**** BEGIN LOGGING AT Thu Jul 14 23:59:57 2005 Jul 15 00:14:42 <__law__> is there any mplayer version for gpe? Jul 15 00:20:29 hello all ! Jul 15 00:21:13 i'm editing my local.conf for the first time. what are the prefered providers ? Jul 15 00:27:25 glrp... i don't like the "see the conf directory for blabla know to OpenEmbedded" line : there is nothing but my local.conf in this directory... Jul 15 00:28:51 ok... found ^^' (it was the /org.openembedded.dev/conf/...) Jul 15 00:45:07 morning Jul 15 00:46:11 yop yop XorA Jul 15 01:01:39 ok. first build, first troubles : "ImportError : No module named parse_py" Jul 15 01:01:47 any idea what i did wrong ? Jul 15 01:02:26 alan: did you take a tagged version of bitbake? Jul 15 01:02:36 tagged ? Jul 15 01:02:40 ^^' Jul 15 01:02:52 alan: in the svn under directory tagged Jul 15 01:03:00 excuse me, but i'm still in thze learnig phase... Jul 15 01:03:54 alan install the one in bitbake/tags/bitbake-1.3.1/ Jul 15 01:04:41 [09:53] howdy people Jul 15 01:04:41 [09:54] mickeyl: leaving Frankfurt for Interlarken today Jul 15 01:04:41 [09:56] sitting in a train station using the wireless Jul 15 01:05:32 mmmm... i did everything that was told to do in the GettingStarted page... Jul 15 01:05:45 XorA: where is the svn ? Jul 15 01:05:57 alan: what method did you use to get bitbake? Jul 15 01:06:17 monotone, i guess Jul 15 01:06:47 * alan studies philosophy, so please, be patient with him... Jul 15 01:06:56 alan: http://oe.handhelds.org/cgi-bin/moin.cgi/GettingStarted#head-edc470c3655ac7ebac7ae7928ea574819170fa1b Jul 15 01:07:37 yeah, that is what i did... Jul 15 01:07:53 alan: which bit, the fetching from subversion? Jul 15 01:08:17 the " svn co svn://svn.berlios.de/bitbake/trunk/bitbake" part. Jul 15 01:08:37 alan: ah, change that to svn co svn://svn.berlios.de/bitbake/ Jul 15 01:08:45 ok. Jul 15 01:08:52 alan: and under bitbake directory there is one called tags/bitbake-1.3.1 Jul 15 01:09:00 alan: install the version in that dir Jul 15 01:09:43 ok. may i say this sould be changed in the getting started pdoc ? Jul 15 01:10:18 alan: you may :-) Jul 15 01:10:53 mmm... should i uninstall the wrong version ? ./setup.py uninstall won't work... Jul 15 01:11:11 alan: just overwrite it Jul 15 01:11:18 ok. Jul 15 01:16:00 grrr... i wonder why there is a shepherd.conf in /conf/machine if i have a " Sorry, There is no dedicated configuration for the Sharp SL-C750 because of the similarities of the clamshell series. Use MACHINE = "c7x0"" message... Jul 15 01:17:49 ok... nano is building. Jul 15 01:17:56 * alan nervous Jul 15 01:18:46 hey XorA : thanks :) Jul 15 01:20:49 alan: well done Jul 15 01:21:52 XorA: first it's not done, yet, then, i didn't do anything ! that's amazing how easy it seems to be ! Jul 15 01:22:22 alan: bitbake and OE are amazing bits of development, and getting better al the time Jul 15 01:22:30 ~hail bitbake developers Jul 15 01:22:30 * ibot bows down to bitbake developers and chants, "I'M NOT WORTHY!!" Jul 15 01:22:46 hehe ^^ Jul 15 01:25:06 mmm... 5 minutes or so to build nano.... i wonder how long gpe-image build will take... Jul 15 01:26:09 alan: a few hours Jul 15 01:26:32 alan: but once the core glibc/gcc is built thats the main time Jul 15 01:27:07 ok. thanks. Jul 15 01:29:11 mmm... i det "openzaurus-3.5.4" as the distribution. i wonder if it's a good idea... Jul 15 01:29:32 alan: that one builds almost to a a comlete image, only minimo doesnt build Jul 15 01:29:39 alan: so yes, its the best idea Jul 15 01:29:45 ok. Jul 15 01:30:27 <__law__> can i only flash a new kernel without changing anything else on my c7x0 ? Jul 15 01:34:33 __law__: yes, just dont put initrd.bin on cf card Jul 15 01:35:04 XorA: i'm not sure, but i think it's quite difficult : you'll need to upgrade the kernel modules too and edit the init script.... maybe this could work, but it requires some work. Jul 15 01:35:38 <__law__> XorA, thanks Jul 15 01:35:51 re Jul 15 01:36:00 re zecke Jul 15 01:55:21 mmm why does compiling nano also compiles things like binutils ? The dependencies are automatically compiled ? Jul 15 01:57:20 alan: yes, thats the magic Jul 15 01:58:37 woooow... Jul 15 01:59:54 <__law__> juhuu ppp is working again!!! Jul 15 01:59:58 03jbowler 07org.openembedded.nslu2-linux * r55a9cfb8... 10/packages/ (5 files in 2 dirs): (log message trimmed) Jul 15 01:59:58 mpd does not build on osuclibc because libvorbis will not build, that Jul 15 01:59:58 happens because it attempts a static link with the -lm library in the Jul 15 01:59:58 wrong place. Jul 15 01:59:58 monotone split into two versions, one for the netsync-4 versions (up to Jul 15 01:59:59 and including 0.19) one for netsync-5 versions. Both may be installed Jul 15 02:00:01 on one NSLU2, allowing the system to handle both <=0.19 and >=0.20 Jul 15 02:00:13 <__law__> sending patch to bugs.oe.org Jul 15 02:06:15 <__law__> submit :-) Jul 15 02:08:40 hi Jul 15 02:15:03 XorA: lots of progs that are curently being built wikk be needed for next buit (gpe-image, for example). Will they be kept ? Jul 15 02:18:20 alan: yes they will Jul 15 02:18:52 houhouhou.... coooooooooooooool !!! Jul 15 02:55:19 03rwhitby 07org.openembedded.nslu2-linux * r068a291c... 10/packages/meta/unslung-packages.bb: Added package-index to unslung-packages Jul 15 02:57:05 FYI for other distros: we added a 'package-index' package, which rebuilds the Packages file in tmp/deploy/ipk after all other packages have been built. it normally only gets rebuild after the firmware image is built, but we build other packages for our feeds after the firmware, so we put 'package-index' as the last DEPENDS to make sure the Packages file is rebuilt. Jul 15 03:45:15 hey pb_ Jul 15 03:46:19 kergoth: ping Jul 15 03:46:22 "morning" Jul 15 03:47:01 good morning mickeyl Jul 15 03:49:34 zecke: good morning Jul 15 03:49:36 mickeyl: good morning Jul 15 04:20:37 tadaaaaa ! i just finished my very first build !! takes like forever on a duron 1.3ghz 512 mo ram, but it works !!! Jul 15 04:27:01 congrats Jul 15 04:31:30 ^^ well... the build is not absolutely finished, it's now packaging everything... Jul 15 04:33:10 congrats to all the developpers ! this is so easy to do, and i'm not a linux expert at all ! Jul 15 04:43:35 <_law_> if someone has droubles with todays gnu-config: i will post an patch to bugs.oe.org Jul 15 04:51:00 OT, is anyone here sitting at an amd64 running debian? Jul 15 05:07:04 koen|ewi: has one available Jul 15 05:12:31 afternoon Jul 15 05:19:17 XorA: nano and deps are compiled !!! Jul 15 05:19:23 * alan happy :) Jul 15 05:23:12 lt's bitbake gpe-image... Jul 15 05:24:11 just for curiosity : how do you edit the progs you want in your image ? Can this be done ? Jul 15 05:26:19 alan: gpe-image is just another bbfile Jul 15 05:26:32 alan: you can create an alan-image.bb and control the files to be installed Jul 15 05:26:56 wow, that is cool ! Jul 15 06:40:53 zecke: apropos the mail you sent to oe@ last night, is there any compelling reason not to relax the grammar so that quotes are optional in an rvalue? Jul 15 06:42:07 pb_: One could allow A = "abc" # really important comment Jul 15 06:43:46 IMHO it does not make sense to allow S = "abc" and PACKAGES += moo in one bbfile Jul 15 06:44:01 it looks incosistent but that is my opinion Jul 15 06:45:00 * mickey|lunch tends to agree Jul 15 06:46:07 pb_: I've the impression some people got to make believe PACKAGES and S can only take unquoted strings Jul 15 06:47:47 mickeyl: how do you like new zealand? http://www.hoult.org/~canllaith/south.html Jul 15 06:49:05 it does look a bit inconsistent, but I don't think that's a strong enough reason to reject such .bb files. I'd be happy with a warning. Jul 15 06:49:30 pb_: if it is worded offending I can agree :} Jul 15 06:49:47 we could claim lex suffers from the torette syndrom Jul 15 06:49:52 heh Jul 15 06:50:26 I don't think the wording is offensive. It just seems a bit unfriendly to break .bb files that previously worked, in absence of any obligation to do so. Jul 15 06:50:45 pb_: sorry the wording of the warning message Jul 15 06:51:01 "You got damn idiot your fucking bbfile is broken" :} Jul 15 06:51:47 pb_: My understanding was quotes were introduced later Jul 15 06:52:32 zecke: oh, heh, right. Jul 15 06:52:42 pb_: I always thought unquotes RVALUES were deprecated by inventing quoted ones Jul 15 06:53:00 I guess we could live with a warning like that. it'd be like having a little bit of kergoth in your own computer. Jul 15 06:53:22 pb_: so if I switch the parser it looked like a good idea to remove the support for unquoted rvalues Jul 15 06:53:28 (no technical reason involved) Jul 15 06:53:52 hehe Jul 15 06:53:58 i like that picture of a little kergoth Jul 15 06:54:12 * [g2] does miss kergoth Jul 15 06:55:05 pb_: but you can agree that for future bbfiles only quoted RVALUES should be allowed? Jul 15 06:55:12 ~botmail for mreimer: did you move files to arch/arm/plat-ipaq? I was expecting it to appear when I did cvs update -d. Jul 15 06:55:34 <[g2]> I'm sure he'd get a big kick out of us running bitbake natively on OpenSlug Jul 15 06:56:31 <[g2]> BTW do you know if any else does that on a cross (non-x86) target ? Jul 15 07:00:04 i don't think so. since zecke made it stop to suck it needed at least 1G of memory Jul 15 07:00:12 s/since/before/ Jul 15 07:01:30 <[g2]> mickeyl nod. I'd imagine the PPC folks would have some boxen with big memories/fast CPUs Jul 15 07:02:16 <[g2]> zecke, that mem reduction does ROCK btw... Jul 15 07:02:39 <[g2]> my FatSlug only has 64MB Jul 15 07:04:51 wait until he completed the C parser ;) Jul 15 07:04:57 i will build openzaurus on my slug then Jul 15 07:05:02 :D Jul 15 07:05:17 urghs Jul 15 07:05:21 ~lart monotone Jul 15 07:05:21 * ibot plops monotone into a giant vat of herring Jul 15 07:05:27 doesn't compile w/ gcc 4.0.1 Jul 15 07:05:40 * mickeyl kickst boost:: Jul 15 07:06:10 ah righto. it needs 1.32 Jul 15 07:06:44 I compile on ppc1000mhz, it's not fast, but better than using VIA EDEN/C3 Jul 15 07:06:55 zecke: yeah, I guess Jul 15 07:07:36 arf... my gpe-image build failed... Jul 15 07:07:37 NOTE: Task failed: /mnt/oe/build/tmp/work/module-init-tools-cross-3.1-r3/temp/log.do_compile.23082 Jul 15 07:07:37 NOTE: package module-init-tools-cross-3.1-r3: task do_compile: failed Jul 15 07:07:47 any idea ? Jul 15 07:08:00 not without error messages Jul 15 07:08:08 /usr/bin/ld: cannot find -lc Jul 15 07:08:12 aha Jul 15 07:08:20 oho ? Jul 15 07:08:20 guess you need to install libc-static-devel Jul 15 07:08:29 static, really? Jul 15 07:08:31 (wild guess here) Jul 15 07:08:33 what crazy distro is this? Jul 15 07:08:39 well Jul 15 07:08:50 i guess all distros have the shared one installed by default Jul 15 07:09:08 or is there an additional shared one for developing? Jul 15 07:09:12 micropal: You can trade your PPC against my PC Jul 15 07:09:27 mickeyl: there's some additional stuff you need for developing Jul 15 07:09:28 ok. thanks mickeyl. got to go for a walk. back in 30 mins. Jul 15 07:09:35 pb_: i see. Jul 15 07:09:36 never :) Jul 15 07:09:42 alan: so either libc-devel or libc-static-devel. Jul 15 07:09:50 ok. Jul 15 07:09:53 thanks Jul 15 07:10:23 by the way, i use mandriva 2005 (mandrake 10.2) Jul 15 07:10:31 good choice :D Jul 15 07:10:32 bbasap Jul 15 07:14:55 <[g2]> micropal, you bitbake on the ppc1000 ? Jul 15 07:18:09 on a mac mini Jul 15 07:18:24 (with debian-ppc) Jul 15 07:26:05 <[g2]> micropal, ahh.. We're bitbaking armeb cross on x86, and then installing on the armeb and bitbaking on that target. Jul 15 07:27:30 wow - armeb Jul 15 07:27:59 <[g2]> armv5teb Jul 15 07:28:35 <[g2]> which brings me to my current "Peter Principle" problem Jul 15 07:49:41 mickeyl: could you mail me your experiences with upgrading monotone to 0.20? Jul 15 07:49:52 I'll try to upgrade to 0.20 on monday/tuesday Jul 15 07:50:34 will do Jul 15 07:51:28 and can someone please add a fixed cvsdate version of gnu-config? Jul 15 07:54:04 koen|ewi: you got access to an amd64 running debian? Jul 15 07:54:16 yes Jul 15 07:54:20 running a 32bit kernel Jul 15 07:54:27 due to iptables breakage on 64bit Jul 15 07:55:13 koen|ewi: damn, wanted to know what the /lib64 entry was, I upgraded from ubuntu to debian and that dir was a casualty Jul 15 07:55:18 koen|ewi: did he upgrade? Jul 15 08:04:57 * koen|ewi -> breakfast & downtown Jul 15 08:05:00 later all Jul 15 08:25:50 niarf ! no libc-static-devel nor libc-devel availiable here... Jul 15 08:26:08 i still can't build GPE image for zaurus Jul 15 08:27:25 /usr/bin/ld: cannot find -lc Jul 15 08:28:00 alan: glibc-static-devel? Jul 15 08:28:17 alan: and glibc-devel Jul 15 08:28:50 looks like they are already installed... Jul 15 08:29:24 haaaaaaaa... glibc-static was missing !! Jul 15 08:29:29 XorA: thanks. Jul 15 08:29:41 alan: google Jul 15 08:32:06 XorA: yeah, i've already heard that sound... :) Jul 15 08:34:29 does the radius ppp plugin currently a.) compile and/or b.) work? on mips it doesn't even compile as "CC=gcc" is hardcoded in some Makefile. did anyone recently changed something on the package? otherwise i'm going to fix it Jul 15 08:37:18 tmbinc: i think i heard someone had fixed something about ppp here some hours ago and sent a mail to the buglist or something (i'm not aware of all this :( )... Jul 15 08:37:48 only thing i saw is http://bugs.openembedded.org/show_bug.cgi?id=149, and that's something kernel-related Jul 15 08:55:01 mickeyl: we've been having problems in NSLU2-land with /etc/init.d/populate-var.sh because we change how /var is mounted according to how the NSLU2 is configured (diskless, diskful, NFS etc). Jul 15 08:55:40 So we need to dynamically configure the /var to have some sub-dirs as links to tmpfs and some one disk Jul 15 08:56:15 We're experimenting with a solution which makes populate-var.sh a normal startup script reading configuration information. Jul 15 08:57:06 So we can change the config as required... Anyway it's a core OE thing, so we're interested in whether anyone else is working on this and how to coordinate any work. Jul 15 08:58:19 I've already done one set of changes to the initscripts in an openslug specific way (packages/initscripts/initscripts-openslug), but those changes really aren't openslug specific, to that's probably a bad approach. Jul 15 08:58:48 jbowler: the general principle sounds good to me. Could you please write a mail to the list to see if anyone objects? Jul 15 08:58:53 On the other hand, hacking initscripts and only testing on NSLU2 is obviously bad. Jul 15 08:59:25 *nod* Jul 15 08:59:34 i need to leave for a couple of hours Jul 15 08:59:34 bbl Jul 15 09:00:25 One thing we could do here is to make new releases of base-files and initscripts (since these don't have a external SRC_URI we can rev PV), then DEFAULT_PREFERENCES=-1 until they are tested Jul 15 09:01:09 That's a method to allow changes in nslu2-linux safely, but it isn't a protocol to allow on how we should go about changing core OE stuff like this - obviously other people may be working on these packages outside nslu2-linux. Jul 15 09:03:22 righto. initscripts and populating dev has always been pretty ad-hoc yet. thoughts to make it more versatile are really welcome and I'd like to change these things in a central way. that's why I'd like to discuss it on the list Jul 15 09:03:23 * mickey|bbl runs Jul 15 09:04:06 righto. initscripts and populating dev has always been pretty ad-hoc yet. thoughts to make it more versatile are really welcome and I'd like to change these things in a central way. that's why I'd like to discuss it on the list Jul 15 09:04:18 problem is not every device developer may want to use the same dev/init setup as every otehr device Jul 15 09:04:48 so why not keep them machine specific? Jul 15 09:05:06 mickeyl|bbl: I'll copy the above to the mailing list (that's what you meant by 'list' I believe.) Jul 15 09:05:25 you wouldnt generally have the same init setup anyway for a POS device as you would for a kiosk Jul 15 09:05:34 emte: the problem is there is a lot of stuff there and initscripts just puts in a one-size-fits-al version Jul 15 09:05:50 yeah i know Jul 15 09:06:08 but for the most part OE is a generic base Jul 15 09:06:16 initscripts-openslug is a real hack - it actually pulls in initscripts, checks it hasn't changed, then hacks it in place... Jul 15 09:06:50 yeah that sounds liek my case in point Jul 15 09:07:26 Hum, I should probably be saying this on the ML, not here - solutions are fairly obvious, and nslu2-linux is motivated to fix it... Jul 15 09:07:48 fork an openslug init package would be my thoughts Jul 15 09:08:12 i'd advise making the initscripts packaging more granular, so distros can more easily pick & choose components they need. Jul 15 09:09:02 thats a hard call tho ... there are what .... 8 init styles nm the scripts themselves Jul 15 09:12:03 anyone in here have a Zaurus C7/8x0 running GPE with a CE-RH1 audio remote? Jul 15 09:15:45 pb_: glib-2.0-native_2.6.2 currently doesn't work with gcc4, is there anything against upgrading to 2.6.5? Jul 15 09:16:30 no, go ahead Jul 15 09:16:42 ok Jul 15 09:30:59 03tmbinc 07org.openembedded.dev * r9674a0cd... 10/packages/glib-2.0/glib-2.0-native_2.6.5.bb: add glib-2.0-native 2.6.5 in order to fix gcc4 compile problems Jul 15 09:34:07 damn I killed my work machine yesterday :} Jul 15 09:34:20 mount: libblkid.so not found Jul 15 09:40:47 hm, I get xf86Bus.c:3128: internal compiler error: Segmentation fault, when compiling xfce-image Jul 15 09:42:35 that's bad.. Jul 15 09:45:21 It ain't good Jul 15 09:46:51 it said I should send a bug report to the arm-linux developers.. Jul 15 09:47:02 ibot: test Jul 15 09:47:02 Test Passed! Jul 15 09:49:54 is ibot made of white plastic with a glowing apple in the cover? :) Jul 15 09:59:44 what dependency should I put for something which requires Xlib? "xlib" and "Xlib" didn't work Jul 15 10:00:54 note: this programs compiles and works fine I'm just trying to set up the dependencies correclty. Jul 15 10:04:25 * JustinP tries diet-x11... Jul 15 10:05:31 "x11" Jul 15 10:05:55 ok, I'll try that Jul 15 10:11:56 you know what would be great in bitbake ? an estimated time remaining indicator when compiling... or something like a "number of bb to compile left"... Jul 15 10:12:25 an eta is practically impossible Jul 15 10:12:33 number of bbs would be a nice addition Jul 15 10:13:21 cept it doesnt count that way Jul 15 10:14:11 I dont recall if it computes the entire dependency list before starting or not :) Jul 15 10:14:29 unless something was changed i dont think it does Jul 15 10:14:51 that part of the code just scares me Jul 15 10:14:55 wasnt that the whole think about dependancies not really working? Jul 15 10:15:01 thing* Jul 15 10:15:19 I didn know dependencies werent working :) Jul 15 10:15:42 Thats kind of like saying "I drove to the store, but I don't know if my engine was working" Jul 15 10:15:45 well if you had two packages with the same deps they would both be included Jul 15 10:16:14 even though they only get built once you had two sets listed Jul 15 10:16:16 CosmicPenguin: well frankly I havent touched bitbake in a month or two Jul 15 10:16:26 so I dont have a clue how fucked up it has gotten Jul 15 10:17:09 i think that was one of the issues to be solved in bb2 Jul 15 10:17:17 by using a dtabase Jul 15 10:17:45 i think that was one of the issues to be solved in bb2 Jul 15 10:17:48 by using a database Jul 15 10:22:06 the world graph is generated on parsing Jul 15 10:22:06 treke: its less fucked up then its ever been, actually Jul 15 10:26:26 * zecke tries to decide to fix his debian installation or reinstally ubuntu Jul 15 10:27:08 debian is more developer friendly Jul 15 10:27:37 gentoo is even more developer friendly Jul 15 10:27:46 no, i dissagree Jul 15 10:28:43 not for non-pc stuff at any rate Jul 15 10:29:07 why do you say that? Jul 15 10:29:49 Gentoo is definately not developer friendly Jul 15 10:30:03 how is debian more developer friendly? Jul 15 10:30:12 a hassle and waste of time trying to secure the system, get needed things to compile, change stupid default settings, etc Jul 15 10:30:16 * JustinP is asking because he's curious, not to argue Jul 15 10:30:27 emte: stop it Jul 15 10:30:33 ? Jul 15 10:30:40 he asked Jul 15 10:30:47 emte: it works flawlessly for me. Jul 15 10:31:08 not another my distro is bigger than yours Jul 15 10:32:30 ::sigh:: never mind, this conversation is pointless Jul 15 10:32:49 yeah Jul 15 10:32:54 that more or less describes it Jul 15 10:33:06 ack Jul 15 10:33:48 lets see if i can fry my system buy putting a third soundcard in it ... Jul 15 10:34:11 maybe i should use a fuse ... Jul 15 10:36:14 wonder if i have a transformer around Jul 15 10:39:13 * JustinP stops himself from making a Hasbro reference Jul 15 10:40:17 dont really trust making my soundcard into an oscilloscope too much with out some type to input protection Jul 15 11:02:43 emte: you can check if there is really 50hz on mains with your new 'oszilloscope'.. Jul 15 11:27:02 wow unbelivable xkb is broken in breezy again Jul 15 11:36:31 * zecke kicks his computers and goes to bed Jul 15 11:40:45 good evening ;) Jul 15 11:40:49 hi treke Jul 15 11:45:54 * zecke kicks his computers even more and goes to bed Jul 15 11:46:26 good night zecke Jul 15 12:14:30 what does "${TARGET_PREFIX}strip ${D}${base_libdir}/libgcc_s.so.*" mean? (from gcc-package-cross.inc).. isn't ${base_libdir} my local path into $TMPDIR/cross/lib? why is it appended to ${D}? i'm confused. Jul 15 12:14:55 no, base_libdir is not that. its /lib. Jul 15 12:15:22 its stripping the one in the ipk its creating Jul 15 12:15:28 has nothing to do with tmp/cross Jul 15 12:16:31 $ bbread ../org.openembedded.dev/packages/gcc/gcc-cross_3.4.4.bb | grep "base_libdir" Jul 15 12:16:31 # base_libdir=${base_prefix}/lib Jul 15 12:16:31 export base_libdir="/home/dump/tmb/oe-mono/build/tmp/cross/lib" Jul 15 12:16:35 what's wrong then here? Jul 15 12:17:31 clearly base_prefix isnt correct. Jul 15 12:17:51 I imagine "inherit cross" clobbers it. Jul 15 12:18:26 I think gcc-package-cross probably does need to spell out "/lib" explicitly in that case. Jul 15 12:18:35 the whole point of base_prefix/base_libdir was to back up prefix/libdir for safe keeping so cross.bbclass doesnt blow it away Jul 15 12:18:48 base_prefix comes from prefix, which in turn comes from CROSS_DIR, which is CROSS_DIR=${TMPDIR}/cross Jul 15 12:18:50 a simple base_prefix := ${prefix} before the inherit would do the job, and iirc thats exactly what it deos Jul 15 12:19:01 ah, right, I see Jul 15 12:19:41 hi kergoth, have u find a work ? Jul 15 12:19:48 no :\ Jul 15 12:20:02 plenty of jobs, just not many _here_ Jul 15 12:20:03 heh Jul 15 12:20:08 * kergoth doesnt want to relocate at this point Jul 15 12:20:43 kergoth if u find a decent (not good) that can sponsor a fireign call me ;) :) :) Jul 15 12:20:51 hehe Jul 15 12:22:27 seem the situations isn't different in may country Jul 15 12:33:05 hrm Jul 15 12:33:37 is there any compelling reason to have a floating gnu-config-native_cvs instead of one with a fixed cvsdate? Jul 15 12:33:50 no Jul 15 12:33:57 yes, to blow up the build process every now and then ;) Jul 15 12:35:54 * reenoo_ fixes Jul 15 12:36:02 ~bitchslap lilo Jul 15 12:36:02 * ibot beats the sh*t out of lilo Jul 15 12:36:24 reenoo_: what irc client do you use? Jul 15 12:36:26 03tmbinc 07org.openembedded.dreambox * r52ea131a... 10/packages/sysvinit/sysvinit/opendreambox/inittab: add serial console to runlevel 5 Jul 15 12:36:29 03tmbinc 07org.openembedded.dreambox * r62ee48aa... 10/packages/tuxbox/tuxbox-libtuxtxt.bb: update libtuxtxt, set package version Jul 15 12:36:34 03tmbinc 07org.openembedded.dreambox * rfc68b6c2... 10/packages/tuxbox/tuxbox-plugins.bb: update tuxbox plugins Jul 15 12:36:42 zecke: BX Jul 15 12:38:14 hi zecke Jul 15 12:41:38 03tmbinc 07org.openembedded.dreambox * r426c2ea0... 10/packages/enigma/enigma_cvs.bb: add libtuxtxt dependency to enigma Jul 15 12:44:31 tmbinc: good job.. got that dreambo-image compiled successfully after started from scratch. oh and if i may ask one thing.. do u know anything bout making "MOD" images? Jul 15 12:47:54 <[g2]> kergoth hey! Jul 15 13:00:03 I've set up another package to be built in a bb file (using PACKAGES) and am trying to set up the main package to depend on the extra one (a -lib package) and make the -lib depend on certain things. I set up what I thought would work but it's putting all of the -lib deps on the main package. The lines I used are: Jul 15 13:00:09 PACKAGES += "${PN}-lib" Jul 15 13:00:24 DEPENDS_${PN} = "lineakd-lib" Jul 15 13:00:28 DEPENDS_${PN}-lib = "x11 xext" Jul 15 13:00:40 DEPENDS doesn't make any sense for subpackages. Jul 15 13:00:45 ok Jul 15 13:00:46 You probably want to be poking at RDEPENDS. Jul 15 13:00:51 ok Jul 15 13:01:49 what's the difference exactly? Jul 15 13:02:06 DEPENDS is build time dependencies (i.e. between .bb files) Jul 15 13:02:23 RDEPENDS is run time dependencies (i.e. between output packages) Jul 15 13:02:25 ok and R is runtime Jul 15 13:02:28 I see Jul 15 13:02:42 so I should put these in both then? Jul 15 13:03:10 it shouldn't be necessary to specify any libraries in RDEPENDS; oe will sort that out for itself. Jul 15 13:03:22 Just put all the packages that you need for building in DEPENDS. Jul 15 13:03:42 except it's not for the -lib dependency since it's int he same bb file (I assume...) Or perhaps I should remove them all and see what happens... Jul 15 13:04:13 trying without any depends for the -lib... Jul 15 13:04:47 * reenoo_ eyes at CIA-4 Jul 15 13:05:00 ah, found the manual entry for the bb vars Jul 15 13:06:41 gnu-config: switch to a fixed CVSDATE (20050701 for now). Jul 15 13:06:50 03rw 07org.openembedded.dev * rd2ba131f... 10/packages/gnu-config/ (4 files): gnu-config: switch to a fixed CVSDATE (20050701 for now). Jul 15 13:06:57 heh Jul 15 13:07:14 bbl Jul 15 13:07:30 renoo_: it works with 20050713 Jul 15 13:08:34 It ought to cope with shared library dependencies between subpackages in the same .bb file. If that isn't happening, it's definitely a bug. Jul 15 13:08:58 reenoo_kitchen: very good Jul 15 13:10:42 reenoo_kitchen: does "mt sync" happen in reasonable time? Jul 15 13:11:52 it's finding the dependency on the -lib for the main package but the -lib package isn't getting any X dependencies (although it's getting a few others). Perhaps I'm incorrect in thinking that the lib depends on the same things. Jul 15 13:18:11 If OE is managing to figure out any RDEPENDS at all for the -lib package, it's probably getting them right. Jul 15 13:20:38 ok, I'll leave them then Jul 15 13:21:03 JustinP: you're not trying to depend on plugins or anything like that are you? Jul 15 13:22:11 pb_: Do you think it is common that bitbake should fetch the current version of a cvs/svn/* tree? Jul 15 13:22:50 pb_: for example if one develops/compiles using bitbake I think it is normal to fetch always the latest version Jul 15 13:23:00 CosmicPenguin: no, I'm trying to break out the libs into their own package. It *seems* like the package maintainer meant for them to be available to other programs (although I know of none which use them) Jul 15 13:23:39 JustinP: ok - just making sure Jul 15 13:24:07 the bb file will be ready momentarily Jul 15 13:24:26 just checking ti into my local copy Jul 15 13:25:38 and testing the ipks ;-) Jul 15 13:33:43 zecke: for small changes like the one I just pushed, yes Jul 15 13:40:20 zecke: I don't think that is a terribly common use case, no. Jul 15 13:40:59 zecke: interactive development using bitbake with a floating CVSDATE is a nuisance, because you can only compile once per day unless you manually clean the sources directory. Jul 15 13:42:36 Even then, it wll only pull to midnight, won't it? Jul 15 13:43:02 CosmicPenguin: you could override CVSDATE to use the current full date and time Jul 15 13:43:32 Oh, thats right Jul 15 13:44:53 pb_: yeah floating != CVSDATE Jul 15 13:51:38 CosmicPenguin: you can set CVSDATE to tomorrow to make it fetch the code from "right now". Jul 15 13:51:43 but, of course, that only works once. Jul 15 13:52:07 CVSDATE_package = "now" Jul 15 13:52:44 or that, but the same applies. Jul 15 13:52:50 hmm, right Jul 15 13:53:25 ~lart cellphones Jul 15 13:53:26 * ibot takes out a seltzer bottle and sprays cellphones in the face. You know, one of those old-school seltzer bottles clowns have? Yeah those. Anyway, consider yourself spritzed Jul 15 13:54:17 stupid thing went amok in my pocket again and called someone who's currently in the US Jul 15 13:54:44 reenoo_: no you wull create a taball once Jul 15 13:54:47 reenoo_: name it now... Jul 15 13:55:10 pb_: SRC_URI = "cvs://...;always_current=1" Jul 15 13:55:25 reenoo_: doh, I hate it when that happens Jul 15 13:55:37 zecke: yah, something like that would be nice. Jul 15 13:55:44 zecke: yeah, pb_ already pointed that out and I agreed... Jul 15 13:55:59 or you could special-case a CVSDATE of "now" in the fetcher, which would make reenoo's idea work. Jul 15 14:18:42 * chouimat|ibook is away: pub Jul 15 15:04:07 ploum Jul 15 15:07:14 is there anyone here ? Jul 15 15:07:19 no Jul 15 15:07:41 hey kergoth Jul 15 15:07:49 hey Jul 15 15:07:59 kergoth: why this ? "Connexion vers www.oesources.org[192.216.230.225]:80...failed: Connection timed out." Jul 15 15:12:49 dam it !!!! "--00:14:46-- http://www.oesources.org/source/current/gpe.base.esound_cvs.handhelds.org__20020817.tar.gz Jul 15 15:12:49 (essai: 4) => `/mnt/oe/sources/gpe.base.esound_cvs.handhelds.org__20020817.tar.gz' Jul 15 15:12:49 " Jul 15 15:13:54 just press ctrl+c, it should do a cvs checkout then Jul 15 15:20:19 reenoo_: thanks for tying down gnu-config Jul 15 15:21:09 koen|ewi: np. just trying to get a working build env again :) Jul 15 15:33:42 is there any way to see the dependency graph for something? I'd like to know why ncurses and ncurses-native are being compiled for gmpc (I think it's the gnome-vfs dependency, but I don't know why that would need it either) Jul 15 15:35:39 JustinP: you could use the shell Jul 15 15:35:49 JustinP: but there is no fine grained dependency tracking Jul 15 15:37:10 anyone want to make a gpraphviz plugin for OE? Jul 15 15:37:17 -p somewhere Jul 15 15:38:22 -p? Jul 15 15:38:36 gpraphviz vs graphviz Jul 15 15:38:49 oh Jul 15 15:39:12 plugin? Jul 15 15:39:24 it has been a very long long day. Jul 15 15:39:32 it it time to head home Jul 15 15:39:33 bye bye Jul 15 15:39:33 something like bitdoc, but for graph generation Jul 15 15:39:38 bye france_ Jul 15 15:39:54 koen|ewi: bye Jul 15 15:39:54 graphs of package depends? Jul 15 15:40:01 JustinP: "bitbake --dry-run --verbose" should yield that information, though not in the most convenient form. Jul 15 15:40:12 I have *no* idea how to do this in the shell ;-) Jul 15 15:40:44 oh well it's nearly done I think Jul 15 15:43:45 aw hell....I just figured out why I was having problems yesterday with configure / compile and such doing nothing....I forgot to inherit autotools Jul 15 15:44:02 * JustinP smacks himself Jul 15 15:44:33 That'll do it Jul 15 15:44:38 mm, I have no idea how to do that in the shell either. Jul 15 15:44:47 hehe Jul 15 15:45:06 I'll try the dry-run after my compile is done. Should be interesting to see where it comes in (if it works) Jul 15 15:45:33 dry-run doesn't list already met deps any more IIRC Jul 15 15:45:39 oh...:-( Jul 15 15:45:41 oh well Jul 15 15:46:01 oh, doesn't it? that's a bit of a shame. Jul 15 15:46:37 something like emerge --emptytree would be nice Jul 15 15:47:00 ah emerge --pretend would also be lovely so that we can see the packages to compile Jul 15 15:47:07 and --tree Jul 15 15:47:14 heh, sorry, I'm so used to portage Jul 15 15:47:22 You should read the man pages Jul 15 15:47:25 oe is working wonders, though Jul 15 15:47:30 man pages for which? Jul 15 15:47:37 bitbake Jul 15 15:48:15 bitbake has a man page? eleet. Jul 15 15:48:18 they don't seem to be installed.... Jul 15 15:48:33 heh Jul 15 15:48:35 I "emerged" bitbake Jul 15 15:48:45 figure of speech Jul 15 15:48:54 surprisingly the man pages weren't installed. Gentoo is usually good about that Jul 15 15:48:56 ah Jul 15 15:48:59 not so eleet Jul 15 15:49:02 It should have a man page though Jul 15 15:49:07 Can Jul 15 15:49:13 't be worse then half the man pages out there Jul 15 15:49:17 like quilt Jul 15 15:49:18 heh Jul 15 15:49:22 quilt's man page has gotten worse Jul 15 15:50:09 Anyway, --dry-run is bitbake's answer to --pretend Jul 15 15:50:40 ok. --emptytree would be the next logical step then Jul 15 15:50:58 basically is just has to ignore the stamps Jul 15 15:51:03 It used to do --emptytree, but that got fixed Jul 15 15:51:08 ah Jul 15 15:51:19 well, both are useful Jul 15 15:51:25 There's a way around it though, and I'll be damned if I remember what it is Jul 15 15:51:28 kergoth knows Jul 15 15:52:40 mv tmp/stamps tmp/stamps.bak; bitbake --dry-run gmpc; mv tmp/stamps.bak tmp/stamps Jul 15 15:52:42 perhaps? Jul 15 15:53:32 thats one way Jul 15 15:53:56 * JustinP waits further for samba to compile.... Jul 15 16:06:46 FYI: monotone.vanille.de now running on 0.20 - please test Jul 15 16:06:47 g'night Jul 15 16:07:41 hmmm? does this mean I need to update my client in order to pull? Jul 15 16:07:52 is there anything else I should know? Jul 15 16:11:00 not much Jul 15 16:11:09 you'll need 0.20 to pull Jul 15 16:15:11 ok Jul 15 16:15:14 I'll upgrade then Jul 15 16:15:29 (even though it's marked "unstable") Jul 15 16:16:15 iirc, the task executing code checks for a 'force' attribute which ignores stamps. maybe you can use that in combination with dry-run? Jul 15 16:19:48 ewi546.ewi.utwente.nl is also running monotone 0.20 now Jul 15 16:21:47 oops. forgot to read the release notes, apparently my update alias needs changing. Jul 15 16:26:46 the 'collections' are globs now Jul 15 16:27:02 so org.openembedded -> 'org.openembedded.*' Jul 15 16:27:18 hmmm....one of my bb files is including the .desktop file automatically when I put it in the SRC_URI and the other isn't Jul 15 16:27:23 even with that it says 'no branch pattern given and no default pattern set' Jul 15 16:27:50 it gets "unpacked" but not packaged Jul 15 16:27:59 perhaps it is interpreting my pattern as a server address Jul 15 16:30:04 any idea why a .desktop file might now get included automatically? Jul 15 16:30:13 and do I need org.openembedded.* or just org.openembedded.dev* ? Jul 15 16:30:41 'org.openembedded.*' to be safe Jul 15 16:30:56 koen|ewi: nslu2-linux will start server migration now too .... Jul 15 16:32:14 rwhitby: I was just about to ask :) Jul 15 16:32:26 how does mtn perform on a turboslug? Jul 15 16:32:31 Wow - I like that tar archive Jul 15 16:32:37 does it make a tar on the spot? Jul 15 16:32:41 it does Jul 15 16:32:50 Does that not kill your server? Jul 15 16:32:57 dual opteron :) Jul 15 16:33:02 good man Jul 15 16:33:38 'budget need to be spent completely' university purchase Jul 15 16:33:54 koen|ewi: I think it took about 18 hours for an initial pull on a TurboSlug Jul 15 16:34:01 so it has unlimited bandwidth too Jul 15 16:34:25 ok, that will be a port forward too :) Jul 15 16:35:49 rwhitby: any experience with sratchbox on a slug? Jul 15 16:37:55 koen|ewi: not that I know of Jul 15 16:38:21 koen|ewi: any gotchas on the mt upgrade - do we just upgrade the binary, and point it at the database, or is there a migration step required? Jul 15 16:43:27 ~lart boost Jul 15 16:43:27 * ibot whips out a hot clue gun and makes sure that boost is stuck to the floor Jul 15 16:43:43 that build system really scares me. heh Jul 15 16:44:53 damn it! Jul 15 16:45:33 boost scares me in generla Jul 15 16:45:38 general even Jul 15 16:50:51 rwhitby: the only change I made was org.openembedded -> 'org.openembedded.*' for mt serve Jul 15 16:52:00 I was pleasantly surprised with the easy of upgrading Jul 15 16:52:06 1) install deb Jul 15 16:52:24 2) change collection param Jul 15 16:52:26 3) done Jul 15 16:52:27 :) Jul 15 17:02:51 ok, nslu2-linux server is now updated to 0.20 too Jul 15 17:03:17 we're using "org.openembedded.* org.nslu2-linux.*" Jul 15 17:04:15 are spaces allowed? Jul 15 17:04:20 I though it was a glob Jul 15 17:04:43 "org.foo.*:org.bla.*" Jul 15 17:04:56 or am I confising it with BBFILES? Jul 15 17:05:16 seemed to work on the pull command line Jul 15 17:09:49 anyone? If I want to include a .desktop file when installing and it's not being automatically done, how should I include it? I've tried pretty much everything I can think of, adding tasks left and right, and nothing does it. Jul 15 17:10:13 JustinP: you mean you want the .desktop packaged? Jul 15 17:10:21 well it needs to be in the ipk.... Jul 15 17:10:29 so add it to the FILES_ Jul 15 17:10:30 it's working automatically for one package but not another Jul 15 17:11:15 install -m 644 bla.desktop ${D}/usr/share/applications/ Jul 15 17:11:34 FILES_${PN} += "usr/share/applications/" Jul 15 17:11:42 that's the quickest way I know off Jul 15 17:11:51 where do you put the "install" Jul 15 17:12:21 do_install() { Jul 15 17:12:25 Jul 15 17:12:26 } Jul 15 17:12:48 what are you inheriting in the .bb? Jul 15 17:12:58 autotools Jul 15 17:13:10 that's all the other one is inheriting and it gets the .desktop.... Jul 15 17:13:21 Ok - you'll need to do do_install_append() instead of do_install() Jul 15 17:13:25 this package puts stuff in usr/share, though Jul 15 17:14:18 ok, trying that Jul 15 17:14:20 thanks for helping Jul 15 17:22:15 'night all Jul 15 17:43:42 03rw 07org.openembedded.dev * rea0f0d66... 10/packages/gpe-mileage/gpe-mileage_0.1.bb: gpe-mileage now resides in GPE CVS and tarballs can be fetched from the usual location. Jul 15 17:43:46 03rw 07org.openembedded.dev * r52308134... 10/packages/gpe-life/gpe-life_0.1.bb: Add gpe-life, an implementation of Conway's Game of Life for GPE. Jul 15 17:46:20 cool, monotone works pretty good on OSX Jul 15 17:50:43 * chouimat|ibook is back. Jul 15 17:52:46 D'oh Jul 15 17:53:00 monotone 0.20 requires a small change to the monotonerc too **** ENDING LOGGING AT Fri Jul 15 18:44:43 2005 **** BEGIN LOGGING AT Fri Jul 15 18:44:52 2005 Jul 15 18:46:37 koen|ewi: ? Jul 15 19:23:25 luke-jr_: what's your question? Jul 15 19:23:59 koen|ewi: can you send me the new monotonerc? Jul 15 19:24:26 rwhitby-away: remove the 'collection' arguments from netsync_foo_premitted Jul 15 19:24:46 rwhitby-away: change it to: function get_netsync_write_permitted (identity) Jul 15 19:25:37 rwhitby-away: http://pastebin.com/314286 Jul 15 19:28:11 koen|ewi: thx - so we can't restrict access based on branch now? Jul 15 19:28:43 according to the monotone people you can, but I haven't looked at it Jul 15 19:34:03 hey bluelightning Jul 15 19:34:31 hi koen|ewi Jul 15 19:34:33 how goes it? Jul 15 19:34:35 hello everyone else too Jul 15 19:37:58 bluelightning: pretty good, enjoying my next to last holiday in the US right now Jul 15 19:38:46 s/holiday/day/ Jul 15 19:45:44 * chouimat|ibook is away: shower Jul 15 19:57:20 mickey|zzZZzz: you'll have to change --depth into --last in monotone.py to get viewmtn working again Jul 15 20:02:49 * chouimat|ibook is back. Jul 15 20:49:15 rwhitby-away: I just ordered a slug + usb 100Mbit nic :) Jul 15 20:54:47 * koen|ewi goes to bed now Jul 15 20:54:50 'night all Jul 15 21:05:33 night koen|ewi Jul 15 22:18:31 monotone 0.20 seems to be working fine for me **** ENDING LOGGING AT Fri Jul 15 23:59:56 2005