f5ef40ed75d89543b9cd086185121dead9fc7de850ca4d2edd9a0cec12205' Oct 18 20:42:36 args Oct 18 20:42:57 dman cut and paste Oct 18 20:43:40 ah Oct 18 20:43:46 can happen Oct 18 20:43:57 just verified by redownloading and it was ok for me Oct 18 20:44:57 stupid idiots Oct 18 20:46:02 why the hell and what for they use autotools Oct 18 20:46:06 agree (they also have a version 1.11 which is older than this one and not in the downloads dir in sourceforge, and apparently for never versions do svn only) Oct 18 20:46:51 oy my god Oct 18 20:46:54 where is my whip Oct 18 20:47:04 these guys will need ut Oct 18 20:47:46 03Michael Smith  07master * r463a4f992d 10openembedded.git/classes/package.bbclass: (log message trimmed) Oct 18 20:47:47 package.bbclass: copy dotfiles in root D to PKGD Oct 18 20:47:47 The previous system("cp %s/* ...") missed dotfiles/dirs at the Oct 18 20:47:47 top-level. Oct 18 20:47:47 Signed-off-by: Michael Smith Oct 18 20:47:47 Signed-off-by: Frans Meulenbroeks Oct 18 20:47:47 Acked-by: Khem Raj Oct 18 20:47:56 effem they are doing totaly utter crap Oct 18 20:48:14 hm Oct 18 20:48:31 hm ah okay Oct 18 20:48:46 # We're in a transition between the bloated, complex GNU Oct 18 20:48:47 # Autoconf/Automake style of build, in which 'configure' creates all Oct 18 20:48:47 # the make files, to simpler static make files. Some directories have Oct 18 20:48:54 *sigh* Oct 18 20:49:02 seems they got stucked Oct 18 20:49:07 so better revert to this? http://www.pastie.org/1230932 Oct 18 20:49:20 effem sorry Oct 18 20:49:32 hm Oct 18 20:49:39 mom Oct 18 20:49:53 simpler static Makefiles? Oct 18 20:50:07 which project? Oct 18 20:50:55 woglinde: this is their svn head, seems they haven't really improved on the auto stuff: https://xmlrpc-c.svn.sourceforge.net/svnroot/xmlrpc-c/advanced/ Oct 18 20:51:03 Crofton|work: xmlrpc-c Oct 18 20:51:09 working on creating that Oct 18 20:51:14 jesus Oct 18 20:51:17 they are nuts Oct 18 20:51:31 yeah, but I have a recipe which needs this package Oct 18 20:51:38 heh Oct 18 20:51:47 I hate it when that happens Oct 18 20:52:05 crofton yeah they didnt understand autotools Oct 18 20:52:12 that happens Oct 18 20:52:16 wtorrent a web based rtorrent frontent (unless someone now says he (or she!) knows a better web based torrent client Oct 18 20:52:21 but now its really worse Oct 18 20:52:24 in this state Oct 18 20:52:37 how can some release such shit Oct 18 20:52:49 instead of waiting the transition Oct 18 20:53:27 effem hm okay Oct 18 20:53:55 at least use cmake Oct 18 20:54:11 if you can't handle autfoo Oct 18 20:54:32 effem we cannt use autoreconf here Oct 18 20:54:46 that would worsen the stuff more Oct 18 20:55:13 try to do what libtoolize -copy -force bla is doing Oct 18 20:55:19 ok, (btw project does not seem to be too active any more; see their dev ml http://sourceforge.net/mailarchive/forum.php?forum_name=xmlrpc-c-devel) Oct 18 20:55:27 and run oe_runconf only Oct 18 20:55:49 ah damn shit Oct 18 20:56:00 the gnuMakfile is calling configure Oct 18 20:56:07 thats need to be patched Oct 18 20:56:37 *sigh* Oct 18 20:57:14 to be honest I dont like to work on it Oct 18 20:57:27 i can understand Oct 18 21:01:12 effem how familiar are you with autotools? Oct 18 21:03:10 effem I would restore whole autotools system Oct 18 21:04:57 effem do you just need the lib? Oct 18 21:05:01 for linking Oct 18 21:05:01 yes Oct 18 21:05:10 shouldnt be hard to restore this Oct 18 21:05:29 at least I think so Oct 18 21:06:22 what does it mean when you type bitbake foo and it returns to the prompt .... Oct 18 21:06:36 segfault? Oct 18 21:06:57 no sign Oct 18 21:07:04 dont know Oct 18 21:07:09 echo $? - what's the return code? Oct 18 21:07:11 try increase debug stuff Oct 18 21:07:34 return code is 1 Oct 18 21:07:40 I got some segfaults from python yesterday Oct 18 21:08:05 hrm, rc 1 usually means an app exited rather than a segfault Oct 18 21:08:12 woglinde: don't think I have the capabilities to do that, my auto tools is limited to 0 Oct 18 21:08:13 unless some wrapper caught the fault Oct 18 21:08:19 urg Oct 18 21:08:31 I know this is some config issue .... Oct 18 21:08:50 effem Oct 18 21:08:55 hm remove autotools Oct 18 21:09:10 or no wait Oct 18 21:09:26 overwrite do_configure with juse oe_runconf Oct 18 21:09:30 that should work Oct 18 21:10:17 woglinde: so move to http://www.pastie.org/1230932 Oct 18 21:10:20 I dont see any libtool call at all Oct 18 21:10:53 as far as I can see that is also what debian does Oct 18 21:10:53 instead of configure Oct 18 21:10:57 oe_runconf Oct 18 21:11:18 the rest autotools.bbclass is doing for you Oct 18 21:11:21 ah ok, and for do_install I could not use oe_runmake, it would not pick up DESTDIR Oct 18 21:11:38 hm? Oct 18 21:11:50 you only need to override do_configure Oct 18 21:11:52 woglinde: ah ok understood Oct 18 21:11:58 rest should work Oct 18 21:12:31 actually the version I just commited has: Oct 18 21:12:32 + gnu-configize Oct 18 21:12:32 + oe_runconf Oct 18 21:12:39 but khem told me it was no good Oct 18 21:12:45 and would break with libtool 2.4 Oct 18 21:12:59 that's why I tried to change in the first place Oct 18 21:13:13 i think the gnu-configize was needed, not 100% sure though Oct 18 21:13:58 this is the verson now in oe: Oct 18 21:13:59 http://cgit.openembedded.org/cgit.cgi/openembedded/diff/recipes/xmlrpc-c/xmlrpc-c_1.06.41.bb?id=28ba6611077bb6f2f723cb0fed1903d0a6b2dc93 Oct 18 21:14:57 on which khem said: Oct 18 21:14:57 hm let me see what gnu-configize is doing Oct 18 21:14:58 (21:18:31) khem`: eFfeM: hmm xmlrpc why cant it be reautoconf'ed Oct 18 21:14:58 (21:18:40) khem`: I am sure its going to break with libtool 2.4 Oct 18 21:17:39 woglinde: i have to call it a day, have to get up early tomorrow, but please leave a msg to eFfeM_work and I'll see it tomorrow moring Oct 18 21:17:46 anyone have time to test a patch bundle? Oct 18 21:17:47 http://patchwork.openembedded.org/bundle/msmith%40cbnco.com/make-native/ Oct 18 21:17:52 btw can't even find this gnu-configize thing Oct 18 21:18:02 just peeked at another recipe Oct 18 21:18:11 and saw it was often used like this Oct 18 21:18:22 nite all Oct 18 21:18:26 gnight Oct 18 21:18:34 woglinde: & thanks for your help Oct 18 21:35:03 kergoth_: I've not dug into this yet: http://pastebin.com/THVrhbhN Oct 18 21:44:49 is files still a good dor to use for version indeendednt config files? Oct 18 21:46:27 crofton what? Oct 18 21:46:40 all this filespath stufff Oct 18 21:46:56 the vim recipe already uses files, so I will stick with it Oct 18 21:47:07 I need to install a sane PC lke vimrc Oct 18 21:48:07 ah this again Oct 18 21:50:58 yeah Oct 18 21:51:01 I not to fix it Oct 18 21:51:04 er Oct 18 21:51:05 need Oct 18 21:54:16 any ubi expert here? :) Oct 18 21:54:24 I wonder if it's this correct http://pastebin.org/297223 Oct 18 21:57:33 one more question about http://wiki.openembedded.org/index.php/Testing - it seems that in the table I am to put the date when the testing-next branch is last merged, not the latest testing tag right? this is confusing as I see 'testing_2010-10-14' which is not a tag but the date at which the last merge occured Oct 18 21:59:42 cbrake: seems the git repo lost its head how do u reach that point ref HEAD is not a symbolic ref Oct 18 22:00:15 tharvey: testing-next is synced with master that date is what you need to put Oct 18 22:00:35 last time it was synced with master is 10/14 I guess Oct 18 22:00:50 so essentially what we are testing is master from that date Oct 18 22:01:23 khem`, right, thats a date 'not' the tag which is what is confusing me - I would expect to just see 10/14 there isntead of 'testing_2010-10-14' which is what the rev would be tagged as in the future Oct 18 22:01:24 if all goes well then master as of that date will be tagged Oct 18 22:01:54 tag is created on master for successful testing Oct 18 22:03:34 but no tag exists yet of 'testing_2010-10-14' right? last tag was testing_2010-10-08 - sorry if I'm missing somethign I think the convention used for the 'last successful build' column is very confusing Oct 18 22:03:39 well eventually that will be the tag Oct 18 22:03:55 yes, ok I understand correctly then Oct 18 22:04:01 thx Oct 18 22:06:40 and if I want to use oestats to send results to tinderbox http://wiki.openembedded.org/index.php/Tinderbox I don't need any sort of tinderbox account correct? Oct 18 22:06:44 good nite Oct 18 22:07:13 tharvey: correct, tinderbox is wide open I think ... Oct 18 22:08:30 ynezz: looks reasonable to me Oct 18 22:09:28 khem`: but I'm getting "error: (size 8158208 bytes) is not multiple of eraseblock" when ubiformat that image Oct 18 22:11:09 and 8158208 % 15872 = 0 Oct 18 22:12:48 ynezz:http://ldn.linuxfoundation.org/blog-entry/aligning-filesystems-ssd’s-erase-block-size Oct 18 22:14:20 how is this connected to my NAND? :) Oct 18 22:15:19 ynezz: this is similar may not be what you are seeing Oct 18 22:15:46 just was a food for thought Oct 18 22:15:56 ok, thanks Oct 18 22:17:55 it's all magic, need to look at sources :) Oct 18 22:18:35 ah, it's counting it with PEB=16384 Oct 18 22:19:03 8158208%16384=15360 Oct 18 22:19:27 I thought, that it needs LEB, d'oh Oct 18 22:20:17 khem`: some weeks back we were talking on the mailing list about make 3.82 vs 3.81 Oct 18 22:20:21 do you have time to take a look at http://patchwork.openembedded.org/bundle/msmith%40cbnco.com/make-native/ Oct 18 22:21:27 mzs: yes I have seen the patches Oct 18 22:22:22 this is going to punish the dev/users who use sane distros with make 3.81 natively Oct 18 22:22:33 the patches look ok Oct 18 22:22:35 as it is Oct 18 22:23:15 but I think adding more and more native packages without giving an easy way for most users is bad Oct 18 22:23:34 I guess. make-native is a pretty quick compile though Oct 18 22:23:46 and this way nobody gets hit with bugs from make 3.82 or 3.80 Oct 18 22:23:53 right now we build almost a native distro before we start building target really Oct 18 22:23:56 this has to stop Oct 18 22:24:02 Is there a better way? Oct 18 22:24:17 (i'm not sure why we build e.g. core-native but there must have been a reason) Oct 18 22:24:52 there was a talk about creating a sane packages which would identify your build machine packages and only build native packages which are not sane Oct 18 22:25:04 in your case then it would be make-native Oct 18 22:25:11 but not everyone will get it Oct 18 22:25:38 PACKAGES =+ "${PN}-vimrc" Oct 18 22:25:41 hmm. something like checking the output of --version, parsing it, and deciding whether to build -native? Oct 18 22:25:50 seems to be different from using += ... Oct 18 22:26:12 mzs: such checks in sanity.bbclass Oct 18 22:26:31 or may be a new class to determine what can be used from the build system Oct 18 22:26:32 mzs: we've talked about creating such a class, yes Oct 18 22:26:38 could automatically populate ASSUME_PROVIDED Oct 18 22:26:46 no one has taken the time to do it yet Oct 18 22:27:29 hmm. interesting idea Oct 18 22:27:43 Stuff could get broken in the future when tools change their versioning format Oct 18 22:27:44 shouldn't be terribly difficult Oct 18 22:27:50 but for the most part it'd be stable Oct 18 22:27:54 the question is how best to add elements to it via the metadata Oct 18 22:27:59 and the worst that'd happen is a -native tool would be built Oct 18 22:28:18 yep Oct 18 22:28:50 yes, how do you say a recipe needs make == 3.81 and not 3.82? Oct 18 22:28:59 or do you just say, across the whole tree, we need make == 3.81? Oct 18 22:29:05 can't, bitbake can't handle version specific DEPENDS yet Oct 18 22:29:15 not intelligently, anyway Oct 18 22:29:21 yeah, and you can't have two versions of make installed, either. :) Oct 18 22:29:37 that's true for everything, its never safe to build more than one version of anything with oe Oct 18 22:29:43 they'll stomp all over one another in sysroot Oct 18 22:30:08 I guess for some tools at least you could check that `make --version` matches PREFERRED_VERSION_make-native Oct 18 22:30:20 and if that's not set, the default pref for make-native Oct 18 22:30:43 i don't think that would be ideal Oct 18 22:30:55 well, for make okay, but most of the time its not a ==N requirement Oct 18 22:31:01 its >= N < M, or something Oct 18 22:31:10 i think ideally they'd check for actual functionality Oct 18 22:31:13 rather than going by versions Oct 18 22:31:16 but that'd be more difficult to implement Oct 18 22:31:35 sounds like you'd need a custom check for each tool. Oct 18 22:31:57 And if you're running that on every build, it might be slower in the long run than just building the -native :) Oct 18 22:32:06 and what sucks is, key=value pairs in metadata isn't particularly conducive to complex data Oct 18 22:32:45 also, we're trying to move away from doing many things in anonymous python functions, for performance and stability reasons Oct 18 22:32:59 kergoth_: we could also build tribal knowledge of what works on which distro Oct 18 22:33:25 instead of generalization Oct 18 22:33:42 sometimes same version shipped by two distros may not work sameway Oct 18 22:33:59 eek Oct 18 22:34:05 mzs: you wouldn't need to run it on every build, though. Oct 18 22:34:14 just use persist_data to cache it in a persistent way Oct 18 22:34:17 khem`: that could cover most common cases though, so it might be a way to start Oct 18 22:34:31 or whatever else, really Oct 18 22:34:31 kergoth_: true enough, and we're probably talking aobut some milliseconds here anyway Oct 18 22:34:50 it adds up though, ignoring it is how it got as slow as it is :) Oct 18 22:35:06 yup. multiply by 8000 and anything adds up Oct 18 22:35:42 yeah. course in this case, even if you use python to do it, we'd want to use an event Oct 18 22:35:47 rather than anonymous python Oct 18 22:36:06 its not obvious, but python () {} is equivalent to code run at RecipeParsed event firing, that is, at the end of the parse of every recipe Oct 18 22:36:17 in this case, we'd want to do it at ConfigParsed or BuildStarted time instead Oct 18 22:36:56 i see, so it'd only run once per build instead of once per recipe per build Oct 18 22:38:14 for now I guess I'll amend my local build to make something like glibc-initial depend on make-native.. that'll do the trick for our sles 10 build box Oct 18 22:39:11 * kergoth_ has been using ~/.oe/conf/site.conf to include conf/site/${HOSTNAME}.conf, ading ~/.oe to BBPATH, and setting ASSUME_PROVIDED as appropriate for the build environment in the per host site file, to try to speed up the builds / add extra base deps Oct 18 23:53:31 03Paul Eggleton  07org.openembedded.dev * rf524002246 10openembedded.git/recipes/opie-taskbar/opie-taskbar/server-pro-cvs.patch: opie-taskbar: fix server-pro-cvs.patch to apply against current git Oct 18 23:53:38 03Paul Eggleton  07org.openembedded.dev * re133b6e71f 10openembedded.git/ (194 files in 179 dirs): Oct 18 23:53:38 Build development versions of Opie from git instead of CVS. Oct 18 23:53:38 Note that this change requires a version of BitBake with the "git subpath" Oct 18 23:53:38 patch applied. Oct 18 23:53:38 Signed-off-by: Paul Eggleton Oct 18 23:53:38 03Paul Eggleton  07org.openembedded.dev * re631e4415a 10openembedded.git/recipes/libqpe/libqpe-opie_cvs.bb: libqpe-opie: add unbreak-logging-2.patch which applies against current git version Oct 18 23:53:39 03Paul Eggleton  07org.openembedded.dev * r690f51b800 10openembedded.git/recipes/opie-networksettings/ (3 files): Oct 18 23:53:55 opie-networksettings: do not apply wireless.patch for git version Oct 18 23:53:56 This patch has been applied upstream. Oct 18 23:53:56 03Paul Eggleton  07org.openembedded.dev * r91c2880a60 10openembedded.git/recipes/opie-taskbar/opie-taskbar_cvs.bb: opie-taskbar: apply force-firstuse-calibrate.patch against git version Oct 18 23:53:56 03Paul Eggleton  07org.openembedded.dev * rcd74274c7b 10openembedded.git/conf/distro/include/preferred-opie-git-versions.inc: preferred-opie-git-versions: add opie-notes and opie-taskbar-images Oct 18 23:53:56 03Paul Eggleton  07org.openembedded.dev * rcee19ddead 10openembedded.git/recipes/opie-textedit/opie-textedit_cvs.bb: opie-textedit: fix git version for move of text editor to "Applications" tab in launcher Oct 18 23:53:57 03Paul Eggleton  07org.openembedded.dev * r3fbdbdb8ae 10openembedded.git/recipes/tasks/task-opie.bb: task-opie: add opie-notes to task-opie-base-pim and move opie-textedit to task-opie-base-apps Oct 18 23:53:57 03Paul Eggleton  07org.openembedded.dev * r64468ed2e6 10openembedded.git/recipes/opie-packagemanager/opie-packagemanager_cvs.bb: opie-packagemanager: ensure libopkg is used for git version Oct 18 23:54:21 03Paul Eggleton  07org.openembedded.dev * rf4b69dccd2 10openembedded.git/conf/distro/include/preferred-opie-git-versions.inc: Bump Opie git revision Oct 18 23:54:22 03Paul Eggleton  07org.openembedded.dev * r42ad5ee920 10openembedded.git/recipes/opie-recorder/opie-recorder_cvs.bb: opie-recorder: add dependency on libopiemm2 Oct 18 23:54:22 03Paul Eggleton  07org.openembedded.dev * r14659e80f8 10openembedded.git/recipes/opie-vmemo/opie-vmemo_cvs.bb: opie-vmemo: add dependency on libopiemm2 Oct 18 23:54:22 03Paul Eggleton  07org.openembedded.dev * r85283a8fa1 10openembedded.git/recipes/opie-taskbar/opie-taskbar_cvs.bb: opie-taskbar: install /etc/opie_sysevents.conf Oct 18 23:54:22 03Paul Eggleton  07org.openembedded.dev * r81de05db94 10openembedded.git/recipes/tasks/task-opie.bb: task-opie: bump PR for previous change Oct 18 23:55:22 03Paul Eggleton  07org.openembedded.dev * red63e62360 10openembedded.git/conf/distro/include/preferred-opie-git-versions.inc: Bump Opie git revision Oct 19 00:19:11 call it a day Oct 19 00:19:13 nite Oct 19 00:33:24 Hi, I just received a Zaurus SL-5600 and am entirely new to linux on handhelds. I'm confused by the numerous projects/sites, dead links, etc. Apparently I should give Ångström images a try (from 2007?), but obviously OE is where the real action is at (bugzilla, irc, etc.) Oct 19 00:33:49 angstrom is built by OE. OE is the tool, angstrom is the distribution. Oct 19 00:33:59 Ångström image is downloading at < 4Kbps, gonna take hours. maybe I ought to dive into the OE source instead? Oct 19 00:34:00 you can use OE to build a current angstrom image for your device Oct 19 00:34:19 cool that's what i was thinking Oct 19 00:40:52 CrazyFoam: but that'll probably take more hours Oct 19 00:41:42 vorporeal, true :P Oct 19 00:42:03 03Khem Raj  07master * r1f279775ab 10openembedded.git/classes/opie.bbclass: Oct 19 00:42:03 opie.bbclass: Define OPIE_SRCREV and OPIE_GIT_PV if not already defined Oct 19 00:42:03 * Without this the distros which do not include Oct 19 00:42:03 preferred-opie-git-versions.inc end up in parsing errors Oct 19 00:42:03 Signed-off-by: Khem Raj Oct 19 00:57:34 should i try angstrom-2008.1 or angstrom-2010.x ? Oct 19 01:27:46 anyway to get 'make -j4' behavior in bitbake? Oct 19 01:32:16 Look up PARALLEL_MAKE and BB_NUMBER_THREADS in the docs Oct 19 01:32:35 mwester, thanks Oct 19 01:32:49 at the risk of sounding like an idiot Oct 19 01:32:54 PARALLEL_MAKE ftw Oct 19 01:33:12 makes build go from OMG I AM GOING TO KILL SOMEONE to ... AWESOME Oct 19 01:37:13 oh geez.. been searching for docs but it's right here in front of me in local.conf comments Oct 19 02:28:07 anyone have any ideas why libsdl-mixer is turning "-lmad" into "/usr/lib/libmad.so"? Oct 19 02:28:43 instead of leaving it as -lmad and it correctly getting linked with the sysroot files? Oct 19 02:29:19 i'm using libtool 2.4, which has fixed a bunch of those issues already, but not this one Oct 19 02:33:20 the Makefile has -lmad but the command line output (when it fails) has /usr/lib/libmad.so instead **** ENDING LOGGING AT Tue Oct 19 02:59:57 2010