**** BEGIN LOGGING AT Sat Sep 01 02:59:56 2007 Sep 01 06:13:08 RP: Angstrom-angstrom-x11-image-glibc-deb-2008.1-test-20070831-c7x0-installkit.tgz :) Sep 01 06:13:43 RP: didn't boot it yet, but image generation with .debs work my my local copy now Sep 01 06:53:12 03koen 07org.oe.dev * r728db2a5... 10/ (1 packages/xoscope packages/xoscope/xoscope_1.12.bb): xoscope: add 1.12 Sep 01 06:53:18 03koen 07org.oe.dev * r0805498f... 10/ (1 packages/opentom packages/opentom/ttimgextract.bb): ttimgextract: add svn version Sep 01 07:27:42 03rwhitby 07org.oe.dev * r89f3ef8f... 10/ (1 packages/tasks/task-openmoko-feed.bb): openmoko-feed: Added ntpclient Sep 01 08:35:43 03rwhitby 07org.oe.dev * r1af7742e... 10/ (1 packages/tasks/task-openmoko-feed.bb): openmoko-feed: Added timezones Sep 01 08:49:38 hi all Sep 01 08:49:58 * * OE Bug 2906 has been created by mitroshin(AT)gmail.com Sep 01 08:50:00 * * /proc/arm stays permanent Sep 01 08:50:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2906 Sep 01 08:50:07 i got a problem with eglibc ... but i suppose is only a OE error .... Sep 01 08:50:14 take a look -> NOTE: package eglibc-intermediate-2.6+svnrNOT IMPLEMENTED: started Sep 01 08:51:38 tha svn date is completelly wrong (and with a space) so it fail to decompress tha tar.gz, despite it's creadetd with the right name, with a true date Sep 01 08:52:15 that means that package is using a bitbake feature that your bitbake does not have Sep 01 08:52:30 SRCREV functionality I believe Sep 01 08:53:53 but i previous days it worked ... Sep 01 08:56:38 dunno - just giving you any hints that I can think of Sep 01 08:56:46 could somebody be so kind an and commit upgraded recipe for rt73 driver module? I built and tested it for angstrom/akita- working fine. http://www.pastebin.ca/677736 Sep 01 08:57:42 sure, thanks anyway rwhitby :) Sep 01 09:35:05 gm all Sep 01 09:36:38 hi likewise Sep 01 11:14:50 good morning Sep 01 11:15:10 hi Sep 01 11:36:17 gm Sep 01 11:36:35 hi Crofton Sep 01 12:33:37 anyone familiar with meta/meta-sdk and/or meta/meta-toolchain ? Sep 01 12:34:54 mickey|mdd: you really want to wait for the poky classes to get merged so we can use their toolchain stuff Sep 01 12:38:17 the diffs for the poky classes are on the oe-dev ml as RFC Sep 01 12:40:41 are they Sep 01 12:40:42 good Sep 01 12:40:48 i'll review when I'm back then Sep 01 12:40:50 i need a toolchain Sep 01 12:41:03 will the poky patches make it relocatable? Sep 01 12:41:11 not sure about that Sep 01 12:41:16 k Sep 01 12:41:30 it would make it pluggable into OE as ASSUME_PROVIDED Sep 01 12:41:55 ah, right, yes, that's a plus Sep 01 12:42:05 but primarily I'm looking for a standalone toolchain Sep 01 12:42:13 their change to image.bbclass is sweet Sep 01 12:42:30 all you need is IMAGE_INSTALL = "udev dropbear busybox" in an image recipe Sep 01 12:45:32 hmm, this toolchain here does not quite produce working binaries Sep 01 12:45:34 ./hello-world: ./hello-world: 1: Syntax error: word unexpected (expecting ")") Sep 01 12:45:38 does that ring a bell? Sep 01 12:46:03 x86 binary? Sep 01 12:46:17 no, arch is ok Sep 01 12:46:18 mickey@harrypotter:/tmp$ file ./hello-world Sep 01 12:46:18 ./hello-world: ELF 32-bit LSB relocatable, ARM, version 1 (SYSV), not stripped Sep 01 12:46:20 and c++ is know broken Sep 01 12:46:27 it's gcc Sep 01 12:46:32 C, even Sep 01 12:46:41 is it an eabi binary? Sep 01 12:46:51 yep, produced with the output of meta-sdk Sep 01 12:47:04 (since meta-toolchain does not package any compilers) Sep 01 12:49:52 tried disambling it? Sep 01 12:50:12 disassembling? Sep 01 12:50:16 no, not yet Sep 01 12:50:34 or just stracing it Sep 01 12:51:04 (normally I would say "use OE, it works" at this point) Sep 01 12:51:24 hehe Sep 01 12:51:33 unfortunately this doesn't work this time Sep 01 12:51:47 root@fic-gta01:/tmp$ strace ./hello-world Sep 01 12:51:47 execve("./hello-world", ["./hello-world"], [/* 13 vars */]) = -1 ENOEXEC (Exec format error) Sep 01 12:51:54 wow Sep 01 12:52:17 I'd still vote for handing out qemu images containing a native toolchain Sep 01 12:53:42 hmm Sep 01 12:54:13 what do you think about making company maintainers (e.g. the mysterious openmoko one) part of the angstrom core team? Sep 01 12:55:08 sounds good to me Sep 01 12:55:12 if he wants :) Sep 01 12:55:26 or she :) Sep 01 12:55:38 right Sep 01 12:55:42 :D Sep 01 13:18:14 Who in OE uses meta-sdk and/or meta-toolchain atm? Sep 01 13:18:26 flo_lap: Are you the main user of the above? Sep 01 13:19:38 RP: hum well, I wouldn't say "main user" but I use it from time to time. Sep 01 13:19:54 flo_lap: What for exactly? To make a toolchain you can use outside OE? Sep 01 13:19:55 it would be very beneficial if we could merge the poky stuff Sep 01 13:20:10 RP: yes Sep 01 13:20:15 flo_lap: I changed it quite a bit in Poky and I'm wondering about merging back the changes Sep 01 13:20:37 It would still work as an external toolchain but also as one you can ASSUME_PROVIDED in OE too... Sep 01 13:21:10 RP: Sounds pretty good... Sep 01 13:21:13 yah Sep 01 13:21:13 mickeyl: I am aware of bugs with meta-toolchain btw. It can't cope with armv5te vs armv4t etc :-( Sep 01 13:21:40 well, gradual improvement :) Sep 01 13:21:47 i'd say go for it then Sep 01 13:21:56 RP: What do you expect to break merging in the changes? Sep 01 13:22:29 flo_lap: Lots of things will get changed and its just possible things will break. I can't point to anything specifically Sep 01 13:22:43 flo_lap: I rewrote the binutils-sdk .bb files for example Sep 01 13:22:49 and gdb Sep 01 13:23:25 We will also need Poky's pkgmap changes which mean people will have to wipe tmp for that change Sep 01 13:23:45 well, since it works for Poky it can't be too bad ;) Sep 01 13:23:59 flo_lap: right :) Sep 01 13:24:08 RP: lets synchronize these changes with the SRCREV changes, so people will only have to wipe tmp once Sep 01 13:24:26 mickeyl: You don't have to wipe tmp for the SRCREV changes Sep 01 13:24:27 ah well, SRCREV doesn't mean they _have_ to rebuild, but most of them will probably Sep 01 13:24:47 then again, not too many changes at once Sep 01 13:25:01 Technically you don't have to wipe tmp for the pkgmaps changes either, just the install stamps Sep 01 13:25:21 Its probably safer to wipe tmp for most people though Sep 01 13:26:26 mickeyl: I also split the package_write_deb and package_write_ipk functions into separate tasks so if you add a new packaging type, it just builds any new packages needed automatically Sep 01 13:26:41 whew, nice! Sep 01 13:26:58 another meta-toolchain bug is it only works with ipk builds not deb ones :-( Sep 01 13:27:19 * RP wishes he had time to fix these kind of things... Sep 01 13:32:08 well, just go for the basics, and lets see whether we can improve then Sep 01 13:32:19 OM has a strong interest into an installable toolchain Sep 01 13:32:30 We can improve them, I know how to do it, its just a question of time :/ Sep 01 13:33:01 mickeyl: Have you looked at bitbake trunk btw? It should be playable with now ;-) Sep 01 13:33:13 RP: no time yet Sep 01 13:33:15 will do Sep 01 13:33:28 mickeyl: bitbake -u ncurses Sep 01 13:34:01 mickeyl: There is a lot of bits I don't like and it could do with being made more pythonic but I think we have a good base there to work from with UIs now Sep 01 13:34:38 * RP needs to go, back later (I might try and merge some bits from poky) Sep 01 13:34:44 * flo_lap too Sep 01 13:34:47 bbl Sep 01 13:34:59 RP: awesome! thanks. see ya Sep 01 13:56:18 RP: I have some fixes for rootfs_deb we can merge on top of http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-August/002931.html Sep 01 13:57:20 it shouldn't be hard to call rpmbuilder as well should it? Sep 01 13:57:27 (for an external toolchain) Sep 01 14:02:29 03mickeyl 07org.oe.dev * rd3620bd2... 10/ (4 files in 3 dirs): Sep 01 14:02:29 imlib2: explicitly state that we're building a version with X11 support Sep 01 14:02:29 add patch to remove using local X11 includes Sep 01 14:05:02 RP,mickeyl: update diffs: http://dominion.kabel.utwente.nl/koen/OE/poky/ Sep 01 14:06:28 oki, will have a look Sep 01 14:07:06 that is able to build a .deb based angstrom-x11-image for me Sep 01 14:07:42 (with adapted image recipes, of course) Sep 01 14:09:15 afaik the armv4t vs armv5te is just what gcc will choose as default in its builtin spec file, so explicitly passing -march=armv4t works Sep 01 14:09:25 (which is how it 'works' in OE) Sep 01 14:09:53 *nod* Sep 01 14:11:20 we should teach package_deb.bbclass to use debtags for the real architecture Sep 01 14:18:01 sakoman: the scanner arrived: http://www.flickr.com/photos/koenkooi/tags/yashicamat124g/ Sep 01 14:57:23 koen|away: Its tricker than that, the version of glibc we ship with the external toolchain (armv5te or armv4t) determines which type of builds the toolchain will work for Sep 01 15:02:35 morning Sep 01 15:03:16 hi chouimat Sep 01 15:59:55 03mickeyl 07org.oe.dev * red77cd74... 10/ (4 files in 3 dirs): uboot-openmoko: create known-good version as seperate bb, make git/svn really uptodate again Sep 01 17:36:54 RP: so we need to make sure the external toolchains are armv4t? Sep 01 17:38:09 RP: people that care about OMG OPTIMIZED! can use OE proper Sep 01 17:49:33 hello Sep 01 17:50:38 privet :> Sep 01 17:53:34 koen: Yes, or we fix the external toolchains to handle this. It shouldn't be that difficult... Sep 01 17:53:35 Crofton|work: still could not build your image because of the damn uboot git that is offline Sep 01 17:53:51 can I turn uboot off somehow and build without it? Sep 01 18:33:11 koen: Nice photos! The new scanner makes an amazing difference in quality over the ghetto scans :-) Sep 01 18:33:44 Though for some types of photo the ghetto scan could work -- nice & gritty Sep 01 18:38:02 grainy works quite nice for B&W Sep 01 18:39:07 and even a 10MPixel it has detail to spare Sep 01 18:39:31 * koen will scan the next roll at 40MPixel (3200DPI) Sep 01 18:43:04 hi florian Sep 01 18:44:29 re Sep 01 18:47:59 koen: which scanner did you end up getting? Sep 01 18:48:12 canon canoscan 8600f Sep 01 18:48:25 I wanted the lide600f (smaller, usb bus powered) Sep 01 18:48:47 but the 8600f has an integrated 120 film scanner util thingy Sep 01 18:50:19 * koen discovered that fotolabs print in 8bit/channel nowadays Sep 01 18:50:27 Nice, really reasonable price too Sep 01 18:50:57 it has scratch& dust removal hardware as well Sep 01 18:51:21 but that uses UV (or IR, I always confuse the two) which doesn't like the silver in b&w films Sep 01 18:51:21 Makes me sorry I sold my Yashica-MAT :-( Sep 01 18:51:54 Have to drag out the old 8 x 10 to play with Sep 01 18:52:11 a tlr makes people come out and ask to be photographed :) Sep 01 18:52:28 Cool if there was an 8 x 10 image sensor for it :-) Sep 01 18:52:36 :) Sep 01 18:52:43 Yes, there's something really cool about old photo tech Sep 01 18:55:40 a small downside is that photoshop CS2 is still powerpc Sep 01 18:56:44 Don't have an old Mac lying around? Sep 01 18:58:02 yes, but that has only 1.2GB of ram and is low on disk space Sep 01 18:58:15 (and is property of my girlfriend) Sep 01 18:59:13 You could always buy her a new one and upgrade her old machine for photoshop :-) Sep 01 18:59:33 or wait for CS3 :) Sep 01 18:59:52 Took the words out of my mouth ! Sep 01 19:00:46 or poke pippin to finish the gimp port to gegl so it can use >8 bit/channel Sep 01 19:01:26 03rpurdie 07org.oe.dev * r96da4aa7... 10/ (1 conf/bitbake.conf): bitbake.conf: Define various variables (from poky) Sep 01 19:01:33 03rpurdie 07org.oe.dev * ra25a7f85... 10/ (1 classes/package.bbclass): package.bbclass: Improve strip function and convert to python (from poky) Sep 01 19:01:38 03rpurdie 07org.oe.dev * r2db32af5... 10/ (1 packages/jards/jards_1.0.bb): jards: Remove insane pkgdata references Sep 01 19:01:43 03rpurdie 07org.oe.dev * r1550bfad... 10/ (5 files in 2 dirs): classes: Convert to PKGDATA_DIR variable (from poky) Sep 01 19:01:50 03rpurdie 07org.oe.dev * r56e0d68d... 10/ (1 classes/multimachine.bbclass): multimachine.bbclass: Add PKGDATA_DIR defintion, disabled for now (from poky) Sep 01 19:01:55 03rpurdie 07org.oe.dev * rb89d5fe7... 10/ (1 classes/base.bbclass): base.bbclass: Fix SRC_URI overrides handling, lower the priority of local file url checksum messages, correct populate_staging dependencies (from poky) Sep 01 19:02:01 03rpurdie 07org.oe.dev * r91de550a... 10/ (1 classes/rootfs_deb.bbclass classes/rootfs_ipk.bbclass): rootfs_{deb|ipk}.bbclass: Fix whitespace damage, remove bashisms, sync with poky Sep 01 19:02:07 03rpurdie 07org.oe.dev * r1d4fc7d0... 10/ (53 files in 5 dirs): images: Clean up image .bb files, moving functionality to image.bbclass (LICENCE default, setup of RDEPENDS and default IMAGE_BASENAME) Sep 01 19:02:13 03rpurdie 07org.oe.dev * r6fb8b044... 10/ (1 classes/native.bbclass): native.bbclass: PACKAGE_DEPENDS no longer need be unset, add extra options to PKG_CONFIG_PATH if EXTRA_NATIVE_PKGCONFIG_PATH is set (from poky) Sep 01 19:02:19 03rpurdie 07org.oe.dev * r2707256d... 10/ (1 classes/image.bbclass): image.bbclass: Append to do_rootfs[recrdeptask] Sep 01 19:02:25 03rpurdie 07org.oe.dev * r5941cc71... 10/ (1 classes/sdk.bbclass): sdk.bbclass: Use variables from bitbake.conf (from Poky) Sep 01 19:02:28 wau Sep 01 19:02:31 03rpurdie 07org.oe.dev * r35178105... 10/ (6 files in 2 dirs): gcc4-sdk: Use sysroot to workaround various path problems with external toolchains (from Poky) Sep 01 19:02:32 spam! Sep 01 19:02:36 03rpurdie 07org.oe.dev * r4e2bc629... 10/ (1 packages/gcc/gcc-cross-sdk_4.1.2.bb): gcc-cross-sdk-4.1.2: Second try and cleaning up SRC_URI (if this fails for anyone, please post the errors to the list) (from poky) Sep 01 19:02:43 03rpurdie 07org.oe.dev * rd3e55a61... 10/ (3 files in 2 dirs): binutils-cross-sdk: 2.17.50.0.5 -> 2.17.50.0.12 and merge in sysroot fix from poky Sep 01 19:02:49 03rpurdie 07org.oe.dev * re3299c40... 10/ (1 packages/gcc/gcc-package-sdk.inc): gcc-package-sdk.inc: Update packaging after sysroot changes (from poky) Sep 01 19:03:00 03rpurdie 07org.oe.dev * r476f9bcf... 10/ (11 files in 2 dirs): gdb: Sync with poky (metadata cleanups, addition of gdb-cross-sdk) Sep 01 19:03:07 03rpurdie 07org.oe.dev * ra8fc829c... 10/ (1 packages/gdb/files/sim-install-6.6.patch): gdb: Add missing patch Sep 01 19:04:24 * koen is going to find out how similar RP and me make changes Sep 01 19:13:27 server -> cable modem -> internet -> adsl -> wifi -> osx Sep 01 19:13:40 that's not a good scenario for running meld over X Sep 01 19:24:54 Is anyone currently using the Nokia N800 here? Sep 01 19:26:17 yes, one just spontaniously rebooted itself behind me Sep 01 19:27:15 Koen, LOL what are your thought on it? I've been checking out all the handhelds out there for one to start hacking on and I, so far, keep getting drawn back to it. Sep 01 19:27:35 the default nokia software is buggy Sep 01 19:27:39 nice hardware Sep 01 19:27:58 Koen, yes I'm not very concerned about the software since I plan to OE it. Sep 01 19:29:19 Koen, I've only really used the PXA270 so that is another reason I like it as it would be a good opertunity to learn a new processor. I saw it is only 330 MHz but that is the TI processor which I believe has a graphics acc. plus a dsp on board.. Sep 01 19:30:10 all the specs I can find say it has USB high speed ports but does not say if it is client or host ports. Sep 01 19:30:46 12mbit client Sep 01 19:30:58 12mbit unpowered host with kernel patch and funky cable Sep 01 19:31:09 I also have read the video on it is very buggy but I'm assuming that is do to the default software. Sep 01 19:31:32 svolpe_gerrath I have a n770, pretty nice hardware, but is a little to slow, and not enough RAM. The n800 should fix all that Sep 01 19:31:50 svolpe_gerrath: I'm still running nokia's sw on it Sep 01 19:32:14 koen, thats a shame, I was hoping for a USB host, however a funky cable is not an issue to me. Sep 01 19:32:29 gah Sep 01 19:32:46 I fixed the oe website and after enabling one module its broken again Sep 01 19:32:50 koen, and I can always use an external power source for USB devices. Sep 01 19:32:51 * koen did narrow it down Sep 01 19:33:08 cbrake, never got time to hack it huh :-) Sep 01 19:33:44 svolpe_gerrath: yeah, I've got a long list of device to hack before the 770 -- AT91SAM9260 EK, AVR32 board, openmoko phone ... Sep 01 19:34:07 urgs Sep 01 19:34:08 svolpe_gerrath: the n770 was worth buying for one reason -- the power management is very interesting. Sep 01 19:34:09 avr32 Sep 01 19:34:35 svolpe_gerrath: they never suspend the device, they just agressively power manage everything. Sep 01 19:34:56 svolpe_gerrath: the battery life is pretty impressive, and the thing always responds instantly Sep 01 19:35:41 cbrake, thats how our product is :-) we do suspend but only to the mem level so it wakes up instantly. Sep 01 19:36:00 koen: what is the status of the avr32 stuff in OE? Sep 01 19:36:42 cbrake, I would rather get the openmoko phone but I don't seem to be able to find one :-/ Sep 01 19:36:53 svolpe_gerrath: cool -- that is the way to do things. Sep 01 19:36:58 hi crofton Sep 01 19:37:15 hi Sep 01 19:37:27 sitting in a field watching people do weird things Sep 01 19:38:02 svolpe_gerrath: I really like the neo1973. Guess you could get a n800 now, and the next rev of the neo when it comes out :-) Sep 01 19:38:27 cbrake: see privmesg Sep 01 19:38:45 cbrake, did you have a neo1973? Sep 01 19:39:54 svolpe_gerrath: yes, I have a neo1973 Sep 01 19:40:06 cbrake, were did you get it from? Sep 01 19:40:13 cbrake, does it work as a phone? Sep 01 19:40:32 svolpe_gerrath: bought it from FIC -- I must have ordered early enough ... Sep 01 19:40:51 Crofton: I've not tried the phone part yet -- still need to get a sim phone. Sep 01 19:41:09 ah Sep 01 19:41:18 I have a sim phone already Sep 01 19:41:20 Crofton: I like the HW pretty well, but looks like it will be a bit before the sw is good enough to throw away my Alltel phone .. Sep 01 19:41:35 need to finish th is friggin MS thesis :) Sep 01 19:42:21 Crofton: I'm working with the mono folks to get mono running on it. There is some interest there, and I have a customer running mono. Sep 01 19:42:41 ah Sep 01 19:43:08 cbrake: talked to alp toker yet? Sep 01 19:43:15 koen: no I have not Sep 01 19:43:25 cbrake: that's a mono dude with a clue Sep 01 19:43:31 crap Sep 01 19:43:39 the gumstix have been delivered Sep 01 19:43:48 but I won't be home until Monday ... Sep 01 19:43:53 heh Sep 01 19:43:56 koen: is he interested in ARM, or what does he do with mono? Sep 01 19:44:10 need to see if my wife saw the package and took it in Sep 01 19:44:19 cbrake: he wants to push mono (minus winforms and other crap no one cares about) into GMAE Sep 01 19:44:44 cbrake, Looks like you can still buy them from openmoko site. Sep 01 19:44:58 koen: great. Does he have a blog or anything to keep tabs on what he is doing? Sep 01 19:45:19 cbrake: http://www.atoker.com/blog/ Sep 01 19:45:27 cbrake: he also hacks on webkit/gtk :) Sep 01 19:46:00 I wish she had bought them up here Sep 01 19:46:12 koen: excellent, thanks! Sep 01 19:50:40 re zecke Sep 01 19:56:59 hi Sep 01 19:57:14 hi likewise Sep 01 19:57:52 hey likewise Sep 01 19:58:25 crofton: good thing she didn't bring them or you wouldn't get a thesis work done Sep 01 19:58:36 heh Sep 01 19:58:42 having fun? Sep 01 19:58:44 yeah Sep 01 19:58:50 it is not raining Sep 01 19:59:06 now we are telling canadian jokes to a Canadian Sep 01 19:59:10 Very hot here -- 105F Sep 01 19:59:16 urg Sep 01 19:59:36 Canadian jokes are always fun, eh? Sep 01 19:59:40 hi ph5 Sep 01 19:59:50 eh Sep 01 19:59:50 hej woglinde Sep 01 19:59:58 yeah, he has a very subtle eh Sep 01 20:00:15 varies with area of origin Sep 01 20:00:39 cbrake, I've been going for a month now on trying to decide which portable device to buy. I just want to hack it but its a tough choice :-). Sep 01 20:01:25 I lived in Hamilton for a few years Sep 01 20:01:44 My Aunt is on Toronto Sep 01 20:02:22 svolpe_gerrath: in my opinion, the openmoko is the most hackable because the BL and everything is open, and fairly clean. But, then if you like reverse engineering, maybe its not the thing to get :-) Sep 01 20:02:36 Never lived there, had lots of Canadian co-workers Sep 01 20:02:46 cbrake, no I prefer not to reverse engineer. Sep 01 20:03:19 svolpe_gerrath: yeah, I find it refreshing to work on a device where the BL is open, after working on all these compulab projects ... Sep 01 20:04:47 cbrake, so it does have a gps in it? Sep 01 20:04:56 svolpe_gerrath: dunno -- looking ... Sep 01 20:04:57 koen: What are the PREFERRED_PROVIDERS doing in the angstrom image files? Sep 01 20:05:45 RP: the thought was that different images need different PREFERRED_PROVIDERS Sep 01 20:05:54 e.g opie-image and x11-image Sep 01 20:06:17 koen: It won't have any effect at all Sep 01 20:06:27 cbrake, just checked it out, nice even has USB host. Sep 01 20:06:31 RP: ah Sep 01 20:06:41 cbrake, a little slow on the processor but I guess that is because it IS a phone. Sep 01 20:06:44 RP: you noticed I don't build opie-images :) Sep 01 20:07:21 svolpe_gerrath: yeah, that is nice. Looks like GPS in included, but the driver is currently binary only ... Sep 01 20:07:32 koen: ;-) Sep 01 20:07:44 cbrake: http://projects.linuxtogo.org/plugins/scmsvn/viewcvs.php/?root=sphyrna :) Sep 01 20:08:26 koen: heh, I could have assumed ... Sep 01 20:08:52 svolpe_gerrath: you could still get a little bit of reverse engineering in on the neo :-) Sep 01 20:11:35 cbrake, have you looked into its processor yet? (Samsung S3C2410AL-26 )? Sep 01 20:12:01 svolpe_gerrath: 2410 is old and slow, but its a decent processor Sep 01 20:12:20 looks like it is an amr9 Sep 01 20:12:57 arm920t Sep 01 20:13:11 03rpurdie 07org.oe.dev * rbcabf2ef... 10/ (1 packages/meta/meta-toolchain.bb): meta-toolchain: Fix whitespace and merge in some improvements from poky Sep 01 20:13:19 03likewise 07org.oe.dev * r155c9f46... 10/ (1 packages/glibc/eglibc_svn.bb): eglibc_svn: get_glibc_fpu_setting must be .._eglibc_.. Sep 01 20:13:20 svolpe_gerrath: the next version of the neo will have the 2442, which should be a good bit faster. Kind of like the n770 vs the n800 Sep 01 20:14:51 svolpe_gerrath: but the big difference will be 2d acceleration. A VGA resolution screen really needs accelaration when driven by an ARM ... Sep 01 20:15:59 03koen 07org.oe.dev * r3620da36... 10/ (1 classes/rootfs_deb.bbclass): rootfs_deb: make sure directories exist before trying to symlink them Sep 01 20:16:30 cbrake, have they posted when the 2d will be out? Sep 01 20:16:56 svolpe_gerrath: only that it will be in the gta02 model Sep 01 20:17:12 svolpe_gerrath: That is very strange. I would copy a simple php script into main_site and see if that runs. You should not have to be in the root directory to run a php script, unless that is a godaddy restriction -- you may want to contact their tech support. Sometimes there are .htaccess files in directories that enable/disable various features. Sep 01 20:17:25 argh, wrong past buffer!!! Sep 01 20:18:10 svolpe_gerrath: http://wiki.openmoko.org/wiki/FAQ Sep 01 20:20:39 ah, my second buildhost is without local diff again :) Sep 01 20:21:11 cbrake, so in Sept. they will let us know what is coming out and when . Sep 01 20:22:10 given the past, that will be in oct. Sep 01 20:22:34 'night all Sep 01 20:29:26 cbrake, lol, I'm still no further along in making a choice :-/ Sep 01 20:30:10 svolpe_gerrath: that is the problem, you can wait forever as there is always something better just a few months in the future ... Sep 01 20:31:14 ping anyone. freenode.org went down? Sep 01 20:31:24 cbrake, exactly, I like the N800 because it has wifi, a nice screen and decent processor, but no USB host and no GPS (unless you pay a lot for the bluetooth GPS module.) Sep 01 20:31:49 cbrake, the openmoko is great but its a little slow and no wifi. Sep 01 20:32:02 sorry for resend, was disconnected: I am getting these opie errors again, I thought this was solved lately? Sep 01 20:32:03 "${@[”, 'opie-help-en (>= ${OPIE_CVS_PV})'][packaged(bb.data.getVar('PN', d, 1), d) == True]} Sep 01 20:32:05 ERROR: EOL while scanning single-quoted string (" Sep 01 20:36:06 flo_lap: hi, do you have some time to help me with oe bug nr 2598 ? Sep 01 20:36:30 hi mickeyl Sep 01 20:36:44 hi woglinde Sep 01 20:38:53 mr_nice: muhhaha "time" ;) Sep 01 20:39:03 * flo_lap takes a look Sep 01 20:39:48 !oebug 2598 Sep 01 20:39:49 * * Bug 2598, Status: NEW, Created: 2007-07-04 07:06 Sep 01 20:39:51 * * : angstrom simpad gpe: metabug suspend with power button does not work Sep 01 20:39:52 * * http://bugs.openembedded.org/show_bug.cgi?id=2598 Sep 01 20:40:52 mr_nice: right, I noticed this too when I got rid of my last Familiar two weeks ago :) Sep 01 20:41:02 cbrake, and lets not forget the Sharp Zaurus SL-C1000 which is a nice device as well. Sep 01 20:42:55 flo_lap: hehe Sep 01 20:43:12 mr_nice: I left my Simpad at work - we need to check which key event the button is mapped to. Sep 01 20:43:54 keylaunch should die... like ipaq-sleep Sep 01 20:45:09 flo_lap: the kernel key event ? Sep 01 20:45:13 mickey|mdd: Is there something to replace this in OpenMoko already? Sep 01 20:45:50 mr_nice: well... yes Sep 01 20:46:11 what does ipaq-sleep ? Sep 01 20:46:14 +do Sep 01 20:46:19 flo_lap: the kernel driver keycode for the power button is 116 Sep 01 20:49:01 mickey|mdd: It dims lights and suspends the device if it is idle. Sep 01 20:49:29 mr_nice: ok, do you know to which X event this is maped? Sep 01 20:50:28 flo_lap: ah. neod does this currently, but it is hardcoded to the Neo. This functionality is supposed to be taken care of by the forthcoming hardware abstraction -- be that OHM, deviced, machined, or whatever Sep 01 20:52:45 flo_lap: hm, I think no X event is maped to it. I have maped it manualy to XF86PowerDown. Sep 01 20:54:36 mickey|mdd: ok... time for me to take a look at ohm, i need to get a concept for this into both gpe and g(pe)^2 Sep 01 20:55:41 :) Sep 01 20:56:35 mr_nice: ok, then we need to do this in the simpad xmodmap and ship a matching keylaunchrc Sep 01 20:58:14 flo_lap: well I have some trouble with keylaunch oe bug 2599. and I have no clue why the mapping is wrong for keylaunch oe bug 2600 Sep 01 20:59:24 mr_nice: Currently we map 102 to XF86PowerDown, i'll change this to 116 now Sep 01 21:01:00 flo_lap: in simpad.xmodmap? Sep 01 21:01:37 mr_nice: yes Sep 01 21:01:51 flo_lap: ok Sep 01 21:04:35 flo_lap: ya, if you have time to review OHM taking our needs into account, I'd be grateful for an analysis Sep 01 21:04:58 Assuming I have a bitbake recipe named "mwester-something-1.0+cvs.bb" -- how does one build that with the bitbake command, and how can I specify precisely what SRCDATE to fetch? Sep 01 21:05:35 i'd suggest using an underscore Sep 01 21:07:40 Er, yes, mistyped. More specifically, it would be "mwester-native_1.0+cvs.bb" Sep 01 21:08:00 mickey|mdd: ok, maybe i can do this... we need a solution. Sep 01 21:09:33 ya Sep 01 21:09:48 ;) and no, I don't know how to build "mwester_1.0+cvs.bb" either -- nothing I've tried results bitbake actually finding the recipe and building anything. Sep 01 21:12:28 mickey|mdd: I try to fit this in my work before OSiM, free time hacking is cancelled for renovations this month. Sep 01 21:38:23 hi Sep 01 21:38:36 mwester: you can set SRCDATE in recipe itself Sep 01 21:39:40 * hrw -> sleep rather Sep 01 21:46:53 Would it be possible to port maemo to Angstrom? Sep 01 21:56:41 mwester: would you like me to post your kecec patchec to the bt/list ? I think they would be very useful :) Sep 01 21:56:58 zap: to some extent yes, we have some components in oe already Sep 01 21:57:17 zap: but only a few are up to date Sep 01 22:02:36 mwester: bitbake mwester-native should do the job... Sep 01 22:06:31 mr_nice: I would expect it to work with the default keylaunchrc now. Sep 01 22:09:04 flo_lap: ? Sep 01 22:10:08 flo_lap: with xmodmap.simpad changed to 116? Sep 01 22:12:00 mr_nice: yes Sep 01 22:14:13 flo_lap: I have tryed this allready without success. but I will try it now a second Sep 01 22:19:04 mr_nice: how does your keylaunchrc look like? Sep 01 22:21:35 flo_lap: http://pastebin.ca/678380 Sep 01 22:26:03 03rpurdie 07org.oe.dev * ra2c72b19... 10/ (6 files in 3 dirs): images: Remove bogus PREFERRED_PROVIDERs (they don't work in .bb files) Sep 01 22:26:33 mr_nice: eeks! # Sep 01 22:26:33 *** glibc detected *** keylaunch: free(): invalid pointer: 0x0001ab11 *** Sep 01 22:26:33 # Sep 01 22:26:33 Aborted Sep 01 22:27:40 mr_nice: I see, this won't work Sep 01 22:32:45 ~lart keylaunch Sep 01 22:32:45 * ibot offers keylaunch some herring Sep 01 22:32:46 flo_lap: do you have any Idea what I can do to make it work? Sep 01 22:32:53 flo_lap: I could write a low level demon for it which would be x11 independent. Sep 01 22:33:27 flo_lap: but I have no clue if this is a good idea Sep 01 22:34:02 mr_nice: that would be an idea, but i'm not sure if it is easy to handle some events before they are passed to x Sep 01 22:35:09 mr_nice: try what happens if you clear your keylaunchrc and remove all these lines handling keys that are not defined. Sep 01 22:35:48 flo_lap: the power buttons driver is gpio key which is only used for power button. I can read out the /dev/ file and block the driver. Sep 01 22:36:56 * flo_lap would ditch the keylaunch implementation and write a new daemon like this Sep 01 22:37:15 that could handle hinge sensor input too Sep 01 22:37:24 flo_lap: ok I will write such a daemon Sep 01 22:37:38 * RP mentions zaurusd and its switch event daemon Sep 01 22:37:57 :-) Sep 01 22:38:22 flo_lap: simpads switches driver is evdev based Sep 01 22:38:40 * flo_lap is convinced tat this would be way better Sep 01 22:38:41 ups I meant RP Sep 01 22:39:13 mr_nice: Add simpad support to zaurusd, it should have everything you need for the hinge sensor ;-) Sep 01 22:39:27 hehe Sep 01 22:39:30 A power button daemon would be nice too... Sep 01 22:40:32 I will write such a daemon. basically I have halv of the code already developt for my thumbscript userspace driver Sep 01 22:41:39 I should take a look at zaurusd as well Sep 01 22:42:34 mr_nice: Please do since there is a ready made switch event daemon there already Sep 01 23:33:40 03rpurdie 07org.oe.dev * r4ab5dc0f... 10/ (1 classes/package_ipk.bbclass): package_ipk.bbclass: Sync with poky for locking, code to add package_write_ipk task (not activated yet), change OE: field to use PN (not P), cleanups and conf and index generation functions Sep 01 23:33:46 03rpurdie 07org.oe.dev * r2330dd11... 10/ (1 conf/bitbake.conf): bitbake.conf: Add STAGING_ETCDIR_NATIVE variable Sep 01 23:33:56 03rpurdie 07org.oe.dev * r41ff5b51... 10/ (1 classes/rootfs_ipk.bbclass): rootfs_ipk.bbclass: Convert to use package_ipk function Sep 01 23:34:04 03rpurdie 07org.oe.dev * r3154e628... 10/ (1 packages/meta/meta-toolchain.bb): meta-toolchain: Convert to use package_ipk fnuctions Sep 01 23:34:10 03rpurdie 07org.oe.dev * r669fa4ac... 10/ (3 files in 3 dirs): meta-toolchain: Switch to use task-sdk-host and allow TOOLCHAIN tasks to be overridden by the distro Sep 01 23:45:46 ok the power off daemon is comlete for simpad Sep 01 23:45:55 and works well Sep 01 23:46:05 I will upload the source now Sep 01 23:46:35 mickey|mdd: Did the splash screen issue with the /dev entries get fixed? Sep 01 23:46:48 through what? Sep 01 23:47:12 mickey|mdd: I don't know, I just mean is it still broken? Sep 01 23:47:19 i mean, did anything do anything to it? Sep 01 23:47:35 i have no idea, i reverted your patch on buildhost. i can build again tomorrow and have a look Sep 01 23:48:12 It went quiet, I just wondered if that meant it was fixed :} Sep 01 23:48:41 mickey|mdd: meta-toolchain in OE is near identical to the one in poky now btw Sep 01 23:48:47 great! Sep 01 23:48:49 thanks a lot Sep 01 23:48:56 does that package up bin now? Sep 01 23:49:02 probably, i'll try as well tomorrow Sep 01 23:49:55 mickey|mdd: It doesn't work by packaging up bin, it builds a separate new compiler which can be installed in a specific location Sep 01 23:50:27 mickey|mdd: The specific location being SDK_PREFIX in bitbake.conf. Poky uses /usr/local/poky Sep 01 23:50:28 ya, i know, but the last tar was missing bin Sep 01 23:50:51 This should work better... Sep 01 23:50:54 cool Sep 01 23:51:00 it's still not relocatable though, right? Sep 01 23:51:11 Sadly not Sep 01 23:52:12 mickey|mdd: If you want to ASSUME_PROVIDED this toolchain, you need the other half of the magic in the form of poky-external-toolchain.bb Sep 01 23:52:23 ok Sep 01 23:52:33 most important is standalone now Sep 01 23:52:36 (for me) Sep 01 23:52:51 If OE can use it, it works standalone too :) Sep 01 23:53:00 how do we proceed with different sdk's based on that Sep 01 23:53:05 do we include and then add packages? Sep 01 23:53:59 If you look at the file, it sets TOOLCHAIN_TARGET_TASK ?=. If the distro overrides that to its own SDK task it should be customisable Sep 01 23:54:37 You also have TOOLCHAIN_HOST_TASK but I doubt people will need to touch that one Sep 01 23:55:40 excellent _) Sep 01 23:59:27 Note that I've tried not to change any behaviours in the commits so far. There are some behaviour changes we should enable at some point. I guess I'd better mail the list about that... Sep 02 00:00:14 The empty /dev problem still exists btw :-( Sep 02 00:00:59 darn Sep 02 00:01:40 * RP has a sudden hunch Sep 02 00:04:49 :) Sep 02 00:05:00 * mickey|mdd goes on to some ezx hacking w/ Stefan... cu tomorrow Sep 02 00:05:03 thanks for alL Sep 02 00:05:04 ! Sep 02 00:18:09 here is the source for the power off daemon. http://slackpad.svn.sourceforge.net/viewvc/slackpad/trunk/powerd/ . I will create a init and .bb file for the simpad tomorrow. Sep 02 00:37:30 'night all Sep 02 00:54:57 * * OE Bug 2907 has been created by cesarb(AT)cesarb.net Sep 02 00:54:59 * * IndentationError on package_ipk.bbclass Sep 02 00:55:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2907 Sep 02 01:09:55 hi, good night, i don't access openemdded documentation, how create login for me ? Sep 02 01:36:57 * * OE Bug 2907 has been RESOLVED (FIXED) by cesarb(AT)cesarb.net Sep 02 01:36:59 * *  IndentationError on package_ipk.bbclass Sep 02 01:37:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2907 **** ENDING LOGGING AT Sun Sep 02 02:59:57 2007