**** BEGIN LOGGING AT Thu Oct 09 02:59:57 2008 Oct 09 04:54:53 do we have a bitkeeper export which contains the full history? Oct 09 04:58:15 hrw sent me one, but it was not complete :/ Oct 09 05:20:13 RP: ? Oct 09 06:17:25 how does the git fetcher work? As far as I see it gets a git repository as tar.gz over http... Oct 09 06:26:16 anyone else getting a dependency loop w/ HEAD? Oct 09 06:30:51 * rwhitby just had a beagle board handed to him for custodianship ... Oct 09 06:31:48 mickey|tw: Yes, I just checked the nslu2-linux autobuilders. Oct 09 06:33:46 rwhitby: congrats :) Oct 09 06:33:54 mwester: bummer, what triggered that? Oct 09 06:34:39 mickey|tw: I also got that second Treo650 today :-) Oct 09 06:35:09 rwhitby: whew, equipment day! Oct 09 06:36:00 the funny thing is that zedstar's old Treo650 is in much better condition than my workhorse, so I'll probably swap the cases as the first job :-) Oct 09 06:36:24 mickey|tw: Dunno, the log doesn't say -- or it might be in there, but it'll take a little digging to sort. Oct 09 06:36:54 * mwester just put a new case on one today. Easier than he thought... Oct 09 06:36:58 rwhitby: think you got time to add FSO support for it? Oct 09 06:37:26 mickey|tw: we'll see Oct 09 06:37:51 btw., when i'm back in .de I'm going to carry out some changes in fso image recipes Oct 09 06:37:55 just to inform you beforehand Oct 09 06:38:02 i will refactor tasks out Oct 09 06:38:10 and will add three additional image recipes Oct 09 06:38:15 or tweak, that is Oct 09 06:38:28 end result should be fso-console-image, fso-illume-image, fso-zhone-image. Oct 09 06:38:47 so Om can adopt two images while ignoring the third one Oct 09 06:38:55 :) fso-console-image might be one that could support the framework but no X Oct 09 06:38:56 ? Oct 09 06:38:59 yep Oct 09 06:39:09 mwester: actually, i thought about something today Oct 09 06:39:10 :) I'm hacking about creating that one... Oct 09 06:39:24 i'll add some module to co-work w/ X Oct 09 06:39:30 someone could add a module to co-work w/ QPE Oct 09 06:39:38 for some functions where it makes sense :) Oct 09 06:39:46 :) That's exactly what I'm thinking. Oct 09 06:39:53 *nod* good thinking Oct 09 06:40:18 the relationship between FSO and Om will also be a bit redefined soon Oct 09 06:40:21 all good though Oct 09 06:40:31 (imo) Oct 09 06:43:07 cool Oct 09 07:43:20 morning Oct 09 07:48:40 how does the git fetcher work? As far as I see it gets a git repository as tar.gz over http... Oct 09 07:48:57 03  07org.openembedded.dev * ra048be2b3c 10org.openembedded.dev.git/packages/linux/ (linux-omap/fix-irq33.diff linux-omap_git.bb): linux omap git: add patch to fix irq -33 problem Oct 09 07:51:25 hi Oct 09 07:54:38 When I try to build an x11-image, I have an error building Gtk+-2.14.2 Oct 09 07:54:54 /home/drakou/OE/tmp/cross/armv5te/lib/gcc/arm-angstrom-linux-gnueabi/4.2.4/../../../../arm-angstrom-linux-gnueabi/bin/ld:.libs/libgdk_pixbuf-2.0.ver:2: syntax error in VERSION script Oct 09 07:54:54 | collect2: ld returned 1 exit status Oct 09 07:55:30 Identifying dependency loops (this may take a short while)... Oct 09 07:55:48 03  07org.openembedded.dev * r871fd71c63 10org.openembedded.dev.git/: Oct 09 07:55:48 merge of '2e9358fe57665ae0d14365c94a62792b10e1b578' Oct 09 07:55:48 and 'b926346694764187095a5101a505ba5ad9128a78' Oct 09 07:55:58 03  07org.openembedded.dev * rdcd7683e84 10org.openembedded.dev.git/classes/gettext.bbclass: revert last patch as it leads to a build loop. woglinde, please try again. Oct 09 07:56:02 how long is a short while? (more than an hour?) Oct 09 07:56:12 you won't be able to build anything Oct 09 07:56:13 just update Oct 09 07:56:18 i have just fixed that Oct 09 07:57:32 morning Oct 09 07:57:47 morning :) Oct 09 07:58:43 I have a question about getty from tinylogin proposed by OE: is there a limitation about the serial port speed ? Because I cannot get it to work at 921600 bps (and the port works at this speed, I get all kernel messages fine). ? Oct 09 07:59:09 thanks, mickey|tw Oct 09 07:59:18 np Oct 09 07:59:23 sorry for the inconvenience Oct 09 07:59:28 will kick woglinde next time :D Oct 09 07:59:31 ~seen woglinde Oct 09 07:59:33 woglinde was last seen on IRC in channel #oe, 8h 8m 52s ago, saying: 'nite now really'. Oct 09 07:59:37 np Oct 09 08:00:07 hi Genesis ! Oct 09 08:00:50 Hey Oct 09 08:01:16 Does someone know what the PE variable means? Oct 09 08:02:35 bonjour Oct 09 08:04:52 Someone could take a while to help me ? I really would like to use OE instead of ltib ( we are currently using Ltib to build our systems ) . I can't build a simple x11-image, I get an error while building gtk+ Oct 09 08:13:57 03  07org.openembedded.dev * r363d86fd66 10org.openembedded.dev.git/: Oct 09 08:13:57 merge of '0cc46a3e4be22a50518f2ca80e16fc733b1d9ca0' Oct 09 08:13:57 and '300c8b716e9fdd8499a7197f71dbcea5255e3b5a' Oct 09 08:14:05 03  07org.openembedded.dev * radb64d8b3d 10org.openembedded.dev.git/classes/gettext.bbclass: gettext.bbclass: revert part of a cset that breaks OE with inifinite loops Oct 09 08:15:24 quickdev: Package Epoch Oct 09 08:15:35 quickdev: same as epoch in debian packages Oct 09 08:15:52 XorA, thanks. Oct 09 08:28:26 XorA, do you know about any full bitkeeper exports? Oct 09 08:29:33 shoragan: no, bitkeeper was like centuries ago :-) Oct 09 08:29:41 :) Oct 09 08:29:53 shoragan: Laibsch was the only guy other than RP to fiddle with them I think Oct 09 08:29:58 ah Oct 09 08:30:14 ~seen RP Oct 09 08:30:17 rp is currently on #maemo #handhelds #oe #openmoko, last said: 'Tartarus: It works in Poky, I've not tried OE for a while'. Oct 09 08:30:37 ~seen Laibsch Oct 09 08:30:38 laibsch was last seen on IRC in channel #debian, 16d 19h 41m 2s ago, saying: 'Try a different port'. Oct 09 08:31:04 everyone has dissapeared while I was on hols Oct 09 08:31:49 heh, yeah Oct 09 08:31:53 laibsch is probably gone for good Oct 09 08:31:58 but RP should still be around somewhere Oct 09 08:32:02 probably pretty busy Oct 09 08:32:10 RP said nothing since 2008-10-02 Oct 09 08:32:13 Intel aquiring OH shows their face Oct 09 08:32:17 mickey|tw: OH, stuff happened while I was away? Oct 09 08:32:18 ;) Oct 09 08:32:32 I guess RP is busy cursing RPM still Oct 09 08:32:44 heh Oct 09 08:35:44 I was cursing rpm/yum Oct 09 08:36:13 shoragan: bkbits.net offers snapshots of bktrees Oct 09 08:36:23 it does? Oct 09 08:36:50 at least it was Oct 09 08:37:06 i cant seem to find it Oct 09 08:40:38 neither do I Oct 09 08:42:00 i've requested a new export via http://kernel.bkbits.net/cgi-bin/cvsexport.cgi Oct 09 08:42:29 I thought we had exports galore Oct 09 08:43:45 XorA, i found only exports which didn't have the latest commits Oct 09 08:44:06 ok, it got an export Oct 09 08:47:36 does busybox also provide getty and why is tinylogin chosen "by default" ? Oct 09 08:48:00 steliosk: pong Oct 09 08:49:13 morning all Oct 09 08:49:28 XorA, mickey|tw: I'm still here ;-) Oct 09 08:49:36 Just very busy :( Oct 09 08:50:02 RP: we just assume M had attached itself and kept you under its control :-D Oct 09 08:50:23 XorA: M? Oct 09 08:50:31 RP: i can imagine. good to hear. Oct 09 08:50:36 (you're still there) Oct 09 08:50:37 RP: RPM :-D Oct 09 08:50:46 hahaha Oct 09 08:50:59 XorA: I'm not an rpm fan... Oct 09 08:51:18 RP: your bosses are though? Oct 09 08:51:49 XorA: Moblin is an rpm distribution! Oct 09 08:52:17 Anyhow, that isn't a problem Oct 09 08:56:56 is there a way to use the busybox getty instead ot the tinylogin one in an OE based distribution ? Oct 09 08:57:23 Longfield: not yet, feel free to produce patches Oct 09 08:57:37 Longfield: we werent to convinced about busybox running SUID Oct 09 08:57:52 I'd like to see that happening Oct 09 08:57:58 bb has fine grained control these days Oct 09 08:58:05 on my list of things to investigate Oct 09 08:58:08 but very down Oct 09 08:58:15 would appreciate if someone beats me to it Oct 09 08:58:30 that would be cool Oct 09 08:58:36 Its not like tinylogin is large though ;-) Oct 09 08:58:44 no, but it's not that Oct 09 08:58:47 tinylogin is broken Oct 09 08:58:53 has known flaws Oct 09 08:58:58 It has? Oct 09 08:58:59 and security issues Oct 09 08:59:00 yep Oct 09 08:59:10 :( Oct 09 08:59:19 that's why i was trying to push that some months ago Oct 09 08:59:31 but then koen came around shooting this idea down Oct 09 08:59:34 well, for me it does not work on my serial port at high speeds (921600) although the kernel is able to use the serial port at that speed Oct 09 08:59:39 and i moved on to other things Oct 09 08:59:56 tinylogin can't be fixed? Oct 09 09:00:00 and I absolutely need the port at that speed Oct 09 09:00:14 I must admit making busybox suid does feel a bit risky Oct 09 09:01:12 well the other option was making busybox-secure .bb and only enabling su login in that Oct 09 09:02:48 XorA: ya Oct 09 09:03:21 re Oct 09 09:03:23 I guess for OM makes no difference anyway :-D Oct 09 09:04:17 so what's the best ? tinylogin is quit old, fixing it would make sense ? Oct 09 09:05:13 on the otherside, busybox is still in development, so better chances that things get fixed (regarding my problem) Oct 09 09:05:17 Longfield: the best would be if you had time to do a quick investigation on both options Oct 09 09:05:27 Longfield: and do a little report for us, but I know thats asking a lot Oct 09 09:05:58 making the whole of busybox SUID feels wrong to me, but is there a way around this Oct 09 09:07:31 My package is build with the name "libframeworkd-phonegui0". Where is the 0 coming from? My recipe: http://rafb.net/p/uDICXg15.html Oct 09 09:08:04 quickdev: debian renaming Oct 09 09:09:45 where do I have to change things if I want to have OE build getty from busybox and remove tinylogin ... I guess changing the defconfig of busybox but then for removing tinylogin ? Oct 09 09:10:27 Longfield: I guess probably task-base to remove tinylogin Oct 09 09:10:34 Longfield: but I didnt look Oct 09 09:11:05 XorA, it is inherited then, right? Oct 09 09:11:24 quickdev: its default in angstrom Oct 09 09:11:45 XorA, how can I disable it, if it's default? Oct 09 09:12:05 quickdev: why do you want to? Oct 09 09:12:58 XorA, because almost all packages are built without any appended digits Oct 09 09:13:47 quickdev: its a library thing, same as in debian Oct 09 09:14:20 XorA, ah, then everything is fine, thanks :) Oct 09 09:17:21 XorA: busybox can be configured to have only selected applets suid Oct 09 09:17:29 hrw: then all is cool Oct 09 09:17:37 hrw: thanks for the info Oct 09 09:20:33 all is stored in /etc/busybox.conf iirc Oct 09 09:23:46 in task-boot, a variable DISTRO_LOGIN_MANAGER ?= "tinylogin" is defined, so this should be enough to change where my getty comes from ? Oct 09 09:47:56 mickey|tw: can python-manifest drop playing with PR? Oct 09 09:53:21 Can anyone help: how does the git fetcher work? As far as I see it gets a git repository as tar.gz over http... Oct 09 09:53:43 Or where is the code for it so I can try to find out on my own?? Oct 09 09:54:11 Kleckse, that's what tried first. If that fails the git repository is used. Oct 09 09:54:51 hmmm... I couldn't see anything about git access failing in the log.. Oct 09 09:55:25 hmm.. ok http is tried first.... Oct 09 09:56:04 why this... so someone has to create all the tar.gz and put them on the server... Any good reason for this? Oct 09 09:57:38 has this anything to do with DEPEND = "git-native" ???? Oct 09 09:58:03 mickey|tw: http://pastebin.ca/1223603 Oct 09 09:58:15 Kleckse: the reason for it is simple Oct 09 09:58:52 Kleckse: git/svn/cvs fetchers first check on SRC_TARBALL_STASH for archives and it this fail then do checkout/update via own methods Oct 09 09:58:59 03  07org.openembedded.dev * r6cd3e98474 10org.openembedded.dev.git/: Oct 09 09:58:59 merge of '4b614ea9dfb14d8f67ed00b35c59a2f23b49e2da' Oct 09 09:58:59 and '9a0857ebdef17f18989f0702e43595ce522a7a24' Oct 09 09:59:00 03  07org.openembedded.dev * r8d7b294296 10org.openembedded.dev.git/packages/linux/linux-rt_2.6.24.bb: linux-rt 2.6.24: Revert to 2.6.24.3-rt3 for mpc8315e-rdb. .7+ needs rebase. Oct 09 09:59:18 Kleckse: this gives us tarballs if upstream repo fails and make them less occupied by us Oct 09 10:00:38 ok.. then I am doing everything correctly :-) Oct 09 10:01:20 Can I have a look at the git fetcher code... I couldn't find it.. Is it part of OE or part of bitbake? Oct 09 10:03:25 Kleckse, I think it isin bitbake Oct 09 10:03:51 bitbake Oct 09 10:03:56 lib/bb/fetch/git.py Oct 09 10:04:52 ok.. thank you... I guess SRC_TARBALL_STASH is normaly set in distro configuration, right? Oct 09 10:05:11 yes Oct 09 10:09:12 It seems like we could benefit from a complete review of all the places we do not use busybox for something Oct 09 10:09:25 mount comes to mind also Oct 09 10:09:54 Crofton: and init Oct 09 10:11:38 need to figure out why we didn't use the busybox equivalent in the first place and if the reason is still valid Oct 09 10:11:54 * Crofton|work is not volunteering Oct 09 10:12:37 XorA, ping Oct 09 10:12:43 I quess I will have to look at the init stuff, as I am struggeling with sysvinit.... Oct 09 10:13:22 Crofton: gnop Oct 09 10:13:29 Crofton|work: gnop as well Oct 09 10:13:31 how was America? Oct 09 10:13:37 very american :-D Oct 09 10:13:45 hopefully in a good way Oct 09 10:13:51 I want an Ihop in Edinburgh Oct 09 10:14:00 heh Oct 09 10:14:01 open one Oct 09 10:14:04 Ponderosa sucks ass though Oct 09 10:14:08 yeah Oct 09 10:14:08 never going there again Oct 09 10:14:17 stupid inlaws saying it was good Oct 09 10:14:44 heh Oct 09 10:15:12 I think my mother likes it, but she just gets the salad bar Oct 09 10:15:44 Disney is creepy Oct 09 10:15:51 Universal Studios is very good though Oct 09 10:15:59 TNA Wrestling is amazing Oct 09 10:16:05 I've only been to Duisney Wolrd Oct 09 10:16:11 a very long time ago Oct 09 10:16:19 I wont ever go to disney again Oct 09 10:16:43 I was however the only guy in New Rocks, Black Jeans and shirt in Universal studios Oct 09 10:16:48 I caused quite a stir :-D Oct 09 10:16:55 heh Oct 09 10:17:18 whole groups of people just staring as I walked past, wish I had done the makeup as well now :-) Oct 09 10:17:49 but that was because I was going to be on TV, I refuse to be on tv looking like a holiday maker Oct 09 10:18:47 sounds like a fun time Oct 09 10:51:15 * raster hands XorA some hairgel Oct 09 11:45:12 hi Oct 09 11:45:18 hi woglinde Oct 09 11:56:01 hms I did broke it again Oct 09 12:13:30 hrw: pastebin looks good, please add Oct 09 12:13:39 hrw: yes, i will drop PR for python 2.6 Oct 09 12:13:47 need to run now Oct 09 12:13:52 l8r Oct 09 12:22:17 ok Oct 09 12:49:11 OK, I now have the way to use busybox instead of tinylogin for the "login" utils (getty, su, sulogin and their friends). Are you interested in a patch ? Oct 09 12:53:18 hi Oct 09 12:55:14 hey cbrake Oct 09 12:55:29 hm, been a while that i've been around here Oct 09 12:57:12 http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=commitdiff;h=c706a3191cfcfa843fb9f86a5fceb60da69d31ae Oct 09 12:57:56 things have changed and I am somewhat lost here ... trying to build angstrom-2008 for ARM I run into a problem when bitbake tries to use an external toolchain in /usr/local/angstrom which I haven't set up .. when I last worked on my stuff the toolchain was built per default instead of expecting an external one Oct 09 12:58:01 can someone give me a hint on this? Oct 09 13:00:23 what machine are you building for? Oct 09 13:01:37 Crofton, arm Oct 09 13:01:46 (xscale) Oct 09 13:01:58 what is MACHINE var set o? Oct 09 13:02:18 ah sorry Oct 09 13:02:24 that would be cm-x270 Oct 09 13:03:26 angstrom-2008.1 Oct 09 13:03:26 maybe look in the machine .conf file and see if it is trying to use an external toolchain .. Oct 09 13:04:00 cbrake, you are maintaing that aren't u? (cm-x270) Oct 09 13:05:34 hi Oct 09 13:05:34 can't spot anything related in the machine file Oct 09 13:06:17 anything in your localconf? Oct 09 13:06:56 I doubt it since that one used to work just fine .. however that was about half a year or more ago .. so who knows what changed Oct 09 13:07:48 I would go through it carefully Oct 09 13:07:59 somehow you are telling it to use an external toolchain Oct 09 13:08:06 it's the dev doing "Identifying dependency loops", do you know hot to get this? Oct 09 13:08:26 working? Oct 09 13:10:09 gilligan_: yes, I maintain the cm-x270 stuff Oct 09 13:10:34 Crofton|work, http://pastebin.com/m4f9c3d1c Oct 09 13:10:43 Crofton|work, it is pretty simple Oct 09 13:10:44 gilligan_: I've been doing builds from oe.dev and they seem to work as of several weeks ago Oct 09 13:11:21 i just clean the tmp env and try to build Oct 09 13:11:26 it gives me this errors Oct 09 13:11:37 cbrake, I'm trying the stable branch right now Oct 09 13:12:00 gilligan_: I've never build cm-x270 in the stable branch, so I can't say if it works there Oct 09 13:12:05 methril, that has been fixed in the past few hours i think Oct 09 13:12:11 ok Oct 09 13:12:15 i pull again Oct 09 13:12:18 i'm on git Oct 09 13:12:25 cbrake, in that case I think i'll first give it a try with the dev branch Oct 09 13:12:44 gilligan_: yeah, dev has been pretty stable for me, for that is what I would recommend Oct 09 13:12:55 except when it isn't :) Oct 09 13:13:02 Crofton|work: :-) Oct 09 13:13:10 Crofton|work: then I wait a week ... Oct 09 13:13:12 cbrake, well the last time I used oedev stuff (maybe you actually remember me hehe.. its' been a while ) there was no stable branch at all Oct 09 13:13:53 or I was just too ignorant to see it? hehe.. at least i can't recall it Oct 09 13:14:58 gilligan_: yes, its been awhile Oct 09 13:15:28 i see oedev is switching to git .. THANK GOD for that Oct 09 13:15:34 yes! Oct 09 13:15:37 *g* Oct 09 13:15:45 god i hate(d) using mtn Oct 09 13:15:45 thank zecke and mickeyl Oct 09 13:16:16 thanks Oct 09 13:16:18 it works Oct 09 13:17:46 hi mr_nice Oct 09 13:17:54 woglinde hi Oct 09 13:18:49 does anyone knows a distribution build with oe which supports buffalo linkstation pro duo? i saw a machine.conf for lsarm Oct 09 13:23:30 03  07org.openembedded.dev * rc706a3191c 10org.openembedded.dev.git/packages/busybox/ (3 files in 2 dirs): busybox 1.11.1: update angstrom config to be able to replace mount, init, tinylogin and add more coreutils apps Oct 09 13:25:26 03  07org.openembedded.dev * r43c6479e40 10org.openembedded.dev.git/packages/tinylogin/tinylogin_1.4.bb: tinylogin: have all applets use update-alternatives so they can be replaced by other packages (e.g. proper login, busybox, etc) Oct 09 13:26:10 03  07org.openembedded.dev * r84eedfba7e 10org.openembedded.dev.git/packages/sysvinit/sysvinit_2.86.bb: sysvinit: add poweroff to update-alternatives Oct 09 13:26:20 03  07org.openembedded.dev * r37f391e1b4 10org.openembedded.dev.git/packages/makedevs/makedevs_1.0.0.bb: makedevs: use update-alternatives Oct 09 13:26:28 03  07org.openembedded.dev * r1a089efb1f 10org.openembedded.dev.git/conf/checksums.ini: checksums: add some more checksums from angstrom autobuilders Oct 09 13:32:04 argh.. lost yet again :) Oct 09 13:32:11 brb Oct 09 13:34:46 BBPATH=/home/gilligan/projects/oedev/pxa270_nor/build:/home/gilligan/projects/oedev/pxa270_nor/org.openembedded.dev bitbake -v -d atm-nor-image Oct 09 13:34:46 ERROR: Unable to parse conf/bitbake.conf (conf/bitbake.conf:312: unparsed line: 'export PATH') Oct 09 13:34:46 make: *** [image] Error 1 Oct 09 13:35:00 i'm seriously clueless about that one, sorry Oct 09 13:37:23 don't see how 'export PATH' at that point could lead to any errors Oct 09 13:47:40 03  07org.openembedded.dev * r05c2f24c53 10org.openembedded.dev.git/packages/util-linux-ng/ (util-linux-ng.inc util-linux-ng_2.14.bb): util-linux-ng: add readprofile to update-alternatives Oct 09 13:47:50 03  07org.openembedded.dev * r5bc9c5c585 10org.openembedded.dev.git/conf/distro/include/angstrom-2008-preferred-versions.inc: angstrom 2008: prefer busybox 1.11.1 Oct 09 13:58:58 still stuck with bitbake complaining about not being able to parse "export PATH" in conf/bitbake.conf ... x_X Oct 09 14:00:27 gilligan_ why you set it where? Oct 09 14:00:37 bitbake reads the PATH from env Oct 09 14:01:31 I suggest you reread the manual Oct 09 14:01:33 http://bec-systems.com/oe/html/index.html Oct 09 14:03:17 sigh.. the whole setup worked just *fine* some months ago Oct 09 14:04:55 gilligan_: well, what did you change? Oct 09 14:05:05 mothns? Oct 09 14:05:16 we improve it every day Oct 09 14:06:02 your config is wrong, bitbake is not good at reporting the right error, I could change it, but I'm too busy editing wiki's (which even less people can do) Oct 09 14:06:50 zecke, yeah I suppose so.. Oct 09 14:07:13 gilligan_: anyway, what do you attempt to do? Oct 09 14:09:24 zecke, trying to built a slightly customized angstrom image .. but I don't get far because the error above is what I get right after executing bitbake and I don't understand what the actual issue is Oct 09 14:09:47 he zecke Oct 09 14:09:52 zecke, it says it fails to parse 'export PATH' .. which is rather weird Oct 09 14:14:04 re Oct 09 14:15:57 hi ant Oct 09 14:16:32 hey woglinde Oct 09 14:16:55 any OM dev here now? XorA? Oct 09 14:17:11 Any hints about booting from ext3? I compiled support in (kexecboot) kernel but get this: VFS:cannot open root device "mmcblk0p2" or unknown block(179,2) Oct 09 14:17:39 I see there are tickets in Openmoko about boot from SD / ext3 Oct 09 14:20:06 well, I'm using this device table: [org.openembedded.dev.git] / files / device_table-kexecboot.txt Oct 09 14:21:50 * ant|work realizes now the kernel booted by kexecboot has NO ext3 support... Oct 09 14:21:57 this is really fucked up Oct 09 14:22:21 *g* Oct 09 14:31:09 zecke, okay.. it turns out the bitbake version I had installed was too old, but only when I removed "export PATH" at line 312 it would get far enough to let the sanity checker report that bitbake is indeed too old Oct 09 14:47:04 How can I debug why one of my BBFILE_COLLECTIONS is not masking upstream? Oct 09 14:47:52 I can see unique package in my collection but my linux package in my collection will not mask upstream. Oct 09 15:02:50 * mwester missed the busybox discussion; a shame because SlugOS switched to busybox for most everything (including tinylogin, but not mdev). Oct 09 15:03:18 :( Oct 09 17:12:13 hm.. trying to build an angstrom2008.1/i686 image -- compilation of glibc-2.6.1 fails tho -- http://pastebin.com/m1c61672c Oct 09 17:12:23 anyone knows anything about that? Oct 09 17:17:05 gilligan_: new one on me, but I only ever do arm builds Oct 09 17:27:27 what is the chance of meta-toolchain making something "useful"? Oct 09 17:29:42 define "useful"? last I checked (10 months ago) it was working fine Oct 09 17:30:00 hhe Oct 09 17:30:14 what I don't like about that remark is the ten months :) Oct 09 17:30:28 * mwester just generated an fso toolchain two days ago... Oct 09 17:30:34 if it builds, I am going to see if any of the beagle users will try it Oct 09 17:31:02 the headache is, this will create alot of questions I can't answer :( Oct 09 17:31:20 Lots of questions, but expect you can answer them! Oct 09 17:31:22 but may be better than having people creating other (non OE based) solutions Oct 09 17:32:15 The number one question topic for Openmoko's toolchain seems to be "How do I add libabcdefg to the toolchain?" Oct 09 17:32:23 heh Oct 09 17:32:29 and the answer is? Oct 09 17:33:03 good luck Oct 09 17:33:05 I've always used the answer: "set up a full OE environment; you're beyond what the toolchain can offer" Oct 09 17:33:13 but apparently one can use opkg. Oct 09 17:33:24 heh Oct 09 17:33:25 and add the libabcdefg-dev pacakge to the toolchain. Oct 09 17:33:32 Dunno how that works, but it works. Oct 09 17:34:05 we had the question, I built something with csl gcc, why am I having issues with stdlibc++ when I run on the beagle Oct 09 17:34:09 The other solution is just to add the library to the toolchain itself in OE, so that the next version of the toolchain has the lib -- makes sense for common ones (libssl for example) Oct 09 17:39:11 adding libs to toolchain ? Oct 09 17:43:03 khem: Yes -- it seems that users have an expectation that the toolchain for a device contain the headers and lib.a files matching the standard rootfs, so they don't have to build libssl (for example) themselves. Oct 09 17:44:13 mwester: well I think what they need is a dev environment Oct 09 17:44:24 mwester: which is toolchain+various dev libs Oct 09 17:44:27 and headers Oct 09 17:44:42 khem: basicly an SDK Oct 09 17:44:45 Ok, so it might be a terminology issue, then. The need is the same though. Oct 09 17:45:09 yes essentially toolchain is just a think that can compile Oct 09 17:45:29 you need a dev env which is superset Oct 09 17:55:09 03  07org.openembedded.dev * rf4993bf2a5 10org.openembedded.dev.git/: Oct 09 17:55:09 merge of '1415f32ae312fabcfb6f1452dbd7ce897665ca88' Oct 09 17:55:09 and '6a04f6b10c819319120eee8f1e058fe39a818c6f' Oct 09 17:55:09 03  07org.openembedded.dev * r64d6ea635a 10org.openembedded.dev.git/packages/freesmartphone/ (frameworkd/om-gta01/frameworkd.conf frameworkd_git.bb): Oct 09 17:55:10 frameworkd: Use GllinChannel for GTA01. Together with gllin version 1.1 (or Oct 09 17:55:12 any initscript in /etc/init.d/gllin) GPS will work out of the box for GTA01. Oct 09 18:33:00 03  07org.openembedded.dev * rc860a35b6c 10org.openembedded.dev.git/packages/mozilla/fennec_hg.bb: fennec: bump SRCREV Oct 09 19:05:00 evening Oct 09 19:06:35 I have a mysterious problem. bitbake console-image with added INSTALL packages, fails because of unknown package. bitbake does well: http://pastebin.com/m600c824b Oct 09 19:06:41 thanks in advance Oct 09 19:26:52 added INSTALL packages... Oct 09 19:26:58 show which ones Oct 09 19:31:24 hm Oct 09 19:31:43 seems like there is a broken dependency.. nfs-utils requires libblkid Oct 09 19:37:18 and there is no libblkid package.. grmbl Oct 09 19:37:23 hrw: thanks for answering. As in the pastebin is mentioned one of the INSTALL packages is ttf-dejavu. So bitbake says: RDEPENDS/RRECOMMENDS or otherwise requires the runtime entity 'ttf-dejavu' but it wasn't found in any PACKAGE or RPROVIDES variables Oct 09 19:43:02 check which packages does ttf-dejavu generate Oct 09 19:43:44 because IMAGE_INSTALL lists packages not recipe names Oct 09 19:44:25 szsoftware: and instead of abusing console-image create rather szsoftware-image which will be based on console-image. this way we will know that this is custom Oct 09 19:47:16 gilligan_: which version? I do not see such dependency in OE Oct 09 19:48:45 hrw, exactly Oct 09 19:48:51 hrw, and that is why nfs-utils fails Oct 09 19:49:00 hrw, as configure checks for the library Oct 09 19:49:17 libbkid = block device id library needed by nfs utils Oct 09 19:49:21 hrw: thanks for your hint and advice. Will do something about that... Oct 09 19:49:41 gilligan_: iirc bbkid is needed if you want nfs4 Oct 09 19:49:53 and by default nfs4 is disabled Oct 09 19:51:28 I remember going down this road, attempting to get nfs4 support for SlugOS -- IIRC blkid is provided by either udev or a dependency of udev, but something didn't work right for me and I never got all the dependencies to be satisfied. :( Oct 09 19:52:39 hrw, well I added nfs-utils to my image .. and it fails because it requires bbkid and the nfs-utils package does not depend on bkkid let alone does a bkkid library exist at all -- as for any details on that matter I am afraid I have no clue (only nfs4 requiring it etc..) Oct 09 19:52:52 in either way nfs-utils can't be compiled right now Oct 09 19:53:37 mwester: I had nfs4 once but lost recipes Oct 09 19:53:50 it was building but there was something broken Oct 09 19:54:14 i don't even need nfs4 :) nfs3 would suit me just fine Oct 09 19:55:26 I do not have OE based build now so will not test... Oct 09 20:33:54 bye Oct 09 20:40:25 re Oct 09 20:41:46 Mr. ant__ : I hear you are working on a kexec-based solution to permit a user to select a kernel and rootfs to boot from? Oct 09 20:42:18 hi, yes...actually thesing wrote it Oct 09 20:42:36 me and Jay are debugging/extending Oct 09 20:42:52 (checking now ext3 from sD) Oct 09 20:46:04 wb Jay Oct 09 20:47:03 ant__: hi Oct 09 20:47:10 What image would I build to see how this is done (I need to do a similar thing for the FSO image for the Openmoko hardware) Oct 09 20:47:37 if you have framebuffer, just 'linux-kexecboot' Oct 09 20:47:39 something strange with my ISP :) Oct 09 20:47:52 but you have to enable your machine Oct 09 20:47:54 Thanks, I'll take a look! Oct 09 20:47:58 np Oct 09 20:48:14 be around for a couple of hours Oct 09 20:48:25 s/be/I'll be/ Oct 09 20:49:19 ant__: I'll end my current project at Oct 19 Oct 09 20:49:35 then I'll be back to game ;) Oct 09 20:49:50 great Oct 09 20:50:21 I hope :) Oct 09 20:50:52 hello Oct 09 20:51:15 is there any ports of OE to windows? Oct 09 20:51:31 using mingw32? Oct 09 20:52:17 howling_mad: it's easier to have virtual machine with some linux inside, IMHO Oct 09 20:53:13 Jay7: I though about that, but I am really keen on some code editors in windows, makes the job a lot easier Oct 09 20:53:16 I don't know anything about such port and possibility to create such port.. Oct 09 20:53:38 also, what glibc are you guys using now? Oct 09 20:56:17 I need to setup a build environment for the open pandora Oct 09 20:56:41 something that will give me fast hability to recompile the kernel Oct 09 20:56:51 and normal code as well Oct 09 20:57:25 mwester: pls note you have to require conf/machine/include/kexecboot.inc Oct 09 20:57:39 in your machine.conf Oct 09 21:11:15 03  07org.openembedded.dev * rb5781bc692 10org.openembedded.dev.git/packages/gcc/ (10 files in 2 dirs): Oct 09 21:11:15 gcc cross: add 4.3.2+csl2008q3, the latest csl release Oct 09 21:11:15 * it builds, but needs Khems sysroot and flags patches to work properly Oct 09 21:14:26 ah Oct 09 21:14:31 cls2008 Oct 09 21:14:46 csl Oct 09 21:14:52 sorry Oct 09 21:15:03 just downloaded and installed it yesterday Oct 09 21:15:11 where are these patches? Oct 09 21:15:20 git? Oct 09 21:30:26 any really good c editors for linux then? Oct 09 21:31:02 howling_mad: KDevelop :) Oct 09 21:31:02 The best Oct 09 21:31:22 no minimalist thing? Oct 09 21:31:26 Eclipse, Emacs, VIM, to name a few Oct 09 21:31:59 jed Oct 09 21:33:00 heres what I find good in an editor: code completion, customizable tools, and off course, a good source file brower with function definition/declaration separation Oct 09 21:33:29 Thats really sort of conflicts with the minimalist critera Oct 09 21:33:45 heh! Oct 09 21:33:52 the thing with KDE is my ati card in my laptop Oct 09 21:34:15 blame ms for being b*******ds but vs works great for code editing Oct 09 21:34:54 Gack! But how do you classify Visual Studio as "minimalist"??? Oct 09 21:35:18 Minimalist as in it does not make my pc overheat Oct 09 21:35:29 but it has been revamped since then Oct 09 21:35:57 vim :) Oct 09 21:36:01 Well, ok, let's put definitions aside. Oct 09 21:36:04 or even gvim :) Oct 09 21:36:21 but with tons of scripts :) Oct 09 21:36:21 emacs/xemacs :) Oct 09 21:36:38 Eclipse Oct 09 21:36:46 eclipse is good Oct 09 21:37:01 used it a couple of times Oct 09 21:37:11 but something a bit more native Oct 09 21:37:16 no java stuff Oct 09 21:38:42 ok Oct 09 21:38:49 a minimalist distro then Oct 09 21:38:57 basically for compilling Oct 09 21:39:23 * howling_mad do not recomend gentoo, had it for a year Oct 09 21:40:50 Jay7:mwester: debug logs about boot from ext2/ext3 SD cards: http://rafb.net/p/Xjua9572.html http://rafb.net/p/mY9u7s59.html  Oct 09 21:41:22 it was my fault...kernel in /boot had no ext3 compiled in... Oct 09 21:41:32 retrying now Oct 09 21:43:38 ant__: did you using standard nand partitioning now? Oct 09 21:43:49 yes atm Oct 09 21:44:02 ah, ok Oct 09 21:44:15 I just have not seen mtdparts in cmdline :) Oct 09 21:44:39 he..kexec works perfectly ... Oct 09 21:45:01 howling_mad: don't agree with you about Gentoo! used it for about 6 years (since its creation) and love it! Oct 09 21:45:14 HokieTux: same here Oct 09 21:45:22 howling_mad: same here Oct 09 21:45:39 denix, but if you have to build every package you have it becomes boring after a while Oct 09 21:46:19 * Jay7 was using gentoo about one year Oct 09 21:46:47 do not use unstable branch! :) Oct 09 21:48:08 I don't mind compiling stuff, but it gives me much more control over my system Oct 09 21:48:34 btw, if you don't like building/rebuilding stuff, what are you doing in OE? :) Oct 09 21:49:08 I do like compiling/building stuff, the stuff I am working on, not every damn app ;-) Oct 09 21:49:38 anyway, my pc builds something in the background all the time... Oct 09 21:51:16 how often do you need to rebuild every app on your system? Oct 09 21:52:10 now not at all Oct 09 21:52:26 but it was nice playing with e17 Oct 09 21:56:32 denix: ever used proteus/labview, microchip stuff_ Oct 09 21:56:58 evening again Oct 09 21:57:37 howling_mad: nope Oct 09 21:58:29 that is one of the main reasons I use windows Oct 09 21:58:36 electronic simulation Oct 09 21:58:52 I moved the oe tmp directory. I've already replaced all old_tmp with new_tmp strings in every neccessary files. Nevertheless I get this error: /bin/sh: /home/sven/oe/stuff/build/tmp/staging/i686-linux/usr/bin/pod2man: not found Oct 09 21:59:00 how to fix that/rebuild? Oct 09 21:59:35 thanks in advance Oct 09 22:00:01 howling_mad: there are some specialized tools only available for windows... Oct 09 22:01:29 denix: precisely why I will have to vm my linux system Oct 09 22:02:13 Jay7: booting now from ext3 ! Oct 09 22:02:35 hi Oct 09 22:02:42 hello woglinde Oct 09 22:02:44 woglinde: evening Oct 09 22:03:01 ~seen kalev Oct 09 22:03:03 kalev is currently on #oe. Has said a total of 23 messages. Is idling for 1d 5m 41s, last said: 'thanks'. Oct 09 22:04:05 woglinde: yesterday you helpedmea lot. Do you see my line 4 mins ago? Oct 09 22:04:35 no Oct 09 22:04:40 szsoftware: after moving tmpdir you have to rebuild from scratch Oct 09 22:04:44 because I wasnt here 4 minutes ago Oct 09 22:04:51 woglinde: andsorry for the missing spaces, a lot of jam in the keyboard ^^ Oct 09 22:06:04 woglinde: you are always with us, virtually :) Oct 09 22:06:14 ant__: do you successfully booted? or booting in progress? :) Oct 09 22:06:24 ant__: please explain why. I've replaced a very lot of pathes (tmp related) with sed and everything seems to be ok, except the above mentioned error Oct 09 22:06:25 booted twice...loging now Oct 09 22:06:32 good Oct 09 22:06:59 hehe Oct 09 22:07:05 should I break oe again Oct 09 22:07:15 Yes yes! It was fun! Oct 09 22:07:34 mwester sorry Oct 09 22:07:48 hm Oct 09 22:07:50 23:57 < szsoftware> how to fix that/rebuild? Oct 09 22:07:54 why it is working here Oct 09 22:07:54 ups Oct 09 22:08:13 :) That's why they're autobuilders -- they pick up from where they left as soon as things get put right, no big deal at all. Oct 09 22:08:50 woglinde: I moved the oe tmp directory. I've already replaced all old_tmp with new_tmp strings in every neccessary files. Nevertheless I get this error: /bin/sh: /home/sven/oe/stuff/build/tmp/staging/i686-linux/usr/bin/pod2man: not found Oct 09 22:08:55 23:57 < szsoftware> how to fix that/rebuild? Oct 09 22:09:56 Jay7: http://rafb.net/p/ACquz520.html Oct 09 22:10:11 hm Oct 09 22:12:13 hi woglinde Oct 09 22:12:42 I'm sorry, seems I don't have my mmc device any more, couldn't test it Oct 09 22:13:20 I looked at your pmount gettext patch. Silly developers who thought that everyone wants to use gettext :) Oct 09 22:13:29 ant__: did you try build kernel without net and use -x option of kexec? Oct 09 22:13:44 yes, no net for linux-kexecboot Oct 09 22:13:51 and this last patch: http://rafb.net/p/UCw7kj12.html Oct 09 22:14:39 for *both* kernels ;-) Oct 09 22:14:39 ah, yes Oct 09 22:15:20 hi raster Oct 09 22:15:24 woglinde: you should try to sent the gettext patch upstream I think Oct 09 22:15:26 kalev does it works? Oct 09 22:15:35 beep Oct 09 22:15:54 kalev I mean runtime Oct 09 22:18:19 hm what I am doing wrong Oct 09 22:18:33 here my gettext patch works with and without locale Oct 09 22:18:38 dont get it Oct 09 22:18:50 woglinde: I don't know Oct 09 22:18:52 can anyone try at his side? Oct 09 22:19:05 hm maybee glibc Oct 09 22:19:09 mom Oct 09 22:19:17 I tried to do a full rebuild of my tree last night to check if another one of my patches didn't break it, but it didn't build Oct 09 22:19:30 so I couldn't test it, sorry Oct 09 22:19:33 it's compiling right now Oct 09 22:23:01 args okay Oct 09 22:23:03 glibc breaks Oct 09 22:24:31 glibc is shitty *g* Oct 09 22:24:54 woglinde: use eglibc :) Oct 09 22:25:11 khem nope uclibc Oct 09 22:25:51 for embedded envs I think eglibc would be best once the kconfig is implemented Oct 09 22:26:01 khem hm Oct 09 22:27:08 I wish I had enough time Oct 09 22:27:19 khem me too Oct 09 22:27:25 with two kids and closed source job it becomes hard Oct 09 22:29:09 * denix is away: I'm not here Oct 09 22:47:51 good night! Oct 09 22:52:58 hm Oct 09 22:53:03 I found what is wrong Oct 09 22:53:15 binutils-cross dependency Oct 09 22:53:29 hmm Oct 09 22:55:47 yes binutils-cross depends on virtual/libintl and virtual/libiconv Oct 09 22:55:53 makes no sense Oct 09 22:56:09 woglinde: it should not Oct 09 22:56:11 khem what now with uclibc install utils in staging? Oct 09 22:57:04 hm I will rewrite gettext.bbclass Oct 09 22:57:11 that it checks if the package is cross Oct 09 22:57:43 utils have some issue ? Oct 09 22:58:08 khem what will you do with crosscompiled libc utils in staging? Oct 09 22:58:17 unless you are using qemu Oct 09 22:58:24 for ldconfig Oct 09 22:58:54 netherless ldd is installed in the wrong place stagingpath/usr/usr/bin Oct 09 23:03:44 * Jay7 -> sleep() Oct 09 23:03:55 'nite Jay7 Oct 09 23:20:00 03  07org.openembedded.dev * r49886fe027 10org.openembedded.dev.git/packages/konqueror/konqueror-embedded-20070316/fix_configure.patch: konqueror-embedded: add patch missing from previous commit Oct 09 23:20:11 03  07org.openembedded.dev * r5b2c4a988b 10org.openembedded.dev.git/: Oct 09 23:20:11 merge of 'aa332b0757a7f250aec59dac51d57c4cdfda5e6e' Oct 09 23:20:11 and 'b74f88e907776bb75ebf406e06efc1fa0a2ea1c0' Oct 09 23:25:35 nite Oct 10 00:13:24 nite **** ENDING LOGGING AT Fri Oct 10 03:00:06 2008