**** BEGIN LOGGING AT Thu Mar 23 10:59:58 2006 Mar 23 11:01:03 zecke: the gnome people use xml-rpc to talk to their bugzilla, maybe we can use their patches? Mar 23 11:01:23 using which client? Mar 23 11:02:09 zecke: bug-buddy Mar 23 11:03:06 koen: only for reporting issues? Mar 23 11:03:48 zecke: afaik, yes Mar 23 11:04:04 koen: KBugBuster allows you to manage your bugs with a nice GUI Mar 23 11:04:15 koen: BTW: KDE looks so much nicer with the Apple Fonts Mar 23 11:05:18 zecke: lets write lua hooks to manage bugs in monotone :) Mar 23 11:09:39 Yay. LED and RTC subsystems marked as "Will merge." by akpm :) Mar 23 11:09:58 cool Mar 23 11:11:43 Nice :) Mar 23 11:25:06 RP, : XLDFLAGS=-L/usr/X11R6/lib Mar 23 11:25:06 in apmd-3.2.2-r7/apmd-3.2.2.orig/Makefile Mar 23 11:25:33 that could come from here ? Mar 23 11:26:32 Genesis: How does it get into matchbox-dekstop from there? Mar 23 11:28:06 it's not relevant but it's the only one makefile that contain -L/usr/X11R6/lib in the entire project Mar 23 11:28:25 perharps i should searsh differently Mar 23 11:30:02 i go to eat , i 'll be back soon Mar 23 11:30:18 don't hesitate to ask me for doing something . Mar 23 11:30:21 Need to get a CFLAGS_append to work depending on kernel version (2.4 or 2.6) . For example for a 2.6 kernel CFLAGS_append =" -D__Linux26__" Mar 23 11:30:31 hwo should i do it ? Mar 23 11:30:51 Ifaistos: check KERNEL_MAJOR_VERSION Mar 23 11:39:02 Genesis: You might be better off searching the .pc files Mar 23 11:39:33 Ifaistos: There is an example somewhere of that - do a search for the variable hrw mentioned Mar 23 11:42:51 Ifaistos: CFLAGS_append = '${@base_conditional("KERNEL_MAJOR_VERSION", "2.6", " -D__Linux26__ ", " _D__Linux24__ ")}" Mar 23 11:43:11 Ifaistos: only need to check does it is KERNEL_MAJOR_VERSION or KERNEL_VERSION_MAJOR Mar 23 11:50:22 * chouimat is back. Mar 23 11:50:35 morning Mar 23 11:50:46 hey chouimat Mar 23 11:57:25 Thanks guys ! This did the trick Mar 23 11:57:26 CFLAGS_append = '${@base_conditional("KERNEL_MAJOR_VERSION", "2.6", " -D__Linux26__ ", " -D__Linux24__ ",d)}' Mar 23 11:58:56 its for a new package cpusage.bb that returns the kernel cpu usage numbers and averages Mar 23 11:59:13 Ifaistos: base_conditional is usefull thing which is not so often used Mar 23 12:00:02 hrw|work : I think i will update the wiki with the solution. it can be very handy in many situations Mar 23 12:08:28 hrw|work, RP: http://maemo.org/pipermail/maemo-developers/2006-March/003373.html Mar 23 12:21:11 zecke: does the cache work again in bbsvn? Mar 23 12:23:13 koen: yes, handlers are added again Mar 23 12:23:36 suck - umbrello gets sequence diagrams wrong :( Mar 23 12:25:53 03rw 07org.oe.dev * rcdf56cd2... 10/packages/binutils/binutils_2.16.91.0.6.bb: binutils: lower DEFAULT_PREFERENCE of 2.16.91.0.6 CVS snapshot for non-arm-linuxeabi builds Mar 23 12:29:44 hrw|work: is there a document or a *.bb file where i can look into how to extract all from the ipk files ? Mar 23 12:29:46 reenoo|z1zZ: whatever Mar 23 12:29:55 reenoo|z1zZ: I'm pissed off now! Mar 23 12:29:56 * france is back (gone 23:38:05) Mar 23 12:30:07 zecke: why? Mar 23 12:30:14 reenoo: on which builds did you test your change? Mar 23 12:30:24 reenoo: I'm currently testing for 4 builds Mar 23 12:30:28 reenoo: how many did you test? Mar 23 12:31:07 reenoo: and while you have the time, what about updating familiar-unstable to not use a random binutils? Mar 23 12:31:11 Zaurusd is still unbuildable with the following error: http://pastebin.com/617856. Any ideas? Mar 23 12:31:59 lardman|work: Connecting to wwwcache.bath.ac.uk|138.38.32.181|:3128... connected. Mar 23 12:32:07 lardman|work: could your proxy be spoiling it? Mar 23 12:32:07 reenoo: I figure you did not read emails? Mar 23 12:32:13 zecke: I did Mar 23 12:32:42 zecke: I guess I just didn't feel like responding to koen insults Mar 23 12:32:49 koen's even Mar 23 12:32:50 reenoo: thanks for wasting my time, I could have spend the time better... Mar 23 12:33:00 * zecke kills the familiar build Mar 23 12:33:01 koen: I'm not sure what they're on about in those benchmarks. The glibc arm5 memcpy should be about as fast as you'll get... Mar 23 12:33:41 RP: I just reported what I got from the benchmark Mar 23 12:33:48 reenoo: what was unclear about. I will do 'this' now if no one objects Mar 23 12:34:39 lardman|work: I know the implementation in OE uses the optimised arm5 memcpy and friends :) Mar 23 12:35:01 zecke: sorry. I must have misread that. I had that change in my tree for a few days and committed it now. I can revert it if you like. Mar 23 12:35:13 I also told Nokia what they should be using. I have no idea if they actually listened... Mar 23 12:35:15 reenoo: no, I just hate when communications fails Mar 23 12:35:21 lunch time Mar 23 12:35:23 cya Mar 23 12:36:41 RP: Well the benchmark for the native c750 functions was faster than his functions Mar 23 12:36:50 RP: So there's no issue there Mar 23 12:37:19 koen: I don't know, is this a known problem if one is behind a proxy? Mar 23 12:37:33 RP: does the memcpy patch still get applied? Mar 23 12:37:47 lardman|work: no idea, I haven't been behind a proxy in years Mar 23 12:37:50 koen: Is it not part of glibc? Mar 23 12:38:05 RP: afaik not, but I could be wrong Mar 23 12:38:48 The patch is in there, and is applied when I looked Mar 23 12:39:58 koen: You're wrong - look in ports/sysdeps/arm/memcpy.S ;-) Mar 23 12:40:14 RP: ah, cool Mar 23 12:40:25 or its coming from somewhere anyway :) Mar 23 12:40:25 collie implementation is slower than his memset8 and memcpy16 though, (as are all of the Nokia stuff) Mar 23 12:41:02 collie is armv4 and I think the optimised versions in OE only apply to armv5 Mar 23 12:41:50 Looking at the .bb, we do patch the memcpy code into glibc Mar 23 12:42:23 03koen 07org.oe.dev * r046a6f9b... 10/packages/binutils/binutils_2.16.91.0.6.bb: disapproval of revision 'cdf56cd23df8324d279945d07a8a2c711f55b025' Mar 23 12:42:36 what in all heavens and earth is monotone doing? Mar 23 12:42:56 koen: ??? Mar 23 12:43:01 aa_: That's something many a monotone user has pondered while waiting for a pull... Mar 23 12:44:05 nevermind Mar 23 12:44:44 if people want binutils cvs snapshots by default so be it Mar 23 12:45:19 reenoo: familar should really be setting a preferred version Mar 23 12:45:33 no Mar 23 12:45:43 setting DISTRO shouldn't be required at all Mar 23 12:47:13 yes and no Mar 23 12:47:56 NAbyss: is it like this all the time? Mar 23 12:48:08 I mean, once I have most things it will be quick right Mar 23 12:48:12 aa_: If you don't do it every day or two, yes. Mar 23 12:48:32 aa_: Just keep doing a pull every 1-3 days, and it'll only take a minute Mar 23 12:48:45 oh that is fine Mar 23 12:48:51 aa_: You're probably better off downloading the OE.db directly from the server if you're just starting out Mar 23 12:49:26 well, I was just starting out yesterday, but it is still going, I can't face wasting all that time Mar 23 12:49:35 Ah, okies Mar 23 12:49:51 aa_: Download the initial database Mar 23 12:50:12 * koen thinks the GettingStarted wiki is quite clear about that Mar 23 12:51:31 RP: the new wiki layout looks nice Mar 23 12:52:05 RP: stuff like this used to be sorted out with COMPATIBLE_HOST, overrides and preferences Mar 23 12:52:08 koen: thanks. Hopefully people can find things easier :). Mar 23 12:52:49 reenoo: If that cvs package doesn't work on a given system, its default preference for that system should be lowered, I agree with that much Mar 23 12:52:57 koen: cheers for the tip - I needed to edit my ~/.subversion/servers Mar 23 12:53:53 RP: ??? how come you guys want everything that doesn't work for you to be opt-in and the rest to be opt-out? Mar 23 12:54:41 RP: it's a _cvs snapshot_ of _binutils_. it shouldn't be the default for anything that doesn't actually require a cvs snapshot. Mar 23 12:56:17 reenoo: I like the approach that without intervention you get the latest everything. If its found not to work, we degrade it. That way we find out about issues that are going to cause us problems Mar 23 12:56:33 no .pc files contain the /usr/X11R6/lib pattern Mar 23 12:58:06 Genesis: Did you check in tmp/staging? Mar 23 12:58:38 in the entire poky Mar 23 12:58:47 i made a little bash script Mar 23 13:01:47 uf... I'm glad to not use .dev ;) Mar 23 13:05:38 morning all. does florian ever hang out here? Mar 23 13:06:10 how depressing: monotone: misuse: database schemas do not match: wanted 1509fd75019aebef5ac3da3a5edf1312393b70e9, got bd86f9a90b5d552f0be1fa9aee847ea0f317778b. try migrating database Mar 23 13:07:25 yes, bad version Mar 23 13:07:53 HopsNBarley: florian as in florian boor? Mar 23 13:08:00 HopsNBarley: he usually hangs out in #gpe Mar 23 13:08:02 ibot: botmail for zecke I've replied to your message on oe@. I'll leave it up to you to sort this out. I believe things should work by default. Mar 23 13:08:14 i think so. hey, thanks koen! Mar 23 13:17:08 koen: reenoo|afk : play well together - please Mar 23 13:17:51 zecke: I've sent a mail to oe@ explaining the disapproval Mar 23 13:18:07 zecke: but as usual hh.org is taking its time to deliver the mail Mar 23 13:18:29 ok, now I am sad. I can't use the monotone db, and my old db doesn't remember that it already downloaded itself... oh well Mar 23 13:18:56 aa_: ? Mar 23 13:19:05 only after failing to install BK did I even consider that you might have stopped using it Mar 23 13:19:21 zecke: my monotone version is too old for the snapshot Mar 23 13:20:20 anyway, my fault. I should hve just waited Mar 23 13:20:31 aa_: try a static build like http://venge.net/monotone/downloads/monotone-0.25.2-linux-x86.bz2 Mar 23 13:20:44 koen: you hero! Mar 23 13:21:06 koen: what was wrong about reenoo's change - besides the lack of communication? Mar 23 13:21:46 zecke: the lack of communication and discussion Mar 23 13:22:04 gcc Mar 23 13:22:18 oops wrong mouse movement Mar 23 13:22:54 koen: do you technically agree in lowering the DEFAULT_PREFERENCE? Mar 23 13:23:54 zecke: yes, as I've said in http://handhelds.org/hypermail/oe/current/7260.html Mar 23 13:24:44 koen: okay :) Mar 23 13:25:58 Genesis: A tip I've just been given is too look through the .la files in staging Mar 23 13:26:07 zecke: It's annoying if people say "stay away from it from now on" and then say "you must fix it" Mar 23 13:26:28 Genesis: Specifically tmp/staging/arm-linux Mar 23 13:26:46 i 'm always checking that Mar 23 13:26:53 i made a mistake in my script Mar 23 14:07:35 crappit running out of cpu power and diskspace on Angstrom already :-) Mar 23 14:09:50 hi Mar 23 14:10:59 koen: read your mail, didn't pbrook mean http://sourceware.org/ml/libc-ports/2005-11/msg00028.html ? Mar 23 14:14:19 pH5: let's ask :) Mar 23 14:21:29 koen: I have thunderbird and firefox now gcc4 patched :-) Mar 23 14:23:28 XorA: cool! Mar 23 14:32:27 XorA: NOTE: package gthumb-2.6.9-r0: task do_package: started Mar 23 14:34:39 koen: gthumb requires 24bit X :-( Mar 23 14:35:38 hmm Mar 23 14:36:04 koen: I did a .bb months ago, but dont think I ever commited it Mar 23 14:36:11 koen: lost it in the mists of time anyway Mar 23 14:44:12 NOTE: package glibc-2.4-r4: task do_compile: started Mar 23 14:44:13 /me leaves for a walk around the lake Mar 23 14:45:10 * koen is still at NOTE: package glibc-2.4-r4: task do_unpack: started Mar 23 14:46:01 NOTE: Applying patch 'eabi-socket.patch' Mar 23 14:46:17 * XorA wonders where he can liberate an Ipaq from Mar 23 15:16:23 pH5: NOTE: package glibc-2.4-r4: task do_compile: completed, so it atleast compiles with the socket patch Mar 23 15:18:11 pH5: I'm eagerly awaiting the results of a boot Mar 23 15:18:11 * koen tries to locate a CF card the bootldr likes Mar 23 15:18:12 is there a wiki entry on altboot? Mar 23 15:18:43 03koen 07org.oe.dev * r7fa7b770... 10/packages/gthumb/gthumb_2.6.9.bb: gthumb: add 2.6.9 **** BEGIN LOGGING AT Thu Mar 23 15:22:33 2006 Mar 23 15:23:40 gosh my University trains Security Nightmares Mar 23 15:25:58 security nightmares? Mar 23 15:26:54 Bernardo: 6 groups created DB implementations Mar 23 15:27:11 Bernardo: all 6 have possible sql injections Mar 23 15:27:22 nice Mar 23 15:27:25 Bernardo: my group hasn't and the eights group didnt manage to write DB code Mar 23 15:27:54 let me guess, the 6 groups that didn't care about security got a decent grade all the same? Mar 23 15:28:18 Bernardo: tomorrow is presentation Mar 23 15:28:31 Bernardo: and I'm currently consdering giving me a login name like DROP TABLE... Mar 23 15:28:38 lol Mar 23 15:30:24 you are going to test everyone else's implementation? Mar 23 15:31:19 yes Mar 23 15:32:26 03xora 07org.oe.dev * r8012e4bd... 10/packages/mozilla/ (19 files in 7 dirs): Mar 23 15:32:26 packages/files/mozilla : cleaned out old versions and add a new version Mar 23 15:32:26 I had left sitting on my HD locally by accident. Mar 23 15:32:26 firefox_1.0.7.bb : add gcc4 patch Mar 23 15:32:26 thunderbird_1.0.7.bb : add gcc4 patch Mar 23 15:32:34 can you use the sql injection bugs to access the filesystem? Mar 23 15:32:47 zecke: consider also using names like DELETE or UPDATE - maybe they do not give all grants for user Mar 23 15:33:06 hrw|work: they have only one DB connection :} Mar 23 15:33:14 well some of them use db4free.org :} Mar 23 15:33:19 *nightmare* Mar 23 15:33:21 zecke: fu... Mar 23 15:33:22 Bernardo: I can just imagie 'SHELL rm -rf ~' Mar 23 15:33:47 XorA: postgres can do that? Mar 23 15:33:59 zecke: no idea, it was just a thought Mar 23 15:34:42 crap after all that mozilla testing only 193M left on /home Mar 23 15:36:38 koen: did you remove reenoo access again? Mar 23 15:36:42 koen: don't do that please Mar 23 15:42:52 zecke: I removed push access for virtually everybody on ewi and dominion a while ago Mar 23 15:43:31 zecke: to avoid multiple heads and stuff Mar 23 15:43:43 koen: ah okay Mar 23 15:44:13 koen: just take care to piss off rw not too much *please* Mar 23 15:44:41 he should stop pissing of the rest of us Mar 23 15:44:42 * Bernardo recalls getting some open source (had to buy a book but could use it for everything) btree code in 1990 Mar 23 15:45:12 zecke: lets revert every cset that wasn't tested on *every* config Mar 23 15:45:55 zecke: and lets insinuate that everything !familiar is a hack Mar 23 15:46:23 koen: Or follow jnc and only worry about OpenZaurus3.5.4 and 'his' toolchain Mar 23 15:46:27 zecke: and lets post on familiar-dev how much OE developers suck Mar 23 15:46:45 koen: stop that - this is just leading us nowhere Mar 23 15:47:02 koen: i have include inherit, how do i tell bitbake that now all packages will be tar.gz ? do i have to run all new ?(hopfully not) Mar 23 15:47:22 zecke: it is indeed leading us nowhere Mar 23 15:47:36 grma: that didn't make any sense to me :) Mar 23 15:47:43 grma: ehm, I'm afraid the packaging step has to run again for each package Mar 23 15:47:48 grma: did you add INHERIT += "package_tar" to local.conf? Mar 23 15:48:40 yup, but bitbake meta-opie, dose not make tar.gz files it tells finished Mar 23 15:49:02 zecke: I'm just getting sick of all those negative comments coming from reenoo without seeing him doing enough work to justify such an asshat attitude Mar 23 15:49:27 NOTE: build 200603231648: completed Mar 23 15:49:33 grma: did you ask it to do tar.gt? Mar 23 15:49:43 grma: do you want to have a tar.gz image?, tar.gz files? Mar 23 15:50:45 tar.gz image or files never mind, i want easy copy the build process to my target ... Mar 23 15:51:30 grma: put INHERIT += "package_tar" in you local.conf, but see local.conf sample Mar 23 15:51:38 zecke: which email works best, @inf.fu or @gmx.net? Mar 23 15:51:50 koen: any Mar 23 15:52:02 koen: for public lists, I prefer @gmx.net Mar 23 15:52:13 i have done this alread, but before that i had run bitbake meat-opie without INHERIT += "package_tar" have in my local.conf file, so i have to build all new ? Mar 23 15:52:33 s/meat/meta/ Mar 23 15:52:40 grma: tar.gz image are probably not built until you build opie-image Mar 23 15:52:45 zecke: you have new mail at @inf.fu Mar 23 15:53:06 ok, and tar.gz files ? Mar 23 15:53:27 grma: dunno, never built tar.gz files -- I assume you are talking for every package? Mar 23 15:54:10 grma: it will create tar.gz :) Mar 23 15:54:37 grma: but see local.conf.sample and search for tar Mar 23 15:54:48 grma: bitbake saves stuff in tmp/deploy/ after the build Mar 23 15:54:58 cbracke: yup tar.gz for every package or build a tar.gz file out from all *.ipk files. Mar 23 15:55:01 grma: you might already have tars Mar 23 15:55:06 where ? Mar 23 15:55:18 grma: see above :) Mar 23 15:55:42 grma: INHERIT += "package_tar" in local.conf and happy rebuilding Mar 23 15:55:51 in tmp/deploy on ipk directory is ! Mar 23 15:56:01 grma: see above (again) Mar 23 15:56:08 grma: you didn't tell it to build tars then Mar 23 15:56:10 ok, happy rebuilding !!! now i know where the error is ! Mar 23 15:56:27 grma: do you have package_tar in your local.conf? Mar 23 15:56:38 grma: you could do bitbake meta-opie -cpackage -f Mar 23 15:56:50 grma: this should recreate all packages (no warranty though) Mar 23 15:56:58 zecke: yup, they are all there in deploy/tar -- just never looked for them before ... Mar 23 15:57:28 zecke, yes but i have included it afer bitbake meta-opie, and i wanted to have an answer if i have to build all new to get tar.gz files ? Mar 23 15:57:54 grma: before rebuilding try the above bitbake command Mar 23 15:59:04 zecke: ok thanks, i will try Mar 23 15:59:53 koen: being disabled/sick helped me to recognize. We all suck, at different things though Mar 23 16:00:07 koen: rw might suck more on communication, we have to live with that Mar 23 16:01:06 Oh, I get an impression why packaging glibc is soo slow Mar 23 16:01:30 zecke: the command didn't help, how do i recompile all ? Mar 23 16:01:54 grma: what did it try? Mar 23 16:02:09 grma: do you have time, or not? Mar 23 16:02:44 zecke: i tried bitbake meta-opie -c package -f --> no tar directory in deploy :-( Mar 23 16:03:07 17:08 < zecke> grma: do you have time, or not? Mar 23 16:03:21 zecke: time ? for what ? Mar 23 16:03:22 grma: your TCP/IP connection must suck, it drops every second pakage Mar 23 16:03:33 grma: time for recompiling, or do you need a hacky shortcut? Mar 23 16:04:03 03pH5 07org.oe.dev * rff962ba2... 10/conf/distro/angstrom.conf: angstrom.conf: use python instead of overrides to select arm eabi on arm builds Mar 23 16:04:07 03pH5 07org.oe.dev * r8465a06c... 10/packages/glibc/ (glibc-2.4/fixup-aeabi-syscalls.patch glibc_2.4.bb): Mar 23 16:04:07 glibc-2.4: Fix up ARM EABI for removed syscalls Mar 23 16:04:07 - patch by Daniel Jacobowitz from http://sourceware.org/ml/libc-ports/2006-03/msg00029.html Mar 23 16:04:11 hacky shortcut :-) copiling would be ok over night Mar 23 16:04:33 grma: look into tmp/stamps, and possible sub directories Mar 23 16:04:52 grma: rm *do_package in the subdirs Mar 23 16:05:00 touch local.conf and rebuild Mar 23 16:05:12 ok will try Mar 23 16:05:17 grma: did you add INHERIT += "package_tar" to the local.conf? Mar 23 16:05:24 grma: if not it will not build tars Mar 23 16:05:42 zecke: yup Mar 23 16:07:17 zecke: ehat is with my tcp/ip connection ? Mar 23 16:07:27 s/ehat/what/ Mar 23 16:07:37 rebuild started :-) Mar 23 16:11:15 didn't rebuild? i will delete and build new :-( Mar 23 16:13:14 dunno, i can't see the output Mar 23 16:22:34 hmm opie fails Mar 23 16:22:45 ah okay Mar 23 16:25:47 cu Mar 23 16:31:26 * chouimat is away: lunch time!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 Mar 23 16:32:00 koen|food: "quick" and "rebuild upward dependencies" in the same sentence. Think of the problems "simple" RDEPENDS caused :) Mar 23 16:33:36 RP: 'caused'? Mar 23 16:34:36 zecke: The handling in bitbake wasn't exactly simple or 100% right at first. Neither was the metadata though which didn't help :) Mar 23 16:35:23 RP: rebuilding upward deps is the right thing to do, but will feed the 'OE is slow' crowd Mar 23 16:35:44 RP: well, I thought 'caused' means it is working now :) Mar 23 16:36:07 zecke: It does. No open bug reports ;-) Mar 23 16:36:28 RP: irc channels are not bugtrackers :) Mar 23 16:36:36 koen: Maybe. I did look at implmenting it but we currently don't track enough information to make it possible Mar 23 16:36:45 zecke: indeed :) Mar 23 16:37:12 RP: we'd have to store graphs on disk somewhere Mar 23 16:37:45 koen: correct. A nice can of worms :) Mar 23 16:37:48 koen: we do that already, /dev/kmem Mar 23 16:38:09 RP: lets do a .dot generator first Mar 23 16:38:28 RP: we could sell the generated graphs as wallposters ;) Mar 23 16:39:31 koen: My current thinking is a module to dump it into an SQL database Mar 23 16:40:08 I think we will be schocked by the cached bitbake speed Mar 23 16:40:24 * koen braces himself Mar 23 16:40:41 zecke: You've found some speed improvements? Mar 23 16:41:10 RP: yes, most of the time is spend in recompiling the functions Mar 23 16:41:25 RP: well, if we store the compiled result, instead of the function string Mar 23 16:41:36 RP: we could be able to save 40-50 seconds from 59 Mar 23 16:41:56 zecke: That would be extremely impressive :) Mar 23 16:43:20 03freyther 07org.oe.dev * rdd750da9... 10/ (2 files in 2 dirs): (log message trimmed) Mar 23 16:43:20 conf/distro/familiar-unstable.conf: Mar 23 16:43:20 Make older binutils snapshot the default Mar 23 16:43:20 for familiar distribution in the dev Mar 23 16:43:20 tree to avoid further errors when newer Mar 23 16:43:21 versions gets added. Mar 23 16:43:23 packages/binutils/binutils: Mar 23 16:46:29 zecke: that commit is wrong Mar 23 16:46:50 zecke: you'd want binutils 2.16 Mar 23 16:47:11 zecke: didn't you read that binutils cvs snapshots are bad? Mar 23 16:48:07 zecke: and the override is also wrong Mar 23 16:48:41 zecke: it's 'arm-linux-gnueabi', not 'arm-linuxeabi' Mar 23 16:49:48 zecke: could you fix both, before 'someone' reverts a zillion csets again? Mar 23 16:50:04 koen: my OVERRIDES is "local:ipaq-pxa270:angstrom:linux-gnueabi:arm:build-linux" Mar 23 16:51:09 ok, so it should be 'linux-gnueabi' Mar 23 16:51:23 pH5: did the patch fix the bootstrap-image problems? Mar 23 16:51:54 * pH5 wants a 10k rpm laptop drive Mar 23 16:52:04 my gpe-image build just finished Mar 23 17:00:04 koen: well, I won't comment on the used version Mar 23 17:00:20 koen: I will fix the commented out override Mar 23 17:00:59 I'll refrain from commenting on the used version as well :) Mar 23 17:00:59 nice Opie is broken and hostap-modules Mar 23 17:01:11 koen: it is the one used by OZ Mar 23 17:01:24 at least my grepped told me that Mar 23 17:01:36 Does anyone have a copy of linux-sl5500-20030509-rom3_10.tar.bz2 as the openzaurus.org sources mirror is running at ~0.4kB/s? Mar 23 17:03:47 zecke: I won't comment about putting OZ versions in familiar Mar 23 17:06:31 koen: fix the argstrom build instead :) Mar 23 17:06:54 angstrom still build fine over here Mar 23 17:07:10 heh right, the *build* is not the problem, currently ;) Mar 23 17:07:11 koen: even the hostap modules? Mar 23 17:07:20 * koen had the common sense to lockdown toolchain bits Mar 23 17:07:56 zecke: I use the builtin ones from the kernel Mar 23 17:08:44 pH5: are you flashing your hx? Mar 23 17:08:54 * koen can't locate a usuable cf card Mar 23 17:09:17 koen: yup. Mar 23 17:09:22 * pH5 found one. took some time, tough Mar 23 17:09:32 bleh Mar 23 17:09:41 gnome-vfs-dbus_svn fails Mar 23 17:09:49 * koen uses the regular on instead Mar 23 17:11:32 bye Mar 23 17:11:56 hi Mar 23 17:13:48 thanks, build started :-), cya Mar 23 17:17:05 hey CoreDump|home Mar 23 17:22:00 * france_ is away: Away Mar 23 17:28:36 hmmm Mar 23 17:28:48 orinoco-modules-0.13e-r4 doesn't compile on 3.5.4.1 Mar 23 17:29:09 koen: started booting, looks like the socket errors are gone :) Mar 23 17:34:49 pH5: yay! Mar 23 17:36:58 pH5: the touch utmp errors are still there? Mar 23 17:38:13 koen: yes, and just now I got the calibration tool. Mar 23 17:38:45 gnn, the touchscreen doesn't work Mar 23 17:38:53 http://en.pastebin.ca/46707 Mar 23 17:39:41 oh, and strace doesn't work either: "syscall: unknown syscall trap 0xef000000" Mar 23 17:39:49 heh Mar 23 17:41:08 is the touchscreen driver loaded? Mar 23 17:41:49 ah, line #355 Mar 23 17:41:53 no, "hx4700_ts: falsely claims to have parameter poll_sample_time" Mar 23 17:52:12 CoreDump|home: I noticed that this morning Mar 23 17:52:33 CoreDump|home: but 0.15rc2 builds Mar 23 17:52:45 Bernardo: ah great =) Mar 23 17:52:59 but what change since yesterday caused it Mar 23 17:53:09 isn't 0.15 broken on arm? Mar 23 17:54:14 I don't even need that module heh, but it's a dependency (probably task-bootstrap) Mar 23 17:59:55 CoreDump|home: I don't know, I started the build yesterday night, at about 21h30 (here is now 18:05) and this morning found it had failed and tried to fix it Mar 23 18:00:25 hrw concluded it was some error in akita.conf and updated it Mar 23 18:00:34 have to run now, see you later Mar 23 18:11:15 Bernardo|away: lol, touch'ing local.conf "removed" the orinoco problem Mar 23 18:13:32 Anyone aware of any pcmcia issues with 2.6.15, havent been able to use my wireless. Mar 23 18:14:10 Ken|JLime: FWIW, CF works perfectly fine on Akita with storage and Wifi cards Mar 23 18:15:05 FWIW? Mar 23 18:15:21 From what i know? Mar 23 18:15:23 Ken|JLime: I think you have to use pcmciautils instead of cardmgr with 2.6.15 Mar 23 18:15:31 for what it's worth Mar 23 18:15:34 For What Its Worth ;) Mar 23 18:15:48 Aha, pcmciautils..oki, I used cardmgr (which didnt work) Mar 23 18:16:02 Will give that a go then, whats the command? Mar 23 18:22:38 koen: with this patch I got into gpe: http://en.pastebin.ca/46720 Mar 23 18:22:59 but there are more things broken. many alignment traps and programs not starting Mar 23 18:26:02 pH5: we can fix those apps one by one I guess Mar 23 18:26:09 pH5: but at least it boots into X Mar 23 18:45:27 I want to include more programs in my bootstrap-image, however when adding to extra-bootstrap-rdepend and extra-bootstrap-depend, they dont appear in the finished rootfs Mar 23 18:53:05 anyone? Mar 23 18:55:18 did you add them to IPKG_INSTALL as well? Mar 23 18:56:04 No i didnt, what does that do, and where should i put it? Mar 23 18:57:36 IPKG_INSTALL contains the list of programs to be installed with ipkg in your image Mar 23 18:57:53 bootstrap-image should already contain that variable Mar 23 18:58:24 oki, so if i set IPKG_INSTALL in my distro file it should add those programs after the bootstrap is finished? Mar 23 18:58:50 hmm? Mar 23 18:59:01 not in the .conf, in $your-image.bb Mar 23 18:59:23 bootstrap-image.bb? Mar 23 18:59:55 http://oz.hentges.net/openembedded/org.openembedded.dev/packages/meta/bootstrap-image.bb Mar 23 19:00:04 indeed Mar 23 19:01:22 big thx Mar 23 19:02:16 np ;) Mar 23 19:05:29 hi zecke Mar 23 19:08:48 Ive added pcmciautils but it says it cannot find it (and its in packages/pcmciautils) Mar 23 19:08:54 the rest worked fine Mar 23 19:09:40 doh, stupid question, hadn't compiled it yet Mar 23 19:22:03 Does wireless-tools install pcmcia-cs?.. And should i delete that if i want to use pcmciautils? Mar 23 19:33:49 * koen builds an eabi image for his c700 Mar 23 19:35:06 Anyone familiar with how i can remove the pcmcia-cs and make it use pcmciautils instead? Mar 23 19:36:14 PCMCIA_MANAGER = "pcmciautils" in you machine.conf Mar 23 19:36:16 or local.conf Mar 23 19:36:41 Great, thx Mar 23 19:37:53 Putting it in the distro file wouldnt do the trick? Mar 23 19:40:08 it would, but I'm not sure if that is the right place Mar 23 19:41:31 hi Mar 23 19:41:40 RP: we have first terrier tester Mar 23 19:48:03 hey hrw Mar 23 19:48:27 zecke: what kind of voodoo would I need to send the tinderbox builddata? Mar 23 19:48:40 www.paypal.com Mar 23 19:48:47 lol Mar 23 19:48:48 Donation -> freyther@inf.fu-berlin.de Mar 23 19:48:49 heh Mar 23 19:49:25 koen: check conf/local.conf in buil/zaurus, build/ipaq or build/arg Mar 23 19:51:18 hi Mar 23 19:51:24 hrw: you fix did it Mar 23 19:51:35 now I'm just stuck trying to d/l zaurusd Mar 23 19:51:56 for some reason svn won't work on my mandriva build machine when called from inside bitbake Mar 23 19:52:00 works fine by itself Mar 23 19:52:35 grr bloody thumb crap Mar 23 19:53:54 mmm... it seems I can't get to svn.o-hand.com from here Mar 23 19:55:55 anyone has a copy of zaurusd-0.0+svn20060319 that can give me? Mar 23 19:56:02 trying to fetch now Mar 23 19:56:06 failed Mar 23 19:56:36 a traceroute here dies after po1.car04.dllstx5.theplanet.com Mar 23 19:59:56 hrw: you don't have the zaurusd tarball lying there somewhere, by any chance? :) Mar 23 20:01:09 bbl Mar 23 20:01:18 Bernardo|away: nope Mar 23 20:01:35 http://ewi546.ewi.utwente.nl/tmp/zaurusd_svn.o-hand.com_.repos.misc.trunk__20060323.tar.gz Mar 23 20:01:39 dinner time... maybe the site will be back up after dinner Mar 23 20:01:44 koen: thanks! Mar 23 20:01:46 fetched 20 minutes ago Mar 23 20:02:48 I'll try and fool oe by renaming it to 20060319 (the preferred version for oz 3.5.4.1) Mar 23 20:02:59 really have to go no, see you guys later Mar 23 20:04:49 NOTE: package opie-bluetoothmanager-1.2.1+cvs-20060323-r0: task do_compile: failed Mar 23 20:04:57 | filelistitem.h:16:20: client.h: No such file or directory Mar 23 20:05:16 gremlin[it]: yes? Mar 23 20:06:00 gremlin[it]: we have a new opie contributor extending our bluetooth support Mar 23 20:09:38 zecke : nothing special just state i got another error doing opie-image :) ... Mar 23 20:11:33 gremlin[it]: http://ewi546.ewi.utwente.nl/tinderbox/showbuilds.pl?tree=OpenEmbeddedBuilds Mar 23 20:11:51 gremlin[it]: two of them broke at the bluetooth manager :) Mar 23 20:12:23 * gremlin[it] blame one of the two ... chosing at random :) Mar 23 20:16:07 Even though setting PCMCIA_MANAGER="pcmciautils" it still grabs and installs pcmcia-cs thus creating package conflicts. Mar 23 20:17:48 Could it be that the cache isnt cleared correctly? Would rather not do a complete tmp deletion. Mar 23 20:18:27 Ken|JLime: are you building a gpe image? i think gpe-conf still explicitly depends on pcmcia-cs Mar 23 20:18:43 no, its based on bootstrap-image Mar 23 20:19:04 dont know if task-bootstrap contains any reference to pcmcia-cs Mar 23 20:19:10 and if so, how to work around it Mar 23 20:21:12 if nobody knows, you can Mar 23 20:21:19 always do a bitbake -v Mar 23 20:21:36 Ken|JLime: rebuild task-bootstrap after config changes Mar 23 20:23:57 doh, that did the trick Mar 23 20:24:06 :) thx Mar 23 20:24:50 Ken|JLime: thats always needed Mar 23 20:25:00 Ken|JLime: how goes jlime distro? Mar 23 20:28:19 Quite well actually, the 2.6 kernel is looking really nice. Only Irda left, otherwise its 100% supported. And I must say that Openembedded makes creating a distro alot(!) easier. :) still having some issues with some packages (sh3 specific) but other than that things are looking better all the time. Mar 23 20:29:18 Should have jlime ported to the j7xx in a week or so Mar 23 20:30:32 koen: could you roll another tarball of 3.5.4 please? I need to resync the mirror Mar 23 20:30:55 CoreDump|home: cant you just compare dirs? Mar 23 20:31:39 umm comparing dirs containing thousands of ipks wouldnt be fun heh Mar 23 20:36:53 hrw: Was it successful on terrier? Mar 23 20:37:17 RP: he will install Mar 23 20:37:26 CoreDump|home: hmm. perhaps not Mar 23 20:38:30 hrw: btw, in answer to your question the other day, edit/remove /etc/zaurusd/scripts/zaurus-hinge Mar 23 20:39:07 RP: noticed that Mar 23 20:39:34 RP: any solution to make it screen off on close, restore on open? Mar 23 20:40:54 hrw: You can change the script to make any calls you like - echo to /sys/class/backlight ? Mar 23 20:41:20 but we can't make it resume on hinge openening due to hardware limitations Mar 23 20:41:36 good to know Mar 23 20:42:18 good morning Mar 23 20:42:46 hi TheMasterMind1 Mar 23 20:43:26 hrw: eeek, _please_ make suspend-on-hinge-close optional Mar 23 20:43:39 why monotone for oe? Mar 23 20:44:00 CoreDump|home: s/hrw/RP Mar 23 20:44:05 TheMasterMind1: because we could not afford bitkeeper licenses? Mar 23 20:44:26 njs: what is the state of monotone 0.26? Mar 23 20:44:33 speaking of which, zecke, any progress on the scm comparisons? Mar 23 20:44:38 * kergoth outta teh loop Mar 23 20:45:06 well yea, just never heard of monotone before, why not just svn though? what're the advantages of monotone Mar 23 20:45:27 svn is client/server, by itself anyway Mar 23 20:45:40 zecke: did you see the monotone dumb progress report on the monotone mailinglist? Mar 23 20:46:03 no Mar 23 20:46:35 TheMasterMind1: we wanted to avoid a workflow change Mar 23 20:46:45 zecke: http://thread.gmane.org/gmane.comp.version-control.monotone.devel/6293 Mar 23 20:46:58 kergoth: well, no one wanted a darcs, hg account from me Mar 23 20:47:50 kergoth: git is the fastest but needs regular packing Mar 23 20:48:04 kergoth: git is not portable - keithp claims they are working on that Mar 23 20:48:12 brb Mar 23 20:48:20 kergoth: git has a horrible interface - I hate to merge with it Mar 23 20:48:30 hmm Mar 23 20:48:33 kergoth: monotone 0.26 is getting way faster already Mar 23 20:48:37 kergoth: has lua support Mar 23 20:48:50 kergoth: forget about darcs for OE - ghc isn't yet ready Mar 23 20:49:11 how so? Mar 23 20:49:22 kergoth: hg, well after a push it sits their minutes, doing a CTRL+C while pulling/pushing requires a hg recover on the server Mar 23 20:49:30 ow Mar 23 20:49:39 kergoth: darcs? it can not open more than 100 files Mar 23 20:49:44 ah Mar 23 20:49:53 kergoth: bisect takes more space than evolution Mar 23 20:50:22 kergoth: on a pull you could get $HOME/.ssh/authorized_keys Mar 23 20:50:25 patched Mar 23 20:51:00 and a 'rebase' that works will all other systems doesn't work darcs Mar 23 20:51:35 co .dev until the branch point,pull oz354fam083 doesn't suceed :) Mar 23 20:51:56 kergoth: and svk is quite slow and would require database dumps like we do with monotone Mar 23 20:52:21 kergoth: measuring the resonance, I think we will stick with monotone and be happy with the speed increase Mar 23 20:52:49 I like monotone's UI but not its speed. I like git's power but not its UI. Mar 23 20:53:06 and we have a monotone developer on stand by here in #oe :) Mar 23 20:54:10 yeah monotone needs to improve speed, firewall friendly nes Mar 23 20:54:31 I hate to not be able to pull monotone from uni and my office at uni :} Mar 23 20:55:30 zecke: ssh port forwarding to vanille or ewi? Mar 23 20:55:45 zecke: sounds like the monotone-dumb extension was made for you Mar 23 20:56:13 koen: I'm using git at the office :} Mar 23 20:56:18 "This will take a VERY long time (hours, if not days) when run for the first time." what's that about Mar 23 20:56:30 TheMasterMind1: security :) Mar 23 20:56:31 that it will take very long Mar 23 20:56:41 the solution is mentioned a few words later Mar 23 21:00:35 well i'm attemping to convince the gumstix project to switch to OE, but they make a good point, the time and effort needed to get started with OE is so much greater than a 'svn checkout; make' Mar 23 21:00:47 koen: what do you want to have set beside "arm-linuxeabi"? Mar 23 21:01:10 TheMasterMind1: look at the NSLU2 project Mar 23 21:01:18 TheMasterMind1: they only need wget ; make Mar 23 21:01:19 zecke: *eabi* Mar 23 21:01:31 koen: just eabi? Mar 23 21:01:45 koen: that doesn't sound too good Mar 23 21:01:47 *eabi Mar 23 21:01:58 what about GNU? Mar 23 21:02:19 there is linux-gnueabi and iwmmxt-linuxeabi, etc Mar 23 21:03:25 zecke: and as bero pointed out, that binutils works fine for glibc 2.4/arm, so default_preference_arm = "-1" is wrong Mar 23 21:03:53 koen: arguable, what is the default preference of glibc2.4? Mar 23 21:04:02 -1 right now Mar 23 21:04:09 okay Mar 23 21:04:14 for everything Mar 23 21:04:17 if you change that, you may change binutils as well Mar 23 21:04:40 I won't change that as glibc 2.4 still needs some patches Mar 23 21:05:08 koen: 'if you change the default preference of glibc2.4 you may change the one of binutils as well" Mar 23 21:07:51 Im having some issues with getting my pcmcia wifi working. Using 2.6.15, and I get "failed to initialize firmware", now i see that you have some packets for it. Can i add them to the image file without risking conflict with pcmciautils? Mar 23 21:08:09 zecke: hrm i can't seem to find that. the wiki articles i looked at all refer to template.mk and not oe Mar 23 21:08:52 Ken|JLime: which driver? hostap or orinoco? Mar 23 21:08:58 TheMasterMind1: http://www.nslu2-linux.org/wiki/Development/MasterMakefile Mar 23 21:09:21 Orinoco Mar 23 21:09:37 a hermes card Mar 23 21:09:53 koen: the right override would have been linux-gnueabi :} Mar 23 21:09:56 Ken|JLime: in OZ with 2.6.15 we use in-kernel orinoco and in-kernel hostap Mar 23 21:10:19 hmm, anyway of getting the firmware in place then? Mar 23 21:10:38 Ken|JLime: no idea - hostap only here... Mar 23 21:10:42 03freyther 07org.oe.dev * r46d4bd0c... 10/packages/binutils/binutils_2.16.91.0.6.bb: Mar 23 21:10:42 packages/binutils/binutils_2.16.91.0.6: Mar 23 21:10:42 The GNU EABI override for ARM Mar 23 21:10:42 was wrong. Simply remove it. If we Mar 23 21:10:42 change the DEFAULT_PREFERENCE for glibc2.4 Mar 23 21:10:43 we can update again. Mar 23 21:10:49 Ken|JLime: mickeyl is local wifi guru Mar 23 21:11:08 how does it work with hostap then? Mar 23 21:11:19 Ken|JLime: most of hostap cards contain firmware Mar 23 21:11:41 Ah oki Mar 23 21:11:52 Ken|JLime: those ones which does not (connect+ by sandisk for example) require prism3-support package which download firmware on insert Mar 23 21:12:26 Oki, does OE supply those firmwares? Mar 23 21:12:48 Ken|JLime: yes Mar 23 21:12:59 with the prism3-support packet? :) Mar 23 21:13:02 Ken|JLime: prism3-firmware (which can be applied to prism2 cards too) Mar 23 21:13:12 prism3-support rdepend on prism3-firmware Mar 23 21:14:12 oki, need to add that anyways. thx Mar 23 21:14:15 Ken|JLime: There are several firmware packages in OE such as spectrum-fw Mar 23 21:14:55 The basic idea is that its hard to download firmware without a working wifi, so i want to have as a many firmwares in place as possible. Mar 23 21:15:56 spectrum-fw.. maybe that is missing for Ken|JLime Mar 23 21:16:33 Ken|JLime, hrw It supports spectrum_cs which is orinoco derived Mar 23 21:16:44 I hate when programs try to map sh3 to an arch Mar 23 21:16:49 :) nicy Mar 23 21:17:03 ok. I have one prism card and who knows - maybe will get another ond ays Mar 23 21:17:07 on days Mar 23 21:17:54 hrw: My spectrum card was given to me so I'd stop breaking orinoco/spectrum in OE ;-) Mar 23 21:18:06 :) Mar 23 21:18:22 RP: good reason Mar 23 21:18:24 Its worked too as I now know how it all works... Mar 23 21:18:35 (and do test with it) Mar 23 21:20:44 RP: soon I will probably work on improving irda keyboards support in OZ Mar 23 21:21:17 hrw: Sounds good. One of these days I might actually test the irda code I maintain :) Mar 23 21:21:20 if posts and customs will pass package Mar 23 21:22:09 away Mar 23 21:24:21 Just for interest. what kind of computers are you guys using for compiling? Mar 23 21:24:54 Ken|JLime: athlon1700+, athlon2200+, 2xopteron in 32bit Mar 23 21:25:19 <_law_> Ken|JLime: sempron 2600+ (2gb ram) Mar 23 21:25:28 Ken|JLime: 2.6Ghz P4, 2x2.8ghz Xeon Mar 23 21:25:42 Ken|JLime: Athlon2000+ here, soon to be a P4 D 2.8 as well. Mar 23 21:25:54 2Ghz Centrino works too Mar 23 21:26:18 Im sitting on a 3Ghz P4. but still feels really slow compiling. Maybe its more up to the Ram? Mar 23 21:26:29 Ken|JLime: How much ram? Mar 23 21:26:32 512MB Mar 23 21:26:40 More will help a lot Mar 23 21:27:01 I have 1GB as a minimum... Mar 23 21:27:14 I've got 2gb in this box.. flies with make -j4. Mar 23 21:27:51 Btw, about ram usage, I talked to a guy last night that had soo much problems crosscompiling that he intended to compile the entire debian source on his Jornada 7xx Mar 23 21:28:52 Bloody hell. That's.. insane. Mar 23 21:29:18 NAbyss: no it's called 'gentoo' Mar 23 21:29:43 you spends months compiling on your pda for a 0.00001% speed improvement Mar 23 21:29:59 koen: True, true.. but compiling it on a Jornada!? I had one of the 7xx models.. 755 or so, IIRC.. fast it was not. Mar 23 21:31:37 It got like 32MB ram and 200Mhz ARM processor me think. :) I told him to give it a try and get back to me after christmas 2006. Mar 23 21:31:50 2006? I'd say 2007 ;) Mar 23 21:31:54 :D Mar 23 21:32:38 ouch Mar 23 21:32:57 how long does compiling 3234*666 lines of python code take? Mar 23 21:33:28 hell of time? Mar 23 21:33:52 haha Mar 23 21:35:55 hrw: okay I think avoiding this will lead to a 40 seconds speed improvement Mar 23 21:36:03 when loading from the cache Mar 23 21:36:26 zecke: It sounds possible :) Mar 23 21:36:48 zecke: so you got <20s with it? Mar 23 21:37:06 ok - time to bed Mar 23 21:37:07 cu Mar 23 21:37:17 'night hrw Mar 23 21:37:28 'night hrw Mar 23 21:38:09 I have a question concerning 'time' Mar 23 21:38:21 real is how much time was spend, user in userspace, sys in kernel Mar 23 21:38:29 and the delta is the time where the task didn't run? Mar 23 21:38:48 real 2m22.479s Mar 23 21:38:48 user 0m59.544s Mar 23 21:38:49 sys 0m1.104s Mar 23 21:39:35 zecke: correct - it accounts for how long it wasn't scheduled in for afaik Mar 23 21:39:47 zecke: or in this case perhaps it was waiting on IO? Mar 23 21:40:03 RP: I'm doing a bitbake build Mar 23 21:40:10 real 0m40.238suser 0m23.595ssys 0m1.033s Mar 23 21:40:27 hmm how did I paste that in one line Mar 23 21:40:43 pasted from and osx terminal? Mar 23 21:40:54 koen: well I did that twice Mar 23 21:41:04 osx pasting sucks Mar 23 21:41:20 it kills newlines in 70% of my pastes Mar 23 21:41:50 okay it looks like I can do 30 seconds better, I need to change the data structure and caching a bit Mar 23 21:41:59 hey gerwinin Mar 23 21:42:03 hey Koen Mar 23 21:42:10 Alles goed ? Mar 23 21:42:19 tot nu to wel :) Mar 23 21:42:30 Vertel zit er iets ernstigs aan te komen dan ? Mar 23 21:43:10 Trouwens koen ik ben op de cebit geweest en zag daar zeer veel oe met opie :) Mar 23 21:43:24 tsssk Mar 23 21:43:25 opie Mar 23 21:43:51 ik heb ook wel gpe zien draaien op een paar coole apparaten Mar 23 21:43:55 Got an issue with packet orinoco-modules. It wants to map the Arch to "sh3" which is not exactly the one im using (its arch=sh with -m l..) so of course it cannot find a sh3 arch, and drops Mar 23 21:44:09 RP: ~5 seconds could be due IO, but it is unlikely Mar 23 21:44:29 Hey zecke long time no see ! Mar 23 21:44:32 zecke: Hm, any other tasks competing for CPU time? Mar 23 21:44:44 NAbyss: just two or three bitbake builds Mar 23 21:44:58 the EWI monster Mar 23 21:45:26 5 sec.. sounds plausible then Mar 23 21:46:00 Koen we are organising a new linux and opensource congress in Eindhoven Mar 23 21:46:11 it would rock to have a BitBake that reads from a cached parsed in < one minute Mar 23 21:46:16 I was wondering if some guys would like to come Mar 23 21:46:41 We got a speakers budget so travelling costs shouldn't be a problem this time Mar 23 21:46:54 gerwinin: when does it take place? Mar 23 21:47:12 Koen it will be in september Mar 23 21:47:28 I was thinking to have one seperate room for oe Mar 23 21:47:36 and related projects Mar 23 21:47:38 if its in the weekend I can travel for free :) Mar 23 21:47:50 Okay how about you zecke ? Mar 23 21:48:00 it depends :) Mar 23 21:48:40 Zecke we can participate in the travelling costs Mar 23 21:49:03 does OE have all the packages required to use bitbake and OE? Mar 23 21:49:07 luckily that is currently not the biggest issue Mar 23 21:49:16 TheMasterMind1: included in OE? Mar 23 21:49:23 Zecke how about your health ? Mar 23 21:49:25 TheMasterMind1: yes, thanks to the NSLU2 folks Mar 23 21:49:33 gerwinin: it is okay, ups and downs Mar 23 21:49:43 ok cool Mar 23 21:49:45 gerwinin: I'm still struggeling with a cold I got at the FOSDEM Mar 23 21:49:59 Zecke: okay I hope you get better soon Mar 23 21:50:01 TheMasterMind1: they build OE natively on their OE distris Mar 23 21:50:08 that's pretty cool Mar 23 21:50:26 i'm thinking of making a small OE distro to run in vmware player that can be used Mar 23 21:50:42 TheMasterMind1: sure, sounds reasonable Mar 23 21:50:52 Koen we got the idea to organise a design challenge as well do you think that it should be a good idea to focuss one design challenge on oe ? Mar 23 21:51:30 that nslug2 master Makefile is pretty crazy though Mar 23 21:51:40 ~lart pb_ Mar 23 21:51:41 * ibot throws pb_'s poor little doggy off a cliff Mar 23 21:51:42 gerwinin: that could be an idea Mar 23 21:52:11 Via promissed me to take part if I write an article about how to create an embedded pabx in oe Mar 23 21:53:31 a voip box? Mar 23 21:54:08 yes I did it once and they were pretty impressed by that Mar 23 21:55:09 I did it for my current job , it was quite funny , I did with oe in one week what they have een trying to do with other tools in three months :)tools in Mar 23 21:55:42 ~lart pb_ Mar 23 21:55:42 * ibot puts on some milking gloves. "All right, now, pb_, this won't hurt a bit...." Mar 23 21:55:50 until he wakes up :) Mar 23 21:56:02 is there a cvsview type web interface for the OE monotone repo? Mar 23 21:56:26 TheMasterMind1: http://ewi546.ewi.utwente.nl/tmp/viewmtn/ Mar 23 21:56:41 or http://monotone.vanille.de/viewmtn/ Mar 23 21:58:00 mucha gracias Mar 23 21:58:19 Koen I am busy with some webcam things now in gpe but I am wondering if there is a similar library as qtvision I can talk to Mar 23 21:58:56 what's the dreambox branch? Mar 23 21:59:01 gerwinin: did you make pictures? Mar 23 21:59:07 gerwinin: gstreamer? Mar 23 21:59:07 TheMasterMind1: Sega Dreambox Mar 23 21:59:18 gerwinin: pictures of Opie at vendors Mar 23 21:59:21 TheMasterMind1: for the dreambox PVR Mar 23 21:59:26 Zecke : yes Mar 23 21:59:34 oh it is the pvr :} Mar 23 21:59:35 ouch Mar 23 21:59:48 Zecke: they are going to send me some devices :) Mar 23 22:00:14 Zecke: the Taiwanese are using it a lot for there wifi phones and stuff Mar 23 22:00:41 I'm doing my DVB-H research on Opie Mar 23 22:01:16 Zecke: I am busy now with a videoconference app in opie + gpe + java Mar 23 22:02:04 gerwinin: a friend of mine has jikes + classpath and swing running on his SIMpad Mar 23 22:02:28 Zecke at this moment I constructed my own webpad with a via epia mainboard Mar 23 22:03:01 And an lcd screen we still had laying arround at my work Mar 23 22:03:08 ah, why a completely different branch though, and not a new machine/distro in the same tree Mar 23 22:03:38 TheMasterMind1: this makes it easier to always build a stable product from OE Mar 23 22:03:42 Zecke I am doing the java thing for windows based devices , tried it first in .net but that is really not working Mar 23 22:04:06 .net is a piece of shit Mar 23 22:04:16 I hope we will have a Qt on Windows.CE, Symbian.. this year :) Mar 23 22:04:40 Yeah I hope so as well becayse .net is the worst that can happen Mar 23 22:05:11 My boss bought eight licenses and 1 of them I send back to micro$oft Mar 23 22:05:30 * NAbyss grins Mar 23 22:16:50 koen: Does Angstrom plan to use PCMCIA_MANAGER = "pcmciautils" ? Mar 23 22:17:09 If so, can I set that for the distro? Mar 23 22:17:50 (or at least set if for the capable machines?) Mar 23 22:18:41 RP: set it for the capable machines if you want Mar 23 22:18:59 what's the latest bitbake version? Mar 23 22:19:11 TheMasterMind1: trunk Mar 23 22:19:31 TheMasterMind1: developers.berlios.de/projects/bitbake should have the last released version Mar 23 22:20:35 k. gentoo's old then. probably best to use trunk anyway right? Mar 23 22:20:49 TheMasterMind1: currently yes Mar 23 22:27:30 Hmm. The bitbake manual doesn't detail ?= Mar 23 22:28:20 At which point does ?= check for existance and set the variable if not found? Mar 23 22:30:28 I think i will settle for the fact that the orinoco drivers somehow isnt working in 2.6.15 Mar 23 22:31:55 RP: At the same time as := expands (during parsing) Mar 23 22:32:15 pH5: Not according to my tests :-/ Mar 23 22:33:17 my test was flawed :) Mar 23 22:37:34 I forgot how much diskspace oe took... Mar 23 22:37:48 nevarrie: Yep, it takes a fair bit.. Mar 23 22:40:08 About 7.5gb here, and that's just for one target. Mar 23 22:40:30 nevarrie: as much as you want Mar 23 22:44:17 I ran out od space compiling vim...I am not moving it to my 300gb partitions instead of my 5gb partition.. Mar 23 22:44:53 rwhitby: hey Mar 23 22:45:52 hey Mar 23 22:46:05 rwhitby: your builds dropped of? Mar 23 22:46:52 dunno, haven't looked lately Mar 23 22:49:10 easiest way to update the ipkg file list? Mar 23 22:50:06 bitbake package-index Mar 23 22:50:19 thx Mar 23 23:05:24 bitbake caches are broken again Mar 23 23:19:02 one might have guessed that Mar 23 23:19:05 "This commit breaks the cache again " Mar 23 23:20:26 the format is changed Mar 23 23:20:31 and will change one more time Mar 23 23:20:54 nice Mar 23 23:22:31 NOTE: Parsing finished. 3167 cached, 0 parsed, 65 skipped, 0 masked. Mar 23 23:22:31 Requested parsing .bb files only. Exiting. Mar 23 23:22:31 real 0m24.940s Mar 23 23:22:31 user 0m23.301s Mar 23 23:22:32 sys 0m0.997s Mar 23 23:22:34 woot Mar 23 23:22:49 nice Mar 23 23:23:00 N g F in t A :} Mar 23 23:23:47 why does is still take 24 seconds? Mar 23 23:24:08 koen: a) becuase gpe's bb files are fishy (just kidding) Mar 23 23:24:28 IO bound? Mar 23 23:24:29 koen: b) I still recompile the methods Mar 23 23:24:37 koen: c) we will find out Mar 23 23:24:50 koen: I think we will see it drop down to 20 quite easily Mar 23 23:43:49 okay 17.91 seconds are spend in unpickle Mar 23 23:44:30 about 10 seconds is spend in python_sub Mar 23 23:44:39 which si responsible for expanding python expressions Mar 23 23:46:31 ~whaleslap zecke Mar 23 23:46:33 * ibot beats zecke upside and over the head with a freakishly huge killer whale named Hugh Mar 23 23:47:13 ~herring pb_ Mar 23 23:47:14 * ibot whacks pb_ on the side of the head with a large red herring named alfred Mar 23 23:47:23 ibot: make jokes about phils mumma Mar 23 23:47:25 make: *** No rule to make target `jokes about phils mumma'. Stop. Mar 23 23:47:56 pb_: good evening, did you just return home? from the pub or work? Mar 23 23:48:18 zecke: good evening. indeed I did, from the pub. Mar 23 23:48:32 pb_: mickeyl claimed that __functions__ handling was created by you :) Mar 23 23:49:00 zecke: you shouldn't believe everything mickeyl says Mar 23 23:49:09 but, in this instance, I think he might be correct Mar 23 23:49:30 pb_: there was hardly anything I could optimize ;) Mar 23 23:50:13 just about 60% :) Mar 23 23:50:17 heh Mar 23 23:50:35 pb_: have a good nite :) Mar 23 23:50:57 zecke: you too :-) Mar 23 23:51:34 * zecke needs to create some more UML diagrams Mar 23 23:51:40 the Software project at Uni failed Mar 23 23:51:49 now we need to do a marketing presentation Mar 23 23:52:08 Do you want to see a Demo or Funky Diagrams? and then we go to the Diagrams Mar 23 23:54:12 so why is it monotone is so slow? Mar 23 23:55:23 you dont have the roster patch applied Mar 23 23:55:32 Roster patch? Mar 23 23:55:46 and its down to ~ 30seconds a rev now without that patch Mar 23 23:55:54 yeah Mar 23 23:56:02 TheMasterMind1: the way it handles revisions is kind of slow Mar 23 23:56:13 its a ... beta patch taht helps speed things up immensly Mar 23 23:56:31 emte: where might one find this patch Mar 23 23:56:37 TheMasterMind1: in contrast to any other SCM signature etc... are controled Mar 23 23:56:48 the wiki says download the snapshot and continue from step 6, what about step 4, generating a public key Mar 23 23:56:51 TheMasterMind1: monotone 0.26 will be way faster Mar 23 23:57:09 TheMasterMind1: if you want to commit to your db, you will need a public key Mar 23 23:57:18 unless you need commit or specific repo access you can skip the kep Mar 23 23:57:31 since that command actually no longer exists Mar 23 23:58:24 it looks like it was removed in the last version of monotone Mar 23 23:58:51 last/current stable Mar 24 00:00:23 emte: is this patch in trunk? Mar 24 00:00:38 TheMasterMind1, i do not belive so Mar 24 00:00:53 njs, if your around do you knwo where the roster patch lives? Mar 24 00:00:57 know* Mar 24 00:01:10 emte: I think you can not get it seperately Mar 24 00:01:16 emte: rosters will be in monotone 0.26 Mar 24 00:01:57 yeah, but i was guessing it was seperate from the way others have been talking about it for the last few mo Mar 24 00:02:31 03rpurdie 07org.oe.dev * r5f22b55f... 10/packages/linux/ (7 files in 2 dirs): linux-oz-2.6: Updated defconfigs to add CONFIG_TUN=m (refresh c7x0 and cxx00 against 2.6.16). Add patch to ignore invalid MMC SCR structures. This should be tested carefully as its effect is unknown. Mar 24 00:02:35 03rpurdie 07org.oe.dev * r1aeba3dd... 10/packages/dates/dates_svn.bb: dates: Sync with poky and fix icons issue Mar 24 00:02:40 03rpurdie 07org.oe.dev * r1fd90356... 10/packages/contacts/ (contacts_0.1.bb contacts_svn.bb): contacts: Sync with poky and fix icons issue Mar 24 00:02:46 03rpurdie 07org.oe.dev * ra8d1f60a... 10/packages/autoconf/ (autoconf-2.59/autoconf-x.patch autoconf_2.59.bb): autoconf: Correct the xlibs detection to correctly detect xlibs (from poky) Mar 24 00:02:51 03rpurdie 07org.oe.dev * rb76b6bfe... 10/conf/ (7 files in 4 dirs): PCMCIA_MANAGER: Correctly handle this for various machines in a consistent way and set a default in bitbake.conf for backwards compatibility in such a way it can be overridden in a variety of ways. Mar 24 00:02:56 03rpurdie 07org.oe.dev * r1fa9e864... 10/packages/pcmanfm/ (8 files in 2 dirs): Add pcmanfm 0.1.9.8 (from poky) Mar 24 00:03:08 03rpurdie 07org.oe.dev * r487614c7... 10/packages/galago/ (eds-feed/disable-bonobo.patch eds-feed_0.3.2.bb): Add eds-feed 0.3.2 (from poky) Mar 24 00:03:14 03rpurdie 07org.oe.dev * refa3876b... 10/ (.mt-attrs packages/gnome/libgnomecanvas_2.6.1.1.bb): libgnomecanvas: Correct DEPENDS Mar 24 00:03:21 03rpurdie 07org.oe.dev * rf7ac5e5e... 10/packages/ipkg/ipkg-native_0.99.159.bb: Add ipkg-native 0.99.159 Mar 24 00:03:23 03rpurdie 07org.oe.dev * r79664cfb... 10/packages/gtk+/ (15 files in 2 dirs): Add gtk+ 2.8.16 (from poky). Only a subset of patches are being used against this by poky but the treeview fix is included Mar 24 00:03:27 03rpurdie 07org.oe.dev * r79cefc4c... 10/packages/gnome/libsoup_2.2.7.bb: Add libsoup 2.2.7 (from poky) Mar 24 00:03:35 03rpurdie 07org.oe.dev * r5c2e54a0... 10/packages/gcalctool/ (gcalctool/makefile-fix.diff gcalctool_5.7.32.bb): Add gcalctool 5.7.32 Mar 24 00:03:38 03rpurdie 07org.oe.dev * rfb171e42... 10/packages/js/ (files/host-cc.patch files/jsautocfg.h js_1.5.bb): Add js 1.5 Mar 24 00:03:42 03rpurdie 07org.oe.dev * rbda91375... 10/ (4 files in 3 dirs): Add kf 0.5.4.1 Mar 24 00:03:46 03rpurdie 07org.oe.dev * r99615a35... 10/ (4 files in 3 dirs): Add leafpad 0.8.4 Mar 24 00:07:07 oh Mar 24 00:09:48 RP: touching documentation.conf for this PCMCIA variable would be nice Mar 24 00:10:17 zecke: ok, will do Mar 24 00:20:06 any existing i686 machine confs i can copy and modify? Mar 24 00:20:32 TheMasterMind1: via boards Mar 24 00:20:46 TheMasterMind1: I'm too tired something like epia is it named Mar 24 00:22:24 k Mar 24 00:22:30 native.conf looks like it might work too Mar 24 00:27:26 TheMasterMind1: well, I can build for epia on a PowerPC (cross compile) Mar 24 00:27:36 TheMasterMind1: native limits me :) Mar 24 00:29:04 RP: how do you upgrade OE at o-hand.com? Mar 24 00:29:11 RP: aeh bitbake... Mar 24 00:29:30 RP: don't tell anyone that it got faster, I want to see if people recognise it is faster Mar 24 00:30:28 zecke: Its updated manually and is a nightmare. I can't afford to break builds so I have to be very careful Mar 24 00:32:30 at ROAD I used to create debs Mar 24 00:34:43 svn works quite well. It leaves all the nightmares for me :-/ Mar 24 00:35:32 bitbake is now really about ~60 faster when parsing (that wouldn't hold in court...) Mar 24 00:35:59 I'm trying it out :) Mar 24 00:36:27 zecke: 60x faster? 60s faster? 60% faster? hoping for the first :-) Mar 24 00:36:56 60% Mar 24 00:37:20 that's still a lot. good work! Mar 24 00:37:34 mreimer: loading from cache is now at 24 seconds, it used to be roughly 59s to 1.1m Mar 24 00:37:42 very good Mar 24 00:38:00 mreimer: and parsing the complete set is one minute faster Mar 24 00:38:02 (on ewi) Mar 24 00:38:08 hey zecke, do you know what causes this? Mar 24 00:38:10 so it could be even faster on other machines Mar 24 00:38:11 ./configure: line 10881: syntax error near unexpected token `1.2' Mar 24 00:38:11 ./configure: line 10881: `GTK_DOC_CHECK(1.2)' Mar 24 00:38:27 emte: you need some gtk.m4's ;) Mar 24 00:38:29 i am guessing an autotool mismatch Mar 24 00:38:44 emte: or a mismatch Mar 24 00:38:46 good nite! Mar 24 00:38:53 no, it errors an all things defined in that fashion Mar 24 00:39:00 night Mar 24 00:39:04 'night zecke - nice work :) Mar 24 00:39:23 RP: unpickling takes 17s of the 23s seconds Mar 24 00:39:36 zecke: cpickle? Mar 24 00:39:57 RP: the other 8 (something is wrong with my math) are spend in expanding python expressions Mar 24 00:40:10 mreimer: at least it tries to use it, and fallsback to stock pickle Mar 24 00:40:22 zecke: That might still leave some room for optimisation then :) Mar 24 00:41:03 RP: we could write a unserpickel or meinpickel which is better suited for OE's usage - but I doubt it is giving us anything Mar 24 00:41:17 RP: we should look into reducing the data we store Mar 24 00:41:31 zecke: Agreed, that's likely to be more productive Mar 24 00:42:15 RP: I think I will speed up glibc packaging now Mar 24 00:42:24 'now' as in the next couple of month Mar 24 00:42:26 zecke: yay :) Mar 24 00:42:30 NOTE: exceptions.SyntaxError:EOL while scanning single-quoted string (line 1) while evaluating: - bunch of those Mar 24 00:42:43 TheMasterMind1: on dvb stuff? Mar 24 00:42:57 quiten them is on my todo :} Mar 24 00:43:02 okay cya tomorrow Mar 24 00:54:01 hrm won't build anything, can't satisfy diffstat-native Mar 24 02:46:00 i686-oe-linux-uclibc/bits/ctype_noninline.h:121: error: cannot convert 'const __ctype_touplow_t*' to 'const int*' in assignment Mar 24 02:59:50 hrm why is bitbake building gcc 4.0.1 when 4.1 is available Mar 24 03:03:10 why does bitbake do any of the things it does? Mar 24 03:03:27 why does it insist on building glibc-initial for a powerpc toolchain? Mar 24 03:03:44 nature of the beast Mar 24 03:04:10 you need to compile a native version of glibc then compile a cross version adn finally use taht to compile a target version Mar 24 03:04:41 you have to build the tools before you can build the tools Mar 24 03:04:53 chicken-egg syndrom Mar 24 03:06:13 emte, well, I don't wanna be rude, but the arm toolchain for nslu2 does not build a glibc-initial Mar 24 03:06:19 TheMasterMind1, you probably need to explicitly specify 4.1 Mar 24 03:06:22 I know this for a fact Mar 24 03:06:57 nslu2 has quite a few non-standard customisations from what i gather Mar 24 03:07:25 bitbake/OE is a generic buildsystem trying to encompass all variables Mar 24 03:08:11 ex building for x86 on an x86 host will do the same as compiling for arm on an x86 host Mar 24 03:08:33 unless you tell bitbake to behave differently Mar 24 03:09:39 correct me if i am wrong but nslu2 already has an emulator in place which would hlep skip some crosstool stages Mar 24 03:09:43 help* Mar 24 03:09:58 qemu has only been introduced into OE in the last few weeks Mar 24 03:14:06 that's cool about qemu - I'll have to check that out, but nslu2 doesn't use that Mar 24 03:15:05 there are other emulators besides qemu Mar 24 04:30:08 emte: what are the other emulators? Mar 24 04:37:02 what's the status of ipkg Mar 24 04:37:12 ever get rewritten to not suck so much? Mar 24 04:52:16 TheMasterMind1, your better off to search and investigate for yourself Mar 24 04:52:40 there are many emulators, some with more device supoprt than others Mar 24 04:53:00 and there is very little worng with ipkg Mar 24 04:53:21 the problem is/are the interfaces written for it Mar 24 05:01:12 emte: there used to be a ton wrong with ipkg, but that was almost 4 years ago. looking at the changelog now, looks like a lot of progress since then Mar 24 06:07:56 run.do_patchcleancmd.15255: line 455: syntax error near unexpected token `'INHIBIT_PACKAGE_STRIP',' Mar 24 06:08:03 any ideas? Mar 24 06:09:55 emte: rosters used to be in a branch (what use would we have for a patch? ;-)), but it's been merged into mainline for a few months now... Mar 24 06:46:32 morning Mar 24 06:46:44 morning grma Mar 24 07:27:58 hi Mar 24 07:32:12 morning all Mar 24 07:46:01 RP: how goes merging zaurus patches? Mar 24 07:50:28 good morning all Mar 24 07:50:38 RP: thanks for syncing poky <> OE Mar 24 07:53:20 morning all Mar 24 07:53:41 hey ade|desk Mar 24 07:54:53 hrw|work: Not enough time to sort anything much out yet :-/ Mar 24 07:55:53 koen: I'm not finished yet although most things are probably in OE now. I've ignored certain problem cases such as opensync as I don't know who's implementation is better :-/ Mar 24 07:56:02 hi ade|desk Mar 24 08:00:41 "Inge Arnesen" - someone know who he is? Mar 24 08:02:12 RP: and we need a policy on the MAINTAINER field Mar 24 08:02:38 RP: IMO every MAINTAINER must have push access Mar 24 08:02:44 koen: Yes. I'm aware I pushed some oh addresses Mar 24 08:03:25 koen: not necessary Mar 24 08:03:26 koen: If we state that, it just gives me more merging nightmares. Effectively people at OH do have access, just via me Mar 24 08:03:44 koen: we need maintainers and "sponsors" Mar 24 08:04:31 Joe Smith maintain umbaumba in OE and Hank Gar sponsors Joe by pushing his changes to umbaumba Mar 24 08:04:59 that should be in the .bb Mar 24 08:06:09 I really don't want to have to do a serach and replace on the MAINTAINER field when merging packages to/from OE as I have enough problems as it is... Mar 24 08:08:04 hi all Mar 24 08:08:39 koen: if every maintainer will have access to repo who will work on fixing their bad NMC? (NMC = Non Maintaner Commit) Mar 24 08:08:43 hi alan Mar 24 08:09:03 hi hrw Mar 24 08:09:42 hrw|work: what's an NMC? Mar 24 08:09:50 Non Maintaner Commit Mar 24 08:10:09 what's an 'Non Maintaner Commit'? Mar 24 08:10:37 any changeset which touch stuff maintained by somebody else Mar 24 08:10:49 if I change your package then it is NMC Mar 24 08:11:20 you can disapprove this, you can shout at me that it is your package and I should not even think about touching it etc Mar 24 08:12:40 I see Mar 24 08:13:31 think about jnc changes for example Mar 24 08:14:13 if everyone get r/w access then someone will have to work only on checking does all changes are ok Mar 24 08:14:45 Personally, I see the MAINTAINER as someone who is prepared to look after the package, be that take bugs, apply patches etc. - someone to talk to about the package. If they have indirect commit access but are active, I don't see a problem Mar 24 08:15:05 fuck i'm hungry. wonder if they're still serving dinner.. Mar 24 08:15:27 RP: thats why I also like Mar 24 08:16:03 TheMasterMind1: you can find a restaurant serving breakfast, here, but i guess it is quite far from where you are... ^^ Mar 24 08:16:42 heh. we have a diner type thing about 10 yards from me on campus, but they stop serving dinner and start breakfast at some point. they also close at 4am Mar 24 08:16:47 and its cold out. Mar 24 08:17:32 anyone done any [modular] xorg7 stuff with OE? Mar 24 08:17:48 TheMasterMind1: we have xorg7 in OE iirc Mar 24 08:17:53 the ideal and unfeasable solution is that each package would have its own branch with maintiner control of that package and regular merges into a main trunk ... Mar 24 08:19:04 <_law_> if i try to build a gpe-image from .dev i get the following error http://pastebin.com/619578 Mar 24 08:19:23 <_law_> are the hostap-modules really required? Mar 24 08:19:35 RP: about opensync: OE has newer (0.18) packages, which I know that they work, because I tested them myself :) Mar 24 08:20:15 koen: right, so I really need to sync that into poky. I need to run that past chris first though Mar 24 08:20:39 koen: I'm trying to deal with the xlibs atm... Mar 24 08:21:02 instead of more senseless comment by me i think i'll go to sleep Mar 24 08:21:47 hrm there's an xorg-xserver_cvs bb that looks like it should work Mar 24 08:23:06 morning all Mar 24 08:23:09 hey do13_ Mar 24 08:23:45 hi koen Mar 24 08:24:00 hi Dirk Mar 24 08:24:12 how do you force a version from commandline bitbake? bitbake =package_ver work like it does with emerge? Mar 24 08:24:52 TheMasterMind1: I prefer 'bitbake -i' (bitbake shell) where you can do it easier Mar 24 08:24:57 hey Marcin Mar 24 08:24:58 'build package-version' Mar 24 08:25:23 sweet Mar 24 08:25:25 03hrw 07org.oe.oz354x * r747076e6... 10/packages/groff/groff_1.19.1.bb: Mar 24 08:25:25 groff: fixed paths so manpages are packaged Mar 24 08:25:25 - removed not-needed RDEPENDS on libstdc++ Mar 24 08:25:25 - disabled PARALLEL_MAKE Mar 24 08:26:44 oh wow. it doesn't have to re-parse cache everytime in interactive mode Mar 24 08:28:42 Why does gnome-vfs-dbus DEPEND on samba? Mar 24 08:29:56 to use smb:// in apps Mar 24 08:31:01 I guess that makes sense :) Mar 24 08:33:41 hrw|work: hrm, i'm trying to build xserver-xorg cvs Mar 24 08:33:48 RP: but on a gnome-vfs-module ipkg depends on samba Mar 24 08:33:53 s/on/only/ Mar 24 08:33:56 doesnt work when i do package-version Mar 24 08:34:01 shows up in which though Mar 24 08:34:12 so you won't see gnome-vfs RDEPENDING on samba Mar 24 08:34:30 <_law_> hrw|work, the matchbox-keyboard-collie-stuff into .oz354x ? Mar 24 08:34:46 koen: Yes, I realise that. Its just a headache when samba isn't in your repository :) Mar 24 08:34:53 _law_: dont remember Mar 24 08:35:11 RP: use 3.0.14 to avoid some api mismatches Mar 24 08:35:47 <_law_> hrw|work, that /etc/matchbox/kbd stuff (you have removed collie specific keys) afer that ctrl and alt should work Mar 24 08:37:13 _law_: that mapping ctrl to calendar, alt to contacts thing? Mar 24 08:38:09 <_law_> hrw|work, yes Mar 24 08:38:23 RP: that new gtk requires glib 2.10.1, which isn't in OE Mar 24 08:40:10 koen: I'll merge along with a newer pango Mar 24 08:40:12 RP: another thing: it's bad form to do 'cp foo_2.6.bb foo_2.8.bb' without zapping the maintainer Mar 24 08:40:54 koen: I'll try and educate the peole within OH. I zap it when I spot it Mar 24 08:41:18 notifying the maintainer would also be ok I guess Mar 24 08:46:16 morning Mar 24 08:47:10 hey XorA Mar 24 08:47:21 hey koen I saw some signs of success last night Mar 24 08:47:50 XorA: you did? Mar 24 08:47:59 koen: you and ph5 last night Mar 24 08:48:11 ah yes Mar 24 08:48:20 we solved the socket() problem Mar 24 08:48:52 but I suspect we need some further patches from the csl gang to make it actually work Mar 24 08:49:00 (or wait for glibc 2.4.1) Mar 24 08:49:28 if glibc like the kernel used to be then, release x86 then let the platforms catch up Mar 24 08:49:33 s/if/is/ Mar 24 08:53:13 XorA: the problem is that the CSL people don't really care about upstream Mar 24 08:53:36 "we have patches in our CSL toolchains for that, which we will release soon" type of shit Mar 24 08:53:43 :-( Mar 24 08:54:34 and seeing the previous csl-arm branch stuff in OE, I'll be happy to stick with the upstream releases Mar 24 08:56:51 csl does give stuff upstream all the time Mar 24 08:57:31 just seems like they are giving more uclibc stuff up to gcc atm Mar 24 08:57:44 ade|desk: but everytime they do they 'forget' patches Mar 24 08:57:55 no idea if the 'forgetting' is on purpose or not Mar 24 08:58:15 i dont think it is on purpose Mar 24 08:58:36 certainly get more stuff from them than mvista Mar 24 08:58:58 heh Mar 24 08:59:02 ~kill mvista Mar 24 08:59:05 * ibot shoots a inverse anti-meson gun at mvista Mar 24 09:10:47 I would like to create a makefile for a shared library I'm makeing... pretty new at this so don't die on me ;) I need to spesify in this makefile first a compilation routing, and then a library creating routine? And what about the macros? Don't I need to add some -shared etc for my $CFLAGS ? Mar 24 09:12:05 I've been told these macros are previously defined, but where is it determined wich to use for a shard library and wich to use for a normal application ? Is it at all? Mar 24 09:22:45 hi guys ! Mar 24 09:26:44 hey grma Mar 24 09:27:33 hi koen Mar 24 09:28:03 i still don't have all *.tar.gz files from all packages... Mar 24 09:29:24 in the ipk dir there are over 1000 *.ipk files and in the tar dir there are only 500, so i tlooks like that the toolchain and libs are build as *.tar.gz but not meta-opie and and not meta-gpe, any solltution ? Mar 24 09:31:36 grma: no idea on that, sorry Mar 24 09:31:42 ok Mar 24 09:38:38 hi Mar 24 09:41:50 hi mickey|saks06 Mar 24 09:43:02 mickeyl: you saks? Mar 24 09:43:08 saks fonetic Mar 24 09:47:17 heh Mar 24 09:54:05 french people speaking english giving me headaches Mar 24 09:54:27 hey mickey|saks06 Mar 24 09:54:30 hi koen Mar 24 09:54:44 hi mickey|saks06 Mar 24 09:54:59 anybody attending SPS/DART next week? Mar 24 09:55:14 03rpurdie 07org.oe.dev * r45e00484... 10/packages/glib-2.0/glib-2.0_2.10.1.bb: Add glib-2.0 2.10.1 Mar 24 09:55:14 03rpurdie 07org.oe.dev * r0611ba84... 10/conf/documentation.conf: documentation.conf: Add PCMCIA_MANAGER Mar 24 09:55:18 hi RP. Mar 24 09:55:24 03rpurdie 07org.oe.dev * rbcb91ddb... 10/packages/pango/pango_1.12.0.bb: Add pango 1.12.0 Mar 24 10:07:20 mickey|saks06: what about Polish people speaking english? Mar 24 10:08:24 interestingly, they're not bad Mar 24 10:08:31 at least those who i've heard Mar 24 10:24:01 ~lart tosa updater.sh for lack of zImage.bin Mar 24 10:24:02 * ibot puts on some milking gloves. "All right, now, tosa updater.sh, this won't hurt a bit...." for lack of zImage.bin Mar 24 10:24:40 03mickeyl 07org.oe.dev * rff499565... 10/packages/uicmoc/ (uicmoc-native-2.3.10/gcc4_1.patch uicmoc-native_2.3.10.bb): Mar 24 10:24:40 uicmoc-native 2.3.10: apply patch to make it work with gcc 4.1 Mar 24 10:24:40 fixes bug #771 Mar 24 10:27:13 Hey Liam Mar 24 10:27:52 hey Dirk Mar 24 10:29:09 lrg: Does the image works? Mar 24 10:31:52 do13: I've not had a chance to try it yet. I'll be able to do so this weekend. Mar 24 10:33:05 lrg: ok Mar 24 10:35:46 03hrw 07org.oe.oz354x * rc89c1106... 10/packages/zaurus-updater/ (tosa/updater.sh zaurus-updater.bb): zaurus-updater: added zImage.bin to tosa updater.sh Mar 24 10:36:15 ~lart updater.sh idea Mar 24 10:36:15 * ibot overclocks updater.sh idea until updater.sh idea burns out Mar 24 10:37:06 RP my opie-image wants to build hostap-modules but I can't see where this dependency comes from. do you have any idea? Mar 24 10:37:12 bitbake -nv shows Mar 24 10:37:38 heh Mar 24 10:37:48 task-bootstrap -- hostap-modules (hostap-conf) Mar 24 10:37:52 RP: I'm pretty sure the nokia770 doesn't need pcmcia-utils ;) Mar 24 10:37:53 what's that supposed to mean? Mar 24 10:38:31 this is for c7x0 with DISTRO=angstrom-2006.9 btw. Mar 24 10:39:03 mickey|saks06: it seems hostap-conf depends on the external modules Mar 24 10:39:03 koen: At the moment, the nokia770 images depend on it. Fix that and we can fix the other issues Mar 24 10:39:28 koen: i can't find that in hostap-conf.bb though Mar 24 10:39:45 mickey|saks06: bitbake -n -D -D -D > logfile and search for hostap-modules Mar 24 10:40:32 mickey|saks06: At a guess set PREFERRED_PROVIDER_hostap-conf = "hostap-conf" Mar 24 10:41:09 For some reason it prefers hostap-modules over hostap-conf by default Mar 24 10:47:09 RP: I think I fixed the 770 stuff Mar 24 10:47:31 | + ipkg-cl -f /oe/tmp/work/bootstrap-image-1.0-r0/temp/ipkg.conf -o /oe/tmp/rootfs update Mar 24 10:47:31 | Updated list of available packages in /oe/tmp/rootfs//oe/tmp/rootfs//usr/lib/ipkg/lists/oe Mar 24 10:47:44 its repeating the first part of the path twice and putting the ipkg list in the wrong place =\ Mar 24 10:47:54 koen: ah, ok. We can remove pcmcia then Mar 24 10:49:18 03koen 07org.oe.dev * r2d22f5a5... 10/conf/machine/nokia770.conf: Mar 24 10:49:18 nokia770.conf: remove cruft Mar 24 10:49:18 * GPE_EXTRA_INSTALL should be in a distro.conf -> removed Mar 24 10:49:18 * handheld-common is largely distro related -> removed Mar 24 10:49:18 Users will be still able to 'bootstrap' without the above, thanks to the initfs present in the 770. Mar 24 10:51:59 koen: Keep in mind that locally some of us don't use the initfs... Mar 24 10:53:21 RP: OE assumus you have the initfs Mar 24 10:53:27 assumes even Mar 24 10:53:57 the 770 kernel isn't in OE yet Mar 24 10:54:08 patches welcome :) Mar 24 10:54:27 RP: wow, that fixed it. Mar 24 10:59:38 03koen 07org.oe.dev * r879120a7... 10/conf/distro/angstrom-2006.9.conf: angstrom 2006.9: add PREFERRED_PROVIDER_hostap-conf = "hostap-conf" **** ENDING LOGGING AT Fri Mar 24 10:59:56 2006