**** BEGIN LOGGING AT Tue May 02 09:59:56 2006 May 02 10:04:56 zecke: what's the Survey? May 02 10:05:02 * pb__ baffled May 02 10:05:23 pb__: sorry, you are not allowed ;) May 02 10:06:34 drat May 02 10:06:59 pb__: your contact with the force must be heavily disturbed ;) May 02 10:13:26 hi May 02 10:13:34 hi May 02 10:17:04 booting my SL6000, after flashing it with a fresh openzaurus-unstable kernel 2.6 build, it tells me "Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0), did someone experienced the same or has an idea of how to work around ? May 02 10:19:51 morning all May 02 10:19:56 hi dirk May 02 10:20:06 hey Marcin May 02 10:20:09 sws: ask again May 02 10:20:14 hey do13_ May 02 10:20:21 hi koen May 02 10:21:15 sws: I think the kernel CMDLINE is a bit wrong May 02 10:21:31 sws: i suspect it points to /dev/ram0 or something May 02 10:21:44 sws: if not, do13_ will probably know the answer :) May 02 10:22:00 booting my SL6000, after flashing it with a fresh openzaurus-unstable kernel 2.6 build, it tells me "Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0), did someone experienced the same or has an idea of how to work around ? May 02 10:22:38 sws: are you using gcc4? May 02 10:23:21 do13_: no, using gcc 3.3.5 May 02 10:23:59 do13_: on my host? else OE is using gcc 4.0.2 I think May 02 10:24:40 When I try to compile trai I get the following errors May 02 10:24:46 does some have an Idea? May 02 10:24:47 http://pastebin.de/6377 May 02 10:25:07 sws: the tosa nand driver has a issue with gcc4. so the nand flash won't work. May 02 10:25:24 Cobelius: yes? May 02 10:25:31 http://pastebin.de/6378 May 02 10:25:34 sws: does the driver detect the flash? May 02 10:25:39 this one May 02 10:25:45 I pasted the wrong stuff May 02 10:26:03 Cobelius: what did the FAQ say? May 02 10:26:10 Cobelius: I fear it defined __KERNEL__ or such May 02 10:26:37 hrw: nice work on rc2 May 02 10:26:45 do13_: which driver? all i see after reseting is the above Message May 02 10:27:33 sws: do you see "No NAND device found!!!" in the log? May 02 10:27:41 the FAQ says nothing useable May 02 10:28:09 Cobelius: I suspect your package is using the libc-headers instead of the kernel headers May 02 10:28:13 good morning May 02 10:28:17 do13_: sorry have overseen your nand driver comment, which log do u mean? May 02 10:28:22 Cobelius: I don't know the solution to that May 02 10:28:22 morning May 02 10:28:24 hey Bernardo May 02 10:28:33 hey Bernardo May 02 10:28:37 koen: thx May 02 10:28:52 do13_: kernel build log? May 02 10:28:53 sws: the boot message. are you using an serial lead? May 02 10:29:17 I'm using ${STAGING_KERNEL_DIR} May 02 10:29:21 sws: not the build log May 02 10:29:34 does there exist any other variable to these header files? May 02 10:31:24 do13_: no i'm not using a serial lead, just flashed the zaurus using cf and reset after it said reset May 02 10:31:33 qemu builds ok with gcc 3.4? May 02 10:31:48 ahhh May 02 10:31:54 or do I need 3.3? May 02 10:31:55 hrw: is should do May 02 10:31:57 Cobelius: is that a module? May 02 10:31:59 it even May 02 10:32:03 Cobelius: check what the modules use ;) May 02 10:32:10 ade|desk: thx. installing 3.4 then May 02 10:32:22 do13_: do u think the issue is with oe using gcc 4? May 02 10:32:30 hmn May 02 10:32:36 sws: yep. the nand isn't detected May 02 10:32:38 sws: for 2.6 on !clamshells I suggest not trying without serial cable May 02 10:32:40 i found some header files that seems to be ok May 02 10:32:47 hrw: its only gcc4 that has the problems May 02 10:33:07 ade|desk: or rather its only qemu having problems with gcc4 May 02 10:33:17 hrw: thanks for the suggestion, so i have to go buy a serial cable May 02 10:33:29 I just have to find the undocumenten variable which contains the path to them May 02 10:36:05 hrw: yeah,ok bad bad qemu .. good good gcc4 May 02 10:36:22 well gcc4.1 is better May 02 10:36:56 gcc 4.1 rocks on x86 May 02 10:37:13 hrw@home:~/devel/build/oz3541$ file tmp/cache/bb_cache.dat May 02 10:37:13 tmp/cache/bb_cache.dat: 8086 relocatable (Microsoft) May 02 10:37:15 ;P May 02 10:38:18 didn't kergoth build an arm-wince-pe toolchain with OE a while ago? May 02 10:40:16 can i still use gcc3 instead of gcc4 to compile OE? May 02 10:40:43 okay May 02 10:40:54 there are 2 time similar header files in the staging dir May 02 10:41:22 koen: wiki for OZ was good decision May 02 10:41:37 one set is under staging/i586-linux/include/linux May 02 10:41:38 hrw: indeed May 02 10:41:40 they are not working May 02 10:41:47 hrw: although berlios wasn't ;) May 02 10:41:53 * koen needs more coffee May 02 10:42:02 'the him' May 02 10:42:17 the other one are unter i586-linux/kernel/include/linux May 02 10:42:19 actually, I need breakfast May 02 10:42:22 I think they should work May 02 10:42:31 how can I say that this header file are be used? May 02 10:43:00 Cobelius: staging/include/linux is linux-libc-headers May 02 10:43:13 Cobelius: staging/kernel/include/ is what you need (headers from kernel) May 02 10:43:44 Cobelius: ${KERNEL_SOURCE} should work May 02 10:43:51 NOTE: Machine ID assigned by tinderbox: 19 May 02 10:44:04 STAGING_KERNEL_DIR iirc was proper May 02 10:44:14 hrw: 19? May 02 10:44:16 export KERNEL_SOURCE = "${@base_read_file('${STAGING_KERNEL_DIR}/kernel-source')}" May 02 10:44:16 I think this has somethink to do with the include path of the envorionment May 02 10:44:25 hrw: do you want to be a sheriff/admin as well? May 02 10:44:34 (from classes/module-base.bbclass) May 02 10:44:55 zecke: I do not understand tinderbox yet May 02 10:44:55 zecke: how 'themeable' is tinderbox? May 02 10:45:01 the configure script needs the kernel source do get som information May 02 10:45:07 this things are wokring May 02 10:45:20 koen: very :) May 02 10:45:21 but while it tries to compile it there comes these error May 02 10:45:32 zecke: it would be nice if we used the same css for all OE site (survey, bugs, tinderbox, viewmtn, www, etc) May 02 10:45:52 koen: well, we have a lot of work with tbox3 ahead May 02 10:45:59 koen: but one step after another May 02 10:46:22 zecke: possibility to select one build to watch would be nice May 02 10:46:24 zecke: I tend to go for the easy tasks :) May 02 10:46:48 koen: they are all eady ;) May 02 10:46:50 zecke: can you kick out build#18? May 02 10:51:34 NOTE: package qemu-native-0.7.0: started May 02 10:51:46 3.5.4.1 will get binary locales May 02 10:52:43 hrw: 'kick'? May 02 10:52:55 hrw: 'kick' means remotely restarting the build, I can't ATM May 02 10:53:00 zecke: ah.. May 02 10:53:10 zecke: I meant kick as 'remove' May 02 10:53:34 http://openzaurus.berlios.de/ReleaseNotes/Oz3541 time to start writing May 02 10:53:40 hrw: what is the name? May 02 10:55:30 What was the command to create a patch file? May 02 10:55:34 zecke: ewi-OZ-3.5.4.1-rc2 Linux 2.6.16-ck1 May 02 10:55:39 Cobelius: diff? May 02 10:55:44 yes May 02 10:56:25 which qemu I need for binary locales? May 02 10:57:42 hrw: any of the two May 02 10:57:48 qemu-native-0.7.0 should be fine May 02 10:58:05 it has this inexplicable segfault on redhat, but the cvs version suffers from that too May 02 10:58:25 yesterday bitbake selected the nogfx-foo version May 02 10:58:30 that worked as well on my debian... May 02 10:58:44 pb__: what do you expect on redhat? was it compiled with gcc 4.96? May 02 10:59:09 oddly, no May 02 10:59:09 gcc version 3.3.2 20031022 (Red Hat Linux 3.3.2-1) May 02 10:59:36 I guess something about RH's glibc and/or gcc must be upsetting it, but I don't know which May 02 10:59:44 actually, thinking about it more, it must be glibc not gcc May 02 10:59:55 if I compile qemu on redhat and then run that binary on debian, it works fine May 02 11:00:15 pb__: lol, so glibc RH has more drepper included? May 02 11:01:25 someone has osb-nrcit-0.5.0 cvs20060212 snapshot avaiable? May 02 11:01:39 the output of diff is looking different from the othe patch files in the project May 02 11:02:12 do13_: http://ewi546.ewi.utwente.nl/mirror/www.openzaurus.org/official/unstable/3.5.4/sources/ May 02 11:02:21 hrw: thx May 02 11:02:23 do13_: http://ewi546.ewi.utwente.nl/tmp/sources/ May 02 11:02:27 * koen too slow May 02 11:02:29 zecke: I suppose it must May 02 11:02:43 do13_: you can add it into CVS_TARBALL_STASH May 02 11:02:43 koen: thx :) May 02 11:02:45 perhaps Uli has some kind of grudge against the qemu developers May 02 11:03:25 pb__: does 0.8.0-native work on RH? May 02 11:04:20 pb__: I propose, you install lsbX.Y.Z May 02 11:04:30 pb__: compile qemu against it, run it on RH May 02 11:04:47 pb__: and ping Uli that qemu LSB is not running on a RH LSB certified system May 02 11:05:33 ~seen bero May 02 11:05:40 bero <~bero@zux176-164.adsl.green.ch> was last seen on IRC in channel #familiar, 271d 1h 54m 21s ago, saying: '| ${PIVOTBOOT_EXTRA_RDEPENDS}'. May 02 11:05:48 koen: he is online May 02 11:06:04 koen: msg him or go to #arklinux May 02 11:06:08 Okay, how can I make patchfiles in this format like all patch files are in OpenEmbedded? May 02 11:06:28 Cobelius: we use quilt for it May 02 11:06:37 Cobelius: cd work/foo-X-Y-Z/foo May 02 11:06:47 Cobelius: quilt new my-funky-patch.patch May 02 11:06:58 Cobelius: quilt edit the/file/to/patch.cs May 02 11:07:02 Cobelius: quilt refresh May 02 11:07:11 Cobelius: vim patches/my-funky-patch.patch May 02 11:10:35 OKay now imageine I have 2 files. One is the original one, the antother one i edited my me. How do I make a patch out of them. I know there was a command, but I can't remember May 02 11:10:58 LOL: http://angryflower.com/bobsqu.gif May 02 11:11:17 Cobelius: diff? May 02 11:11:52 yes, but I don't find the right option May 02 11:12:14 diff -Nurd? May 02 11:12:36 yes May 02 11:12:45 This is the one I was looking for May 02 11:12:47 thank you May 02 11:14:12 it if it C code then "-p" can be useful May 02 11:14:30 it's a shell script May 02 11:15:08 -p is ignore whitespace ? May 02 11:15:25 koen: -p --show-c-function Show which C function each change is in. May 02 11:15:40 koen: -w --ignore-all-space Ignore all white space. May 02 11:16:06 * koen wget http://eng.iaudio.com/download/gnu/cowon_a2_bsp_source.tar.gz May 02 11:16:14 I wonder if they use directfb May 02 11:26:10 re May 02 11:26:14 again keyboard died in x11 May 02 11:36:58 Hmn May 02 11:37:03 it still don't work May 02 11:37:10 I also tried "inherit module" May 02 11:37:23 because some of the stuff in the package are modules May 02 11:38:00 Cobelius: you include the wrong headers :) May 02 11:38:02 Does somebody know a working package of a module in which I can look? May 02 11:38:21 zecke: I already know that. But I don't know how to fix it May 02 11:38:50 Cobelius: do you inherit module-base? May 02 11:39:10 I tried this before but it doesn't helped May 02 11:40:16 zecke: Do you know how to solve this problem? May 02 11:40:39 not knowin rtai, sadly not May 02 11:43:03 Howdy! May 02 11:43:25 hmn May 02 11:43:42 Is there something like a variable where I can configure the Inlucde Path? May 02 11:44:14 in the crated do_compile script are exports like CPPFLAGE="-I/var/...." May 02 11:44:16 hey Laibsch May 02 11:44:23 how can I manipulate this? May 02 11:44:39 Because this statements are direct to the wrong directory May 02 11:46:29 cu May 02 11:47:00 Cobelius: use quilt for editing May 02 11:47:07 Cobelius: otherwise edit the file again, diff again May 02 11:57:03 That was not about the patch problem. May 02 11:57:13 It was about the Problem of the not compiling rtai May 02 11:59:54 hey Liam May 02 12:00:00 hey lrg May 02 12:00:06 hey Dirk, koen May 02 12:00:22 lrg: did you received my mail? May 02 12:00:52 do13, yep, no problems - I'll reply today May 02 12:00:53 hey lrg May 02 12:01:26 hey XorA, I'll hopefully get a chance to fix your audio soon. I've a customer using the same chip May 02 12:02:03 lrg: you still going to be at lug? May 02 12:02:31 XorA, yes. This thursday - bring your Z and I'll have a look May 02 12:02:33 lrg: Ive never tried to record on c7x0 so never seen any bugs :-) May 02 12:03:03 XorA, I'll bring a mic and my iPod and we can try recording. May 02 12:03:18 lrg: ok, Ill have to make sure arecord is installed then May 02 12:03:28 cya May 02 12:03:30 XorA, cool May 02 12:04:12 * XorA will have to once again brave the depths of gothdom :-) May 02 12:04:28 XorA, it's not that bad ;) May 02 12:05:07 lrg: heh heh, there have been some interesting times in holyrood, like tranny night :-) May 02 12:05:51 XorA, ooer - I didn't know that ! - best be careful then :-) May 02 12:06:06 lrg: that was back in 93/94 May 02 12:15:16 hey woglinde May 02 12:15:31 woglinde: did you compile glibc for the msp yet? ;) May 02 12:15:58 koen no May 02 12:16:29 koen we have to discuss about new variables for thread/no-thread and static/shared May 02 12:17:00 and I have to rework the libgcc install May 02 12:17:07 for this cases May 02 12:17:09 I've been thinking about doing static builds for some packages May 02 12:17:16 e.g. the angstrom-installer package May 02 12:17:52 but I found a solution not to break the actual configure May 02 12:18:31 default should be shared and threaded and you can changed it with the variables May 02 12:18:48 for instance in the machine.conf May 02 12:19:06 but I have to do some homeworks May 02 12:19:18 dont know if can work on it today May 02 12:20:02 homework... May 02 12:20:04 good point May 02 12:20:48 yes it's about threaded processes and atomic operations and so on May 02 12:39:29 morning May 02 12:40:43 Is wxpython too heavy to ever make it into OE (at least for the collie)? May 02 12:40:48 hi chouimat May 02 12:40:56 hi chouimat May 02 12:46:03 Laibsch: it isn't May 02 12:46:08 Laibsch: someone need to add it May 02 12:46:16 Laibsch: we already have wxbase in oe May 02 12:46:31 koen: Cool. May 02 12:47:17 Then it is an option for a python GUI. I will to evaluate that further. May 02 13:07:38 hey zecke May 02 13:08:25 hey, and cya later ;) May 02 13:08:34 hrw|gone: I have some things for you May 02 13:10:34 cya May 02 13:59:03 <_guillermo> is monotone 0.26 supported by oe? May 02 13:59:30 _guillermo: packages/monotone/monotone-6_0.26.bb May 02 13:59:47 _guillermo: so 'bitbake monotone-6' should get you the binary May 02 13:59:50 <_guillermo> I mean the repositry May 02 14:00:06 <_guillermo> repository I mean, it I can use monotone 0.26 with no problems May 02 14:01:45 you can't May 02 14:02:59 <_guillermo> ok May 02 14:03:56 hrw|gone: I have something for the install docs: http://dominion.kabel.utwente.nl/koen/images/install/ May 02 14:05:39 hrw|gone: for people who can't read english or japanese ;) May 02 14:07:25 (yes, they are crappy) May 02 14:46:26 hi May 02 14:56:43 * CoreDump|home got populate-volatile.sh down to 1.3s with "caching" May 02 14:57:15 on an idle system that is May 02 14:58:39 what kind of caching? May 02 14:59:25 caching the mkdir / touch / ln -s commands generated by the scripts logic on running that after the first run created the "cache" May 02 14:59:50 meh, what a sentence May 02 15:00:24 CoreDump|home: can't you just tar up the result of all those commands and then unpack it on subsequent runs? May 02 15:00:30 that sounds a better plan than caching the commands themselves May 02 15:01:02 in theory yeah, but populate-volatile is flexible enough to mess with all kinds of files, even outside /var if desired May 02 15:01:26 Birth is the leading cause of death. So far, it has a 100% fatality rate. ;) May 02 15:02:26 boot-time down ~12s May 02 15:02:55 that said, it would be even better to just generate the tarfile offline at build time. May 02 15:02:55 mithro: oxygen is poisonous as well, fatal in concentrations of 0,00000000000000001% May 02 15:03:08 does the result of populate-volatile vary much from one machine to another? May 02 15:03:10 pb_: we did that ages ago May 02 15:03:36 oh, very good May 02 15:03:40 why is it still an issue then? May 02 15:03:44 there has to be _some_ reason we did away with that May 02 15:04:03 true, though not necessarily a good reason May 02 15:04:12 * CoreDump|home dunnos May 02 15:04:14 it might have been some well-intentioned but ultimately misguided change May 02 15:04:58 well, the script is quite flexible. Changes are easy etc. May 02 15:05:01 * CoreDump|home shrugs May 02 15:05:08 or, conceivably, some DISTROs need the full populate-var thing but others don't. May 02 15:05:27 I'm pretty sure we don't need that flexibility in familiar, for example, and I doubt you need it in openzaurus May 02 15:06:17 ibot, botmail for reenoo: I flashed 0.8.4-rc2 onto my hx4700 and the keymap does indeed seem wrong. haven't yet figured out why but will keep you posted. May 02 15:06:50 true, there is one single link outside of var (/etc/resolv.conf)...and IMO that shouldn't even be a link, causes all kind of annyoing problems during development May 02 15:07:25 I definitely want /etc/resolv.conf to be a link in Familiar. May 02 15:07:58 I can see that this might not be universally true though. May 02 15:08:37 http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=75cf64ddd3f313d225930408f46ff10f7208d0cb May 02 15:08:41 it's not written that often and the flash should care about not always using the same blocks. What's the reason of it beeing a link? May 02 15:08:58 CoreDump|home: every ten minutes on my uni wlan...... May 02 15:09:06 it prevents you from mounting / readonly, and it does get written fairly frequently in some environments. May 02 15:09:07 hmm May 02 15:09:12 CoreDump|home: it depends on your dhcp lease time and dns settings May 02 15:09:17 udhcpc is not terribly careful about that May 02 15:09:46 ah thanks, I didn't think about crazy lease times or RO rootfs =) May 02 15:11:30 ideally, the system should not ever be writing to flash except in response to a direct user request. Any time you have a background process that writes to volatile storage, it is almost always a bug. May 02 15:12:31 ideally, yeah May 02 15:14:46 how easy is it to override that behaviour on runtime? May 02 15:15:00 for root-on-sd, root-on-cf and cxxxx? May 02 15:15:13 the cache? rm /etc/volatile.cache May 02 15:15:33 koen: root-on-* wouldn't be affected by it May 02 15:16:33 I meant disabling the whole volatile stuff May 02 15:17:05 i am a bit confused here regarding this whole branch divergence issue. Developers now need to have all 4 branches and need to patch against all 4 and recommit to each branch? May 02 15:17:07 none, ignore it. Costs 1.3s on boot =) May 02 15:17:49 koen: the old script would run and see that all the links are already present (since /var is not a tmpfs there) May 02 15:19:26 well, the cache would probably throw errors, yes May 02 15:20:39 maybe a VAR in /etc/defaults/rcS (which is sourced by that script) could be used to deactivate it? May 02 15:21:39 OTOH you have to run it once at least, mehh May 02 15:22:37 emte: no, the point is that bugfixes should be in .dev as well, not just in org.oe.somebranch May 02 15:23:33 koen, i am just trying to figure out how you decide that a direct commit woulnt mangle they system due to other changes/patches taht may not exist in the original branch May 02 15:23:37 emte: some people commit to org.oe.somebranch *only* May 02 15:24:04 * emte thinks i need to fix my keyboard ... May 02 15:24:52 which is what branches are for, but stuff 'for the general good' should be in .dev as well May 02 15:25:44 so why not change policy and keep non-core packages out of branches? May 02 15:26:26 (this is based on me thinking it will only be core packages that have radical changes) May 02 15:26:28 it's not ours to say what people want in their branches May 02 15:27:00 who controls the SCM and branching then? May 02 15:27:12 people control their own branches May 02 15:27:25 you can push a org.oe.emte branch to server if you wish May 02 15:28:15 ah ... perhaps i'll propose that non-core stuff shouldnt be in branches on the mailing list then, as it seems it would cut down on issues May 02 15:29:00 emte: the problem is that when you commit something like "fix huge security hole in glibc" to org.oe.emte, which would apply cleanly to .dev as well and we ask "emte, could you apply that to .dev as well", you say "no, fuck .dev" May 02 15:29:33 which means one of the OE devs has to extract the diff an apply it .dev himself May 02 15:29:44 which is a problem with a lot of branches, with a lot of commits May 02 15:30:02 okay that is the case i was curious about, who decides/tests taht it will/would work without having multiple source repos? May 02 15:30:16 I don't get the question May 02 15:30:28 since glibc would be concidered a core package May 02 15:31:28 well if you fix it in a branch that was forked ... 8 moths ago and you exclusivly use that branch, in order to put it back in the main you would have to pull .dev apply it and test it May 02 15:31:40 moths/months* May 02 15:32:08 since you would not know if glibc had chnaged without reading 8months of logs May 02 15:33:14 ( on a side note i was wondering if a multiple diff tool and multiple branch commits might be useful for this ) May 02 15:33:18 suppose I wanted to compile something for the Z and statically link it May 02 15:33:42 how would I go about doing that? May 02 15:34:12 emte: that would be an issue, but in practise the commits are pretty simple (2-3 lines) and you can view the current repo online May 02 15:34:24 set the static flag in your extra flags? May 02 15:34:58 emte: let me be more clear. I have *no idea* how to cross-compile :) May 02 15:35:15 look at the bb metafile for your package May 02 15:35:45 but the static linking is important because I have no idea what libraries it needs, and suspect that even if I did they might not be available for the Z. May 02 15:35:57 what's a bb metafile, and this isn't from a package, it's from source. May 02 15:36:05 (I'm trying to put pork on my Z) May 02 15:36:06 a+ May 02 15:36:07 StarKruzr, if they are not availible your package will not compile ... May 02 15:36:19 StarKruzr: OE takes care of dependencies itself May 02 15:36:37 (the whole point of OE) May 02 15:36:45 koen, thanks for the info May 02 15:36:48 ... May 02 15:36:52 but May 02 15:37:52 ok, I really don't get it. this program is available as a .deb for Ubuntu and Debian, but that's obviously not going to help me because a) it's a .deb and b) it's for x86/PPC/whatever. yes? May 02 15:38:23 StarKruzr, have you read GettingStarted? May 02 15:38:28 ?????? May 02 15:38:59 where are my fishes May 02 15:39:06 StarKruzr, http://oe.handhelds.org/cgi-bin/moin.cgi/GettingStarted May 02 15:40:03 oh. May 02 15:40:10 well, this looks helpful :) May 02 15:40:52 but what if OE has never heard of the package I am making? May 02 15:41:06 StarKruzr, the branch you want is the one for the distro version your using May 02 15:41:13 then you add it May 02 15:41:16 OpenZaurus. May 02 15:41:18 starkruz you can kindly ask and if someone has the tim he will do it May 02 15:41:25 time May 02 15:41:34 most packages are very simple May 02 15:41:37 honestly, I'd rather help contribute to the project May 02 15:41:42 other require changes ... May 02 15:42:00 disclaimer: I'm functioning on almost no sleep and am somewhat muzzyheaded May 02 15:42:08 to start look at the packages dir in the repo for examples and get it to start compiling May 02 15:42:16 ok May 02 15:46:00 muhaha, pop-volatile down to 0.99s =) May 02 15:46:06 Ok - its time for matchbox and the Geode to get more friendly May 02 15:54:03 I have a bb file that complains abut some files that are installed but not shipped in any package May 02 15:54:13 I need them in the ipk May 02 15:54:23 what do I grep for in the bb files? May 02 15:54:26 FILES May 02 15:54:46 re May 02 15:54:47 maxr, I just asked May 02 15:54:55 ok May 02 15:55:18 grep for FILES in packages/*bb May 02 15:56:05 looks like my attempts to get binary locales in .oz354x are progressing.. May 02 15:56:12 NOTE: generating locale es_NI (UTF-8) May 02 15:56:16 finally ;) May 02 15:58:57 someone remember how much time does takes generating binary locales? few hours or eons? May 02 15:59:19 couple of hours, I think May 02 16:00:14 ok. so I will do something other in that time May 02 16:01:31 hrw: it depends on whether you're sitting around waiting for it, or if you're busy doing something else while they build May 02 16:01:44 hrw: it takes bloody forever, in the former case May 02 16:01:50 hrw: not too long, in the latter case May 02 16:02:41 raduga: I know, I know ;) May 02 16:03:27 hrw: "few hours" modulo your cpu/disk speed probably :) May 02 16:04:54 raduga: today I can leave it for whole night May 02 16:07:37 Is it just me or is qpobox not compiling in the current dev branch? I know about bugs 842 and 843 but I am unsure as to what has been incorporated into OE already and what's left to be patched in. May 02 16:08:26 Laibsch: qte configs are updated for it? May 02 16:11:41 hrw: Thank you for the reply. That is what I was talking about. I was not sure if that still needs to be done since apparently, most of what is in bug 843 is in .dev already. Were you able to compile qpobox? May 02 16:11:56 Laibsch: I do not have .dev builds May 02 16:12:13 hrw: Where do I "QPOBox requires QEucJpCodec to be enabled in libqte"? May 02 16:12:25 hrw: Just oz354x? May 02 16:12:30 I could try there. May 02 16:13:00 someone has 4.x builddir and can check one build? May 02 16:13:09 sure May 02 16:13:42 koen: try to build /tmp/openttd_0.4.7-r4651.bb May 02 16:14:12 koen: 3.4.4 gives ICE May 02 16:16:13 koen: http://oe.pastebin.com/694406 is what I got May 02 16:17:03 NOTE: package openttd-0.4.7-r4651-r0: task do_compile: started May 02 16:18:20 NOTE: package openttd-0.4.7-r4651-r0: task do_package: started May 02 16:18:24 and it fails there May 02 16:18:29 so gcc 4.1 works May 02 16:18:43 coool May 02 16:18:58 | install: cannot stat `/data/build/koen/OE/build/tmp/angstrom/work/armv5te-linux-gnueabi/openttd-0.4.7-r4651-r0/OTTD-source-nightly-r4651//scenario/*.scn': No such May 02 16:19:26 ok May 02 16:23:43 hrw: did you seen the images for the c7x0 flash menu? May 02 16:24:03 yes May 02 16:24:22 installation for total dummies May 02 16:25:01 I'll make pictures of the battery switch when I have some more time May 02 16:26:42 I have bloodhound gang day May 02 16:26:58 you do? May 02 16:27:47 np: The Bloodhound Gang - Uhn Tiss Uhn Tiss Uhn Tiss (0:45 / 0:00) May 02 16:28:11 that on is in my playlist as well :) May 02 16:30:33 ~seen _law_ May 02 16:30:38 _law_ was last seen on IRC in channel #openzaurus, 4d 22h 32m 34s ago, saying: 'MrJesse: it is also i good idea to link /var/tmp and /var/lib/ipkg to you sd'. May 02 16:30:46 ~seen __law__ May 02 16:30:48 __law__ was last seen on IRC in channel #oe, 9h 52m 37s ago, saying: 'hi koen '. May 02 16:30:57 ~seen *law* May 02 16:30:59 hrw: i haven't seen '*law*' May 02 16:31:07 ibot: you suck May 02 16:31:08 Thank you! That will be $5 May 02 16:31:13 hihi May 02 16:31:19 5 zloty May 02 16:31:20 southpark fans May 02 16:31:44 woglinde: I should have 5 pln coin in wallet May 02 16:44:19 ~hail SFTP KIO May 02 16:44:22 * ibot bows down to SFTP KIO and chants, "I'M NOT WORTHY!!" May 02 16:44:32 saving patches from kmail directly to ewi roxx May 02 16:59:10 CosmicPenguin: sorry to bother you again - do you have any tools/... available for creating a partition table / fat boot partition now? May 02 16:59:31 dosfstools May 02 16:59:32 ? May 02 16:59:59 ..in an image file :) May 02 17:00:06 i.e. scriptable without being root May 02 17:00:42 i know it's not a big deal (mtools can create the dos partition, and creating a partition table is not that hard), but i'd like to stick to standard tools May 02 17:00:44 (if possible) May 02 17:02:09 iirc we added kind of this into .dev recently May 02 17:06:15 tmbinc: ah, yes May 02 17:06:45 tmbinc: http://cosmicpenguin.net/pub/sources/dosfstools-patches/ May 02 17:06:50 Ok - thats the dofstools patches May 02 17:06:59 that will create a dos partition from a directory May 02 17:07:09 somewhat more OE friendly then mtools May 02 17:07:11 Hi May 02 17:07:12 at least I think May 02 17:07:32 tmbinc: http://cosmicpenguin.net/pub/sources/imgloader/ May 02 17:07:58 tmbinc: and that is a two part system - on one side it makes a binary image file with a generic partition table + your filesystem contents (gziped) May 02 17:08:05 hrw: Anything new regarding the MMC kernel problem ? May 02 17:08:12 and on the other side it writes that image to a LBA addressed device May 02 17:08:36 taking into account geometry and other things that partition tables are picky about May 02 17:08:41 uv1: will test another patch now May 02 17:08:48 hrw: OK May 02 17:09:02 Did you read my message regarding ALSA yesterday evening ? May 02 17:10:35 CosmicPenguin: that's great, exactly what i've been searching. thanks! May 02 17:10:55 tmbinc: please bother me early and often with questions May 02 17:12:22 tmbinc: Why did you not use fdisk ? May 02 17:12:43 uv1: sound not working? May 02 17:13:22 hrw: partly. mplayer plays ogg via headphone though they are a bit muted. May 02 17:14:10 Speaker is very muted (nearly unhearable). Could it be, that for some reason the speaker signal is rootet to the microphone ? May 02 17:14:31 uv1: using sfdisk is ok, but you still need something to assemble the disk image then. May 02 17:14:47 Thats exactly the point too - you want to eliminate the concept of partitioning from the install stage May 02 17:15:03 or at least, hide it May 02 17:15:08 for the lusers of the world May 02 17:15:47 tmbinc: What do you mean with assemble diskimage ? May 02 17:17:08 uv1: you need to pad the parttable to a cylinder boundary, then append the first image, pad again, append the second image. I didn't wanted to do that with "dd", especially as you don't know the exact boundaries (as they relate to a choosen CHS) May 02 17:21:05 tmbinc: Understood. So dosfstools helps ? May 02 17:21:47 uv1: dosfstools doesn't know anything about partitions, iirc. just about the filesystem on it. May 02 17:21:58 tmbinc: http://oe.pastebin.com/694521 May 02 17:22:11 Thats bootimg.bbclass - we use that to make the bootable dos image w/ syslinux May 02 17:22:19 http://oe.pastebin.com/694519 May 02 17:22:29 And thats syslinux.bbclass - we use that to make automated configuration files and menus May 02 17:24:17 CosmicPenguin: did i miss the "mkimg" call there? May 02 17:25:04 I haven't posted it yet May 02 17:26:32 hi greentux May 02 17:26:49 hi hrw, and bugsquashing was good? May 02 17:27:01 tmbinc: http://oe.pastebin.com/694530 May 02 17:27:04 something to that effect May 02 17:28:02 greentux: fixed some bugs so yes May 02 17:28:08 _law_: hi May 02 17:28:15 <_law_> hi hrw May 02 17:28:52 _law_: I posted to winner of gfx contest to get some extra infos from him (name/surname/website) May 02 17:29:09 <_law_> hrw: okay thanks May 02 17:35:54 <_law_> koen: staging/arm-linux/include/libcroco-0.6/libcroco/libcroco.h couldnt be found from librsvg-2.11.1 May 02 17:36:05 <_law_> rsvg-styles.c:838:31: libcroco/libcroco.h: No such file or directory May 02 17:39:46 _law_: looks like badly staged libcroco May 02 17:41:46 <_law_> hrw: if i change back the staging stuff from 0.6.0 everything works again.. May 02 17:42:04 <_law_> hrw: btw. 5500 is on the way :-) May 02 17:44:08 _law_: cool May 02 17:52:23 _law_: ooops May 02 17:52:36 _law_: I guess I had 0.6.0 still installed May 02 17:56:24 03koen 07org.oe.dev * r53b45ad8... 10/packages/libcroco/libcroco_0.6.1.bb: libcroco: fix staging May 02 17:58:20 <_law_> koen: thanks May 02 17:58:43 Hi, May 02 17:59:00 I compiled anthy successfully now. May 02 17:59:05 YEAH! May 02 17:59:15 cool May 02 17:59:18 Trying to get the necessary front-end into OE now. May 02 17:59:28 The name is imkit-anthy. May 02 18:00:12 I have a bb file (basically just the source URI right now) but it fails with "No rule to make target `install'." May 02 18:00:23 Trying to take a look at the make files now. May 02 18:00:53 But I do not even know where to start in the somewhat complicated OE dir structure. May 02 18:01:33 Laibsch, thats a bb file error May 02 18:01:42 did you atleast inherti autotools? May 02 18:01:44 Question 1: Where is the place to look? build/tmp/work/arm-linux/imkit-anthy-0.4.4-r0/? May 02 18:01:47 inherit* May 02 18:02:01 emte: No May 02 18:02:05 I will fix that. May 02 18:02:09 if not it does not know what type of build system the package uses May 02 18:02:36 Question 2: If I needed to make patches where should I apply them? May 02 18:02:58 Laibsch, you should really look at another bb file for examples May 02 18:02:59 koen: is the libc bug still there? May 02 18:03:13 everything you have asked so far is there May 02 18:03:14 I had "inherit autotools pkgconfig" May 02 18:03:22 So that is not it. May 02 18:04:30 why does libxsettings need coreutils-native? May 02 18:04:36 sigh May 02 18:04:38 emte: It is not always easy to understand where to look. All bb files are different. Maybe I should also add that I do not really have much of a clue about compilation ;-) May 02 18:04:59 Laibsch, packages/nano/nano_1.2.1.bb May 02 18:04:59 is a pretty basic bb file May 02 18:05:44 it has the minimum, a patch, and an extra file it needs to compile May 02 18:06:01 Ken|JLime: which bug is that exactly? May 02 18:06:36 hmm missing a maintainer as well ... May 02 18:06:59 too bad i hate using it May 02 18:08:05 bah May 02 18:09:56 Ken|JLime: I'll push your #930 in a minute May 02 18:10:37 03koen 07org.oe.dev * rbfa05745... 10/conf/machine/jornada6xx.conf: jornada6xx.conf: sh -> sh3 and remove PREFERRED_VERSIONS, those should be in the $DISTRO.conf May 02 18:10:44 or not :) May 02 18:10:44 hrw: are you sure? May 02 18:10:44 03koen 07org.oe.dev * rd1e9c567... 10/conf/machine/lite5200.conf: lite5200.conf: remove stuff that should be in $DISTRO.conf May 02 18:11:08 koen: add 'close #NUMB' next time please May 02 18:11:20 crap, forgot to add that May 02 18:13:38 koen: Eric ircnick is? May 02 18:15:20 uE May 02 18:15:34 and s/Eric/Erik/ May 02 18:15:52 uE? May 02 18:15:59 ~seen uE May 02 18:16:05 ue was last seen on IRC in channel #handhelds.org, 88d 17h 50m 33s ago, saying: ''night all'. May 02 18:16:15 that's him May 02 18:16:26 added May 02 18:16:44 thanks May 02 18:16:48 koen: still no koen@hh.org? May 02 18:16:59 nope May 02 18:17:33 we need cms for oe website... May 02 18:18:44 lol May 02 18:19:02 wow - this is so 2004 May 02 18:19:07 hrw, arnt those called HOST backups? May 02 18:19:50 koen: 'You are now a registered user on LTG Projects' May 02 18:25:59 hrw: any suggestions for the scm? May 02 18:26:29 koen: scm or cms? May 02 18:26:35 ehm May 02 18:26:37 cms May 02 18:27:13 koen: I know wordpress, you know drupal May 02 18:28:01 koen: I also saw many projects which used wiki also for mainpage May 02 18:28:36 that could work as well May 02 18:28:36 evening May 02 18:28:42 hey cyrilRomain May 02 18:28:56 time to do some homecleaning May 02 18:29:05 hrw|afk: we just need someone to to the styling May 02 18:30:20 koen: create few snapshots as gfx and I will try to get someone for layouting May 02 18:30:35 koen: remember that I do websites at work ;D May 02 18:30:39 :) May 02 18:30:55 hrw|afk: I know that, I just don't know if you *want* to do it May 02 18:31:17 koen: finally, I remade the script in bash (less quote and comma than the python one), with the md5sum check: http://oe.pastebin.com/694676 May 02 18:31:47 * cyrilRomain usually prefer python but for simple script like that, bash script are maintainable May 02 18:32:21 cyrilRomain: s/donw/down/ May 02 18:32:26 man - nothing like building a GUI to make you pine for a USES behavior, huh? May 02 18:32:46 koen: BTW, there is a problem in the OE.db.bz2.md5 : it refers to the /var/www/OE.db.bz2 instead of just OE.db.bz2 May 02 18:33:32 koen: you have phoenix eyes ;) May 02 18:34:13 and s/ipi/ippi/ May 02 18:34:36 koen: two phoenix eyes ;) May 02 18:35:18 cyrilRomain: md5sum corrected May 02 18:35:22 http://ewi546.ewi.utwente.nl/OE/OE.db.bz2.md5 May 02 18:35:29 the script now seds out the path May 02 18:37:56 koen: 'seds out the path' ? May 02 18:38:41 * emte shoots ncurses and then runs it over with a steamroller May 02 18:38:42 cyrilRomain: http://oe.pastebin.com/694686 May 02 18:41:19 koen: great May 02 18:41:25 koen: http://oe.pastebin.com/694692 May 02 18:42:32 thanks May 02 18:42:35 hrw|afk: even better :) May 02 18:43:00 emte: I still have difficulty understanding the file structure. May 02 18:43:13 emte: That stuff is not in the sample bb files. May 02 18:43:17 koen: zecke mentioned yesterday the 'mnt db check' commande, it is a monotone-0.26 command ? May 02 18:43:17 ok. now worse part - bathroom. May 02 18:43:50 RP: http://www.oesf.org/forums/index.php?showtopic=19184 May 02 18:43:51 cu May 02 18:44:03 What is for example the difference between tmp/ and build/tmp/? May 02 18:44:56 hrw|afk: :-/ May 02 18:45:49 Laibsch: should be the same directory I think. If you can have both, I guess it's because you 'bitbake somepackage' from the ./ and 'bitbake somepackage' from the build/ directory May 02 18:47:30 Laibsch: so be sure to always call bitbake from the build/ directory, or set the TMPDIR in your local.conf file May 02 18:50:06 OK May 02 18:50:25 But I think TMPDIR is set. May 02 18:52:12 Actually, it was not. May 02 19:13:29 ~botmail for zecke: Does http://www.rpsys.net/openzaurus/temp/bb_ncurses_fix-r0.patch look like the right solution? May 02 19:15:54 re May 02 19:16:05 RP: new patch to add to linux-oz May 02 19:16:20 mmc 1/2 + 2/2 May 02 19:16:25 re hrw May 02 19:16:30 hrw: ? May 02 19:17:23 raduga: read lakml May 02 19:17:30 ok May 02 19:19:53 raduga: basically: mmc timeouts fixes May 02 19:20:13 raduga: 2/2 was tested be me on card which does not work under 2.6 May 02 19:20:30 great May 02 19:21:14 NOTE: package linux-openzaurus-2.6.16-r19: task do_patch: completed May 02 19:30:41 03hrw 07org.oe.oz354x * rd17bc11d... 10/packages/linux/ (4 files in 2 dirs): linux-openzaurus: added 2 MMC patches from RMK - should help with timeouts a bit May 02 19:30:44 03hrw 07org.oe.dev * r06d21bcf... 10/packages/linux/ (3 files in 2 dirs): linux-openzaurus: added 2 MMC patches from RMK - should help with timeouts a bit May 02 19:31:59 hrw: That's good news. It might even fix that error on the forum :) May 02 19:34:11 RP: may but does not have to May 02 19:34:21 RP: my cards still fail with 2/2 patch May 02 19:34:40 > cards still fails: May 02 19:34:43 re zecke May 02 19:34:43 Indeed it does, but with this patch we can continue to talk to the card, May 02 19:34:43 so it's a positive step forward in terms of handling data timeout errors May 02 19:34:43 correctly. May 02 19:34:44 Thanks for confirming that change. May 02 19:34:52 rmk mail May 02 19:34:57 wb zecke May 02 19:35:19 zecke: out of revdns entries? May 02 19:35:56 RP: which gcc version do you use for the kernel? May 02 19:36:18 do13: 3.4.3 May 02 19:36:36 RP: weird this patch didn't work for me :} May 02 19:37:13 do13: I use 3.4.4 May 02 19:37:25 do13: Although I've booted 4.0 and 4.1 compiled versions (EABI though) May 02 19:37:32 RP,hrw: thx. 4.0.1 is evil May 02 19:37:38 zecke: Its working here now... May 02 19:39:13 hrw: At least you're making some headway :) May 02 19:39:29 do13: I booted 4.0.2 compiled kernels without problems May 02 19:39:32 RP: ? May 02 19:39:41 hrw: With rmk on that problematic card May 02 19:40:18 RP: and I fear we need a test case as well ;) May 02 19:41:03 zecke: That'll be tricky without hooking into bitbake internals :-/ May 02 19:41:20 hrw: my 4.0.1 kernel works without nand:) 3.4.4 generates correct code and koens 4.1 eabi kernel works May 02 19:41:31 RP: in next days I will get sharp serial cable - next debug session will be more detailed May 02 19:41:50 RP: that is easy ;) May 02 19:42:00 RP: mickeyl has some unused python magic in shell.py May 02 19:42:12 RP: e.g. we can load bin/bitbake like any other module :} May 02 19:42:13 hrw: For kernel bugs they're near essential May 02 19:42:39 zecke: I didn't know that. I might have a use for it :) May 02 19:42:48 cya in a while I need to get my physics assignment finished May 02 19:43:09 ok. I'll probably go and watch a film shortly... May 02 19:43:12 RP: you could even 'inject' your version of a function :) May 02 19:43:27 I think I will be around some more hours May 02 19:43:28 zecke: python is scary ;-) May 02 19:44:04 RP: dynamic ;) May 02 19:45:04 JustinP: Erik has no write access at all :) May 02 19:46:11 zecke: Erik H? May 02 19:46:27 zecke: :-( May 02 19:46:38 this was my impression from the mail May 02 19:46:46 pb_: you don't have access as well ;) May 02 19:46:55 :-) May 02 19:46:56 pb_: why should we grant it to Erik then ;) May 02 19:47:21 well, at least Erik seems to have asked for it. I guess that puts him one step ahead. May 02 19:47:34 pb_: right, and it should be granted of course May 02 19:47:45 koen and mickeyl will add him on days May 02 19:48:14 let's hope mickey|exhausted soon recovers sufficiently to do the needful May 02 19:49:54 zecke: getting any closer to fixing the glibc bug? May 02 19:50:36 Ken|JLime: ah, didn't the patch work? May 02 19:50:53 Ken|JLime: alternatively try RPs patch May 02 19:56:40 cd 12 May 02 20:01:27 zecke: havent used the patch yet, just wonderd if any patch has gone into the repository May 02 20:27:14 koen: could we install a key from erik? May 02 20:27:44 any documentation on how to add a new machine and distro into oe? May 02 20:28:12 chouimat: not yet May 02 20:28:24 chouimat: edit text files is the closest thing, and mickeyls fosdem talk May 02 20:28:56 chouimat: pick something close, make sure it builds and then change it to fit your needs. May 02 20:29:16 On the varpoware list are screencasts May 02 20:30:04 zecke: cbrake thank guys ... May 02 20:56:24 night all May 02 21:05:38 hrw|gone: cya May 02 21:05:45 hrw|gone: I will send you some files tomorrow May 02 21:10:00 chouimat, copy someone else machine and distro files May 02 21:10:21 I am sort of doing that for the osk machine and openmap distro May 02 21:11:00 zecke: way ahead of you May 02 21:12:29 added him? May 02 21:12:31 zecke: and actually it was me who approached Erik about access a month ago, not the other way around :) May 02 21:12:44 zecke: I've been the past two hours in the cinema May 02 21:14:08 for the last hour i have been writing bullshit for my physics assignment May 02 21:15:33 what were you assigned? May 02 21:15:59 zecke: it would also help if I actually had Eriks key to add... May 02 21:16:21 raduga: boring mechanics/dynamics assignment for minors May 02 21:19:37 zecke: I'll add the key as soon as it lands in my inbox && I'm awake May 02 21:21:57 koen: did you inform erik as well? May 02 21:22:15 zecke: yes I did May 02 21:22:25 nice May 02 21:22:27 zecke: hrw got a CC: to add him to people.php May 02 21:22:38 awesome May 02 21:22:47 zecke: I forgot to follow up on the email conversation we had a month ago May 02 21:23:20 ~lart koen May 02 21:23:20 * ibot gives koen an extra strength ACME sleeping pill, sending koen to sleep for 150 years, and awakening to seven strange dwarfs and a large apple May 02 21:23:22 that happens May 02 21:23:26 :) May 02 21:23:45 Thats an interesting lart - combining Looney Tunes and Disney May 02 21:24:26 and sleeping, given koan's comment about being awake May 02 21:29:11 RP: I'm back May 02 21:30:32 zecke: can tinderbox create groups? May 02 21:30:43 koen: groups? May 02 21:30:56 koen: What makes tbox3 good = it doesn't crash like tbox2 does May 02 21:31:25 koen: you want to group by built software? May 02 21:31:32 zecke: like a link to a MACHINE (say, c7x0) which shows all the builds going on using that machine (oz, angstrom) May 02 21:31:52 and a link to $DISTRO, which would should for what machines it is getting built May 02 21:31:58 koen: it will work that way (including ATOM, irc bot...) May 02 21:32:17 koen: I'm happy we receive the data ATM May 02 21:32:17 zecke: I just fired up a multimachine build for angstrom :) May 02 21:32:42 oh May 02 21:33:08 and I put ${DISTRO}-${MACHINE} in the tinder name May 02 21:33:35 INDER_MACHINE = "ewi546-${DISTRO}-${MACHINE}-bootstrap" May 02 21:33:38 to be exact May 02 21:33:49 (although I'm building gpe) May 02 21:34:27 koen: I will start to send 'hostname+distri+WHAT-YOU-BUILD' May 02 21:34:49 and machine? May 02 21:39:00 RP: "Installed 3.5.4-rc2 and it now detects my card as 1.9gig free " May 02 21:40:10 zecke: how easy can we pull info like "what was the last succesfull build for machine foo" from tinderbox? May 02 21:40:49 zecke: I'd like to integrate it with the CMS we'll use for the new OE site May 02 21:43:44 koen: once I hacked on tbox it will be possible May 02 21:44:15 koen: current agenda, Fix BitBake, Create BitBake QA, Consider OE Foundation May 02 21:44:22 koen: make TBox the way I want May 02 21:45:39 decide what scm to use for oe metadata ... May 02 21:49:12 wow May 02 21:51:10 CosmicPenguin: wow? May 02 21:52:25 koen: heh - no, matchbox just went boom May 02 21:52:29 so I was vocalizing May 02 21:52:50 did you wrap tinfoil around it and heat it? May 02 21:53:37 uhmmm.... a couple of questions May 02 21:53:46 what's the normal transfer rate for an SD card ? May 02 21:53:58 depends on the card May 02 21:54:15 anywhere between 10kB/s and 3MB/s May 02 21:54:31 i'm getting 60kb/s on a 1gb sandisk May 02 21:54:33 it also depends on the controller :) May 02 21:54:36 RP: okay, I need to go to bed (tired) and need to wake up in 7h again May 02 21:54:38 koen: so it doesn't depend on the OS May 02 21:54:43 ah, there you go May 02 21:54:45 =P May 02 21:55:05 zecke: sleep tight May 02 21:55:14 second question, i haven't found vnc in the .dev branch May 02 21:55:21 is it even part of oe ? May 02 21:56:04 CSMan: try looking in ./packages/vnc May 02 21:56:40 org.openembedded.org/packages/vnc ? May 02 21:57:01 *dev May 02 21:57:06 probably May 02 21:57:21 ok May 02 23:50:03 nite all May 02 23:54:37 hi May 02 23:55:46 I'm looking for openzaurus's kernel source and config. May 02 23:55:52 can someone help me please ? May 03 01:16:50 magnet: look in OE May 03 01:17:42 * emte suspects that magnet wants to try and compile it outside of OE May 03 01:17:47 magnet: the URLs for the source and patches are in OE. There's no tarball of the patched source that I know of May 03 01:18:05 I saw his comments in #oz, he wants to compile a module.....on his Z.... May 03 01:18:18 but i guess he didnt want to read the oz site? May 03 01:18:33 or atleast the old links were pretty clear May 03 01:18:54 * JustinP doesn't know which links you mean May 03 01:19:00 but all of the URLs are in OE May 03 01:20:52 mostly the openzaurus.org wiki link May 03 01:21:09 which has http://openzaurus.berlios.de/Developing May 03 01:21:21 which points to oe getting started :) May 03 01:22:00 uhmmm... my ipaq won't restart after going into sleep mode May 03 01:22:12 i have to click the reset button May 03 01:22:13 which ipaq and distro? May 03 01:22:20 and version May 03 01:22:26 h5550 familiar 0.8.4rc1 May 03 01:22:54 but it happens only when mplayer is in the middle of playback May 03 01:23:04 is there anything in the familiar bug list about it? May 03 01:23:11 ... May 03 01:23:13 i doubt it May 03 01:23:23 why would you suspend without killing mplayer? May 03 01:25:03 i forget, doesnt mplayer lock memory area while running? May 03 01:25:33 locak/allocate May 03 01:25:36 lock* May 03 01:26:22 and when you suspend you allocate and lock "floating" data into memory May 03 01:27:10 so if you try to save floating system data and it has nowhere to save it, will the system lock up? May 03 01:28:07 bah thats too much thinking for me ... i need to figure out how to break out 80 pins to a bradboard May 03 01:28:19 bread too May 03 01:31:28 oz runs so well on my small handheld, could I run it on my old panasonic 75mhz 32meg 2gig toughbook? May 03 01:32:25 if you could convince it to compile May 03 01:43:35 emte: i see May 03 01:44:00 will the same happen if i use xmms-embedded May 03 01:44:19 by my wonky logic i would suspect so May 03 01:44:35 but that by no means means i am correct May 03 05:14:06 anyone know if QTopia Desktop works with OZ, and in what package the code for that lives? May 03 05:14:41 also if there is support for something like Evolution May 03 06:46:49 bonjour May 03 06:51:42 moin May 03 07:40:25 morning May 03 07:55:04 Good morning May 03 07:55:31 Guys, is still bug squashing time? May 03 07:55:37 * Laibsch points to topic. May 03 07:57:19 Laibsch: moin May 03 08:01:46 morning all May 03 08:01:53 hey RP May 03 08:07:31 morning May 03 08:07:39 RP: hey May 03 08:07:52 * zecke missed the class this morning May 03 08:25:10 RP: Woul you mind if I moved rcS/S03udev to rcS/S38udev? That fixes the "sometimes /etc/resolv.conf is empty on boot" problem May 03 08:25:50 depending on how fast you get the lease /var is not populated yet May 03 08:26:12 CoreDump|home: That sounds like quite a bit of a change in init level and I'd be a bit concerned about doing that :-/ May 03 08:26:27 CoreDump|home: udev should be one of the first things running... May 03 08:27:08 none of the other rcS scripts depend on udev AFAICT, and FWIW it's working nicely at S38 on OZ. May 03 08:27:30 No init scripts access /dev? May 03 08:28:16 sure some do but they use devicefiles created by rs/S??devices, not udev May 03 08:28:29 udev should be creating those devices May 03 08:28:46 well, we could try moving mount-all and pop-volatile at S01 and S02 May 03 08:29:00 mount-all comes after udev May 03 08:29:09 as it is now yes May 03 08:29:24 the problem is that udev enables WLAN / LAN cards May 03 08:29:31 I'm not familiar enough with pop-volatile to comment about that May 03 08:29:39 Well, this is strange. If I build bitbake -b ../org.openembedded.dev/packages/qpobox/qpobox_0.5.4.bb, OE completes. But just bitbake qpobox fails. Why is that? May 03 08:29:40 should hotpluging networks not be delayed until /etc/init.x/networking is up? May 03 08:29:47 and if /var is not populated yet (as it is run at S38) resolv.conf will be empty May 03 08:30:10 Shouldn't bitbake -b fail if the dependencies are not fulfilled? May 03 08:30:25 Laibsch: nom -b doesn't check depends May 03 08:30:27 Other than that I cannot think of any difference between the two. May 03 08:30:44 udev must be one of the first things run and the other things should come later. I'm absolutely certain about that. If that causes problems with networking, we need to fix that in some other way May 03 08:30:46 CoreDump|home: yes, of course not. But I thought the build would fail inherently. May 03 08:31:14 RP: see above, Im pretty sure right thing to do is not to hotplug networks until in init.d/networking May 03 08:31:24 Laibsch: only if the dependency is a compile-time dependency *shrug* then yeah, it should fail May 03 08:32:03 XorA: I agree but the way udev handles hotplug that is not an easy task May 03 08:32:09 XorA: udev is right to load the modules - we just should't be trying to configuring them until the networking scripts start... May 03 08:32:26 RP: yes, that is what I meant, I having bad explaining day May 03 08:32:33 CoreDump|home: So the correct solution is to have network.sh exit until the networking scripts have run May 03 08:33:10 RP: which would result in WLAN cards never be working out-of-the-box if the device is booted with that card inserted May 03 08:33:35 as udev doesn't get a 2nd hotplug event for the CF device May 03 08:34:10 echo $DEVICE >>/var/queded-network-start May 03 08:34:21 CoreDump|home: udev should load the modules. Or boot time networking code should then tun network.sh over any devices it finds to configure them May 03 08:34:38 RP: which it does, yes May 03 08:34:49 CoreDump|home: So why doesn't that work? May 03 08:35:33 CoreDump|home: I'm saying that running network.sh to configure the devices should happen later in the boot, not running udev later May 03 08:35:40 because udev is started wayyyyy before /var is populated, which means DNS entries can not be written to /etc/resolv.conf (which is a link to /var/*) May 03 08:36:08 So we need to rerun the network.sh calls later May 03 08:37:21 might just as well scrap network.sh and use XorA's qued-network-start thingy from within init.d/network May 03 08:37:48 CoreDump|home: right. For booting that's the correct approach May 03 08:37:56 Once we have udev running properly, there will be three files in /dev before udev runs. This means udev needs to run early in the boot process. We can't change that. May 03 08:38:29 ok, I will implement the delayed NIC config then May 03 08:39:30 after that we should have a good hard look what devicefiles are created by udev to make it faster (ie: it's creating like 50 tty's and stuff) May 03 08:39:59 CoreDump|home: verison 089 is supposedly a lot faster May 03 08:40:09 but we should also try and trim it down a bit... May 03 08:40:12 but is disabled in .dev =) May 03 08:40:22 or was last time I checked May 03 08:40:41 Just because of lack of testing - if it works on the Z happily, I'll change it :) May 03 08:41:42 tuning udev would lower boot-time quite a bit. I managed to lower it from 1:26m to 1:06m already w/o touching udev May 03 08:44:30 Sounds good :). Or boot process is far from optimal atm :-/ May 03 08:44:34 s/Or/Our/ May 03 08:45:45 I'm staring a few rcS scripts in parallel and hacked pop-volatile. udev could drop boot-time to < 50s May 03 08:46:26 I wiped 5 seconds off that with the kernel zlib changes. We can also benefit from jffs2 summary and prelink May 03 08:46:38 prelink :) May 03 08:46:52 yeah, summary on Akita and c7x0 would be a good thing May 03 08:47:02 zecke: prelink :)? May 03 08:47:05 always takes up to 5s just to mount /home May 03 08:48:26 damn, there comes another problem May 03 08:49:07 can't use /var for the network queue as it is not populated or a tmpfs yet mehh, gotta write to flash for that May 03 08:50:23 Can't a script just scan for all network interfaces present? May 03 08:52:57 ahem...another problem =) We can't scrap network.sh since it woul still be needed for hotplugging a NIC May 03 08:53:02 mehh May 03 08:53:23 CoreDump|home: Add a check to the start of it which just exits if the system is in early bootup May 03 08:53:41 and a comment explaining why ;-) May 03 08:54:01 the problem there is a reliable check for beeing in runlevel S May 03 08:54:19 I'm avoiding that issue :) May 03 08:54:24 can't check /sys or /proc beeing mounted, can't use "runlevel" May 03 08:55:40 time to queue stuff in a fifo May 03 08:56:00 hmm? May 03 08:56:35 if I recall, stuff written to a fifo isnt actually stored on disk May 03 08:56:51 ~lart cvs.sf.net May 03 08:56:51 * ibot slaps a compatible dib on cvs.sf.net's head May 03 09:18:24 I'll just check if /tmp is umounted to indicate we are at rcS, screw users hacking /tmp to be non-tmpfs May 03 09:21:51 03mickeyl 07org.oe.dev * r0ec23d05... 10/packages/qt/qt4-x11-free_4.1.1.bb: qt4: improve packaging May 03 09:21:59 03mickeyl 07org.oe.dev * raa79837f... 10/packages/python/ (python-pyqt4/optional-features.patch python-pyqt4_0.9.bb): pyqt4: make packaging more standard May 03 09:22:00 03mickeyl 07org.oe.dev * refe62767... 10/classes/ (sip.bbclass sip4.bbclass): sip(4).bbclass: fix appending/prepending dependencies May 03 09:29:53 zecke: Maybe it would be a good idea to point http://handhelds.org/~zecke/downloads/bitbake/html/ to the more current version at bitbake.berlios.de? May 03 09:30:05 Laibsch: hehe May 03 09:30:21 Laibsch: well, as any userpage May 03 09:30:37 Is it still bug squashing time? May 03 09:30:42 * Laibsch points to the topic May 03 09:30:52 Laibsch: change it, at your will May 03 09:30:53 zecke: What is the reason for the duplication? May 03 09:31:04 Laibsch: we always try to close bugs May 03 09:31:14 I do not even know how to change the topic May 03 09:31:15 Laibsch: I wasn't project admin at berlios for bitbake May 03 09:31:23 Laibsch /help topic May 03 09:31:29 Aha May 03 09:31:56 topic [OpenEmbedded Developer Lounge | OE and bitbake were made to match, keep them both up to date | BitBake 1.4.0 has been released - only use HEAD if you are developing it] May 03 09:32:04 Ok, that did not work ;-) May 03 09:32:51 zecke: Did you change any part of the manual? May 03 09:32:59 Laibsch: yes May 03 09:33:07 Laibsch: the fetcher section is new May 03 09:33:08 I see. May 03 09:33:12 but not reviewed May 03 09:35:23 I was able to compile qpobox yesterday. Today, I can't seem to do it. How can I clean a package and its dependencies? I think I read about that some day but can't seem to find it now. May 03 09:35:50 Laibsch: there is no recursive way May 03 09:35:57 :-( May 03 09:35:57 Laibsch: bitbake -b file.bb -cclean May 03 09:36:06 Yeah, I did that. May 03 09:36:13 qpobox still does not compile. May 03 09:36:21 Changes from bug 842 are applied. May 03 09:36:23 Laibsch: recursive would mean cleaning everything to quilt-native May 03 09:36:33 Laibsch: try to figure out the reason ;) May 03 09:36:38 indeed. May 03 09:36:40 sorry need to do some work now May 03 09:36:50 As a non-programmer that is very hard for me. May 03 09:37:08 Laibsch: well, try classifying the error May 03 09:37:20 Laibsch: does it happen on patch, configure, compile, package, stage May 03 09:37:26 Laibsch: is it a missing file May 03 09:37:38 Laibsch: is it some weird compiler error and it is barfing about the syntax May 03 09:38:22 Most of that information has scrolled off the screen. May 03 09:38:30 I will run it inside less or something. May 03 09:38:39 What is staging anyways? May 03 09:38:50 Laibsch: installing headers to tmp/staging/* May 03 09:38:51 RP: could you pleae check if /tmp is a tmpfs on your spitz? May 03 09:39:13 kergoth`zzz: sleep well May 03 09:39:32 CoreDump|home: its not May 03 09:39:36 err May 03 09:40:01 god damn this May 03 09:40:32 3 machines out of a bazillion use a real /var and that screws anything up grr May 03 09:41:19 Also keep in mind OE != Zaurus :-/ May 03 09:41:37 there is no way to see if we are at rcS May 03 09:41:41 none May 03 09:42:21 can't use /proc or /sys or /var or /tmp, can't use `runlevel`. Nothing damnit. May 03 09:43:07 CoreDump|home: write a saner init system May 03 09:43:49 init has nothing to do with sanity =) May 03 09:44:28 CoreDump|home: Have the network script touch /dev/networkingup and network.sh check for it May 03 09:45:12 03mickeyl 07org.oe.dev * r1e5d155b... 10/packages/nonworking/imkit/imkit_0.4.0-pre2.bb: May 03 09:45:12 add imkit, a japanese inputmethod for Opie. May 03 09:45:12 in nonworking, because it needs a c++ template syntax patch to comply with modern compilers May 03 09:45:15 udev mounts /dev/ tmpfs by its own? May 03 09:46:51 CoreDump|home: yes May 03 09:47:07 morning all May 03 09:47:14 hi Liam May 03 09:48:25 RP: I'm meeting up with XorA tomorrow to fix that c7x0 audio record issue. hopefully It will be something easy :) May 03 09:48:51 lrg: Lets hope so. I've never had a chance to look into it properly... May 03 09:49:37 RP: We'll be at the local Lug, so we may be rather intoxicated and fail to easily see the bug ;) May 03 09:51:40 lrg: Worryingly, I've written some of my best code whilst intoxicated! ;-) May 03 09:52:53 heh May 03 09:54:00 you code without being intoxicated? May 03 09:54:00 lrg: will you have a build system there? May 03 09:54:05 englishs are sick... May 03 09:54:32 zecke: you are close to a beating probably for that insult :-) May 03 09:54:54 XorA: sick and violant ;) May 03 09:55:04 zecke, XorA and I are from Scotland May 03 09:55:14 zecke: people in scotland dont like the word glish :-) May 03 09:55:22 English, this keyboard is shit May 03 09:55:22 XorA, I'll try May 03 09:55:40 lrg: cool, you can flash some kernel changes then May 03 09:56:06 XorA: hehe, I knew May 03 09:56:10 * RP is nearly in Scotland... May 03 09:56:11 * XorA is technically english though, only 20/31 years in scotland May 03 09:56:17 I considered british or english May 03 09:58:16 XorA: lrg : RP : all: apologize I will not use this disgusting word again ;) **** ENDING LOGGING AT Wed May 03 09:59:56 2006