**** BEGIN LOGGING AT Tue Aug 12 02:59:56 2008 Aug 12 06:57:29 moin Aug 12 07:44:29 hi, i wonder why the xf86-video-i810 packages will not be build if I try to build xorg instead of kdrive in ord.openembeded.stable? Aug 12 07:44:32 any clue Aug 12 08:25:50 bonjour Aug 12 09:44:47 florian: good morning Aug 12 09:47:02 hi all Aug 12 11:26:31 oe.org back online!! :P Aug 12 11:27:04 (at least apache tells me that it works ^^ ) Aug 12 11:30:50 josch|nsn: We are working on a waorkaround to get a sane message instead of a timeout. :-( Aug 12 11:31:20 take your time :) Aug 12 11:35:05 josch|nsn: unluckily we do not yet have an idea what the reason of the failure is. Aug 12 11:35:15 huh? Aug 12 11:35:20 why's that? Aug 12 11:41:02 * XorA wonders if anyone checked is the machine still in the rack :-) Aug 12 11:42:48 :) Aug 12 11:43:52 mickey|tv is the owner of the machine - its hosted by Laf0rge somewhere in the Berlin area iirc. mickey|tv is located in Frankfurt so he does not have easy access to the machine. Aug 12 11:44:21 it was caught in a GPL violation :-D Aug 12 11:45:14 heheh Aug 12 11:45:21 Its likely that its just a matter of pressing the reset switch, but obviously noone checked so far. Aug 12 11:45:53 time to send zecke Aug 12 11:46:03 send a ninja Aug 12 11:46:12 send rms with two katanas! Aug 12 12:04:02 josch|nsn: http://xkcd.com/225/ ? Aug 12 12:06:48 sending rms should alone be enough to scare the server to reset itself Aug 12 12:42:01 morning Aug 12 12:42:35 morning Aug 12 12:42:43 bye ;) Aug 12 12:46:58 florian: ltg needs to enable the oe.org web redicrection again, otherwise a dns change won't lead to oe.linuxtogo.org, but just the ltg main site Aug 12 12:59:12 mickeyl, I emailed this guy www.blacksburg.net to see if he has cheap colocation Aug 12 12:59:31 I can always put something together so we can split the load Aug 12 13:00:14 this thing in Berlin is a very cheap computing centre, that's the problem Aug 12 13:00:28 it's likely not the server, but one of their routers Aug 12 13:00:49 yeah Aug 12 13:01:08 * josch|nsn said cheap collocations goodbye some time ago as he repeatedly had similar errors Aug 12 13:01:11 I am annoyed, and if the price is ok I can help prevent future issues Aug 12 13:01:34 even if this guy is cheap, I usually eat lunch within 200 yds of the location :) Aug 12 13:01:46 that's helpful :) Aug 12 13:01:54 yeah Aug 12 13:02:11 Lyle basically makes a living selling dialup and hosting Aug 12 13:02:18 that's right, dial up Aug 12 13:02:43 eye. selling dialup can pay your debts? Aug 12 13:03:20 he's been doing it for a while Aug 12 13:03:51 we'll see what he says Aug 12 13:05:12 http://thedailywtf.com/Articles/ITAPPMONROBOT.aspx Aug 12 13:08:47 heh Aug 12 13:10:12 so the down of OE.org is more a matter of money ? Aug 12 13:12:13 that would be a bit far fetched Aug 12 13:12:27 we have a setting where i own the server, but someone else donates the hosting cost Aug 12 13:12:42 this is the one responsible for the colocation in a berlin computing centre Aug 12 13:13:22 mickeyl: oh, is oe.org working again now? it seems to ping for me all of a sudden. Aug 12 13:13:48 it seems that my local monotone db is hosed though so a pull still fails :-} Aug 12 13:14:04 no Aug 12 13:14:06 still down Aug 12 13:14:13 i just switched www.oe.net back Aug 12 13:14:20 so that we at least get a site Aug 12 13:14:32 ooh Aug 12 13:14:33 wait Aug 12 13:14:36 IT'S BACK!!!!!!!!!!!!!!!!!!!!!! Aug 12 13:14:38 mickeyl, where is zecke? Aug 12 13:14:42 ~hail mickeyl Aug 12 13:14:43 * ibot bows down to mickeyl and chants, "I'M NOT WORTHY!!" Aug 12 13:14:44 yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyeeeeeeeeeeeeeaaaaaaaaaaaaaaaah Aug 12 13:14:45 \o/ Aug 12 13:14:46 wow Aug 12 13:14:56 ok, this night I'm going to do a _full_ backup Aug 12 13:15:06 heh Aug 12 13:15:08 so that we can move this to a virtual instance if this will ever happen again Aug 12 13:15:16 * Genesis go edit his userspace with his dreams^^^todolist =D Aug 12 13:15:31 mickeyl: See my offer on oe-devel - I'll aim to host a mirror Aug 12 13:15:40 RP: righto, pretty cool Aug 12 13:16:12 * Jay7 can provide some help with system administration if requested Aug 12 13:17:01 Crofton|work: zecke is in .tw, I think Aug 12 13:17:42 I was afraid of that Aug 12 13:17:52 we could have sent him on a mission :) Aug 12 13:18:16 mwester, Crofton|work: Actually he is back in berlin, so you can send him :) Aug 12 13:18:27 too late amethyst came back Aug 12 13:18:34 but then again, maybe he did it Aug 12 13:18:38 heh Aug 12 13:18:53 zecke is back in Berlin Aug 12 13:18:54 ah yes Aug 12 13:19:01 maybe he just power cycled the whole data center from outside Aug 12 13:19:03 ~hail zecke Aug 12 13:19:04 * ibot bows down to zecke and chants, "I'M NOT WORTHY!!" Aug 12 13:22:14 hehehe Aug 12 13:22:38 * mickeyl imagines zecke using his kick-ass projector playing a NES game to produce a power outage in Berlin Aug 12 13:25:21 :-D Aug 12 13:47:14 good morning Aug 12 13:50:25 Jay7: what's up? Any news? Aug 12 13:50:30 Ant_Office: hi! how is new u-boot? :) Aug 12 13:50:45 not yet seen... Aug 12 13:50:48 UFO Aug 12 13:50:58 Ant_Office: I have new ISP. It is main news :) Aug 12 13:52:05 Jay7: have you read the logs I posted to thesing earlier? Aug 12 13:52:37 Ant_Office: no Aug 12 13:52:46 is here web-archive? Aug 12 13:52:49 Jay7: http://hentges.net/tmp/logs/irc/livelogs/%23oe.livelog?lines=500#latest Aug 12 13:53:04 yes, there is archive too Aug 12 13:53:22 morning Aug 12 13:54:27 Jay7: about rotated screen I suppose changing FBANGLE = "270" in kexecboot/initramfs-kexecboot_1.0.bb could help Aug 12 14:03:24 Ant_Office: I've read Aug 12 14:04:42 Ant_Office: kexec does not take u-boot's cmdline and atags? Aug 12 14:05:26 Jay7: no, uses own cmdline, the defaultis in /boot Aug 12 14:06:25 then new cmdline/atags should be prepared before kexec'ing Aug 12 14:06:41 e.g. via reading u-boot's env Aug 12 14:06:56 Jay7: well, I think there are several oprions Aug 12 14:07:12 butreading u-boot env is the best Aug 12 14:07:56 e.g. mtdparts=mtdparts=laze-0:256k(U-Boot)ro,128k(U-BootEnv),5376k(Emergency),1408k(Kernel),-(JFFS2) Aug 12 14:08:16 one would not need regexp Aug 12 14:09:59 Jay7: dunno if passed ATAG_CMDLINE is valid... Aug 12 14:10:17 Ant_Office: we should test new u-boot Aug 12 14:10:47 may be atags there is working and we don't needed to patch cmdline Aug 12 14:11:04 he, hopefully Aug 12 14:11:23 * Jay7 -> some food Aug 12 14:11:31 bbl Aug 12 14:21:49 in which case a python lib should be package as packages/python /python-mylibs instead of being packaged in packages/mylib ? Aug 12 14:26:52 03  07org.openembedded.dev * rba85c02150 10OE.dev/conf/machine/beagleboard.conf: beagleboard: switch to using xorg Aug 12 14:27:29 03  07org.openembedded.dev * rcfb6ed1d29 10OE.dev/packages/linux/linux-neuros_git.bb: linux-neuros: bump SRCREV for more fb fixes Aug 12 14:27:29 03  07org.openembedded.dev * rf6c1c0a798 10OE.dev/conf/distro/ (2 files in 2 dirs): angstrom 2008: prefer a newer glib and gtk+ Aug 12 14:27:30 03  07org.openembedded.dev * r71ad63a930 10OE.dev/packages/tasks/task-base.bb: task-base: bump PR Aug 12 14:27:31 03  07org.openembedded.dev * rb09a7bad1b 10OE.dev/packages/neuros-public/ (neuros-nwm/init neuros-nwm_git.bb): neurow-nwm: add correct initscript Aug 12 14:27:34 03  07org.openembedded.dev * r29fac407cf 10OE.dev/packages/linux/ (linux-neuros/neuros-osd2/defconfig linux-neuros_git.bb): linux-neuros git: enable oprofile Aug 12 14:27:37 03  07org.openembedded.dev * r1156f540f6 10OE.dev/packages/mythtv/ (gmyth-upnp_0.7.0.bb gmyth_0.7.1.bb): gmyth*: fix up .pc files for -dev packages as well Aug 12 14:27:40 03  07org.openembedded.dev * r8137f57885 10OE.dev/: Aug 12 14:27:42 merge of '5a2fb2965a905224af0a1e26d73f342ecf8944e8' Aug 12 14:27:44 and '6fc6bb1fff3ea7b5546ddedf048c33f9a287ac15' Aug 12 14:27:46 03  07org.openembedded.dev * rda961bc161 10OE.dev/: Aug 12 14:27:48 merge of '003d5589c32126d84385935351087e649339386a' Aug 12 14:27:56 and 'faac6daf1f748c6f7109cbc33bb0af2fbe932c77' Aug 12 14:27:58 03  07org.openembedded.dev * r2cea9cc64a 10OE.dev/contrib/source-checker/oe-source-checker.py: oe-source-checker.py: use regular oe_sha256sum since OE doesn't build its own version anymore Aug 12 14:28:01 03  07org.openembedded.dev * r30bf862a42 10OE.dev/packages/u-boot/u-boot_git.bb: u-boot git: bump SRCREV for beagleboard for i2c fixes Aug 12 14:28:04 03  07org.openembedded.dev * ra212160e77 10OE.dev/packages/numptyphysics/numptyphysics_svn.bb: numptyphysics: Add -lX11 to LDOPTS to fix build Aug 12 14:28:07 03  07org.openembedded.dev * r91bea741c1 10OE.dev/packages/ffmpeg/ffmpeg_git.bb: ffmpeg git: bump SRCREV Aug 12 14:28:10 03  07org.openembedded.dev * r273cb190ca 10OE.dev/packages/binutils/binutils-avr32.inc: binutils: fix avr32 prefixing issue Aug 12 14:28:23 03  07master * r30bf862a42 10OE.dev/packages/u-boot/u-boot_git.bb: u-boot git: bump SRCREV for beagleboard for i2c fixes Aug 12 14:28:26 03  07master * r9e4b744670 10OE.dev/packages/oprofile/ (5 files in 2 dirs): Aug 12 14:28:29 oprofile 0.9.3: Fixed armv7a patch to actually distribute armv7 event files. Aug 12 14:28:31 oprofile 0.9.4: New recipe. Aug 12 14:28:33 03  07master * r40552fbe64 10OE.dev/: Aug 12 14:28:35 merge of '453683170be608ea537c8e9133e9808544d8ac3c' Aug 12 14:28:37 and 'dd27045343c0a4288e8141f6cd9ce76afdd583b2' Aug 12 14:28:39 03  07org.openembedded.dev * rd3b1dec3a4 10OE.dev/conf/distro/include/sane-srcrevs.inc: sane-srcrevs: Updated linux-openmoko rev to match XorA's change in OM repo Aug 12 14:28:42 03  07org.openembedded.dev * r8395f3c0fe 10OE.dev/packages/numptyphysics/ (3 files in 2 dirs): numptyphysics: add patch to build it with gcc 4.3 Aug 12 14:28:53 03  07org.openembedded.dev * rd0cc2fd24c 10OE.dev/: Aug 12 14:28:55 merge of '05dd48c35e7fc8bd8ec725a975682953f4e2eb8b' Aug 12 14:28:57 and 'e5d51ef06db323f64e81b4277041a3eb479eda1b' Aug 12 14:28:59 03  07org.openembedded.dev * r9e4b744670 10OE.dev/packages/oprofile/ (5 files in 2 dirs): Aug 12 14:29:01 does this mean that the server is back online? Aug 12 14:29:02 oprofile 0.9.3: Fixed armv7a patch to actually distribute armv7 event files. Aug 12 14:29:06 oprofile 0.9.4: New recipe. Aug 12 14:29:08 (123 lines omitted) Aug 12 14:29:10 03  07master * rf5eeb4b683 10OE.dev/packages/freesmartphone/illume-theme-freesmartphone_git.bb: illume-theme-freesmartphone: Build theme from source. Aug 12 14:29:13 03  07master * r07267c2b9f 10OE.dev/packages/dsplink/ (codec-engine_2.10.bb dsplink.inc ti-paths.inc): dsplink, codec-engine: put common paths in a .inc file so more recipes can share them Aug 12 14:29:22 03  07master * rc73a2ec2c5 10OE.dev/packages/resolvconf/resolvconf_1.41.bb: Aug 12 14:29:22 I guess wolfson synced then :-) Aug 12 14:29:24 resolvconf: add 1.41 Aug 12 14:29:26 * openresolv (http://roy.marples.name/node/343) doesn't need bash, which is more in line what OE expects Aug 12 14:29:28 03  07master * r53f0cf59e8 10OE.dev/conf/machine/om-gta01.conf: om-gtao1: use += for image types Aug 12 14:29:37 03  07master * r82f66e0432 10OE.dev/packages/python/ (3 files in 2 dirs): python-edje cvs add debian patch fixing unicode conversion Aug 12 14:29:40 03  07master * r327318d840 10OE.dev/packages/mythtv/gmyth_0.7.1.bb: gmyth: add mythtv dependency Aug 12 14:29:43 03  07master * r0d9cbce19b 10OE.dev/packages/linux/ (2 files in 2 dirs): linux-openmoko.inc: sync with openmoko so package upgrade doesn't &*$(*@$(@ erase my kernel partition again Aug 12 14:29:49 huhu Aug 12 14:29:50 03  07master * r18d7c73dd0 10OE.dev/packages/bluez/ (bluez-utils/hcid.conf bluez-utils_3.33.bb): bluez-utils: restore device naming behaviour that got lost after removing bluez-utils-3.x from FILESPATH Aug 12 14:29:53 03  07master * r2c57de15fa 10OE.dev/packages/u-boot/u-boot_git.bb: u-boot: fix bump SRCREV for neuros-osd2 to fix HDCP handshake Aug 12 14:29:56 03  07master * re2eb27e900 10OE.dev/packages/ffmpeg/ffmpeg_git.bb: ffmpeg git: switch from 'arm-neon' to 'arm' branch for arm Aug 12 14:30:01 03  07master * ra480a1ffbc 10OE.dev/packages/gstreamer/gst-ffmpeg_0.10.4.bb: gst-ffmpeg: bump PR, ffmpeg changed SOVERSION Aug 12 14:30:04 03  07master * rec3bc08b42 10OE.dev/packages/ffmpeg/ffmpeg_git.bb: ffmpeg git: add note on how to keep software working Aug 12 14:30:07 03  07master * rd114752989 10OE.dev/packages/netbase/ (3 files in 3 dirs): netbase: remove resolv.conf hack from om-gta0* that keeps overwriting working resolv.conf files with broken ones Aug 12 14:30:12 03  07master * re709fa19b3 10OE.dev/conf/machine/adzs-bf548-ezlite.conf: adzs-bf548-ezlite: adjust serial console setting Aug 12 14:30:15 03  07master * r8900e965d2 10OE.dev/conf/distro/angstrom-2008.1.conf: angstrom 2008: fix typo Aug 12 14:30:18 03  07master * rb3945301ba 10OE.dev/classes/distutils.bbclass: distutils.bbclass: allow specifying arguments to the build command Aug 12 14:30:23 03  07master * rcebbc63d04 10OE.dev/packages/totem/ (totem/.mtn2git_empty totem/gst-detect.diff totem_2.23.4.bb): totem: add 2.23.4 Aug 12 14:30:26 03  07master * rb104477b6c 10OE.dev/conf/distro/include/preferred-om-2008-versions.inc: preferred-om-2008-versions.inc: update curl and pycurl Aug 12 14:30:29 03  07org.openembedded.dev * rf5eeb4b683 10OE.dev/packages/freesmartphone/illume-theme-freesmartphone_git.bb: illume-theme-freesmartphone: Build theme from source. Aug 12 14:30:36 03  07master * red991ece35 10OE.dev/packages/mythtv/gmyth_0.7.1.bb: gmyth: add 0.7.1 Aug 12 14:30:39 03  07master * rf2749d3abe 10OE.dev/packages/totem/totem-pl-parser_2.23.3.bb: totem-pl-parser: add 2.23.3 Aug 12 14:30:42 (55 lines omitted) Aug 12 14:39:45 good news we have udev_124 to test now..thx koen Aug 12 14:39:55 but looking at the init...export TZ=/etc/localtime Aug 12 14:40:13 well...angstrom does not provide /etc/localtime Aug 12 15:18:58 bbl Aug 12 15:22:58 http://venturebeat.com/2008/08/08/defcon-excuse-me-while-i-turn-off-your-pacemaker/ Aug 12 15:23:27 koen can kill a lot of people in the street with a portable version of his beagleboard gnu radio :D Aug 12 15:23:40 fear. Aug 12 15:26:09 you have to love the irony http://www.newsweek.com/id/151355 Aug 12 15:57:13 I've got a question regarding FILESPATH: I see where its configured in conf/bitbake.conf to be ${P}:${PN}:files yet when I print it where its used in bitbake (ie bitbakes local file fetcher) I see that there's another set of subdirs that are applied to each of FILESPATH - namely ${TASK},local,${MACHINE},${DISTRO},${ARCH}, build-${OS}, pn-${PN} - I'm not understanding where this is configured Aug 12 16:12:20 nvr mind... I see that its base.bbclass:base_set_filespath where it does this based on OVERRIDES - basically FILESPATH becomes what you set FILESPATH to combined with OVERRIDES Aug 12 16:28:22 morning all Aug 12 16:28:30 yo thesing Aug 12 16:33:50 thesing: hey there:) Aug 12 16:48:56 a+ Aug 12 16:59:03 I've got a do_rootfs where opkg is trying to prompt for an action b/c of a config file thats both created by a script and provided by a package - Clearly this is an error in my recipe but I wonder why opkg shouldn't be failing vs asking for a user-prompt (the prompt goes to log.do_rootfs and the process appears to be hung at first glance) Aug 12 17:01:30 shouldn't rootfs_ipk.bbclass execute opkg to fail instead of prompt? Aug 12 17:02:16 does opkg have an option for that? Aug 12 17:02:20 if so, go ahead and patch the .bbclass to use it Aug 12 17:03:19 not sure... I'm pretty new to opkg Aug 12 17:53:28 mickeyl: could you quickly coordinate the python changes for bsddb with Marcel_M (maybe via query) - for him the _bsddb.so module still does not build Aug 12 17:54:53 no queries please, too many open windows Aug 12 17:55:01 i'll join 1973 in a minute Aug 12 17:55:47 mickeyl: or this way ;) Aug 12 17:56:09 mickeyl: what client do you have that a query creates a new window?? :D Aug 12 17:56:20 well, window, tab, whatever Aug 12 17:56:27 "an entity that requires attention" Aug 12 17:56:34 ^^ Aug 12 17:59:00 hi Aug 12 17:59:55 mickeyl is a windows man, we got pictures of it :-) Aug 12 18:01:04 no longer Aug 12 18:01:06 honestly Aug 12 18:01:12 got rid of windows on all but one machine Aug 12 18:01:26 mickeyl: although my phone is a windows phone :-D Aug 12 18:01:31 heh Aug 12 18:16:31 Tartarus, you use OE via an 'overlay' where you have a local tree of your additions/changes and an untouched oe tree right? Aug 12 18:18:07 Yeap Aug 12 18:18:19 Works like a charm :) Aug 12 18:21:28 Tartarus, how do you deal with overriding local files that are used by packages such as netbase/base-files? Aug 12 18:22:50 tharvey, use the same path/name in the overaly it have the overlay at a higher prioity Aug 12 18:26:45 Crofton, doh... I could swear I tried that yesterday and could not get it to work - thats the intuitive way to do it. I bet I didn't clean something before a rebuild Aug 12 18:26:48 thanks Aug 12 18:27:16 yeah Aug 12 18:27:22 I am pretty sure it does work Aug 12 18:29:15 ya, I just tested it again and it works Aug 12 20:11:05 * * OE Bug 4494 has been created by hans(AT)eds.org Aug 12 20:11:08 * * lzma-native_4.17.bb fails to fetch source tarball Aug 12 20:11:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4494 Aug 12 20:15:06 oe will not do dependency tracking on local files correct? (ie if a local file in SRC_URI changes oe doesn't know to rebuild the package) Aug 12 20:16:00 no, afaik Aug 12 20:17:55 Jay7: what's your opinion about /home if one uses to kexec different distros from sd? Have it on the same partition as / ? Aug 12 20:18:36 ant_: let such distro's fstab decide? Aug 12 20:18:55 or this...remember when you test Aug 12 20:19:16 (I have bigger problems...wrong mtdparts...) Aug 12 20:19:24 almost solved...;-) Aug 12 20:19:36 1st is bad, 2nd is good :) Aug 12 20:19:46 but the, what's the sense of /home on mtd3? Aug 12 20:20:29 I think I'll monunt this as encrypted 'private' partition Aug 12 20:20:33 application's settings stored in home Aug 12 20:20:43 hmm...it depends... Aug 12 20:21:03 if you boots let say angstrom and debian...dunoo Aug 12 20:21:25 when we have single /home for different distros it can be a problem Aug 12 20:21:25 urgh..this keyboard needs real cleaning... Aug 12 20:21:33 exactly Aug 12 20:22:07 I think that's why sash reverted back to one single jffs2 partition Aug 12 20:23:18 may be.. but I want to have angstrom on NAND and angstrom's home on mtd3 Aug 12 20:23:31 when I decide to start debian from sd Aug 12 20:23:51 I'll using home on sd :) Aug 12 20:23:55 I see Aug 12 20:24:33 BTW I'm not sure the kexecboot can yet boot a jffs2 image... Aug 12 20:24:46 I tested on a real ext2 fs Aug 12 20:24:56 on sd (mmcblkp1) Aug 12 20:25:03 I don't want to mix / and /home in single partition.. it kind of paranoia :) Aug 12 20:25:28 * my paranoia :) Aug 12 20:25:39 Jay7: I'm restoring the u-boot partitions... Aug 12 20:25:44 :-) Aug 12 20:25:51 always had /boot ro Aug 12 20:26:03 ant: fwiw, I have been putting my home on a seperate partition on the SD card Aug 12 20:26:21 hacking fstab? Aug 12 20:26:26 ant_: yup Aug 12 20:26:29 ok Aug 12 20:26:39 thesing needs to know... Aug 12 20:27:25 hvontres|work: can you safely remove sd when still working with device? Aug 12 20:27:51 Jay7: no, but I also have my rootfs on SD Aug 12 20:27:56 ah Aug 12 20:28:15 then it's good idea :) Aug 12 20:28:35 * ant_ was seriously hit by a user uploading a full photo-book...quota implemented now... Aug 12 20:28:40 Jay7: makes testing new images much easier, plus with SD cards < $7 now, I'd ratehr wear that out than the internal flash Aug 12 20:28:49 but using sd as home when booted from nand and needed to eject sd.. is not so good for me Aug 12 20:30:09 Jay7: you could create let say 3 partitions: /home-angstro, /home-debian , /home-android Aug 12 20:30:54 but then heavilyuse symlinks Aug 12 20:32:02 btw, is mnt repo on oe.net working now? Aug 12 20:32:07 yup Aug 12 20:32:18 is it uptodate? Aug 12 20:32:28 udev_124 is there...compiling... Aug 12 20:32:39 (btw _126 is out !) Aug 12 20:32:42 ok, going to sync :) Aug 12 20:32:59 I'm tempted to tray with straight _126 Aug 12 20:33:46 btw2, how to rebuild properly after updating? Aug 12 20:34:17 e.g. I have console-image built Aug 12 20:34:21 well, I change angstrom-2008-preferred-versions.inc Aug 12 20:34:33 directly Aug 12 20:34:37 how to rebuild all updated packages? Aug 12 20:34:50 (and console-image too) Aug 12 20:34:53 bitbake -c clean udev_118 Aug 12 20:35:01 bitbake udev_124 Aug 12 20:35:19 ant_: It works for single package Aug 12 20:35:20 (look at its default preference, though) Aug 12 20:35:40 but e.g. after I've synced from mtn repo Aug 12 20:35:46 Jay7: if PR got bumped it is automatic Aug 12 20:36:04 will bitbake console-image rebuild all updated packages? Aug 12 20:36:13 ah, ok Aug 12 20:36:27 otherwhise (similar to Gentoo) you could try #BB_STAMP_POLICY = "whitelist" Aug 12 20:36:49 but IIRC it has issues with packaged-staging. have to check Aug 12 20:37:32 I've forget about oestats Aug 12 20:37:32 ## rebuild always on stamp changes "whitelist" (with pstaging) or "full" Aug 12 20:38:28 Jay7: here you'll have to ask RP Aug 12 20:38:38 in my kernel recipe which uses an initramfs, how do I specify that when the initramfs image is rebuilt the kernel needs to be rebuilt? I thought a DEPENDS += "initramfs-image" would do the trick but it doesn't Aug 12 20:39:18 tharvey: look at kexecboot Aug 12 20:39:58 ant_: did you test new u-boot? I'm asking again ;) Aug 12 20:40:12 I did not receive any file... Aug 12 20:40:24 ant_, yes, I'm looking at that for an example, but I'm not seeing it - I also have a do_compile[depends] += "initramfs-image:do_rootfs" in my kernel but that causes the initramfs-image to be rebuilt when the kernel is recompiled but not visa-versa Aug 12 20:41:06 it's a pity Aug 12 20:41:20 ok, waiting for new u-boot then :) Aug 12 20:41:28 Jay7: I could chat whith InSearchOf on #pdaxrom Aug 12 20:41:52 he confirms he's working on u-boot_1.33 Aug 12 20:42:06 but the patches are not yet public Aug 12 20:42:47 shouldn't the 'do_compile[depends] += "initramfs-image:do_rootfs"' in a kernel recipe cause the kernel recipe to do_compile if initramfs-image:do_rootfs is called? Aug 12 20:42:59 I asked him whether some tosa patch could fly around (even for the old u-boot_1.1.4) but he answered he restarted with u-boot_1.3.3 from scratch Aug 12 20:44:16 well, I hope we will have current u-boot on zaurus'es :) Aug 12 20:44:41 working atags too Aug 12 20:46:07 tharvey: linux-kexecboot.inc sets INITRAMFS_IMAGE = "initramfs-kexecboot-image", which is evaluated in kernel.bbclass Aug 12 20:47:47 Jay7: for sure one could fix that broken atags of u-boot_1.1.4 (or at least confirm these are b0rked) Aug 12 20:48:00 one = one u-boot hacker Aug 12 20:48:37 I really hope that new u-boot have working atags already :) Aug 12 20:48:59 perhaps the issue is in machine.conf...and not in u-boot Aug 12 20:49:08 that's my guess Aug 12 20:49:17 hm.. Aug 12 20:50:05 machine conf = u-boot akita.h corgi.h Aug 12 20:50:21 btw, did you search why u-boot image will be created with XIP enabled? Aug 12 20:50:47 not yet..I confess I'm hardcoding the mtdparts in /boot/cmdline Aug 12 20:51:14 just to see x11-image booting properly from SD Aug 12 20:51:16 and next, do someone tries to compile some zaurus kernel with XIP enabled? Aug 12 20:51:38 does it make sense? doesn't kexec free all the ram? Aug 12 20:51:55 ant_, right - but I specify INITRAMFS_IMAGE in a machine conf file instead of the kernel - it should still setup the dependency correctly so that if I 'bitbake -c clean initramfs-image; bitbake virtual/kernel' it should rebuild the kernel correct? (it doesn't rebuild the kernel) Aug 12 20:52:11 hm.. yes, for kexec it's senseless.. Aug 12 20:52:30 tharvey: bitbake -c clean before Aug 12 20:53:14 ant_, clean the kernel? I'm trying to avoid a full rebuild of the kernel - all that needs to be done is a relink if the initramfs changes Aug 12 20:53:35 I have tomorrow free day more or less Aug 12 20:53:38 tharvey: I'm sorry but Aug 12 20:53:38 kernel_do_compile() { Aug 12 20:53:38 if [ ! -z "${INITRAMFS_IMAGE}" ]; then Aug 12 20:54:01 you need to redo do_compile Aug 12 20:54:14 so I'll try to search why mkimage creates XIP-enabled images Aug 12 20:54:22 Jay7: good Aug 12 20:54:40 Jay7: it is probably some openmoko contamination...he he Aug 12 20:55:11 Jay7: I'll test with u-boot environment Aug 12 20:55:13 ant_, ya... ok Aug 12 20:56:20 seems like I needed other akita :)) Aug 12 20:56:29 one for work and one for testing Aug 12 20:56:42 Jay7: I have two SL-C860... Aug 12 20:56:55 you know why... Aug 12 20:57:01 yep :) Aug 12 20:57:21 my wife don't want akita :) Aug 12 20:57:32 it's main problem :)) Aug 12 20:57:37 try with flowers :-] Aug 12 20:58:15 hm.. colored in red and with swarowsky :) Aug 12 20:58:25 it's good idea :) Aug 12 21:02:26 hi thesing Aug 12 21:02:36 hi everybody Aug 12 21:03:36 ant_: do you now tries to compile image with new udev? Aug 12 21:04:05 yes, but I'm studying how to patch that kexec-kernel cmdline Aug 12 21:04:40 I just want to start new build and go to bed :) Aug 12 21:05:14 wait 5 mins and I'll tell you about udev_124 Aug 12 21:05:18 ok Aug 12 21:10:34 Jay7: image completed...flashing... Aug 12 21:11:17 * Jay7 sits with crossed fingers :) Aug 12 21:18:46 Jay7: hm... Aug 12 21:21:16 ant_: ? Aug 12 21:21:31 broken jffs Aug 12 21:21:39 (rad error?) Aug 12 21:22:27 at what stage it brokes? Aug 12 21:23:13 it doesnt'find anymore the partition on the sd card... Aug 12 21:23:21 old card btw... Aug 12 21:23:34 No bootable device found... Aug 12 21:24:59 ant_: it seem that you got the initramfs running? which partition is your image in? Aug 12 21:26:40 flashing to mtd2 now Aug 12 21:27:02 on sd I have console-image Aug 12 21:27:08 on flash x11-image Aug 12 21:46:16 I've switched to udev_124 Aug 12 21:46:25 now starting build Aug 12 21:46:30 rebooting...for the n time... Aug 12 21:47:26 Jay7: no, not yet fixed Aug 12 21:48:18 ant_: ok, then I'm just checking 'buildability' :) Aug 12 21:48:46 btw, zaurus-2.6.inc should be commited :) Aug 12 21:48:55 hm..I discovered my error: there is a dead symlink in /boot/zImage.bin Aug 12 21:48:59 (kinda reminder) Aug 12 21:49:06 in the deployed images Aug 12 21:49:20 Jay7: he he..I'll repost the big patch Aug 12 21:51:02 Jay7: hardcoding the mtdparts in /boot/cmdline-2.6.26 is a workaround for custom rootfs sizes... Aug 12 21:51:40 it would be nice if dinamically read from u-boot env or /proc/cmdline or atags Aug 12 21:52:06 * * OE Bug 4399 has been RESOLVED (FIXED) by thommycheck(AT)gmx.de Aug 12 21:52:08 * * kernel.bbclass does not run sizecheck on kernel+initramfs Aug 12 21:52:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4399 Aug 12 21:52:25 hehe.. Aug 12 21:52:30 sizecheck :) Aug 12 21:52:47 he he commented out Aug 12 21:53:15 ant_: when tomorrow all will be ok then I back up my current akita nand and try to deal with kexecboot :) Aug 12 21:54:55 thesing: BTW I could not skip the sizechek if [ ! -z "${KERNEL_IMAGE_MAXSIZE}" -a -z "${DONT_CHECK_KERNELSIZE}" ]; Aug 12 21:55:29 setting DONT_CHECK_KERNELSIZE="1" one line before Aug 12 21:56:03 now I'm doing ####addtask sizecheck before do_install after do_compile Aug 12 21:56:11 Jay7: ^^ Aug 12 21:56:16 REM Aug 12 21:56:21 ant_: there was a typo there. So it never checked the size, I was told Aug 12 21:56:27 ant_: in what file? Aug 12 21:56:43 thesing: after correcting the typo... Aug 12 21:56:50 Jay7: kernel.bbclass Aug 12 21:56:57 ant_: ok Aug 12 21:56:58 hm. will check. Aug 12 21:58:05 ant_: where did you add DONT_CHECK_KERNELSIZE="1"? Aug 12 21:58:26 in local.conf (no go) then after do_deploy() Aug 12 21:58:33 one line after Aug 12 21:58:48 wtf..wrong window Aug 12 21:58:59 do_sizecheck() { Aug 12 21:59:00 if [ ! -z "${KERNEL_IMAGE_MAXSIZE}" -a -z "${DONT_CHECK_KERNELSIZE}" ]; Aug 12 21:59:00 Aug 12 21:59:12 betwen do_sizecheck and if Aug 12 22:00:01 and it never worked? Aug 12 22:00:23 it was complaining about the size Aug 12 22:00:32 it was working Aug 12 22:00:41 but no way to opt-out Aug 12 22:01:33 well, now I'm going to bed :) Aug 12 22:01:41 see you tomorrow :) Aug 12 22:01:43 bye Aug 12 22:01:46 so the DONT_... option didn't work Aug 12 22:02:23 yes Aug 12 22:03:12 thesing: I can do it now, oestats is back Aug 12 22:04:48 http://tinderbox.openembedded.net/builds/20092/ Aug 12 22:09:29 ant_: this is intentional. Aug 12 22:10:10 linux-kexecboot will be flashed in the kernel partition. So the size matters. Aug 12 22:10:35 if you wan't to override this you need to edit linux-kexecboot.inc Aug 12 22:14:22 thesing: np, I comment it in kernel.bbclass Aug 12 22:14:38 now, about FBANGLE = "270" Aug 12 22:15:07 this is the Service-Menu - orientation, isn't? Aug 12 22:15:31 is menu rotated for you? Aug 12 22:15:42 on clamshell it appears rotated 90 clockwise Aug 12 22:16:15 ok if you changed to palmtop-position Aug 12 22:17:10 well experiment with different values and choose one. Landscape is probably best. Aug 12 22:17:42 and the cursors? Aug 12 22:18:07 are they rotated too? Aug 12 22:18:27 let me retry..I think I use left to scroll Aug 12 22:20:07 thesing: 2 boot from sd = no keyboard, 3rd boot from sd...stalling Aug 12 22:20:46 ant_: if the boot from sd starts at all the job of kexecboot is done. Aug 12 22:20:57 is it better 'rootwait' or 'rootdelay=' Aug 12 22:21:01 ? Aug 12 22:21:25 no idea what the difference is. I use rootdelay Aug 12 22:23:19 ant_: if the keys are wrong the kernel keymap is wrong. (I use KEY_UP and KEY_DOWN defines) The only problem could be the selection key Aug 12 22:23:55 I use 'ok' Aug 12 22:24:30 ~seen hrw Aug 12 22:24:32 hrw was last seen on IRC in channel #oe, 4d 9h 39m 36s ago, saying: 'Crofton|work: freerunner...'. Aug 12 22:24:57 ~seen hrw|gone Aug 12 22:24:58 hrw|gone was last seen on IRC in channel #oe, 6d 53m 38s ago, saying: '2.6.26 is stable on c7x0'. Aug 12 22:25:19 thesing, ant_ : IIRC, rootwait was added in 2.6.23 and waits for the rootfs instead of rootdelay= where you need to guess how fast it will be there Aug 12 22:27:51 hvontres|work: good to know. Thanks. Aug 12 22:28:50 hvontres|work: thx Aug 12 22:29:21 thesing: I forgot what to change in order to have the zImage in /boot. Was a custom do_deploy? Aug 12 22:29:34 (doing it by hand now) Aug 12 22:32:12 no. That is RDEPENDS_kernel-base_collie = "kernel-image" in zaurus-2.6.inc Aug 12 22:34:37 I'm seeing, thx Aug 12 22:38:39 thesing: 3rd question...sorry :-)...do you think there is actually a way to dinamically pass the mtdparts to the kexeced kernel? Aug 12 22:39:24 if there is a way for userspace prog to get it? Aug 12 22:39:38 it's in the output of /proc/cmdline Aug 12 22:39:53 in the atag_cmdline tag Aug 12 22:40:09 and in u-boot.env ;-) Aug 12 22:40:33 then its doable. kexecboot would only need to read /proc/cmdline and extract the info from there. Aug 12 22:40:46 yes, no regexp Aug 12 22:41:21 it doesn't matter since kexecboot is a C program. You can parse the cmdline in any way you want. Aug 12 22:41:49 I was looking at linux-hh bootmenu and they dinamically look for 'root=' and so on, using 'expr'... Aug 12 22:41:58 you know... Aug 12 22:42:45 but the u-boot variable mtdparts would not need coversion Aug 12 22:43:42 how does a userspace prog access the uboot vars? Aug 12 22:43:45 mtdparts=sharpsl-nand-0:256k(U-Boot)ro,128k(U-BootEnv),5376k(Emergency),1408k(Kernel),-(JFFS2) Aug 12 22:44:06 thesing: u-boot-utils Aug 12 22:44:35 but a read to the atag list should work too Aug 12 22:44:46 too complicated then. parsind /proc/cmdline would be easier. Aug 12 22:45:13 well, the parameter is called 'atag_cmdline' Aug 12 22:45:23 so, back to /proc :-)) Aug 12 22:48:18 you know where the git tree for kexecboot is, don't you? Aug 12 22:48:30 yep, ltg Aug 12 22:50:36 I will gladly take patches for the /proc/cmdline stuff. Aug 12 22:50:43 (or any other patches) Aug 12 22:51:28 ok Aug 12 22:58:48 night Aug 12 23:15:22 Crofton, arg... a clean build and files from local do 'not' take precedence over files from oe dir - any ideas what I may be missing? Its intuitive that BBPATH would be used for local files but I can't find any code which makes that happen in bitbake Aug 12 23:16:47 ask tomorrow, I'm going to have to try something similar Aug 12 23:16:54 also look at gumstix.net Aug 12 23:17:07 I think sakoman has it working there Aug 12 23:17:19 will do, thx Aug 12 23:23:24 tharvey, it seems to work for me Aug 12 23:24:35 I have a question about sysvinit: if I have 5 as default runlevel, will init run all scripts from rcS then from rc1.d, ... and finally those from rc5.d or only those in rc5.d ? Aug 12 23:28:59 Crofton|work, strange... it seems inconsistent to me. I see gumstix bsp doing the same thing I am - package/base-passwd for example Aug 12 23:29:07 yeah Aug 12 23:29:20 that is why I think it works, and it is working fo rme Aug 12 23:29:39 actually... there is a difference, they are using different dirs... where I'm using hte exact same dir: local/netbase/netbase/interfaces vs oe/netbase/netbase/interfaces Aug 12 23:30:53 I am copying stuff out of .dev and modifying depends Aug 12 23:31:01 exact copies are working Aug 12 23:31:14 need to change BBPATH and things in local.conf Aug 12 23:31:37 clean build running now Aug 12 23:31:47 just make sure your local appears before .dev in BBPATH - but what else would you need in local.conf? Aug 12 23:32:46 http://rafb.net/p/rtURpG73.html Aug 12 23:36:56 Crofton|work, hmmm maybe I have something wrong with my priorities - in your example ossie_collection is your local that takes precedence over upstream right? Aug 12 23:39:12 right Aug 12 23:39:37 http://pastebin.ca/1168868 Aug 12 23:40:15 and BBPATH in the env is OK? Aug 12 23:40:20 if I understand collections correctly the order they exist in BBFILES and BBFILE_COLLECTIONS is irrelevant and precedence is based off of BBFILE_PRIORITY* Aug 12 23:40:31 this is my understanding Aug 12 23:40:45 ya - /home/tharvey/oe/commonOS/build:/home/tharvey/oe/commonOS/local:/home/tharvey/oe/commonOS/oe Aug 12 23:41:55 perhaps I'm just doing something really stupid with regards to my build: I'm doing a bitbake -c clean netbase; bitbake netbase then looking at $WORDIR/.../netbase*/interfaces - its the one from oe/packages/netbase and not from local/packages/netbase Aug 12 23:43:17 tharvey, I just don't know :( Aug 12 23:45:38 Crofton|work, well thx for verifying I'm doing the right thing... its got to be something really silly (granted I've wasted almost an entire day on it :( ) Aug 12 23:46:28 I am pretty sure you are on the right track Aug 12 23:46:42 once you get it resolved, you will get the day back many times over Aug 12 23:48:39 Crofton|work, interesting - both you and I specify 'packages' subdir in our BBFILES and collections - yet the bitbake manual does not - http://bitbake.berlios.de/manual/ch04s02.html#id870544 - completely different issue but I would think that keeps our collections from allowing anything other the packages to be overridden? Aug 12 23:49:21 not sure Aug 12 23:49:33 I wonder if I've got some strange cache issue Aug 12 23:49:41 we are also overrinding stuf in conf at times Aug 12 23:49:43 tharvey: maybe its the OEDIR var. someone had problems with using more than on var or so. Aug 12 23:49:51 nuking tmp should reolve cahce Aug 12 23:50:54 ya, I'll do a clean build again - takes over an hour and its about time to head home anyway Aug 12 23:51:20 I am doing the same Aug 12 23:51:22 thesing, ok - I'll try that - I think its more readable w/o the extra var anyway Aug 12 23:51:31 but I have some extra stuff going on Aug 12 23:51:44 we should be able to solve this tomorrow Aug 12 23:52:26 hope so... appreciate the help Aug 12 23:52:30 * Crofton|work is losing track Aug 12 23:52:39 I am talking OE in two channels Aug 12 23:52:40 gn Aug 12 23:53:13 removing extra OEDIR didn't make a diff Aug 12 23:53:43 night Aug 12 23:54:00 thats good. So our parser isn't broken ;) Aug 12 23:55:21 can I blow away cache dir to invalidate cache? Aug 13 00:06:25 morning Aug 13 00:42:53 night all **** ENDING LOGGING AT Wed Aug 13 02:59:57 2008