**** BEGIN LOGGING AT Thu Oct 16 02:59:58 2008 Oct 16 03:38:02 03Jeremy Lainé  07org.openembedded.dev * rf6b3a3a8a0 10openembedded.git/ (13 files in 3 dirs): qt4-x11-free: upgrade to qt 4.4.3 Oct 16 03:39:35 03Jeremy Lainé  07org.openembedded.dev * rfa690d1b82 10openembedded.git/conf/checksums.ini: checksums.ini: add entry for sailcut 1.3.2 Oct 16 06:52:53 hi i want to build gcc 4.2.1 Oct 16 06:53:06 getting linker error Oct 16 06:53:30 in glibc-2.5 intermediate file Oct 16 07:15:21 03Koen Kooi  07org.openembedded.dev * rd68a47a362 10openembedded.git/conf/bitbake.conf: Oct 16 07:15:21 bitbake.conf: removing trailing space from TARGET_HASH to attempt to Oct 16 07:15:21 fix bug4733 Oct 16 07:19:18 why are source packages checked against digests with their source? I mean ftp://ftp.gentoo.mesh-solutions.com/mirrors/gentoo/distfiles/fakeroot_1.7.1.tar.gz has no entry in conf/checksums.ini has no digest and these to have ftp://ftp.pl.debian.org/pub/debian/pool/main/f/fakeroot/fakeroot_1.7.1.tar.gz, ftp://ftp.debian.org/debian/pool/main/f/fakeroot/fakeroot_1.7.1.tar.gz Oct 16 07:48:13 hi!! is the git repo off? Oct 16 07:48:42 methril: same problem here, it looks like the URI changed Oct 16 07:48:49 ups Oct 16 07:48:52 lets see Oct 16 07:50:42 the wiki was updated yesterday with a new URI Oct 16 07:51:00 git://git.openembedded.net/openembedded Oct 16 07:52:17 03Koen Kooi  07org.openembedded.dev * ra765a58dfd 10openembedded.git/packages/hal/hal_0.5.11.bb: Oct 16 07:52:17 hal 0.5.11: move to nonworking: Oct 16 07:52:17 * autoreconf fails Oct 16 07:52:17 * after fixing that it doesn't build Oct 16 07:52:27 03Koen Kooi  07org.openembedded.dev * rca73c43100 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 16 07:58:30 bonjour ! Oct 16 07:59:08 hello Oct 16 08:24:51 good morning Oct 16 08:25:09 yo florian_ :) Oct 16 08:28:28 03Koen Kooi  07org.openembedded.dev * rb2e74ef1ba 10openembedded.git/conf/checksums.ini: checksums.ini: add telepathy-glib Oct 16 08:31:54 03Koen Kooi  07org.openembedded.dev * rb47331e123 10openembedded.git/conf/checksums.ini: checksums.ini: add loudmouth Oct 16 08:31:55 03Koen Kooi  07org.openembedded.dev * r680ec1a6cd 10openembedded.git/packages/xorg-driver/xf86-video-omapfb_git.bb: xf86-video-omapfb: bump SRCREV Oct 16 08:42:42 03Koen Kooi  07org.openembedded.dev * r43efafa7e2 10openembedded.git/packages/telepathy/telepathy-glib_0.7.14.bb: telepathy-glib: add libxslt-native to DEPENDS Oct 16 08:44:37 florian: good morning Oct 16 08:55:38 03Koen Kooi  07org.openembedded.dev * r22008c0422 10openembedded.git/conf/checksums.ini: checksums.ini: add sugar stuff Oct 16 09:18:31 03Daniel Willmann  07org.openembedded.dev * r40ee53650e 10openembedded.git/conf/checksums.ini: checksums.ini: Add fakeroot Oct 16 09:31:01 * broonie pulls in overnight OE changes in only a couple of seconds \o/ Oct 16 09:32:20 yeah, indeed. Oct 16 09:32:35 * pb__ <3 zecke Oct 16 09:34:55 hi Oct 16 09:38:11 while building gcc4.2.1 i am getting Oct 16 09:38:15 configure: error: GNU Fortran is not working; the most common reason for that is that you might have linked it to shared GMP and/or MPFR libraries, and not set LD_LIBRARY _PATH accordingly. If you suspect any other reason, please report a bug in http://gcc.gnu.org/bugzilla, attaching /home/nishant/Desktop/pinky-3.1/build/tmp/work/armv5te-pok y-linux-gnueabi/gcc-cross-4.2.1-r0/gcc-4.2.1/build.i686-linux.arm-poky-linux-gnueabi/arm-poky-linux-gnuea Oct 16 09:38:16 bi/libgfortran/config.log Oct 16 09:38:21 this error Oct 16 09:38:30 can u help Oct 16 09:38:48 linux-kenel: if you are using poky, you should probably take it up with the poky people Oct 16 09:38:58 k Oct 16 09:39:22 no body was there Oct 16 09:39:23 i have a probleme installing openembedded.common.deb it tries to connect to openembedde.org with port 80, but im behind a proxy Oct 16 09:39:39 so i thought of asking u people Oct 16 09:40:00 in the end both are same ..........no Oct 16 09:40:15 now can u help me in this regard....... Oct 16 09:44:06 is there any way how i can say that installroutine to use port 8080? Oct 16 09:50:13 hi lrg Oct 16 09:50:59 ok problem solved Oct 16 09:51:16 hey pb_ Oct 16 09:51:43 03Koen Kooi  07org.openembedded.dev * r71e2574a3e 10openembedded.git/packages/sugar/ (files/acinclude.m4 sugar-toolkit_0.82.1.bb): sugar-toolkit: fix python detection Oct 16 10:02:49 new problem, i use apt-get install openembedded-common on a debian linux. after unpacking OE.mtn.bz2 i get some mtn misuse errors then /usr/src/oe seems to be used for somerthing else and the installation is aborting Oct 16 10:03:17 but that directory doesnt exist before that installation so only oe is using it Oct 16 10:04:58 the error is that OE.mtn is laid out according to an old schema Oct 16 10:05:39 Shanuson: we use git now Oct 16 10:06:55 morning everyone Oct 16 10:07:25 what are all these changes in checksum.ini ? Oct 16 10:07:58 ok,he is migrating that db now. Oct 16 10:08:03 git? Oct 16 10:08:36 git is the source control method we use now Oct 16 10:08:41 changed yesterday Oct 16 10:08:41 ok Oct 16 10:09:05 so i can install OE using git? Oct 16 10:09:11 http://wiki.openembedded.net/index.php/Getting_Started#Obtaining_OpenEmbedded_using_GIT Oct 16 10:10:09 I've edited the page so show that firewall challenged repo at repo.or.cz/r/openembedded.git , can someone add this to the topic please Oct 16 10:11:17 er, can't you add it to the topic yourself? Oct 16 10:12:03 is it not blocked? .. ok then Oct 16 10:12:28 don't think so. this channel never used to be topic locked. Oct 16 10:14:47 /title OpenEmbedded Developer Lounge | Web: http://www.openembedded.net | Bugtracker: http://bugs.openembedded.net | Repository: git.openembedded.net/openembedded or (http and git ro mirror) repo.or.cz/r/openembedded.git | This is not a distro or machine support channel | ADMIN STATUS: all services ok. | Check/fix your entry in http://wiki.openembedded.net/index.php/Talk:Monotone_to_Git_ID_mapping Oct 16 10:15:47 /topic OpenEmbedded Developer Lounge | Web: http://www.openembedded.net | Bugtracker: http://bugs.openembedded.net | Repository: git.openembedded.net/openembedded or (http and git ro mirror) repo.or.cz/r/openembedded.git | This is not a distro or machine support channel | ADMIN STATUS: all services ok. | Check/fix your entry in http://wiki.openembedded.net/index.php/Talk:Monotone_to_Git_ID_mapping Oct 16 10:17:11 03Koen Kooi  07org.openembedded.dev * raf199a3e19 10openembedded.git/packages/gnome/hippo-canvas_0.3.0.bb: hippo-canvas: add DEPENDS Oct 16 10:17:12 03Koen Kooi  07org.openembedded.dev * r6ecf7530ed 10openembedded.git/packages/sugar/sugar-base_0.82.0.bb: sugar-base: fix python header detection Oct 16 10:17:33 ade|desk: well done :-) Oct 16 10:17:53 helps to read manual Oct 16 10:18:47 is there a debian package available which would setup the git oe stuff? Oct 16 10:19:21 what sort of setup do you mean? Oct 16 10:19:39 checking out is just a case of "apt-get install git" followed by the single command from the GettingStarted page Oct 16 10:20:11 the development environment. i've been using the mtn version which installs oe into /usr/src/oe Oct 16 10:21:14 oh, right. no, I don't think that stuff has quite caught up with the git changes yet. Oct 16 10:22:29 i don't mind using the manual method. i was just curious Oct 16 10:32:39 apt-get install git-core might be better than git Oct 16 10:36:10 03Koen Kooi  07org.openembedded.dev * rd140603497 10openembedded.git/contrib/angstrom/build-feeds.sh: angstrom feedbuiler: run entries through 'sort | uniq' Oct 16 10:36:15 why the swap to git from mtn? Oct 16 10:36:27 or is it in the wikis and I just missed it? Oct 16 10:41:03 BlindMan: essentially because git sucks less badly than mtn does Oct 16 10:42:31 hi! do anyone know about how to make mkcramfs ignore directory contents? Oct 16 10:48:39 I see mksquashfs does have an option to exclude files, will it be any problemes switching from cramfs to squashfs? Oct 16 10:51:09 zecke: my hero! Oct 16 10:51:10 * pb__ swoons Oct 16 10:51:26 pb__: what did I break? Oct 16 10:53:08 anyone got any opinion regarding cramfs or squashfs? Oct 16 10:53:11 zecke: nothing, I am just overcome with admiration for the git repository. Oct 16 10:53:35 zecke: this morning I updated to get the new changes, and it ran so quickly that I thought my computer was broken. Oct 16 10:53:37 pb__: My hash-style build finished... and busybox and Qtopia do not honor our LDFLAFGS Oct 16 10:53:41 lol Oct 16 10:53:43 zecke: doh! Oct 16 10:54:14 zecke: I am surprised they build at all in that case. don't our LDFLAGS contain important -L flags? Oct 16 10:54:31 pb__: hehe, yes, just rpath-link... :) Oct 16 10:57:12 mm, I guess busybox and qtopia are maybe self-contained enough that they don't care. Oct 16 10:57:19 oh well, it is good that you uncovered this bug :-) Oct 16 10:57:33 pb__: yes, I will do something about that now... Oct 16 10:57:50 pb__: but speciall with Qt/Qtopia gnu hash style should make a big difference Oct 16 10:57:59 right, I hope so Oct 16 10:58:31 I plan to try the same thing with our helix dna client, which currently takes about two seconds to dlopen() its various shared objects. Oct 16 10:59:06 pb__: it is computer science, it is all about believing :) Oct 16 10:59:19 right :-) Oct 16 10:59:19 NP = P? I hope so :) Oct 16 11:00:06 tcooksey: hello :) Oct 16 11:00:33 Hi zecke Oct 16 11:04:58 pb__: you still need to send a publich ssh key to get push access Oct 16 11:06:05 ah yes, I should do that Oct 16 11:06:19 can I have more than one key? I have different ssh keys for my machine at the office and my laptop at home. Oct 16 11:06:32 I guess I could make a special one for oe but that would be a bit of a nuisance. Oct 16 11:06:37 pb__: as many as fit on our finite turing tape Oct 16 11:06:48 okay, excellent Oct 16 11:07:17 should I send it to your good self? Oct 16 11:07:23 everyone is so young in university :} Oct 16 11:07:29 pb__: mickeyl, koen, rp, me Oct 16 11:07:35 okay Oct 16 11:07:40 zecke: indeed, you kids Oct 16 11:08:08 pb__: haha, no seriously... I'm _old_ _bored_ _gray_ and almost bold Oct 16 11:08:30 heh Oct 16 11:08:45 yah, my hair is getting pretty grey as well. soon I might have to make like kergoth and cut it all off. Oct 16 11:09:01 pb__: just because it's grey? Oct 16 11:09:25 you could make it blue, makes you look young :) Oct 16 11:09:32 pb__: and at least, grey is a sign of wisdom (or so to say), seen the world, traveled countries Oct 16 11:11:07 ah yes, good point Oct 16 11:11:23 hmm Daniel Willmann added missing fakeroot entry to checksums, but still why is checksum checked agains filename AND url instead only filename? all archives are put in one directory in the end, aren't they? Also same archive with different checksum from different mirrors is really strange. Oct 16 11:11:28 not just a sign of decrepitude and senility Oct 16 11:11:41 JaMa: that does seem like a misfeature Oct 16 11:11:55 JaMa: Jeah, thought about that as well when I added it... Oct 16 11:12:02 pb__: you got it! Oct 16 11:12:46 why cant we build dreambox images with bitbake 1.8 Oct 16 11:12:49 ? Oct 16 11:13:06 prokami: dunno, you tell us. who says you can't? Oct 16 11:13:10 prokami: it requires a merge with org.openembedded.dev? Oct 16 11:13:24 how can we merge? Oct 16 11:13:32 is there any documentation? Oct 16 11:14:05 prokami: ask tmbinc for plans, I have little knowledge about the dreambox branch Oct 16 11:16:37 zecke: I see that 4ng5tr0m.inc selects -Os rather than -O2 for FULL_OPTIMIZATION. Do you know what the reason for that was? Oct 16 11:22:35 using git clone git://git.openembedded.net/openembedded if get the error: git.openembedded.net[0: 80.81.242.146]: errno=Connection timed out Oct 16 11:22:47 i am behind a proxy Oct 16 11:24:15 Shanuson, does your proxy allow ssh connections? Oct 16 11:24:33 i think so Oct 16 11:25:48 why? Oct 16 11:26:17 I think git uses ssh Oct 16 11:26:41 hm Oct 16 11:28:46 Crofton|work: Only for your push access Oct 16 11:28:57 Crofton|work: For anonymous it uses the git:// proto Oct 16 11:29:03 ah Oct 16 11:29:06 hm Oct 16 11:29:07 Perhaps even http if enabled Oct 16 11:29:09 thanks for clearing that up Oct 16 11:29:17 Crofton|work: no worries Oct 16 11:29:28 so how does that help me? Oct 16 11:29:34 Shanuson: can you try to exchange git:// with http://? Oct 16 11:29:42 sure Oct 16 11:29:44 Shanuson: see topic Oct 16 11:29:45 tring Oct 16 11:30:21 ade|desk: Ah, a http mirror. Good. :) Oct 16 11:30:46 Shanuson: The topic says it all. There is a mirror with http enabled at epo.or.cz/r/openembedded.git Oct 16 11:31:09 Shanuson: That should work even with a proxy. Way slower then git:// of course, but working Oct 16 11:31:41 thanks Oct 16 11:32:30 03Koen Kooi  07org.openembedded.dev * r9860d3a402 10openembedded.git/packages/gtk+/gtk+-1.2_1.2.10.bb: gtk+ 1.2: use sed to get m4 macros to work with gettext > 0.14.0 Oct 16 11:32:32 maybe you could at proxy to that firewall line Oct 16 11:34:13 anyway, it works, thx Oct 16 11:36:05 anyone knows if there will be some bigger problems switching from cramfs to squashfs? which does OE use? Oct 16 11:40:44 03Koen Kooi  07org.openembedded.dev * r9d97ef8717 10openembedded.git/conf/distro/include/angstrom-glibc.inc: angstrom glibc: -Os seems to trip up almost all non vanilla gcc toolchains (e.g. the harry potter ones from CSL), so take a small size penalty to save us some segfaults Oct 16 11:44:26 re Oct 16 11:44:49 speakman: I use squashfs without any problems Oct 16 11:46:00 on what hardware? Oct 16 11:47:07 no performance drawbacks? Oct 16 11:48:08 is checksums.ini "properly" sorted atm? Oct 16 11:48:24 Crofton|work: if not run it through the sorter Oct 16 11:48:43 I need to add boost, but other stuff was moving Oct 16 11:49:03 but, I see alphaone commited some changes so I will try again :) Oct 16 11:50:13 do anyone know of a more generic channel for embedded discussions? Oct 16 11:50:17 (freenode channel) Oct 16 11:53:59 zecke: forget what I said about -Os, I see that koen has just changed it anyway Oct 16 11:54:15 pb__: I luckily did not read it :) Oct 16 11:54:31 03Philip Balister  07org.openembedded.dev * r043591bc8b 10openembedded.git/conf/checksums.ini: checksums.ini : Add checksum for boost_1_36_0 and sort. Oct 16 11:54:53 alphaone: I've modified base.bbclass to check also on filename only, but now I see why its not so easy when there are "upstream" packages with bad names like this http://www.red-bean.com/~proski/firmware/ Oct 16 11:55:07 zecke: heh Oct 16 11:56:14 alphaone: maybe there should be more variants of file checksums like pn-pv-file and pn-file asi its used for src_uri Oct 16 11:57:01 JaMa: Sounds like a good idea. You should talk to someone about that who knows OE better than I do, though Oct 16 11:57:02 zecke: ? Oct 16 11:57:16 alphaone: not now :) Oct 16 11:57:27 heh Oct 16 11:59:41 drat my x in virtualbox has no input... Oct 16 11:59:51 zecke, alphaone: I'll send patch for it in the evening (in 6 hours or so :)) Oct 16 12:02:39 * zecke needs to review that no one is introducing PR's to bb files anymore... Oct 16 12:03:22 JaMa: if there are multiple recipes whose source files have leafname collisions then you are already going to have a problem, irrespective of checksums.ini. Oct 16 12:04:08 pb__: in this recipes whould be used the variant with pn-pv-file Oct 16 12:04:36 pb__: but still it should be the same file with same checksum not depending on mirror used.. Oct 16 12:04:50 right, but what I mean is that there is no point solving that problem for checksums.ini when the fetcher will still fail for other reasons. Oct 16 12:05:19 disambiguating leafnames across recipes is a good idea but checksums.ini is probably not the place to start Oct 16 12:05:23 you mean like putting the same filename in downloads directory Oct 16 12:05:30 right, exactly Oct 16 12:05:41 I see Oct 16 12:06:02 so [file_name] in checksums.ini should be enough Oct 16 12:06:28 yes, I think so. Oct 16 12:06:30 and it can save sometime using wrong file from downloads Oct 16 12:06:51 if you want to solve the leafname collision problem then you first need to find a way of naming the files in downloads so that they don't overlap, and then (if necessary) change checksums.ini to match. Oct 16 12:07:39 now I have something like this.. http://pastebin.com/m5f33fb27 Oct 16 12:08:00 incidentally, I kind of feel that the whole checksums.ini thing is a bit misconceived and it would be better just to have the md5sums in the recipe file along with the URIs. Oct 16 12:08:24 Yes it would be something like digest in gentoo ebuild directory Oct 16 12:08:42 pb__: with which syntax? Oct 16 12:08:50 pb__: md5sum_filename = "foo" Oct 16 12:09:03 pb__: sha256sum_filename = "foo" Oct 16 12:09:11 zecke: SRC_URI = "http://selfish.org/zecke.tar.gz;md5sum=012345" Oct 16 12:09:19 would be my preference, I think, but more or less anything would be ok Oct 16 12:09:37 pb__: bitbake can do that already, it didn't work with humans :) Oct 16 12:10:56 zecke: mm, but if you had made it mandatory (as you did with checksums.ini) then I guess the humans would have gotten used to it. Oct 16 12:13:01 I don't have a very strong feeling about it, but centralising this stuff in one big file makes it harder to cherry-pick recipes from one tree to another because they are no longer self contained. Oct 16 12:14:13 pb__: the down side of the SRC_URI is that adding md5 and sha sum to the end... Oct 16 12:14:14 also, it seems like you could have confusion if multiple checksums.ini files exist and have different entries for the same file. Oct 16 12:17:21 also also, it seems like the big checksums.ini file is a potentially fertile source of merge errors, though perhaps git can cope fine with all the likely use cases there. Oct 16 12:18:36 pb__: git-mergetool is quite good but yes, any central file is good for merge errors Oct 16 12:21:31 hey guys. I hate setting various variables in *-native recipe just because P and PN are a little different Oct 16 12:21:31 in my java-library.bbclass I added a simple mechanism that involves a BPN variable (base package name) Oct 16 12:21:51 BPN is set through a small python snipped that looks for "-native" suffix and removes it automatically Oct 16 12:22:18 then you can use BP is set as $BPN-$PV Oct 16 12:23:04 then you can use those variables in SRC_URIs/S. removing the need to set those again in the *-native variants Oct 16 12:23:48 hello Oct 16 12:23:55 someone here? :-) Oct 16 12:24:01 I want to move this mechanism into base.bbclass, since I believe it will cause no harm (has no effect if you do not use BPN/B) Oct 16 12:24:23 rschuster: sounds sane, show a patch :) Oct 16 12:24:35 zecke: good. will do so Oct 16 12:25:30 zecke: May I ask you? Which package does stage zlib.h? zlib-native? zlib-sdk? doesn't seem so to me. Thanks in advance Oct 16 12:25:51 szsoftware: zlib would be my guess Oct 16 12:26:05 zecke: no it doesn't to Oct 16 12:26:23 szsoftware: are you sure? Oct 16 12:26:28 find oe_tmp -name zlib.h ---> no result Oct 16 12:26:48 szsoftware: then it was not built? Oct 16 12:27:38 zecke: will rebuild zlib now, wait Oct 16 12:28:49 03Michael 'Mickey' Lauer  07org.openembedded.dev * r772e59441b 10openembedded.git/packages/e17/illume-theme-illume_svn.bb: add illume-theme-illume (from raster) Oct 16 12:28:49 03Michael 'Mickey' Lauer  07org.openembedded.dev * r03becd2a29 10openembedded.git/conf/distro/include/ (preferred-om-2008-versions.inc preferred-xorg-versions.inc): Oct 16 12:28:49 preferred-{om-2008|xorg}-versions.inc: sync with what's available in the tree Oct 16 12:28:52 NOTE: How about a policy that no removals of packages are allowed when anything DEPENDS on 'em? Oct 16 12:28:54 03Michael 'Mickey' Lauer  07org.openembedded.dev * r7cc414eef3 10openembedded.git/packages/tasks/ (task-x11-illume.bb task-x11.bb): add task-x11 and task-x11-illume Oct 16 12:30:58 zecke: we do not have python function definitions in bitbake.conf Oct 16 12:31:15 rschuster: right, they are not allowed in conf files Oct 16 12:31:23 zecke: hm Oct 16 12:31:49 zecke: I need a python function for removing -native (and possibly other suffixes like -cross) Oct 16 12:32:01 zecke: where im tmp path should zlib.h be? Oct 16 12:32:09 zecke: and it needs to be available when setting BPN. Oct 16 12:32:16 zecke: where should I put it then? Oct 16 12:32:18 rschuster: you can use oe_filter_out in the variable assignment Oct 16 12:32:20 a colleague finish my recipe on webif , we have a webif that run on openembedded :) Oct 16 12:32:36 zecke: thanks Oct 16 12:32:49 rschuster: check openssl recipe's for a usage Oct 16 12:35:34 03Koen Kooi  07org.openembedded.dev * rd42435a8c9 10openembedded.git/packages/xmms/xmms_1.2.10.bb: xmms: add sed magic to fix builds with gettext > 0.14.0 Oct 16 12:36:06 zecke: is stuff from base.bbclass available in bitbake.conf or does this come later? Oct 16 12:36:28 rschuster: base.bbclass is always loaded Oct 16 12:36:46 rschuster: and at the moment of the variable evaluation the function should be there Oct 16 12:36:54 zecke:ok Oct 16 12:37:37 zecke, ok it was built, have to patch the Makefile again. Thanks Oct 16 12:38:41 mickey|sofa: quite right. it is even tempting to go further and just say no removals, full stop. Oct 16 12:40:09 git-clone is so fast... :) Oct 16 12:42:36 after years of monotone , we appriaciate this more than the others =) Oct 16 12:44:56 yeah, indeed Oct 16 12:45:06 we should send a crate of champagne to zecke Oct 16 12:47:43 pb__: the only sad thing is there is a lot i like about monotone Oct 16 12:49:34 and OE without disk encryption is faster too... Oct 16 12:50:30 zecke: yeah, that is sad Oct 16 12:53:42 zecke: still, you have made a lot of people very happy Oct 16 12:54:13 pb__: but no girls... I need to change the profession... Oct 16 12:55:06 Gigilo? Oct 16 12:56:11 Crofton|work: pet boy in asia... no idea... I just need to change something... Oct 16 12:59:20 zecke: right, that is a drawback Oct 16 13:03:32 pb__: do you think disabling hwcaps is a sane default for OE based distributions? Oct 16 13:03:57 pb__: in my eight years with hh.org and ltg I have never seen any usage of this feature Oct 16 13:08:11 please comment: http://bugs.openembedded.net/show_bug.cgi?id=4735 Oct 16 13:09:07 zecke: I think familiar did use it at one point, and debian has as well. But yes, as I said in my mailing list posting yesterday, I think the majority of oe distros probably do want hwcaps disabled altogether. Oct 16 13:10:30 if you are building a DISTRO which basically just targets a single hardware platform, hwcaps are just a waste of time. Oct 16 13:11:28 pb__: right but even with OM having armv4 and armv6 we will have two different glibc packages anyway... and will let opkg sort it out Oct 16 13:11:30 and, nowadays, all common platforms have most of the important hwcaps anyway. anything armv4 or newer will have halfword instructions and fast multiply, which is all we used to worry about for debian. Oct 16 13:11:56 zecke: right, in that case you are basically building two different DISTROs (i.e. rootfs images), one for each target platform. Oct 16 13:12:17 the only case where hwcaps would be important would be if you wanted to have a single flash load that could be runtime optimised for the hardware in question. Oct 16 13:12:23 for most embedded people, flash is still too precious for that. Oct 16 13:12:42 the idea of having two copies of glibc there just to save a few cycles is not something they can even bear to think about. Oct 16 13:13:14 this is rather different to the debian situation in days of yore, where it was worth spending an extra 200k of disk space in order to make openssh run twice as fast on strongarm. Oct 16 13:14:16 http://lists.linuxtogo.org/pipermail/openembedded-devel/2008-October/006486.html is what I wrote on this subject yesterday Oct 16 13:16:37 hmm, I need to patch gettext now... Oct 16 13:16:48 why's that? Oct 16 13:17:16 they use open in a wrong way and gcc now knows semantic of open :) Oct 16 13:17:23 ah, heh Oct 16 13:19:59 zecke: anyway, yes, I would be happy enough to see hwcaps disabled by default in oe. even better would be to make it DISTRO-configurable but I don't think that is necessary. Oct 16 13:20:09 if anybody wants that they can implement it for themselves Oct 16 13:20:28 03Koen Kooi  07org.openembedded.dev * re4019bc9cf 10openembedded.git/ (conf/checksums.ini packages/libsdl/ballfield_1.0.bb): ballfield: add SDL demo of a field with balls Oct 16 13:20:50 hm, bitbake is tring to download files , but because of that proxy it cant Oct 16 13:21:49 hmmm... Oct 16 13:22:02 bitbake uses wget , right? Oct 16 13:22:12 can you configure wget to use the proxy? Oct 16 13:22:19 I think it ought to respect $HTTP_PROXY Oct 16 13:22:25 or $http_proxy, whichever it is Oct 16 13:22:45 the latter Oct 16 13:23:32 if its http:// it us the right port, but ftp ist the problem Oct 16 13:23:42 ftp_proxy then Oct 16 13:23:57 where do i set that variable? Oct 16 13:24:05 in the environment Oct 16 13:25:09 I'm not sure whether setting it in local.conf would also work. it might do, depends how exactly bitbake invokes the fetchers. Oct 16 13:25:25 ok Oct 16 13:25:42 well, how do i stop bitbake? Oct 16 13:26:10 its still retrying ctrl+c doent stop it Oct 16 13:29:59 Shanuson: ps, and then kill wget and python processes Oct 16 13:30:22 ok Oct 16 13:32:43 ah I need hal to get keyboard and mouse in x11... what a wonderful world Oct 16 13:38:55 okay, time for food Oct 16 13:39:22 good lunch Oct 16 13:51:53 an export ftp_proxy="bla" did it Oct 16 13:58:37 seriously, how do i keep the configure script for libsdl-image to stop looking for sdl-config? Oct 16 14:01:45 03Koen Kooi  07org.openembedded.dev * r257e641f03 10openembedded.git/conf/checksums.ini: checksums.ini: add gnuplot Oct 16 14:22:59 so i did bitbake gpe-image and installed the result on my tosa. the install seemed to go ok, but when it was done it shut off, and the power button just does nothing. Oct 16 14:23:14 all i can do now is hit ok+power to bring up the maintenance menu again (to re-install) Oct 16 14:23:28 how does the power button get handled? is that a kernel issue? Oct 16 14:41:18 03Koen Kooi  07org.openembedded.dev * ra9ec52021c 10openembedded.git/ (3 files in 2 dirs): gimp: add 2.6.0 Oct 16 14:43:13 03Graeme Gregory  07org.openembedded.dev * r57eddbf7b9 10openembedded.git/conf/distro/include/moko-autorev.inc: mokoautorev.inc : sync with the OM version Oct 16 14:43:14 03Graeme Gregory  07org.openembedded.dev * r318a879dbf 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 16 14:43:15 03Graeme Gregory  07org.openembedded.dev * r7d340cfc05 10openembedded.git/conf/distro/include/preferred-om-2008-versions.inc: Oct 16 14:43:15 preferred-om-2008-versions.inc : update the versions in line with OM Oct 16 14:43:17 branch versions Oct 16 14:43:19 03Graeme Gregory  07org.openembedded.dev * r2172627af9 10openembedded.git/packages/hal/hal_0.5.11.bb: hal_0.5.11.bb : add new version from OM branch Oct 16 14:43:31 fuck Oct 16 14:47:33 pft gimp 2.6.0, 2.6.1 is current ;) Oct 16 14:48:25 I guess I get prize for first arseup in the new git :-) Oct 16 14:48:42 what did you do? Oct 16 14:48:47 wrong email Oct 16 14:48:53 ah no Oct 16 14:49:01 I wonder why you merge hal 0.5.11 again? Oct 16 14:49:24 git config --global user.email you@preferred.address :) Oct 16 14:49:43 broonie: unfortuneately different address for different repos :-( Oct 16 14:49:53 XorA: Yeah, me too. Oct 16 14:49:58 XorA: ... I'm not amused, how hard is reading git-log? Oct 16 14:50:18 http://git.openembedded.net/?p=openembedded.git;a=commit;h=a765a58dfd410eef8147e2e6ad1d0e4cef922c58 Oct 16 14:50:20 XorA: I set the global to the one I use most so at least if I fsck up it'll be a valid address. Oct 16 14:50:34 zecke: it builds Oct 16 14:50:51 XorA: a) you don't credit me Oct 16 14:51:00 XorA: b) you didn't read previous commits Oct 16 14:52:01 XorA: at least remove hal from nonworking... Oct 16 14:52:20 zecke: sorry we cant all be as fucking god like perfect as you Oct 16 14:53:08 hmm Oct 16 14:54:09 heh Oct 16 14:54:25 oh well, I imagine koen will revert that checkin within about the next four minutes in any case. Oct 16 14:54:44 a matter of communication... XorA and me Oct 16 14:55:01 I hate to lose history, now we can git-cherry-pick things from Om... Oct 16 14:59:46 right Oct 16 14:59:53 I wonder why xora is so touchy. Oct 16 15:03:02 pb__: obviously there is a reason Oct 16 15:03:43 its a thursday... its always on a thrsday Oct 16 15:05:06 that must be it Oct 16 15:12:30 git trouble ... Oct 16 15:12:56 I have a little patch which I committed locally already, but when I want to push it I get: Oct 16 15:12:56 error: failed to push some refs to 'git@git.openembedded.net:openembedded' Oct 16 15:13:17 To git@git.openembedded.net:openembedded Oct 16 15:13:17 ! [rejected] org.openembedded.dev -> org.openembedded.dev (non-fast forward) Oct 16 15:13:25 rschuster: git-pull --rebase Oct 16 15:13:31 so I need to 'git pull --rebase' first, right? Oct 16 15:13:33 zecke: ok :) Oct 16 15:13:54 when I do that, however I get no new revisions Oct 16 15:14:15 rschuster: with mtn you would cresate a new head Oct 16 15:14:44 the last one before my local patch is: Oct 16 15:14:45 commit 3fa03adfe19f7fc926671a6353f70e86661366d1 Oct 16 15:14:45 Author: Holger Hans Peter Freyther Oct 16 15:14:45 Date: Sat Sep 20 14:21:25 2008 +0200 Oct 16 15:15:03 rschuster: try harder Oct 16 15:31:08 03Graeme Gregory  07org.openembedded.dev * r368320c838 10openembedded.git/conf/distro/include/preferred-om-2008-versions.inc: preferred-om-2008-versions.inc : move hal back to latest buildable one Oct 16 15:31:08 03Graeme Gregory  07org.openembedded.dev * rb411bda61f 10openembedded.git/packages/hal/hal_0.5.11.bb: Oct 16 15:31:08 Revert "hal_0.5.11.bb : add new version from OM branch" Oct 16 15:31:08 This was already cherry picked then moved to non-working Oct 16 15:31:11 This reverts commit 2172627af94c9b3bfd6278e64b90b0ab664512f6. Oct 16 15:49:31 g'day kergoth Oct 16 15:50:33 hey Oct 16 16:06:35 doh, fso-image build fails due to checksums.ini Oct 16 16:06:37 * pb__ stabs it Oct 16 16:07:10 actually, though... Oct 16 16:07:12 ERROR: fakeroot-native-1.7.1: ftp://ftp.gentoo.mesh-solutions.com/mirrors/gentoo/distfiles/fakeroot_1.7.1.tar.gz has no entry in conf/checksums.ini, not checking URI Oct 16 16:07:37 ... if ever there was a site for which checking the md5sum was justified, this one must surely be it. Oct 16 16:09:38 pb_: When did you last update your oe repository? Oct 16 16:10:04 dunno, a few hours ago I guess Oct 16 16:10:09 pb_: Because I recently added that Oct 16 16:10:22 ah, let me try again Oct 16 16:43:50 | /bin/sh: line 3: /home/balister/oe/tmp/staging/x86_64-linux/usr/bin/install: No such file or directory Oct 16 16:43:58 from gcc-cross-intermediate .... Oct 16 17:19:33 hi Oct 16 17:19:59 hi woglinde Oct 16 17:20:29 pb uclibc has HASH_GNU Oct 16 17:20:37 but we dont have it activated in our confs Oct 16 17:20:40 I will try it now Oct 16 17:30:37 hi jay7 Oct 16 17:30:47 woglinde: hello :) Oct 16 17:32:24 * Jay7 is studying git :) Oct 16 17:33:34 with mnt i have two branches in two directories (.dev and .stable) Oct 16 17:33:45 now I have only one working directory.. Oct 16 17:34:09 can I have similar schema something? Oct 16 17:34:24 *mtn Oct 16 17:40:45 ???? Oct 16 17:40:54 you can either clone locally Oct 16 17:41:01 or branch localy Oct 16 17:41:04 as you want Oct 16 17:41:11 but it is always one dir Oct 16 17:41:22 hm okay with clone Oct 16 17:41:26 you have 2 dirs Oct 16 17:41:49 hm.. so - one .git dir -> one current workdir? Oct 16 17:42:35 I'm just a little confused :) Oct 16 17:43:18 ???? Oct 16 17:43:28 you switch btw. the branches with command Oct 16 17:43:43 git-branch :) Oct 16 17:43:46 git -b shows you which branches are avaible Oct 16 17:43:47 I know already :) Oct 16 17:43:59 did you read the git site on oe wiki? Oct 16 17:44:07 yeah.. Oct 16 17:44:22 that should be enough Oct 16 17:44:24 for startig Oct 16 18:51:39 woglinde: okay, very good Oct 16 18:52:18 if you can activate support for GNU_HASH in uclibc then you should be able to turn on --hash-style=gnu in LDFLAGS. Oct 16 18:52:50 pb_ hm unfornatly it does not really compile Oct 16 18:52:52 or you might want to use --hash-style=both for a while to ensure that you have a workable upgrade path Oct 16 18:52:56 for armv4 at least Oct 16 18:53:01 I am just searching Oct 16 18:53:08 otherwise, obviously, if you build busybox with --hash-style=gnu and it gets upgraded before uclibc, you will be hosed. Oct 16 18:53:20 yes Oct 16 18:53:22 woglinde: okay, very good Oct 16 18:54:12 actually, I guess you could just use a versioned depend on uclibc to avoid upgrade breakage Oct 16 18:54:24 grr, cat is on my lap again Oct 16 18:54:25 pb did you ever such an error? Oct 16 18:54:26 binutils-2.18.50.0.7/bfd/elf.c line 369 in bfd_elf_get_elf_syms Oct 16 18:54:47 that's quite an old binutils Oct 16 18:54:53 can you use a newer one? Oct 16 18:54:58 hu? Oct 16 18:55:05 thats the one we use in oe Oct 16 18:55:32 oh, is it? Oct 16 18:55:37 maybe it's not as old as all that then Oct 16 18:55:37 there is no newer Oct 16 18:55:39 heh Oct 16 18:55:50 in oe at least Oct 16 18:56:01 hm let me see Oct 16 18:56:07 if I find a newer Oct 16 18:56:45 huch 2.19 is out? Oct 16 18:57:11 hm no Oct 16 18:57:17 why he did then mark it Oct 16 18:57:57 oh right. I thought 2.19 had been released, but apparently not. Oct 16 18:58:10 anyway Oct 16 18:58:15 what wsa the actual error you saw? Oct 16 18:58:17 hm release.binutils-2.18.50.0.9 is the last Oct 16 18:58:24 you previously mentioned the line number but not the error message :-} Oct 16 18:58:47 ups sorry Oct 16 18:58:49 arm-angstrom-linux-uclibc-ld: BFD (Linux/GNU Binutils) 2.18.50.0.7.20080502 internal error, aborting at /devel/arm/oetmp2/work/arm-oabi-angstrom-linux-uclibc/binutils-cross-2.18.50.0.7-r4/binutils-2.18.50.0.7/bfd/elf.c line 369 in bfd_elf_get_elf_syms Oct 16 18:59:26 but the compiles go on Oct 16 18:59:37 that is odd Oct 16 18:59:42 no, I haven't seen that message Oct 16 19:04:52 doh, mythfront-image doesn't fit in its ramdisk anymore. Oct 16 19:04:54 stupid bloatware Oct 16 19:05:00 * pb_ shakes his fist Oct 16 19:06:32 hell, I seem to be getting a load of -dev libraries installed in the image Oct 16 19:06:49 no wonder it is too big Oct 16 19:08:03 Package: udev Oct 16 19:08:03 Version: 124-r8 Oct 16 19:08:03 Depends: update-rc.d, libc6 (>= 2.5), libvolume-id-dev (>= 124) Oct 16 19:08:04 ! Oct 16 19:08:40 and, of course, libvolume-id-dev depends on libc6-dev Oct 16 19:08:47 pb muhahaha Oct 16 19:08:49 which then seems to recommend itself for good measure :-} Oct 16 19:08:57 hi flo Oct 16 19:09:06 well, I always thought udev was bad news Oct 16 19:09:19 maybe I can ditch it in favour of mdev Oct 16 19:09:37 I guess I would lose hal, but I don't like hal much anyway Oct 16 19:11:47 re Oct 16 19:12:23 hm, xserver-xorg rdepends on hal. that doesn't seem right. Oct 16 19:12:28 pb could you look at http://www.busybox.net/lists/uclibc-cvs/2008-January/011661.html ? Oct 16 19:12:48 I dont not get why he removed extern long int testandset (int *spinlock); Oct 16 19:12:49 ok Oct 16 19:12:54 in the arm header Oct 16 19:13:18 pb there are serval missleading RDEPENDS Oct 16 19:13:30 woglinde: does arm actually have that function? Oct 16 19:13:35 there is no test-and-set instruction as such on arm Oct 16 19:13:47 its defined Oct 16 19:13:48 at least, not prior to armv6 Oct 16 19:13:48 later Oct 16 19:13:51 ah Oct 16 19:13:53 hm Oct 16 19:13:55 dunno then Oct 16 19:14:03 flo_lap: good evening Oct 16 19:14:06 this your code btw. Oct 16 19:14:11 heh Oct 16 19:14:13 http://sourceware.org/cgi-bin/cvsweb.cgi/ports/sysdeps/arm/linuxthreads/pt-machine.h?rev=1.7&content-type=text/x-cvsweb-markup&cvsroot=glibc Oct 16 19:14:23 yah, but my code from nearly ten years ago Oct 16 19:14:31 actually, maybe more than ten years Oct 16 19:14:36 I think I wrote that in 1996 or 1997 Oct 16 19:15:07 pb so I normal define without extern should work Oct 16 19:15:27 oh, right, it does testandset with swp Oct 16 19:15:35 yeah Oct 16 19:16:34 h this seems a silly commit then http://sourceware.org/cgi-bin/cvsweb.cgi/ports/sysdeps/arm/linuxthreads/pt-machine.h.diff?r1=1.3&r2=1.4&cvsroot=glibc&f=h Oct 16 19:16:56 03Jeremy Lainé  07org.openembedded.dev * r34a12c2ff9 10openembedded.git/ (4 files in 3 dirs): lua5.1: package 5.1 branch of the lua language Oct 16 19:17:54 pb_ hm was there a reason not to write the protofucntion? Oct 16 19:17:59 just lazyness? Oct 16 19:18:05 ah mom Oct 16 19:18:31 this it looks the first Oct 16 19:18:32 extern inline int Oct 16 19:21:47 * pb_ sends mail to the list about bogus rdepends Oct 16 19:23:11 woglinde: I don't really know what was going on with that patch. looks like andreas was just moving code around from one file to another. Oct 16 19:23:26 pb ah I got the problem Oct 16 19:23:33 in uclibc Oct 16 19:23:39 ah right Oct 16 19:23:45 http://www.busybox.net/lists/uclibc-cvs/2008-January/011661.html Oct 16 19:24:00 static inline vs. extern inline Oct 16 19:24:11 now its clear Oct 16 19:24:32 oh right, that old problem Oct 16 19:25:05 hm should be easy to fix Oct 16 19:26:00 Morning... Oct 16 19:26:20 hm Oct 16 19:26:27 *sratch my head* Oct 16 19:26:37 on which branch is this Oct 16 19:26:56 If I want to build a linux system for my target using busybox init instead of sysv-init, what do I need to do? Oct 16 19:27:29 Currently I can build a system using angstrom-2008.1 Oct 16 19:27:31 Klecks remoeve sysv-init deps from your taks and define init in busybox config Oct 16 19:27:37 task Oct 16 19:27:55 tasks are located in packages/tasks Oct 16 19:28:02 grep for sysvinit Oct 16 19:31:55 ok... do I have to delete the sysv-init deps or can I just overwirte it with something like DISTRO_INIT_MANAGER_myboard =? "...." Oct 16 19:32:16 klecks hm not sure Oct 16 19:33:00 how can I change busybox config? I don't want to change behaviour for all boards, just for mine? Oct 16 19:33:54 can I add a subdir in busybox-XX.xx/ with the name of my board and copy my defconfig into it? Oct 16 19:34:34 Klecks are you the only user of your oe dir? Oct 16 19:34:39 then it is okay Oct 16 19:34:52 otherwise you could try either a new recipe Oct 16 19:34:58 or make an overlay Oct 16 19:35:12 overlay is described in the wiki Oct 16 19:35:39 pb really strange with the uclibc Oct 16 19:35:42 * Jay7 have created two local repos to separate tracking .dev and .stable Oct 16 19:35:50 hmmm.. at the moment yes, but I would like to push my stuff back into openembedded Oct 16 19:35:56 it is fixed by commit but the commit never shows up in the trunk Oct 16 19:37:11 moin Oct 16 19:37:18 http://www.angstrom-distribution.org/building-angstrom is the checkout process supposed to work? Oct 16 19:37:24 ah no Oct 16 19:37:26 hms Oct 16 19:37:39 he removed his fix and did forget bring the extern back Oct 16 19:37:41 args Oct 16 19:38:51 I think that wiki page needs review Oct 16 19:39:53 oh "Posted March 30th, 2006 by koen" Oct 16 19:40:07 dcordes: git url is wrong Oct 16 19:40:17 all urls are wrong Oct 16 19:40:36 what is the function of a distro? When would I need to create my own, rather then using e.g. angstrom? Oct 16 19:41:04 distro sets the version of toolchain stuff Oct 16 19:41:15 e.g binutils gcc gettext libtool Oct 16 19:41:17 Klecks: a DISTRO is basically a set of policies. Oct 16 19:41:34 dcordes: it was corrected after Oct 16 19:41:48 some time ago here was mtn instructions Oct 16 19:42:20 Jay7, yea I know the old howto. do you have access to this document? Oct 16 19:42:39 dcordes: no :\ Oct 16 19:42:41 just toolchain stuff? all the other things like busybox, apps etc. are choosen by a task? Oct 16 19:42:55 Jay7, itwould be useful to correct it imho Oct 16 19:45:19 dcordes: koen was made latest corrections afaik Oct 16 19:45:22 so.. Oct 16 19:48:05 "The report makes the following recommendations: Oct 16 19:48:05 a) stimulant drinks should be labelled with an indication that they Oct 16 19:48:06 are unsuitable for children (under 16 years of age), pregnant women Oct 16 19:48:09 and individuals sensitive to caffeine Oct 16 19:48:12 b) they should be classified with other beverages of high caffeine Oct 16 19:48:16 content Oct 16 19:48:19 c) the consumption of stimulant drinks by children under 16 years Oct 16 19:48:22 should be discouraged Oct 16 19:48:22 d) caution should be exercised in the consumption of stimulant drinks Oct 16 19:48:26 with alcohol Oct 16 19:48:50 important safety information there Oct 16 19:50:18 Jay7, ok thanks. I pinged him Oct 16 19:57:47 03Jeremy Lainé  07org.openembedded.dev * rf4dfa98061 10openembedded.git/conf/checksums.ini: checksums.ini: add entry for pysqlite 2.4.0 Oct 16 20:01:32 hm Oct 16 20:01:47 git branch switching is not so easy to understand Oct 16 20:05:06 woglinde: what do you mean? Oct 16 20:06:50 03Jeremy Lainé  07org.openembedded.dev * r4c9035f449 10openembedded.git/conf/checksums.ini: checksums.ini: add hostapd python-django python-flup squidview Oct 16 20:08:55 hm I want to work in zeckes branch Oct 16 20:09:07 but git checkout branchname did not work Oct 16 20:09:49 03Jeremy Lainé  07org.openembedded.dev * rc6535102cb 10openembedded.git/conf/checksums.ini: checksums.ini: add entry for v86d Oct 16 20:11:29 hm Oct 16 20:11:34 seems to work Oct 16 20:11:59 woglinde: remote branches != local branchs Oct 16 20:14:52 cbrake yes Oct 16 20:15:03 but how than I can work with this Oct 16 20:15:21 acutally I can switch to it Oct 16 20:15:29 but than I am in no branch Oct 16 20:17:47 woglinde: git remote --help Oct 16 20:22:24 hrw but how do I know the url? Oct 16 20:28:32 git branch -r Oct 16 20:28:46 hm ah now i am seems to understand Oct 16 20:29:03 I am making a local branch of the remote branch Oct 16 20:29:04 22:28 hrw@home:openembedded$ git checkout origin/holger/hash-style Oct 16 20:29:12 this is shortcut Oct 16 20:29:33 $ git checkout -b mycopyofholgerbranch origin/holger/hash-style Oct 16 20:29:36 is better Oct 16 20:29:53 bye Oct 16 20:30:08 jepp Oct 16 20:30:10 puh Oct 16 20:44:45 hrm. git dead? Oct 16 20:44:48 fatal: The remote end hung up unexpectedly Oct 16 20:46:27 oh the remote origin url change recently? Oct 16 20:49:15 hm Oct 16 20:49:23 when did it last work for you? Oct 16 20:49:46 the testing git ro differs from what we have now Oct 16 20:49:57 Hi! Can I just download+patch the source with "bitbake" (not build). If yes, how or where are the instructions? Thanks. Oct 16 20:50:13 bitbake -c patch moo Oct 16 20:50:38 are these Tasks described somewhere? Or they are custom and defined in my OE tree somewhere? Oct 16 20:52:43 the are standard bitbake Oct 16 20:52:45 i cant remember when i last pulled down changes :) Oct 16 20:52:55 anyhoo fixed now :) Oct 16 20:53:04 timtimred we went yesterday live with git read write Oct 16 20:53:06 hooray for git over mtn btw :) Oct 16 20:53:32 read about the git stuff in the wiki Oct 16 20:53:43 this makes me more likely to request access to cut stuff over from foonas :) Oct 16 20:54:25 actually if thats possible to setup sometime that would be great ;) Oct 16 20:54:29 no hurry though... Oct 16 20:54:57 NAiL used to do it but hes less active now :) Oct 16 20:55:51 mostly just interested in adding new platforms and other packages you dont have Oct 16 20:56:30 hm let me see how do I get a file from another branch into the branch I am currently in Oct 16 20:57:23 hm seems git add -i is my friend again Oct 16 20:57:36 hm not really Oct 16 20:59:42 git-merge-one-file Oct 16 20:59:45 Hmmmmm Oct 16 20:59:54 Is anyone using SDK stuff and opkg, rather than ipkg? Oct 16 21:11:29 hmm, is e2fsprogs busted again? Oct 16 21:11:48 jeremy_laine why? Oct 16 21:12:04 | make[2]: Entering directory `/home/bot/jlaine/oe/build/tmp/work/x86_64-linux/e2fsprogs-native-1.41.2-r1.1/e2fsprogs-1.41.2/e2fsck' Oct 16 21:12:06 | MKINSTALLDIRS /home/bot/jlaine/oe/build/tmp/staging/x86_64-linux/usr/sbin /home/bot/jlaine/oe/build/tmp/staging/x86_64-linux/usr/share/man/man8 Oct 16 21:12:07 | make[2]: MKINSTALLDIRS@: Command not found Oct 16 21:12:28 hm I didnt fix 1.41.2 yet Oct 16 21:12:35 let me see in a moment Oct 16 21:13:10 you could just take the patch from 1.38 Oct 16 21:13:21 m4.patch? Oct 16 21:13:29 nope Oct 16 21:14:51 Oct 16 21:14:55 mkinstalldirs.patch Oct 16 21:16:32 why is m4.patch even present, it doesn't seem to be used? Oct 16 21:16:55 hm this seems an accident commit Oct 16 21:17:05 it puts the 1.34.bb into files Oct 16 21:17:12 I think this can be removed Oct 16 21:17:25 woglinde: the patch applies, let's see if it fixes the problem Oct 16 21:17:38 jeremy_laine there are some problems Oct 16 21:17:48 thats why I set it to -! Oct 16 21:17:51 ups -1 Oct 16 21:17:54 woglinde: if so, I'll move mkinstalldirs.patch from e2fsprogs-1.38 to e2fsprogs Oct 16 21:18:43 woglinde: but why is this not the case for e2fsprogs-native? Oct 16 21:19:02 mom Oct 16 21:19:08 woglinde: 1.41.2 is getting picked up for native Oct 16 21:19:56 hm require e2fsprogs.inc Oct 16 21:19:59 ups Oct 16 21:20:07 okay Oct 16 21:20:10 seems my fault Oct 16 21:20:22 maybe you are lucky and compiles fine for x86 Oct 16 21:20:58 well on x86-64/Debian I can tell you it fails :) Oct 16 21:21:03 heya NAiL... Oct 16 21:21:37 sorry Oct 16 21:21:51 could you do the default_prefrences line to Oct 16 21:22:30 hey timtimred =) Oct 16 21:23:55 woglinde: for now I'll put the default preferences to -1 until we get a proper fix Oct 16 21:24:04 yes Oct 16 21:24:06 oh looks like most of team foonas is in here :) Oct 16 21:24:06 sorry Oct 16 21:24:23 didnt look up that it not includes the target package Oct 16 21:25:11 how are we able to get commit access? Oct 16 21:25:35 03Jeremy Lainé  07org.openembedded.dev * redb26551b4 10openembedded.git/packages/e2fsprogs/e2fsprogs-native_1.41.2.bb: e2fsprogs-native: don't pick up version 1.41.2 for now, it has issues Oct 16 21:26:25 kergoth: ping? Oct 16 21:31:22 NAiL: pong Oct 16 21:35:10 Who "keeps the keys" for the git repo? Oct 16 21:35:34 And, who can I prod to get commit-access for timtimred? :) Oct 16 21:35:57 i haven't been active in the oe world in, literally, like 3 years Oct 16 21:36:01 :P Oct 16 21:36:30 Who is? :P Oct 16 21:41:17 mickey|sofa: ping? Oct 16 21:42:17 RP? Oct 16 21:50:07 woglinde: I have adapted a fedora patch: http://cvs.fedora.redhat.com/viewvc/rpms/e2fsprogs/F-7/e2fsprogs-1.39-mkinstalldirs.patch?revision=1.2 Oct 16 21:52:22 woglinde: and I am applying the patch both to e2fsprogs-native_1.38 and e2fsprogs_1.38 Oct 16 21:52:47 woglinde: also, I am removing the workaround for slugos as the new patch should work everywhere Oct 16 21:55:55 03Jeremy Lainé  07org.openembedded.dev * r72b7ac3a4e 10openembedded.git/packages/e2fsprogs/ (3 files in 2 dirs): Oct 16 21:55:55 e2fsprogs: better fix for mkinstalldirs issue Oct 16 21:55:55 * update mkinstalldirs.patch based on fedora patch Oct 16 21:55:55 * remove workaround for slugos, the new patch should work Oct 16 21:55:55 * apply patches to e2fsprogs-native too Oct 16 22:52:36 so what in an angstrom image would cause the power button to do nothing at all - is that the kernel, or can something else be to blame? Oct 16 22:52:57 kernel Oct 16 22:53:14 acpi code? Oct 16 22:53:22 you didnt say the machine Oct 16 22:53:24 tosa Oct 16 22:53:33 kernel Oct 16 22:53:43 tosa has no acpi Oct 16 22:55:23 how fickle is that code? can i likely rewind the kernel tree by 2 weeks and have it work? or is there a long-known badness there? Oct 16 22:55:41 sorry dont know Oct 16 22:55:47 ok thanks Oct 16 22:55:53 but there is no really work on tosa Oct 16 22:55:59 anymore Oct 16 22:56:09 that's too bad Oct 16 22:56:22 thats life Oct 16 22:57:48 recon' i'll play around with it some more, just not sure where to start when i don't even get any printks :) Oct 16 22:58:48 03Jeremy Lainé  07org.openembedded.dev * r2181f9ee2d 10openembedded.git/conf/checksums.ini: checksums.ini: add entries for gob2 grub libupnp Oct 16 23:18:39 woglinde: I am getting the first bits of openjdk compiling .. Oct 16 23:18:55 woglinde: I had to write a lesstif recipe ... Oct 16 23:35:09 * Jay7 -> sleep Oct 17 01:36:54 how can we update enigma and tuxbox from cvs to org.openembedded.dreambox branch to compile with bitbake **** ENDING LOGGING AT Fri Oct 17 02:59:57 2008