**** BEGIN LOGGING AT Thu Jul 13 02:59:56 2006 Jul 13 02:59:58 its still trying to make the qemu Jul 13 03:31:10 SonicvanaJr, i suspect your BBPATH or others are wrong Jul 13 03:31:20 ok Jul 13 03:31:28 well I just deleted the entire directory Jul 13 03:31:31 and starting from scratch Jul 13 03:31:57 BBPATH is an enviroment variable Jul 13 03:57:41 * bgat is away: sleep, hopefully. Jul 13 05:04:12 sigh Jul 13 05:04:21 why is everybody flocking to modular X.org? Jul 13 05:04:41 from actual experiences people have had, it sounds way too buggy :/ Jul 13 05:04:47 why wouldnt they? we were using it long before it ever released that way Jul 13 05:04:58 fd.o cvs was used in oe for quite some time Jul 13 05:05:42 because end result, chances are good your base install will be smaller Jul 13 05:06:44 emte: Kerwood_: speaking of which, I'm not exactly one of the "new people"-- just don't have time to keep up to date on these things ;) Jul 13 05:08:01 kergoth: yeah, just ranting mostly about Gentoo's decision Jul 13 05:08:15 I can kinda see the point with OE-- modular stuff is needed more with low disk Jul 13 05:09:43 Luke-Jr: less building when updating small parts, more small bugfix releases Jul 13 05:10:03 Luke-Jr: I think that's a bit plus. Built fine for me on Gentoo Jul 13 05:10:14 JustinP: doesn't help much when you *need* more bugfixes ;) Jul 13 05:10:30 JustinP: building isn't what I hear is a big problem-- I'm hearing lack of benefits and speed issues Jul 13 05:10:35 and stability, IIRC Jul 13 05:10:43 well...they're new ebuilds Jul 13 05:10:55 I find that Gentoo tends to stabilize things well Jul 13 05:10:58 actually, they're a few months old IIRC Jul 13 05:11:05 just recently marked as stable Jul 13 05:11:13 well, true, but just marked stable, yeah Jul 13 05:11:26 I imagine they should wait until the problems are solved before doing that Jul 13 05:11:32 not to mention providing a real upgrade path Jul 13 05:11:51 upgraded fine for me....:-P Jul 13 05:11:59 eg, not leaving people without X for hours Jul 13 05:12:06 and you can't solve all problems before something is "stable" Jul 13 05:12:30 I'd expect it to be up to par with the last release Jul 13 05:12:36 it worked for the testers.... Jul 13 05:12:43 unless there's significant improvements Jul 13 05:12:47 there are of course going to be some problems if you break it out... Jul 13 05:13:06 eh...not interested in continuing Jul 13 05:13:27 you can read the Gentoo announcement for the upsides Jul 13 05:13:30 anyway, it really shouldn't be hard for Portage to manage building all the new X.org and then merging it all before removing the old one Jul 13 05:14:06 like with almost every other upgrade Jul 13 05:14:08 * thejapa just doing revdep-rebuild Jul 13 05:14:25 thejapa: does it need that, too? Jul 13 05:14:35 yep Jul 13 05:14:51 only mplayer and xine-lib had to be rebuilt for me Jul 13 05:14:54 when I used gentoo some years ago it never existed Jul 13 05:15:03 but I don't use X too often on this machine any more (not a desktop) Jul 13 05:15:06 I'm starting to dislike gentoo Jul 13 05:15:37 thejapa: I'm of the view that it sucks the least, but still sucks. Jul 13 05:15:43 * JustinP goes to bed and lets portage finish emerging 111 new gnome packages Jul 13 05:15:44 right Jul 13 05:17:10 i had tested cvs e17 Jul 13 05:17:40 now I undone some /etc/portage/package* edits Jul 13 05:17:49 and must unmerge stuff Jul 13 05:17:51 ew Jul 13 05:18:00 ~lart gentoo Jul 13 05:18:00 * ibot resizes gentoo's terminal to 40x24 Jul 13 05:18:39 yeah, I'm planning/prototyping a modular package manager that will hopefully be able to solve Portage's problems Jul 13 05:21:39 haha revdep-rebuild will rebuild the whole machine to get xorg updated Jul 13 05:22:35 why? Jul 13 05:23:49 okay, not completely true. i was dumb and tried to use gentoo's qt to get a working qvfb and my whole qt libs were just a mess Jul 13 05:24:15 now i'm paying the price of a lot of stuff broken Jul 13 05:24:53 hm Jul 13 05:25:05 it'd be nice if I could make kdelibs-native build ;p Jul 13 05:25:37 btw, a lot of requires libXcursor and libXrandr are showing. If you have kde and upgrade xorg you will probably rebuild kde too Jul 13 05:26:08 fun -.-;; Jul 13 05:26:15 I'm thinking I won' Jul 13 05:26:15 t Jul 13 05:26:47 * thejapa has no turning back. already unmerged xorg Jul 13 05:27:45 FEATURES=buildpkg Jul 13 05:59:09 good morning all Jul 13 06:09:15 <_law_> hi koen Jul 13 06:24:35 hey _law_ Jul 13 06:38:33 good morning Jul 13 06:40:07 koen: my bootstrap-image says 'Cannot find package task-bootstrap' but i ran task-bootstrap sccessfully before Jul 13 06:40:18 where is my fault? Jul 13 06:40:50 did it build task-bootstrap or task-bootstrap-unionroot ? Jul 13 06:41:57 task-bootstrap Jul 13 06:42:33 but wait a min. please. maybe i found my mistake Jul 13 06:58:10 still exists Jul 13 06:58:57 it still doesn't find task-bootstrap package Jul 13 07:06:48 koen: do you have an idea Jul 13 07:07:21 sorry, no Jul 13 07:08:44 :-( It happens since i updated my db. the task-bootstrap...ipkg is in tmp/deploy/ipk folder. Jul 13 07:08:59 koen: do you know someone who can help? Jul 13 07:21:36 koen: i figured out where the error comes from Jul 13 07:21:54 koen: do you know where/when ipkg.conf i created? Jul 13 07:22:07 iirc on do_rootfs Jul 13 07:25:01 where do i find do_rootfs? Jul 13 07:25:13 CoreDump|afk: I just finished testing my opie image.. Did you have problems with some of the keys not working in opie? The home key does not switch tabs anymore and the OK button does not seem to work right anymore :( Jul 13 07:25:30 I am turning in now, see you in the moring :) Jul 13 07:29:58 TEMic: classes/rootfs_ipk.bbclass (iirc) Jul 13 07:30:41 <_law_> any ideas why glibc doesnt build? http://rafb.net/paste/results/WCDCAE24.html Jul 13 07:31:34 what version of binutils did get build? Jul 13 07:32:21 <_law_> koen, binutils-cross-2.17-r0 Jul 13 07:32:33 * zwelch tries not to laugh Jul 13 07:32:42 try 2.16 :) Jul 13 07:33:00 PREFERRED_VERSION_binutils-cross = "2.1.6", i beleive Jul 13 07:33:58 it's binutils-cross = "2.16" Jul 13 07:35:05 Hi, I have a question. Say if I modified a pachage, how to rebuild it? Using bitbake interactive mode and rebuild package just bring a brand new download and unpack etc. Jul 13 07:35:33 TEMic: are you sure? i just checked my local.conf and my example is what i have Jul 13 07:35:48 erp.. Jul 13 07:36:06 I mean I just want the bitbake to do ./configure && make. Thank you. Jul 13 07:36:06 sorry, somehow i manged to overlook my typo several times Jul 13 07:36:08 :) Jul 13 07:36:40 Kangkai: rebuild shouldn't download everytime Jul 13 07:37:03 oh, you mean to modify the work copy and have it rebuild? Jul 13 07:37:35 yes, zwelch Jul 13 07:37:47 you can rm tmp/stamps//do_compile* to force a normal build to do that step again Jul 13 07:38:08 except s,><,>/<, Jul 13 07:38:36 bitbake -b file.bb -c compile -f Jul 13 07:39:19 ah, -f is useful Jul 13 07:40:05 iirc I filed a bug that -i doesn't have -f Jul 13 07:40:06 my pattern of development revolves around creating a patch that gets used by bitbake, then just rewriting the patch until i get it right Jul 13 07:40:06 zwelch: yes, i'm sure. I used it sucessfully yesterday Jul 13 07:40:42 TEMic: i missed my typo of "2.1.6" if that's the bit you were correcting; sorry :) Jul 13 07:40:54 For example, I modified something in the oe/build/tmp/work/e680-kernel-2.6-ezx6-r12/linux-2.6.16 Jul 13 07:41:41 zwelch: :-) Jul 13 07:41:50 bitbake -b file.bb -c compile -f will do the "make oldconfig && make" thing? Jul 13 07:43:45 Kangkai: no, it will only do the compile step Jul 13 07:43:51 morning all Jul 13 07:43:59 bitbake -b file.bb -c configure -f would do the configure step Jul 13 07:44:02 hey RP Jul 13 07:45:05 morning Jul 13 07:46:16 I'm afternoon here :) Jul 13 07:46:31 morning all Jul 13 07:46:37 ~ugt Jul 13 07:46:39 well, ugt is Universial Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. Jul 13 07:47:10 Oh, thanks Jul 13 07:47:25 * zwelch laughs Jul 13 07:47:30 that's a good one Jul 13 07:47:54 it's missing the part about how everything in between is "high noon" ;) Jul 13 07:48:33 morning Jul 13 07:50:30 what do i have to put in IPKG_ARCHS? Jul 13 07:50:45 03koen 07org.oe.dev * r76c8df08... 10/conf/distro/openzaurus-unstable.conf: oz-unstable: lock down binutils tils oz switches to a newer glibc Jul 13 07:50:51 (exept i486) Jul 13 07:51:03 TEMic: anything your machine can support Jul 13 07:51:34 RP: does $MACHINE get appended automagically nowadays? Jul 13 07:52:01 koen: i found in an older config from a collegue things like 'all' 'any' 'noarch' and so on Jul 13 07:52:12 koen: Yes Jul 13 07:52:48 TEMic: grep for IPKG_EXTRA_ARCHs Jul 13 07:53:00 Its usually better to set that Jul 13 07:53:03 (in .dev) Jul 13 07:54:08 anyone got any ideas what this means? : Jul 13 07:54:09 http://rafb.net/paste/results/5xcAT114.html Jul 13 07:55:53 tkp: What did you run to cause that? Jul 13 07:56:24 http://rafb.net/paste/results/bTYXix38.html Jul 13 07:56:44 tkp: sudo? Jul 13 07:56:55 tkp: never build as root (or sudo) Jul 13 07:57:11 tkp: you might end up with cross packages in your host filesystem Jul 13 07:57:16 mm. for some reason it wouldn't grant me access to the cache file without it Jul 13 07:57:18 or worse Jul 13 07:57:21 badly designed buildsystems and all Jul 13 07:58:33 I tried to take a copy of the build directory right after building the toolchain Jul 13 07:58:44 I think permissions must have been screwd up in doing that Jul 13 08:02:50 tkp: I'm not sure why it would do that... Jul 13 08:03:22 well when I tried to copy it... it missed out a few files due to permissions Jul 13 08:03:29 su I reran cp as sudo Jul 13 08:03:43 so I..* Jul 13 08:04:46 tkp: Try wiping the bitbake cache file... Jul 13 08:11:11 ok, that seems to be ok noe :) Jul 13 08:11:14 now* Jul 13 08:16:57 yeah, permissions seem to be really screwed Jul 13 08:17:17 the entire build directory is owned by root Jul 13 08:17:38 is it safe to just chown -R on the entire build directory Jul 13 08:17:39 tkp: you copied it with sudo, what did you expect? Jul 13 08:18:09 I copied it with sudo because some files got missed out when I tried copying it as myself Jul 13 08:18:32 tkp: I fail to understand why you are surprised root now owns it Jul 13 08:18:32 should have used cp -a Jul 13 08:18:37 I'm not Jul 13 08:18:48 tkp: but yeah, just chown it Jul 13 08:18:52 I'm jjust wondering if it's ok to chown -R on it Jul 13 08:18:54 cool Jul 13 08:20:22 kate is kool. it changes the konsole dir when you switch the files you're editing :P Jul 13 08:24:13 koen, RP: Thanks a lot! My IPKG_ARCH is now complete and my bootstrap-image finds task-bootstrap package too :-))) Jul 13 08:41:31 he Jul 13 08:41:32 y Jul 13 08:44:21 hi zecke Jul 13 08:44:43 h Jul 13 08:44:45 i zecke Jul 13 08:47:57 good morning Jul 13 08:48:18 JustinP: You were talking about bug 1074 yesterday, right? Jul 13 08:48:29 Just stumbled across it. Jul 13 08:53:19 hmm, I'm tempted to say. File this bug for your bash Jul 13 08:53:29 or quilt Jul 13 08:56:10 JustinP: I think, bitbake is not searching for patch in the PATH ( I could be wrong though ). It looks like quilt found your patch in bitbake/bin/patch Jul 13 08:56:24 JustinP: this means, the autoconf magic of quilt should be better Jul 13 08:58:17 anyone have any ideas why building python would require a build of x11? Jul 13 08:58:55 tkp: to support the tcl/tk extensions Jul 13 08:59:17 hmm... is that really needed to use python? Jul 13 08:59:29 not needed Jul 13 08:59:38 you should be able to disable it Jul 13 08:59:42 so if I want to disable it? Jul 13 08:59:47 either way only the tk plugins depend on x Jul 13 09:00:42 I'm trying to build python-kaa (this is a bb I'm creating for a new python module) Jul 13 09:01:07 when it runs setup.py it tests for thee presence of X11, and if it find's it, it enables some stuff Jul 13 09:01:21 tkp: we definately dont need tcl/tk Jul 13 09:01:27 I don't want any presence of X11 on my system Jul 13 09:02:09 yet building python is building some X11 stuff - which is not what I want Jul 13 09:02:54 tkp: then use --disable-x11 or something in the kaa .bb Jul 13 09:03:20 I'm wondering... even if I have no presence of X1 in the build/staging directory, maybe it will pick up the presence of the X11 on my system Jul 13 09:03:20 otherwise your .bb is subject to heisenbugs Jul 13 09:03:44 koen: true... but this setup.py has no option to --disable-x11 Jul 13 09:03:48 or DEPENDS="libX11" Jul 13 09:04:11 tkp: then setup.py is broken Jul 13 09:04:25 tkp: unless you want to force users to build kaa from scratch Jul 13 09:04:42 hmm. gonna need to patch setup.py then I guess Jul 13 09:05:24 tkp: yeah, there ought to be a way to do that really Jul 13 09:05:35 we can patch it and send it to Tack Jul 13 09:06:43 morning Jul 13 09:07:20 koen: isnt the norm for python stuff to work that sort of thing out on its own? Jul 13 09:07:45 i mean, how do you pass something like --disable-x11 to a python setup.py ? Jul 13 09:08:34 jkp: if that's the norm, it's broken Jul 13 09:08:59 whats the difference between libX11 and diet-x11? Jul 13 09:09:45 diet-x11 is smaller because it stubs out 'unneeded' sstuff Jul 13 09:09:59 TEMic: diet-x11 stubs out s load of stuff not need by most xclients Jul 13 09:10:58 although .jp software does tend to have problems due to input methods Jul 13 09:11:16 thx Jul 13 09:12:39 other libs can provide the layers needed though, but which and how i don't know, asking jg or mallum might give better info Jul 13 09:13:22 ~ seen jg_ Jul 13 09:13:35 ~seen jg_ Jul 13 09:13:35 jg_ was last seen on IRC in channel #gpe, 129d 6h 14m 36s ago, saying: 'joshua_: you too ;-).'. Jul 13 09:13:37 jg_ was last seen on IRC in channel #gpe, 129d 6h 14m 38s ago, saying: 'joshua_: you too ;-).'. Jul 13 09:24:03 03tmbinc 07org.oe.dreambox * r7926bfdb... 10/classes/package.bbclass: classes/package.bbclass: temporary workaround to throw away .pyo files Jul 13 09:24:08 03tmbinc 07org.oe.dreambox * r8a16da21... 10/packages/dreambox/dreambox-dvb-modules.bb: dreambox-dvb-modules/dm7025: update to 20060711 Jul 13 09:24:13 03tmbinc 07org.oe.dreambox * rc7ee859a... 10/packages/enigma2/enigma2.bb: enigma2: update to 20060711 plus some fixes Jul 13 09:24:17 03tmbinc 07org.oe.dreambox * r96a5ca14... 10/packages/mc/mc_4.6.0.bb: mc: ncurses-terminfo is overkill, don't RDEPEND on it Jul 13 09:24:21 03tmbinc 07org.oe.dreambox * r299007f2... 10/packages/meta/dreambox-image.bb: dreambox-image: remove ncurses-terminfo, bash, add satfinder plugin Jul 13 09:24:28 03tmbinc 07org.oe.dreambox * rfeffecbe... 10/packages/tuxbox/tuxbox-common.bb: tuxbox-common: obey CVSDATE Jul 13 09:31:10 pb_: master? which signal is sent by alignment trap? Jul 13 09:31:20 pb_: if I use signal+warn? Jul 13 09:31:26 SIGTRAP? Jul 13 09:31:53 zecke: IIRC, it's SIGBUS.. Jul 13 09:33:21 thanks Jul 13 09:33:24 hmm it gets killed Jul 13 09:34:48 hi Jul 13 09:34:58 hey Jul 13 09:41:22 * koen has a great idea Jul 13 09:41:33 now we have modular xorg, why not modular glibc? Jul 13 09:43:27 hehe Jul 13 09:43:35 no nis? Jul 13 09:44:47 koen: that is called uclibc Jul 13 09:49:05 hm Jul 13 09:49:51 hmm uclibc ain't quite ready for general glibc replacement Jul 13 09:50:09 close but not quite Jul 13 09:59:36 zecke: that's good to know Jul 13 10:02:43 ade|desk: do you have any insights when the next uclibc release will be? Jul 13 10:03:04 koen I asked anderseen last week Jul 13 10:03:23 he has much other things to do Jul 13 10:03:31 so it will be last a while Jul 13 10:07:50 i have two builds working at the moment. one bulit with a very old db and one bulit with an updated db. both use the same local.conf, but the GPE on the new bulit is 10x slower than the old one Jul 13 10:08:20 TEMic: check the gtk versions Jul 13 10:09:13 koen: when sjhill has ported his NPTL bit and mjn3 has played a bit more, psm has been taken seriously and everyone has built buildroot apps perfectly and the moons aligned .. ish Jul 13 10:10:16 both builts have the gtk-doc-1.0-r1 and gtk-doc-native-1.0-r1 packages nothing else with gtk Jul 13 10:10:44 TEMic: libgtk* Jul 13 10:11:35 ade|desk: psm? Jul 13 10:11:58 Peter S. Mazinger Jul 13 10:12:16 gentoo embedded bloke Jul 13 10:12:52 XorA, koen: nothing Jul 13 10:13:12 driving forward changes w.r.t sus3 yet everyone is hammering him for breaking buildroot in the process Jul 13 10:13:42 my personal view is for him to continue, get uclibc better then go fix buildroot Jul 13 10:13:44 * koen wonders how to run gpe with gtk Jul 13 10:13:58 i'm using xorg6.8.1, gpe-dm, matchbox Jul 13 10:13:59 TEMic: 'ipkg list_installed | grep gtk' Jul 13 10:14:37 koen: nada Jul 13 10:14:46 then gpe will never work Jul 13 10:16:00 well, i can see a blue desktop background, some buttons at the bottom and a terinal opened... Jul 13 10:16:13 DEPENDS = "gtk+ cairo libxrender gtk-doc" Jul 13 10:16:37 ade|desk: which package? Jul 13 10:16:54 libgpewidget Jul 13 10:17:13 which most gpe stuff needs, if not all Jul 13 10:17:48 can I set an environment variable from a bb file? Jul 13 10:18:22 tkp: export blob perhaps? Jul 13 10:18:39 export.. ok, cool Jul 13 10:19:26 i have BOOTSTRAP_EXTRA_DEPENDS ="xserver-xorg xfonts-xorg ttf-bitstream-vera xset xrdb rxvt-unicode gpe-dm matchbox" Jul 13 10:19:37 wow i had no idea gpe inherits gettext ... nasty Jul 13 10:19:47 then thats not gpe Jul 13 10:19:59 ok sorry, my fault. Jul 13 10:20:16 what is it if not gpe? Jul 13 10:20:29 TEMic: that is just a window manager and x terminal Jul 13 10:20:41 mmm... k Jul 13 10:21:07 ade|desk: gpe did gain some ifdef #NLS a few months ago Jul 13 10:21:09 you wouldn't even need the gpe-dm there really either Jul 13 10:21:31 then, the wm+term of the new built is much slower than the older one ;-) Jul 13 10:21:43 koen: its just set in gpe.bbclass Jul 13 10:21:51 i used it to start matchbox Jul 13 10:22:14 i'd guess the xserver is different by someway Jul 13 10:22:19 ~lart qtdesigner Jul 13 10:22:19 * ibot executes killall -HUP qtdesigner Jul 13 10:22:58 check for alignment errors in dmesg Jul 13 10:23:04 wow just realised which xserver etc you are using Jul 13 10:23:34 ? Jul 13 10:24:06 TEMic: you are using the full blown versions of xorg Jul 13 10:24:26 TEMic: what is your target ? Jul 13 10:24:33 geode lx Jul 13 10:24:55 and the goal ? Jul 13 10:25:17 linux on geode ;-) Jul 13 10:25:46 TEMic: why not use debian then ;) Jul 13 10:25:58 too easy ;-) Jul 13 10:26:18 it should run with a 128mb CF card Jul 13 10:26:35 (incl. some space for data) Jul 13 10:26:43 TEMic: most folks here use OE to create smaller installs for various targets Jul 13 10:27:01 TEMic: so they use diet-x11 Jul 13 10:27:47 I've chosen xorg because of the geode support patch Jul 13 10:27:47 and kdrive xserver Jul 13 10:28:00 would it run on my target? Jul 13 10:28:28 with matchbox-session and terminal? Jul 13 10:28:57 why not? you can use fb or vesa Jul 13 10:29:21 bbl , meeting Jul 13 10:29:33 cya Jul 13 10:29:48 * TEMic will try a smaller built Jul 13 10:40:28 now thats the meeting style i like Jul 13 10:40:42 speed meetings Jul 13 10:40:56 ade lol Jul 13 10:44:11 * ade|desk looks at calendar and spots another meeting which will be much longer Jul 13 10:44:11 while I'm trying to build a smaller distro, the speed problem between my big distros still exists... Jul 13 10:45:26 TEMic: have you taken a look at the differences between the two setups, used ipkg to get version number changes etc ... that may help Jul 13 10:45:39 else we are firing blind Jul 13 10:47:19 i compared the package versions. the biggest difference, to my mind, is the libx11. in the newer built, it is libx11R7. the older one has no 'libx11' pakages, but an x11-6.2.1 Jul 13 11:11:55 morning all Jul 13 11:12:03 hi hrw Jul 13 11:12:15 10h offline Jul 13 11:13:53 hey hrw|work Jul 13 11:15:13 hrw why? Jul 13 11:15:38 woglinde: ISP problem Jul 13 11:16:05 hrw :( Jul 13 11:30:50 mickey|zzZZzz: you awake? ;) Jul 13 11:30:58 zecke: ping Jul 13 11:31:16 /quit Jul 13 11:31:18 * jkp notes the zzZZzz in that nick :) Jul 13 11:31:22 TEMic: yeah Quit! Jul 13 11:48:07 After cleaning out tmp, it seems that OE might be willing to cooperate again ;-) While recompiling everything, I observed the following: http://rafb.net/paste/results/wdZcjv34.html. Is that something that should be fixed? Jul 13 11:48:34 pong Jul 13 11:49:02 Laibsch: no Jul 13 11:49:19 zecke: OK, thanks. Jul 13 11:49:21 zecke Jul 13 11:58:02 woglinde: yes? Jul 13 12:00:21 hm Jul 13 12:00:23 nothing Jul 13 12:00:29 sorry Jul 13 12:00:36 to bothered you Jul 13 12:00:46 heaeh? Jul 13 12:00:54 it is too hot Jul 13 12:04:39 zecke: was wondering if my problem is due to bitbake cache or not, let me explain .... Jul 13 12:06:26 ade|desk: I have missed your question/problem Jul 13 12:07:13 when i do a build of uclibc on day A giving uclibc-0.9.28+svn20060711-r2 fine, then next day to an oe pull and update, rebuild uclibc , uclibc-0.9.28+svn20060712 not available using 0.9.28. Jul 13 12:07:40 yet if i clear the cache , it can build uclibc-0.9.28+svn20060712 Jul 13 12:08:21 is this bitbake or oe problem ? Jul 13 12:08:35 bitbake Jul 13 12:09:06 would this be classed as a bug ? Jul 13 12:09:17 or user error Jul 13 12:21:23 How do I choose xserver-kdrive in BOOTSTRAP_EXTRA_(R)DEPENDS? Jul 13 12:22:18 PREFERRED_PROVIDER_virtual/xserver="xserver-kdrive" Jul 13 12:36:10 somebody rang? Jul 13 12:44:01 * bgat is back Jul 13 13:00:42 I really like the direction OE seems to be going in :) Jul 13 13:01:15 the split arch work/stamps suggests unification of host/target bb files Jul 13 13:01:47 Luke-Jr: multimachine.inc is over two years old Jul 13 13:02:11 koen: the split build directories isn't Jul 13 13:02:30 and I think OE is still using separate -native bbs Jul 13 13:02:57 Luke-Jr: multimachine *is* the split directories Jul 13 13:03:15 well, the directories certainly haven't been split for 2 years Jul 13 13:03:53 not if you didn't use multimachine.inc (now multimachine.bbclass) Jul 13 13:04:13 I used whatever the OZ was using Jul 13 13:07:11 Luke-Jr: It just never used to be the default Jul 13 13:07:32 morning Jul 13 13:07:40 hi chouimat|ibook Jul 13 13:07:59 hey chouimat|ibook Jul 13 13:08:20 aww, does that mean *-native won't turn into simply regular bbs build for the native arch? Jul 13 13:08:35 RP: you wouldn't have any patches for mpfr or gcc to enable fotran in poky, would you? Jul 13 13:09:03 koen: With my love of fortran? Jul 13 13:09:15 That's a no btw ;-) Jul 13 13:09:18 RP: :) Jul 13 13:09:48 Its the only langauge I was properly taught Jul 13 13:10:29 gcc doesn't find -lmprfr and mpfr only produces libmpfr.a Jul 13 13:10:47 but I don't need fortran Jul 13 13:11:03 Does anyone need fortran anymore? Jul 13 13:11:10 R Jul 13 13:11:25 http://www.r-project.org/ Jul 13 13:11:32 they need the f2c part Jul 13 13:11:58 I was trying to compile it natively, but the OE gcc 4.1.1 doesn't have f2c (or f77, f90, etc) Jul 13 13:12:35 f2c doesn't work very well Jul 13 13:12:42 g77/f77 is better Jul 13 13:13:06 * RP is keeping well clear ;-) Jul 13 13:13:07 I need fortran for Octave too Jul 13 13:13:26 iirc the cross toolchain did have fortran Jul 13 13:13:32 but that could have been gcc 3.x Jul 13 13:13:51 It did used to have it, may still for all I know Jul 13 13:13:52 lardman: EABI also solves your float problems Jul 13 13:14:06 koen: yes :) Jul 13 13:15:10 koen: wrt R, I think cross-compiling is still possible, I must get back to it, now I've got some more time Jul 13 13:16:10 lardman: could be, but I want something to test the native toolchain Jul 13 13:16:19 something more advanced as hello.c Jul 13 13:17:46 I'd say octave, it's a long compile, but doesn't need any patching, etc., nor wierd tools other than g77/f77/f2c Jul 13 13:18:21 actually no, it's a long compile on my p4, so your Z might not like it much Jul 13 13:19:52 good thing it's not giac :) Jul 13 13:19:58 yeah Jul 13 13:20:06 that compiles though, if it's patched... Jul 13 13:20:23 and the -o2 is removed - that killed my build machine Jul 13 13:21:15 Try mathomatic Jul 13 13:21:35 http://mathomatic.orgserve.de/math/ Jul 13 13:32:38 Unable to generate local paths for SRC_URI due to malformed uri: cvs while parsing /home/koen/OE/monotone/org.ope Jul 13 13:32:38 nembedded.dev/packages/linux/handhelds-un-2.6_2.6.16.bb Jul 13 13:34:11 lardman: mathomatic: ELF 32-bit LSB executable, ARM, version 1 (SYSV), for GNU/Linux 2.6.14, dynamically linked (uses shared libs), for GNU/Linux 2.6.14, stripped Jul 13 13:34:31 I'm trying to get bitbake to build my new version of mplayer (from svn), rather than 1.0pre8 Jul 13 13:34:35 root@c7x0:/usr/src/mathomatic-12.5.20# objdump -p mathomatic | grep EABI Jul 13 13:34:35 private flags = 4000002: [Version4 EABI] [has entry point] Jul 13 13:34:52 I have named my file mplayer_1.0pre8_svn.bb as opposed to mplayer_1.0pre8.bb Jul 13 13:35:03 yet it still seems to go for mplayer_1.0pre8.bb Jul 13 13:35:08 isn't there a mplayer_svn.bb? Jul 13 13:35:24 koen: so there is... Jul 13 13:35:29 how can I get it to use that? Jul 13 13:36:39 something todo with PREFERED_VERSION_mplayer right Jul 13 13:36:50 PREFERRED_VERSION_mplayer = "1.0pre8+svn20060713" or something like that Jul 13 13:36:52 = "svn" maybe? Jul 13 13:36:58 ah, ok Jul 13 13:37:14 grep PV mplayer/*.bb and you'll see why Jul 13 13:37:43 mplayer svn isnt upto date with config options Jul 13 13:38:07 no.. I need my own config options and patches anyway Jul 13 13:38:12 tkp: cool Jul 13 13:38:29 tkp: just thought I would warn you updating it was in my TODO Jul 13 13:39:14 it's just the configure bits thats out of date? Jul 13 13:39:45 seems to have only specifed a few of the options Jul 13 13:39:57 tkp: should be, I updated pre8 a few days ago but havent got round to replacing EXTRA_OECONF in svn yet Jul 13 13:40:18 ok.. Jul 13 13:40:22 koen: cool Jul 13 13:40:50 XorA: what did you change in Pre8? seemed to be working ok for me as it was Jul 13 13:41:20 tkp: I added many --disable to OECONF to make sure whats in tmp doesnt affect the mplayer that comes out at end Jul 13 13:41:33 tkp: causes havoc with world build otherwise Jul 13 13:41:46 oh ok, I didn't realize some were missing Jul 13 13:41:51 like kaa-canvas Jul 13 13:42:25 * tkp does a pull & update Jul 13 13:43:04 XorA: thought about adding support for the essential codecs package? Jul 13 13:43:32 may be kind of tricky to do it properly tho I guess Jul 13 13:43:41 my one is a bit of a botch Jul 13 13:47:32 tkp: no, I will never add that to OE, though you can add a --enable-win32 for x86 targets Jul 13 13:47:46 XorA: yeah, thats what I did Jul 13 13:48:35 tkp: feel free to submit a patch I can add to mplayer to make your life easier Jul 13 13:50:27 bleh Jul 13 13:50:32 python makes my eyes bleed Jul 13 13:50:39 * koen closes vi Jul 13 13:51:14 XorA: well, the patch is this one: http://rafb.net/paste/results/lOudku83.html Jul 13 13:52:27 XorA: what's the deal with virtual/libsdl Jul 13 13:52:47 you must choose Jul 13 13:52:52 I don't need sdl in mine... what might be a sensible way to disable it? Jul 13 13:52:59 the ugly trollified libsl-qpe, or the regualr one Jul 13 13:53:07 what about none? Jul 13 13:53:17 --disable-sdl? Jul 13 13:53:26 so just edit the bbb file Jul 13 13:53:28 bb* Jul 13 13:53:30 DEPENDS_mymachine = "blah" Jul 13 13:53:53 oh ok, and that will override DEPENDS for my machine only Jul 13 13:54:23 can I do EXTRA_OECONF_mymachine too ? Jul 13 13:54:27 yes Jul 13 13:54:32 cool Jul 13 13:54:59 EXTRA_OECONF_append_mymachine would append it to the generic ones Jul 13 13:55:12 you can basically override any var in OE Jul 13 13:55:20 tkp: when you are finished, diff it all and stick it in bugtracker and as long as it doesnt interfere with OZ/Fam Ill stick it in OE Jul 13 13:55:22 and what if I wanted to override some... would append do that? Jul 13 13:55:39 sorry for all these noop questions! Jul 13 13:55:45 ;) Jul 13 13:56:09 matchbox-wm: can't open display! check your DISPLAY variable. Jul 13 13:56:16 how do i configure it? Jul 13 13:57:38 export DISPLAY= Jul 13 13:57:40 i switched from xserver-xorg, gpe-dm, matchbox to diet-x11, xserver-kdrive, matchbox Jul 13 13:58:35 koen: :-) sure, but where do i get the number from? Jul 13 13:59:01 it ignores my xorg.conf Jul 13 14:00:36 kdrive doesn't have any configuration file Jul 13 14:00:55 by default it starts on :0 Jul 13 14:03:28 matchboxĂ-panel: failed to open display Jul 13 14:04:55 where does it know from which device my geodefb is? and which screen res? Jul 13 14:06:25 TEMic: that will be determined by whatever your framebuffer settings are Jul 13 14:07:23 k Jul 13 14:10:07 what do i have to do to start matchbox? Jul 13 14:10:09 can I SRC_URI_append_stv too? Jul 13 14:10:16 to apply my own patch Jul 13 14:10:25 tkp: distro in in overrides as well Jul 13 14:10:45 TEMic: to start matchbox, you'll proabbly have to export your display variable Jul 13 14:10:51 koen: what do you mean? Jul 13 14:10:59 just do export DISPLAY=:0 before starting matchbox-session Jul 13 14:11:09 tkp: I though stv was your distro name Jul 13 14:11:10 tkp: I assume that DISTRO == stv Jul 13 14:11:20 koen: thats right Jul 13 14:11:57 so I need SRC_URI_append_stv += " Jul 13 14:12:22 why append and += Jul 13 14:12:49 zecke: thats why I'm asking.. I'm not too sure of the correct method/syntax Jul 13 14:12:59 SRC_URI_append_stv = " something" Jul 13 14:13:04 note the leading space Jul 13 14:13:23 koen: ok, that's why I had the += Jul 13 14:13:27 hi Jul 13 14:13:35 hey Bernardo Jul 13 14:16:40 CosmicPenguin: That doesnt work! DISPLAY=:0/1/2 plus x-window-manager always says 'unable to open display :-(( Jul 13 14:17:04 Are you sure the server is running? Jul 13 14:17:32 CosmicPenguin: no Jul 13 14:19:42 Ok - well, I'm thinking thats the first step to take Jul 13 14:21:03 * koen updates http://www.openembedded.org/openembedded-migration-from-monotone-0-25-to-0-27 Jul 13 14:21:58 bugger that was a long meeting Jul 13 14:22:26 * ade|desk feels like 2 hours of life has been ripped from his existance Jul 13 14:22:44 ade|desk: They're so rewarding, aren't they? =) Jul 13 14:23:28 and at the end, actions: 0 Jul 13 14:24:14 ade|desk: hehe Jul 13 14:24:53 ade|desk: when I was head of the technical department we had a rule if there were no action someone had to bring a pie to the next meeting Jul 13 14:25:17 koen:sounds like a grand idea Jul 13 14:25:18 ade|desk: 5 pies and 5 meetings later we reconsidered that plan Jul 13 14:25:32 the chairman was complaining about getting too fat Jul 13 14:25:57 ah and i missed lunch for this too, now i am really annoyed Jul 13 14:26:20 damn Jul 13 14:26:24 * ade|desk heads to find some form of food, thank god i'm off tomorrow Jul 13 14:26:40 our brand spanking new exchange server keeps the messages when downloaded via pop Jul 13 14:26:58 and the webform lacks a 'select all' Jul 13 14:27:33 and to top it all off: Jul 13 14:27:39 hmm. I have set PREFERRED_VERSION_mplayer = "1.0pre8+svn20060713" Jul 13 14:27:41 but I just get NOTE: preferred version 1.0pre8+svn20060713 of mplayer not available Jul 13 14:27:55 "I don't know this 'fetchmail' client, use another (supported) one, like outlook" Jul 13 14:28:17 * koen decides to hurry up with completing uni Jul 13 14:28:40 tkp: SRCDATE_mplayer = "20060713" Jul 13 14:29:01 tkp: and if you haven't set a global srcdate, harass one of the bitbake hackers Jul 13 14:29:11 koen: in distro.conf? Jul 13 14:29:23 yeah or local.conf Jul 13 14:31:48 I haven't set a global srcdate... and adding SRCDATE_mplayer = "20060713" to DISTRO.conf made no difference :( Jul 13 14:34:03 weird Jul 13 14:37:51 koen: getyour uni to do imap rather than pop Jul 13 14:38:43 ok, it's as soon as I add SRC_URI_append_stv = " ..." that it fails to recognise the correct version Jul 13 14:38:49 remove that and it's fine again Jul 13 14:39:21 ade|desk: it does imap, but I want the pop message to go into my own imap automagically Jul 13 14:41:09 koen: fetchmail --all --flush should work on imap Jul 13 14:42:37 ah, --flush did the trick Jul 13 14:42:39 thanks Jul 13 14:47:13 hi Jul 13 14:55:22 koen: no problem Jul 13 14:56:21 koen: its what being an MCSE is all about, knowing how crap microsoft products are ;) Jul 13 15:02:15 have a nice time guys - vacations starts Jul 13 15:02:48 hrw|work: enjoy Jul 13 15:02:54 hrw|work: enjoy Jul 13 15:05:04 have a fast migration to monotone 0.27 so I will skip it ;) Jul 13 15:05:12 maybe will appear tomorrow for a moment Jul 13 15:05:13 cu Jul 13 15:05:33 Shouldn't task-bootstrap have a DEPEND? Jul 13 15:12:47 zecke: yeah, it could have been quilt. It's not my shell, it doesn't try to run non-executables from the path, as it should. Jul 13 15:12:58 and quilt is from OE, so.... Jul 13 15:13:10 JustinP: lol Jul 13 15:13:21 JustinP: go to your quilt tools Jul 13 15:13:32 JustinP: head * | grep bitbake/bin/patch Jul 13 15:14:17 I sound mean again :} Jul 13 15:14:37 If the main bb file has --disable-directfb in its EXTRA_OECONF, and I then do EXTRA_OECONF_stv=" --enable-directfb", will that override properly? Jul 13 15:16:01 it should Jul 13 15:16:21 it should Jul 13 15:16:59 cool :) Jul 13 15:20:28 later Jul 13 15:21:30 Kerwood: if you are looking for a place to post the tip about binutils, put it on the wiki:OeFaq page. Jul 13 15:27:10 cbrake: ok, I'll move it from CommonProblems Jul 13 15:28:03 Kerwood: I think we should get rid of CommonProblems and move everything into one Faq -- any opinions? Jul 13 15:28:11 Shouldn't task-bootstrap have a DEPEND? If not, why not/ Jul 13 15:29:00 Kerwood: one problem we had w/ the old wiki was too many pages w/ no real content. My suggestion for this round is wait until existing pages get too big before creating more pages. Jul 13 15:30:36 Kerwood: this is a general comment and not directed to anyone in particular Jul 13 15:33:20 cbrake: sorry, I was busy looking at the patch mickeyl put in, which appears to fix the problem my wiki addition was meant to work around. Jul 13 15:33:43 cbrake: The problem was we had two sets of qualitively different data in the wiki before and I split them for that reason Jul 13 15:35:26 cbrake: Anyway, I rather agree with you. they didn't seem like they should be apart Jul 13 15:35:52 cbrake: I'm still looking for an automated way to generate TOCs Jul 13 15:36:35 koen: you must have added the manual TOC to the faq? Its better than nothing, but would be nice if there were links to entries. Jul 13 15:36:35 koen: mediawiki can do that iirc Jul 13 15:36:56 RP: any wiki can Jul 13 15:37:05 RP: just not the plugin we use :( Jul 13 15:37:17 :-/ Jul 13 15:37:38 I was evaluating this wiki and *bam* oe.org went lieve Jul 13 15:37:44 2 weeks to soon Jul 13 15:37:51 s/lieve/live/ Jul 13 15:38:33 koen: what are the advantages of keeping the wiki inside drupal instead of implementing mediawiki, etc? Jul 13 15:38:57 cbrake: the 'search' link at the top Jul 13 15:39:06 koen: nod Jul 13 15:39:14 and everything is in one place Jul 13 15:39:20 and that's about it Jul 13 15:39:44 koen: yeah, there is some benefit to that. Jul 13 15:39:50 moin changed it's storage format, so I gave up on that and converted a few wikipages by hand Jul 13 15:40:08 but if anyone is willing to implement a better system, I'm all for it Jul 13 15:41:24 koen: so you could not move to the latest version of moin on the new server? Jul 13 15:41:44 I had the latest version and installed the backup Jul 13 15:41:58 the user and such were all there, but no pages Jul 13 15:42:16 so I got annoyed and installed the wiki plugin for drupal to play with Jul 13 15:42:42 (and moin blewup after apt-get upgrade) Jul 13 15:42:49 ah Jul 13 15:42:57 I remember why I didn't install mediawiki Jul 13 15:43:03 it doesn't run on postgres Jul 13 15:43:16 * koen doesn't trust mysql Jul 13 15:43:19 koen: yeah, its painful keeping web apps going. Jul 13 15:43:33 ~lart me for looping 2 switches Jul 13 15:43:46 ok, sure enough, mickeyl dealt with the binutils version problem properly, so my wiki addition is dustbin-bound... Jul 13 15:44:14 Kerwood: the binutils problem is solved for a few distro's, not all Jul 13 15:45:10 koen: understood but mine only addressed the .dev branch of OZ Jul 13 15:45:35 s/OZ/OE/ Jul 13 15:45:46 sorry, right Jul 13 15:48:47 koen: well, I'd be glad to work on the wiki once we figure out which way we want to go. I personally prefer mediawiki, but if makes sense to hack on the drupal plugin, I can look at that as well. Jul 13 15:48:59 gpe-image DEPENDs on task-bootstrap and task-gpe Jul 13 15:49:02 Kerwood: and it's only fixed for distro=generic, distro=oz-instable and distro=angstrom-2006.9 Jul 13 15:49:08 neither of the tasks DEPEND on anything else Jul 13 15:49:14 what causes stuff to get built? Jul 13 15:49:19 Luke-Jr: rdepends Jul 13 15:49:45 koen: RDEPENDS are supposed to be built prior to the package? Jul 13 15:49:58 cbrake: it's http://drupal.org/node/53517 with a few patches from the bugtracker Jul 13 15:50:10 Luke-Jr: rdepends get added to depends nowadays Jul 13 15:50:12 shouldn't the images all have a (build-time) DEPEND on anything they contain? Jul 13 15:50:13 oh Jul 13 15:50:42 well, BB doesn't seem to be building the packages before it tries to make the image :( Jul 13 15:50:47 unless I add stuff to DEPEND Jul 13 15:52:05 it was initially, hit a 404 and aborted, and when I tried again it went straight to making an image without completing the packages Jul 13 15:52:40 Luke-Jr: images only have runtimes requirements, since they don't need .h and .o files to get 'built' Jul 13 15:52:46 koen: hmm... ok, well I'll leave my wiki blurb in, but with another qualification. thanks Jul 13 15:53:14 koen: images need the regular packages to build the rootfs Jul 13 15:54:13 I'm also hitting stuff like numerous symlinks being unpackaged with busybox... maybe not related Jul 13 15:54:36 Luke-Jr: yes, and packages are runtime stuff Jul 13 15:54:57 koen: packages are a build-time DEPEND for an image... Jul 13 15:55:04 you can't build the image without the package Jul 13 15:55:15 packages are never build time Jul 13 15:55:24 only .h and .o files are buildtime Jul 13 15:55:29 so you can build the image without packages? Jul 13 15:55:52 I didn't say that Jul 13 15:56:06 I said that packages go into RDPEPENDS Jul 13 15:57:54 All I'm saying is that by definition, the packages are build-time dependencies for an image Jul 13 15:58:09 if runtime dependencies are appended to build-time dependencies, then it's trivial Jul 13 15:58:25 Just need to figure out why the packages aren't being built prior to the image Jul 13 16:31:00 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Jul 13 16:38:44 03kristoffer 07org.oe.dev * r4110ff42... 10/ (7 files in 5 dirs): (log message trimmed) Jul 13 16:38:44 DISTRO Jul 13 16:38:44 ----- Jul 13 16:38:44 distro/jlime-donkey.conf : Setting sysvinit terminals Jul 13 16:38:44 distro/preferred-opie-cvs-versions.inc : Fix of typo Jul 13 16:38:45 KERNEL Jul 13 16:38:46 ------ Jul 13 16:42:33 he koen Jul 13 16:43:12 hey woglinde Jul 13 16:44:27 snooker time, cya folks Jul 13 16:44:39 hm too fast Jul 13 16:44:46 ok, http://www.openembedded.org/wiki/OeFaq should look a little cleaner now and have all "Common Problems" moved to it. Jul 13 16:46:01 Luke-Jr: .dev appends RDEPENDS to DEPENDS, .oz354x does not Jul 13 16:50:18 CoreDump|home: Did you have to do anyting to your keymap to get ctrl arrow to work in opie on poodle? Jul 13 16:52:54 ~seen CoreDump|home Jul 13 16:53:00 coredump|home is currently on #gpe #oe #openzaurus, last said: 'hrw|work: have fun!'. Jul 13 16:53:16 Kerwood: thanks, I did change the wording a bit Jul 13 16:53:34 hvontres|poodle: you mean Fn + arrow? Jul 13 16:53:44 heh Jul 13 16:53:48 great timing :) Jul 13 16:54:05 Kerwood: looks good Jul 13 16:54:38 heh Jul 13 16:55:10 CoreDump|home: Morning Jul 13 16:55:26 cbrake: thanks. let's hope it raises the SNR here :) Jul 13 16:56:36 morning hvontres|poodle Jul 13 16:57:05 hvontres|poodle: there may be keymap changes I didn't push yet.... Jul 13 16:57:15 * CoreDump|home verifies Jul 13 16:57:47 CoreDump|Home: I got my opie image mostly running now.. It seems that some of the front pan Jul 13 16:58:15 CoreDump|home: Ahh, I see...:) Jul 13 16:58:58 * koen crosses fingers for having a usuable 2.6 on poodle + tosa soon Jul 13 16:59:17 Now I just need to take a look at the kbdsim code for fbvncserver and see if I can get that Jul 13 16:59:20 koen and n770? *duck and hide* Jul 13 17:00:41 * Kerwood is impressed by Timelord's wirelessness Jul 13 17:01:06 brb Jul 13 17:01:30 pushing Jul 13 17:02:03 re Jul 13 17:02:23 ~lart admins for closing irc ports at work Jul 13 17:02:23 * ibot teaches admins the basics, including how to RTM for closing irc ports at work Jul 13 17:02:54 he you should work not chat Jul 13 17:02:56 hvontres|poodle: fwiw, while sound dependencies were added, they appear not to work Jul 13 17:03:33 03coredump 07org.oe.oz354x * re1e2c14d... 10/packages/keymaps/ (files/poodle/keymap-2.6.map keymaps_1.0.bb): keymaps: Add some forgotten poodle changes... Jul 13 17:04:08 CoreDump|home: Really? I must have added the right bits to mine from .dev. I will have to check when I get back home..:) Jul 13 17:04:15 I'm trying to get mplayer to compile with directfb support, but to do this it needs to be able to find the directfb headers Jul 13 17:04:42 I created a directfb-native package - very simple, it should work Jul 13 17:04:43 include directfb_${PV}.bb Jul 13 17:04:44 inherit native Jul 13 17:04:58 CoreDump|home: Ahh, muchos gracias.. I will fix my kemaps when I get home :) Jul 13 17:05:04 I now have the headers in tmp/staging/i586-oe-linux/include/directfb/direct Jul 13 17:05:06 np Jul 13 17:05:12 yyet mplayer still can't see them Jul 13 17:05:29 How can I pass extra include directives to mplayer? Jul 13 17:05:39 * CoreDump|home probably needs to rebuild task-bootstrap Jul 13 17:06:26 could it be because the mplayer bb file is using ./configure, rather than oe_configure Jul 13 17:06:29 ? Jul 13 17:08:16 Kerwood: I am sexy because I use bluetooth. XD Jul 13 17:08:44 tkp: if something needs headers to compile, 99.99% of the cases a -native package would be the wrong solution Jul 13 17:08:45 how do I get configure to use the stagingdir for it's includes? Jul 13 17:08:53 koen: oh Jul 13 17:09:04 koen: what would be a better solution Jul 13 17:09:16 DEPENDS += directfb Jul 13 17:09:17 the headers have been put into direcfb-dev Jul 13 17:09:26 koen: I have done that Jul 13 17:09:33 OE doesn't use the headers of the packages (yet) Jul 13 17:09:44 that's what google is paying me for :) Jul 13 17:09:59 so... how could I get it to see these headers?\ Jul 13 17:10:20 fix the do_stage in directfb Jul 13 17:10:39 all it does at the moment is "autotools_stage_all" Jul 13 17:10:52 RP: I just remembered one more issue that we have left to tackele on poodle. Currently the on/off switch is not being used for suspend. That might be a nice thing t Jul 13 17:11:29 RP: unless of course it is as borked as it was in the old 2.4 crappix kernel.. Jul 13 17:12:02 isn't that enough to get the headers installed? Jul 13 17:12:07 hmm. obviously not Jul 13 17:12:09 RP: Of course the "suspend on battery switch" is a nice touch..:) Jul 13 17:13:07 do I need oe_libinstall perhaps? Jul 13 17:13:29 all these oe_ functions are a bit of a mistry to me Jul 13 17:14:09 tkp: autotools_stage_all should work Jul 13 17:14:26 tkp: which headers does mplayer miss? Jul 13 17:14:32 well as I say, the headers are in tmp/staging/i586-oe-linux/include/directfb/direct Jul 13 17:14:34 and are those really missing in staging Jul 13 17:14:35 the directfb ones Jul 13 17:15:09 I have altered mplayer.bb to do --enable-directfb Jul 13 17:17:20 I can see that it's running with -isystem/home/tom/stv/oe-stv/build/tmp/staging/i586-oe-linux/include Jul 13 17:17:40 so I don't understand why it's missing these headers Jul 13 17:19:37 JustinP: I'm using .dev Jul 13 17:29:09 * SonicvanaJr stabs qemu-native Jul 13 17:31:38 ok, so the problem is that it needs the dev headers Jul 13 17:32:16 hmmm Jul 13 17:32:16 which don't seem to get installed by autotools_stage_all Jul 13 17:32:32 it seems package management a la debian is too hard for some users Jul 13 17:33:03 they keep downloading single files and trying to install them, and are surprised about dependency problems Jul 13 17:33:20 autoapt Jul 13 17:33:25 is there another magic oe function that will install dev headers in the stageing dir? Jul 13 17:33:34 tkp: 'cp' Jul 13 17:33:39 hehe Jul 13 17:33:44 thought as much :) Jul 13 17:33:50 hello everyone Jul 13 17:33:54 or 'install' Jul 13 17:33:57 if seen other packages do it Jul 13 17:33:59 hey Greg2 Jul 13 17:34:02 i've* Jul 13 17:34:10 tkp: http://www.openembedded.org/packagedstaging Jul 13 17:34:10 hey koen Jul 13 17:34:46 i'm having a problem compiling qemu-0.7.0 using oz354x for a poodle build Jul 13 17:34:53 could someone check my log: http://pastebin.ca/87141 Jul 13 17:35:06 i've been using dev with gcc 4.xx without any problems, but i though i would try RP's new sound patch and driver with oz354x :-) Jul 13 17:35:15 so i've rm -rf tmp... and start over Jul 13 17:36:45 Greg2: wild guess -> disable locale build in local.conf Jul 13 17:37:39 thejapa: i'll try Jul 13 17:40:16 Greg2 I've been having the same problems with qemu-native using akita :( Jul 13 17:40:39 SonicvanaJr: didn't work disabling locale build? Jul 13 17:40:45 Greg2: you can install qemu from your distro and ASSUME_PROVIDED +="qemu-native" Jul 13 17:40:46 nope Jul 13 17:41:00 the ENABLE_BINARY_LOCALE_GENERATION thing didnt' work Jul 13 17:41:13 Greg2: and if you use gcc 4.x you need to compile the kernel and alsa-lib with -O1 Jul 13 17:41:36 Greg2: see http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27363 Jul 13 17:41:46 koen why did it work ok for dev Jul 13 17:42:24 it will compile, but alsa won't work very good Jul 13 17:44:28 koen: the headers are there - in the staging dir Jul 13 17:44:38 they are in a subdirectory of include Jul 13 17:44:45 which is what I think the problem is Jul 13 17:45:05 {STAGING_DIR}/include in in the -I path Jul 13 17:45:19 but not the {STAGING_DIR}/include/directfb directory Jul 13 17:45:40 is it supposed to recurse into subdirectories? Jul 13 17:45:40 afk Jul 13 17:45:52 or do I need to explicetly tell it to do that? Jul 13 17:46:59 mplayer does: Jul 13 17:46:59 | Checking for extra headers ... -isystem/home/tom/stv/oe-stv/build/tmp/staging/i586-oe-linux/include Jul 13 17:47:27 but not ...include/directfb (which is where they actually are) Jul 13 17:48:14 tkp: CFLAGS+="-i${STAGING_INCDIR}/whateversubdirmplayerwantstohave/" Jul 13 17:48:21 s/i/I/ Jul 13 17:49:19 directfb-config *duck* Jul 13 17:49:44 directfb-config? Jul 13 17:49:51 directfb-config --cflags Jul 13 17:49:56 -I/usr/include/directfb -D_GNU_SOURCE -D_REENTRANT Jul 13 17:50:02 thats the normal way Jul 13 17:50:29 dont know if oe really support this Jul 13 17:50:35 brrr Jul 13 17:50:42 running generated files Jul 13 17:50:50 koen *g* Jul 13 17:51:33 directfb supports pgkconfig too Jul 13 17:51:51 mplayer probably doesn't Jul 13 17:52:13 I added CFLAGS+="-l${STAGING_INCDIR}/directfb to the mplayer bb file... Jul 13 17:52:13 mplayer doesn't support anything that is a) easy to use b) sane Jul 13 17:52:15 no change Jul 13 17:52:21 hehe Jul 13 17:52:28 tkp: -c clean it first Jul 13 17:52:33 i did Jul 13 17:52:54 koen other words mplayer suckz Jul 13 17:53:02 woglinde: yep Jul 13 17:53:31 hmm. mplayer has it's own --with-extraincdir=DIR Jul 13 17:53:50 *sigh* Jul 13 17:54:37 heh, that did it :) Jul 13 17:54:47 damn stupid mplayer Jul 13 17:54:54 *eyesroll* Jul 13 17:56:55 mplayer 1.0.5.5.32.-pre-5-alpha-2-no-developers-rc4-we-suck-4-releasing-is-hard-23.54.3-final2-try4 Jul 13 17:57:08 when they suck, they are good at it Jul 13 17:57:23 hehe :p Jul 13 17:57:54 maybee thats why the founders leave the project Jul 13 18:03:03 Does the package need to be within the OE tree to have RDEPEND copied to DEPEND? Jul 13 18:04:10 Can't create file /tmp/poiecaYP : No such file or directory hmm Jul 13 18:04:20 What did I mess up now :\ Jul 13 18:04:32 nm Jul 13 18:04:34 no disk space Jul 13 18:04:54 I would be suspect of a chroot lacking /tmp Jul 13 18:10:16 Yeah wow, I just rebooted, and it complained about diskspace, but its a 160GB harddrive, and I don't have anywhere near that much data on it Jul 13 18:11:15 O.o Jul 13 18:11:20 Exactly Jul 13 18:11:26 sure? Jul 13 18:11:35 OE builds can use a bit Jul 13 18:11:44 and you can run out of inodes Jul 13 18:11:54 koen: does anyone do that? Jul 13 18:12:05 I'd think it impractical with the block sizes Jul 13 18:13:24 bah what command would I use to check the disk space Jul 13 18:13:31 df -h Jul 13 18:13:55 lol Jul 13 18:14:14 yeah 76GB available Jul 13 18:14:24 on /tmp? Jul 13 18:14:34 false mounpoints? Jul 13 18:14:47 /tmp is usally on root Jul 13 18:14:57 is it? :) Jul 13 18:14:57 or ram Jul 13 18:15:03 RAM for me Jul 13 18:15:20 but default for the unexperienced user it is / Jul 13 18:16:03 anyway Jul 13 18:16:13 I still suspect a chroot without /tmp Jul 13 18:17:56 I didn't chroot though Jul 13 18:18:52 what did you do? Jul 13 18:19:07 Nothing Jul 13 18:19:09 ....... Jul 13 18:19:21 No I mean nothign that could have resulted in this Jul 13 18:19:27 so your computer just randomly spit out an error? =p Jul 13 18:19:40 All I've done today is worked on getting openembedded working Jul 13 18:20:02 Which pretty much means I just used commands that were in the wiki Jul 13 18:20:02 getting OpenEmbedded working can involve a chroot Jul 13 18:20:10 I never chrooted though Jul 13 18:20:16 maybe bitbake did Jul 13 18:20:20 maybe you should simply clean your /tmp Jul 13 18:20:36 < Semi Linux-Noob Jul 13 18:20:41 but then, that assumes you're using root Jul 13 18:20:44 which you shouldn't be Jul 13 18:21:42 he root=the great Administrator Jul 13 18:22:08 without Administrator you cant run programs on windows Jul 13 18:22:23 I think that's at least a bit exaggerative Jul 13 18:22:40 thats the sadly the truth Jul 13 18:23:14 koen, i've installed qemu from my distro and it's building my image Jul 13 18:23:17 koen, thanks Jul 13 18:23:46 Well how would I go about cleaning my /tmp Jul 13 18:25:00 Greg2: how exactly did you install qemu from your DISTRO? Jul 13 18:25:17 cd /tmp Jul 13 18:25:28 Kerwood, apt-get install qemu Jul 13 18:25:28 nad look what files are in there Jul 13 18:25:39 noting as far as I can see Jul 13 18:25:53 whats df telling? Jul 13 18:26:03 ups du Jul 13 18:26:27 Greg2: got it Jul 13 18:26:30 More then 76GB available Jul 13 18:27:54 mkdtemp: private socket dir: Permission denied Jul 13 18:28:05 thats from the xsession errors file Jul 13 18:28:21 ls -lhd /tmp Jul 13 18:28:54 Luke-Jr I have 2 computers fyi :P Jul 13 18:28:59 I noticed Jul 13 18:29:15 anyway, paste the result of that here Jul 13 18:29:57 drwxr-xr-x root root 4.0K 2006-07-13 14:47 /tmp Jul 13 18:30:01 * mrz80 is back (gone 01:59:01) Jul 13 18:30:07 ok, there's your problem Jul 13 18:30:45 sudo chmod 1777 /tmp Jul 13 18:31:04 you mean 777 or 1777 ? Jul 13 18:31:14 1777 Jul 13 18:31:17 what's o+t/0x1000? Jul 13 18:31:20 ok, just making sure Jul 13 18:31:45 stick bit Jul 13 18:31:52 o.O Jul 13 18:31:58 Yay all better now Jul 13 18:32:05 I think I know what happen Jul 13 18:32:11 I'm surprised your OS worked without /tmp Jul 13 18:32:23 if he were using root no problem at all Jul 13 18:32:25 you should probably reboot to fix other stuff Jul 13 18:32:30 s/were/was Jul 13 18:32:34 I think I accidently rm'd /tmp instead of tmp Jul 13 18:32:38 when in the zaurus folder Jul 13 18:32:39 :( Jul 13 18:32:41 thejapa: if he ran as root, there wouldn't be *any* problem =p Jul 13 18:33:02 right Jul 13 18:33:08 anyway... SonicvanaJr: reboot Jul 13 18:33:28 Well thank you guys very much for dealing with my incompetence Jul 13 18:33:29 SonicvanaJr: BTW, that's a good reason not to do OE stuff as root-- non-root can't rm /tmp ;) Jul 13 18:33:44 I'll start doing it as a user from now on Jul 13 18:33:46 people build as root nowadays? Jul 13 18:33:53 tisk :) Jul 13 18:34:00 No, I'm just an idiot : Jul 13 18:34:00 IIRC, doesn't the wiki specifically say not to use root? Jul 13 18:34:34 hehe. its a lesson most have to learn the hard way Jul 13 18:34:58 myself, i had a shell script that did rm -rf $FOO/, and ran the script with sudo, but had a typo in FOO... Jul 13 18:35:10 * Luke-Jr learned the easy way, after surviving building a Lfs as root Jul 13 18:35:15 hehe Jul 13 18:35:23 ouch Jul 13 18:36:27 alias rm='rm --preserve-root' Jul 13 18:36:27 :) Jul 13 18:36:46 Well hopefully everything will work out fine now, I was having some of the same problems as Greg2, but hopefully using his method everything will run smoothly Jul 13 18:36:56 ljr2@!tsurukikun! ~ $ rm -rf / Jul 13 18:36:57 rm: it is dangerous to operate recursively on `/' Jul 13 18:36:57 rm: use --no-preserve-root to override this failsafe Jul 13 18:38:00 I just put the qemu workaround in the OE wiki. hi kergoth. Jul 13 18:38:27 Kerwood: isn't it already in local.conf comments? Jul 13 18:39:39 "or installing Jul 13 18:39:39 # a working third-party qemu (e.g. provided by your distribution) and # adding qemu-native to ASSUME_PROVIDED Jul 13 18:39:39 " Jul 13 18:39:40 Luke-Jr: sure is. and it can't hurt to put it in the wiki, too Jul 13 18:39:51 I suppose Jul 13 18:40:14 * Luke-Jr wonders when his newlines went invisible Jul 13 18:40:20 X-Chat used to show them visually Jul 13 18:40:21 Maybe Greg2 woulda seen it Jul 13 18:46:24 Does the package need to be within the OE tree to have RDEPEND copied to DEPEND? Jul 13 18:46:48 or should the same apply to my bb files outside openembedded/packages Jul 13 18:46:49 ? Jul 13 18:57:12 JustinP: any idea? Jul 13 19:05:30 Luke-Jr: ...? any bbfile should have the same thing Jul 13 19:06:23 JustinP: if I copy gpe-image.bb from openembedded/packages/meta into ljr.pkgs/forks/gpe-image-ljr.bb, it loses the behaviour Jul 13 19:07:03 the only change I made was in packages included (replace figment with minisip) Jul 13 19:08:07 and I'm not sure where RDEPENDS is supposed to be used-- BUILD_ALL_DEPS is not set in *-image.bb :/ Jul 13 19:09:39 though -image do actually DEPEND so maybe that's ok Jul 13 19:10:53 the problem seems to be that task-bootstrap doesn't use either BUILD_ALL_DEPS nor DEPEND Jul 13 19:10:57 s Jul 13 19:11:56 re koen Jul 13 19:15:23 hey woglinde Jul 13 19:29:58 see you all later Jul 13 20:08:53 Geez how many locales are there Jul 13 20:09:19 all of them Jul 13 20:44:54 does openzaurus-unstable use unionfs? o.O Jul 13 20:48:52 03koen 07org.oe.dev * rd955e1e4... 10/conf/distro/openzaurus-unstable.conf: oz-unstable: prefer task-bootstrap over task-bootstrap-onionroot Jul 13 20:49:02 Luke-Jr: there you go :) Jul 13 20:49:15 =p Jul 13 20:49:29 so... not anymore Jul 13 20:50:06 * Luke-Jr now waits the 5 or so mins to pull that update Jul 13 20:50:19 heh Jul 13 20:50:21 onionroot Jul 13 20:52:34 ooops Jul 13 20:52:38 force of habit Jul 13 20:52:43 lol Jul 13 20:52:54 we have a 'student union' over here which we call the onion Jul 13 20:53:12 is there any way to recursively verify that dependencies have been built? Jul 13 20:53:52 I'm sure it's possible, but it probably isn't easy Jul 13 20:53:59 aww Jul 13 20:54:00 bitbake does it by checking the stamps Jul 13 20:55:00 except it doesn't go recursive Jul 13 20:55:13 so if task-bootstrap is done, it assumes its dependencies are Jul 13 20:55:29 or at least that's what I'm guessing my problem is Jul 13 20:56:47 bitbake -c clean task-bootstrap Jul 13 20:57:00 then it rebuilds everything...? Jul 13 20:58:56 * Luke-Jr spanks koen Jul 13 20:59:06 use monotone diff or some equivalent before you commit! Jul 13 20:59:26 you corrupted openzaurus-unstable with a 'h' Jul 13 21:00:54 (line 1 char 1 before the comment) Jul 13 21:05:31 Luke-Jr: that wont cause it to rebuild everything. -c clean cleans only the requested package, not its deps Jul 13 21:06:03 kergoth: so I would need to -c clean everything recursively? Jul 13 21:06:10 hm? Jul 13 21:06:19 the problem is that while some packages are built, their dependencies are gone Jul 13 21:10:59 I had to look 3 times before I could spot the 'h': http://www.openembedded.org/viewmtn/getdiff.py?id1=4110ff42248f7be7629efffc63b72cb06c918f2b&id2=d955e1e46bd54ae897b25fc9204f3dbdf1809039 Jul 13 21:11:15 O.o Jul 13 21:11:24 bitbake should unset CFLAGS etc Jul 13 21:11:40 OMG LOLZORO &*#&$*@ Jul 13 21:11:46 CFLAGS 1337 Jul 13 21:11:51 sorry Jul 13 21:12:02 I always get excited about cflags Jul 13 21:12:03 Luke-Jr: bitbake exports its own cflags. generally thats passed into the package buildsystem correctly. Jul 13 21:12:09 koen: your page is Internal Server Error Jul 13 21:12:11 * CosmicPenguin hands koen a towel Jul 13 21:12:19 kergoth: well, then it should unset the stuff it doesn't set Jul 13 21:12:21 like CPATH Jul 13 21:12:43 patches welcome. Jul 13 21:13:00 I imagine it'd be trivial for someone here to do, but I'd have to find where it goes Jul 13 21:13:29 best way to learn Jul 13 21:13:59 maybe Jul 13 21:14:02 03koen 07org.oe.dev * rb6a2a2fc... 10/conf/distro/openzaurus-unstable.conf: oz-unstable: fix typo Jul 13 21:14:33 * Luke-Jr adds the 'h' back and updates :) Jul 13 21:15:20 Luke-Jr: 'monotone update' runs a merger over modified files Jul 13 21:15:51 koen: yes, but I'm not sure if it can handle my already having hand-patched an update Jul 13 21:16:00 so I added it just in case Jul 13 21:43:59 erm Jul 13 21:44:16 why is glibmm trying to link with x86_64 staging libgobject? Jul 13 21:45:43 ld.alt/tmp/cross/arm-linux/lib -L/home/luke-jr/src/oe2005/build.alt/tmp/staging/arm-linux/lib -L/home/luke-jr/src/oe2005/build.alt/tmp/work/armv5te-linux/glibmm-2.8.4-r1/temp-staging/usr/lib -L/usr/lib -lsigc-2.0 -lgobject-2.0 -lgmodule-2.0 -ldl -lglib-2.0 -L/home/luke-jr/src/oe2005/build.alt/tmp/cross/lib/gcc/arm-linux/4.1.1 -L/home/luke-jr/src/oe2005/build.alt/tmp/cross/lib/gcc/arm-linux/4.1.1/../../../../arm-linux/lib -L/home/luke-jr/s Jul 13 21:45:43 rc/oe2005/build.alt/tmp/work/armv5te-linux/glibmm-2.8.4-r1/temp-staging/home/luke-jr/src/oe2005/build.alt/tmp/cross/lib/gcc/arm-linux/4.1.1/../../../../arm-linux/lib -lstdc++ -lm -lc -lgcc_s /home/luke-jr/src/oe2005/build.alt/tmp/cross/lib/gcc/arm-linux/4.1.1/crtendS.o /home/luke-jr/src/oe2005/build.alt/tmp/cross/lib/gcc/arm-linux/4.1.1/../../../../arm-linux/lib/crtn.o -march=armv5te -mtune=xscale -Wl,-rpath-link -Wl,/home/luke-jr/src/oe Jul 13 21:45:44 2005/build.alt/tmp/staging/arm-linux/lib -Wl,-O1 -Wl,--export-dynamic -Wl,--export-dynamic -Wl,-soname -Wl,libglibmm_generate_extra_defs-2.4.so.1 -o .libs/libglibmm_generate_extra_defs-2.4.so.1.0.20 Jul 13 21:45:49 hm, that was a bit long Jul 13 21:45:49 oops Jul 13 22:30:49 any vim around? Jul 13 22:30:57 I want to reindent sourcecode Jul 13 22:31:04 but it insists to add 8 spaces Jul 13 22:36:27 zecke: how do you want it indented? Jul 13 22:37:28 kergoth: 4 spaces Jul 13 22:37:40 kergoth: I have set et, set sts=4 set ts=4 Jul 13 22:37:45 so something is missing Jul 13 22:37:59 kergoth: is use the visual block to select and then '=' Jul 13 22:42:09 zecke: you didnt set 'et' aka 'expandtab' Jul 13 22:42:21 without that, itll indent with 4 space wide hard tabs Jul 13 22:42:33 :set sts=4 sw=4 et Jul 13 22:42:36 gg=G Jul 13 22:43:01 (= uses shiftwidth, not sts/ts, to determine indentation width, just as the automatic indentation does) Jul 13 22:44:09 sh sw was missing Jul 13 22:47:26 you normally dont need to mess with ts=, unless you want to show existing tabs differently Jul 13 22:49:21 kergoth: I find myself doing set ts=8 on TT code Jul 13 22:49:35 kergoth: I wouldn't be able to configure my vi to mix space and tab Jul 13 22:50:02 ljp: No trolling. How do you do that with vi and emacs? do you have a special init.el? Jul 13 22:54:00 zecke: what format do they use? Jul 13 22:54:26 kergoth: first indention is 4 spaces Jul 13 22:54:31 kergoth: second level is tabs Jul 13 22:54:36 oh, thats easy Jul 13 22:54:41 :set ts=8 sts=4 noet sw=4 Jul 13 22:55:22 hard tabs are still 8, indentation is 4, and theres no expandtab, which means itll use tabs for compression (one tab instead of 8 spaces at indent 2) Jul 13 22:56:50 ah Jul 13 23:22:44 Can somebody please push the stuff from bug 1117 before OE turns read-only? Jul 13 23:32:27 Laibsch: he should use require instead of include Jul 13 23:33:31 Laibsch: ping me tomorrow if it wasn't comitted Jul 13 23:33:53 OK Jul 13 23:34:00 I will update the patch then. Jul 13 23:35:44 You mean require uim.inc? Jul 13 23:35:47 zecke? Jul 13 23:36:05 yes Jul 13 23:44:43 zecke: i haven't seen mention of require before now. is the difference basically "require is include that will fail if it doesn't find it?" Jul 13 23:45:14 * zwelch goes to try it with his .bb files to see Jul 13 23:45:58 zwelch: right, see the 'great' bitbake manual Jul 13 23:46:12 i've read through it, i just must have missed it Jul 14 00:24:58 zecke: Made a patch to update tota's work and attached it to bug 1117. Jul 14 00:25:52 awesome, If I wouldn't be soo tired Jul 14 01:01:33 boy, freewrt sure is taking their time getting off the dime. Are they any warmer to the idea of OE as a preferred environment? Jul 14 01:03:14 Kerwood_: most of the OE devs, including me, would welcome it Jul 14 01:03:28 but it will be up to FreeWRT guys to do what they like Jul 14 01:04:22 i've not heard strong objections from the openwrt devs i talked to in the past.. what do you mean by warmer? has that changed? Jul 14 01:04:32 zecke: I took a peek at their freenode IRC channel and noticed some rather "crackerjack" people who used to be on #openwrt, a good sign Jul 14 01:05:10 kergoth: well, last I heard, there was still some ambivalence Jul 14 01:06:51 last i heard, there were no strong objections, just the amount of work involved to get the kernel and wrt specific packages over into oe was a barrier Jul 14 01:07:09 i'd expect that if we were to add those things ourselves, we'd find them a great deal more ready to switch Jul 14 01:07:13 * kergoth shrugs Jul 14 01:07:36 indeed Jul 14 01:08:14 unfortunately the wrt kernel is a nightmare Jul 14 01:08:25 drivers in the tree referencing shit outside the tree whose licensing is questionable.. Jul 14 01:09:47 true hackers. I guess. there is a sense of adventure with these guys, spending so much time reverse engineering stuff... Jul 14 01:17:15 I like OpenWrt **** BEGIN LOGGING AT Fri Jul 14 02:18:58 2006 **** ENDING LOGGING AT Fri Jul 14 02:59:57 2006