**** BEGIN LOGGING AT Mon Jan 30 02:59:57 2006 Jan 30 03:15:30 morning Jan 30 03:27:29 mehhhh who broke RUNSTRIP? At least have your new STRIP routine drop an error message when it fails Jan 30 03:29:10 CoreDump|home: there was some discussion on this on oe@ unfortuneately I deleted it this morning thinking it wasnt essential :-( Jan 30 03:29:44 I'm not entirely sure who/what caused it to break but a temporary workaround is to set IGNORE_STRIP_ERRORS in local.conf Jan 30 03:29:58 I don't have a problem when someones fixes STRIP for the better, however, having RUNSTRIP fail and the logfile beeing empty stinks in my book Jan 30 03:34:44 is there any need for a tar.gz mirror of bitbake / oe? Jan 30 03:39:04 morning Jan 30 03:39:15 hi hrw|work Jan 30 03:39:38 hi hrw|work Jan 30 03:39:48 CoreDump|home: As a number of people have phobias of monotone, tar.gz mirrors of oe are probably a good thing (I think koen has them on ewi?) Jan 30 03:39:51 CoreDump|home: GPE bugs are something which I cannot control - its good to have them fixed before release but we can fix them via upgrades and/or 3.5.4.1 Jan 30 03:40:37 RP: I was more talking about bitbake / OE itself, not the database =) Jan 30 03:40:48 ~ewi Jan 30 03:40:50 ewi is probably ewi546.ewi.utwente.nl - main backup server for OE monotone. Hosts database snapshot: http://ewi546.ewi.utwente.nl/OE/OE.db.bz2 Jan 30 03:41:08 CoreDump|home: http://ewi546.ewi.utwente.nl/OE/OE-20060130.tar.bz2 maybe? Jan 30 03:41:17 hrw|work: ok ignore that =) Jan 30 03:41:28 CoreDump|home: I thought there were copies of the database and checkouts? Jan 30 03:42:05 I have my server setup to auto sync the db so it wouldn't have been much work to provide .tar.gz's. But since koen is already doing so, there'd be no point Jan 30 03:43:22 hrw|work: I'd love to see some of the more annoying bugs fixed in GPE. It's a great release as it is but its the details that count in the user's eye. Jan 30 03:43:53 CoreDump|home: ok - we could wait for it Jan 30 03:43:59 for example the b0rked font size in the menu or in matchbox's title Jan 30 03:44:09 the non-working hotkeys Jan 30 03:44:35 and a showstopper: gpe-login crashing due to keylaunch on some models Jan 30 03:44:47 fontsize - did not xora reported succes on it? Jan 30 03:45:08 hrw|work, I think I remember something about that on #gpe Jan 30 03:45:16 ok. time to test cryopid to preserve irssi buffers and reboot Jan 30 03:45:25 dunno, I'll build a fresh 3.5.4 image this evening and see Jan 30 03:45:26 11:45:57 up 23 days, 17:22, 2 users, load average: 1.17, 1.24, 1.34 Jan 30 03:45:45 hm.. I should probably rebuild test iamges Jan 30 03:46:01 mhentges@h6563:..l/tmp/do_not_use/.upload >uptime Jan 30 03:46:01 11:46:09 up 408 days, 18:26, 3 users, load average: 0.06, 0.23, 0.43 Jan 30 03:46:07 oops Jan 30 03:46:21 that path was supposed to be hidden *sigh* Jan 30 03:47:00 ;) Jan 30 03:47:04 brb Jan 30 03:47:07 Its only preserved in ibot logs for all to see now :) Jan 30 03:47:14 CoreDump|home: the b0rked menu fonts should be fixed, hopefully matchbox will make another release before oz354 Jan 30 03:47:25 XorA: excellent! Jan 30 03:47:33 RP: time to rename =) Jan 30 03:48:20 Relying on security by obsurity is never a good idea alone... Jan 30 03:48:37 RP: oh that isn't a security thing at all Jan 30 03:49:01 it's just a folder to uploa images to. I don't want users to d/l incomplete files, that's all Jan 30 03:49:33 fair enough :) Jan 30 03:49:52 which reminds me...anyone here w/ a poodle? Jan 30 03:50:31 * CoreDump|home already knew the answer to that Jan 30 03:53:28 XorA: was there any descision if keylaunch will be removed for the release? A non-working login is one hell of a bug Jan 30 03:56:21 CoreDump|home: thats a OZ decision, the fault doesnt appear on Familiar according to GPE guys Jan 30 03:56:23 re Jan 30 03:56:32 cryopid works nice :) Jan 30 03:57:34 what did I miss? Jan 30 03:57:43 well, it's a GPE descision as keylaunch is only in the GPE image =) Jan 30 03:57:52 hrw|afk: have you time to try a perl-native patch for me, on i386, it fixes a multilib problem on x86_64, but I dont want to break i386 Jan 30 03:58:06 XorA: I can Jan 30 03:58:53 hrw|work: cool, Im just double checking here, then I send you a url Jan 30 03:59:04 ok Jan 30 03:59:33 ~lart firefox trunk Jan 30 03:59:34 * ibot moos at firefox trunk Jan 30 04:00:30 hrw|work: http://hyboria.mine.nu/Configure-multilib.patch Jan 30 04:01:46 ok Jan 30 04:07:02 hrw|work: could you pull in initscrits from dev? It completes the keymap for Akita / Spitz / Borzoi and adds support for keymap-locales Jan 30 04:07:32 hmm looks lile ewi's mt is down Jan 30 04:08:03 CoreDump|home: I could Jan 30 04:08:50 until thursday I have quiet days at work - coworkers on vacations or sick Jan 30 04:08:51 initscripts from.dev work fine in branch, I've been using it since day one Jan 30 04:09:00 great =) Jan 30 04:09:16 only project manager may come Jan 30 04:09:18 * RP wishes he could have a quiet day :-/ Jan 30 04:10:11 ~lart kopete and jabber Jan 30 04:10:12 * ibot pries kopete and jabber's back open with a screwdriver and flashes a new bootldr to kopete and jabber Jan 30 04:10:23 ~lart OZ for mount SD cards sync Jan 30 04:10:23 * ibot says "boot to the head" and knocks OZ over for mount SD cards sync Jan 30 04:12:20 ~lart midnightcommander ftp mode Jan 30 04:12:20 * ibot flings poo at midnightcommander ftp mode Jan 30 04:13:05 ibot: that's disgusting Jan 30 04:13:16 :-) Jan 30 04:13:49 XorA: NOTE: package perl-native-5.8.7-r1: task do_compile: started Jan 30 04:14:33 * XorA crosses fingers Jan 30 04:18:16 NOTE: package perl-native-5.8.7-r1: task do_populate_staging: started Jan 30 04:20:07 grrr, I forgot how "barebone" a standard image is Jan 30 04:20:56 XorA: NOTE: package perl-native-5.8.7-r1: task do_build: completed Jan 30 04:21:06 CoreDump|home: collie? Jan 30 04:21:14 akita Jan 30 04:21:26 no fsck, no mkfs no "less" etc Jan 30 04:21:39 hrw|work: safe to checkin do you think? Jan 30 04:21:49 hrw|work: or seeing as its perl should I bug tracker it Jan 30 04:23:23 * XorA looks at maintainer and the magic in .bb and decides to bug tracker it Jan 30 04:23:31 XorA: I do not know Jan 30 04:23:48 CoreDump|home: I was thinking about adding e2fsprogs to base for all Jan 30 04:23:57 hrw|work: got that ip address handy? Jan 30 04:24:57 hrw|work: got my vote for that heh. People are using SD and CF cards so basic FS tools should be there. At least on devices > collie Jan 30 04:28:14 XorA: ? Jan 30 04:28:21 Lets create a task-bootstrap-plus or something Jan 30 04:28:35 bootstrap should be the bare minimum needed to boot Jan 30 04:29:12 I agree Jan 30 04:29:43 hrw|work: for bugs.oe.org Jan 30 04:30:41 XorA: read topic? Jan 30 04:31:25 hrw|work: without specifically going to it, irssi only shows first 80 chars of topic, cheers Jan 30 04:31:38 XorA: /topic show all Jan 30 04:31:43 morning all Jan 30 04:31:46 hi dirk Jan 30 04:31:59 great.. Jan 30 04:32:04 hrw|work: ah pish, now look what you made me do Jan 30 04:32:06 hi hrw|work Jan 30 04:32:47 XorA: "/topic" show all Jan 30 04:32:51 now better Jan 30 04:34:01 do13_: did you had a change to check 2.4 keymap on tosa? Jan 30 04:34:35 hrw|work: Yes, the 2 characters are fixed Jan 30 04:34:53 do13_: ok - I will update keymaps in branch then Jan 30 04:36:38 CoreDump|home: does collie/keymap-2.4.map has all keys working ok? Jan 30 04:36:40 now lets see if the SD card boots Jan 30 04:37:03 hrw|work: for the english keyboard: yes Jan 30 04:37:42 ok Jan 30 04:38:06 hrw|work: have a look at "keyboard-locales" Jan 30 04:38:55 CoreDump|home: so I have to add: initscripts/keymap and keyboard-locales? Jan 30 04:39:32 that would be the best, yes. Jan 30 04:39:51 it isn't required tho but locales comes in handy for localized keymaps Jan 30 04:40:12 and the new initscripts support / auto-load them Jan 30 04:40:34 rshift as altgr? Jan 30 04:40:54 on german collie keyboards that is the case Jan 30 04:41:15 I have normal one - qwerty as it should be ;) Jan 30 04:41:23 =) Jan 30 04:41:43 this mapping shouldn't be in keymap-2.4.map Jan 30 04:42:26 we have 8 revs difference in initscripts Jan 30 04:43:42 CoreDump|home: could you update article about keymaps on oz.org? Jan 30 04:43:55 didn't even know there is one Jan 30 04:44:14 http://openzaurus.org/wordpress/2006/01/09/c7x0-and-cxx00-26-kernel-information/ Jan 30 04:44:52 ah,kernel 2.6. Will do after I fix the 2.6 keymap for SL-Cxxxx Jan 30 04:45:25 CoreDump|home: 2.4 tricks could also be described Jan 30 04:45:34 right Jan 30 04:47:46 RP: did you wire the Mail LED to act as an HDD LED? Jan 30 04:48:42 keymaps pushed Jan 30 04:48:47 ty Jan 30 04:49:13 CoreDump|home: yes Jan 30 04:49:18 nice one heh Jan 30 04:49:47 yay! First altboot SD boot in 2.6 Jan 30 04:50:26 CoreDump|home: yay, you making config files for 2.6 kernels, yay Jan 30 04:51:01 CoreDump|home: Look in /sys/class/leds for fun with the leds class :) Jan 30 04:51:08 RP: will do Jan 30 04:51:34 XorA: I've pushed the changes to ewi but ewi has gone done since then Jan 30 04:51:35 Someone could document it on oz.org actually... Jan 30 04:52:08 btw: "sync" mounted SD + booting off SD stinks Jan 30 04:52:43 I'd expect that... Jan 30 04:53:09 RP: could you give me a hint how to make sound work? alsmixer shows tons and tons of settings Jan 30 04:53:23 CoreDump|home: Did you read the above kernel information page? Jan 30 04:53:47 oops Jan 30 04:53:50 =) Jan 30 04:54:45 CoreDump|home: I assume it will work if SD is one partition with ext2 Jan 30 04:55:03 CoreDump|home: Ive been running with root=/dev/mmcblk0p1 for months now Jan 30 04:55:17 XorA: actually the partition can be configured in /etc/altboot.cfg Jan 30 04:55:22 XorA: heh Jan 30 04:56:02 CoreDump|home: but altboot would be nice so I can have opie on internal flash just to check out now and then :-) Jan 30 04:56:50 uhh, there's no "boot from flash" option, yet as that was the default in the past Jan 30 04:57:02 shouldn't be hard to do tho Jan 30 04:58:12 CoreDump|home: xora boots directly from sd Jan 30 04:58:35 and if he wants altboot to boot his flash instead he would be a "boot from flash" option =) Jan 30 04:58:44 *he would need Jan 30 05:00:14 ~lart phpbb modders Jan 30 05:00:15 * ibot beats phpbb modders senseless with a 50lb Unix manual Jan 30 05:03:47 we should sync the 2.4 and 2.6 keymaps for Akita / Spitz / Borzoi. 2.4 uses Alt to switch VT's, 2.6 uses AltGr (aka Fn). What do other devices use? Jan 30 05:06:50 I vote for continuing to use FN as anything else will drive me insane ;-) Jan 30 05:07:01 well, the oz.org article says it's Fn, so I'll change the 2.4 keymap Jan 30 05:07:21 once ewi is back that is Jan 30 05:07:34 CoreDump|home: That article was 2.6 based (I wrote it) Jan 30 05:07:46 Soon, we won't care about 2.4 anyway... Jan 30 05:08:04 heh in 3.5.4 we do =) Jan 30 05:08:34 hrw|work: well if altboot is working on c7x0 Id probably just stop using root=/dev/mmcblk0p1 as its a pain to maintain Jan 30 05:09:27 actually noone bothered to port it to c7x0 yet Jan 30 05:09:35 CoreDump|home: but 3.5.4.1 will hopefully follow rapidly ;-) Jan 30 05:09:40 indeed Jan 30 05:09:49 CoreDump|home: well spitz/akita/borzoi 2.6 version should be the same Jan 30 05:09:58 XorA: uh no Jan 30 05:10:09 akita maybe but not the HDD models Jan 30 05:10:42 I'll make the 2.6 version usable after work so we can try Jan 30 05:11:03 CoreDump|home: not sure why the HDD models are different Jan 30 05:11:13 The HDD models don't bother with using the flash Jan 30 05:11:18 RP: I will take care of 2.4/collie (not officially) Jan 30 05:11:19 these are a pain w/ kernel 2.4 Jan 30 05:11:42 RP: nor does my c7x0 at the moment Jan 30 05:12:21 I'm thinking to just scrap 2.4 support for SL-C3xxx. 2.6 should be trivial Jan 30 05:12:45 XorA: check altboots .bb for a hint how bad these machines suck in 2.4 Jan 30 05:12:51 CoreDump|home: Most 2.4 support for cxx00 and c7x0 is likely to get scrapped soon... Jan 30 05:12:58 good Jan 30 05:13:11 the default would be to boot off HDD right? Jan 30 05:13:28 CoreDump|home: Yes, they just boot straight to /dev/hda1 Jan 30 05:13:40 makes things straightforward Jan 30 05:13:59 CoreDump|home: I looked at altoboot.bb got scared, hid under the bed Jan 30 05:14:34 soon means when I rename the machines as per the OE ToDo list btw ;-) Jan 30 05:14:40 XorA: heh, _every_ machine is dead simple to support, just not the SL-C3xxx beasts and their retarded booting process in 2.4 Jan 30 05:14:58 CoreDump|home: especially c3000 Jan 30 05:15:06 * RP used that retarted booting process exactly once Jan 30 05:15:07 CoreDump|home: maybe I confused myself Jan 30 05:15:14 CoreDump|home: It was late at night Jan 30 05:15:16 heh Jan 30 05:15:47 pkg_postinst_spitz() { <- check it out and shiver in pain Jan 30 05:15:50 or I had spent too long staring at m4 tring to make it make sense Jan 30 05:17:30 http://pastebin.ca/39117 Jan 30 05:17:42 hrw|work: are their plans to apply the amd64 fixes to uicmoc or whatever its called? Jan 30 05:17:44 RP: virtual/kernel-hostap - which idea was to fix it correctly? Jan 30 05:18:10 XorA: that + customized scripts for almost every boot option Jan 30 05:18:14 Hm.. is .dev broken at present? I remember seeing some discussion on oe@ about RUNSTRIP.. Jan 30 05:19:27 XorA: I didnt noticed problems with branch on ewi (dual opteron) Jan 30 05:19:49 hrw|work: ewi runs i386 unless koen changed it recently Jan 30 05:19:55 Building busybox dies at do_package Jan 30 05:20:01 NAbyss: yes, .dev is b0rked. set IGNORE_STRIP_ERRORS = "1" in local.conf Jan 30 05:20:08 CoreDump|home: Ah, thankye :) Jan 30 05:20:13 np Jan 30 05:25:59 hmm Jan 30 05:26:06 is sourceforge migrating to svn? Jan 30 05:26:25 shadows: not migrating - allow to migrate Jan 30 05:26:34 oh okay Jan 30 05:26:37 hence the outtage Jan 30 05:26:58 oooh, does that mean we can sync OE to sf.net and let them suck some bandwidth for us Jan 30 05:28:11 XorA: if their svn will work like cvs then it will rather suck our time too much Jan 30 05:28:46 hrw|work: I remeber the days sf.net was usable, back when they had massive VC input Jan 30 05:29:11 sf being unusable is a myth Jan 30 05:29:31 they plan outtages according to the USA convenience, though Jan 30 05:29:52 shadows: recently their pipe to the EU seems to be fed through damp string Jan 30 05:30:00 heh Jan 30 05:30:01 shadows: its not uptime I have problems with Jan 30 05:30:01 yeah. Jan 30 05:30:31 shadows: cvs.sf.net 'usability' is not a myth - too much times I was unable to fetch sources from their cvs Jan 30 05:30:35 there are domestic disputes with backbone internet providers MCI and others Jan 30 05:30:51 hrw|work: anoncvs goes down quite often Jan 30 05:30:56 normal cvs worked fine Jan 30 05:31:17 hrw|work: I'm not sure what the right solution for the hostap problem was. I need to look into it again... Jan 30 05:35:10 shadows: in how much projects you are developer? etc.. OE use anoncvs Jan 30 05:36:08 hrw|work: briefly i worked on a project "gaim-vv" Jan 30 05:36:27 normal CVS *never* went down. anoncvs broke frequently Jan 30 05:36:48 it never affected the developer, only the end user who would want anoncvs code Jan 30 05:36:58 shadows: I mostly use anoncvs on sf.net Jan 30 05:37:12 anyway I avoid cvs as much as I can Jan 30 05:39:16 when i was a Gentoo developer, the mantra was "if you rely on anonCVS, expect things to break" Jan 30 05:39:41 later all Jan 30 05:40:20 * shadows waves Jan 30 05:40:59 it is not reasonable to resolve bugreports against a mystery code base Jan 30 05:41:55 the Gentoo way is that "cvs" builds get special attention, and are not allowed to have a bug report submitted about them Jan 30 05:42:21 instead, there are cvs snapshot tarballs that are kept on gentoo mirror servers Jan 30 05:43:04 ~seen koen Jan 30 05:43:07 those we could sometimes see a bugreport and think, oh okay, i have the codebase this user or developer is submitting the bug against, maybe i will be understanding of the problem and have a possibility to help them Jan 30 05:43:07 koen was last seen on IRC in channel #openslug, 14h 51m 28s ago, saying: ''night all'. Jan 30 06:02:08 The post to git@ about shallow cloning of git tree is interesting Jan 30 06:05:44 <[lala]> hi guys :) Jan 30 06:06:22 hi [lala] Jan 30 06:06:43 <[lala]> RP: do u happen to know if i can use the fb device of borzoi with the graphical mode of links? Jan 30 06:07:20 [lala]: The fb device is a standard framebuffer which can do everything a normal graphical framebuffer can do Jan 30 06:07:49 <[lala]> RP: sounds good ... what about the touchscreen? can i use that with gpm? Jan 30 06:07:50 [lala]: So at a guess, yes even though I have no idea about the gaphical mode of links :) Jan 30 06:08:13 No. gpm would need patching to support tslib Jan 30 06:08:36 There was a patch around which did something like that I think Jan 30 06:09:16 <[lala]> not anymore it seems Jan 30 06:09:25 [lala]: What's your email address? Jan 30 06:09:57 [lala]: It was posted on openzaurus-devel on 2nd December Jan 30 06:10:40 [lala]: yes you can Jan 30 06:10:40 <[lala]> i see Jan 30 06:11:26 [lala]: but chanes are links will render in portrait as thats the native direction of fb I beleive Jan 30 06:11:36 [lala]: Ive run links on fb on c7x0 Jan 30 06:12:41 <[lala]> XorA: with or without the tslib patch? Jan 30 06:13:03 [lala]: without as I didnt have that back then Jan 30 06:13:24 Has anyone been to FOSDEM before ? I am thinking of going this year and i could use some tips regarding hotels and such Jan 30 06:13:25 [lala]: The problem is the framebuffer is rotated as XorA has pointed out :-( Jan 30 06:13:38 <[lala]> XorA: do u remember with which parameter u have started gpm then? Jan 30 06:14:04 [lala]: I just pointed at a non existant serial mouse Jan 30 06:14:33 <[lala]> RP: well i would like to try it even with a rotated fb :) Jan 30 06:14:51 hi all Jan 30 06:14:52 [lala]: someone to develop a .bb for tslib gpm would be cool Jan 30 06:15:16 <[lala]> XorA: absolutely Jan 30 06:15:31 It should just be a question of applying that patch I've mentioned Jan 30 06:15:53 <[lala]> RP: i will try that one and report then if it works Jan 30 06:16:43 <[lala]> on another topic: has anybody here tried to build an image for the xxs1500 (mycable mips box)? Jan 30 06:17:11 Ifaistos: I've not been before. I believe quite a few people are staying at the Novotel nearby. There was some other places mentioned as well... Jan 30 06:18:18 <[lala]> RP: i will try your latest patches for the fb later today. just building a new kernel. Jan 30 06:18:50 [lala]: ok, thanks. I'll be interested to see if it helps Jan 30 06:20:01 RP: Ok. I will look either at Novotel or a hotel near by. Any idea what time fosdem starts ? i am on rather tight schedule and i am not sure i can go on friday, so i might have to arrive on Saturday morning Jan 30 06:20:36 Ifaistos: The fosdem website is the best place to look. I don't know more than that Jan 30 06:21:02 RP: Not much there (yet) regarding the program unfortunately Jan 30 06:22:20 Ifaistos: Maybe 2005 helps: http://www.fosdem.org/2005/index/schedule.html Jan 30 06:24:38 RP: unless /dev/input/mice in kernel 2.6 supports the touchscreen Jan 30 06:24:44 then it would work with gpm Jan 30 06:25:13 shadows: the touchscreen is not a mouse Jan 30 06:25:33 i don't see why it couldn't be emulated as a mouse for that function though Jan 30 06:26:02 shadows: calibration Jan 30 06:26:11 ohh. my oversight. Jan 30 06:27:20 i completed my amd64 host build of openzaurus-unstable in OE environment for spitz target this morning Jan 30 06:27:27 am extracting it to the target hardware now Jan 30 06:28:48 of note are two bugs, one is the gcc-3.3.4 amd64 host bug, and the other is related to sqlite using host precision and assuming 64-bit due to my host, when arm is 32-bit Jan 30 06:29:04 do13_: Thanks ! Jan 30 06:29:05 both bugs have standing bugreports in oe Jan 30 06:29:21 should there be an amd64 host compile metabug? Jan 30 06:29:27 shadows: bug number for amd64 bug? Jan 30 06:29:42 shadows: as I build gpe-image on amd64 without problem Jan 30 06:30:28 Ifaistos: np Jan 30 06:30:35 Are you using JIT on the amd64 (32bit mode)? Jan 30 06:31:30 i meant psyco JIT ;) Jan 30 06:33:31 Ifaistos: no i am not Jan 30 06:40:55 Ifaistos: FOSDEM starts at about 10:00, from what I remember. Jan 30 06:41:43 Ifaistos: fwiw, Team Opie is staying in the Astrid hotel in central Brussels. The campus where FOSDEM actually happens is in a fairly desolate area so I don't recommend trying to find accomodation in the immediate vicinity. Jan 30 06:41:45 zecke: hail Jan 30 06:42:19 RP: could you point me to your patch for bitbake again? Jan 30 06:42:27 pb_: http://www.perrier.eu.org/weblog/2005/12/11#bazsucks2 Jan 30 06:42:32 pb_: sounds trustworthy as well?! Jan 30 06:48:53 hey! so whats the solution to who ever introduced the RUNSTRIP problem? Jan 30 06:48:53 shadows: I was wondering if it made any difference in the overall speed of the bitbake Jan 30 06:49:24 zecke: http://www.rpsys.net/openzaurus/temp/bitbake_filtprov_fix-r0.patch Jan 30 06:49:38 pb_: Thanks! Jan 30 06:50:16 Ifaistos: oh, i know it makes a gigantic increase in speed for the "nicotine" python p2p app Jan 30 06:50:21 i don't know about bitbake though Jan 30 06:50:55 zecke: awesome Jan 30 06:51:41 morning Jan 30 06:51:45 holy crap fonts are small, now i see the sqlite problem simcop was talking about Jan 30 06:51:55 pb_: I think you should get used to monotone ;) Jan 30 06:52:21 hmm Jan 30 06:52:32 so no clear workaround to runstrip? Jan 30 06:53:11 emte: Add IGNORE_STRIP_ERRORS = "1" to local.conf for now Jan 30 06:53:36 who is running into problems with runstrip? Jan 30 06:53:51 k, hopefully thats enough so i can test my fakeroot fix Jan 30 06:54:04 everyone who has a recent pull Jan 30 06:54:06 i.e. what config sets that off Jan 30 06:54:08 oh okay Jan 30 06:54:21 lastnight it showed up in glibc and diet-x11 Jan 30 06:54:22 my pull is from early yesterday Jan 30 06:54:49 zecke: :-} Jan 30 06:54:55 jbowler, has a message on the mailing list as well Jan 30 06:55:08 re Jan 30 06:55:21 RP: I will apply the patch soon Jan 30 06:55:32 zecke: thanks Jan 30 06:56:09 zecke: Did we fix that stamps path reference btw? Jan 30 06:56:21 ~monday Jan 30 06:56:22 nope, ph5's patch is pending :} Jan 30 06:56:25 XorA: http://bugs.treke.net/attachment.cgi?id=451 Jan 30 06:56:25 extra, extra, read all about it, monday is when everything breaks for no apparent reason, creating so many problems that it takes you until friday to get back to normal. at which point one more thing breaks that takes you the whole weekend to fix Jan 30 06:56:48 zecke: ok. I've not looked at the possible fixes for that yet Jan 30 06:56:49 XorA: http://bugs.treke.net/show_bug.cgi?id=637 and also http://bugs.treke.net/show_bug.cgi?id=640 Jan 30 06:57:46 shadows: ah, havent seen that efect on c7x0 kernel Jan 30 06:57:55 shadows, http://www.handhelds.org/hypermail/oe/current/5949.html is jbowler's post Jan 30 06:58:29 oh Jan 30 06:58:36 well what the hell are you doing grepping for "not stripped" Jan 30 06:58:43 you should be using errorlevels Jan 30 06:58:56 shadows: though that might just mean that c7x0 kernel isnt as close to the edge, Ill leave kernel for RP :-) Jan 30 06:59:15 shadows: the grep for "not stripped" is okay, it's just the return value logic that seems to be busted. Jan 30 06:59:28 that's an issue for reenoo, anywya Jan 30 06:59:31 anyway, too Jan 30 07:02:16 The c7x0 kernel is quite a bit smaller that that on spitz due to the lack of ext3 compiled in Jan 30 07:03:20 i thought the c7x0 doesn't have the same limit though Jan 30 07:03:27 the test is only for spitz/borzoi/akita Jan 30 07:03:36 no it isn't Jan 30 07:03:46 oh okay Jan 30 07:03:59 is it the same limit in bytes? Jan 30 07:04:05 yes Jan 30 07:04:58 hmm Jan 30 07:05:29 what do you say about bug 540 Jan 30 07:05:39 what's the best way to fix that in OE Jan 30 07:06:26 specifically the sqlite stuff Jan 30 07:06:39 it's compiling code on the host that outputs the size of a type Jan 30 07:06:48 and uses that as the type for the crosscompiled output Jan 30 07:06:57 which on 32-bit host arch is correct for 32-bit arm Jan 30 07:07:05 annnnd on amd64, not so good Jan 30 07:07:35 http://bugs.treke.net/attachment.cgi?id=451 Jan 30 07:09:43 probably needs to make like autoconf and either use the site file, or use a test that doesn't involve executing code. Jan 30 07:12:25 is there not an ISO C type for pointer size there days? Jan 30 07:13:11 I'm not sure I understand what you mean. Jan 30 07:14:20 actualy why cant that be #define SQLITE_PTR_SZ sizeof(void *) Jan 30 07:14:48 or char * or whatever the real pointer is actually supposed to be Jan 30 07:15:41 in any event, i've filed a metabug for amd64 build host troubles, where differing from ia32 build host Jan 30 07:15:43 * XorA is puzzled by they crazy snippet Jan 30 07:15:54 http://bugs.treke.net/show_bug.cgi?id=641 Jan 30 07:17:21 shadows: Tiny GPE desktop fonts related to sqlite what does that mean Jan 30 07:18:14 XorA: desktop fonts depend on sqlite to get the sizes stored in a db Jan 30 07:18:31 XorA: that #540 was multibug type ;( Jan 30 07:18:35 shadows: erm, where? Jan 30 07:18:40 somehow the size of a type in the code is getting messed up, because it is presuming something for 64-bit arch Jan 30 07:18:41 shadows: no Jan 30 07:18:43 only on gpe-image Jan 30 07:18:46 did i get that wrong? Jan 30 07:18:51 simcop was going on about this Jan 30 07:18:51 the tiny desktop fonts problem was a matchbox-panel problem which is now fixed Jan 30 07:19:04 fixed since yesterday? Jan 30 07:19:06 seems someone has been mixing and matching info Jan 30 07:19:10 aye Jan 30 07:19:19 shadows: fixed in matchbox svn version Jan 30 07:19:25 hmm Jan 30 07:19:29 XorA: so which versions (of which gpe/x11 components) should be used for release? Jan 30 07:19:52 hrw|work: should be following gpe 2.7 Jan 30 07:20:56 morning all Jan 30 07:21:02 hi liam Jan 30 07:21:07 hrw|work: or I misunderstand your question Jan 30 07:21:30 lrg|home: DId you mention that there's a way of hiding controls in alsamixer via the config files? Jan 30 07:21:32 RP: got my confirmation for the embedded track at FOSDEM Jan 30 07:21:43 XorA: could you please comment on bug 641 to clarify? i don't want to make a false statement, i took what simcop said Jan 30 07:21:44 lrg|home: Excellent :) Jan 30 07:22:32 XorA: gpe 2.7 had that problem iirc Jan 30 07:22:52 hi Liam Jan 30 07:23:01 Hi Liam Jan 30 07:23:09 hrw|work: but I think the /etc/X11/Xserver for branch says -dpi 100 which is the hack fix Jan 30 07:23:43 RP: yes, we need to create a configuration file for each machine. This file can specifiy things like default values and hidden mixers. Have a look at /ertc/alsa/cards/ on your PC for info. Jan 30 07:24:08 it's not per machine Jan 30 07:24:14 it's per sound driver/chipset Jan 30 07:24:27 though i guess you could make it work per machine using init scripts Jan 30 07:24:28 lrg|home: My PC doesn't have a /etc/alsa/cards :-/ Jan 30 07:24:35 XorA: too bad that it does not looks like that on my c760 ;( Jan 30 07:24:41 RP: /usr/share/asound... i think Jan 30 07:24:56 or s/asound/alsa/ Jan 30 07:24:59 shadows: sorry, I mean per soc and codec/ machine cofiguration. Jan 30 07:25:13 /usr/share/alsa/cards/ Jan 30 07:25:16 on debian at least Jan 30 07:25:32 hrw|work: I beleive that is what 640x480 z's should use as dpi for 2.7 GPE, you'll have to double check with florian/koen/other GPE gurus Jan 30 07:25:40 ok Jan 30 07:25:43 its /usr/share/alsa/cards (on ubuntu) Jan 30 07:25:56 hrw|work: I have 200 dpi working on my Z but with a change to theme.xml and svn matchbox-panel Jan 30 07:26:09 XorA: ok Jan 30 07:26:23 hrw|work: and also florian needs to add another setting to gpe-conf, which I set by hand Jan 30 07:26:41 ok Jan 30 07:26:52 RP: it might be best to wait a day or two as I'm merging in all the stereo mixers. This will mean less mixers with different names. Jan 30 07:26:54 XorA: where are matchbox themes? Jan 30 07:27:02 XorA: is there a runtime-workaround to the tiny fonts problem Jan 30 07:27:06 hrw|work: /usr/share/themes/*/matchbox Jan 30 07:27:31 thx Jan 30 07:27:52 lrg|home: Ok. In these config files is it the "lock true" paramteres that hide them? Jan 30 07:28:27 RP: let me check, I read something in the alsa wiki... Jan 30 07:28:53 heh.. fontconfig.. I do not like it Jan 30 07:33:21 ~curse fontconfig Jan 30 07:33:23 May the fleas of a thousand camels infest your most sensitive regions, fontconfig ! Jan 30 07:34:03 someone know how to tell fontconfig that 'dejavu sans' is sans? and 'dejavu sans mono' is monospace? other then editing /etc/fonts/* files Jan 30 07:37:48 hi all Jan 30 07:38:07 hi here florian Jan 30 07:38:15 Is there any particular reason why openembedded.org is still not connected? Jan 30 07:38:55 is the DNS its only issue Jan 30 07:39:54 Yes... its available via oe.handhelds.org but all links point to openembedded.org. Jan 30 07:53:50 is the "getting started" documentation on the wiki out of date? Jan 30 07:54:30 xumbi: in which part? Jan 30 07:54:54 I couldn't get monotone to checkout anything from the snapshot Jan 30 07:55:11 which ver of monotone you have? Jan 30 07:55:24 and why <0.24? Jan 30 07:56:11 I think we need a Frequently Encoutnered Problems page on the wiki... Jan 30 07:56:31 we had DeveloperFaq Jan 30 07:56:37 RP: cant find on wiki, i'll try on Mainstone and get back to you. Jan 30 07:57:01 lrg|home: ok, thanks. I've also been looking but I can't find any documentation Jan 30 07:57:20 RP: I should have bookmarked it :( Jan 30 07:58:01 hrw|work: version 0.23 Jan 30 07:59:04 oh Jan 30 07:59:10 hi lrg|home Jan 30 07:59:47 hey florian_kc Jan 30 08:00:02 florian_kc: going to FOSDEM ? Jan 30 08:01:36 lrg|home: I would like... but its very unlikely that i make it this year. My wife plans to finish our fork() around this time :-) Jan 30 08:01:38 is a version higher than 0.23 required to checkout the latest branches? Jan 30 08:02:26 xumbi: yes Jan 30 08:02:40 florian_kc: good luck with your fork(), I know they can consume resources ;) Jan 30 08:02:45 florian_kc: ah, ok thanks Jan 30 08:04:26 lrg|home: maybe i can manage to come for one day, its not that far from me... Jan 30 08:05:36 anyone wants to try mercurial? Jan 30 08:06:35 * florian_kc would like to do this, but i don't have time to do it :-/ Jan 30 08:06:46 zecke: There's an interesting post on the git mailing list about developing support for shallow cloning Jan 30 08:06:59 RP: as in bk lclone? Jan 30 08:07:11 zecke: I never tried that Jan 30 08:07:28 RP: let me get the translation of shallow first Jan 30 08:07:36 zecke: it cuts off history from a certain point in the past Jan 30 08:07:46 * Crofton wants to get trac running against the oe svn repo, but not this week :( Jan 30 08:07:57 RP: ah right Jan 30 08:08:13 Crofton: that would be pretty useful... Jan 30 08:08:31 * hrw|work wants good band for wedding... Jan 30 08:08:53 yeah, I have it running on the real svn repo I have, but I need to figure out how to make it work against two repos, without screwing up the one we (as in my research group) use Jan 30 08:09:43 this week I have to go see US immigration and explain I want to become a US citizen to promote regime change from within :) Jan 30 08:10:13 heh Jan 30 08:13:35 RP: I see git,monotone26,svn and hg as true alternatives to monotone025 Jan 30 08:14:08 darcs is the haskell based one? Jan 30 08:14:13 Crofton: yes Jan 30 08:14:29 zecke: Yes. I don't know much about hg... Jan 30 08:14:43 Crofton: well it produces load on the server (darcsweb), it has security issue, it hogs memory and cpu Jan 30 08:15:02 zecke: When I next get some spare time, I'll try and write a using git guide... Jan 30 08:15:04 Crofton: and can not pull patches worth of three days of development on OS X Jan 30 08:15:04 heh Jan 30 08:15:31 zecke: It might be worth putting up a summary of this and state why we're rulling certain ones out Jan 30 08:15:43 Crofton: darcs is a collection of patches. And when applying patches you can leave the working directory Jan 30 08:15:44 (on the wiki) Jan 30 08:15:51 the haskell compiler is an issue also? Jan 30 08:16:00 RP: yes I plan to do that Jan 30 08:16:04 RP: thats a great idea Jan 30 08:16:07 Crofton: yes as bootstraping is 'hard' Jan 30 08:16:13 heh Jan 30 08:16:18 Crofton: I find it challenging though Jan 30 08:16:24 We don't need OE to be any harder to get started with Jan 30 08:16:41 One mans challenge is another mans roadblock Jan 30 08:17:04 The point of OE is to make embedded development easier :) Jan 30 08:17:07 Crofton: yeah, this is why svn+svk is interesting Jan 30 08:18:32 RP: but git is damn fast Jan 30 08:19:11 All alternatives except for svn+svk and monotone "frighten" me Jan 30 08:19:32 I like svn and kergoth says nice things about svk Jan 30 08:19:48 git is frightening. There is power there though Jan 30 08:19:53 I have dealt with monotone long enough to know how to use it (sort off) Jan 30 08:20:05 Heh, I don't need power :) Jan 30 08:20:33 I need to get stuff from the OE tree, merge it into a local tree and share that with a few people Jan 30 08:20:46 the downstream users only care about ease of use Jan 30 08:21:01 and speed Jan 30 08:21:07 and processing power requirements Jan 30 08:22:00 sending around oe.db files is kind of self defeating :) Jan 30 08:25:08 Crofton: svk can really help there (merging) Jan 30 08:25:39 for me it can be any distributed SCM available in debian (or monotone 0.26 which will take time to get into probably) Jan 30 08:26:09 Thats what Kergoth said. I really need to look at this, currently merging branches looks pretty manual. Jan 30 08:26:33 hrw|work, what is in Debian? Jan 30 08:27:44 Crofton: I meant: it can be git/monotone/svk/darcs/mercurial - but I will probably look at git and probably svk Jan 30 08:27:52 lack of time kills Jan 30 08:28:06 :) Jan 30 08:28:18 time... I do not know how to get all these things sorted :} Jan 30 08:28:48 and I wake at 04:50 ;( Jan 30 08:40:35 * RP has a time problem :-/ Jan 30 08:43:58 RP: Sounds familiar :-( Jan 30 08:45:26 shadows: could you please add the two Opie/uicmoc bugs as well (they are amd64 related as well) Jan 30 09:04:45 http://www.alwins-world.de:1180/hg/cgi-bin/hgwebdir.cgi Jan 30 09:04:49 for a mercurial gui Jan 30 09:05:39 zecke: hmm... pretty fast that thingy. Jan 30 09:06:05 -points for no tree view Jan 30 09:06:54 RP: I think it has one... Jan 30 09:06:59 before I fill out a long request is anyone havig a problem with libx11 Jan 30 09:07:07 I will post a darcsweb link as well Jan 30 09:07:18 zecke: you used tailor to create mercurial tree? Jan 30 09:07:19 but please no more than two people access it... Jan 30 09:07:26 hrw|work: yes Jan 30 09:07:41 for the final transformation I will make tailor smarter Jan 30 09:07:53 http://www.alwins-world.de:1180/darcs/cgi-bin/darcsweb.cgi Jan 30 09:08:09 zecke: yep.. current tailor makes revs messed Jan 30 09:08:41 hrw|work: it tries to lineraize tree's Jan 30 09:08:55 hrw|work: which create a file ping pong (added, deleted, added, deleted) Jan 30 09:09:07 I see Jan 30 09:09:21 hrw|work: but before you mentioned it, I did not recognize it. Jan 30 09:09:21 ok. time to go home and get some sleep probably Jan 30 09:09:35 hrw|work: sleep well! stay away from computers Jan 30 09:09:46 hrw|work: and get some flowers for your sweat-heart Jan 30 09:10:06 zecke: did I tell that we will marry 8 July? Jan 30 09:10:20 hrw|work: yes Jan 30 09:10:32 hrw|work: congratulations... Jan 30 09:11:36 hrw|work: Someone pointed me to this... afer it was too late: http://www.ms1.de/wordpress/wp-content/20051114aptgetwifeg.jpg Jan 30 09:14:02 florian_kc: sweet Jan 30 09:14:19 :-) Jan 30 09:14:28 http://maps.google.com/maps?f=q&hl=en&q=szczecin,+poland&ll=53.420735,14.491568&spn=0.002436,0.006432&t=h Jan 30 09:14:37 hrw|work: don't take it too serious :-) Jan 30 09:14:48 florian_kc: yep Jan 30 09:15:12 florian_kc: anyway its good that it is apt-get not emerge/bitbake ;) Jan 30 09:15:41 hrw|work: building a wife from source, hmm, interesting idea Jan 30 09:16:02 USE=-creditcard -bankaccount -nag Jan 30 09:16:13 hrw|work: heh Jan 30 09:17:03 XorA: don't forget to build in the audio mute option. very handy ! Jan 30 09:17:13 cu Jan 30 09:17:22 cu Marcin Jan 30 09:17:26 would this be the correct place to report a .bb file in oz354 having an incorrect url? or should i file a bug report myself? Jan 30 09:17:46 lrg|home: USE=-alsa -oss Jan 30 09:17:48 zecke: instead of flowers I will buy 6th harry potter book Jan 30 09:18:03 simcop2387-lab: bugzilla and add dependency to #350 Jan 30 09:18:24 thanks Jan 30 09:20:05 * XorA is reminded to look at poketop again Jan 30 09:48:11 re: alsa Jan 30 09:48:17 how do you unmute? Jan 30 09:48:40 raduga: in alsamixer use m Jan 30 09:49:17 on which channel? Jan 30 09:49:57 perhaps there's some other problem on my end; I've tried toggling 'm' both ways on all the channels, and still no sound. Jan 30 09:50:55 raduga: make sure your output and pcm volumes are set high enough. Jan 30 09:51:13 yep. Jan 30 09:51:15 raduga: what handheld do you have ? Jan 30 09:51:18 at max :) Jan 30 09:51:20 akita Jan 30 09:51:51 raduga: you will need to unmute the left and right mixers. Jan 30 10:04:42 Ciao all Jan 30 10:04:50 hi Pigi Jan 30 10:05:06 hi florian_kc ! Jan 30 10:06:19 pb_ you around ? Jan 30 10:10:37 Hi Pigi! I never got your ipkg test version you sent about a week ago...If you still want me to test, can you send again to drw@draknor.net? Jan 30 10:19:03 drw, checkout the latest ipkg, and you have it Jan 30 10:19:20 Pigi: ok, will do Jan 30 10:22:00 MX for hh.org was down when I sent the mail, and then I haven't got the time to resend. sorry Jan 30 10:22:59 np Jan 30 10:26:11 pb_: Help :) netstat -a what does the Recv-Q column do? Jan 30 10:27:27 ah the Linux manpage explains that :} Jan 30 10:33:59 :-} Jan 30 10:34:07 Pigi: yo Jan 30 10:41:40 pb_ I would like to discuss a bit your bug report, and patch, if you have time.... Jan 30 10:43:10 for #1488 I mean Jan 30 10:46:24 Pigi: sure Jan 30 10:47:11 I was wondering if it's correct for ipkg to work the way your patch indicate. Jan 30 10:47:39 Basically IMHO isn't a bug from ipkg the slowness of familiar on first boot Jan 30 10:49:45 and adding the code you provided seems to me a bit out of ipkg scope. Jan 30 10:51:36 it's indeed not a bug in ipkg, but I think it would be a useful enhancement. Jan 30 10:51:45 mickeyl: wb Jan 30 10:51:49 hi guys Jan 30 10:51:54 hail mickeyl Jan 30 10:51:58 hi mickeyl ! Jan 30 10:52:01 Is there another way to collect redundant postinst scripts and only run one of them at the end of the package configuration process? I think this can only be done in ipkg Jan 30 10:52:09 (hi everyone) Jan 30 10:52:58 hi ph5 Jan 30 10:53:25 what does not convice me ( hope this is right ) is that this could be a useful enhancement for familiar boot, not for current use IIUC Jan 30 10:53:45 I'm not sure I understand what you mean by "for current use" Jan 30 10:53:47 IMO this is an optimization that should indeed be a part of ipkg. It's not only interesting at boot time, but also when you e.g. install more than one additional module or font packages or libraries lateron. Jan 30 10:54:25 pb_ I mean when installing a small bunch of packages or either when upgrading. Jan 30 10:55:00 I have never had so "hard" slowness when upgrading, while I can understand what's happening on first boot. Jan 30 10:55:39 I must say I'm not against it, just wondering about it. Jan 30 10:56:03 The same issue would exist (albeit in a smaller way) whenever you install or upgrade two or more packages that have a common postinst. Jan 30 10:56:21 (for example, upgrading the kernel and modules, or multiple libraries) Jan 30 10:56:45 pb_ sure. Jan 30 10:57:13 but in a really light way.... I will study it better, and do som more thought on it. Jan 30 10:57:18 s/som/some Jan 30 10:57:45 I'd like it even better if the rootfs images would come preconfigured for most of the packages (kernel-module-* for example). Isn't this possible? Jan 30 10:58:13 pH5: some things have to be run natively Jan 30 10:58:24 pH5: integrating scratchbox would be nice for those Jan 30 10:58:26 That was my idea too, pH5 Jan 30 10:58:27 pH5: anything is possible, yeah, but it would not be trivial Jan 30 10:58:35 pb_ right..... Jan 30 10:58:57 but more in the ipkg and familiar scopes. Jan 30 11:01:33 yeah Jan 30 11:01:56 well, all options are open for consideration, but this is a problem that needs to be addressed for familiar 0.8.3. The current situation is pretty unacceptable. Jan 30 11:03:04 oh, this regarding the shitload of executions of depmod and such? Jan 30 11:03:09 yeah Jan 30 11:03:38 currently, depmod gets run twice (!) for every module package that's configured, and there are many tens of those in the familiar images. Jan 30 11:04:41 any ideas why i always get RUNSTRIP errors? Jan 30 11:04:54 I could patch ipkg for 0.8.3 this way, but would like to remove it after, if possible to remain inside ipkg scope... I'm not sure there is such option in apt-get or deb or whatever..... Jan 30 11:05:09 law_: I think the patch that reenoo checked in yesterday might have been a bit broken. Jan 30 11:06:29 pb_: so i (we) have to wait for reenoo? Jan 30 11:06:44 and/or fix it yourself, yeah Jan 30 11:13:14 bbl Jan 30 11:13:16 Ciao Jan 30 11:55:06 03koen 07org.oe.oz354fam083 * r6580dbe8... 10/packages/figment/ (files/dotdesktop-name-comment.patch figment_0.3.5.bb): figment: add .desktop patch from .dev Jan 30 11:55:10 03hrw 07org.oe.oz354fam083 * r8a055baa... 10/packages/initscripts/ (7 files in 7 dirs): Jan 30 11:55:10 initscripts: updates for Zaurus machines keymaps Jan 30 11:55:10 - added collie 2.4 one Jan 30 11:55:10 - updated cxx00 2.4 ones Jan 30 11:55:10 - fixed tosa 2.4 one Jan 30 11:55:11 - added 'keymap' script which can read 'extensions' keymaps too Jan 30 11:55:13 all taken from few revs in .dev Jan 30 11:55:15 03jbowler 07org.oe.dev * r180653f2... 10/packages/busybox/busybox-1.01/thumb-bsdlabel.patch: busybox: add missing patch in 1.01 Jan 30 11:55:20 03rpurdie 07org.oe.dev * r3e9163d9... 10/packages/linux/linux-openzaurus_2.6.15.bb: (log message trimmed) Jan 30 11:55:23 linux-oz-2.6: 2.6.15 updates: Jan 30 11:55:25 * LEDs updates based on comments from LKML and other sources: Jan 30 11:55:27 * Functions now all follow the name led_* Jan 30 11:55:29 * brightness is now an enum Jan 30 11:55:31 * the timer frequency is in Hz Jan 30 11:55:34 * the time has a duty parameter Jan 30 11:55:36 03jbowler 07org.oe.dev * r05a4c302... 10/conf/distro/slugos.conf: slugos: temporary workround for the RUNSTRIP problem in conf Jan 30 11:55:40 03jbowler 07org.oe.dev * rc85bf07d... 10/ (15 files in 4 dirs): Jan 30 11:55:42 linux: fix missing HOST_CC_ARCH in module compilation in kernel/module all Jan 30 11:55:44 - define TARGET_KERNEL_CC_ARCH to define arch specific compiler options Jan 30 11:55:46 in machine.conf. The default for the kernel is "" (empty), although Jan 30 11:55:48 it should probably be TARGET_CC_ARCH in the future. The default for Jan 30 11:55:52 modules is TARGET_CC_ARCH Jan 30 11:55:54 03justinp 07org.oe.dev * r9e9c1057... 10/packages/ (68 files in 2 dirs): Jan 30 11:55:56 efl, e17: update CVS versions again, fix cpp issues Jan 30 11:55:58 - evas is left at 20060113 because it changed that day and breaks ecore Jan 30 11:56:00 - cpp and gcc paths are hard-coded to /usr/bin/ to fix some issues with OE's binaries in use with edje Jan 30 11:56:02 03jbowler 07org.oe.dev * r45275064... 10/packages/devlabel/ (devlabel_0.45.01.bb devlabel_0.48.01.bb): Jan 30 11:56:09 devlabel: release 0.48.01 Jan 30 11:56:10 - this does actually seem to work (release 0.45.01 simply exited from Jan 30 11:56:12 bash without any output), but it depends on things like the 'raw' Jan 30 11:56:14 command existing and working, so it isn't very useful Jan 30 11:56:16 03xora 07org.oe.dev * r776c1d67... 10/packages/abiword/abiword_2.4.2.bb: abiword_2.4.2.bb : new version Jan 30 11:56:18 03florian 07org.oe.dev * r86dda2fd... 10/classes/gpe.bbclass: gpe.bbclass: Set default maintainer to GPE ml instead of OE. Jan 30 12:01:12 whats a is the bitbake manual the only good place to find information on making .bb files? Jan 30 12:04:31 anyone successful getting raw data from a gps device? Jan 30 12:04:34 specifically i'm wondering what all the "functions" are that i can use Jan 30 12:04:46 bitbake manual, FOSDEM 2005 presentation + thousands of examples, especially look at the classes Jan 30 12:05:09 I have a usb device connected to my akita, but I think the serial port options are wrong and that's why I get nothing Jan 30 12:05:15 bitbake.conf + base class are a good read Jan 30 12:05:25 the stty command from busybox is not so flexible as the real one Jan 30 12:07:09 oh dear, i keep getting errors due to "autotools_do_configure" being replaced by "meta-gpe.bb_do_configure" in run.do_configure. Jan 30 12:07:18 does anybody know where this: http://pastebin.ca/39142 could have happened? Jan 30 12:07:28 mickeyl: thanks i'll check them out Jan 30 12:14:40 hi cedric! Jan 30 12:46:23 mickeyl, please message when you have a few minutes Jan 30 12:46:30 it's about a feed problem Jan 30 12:48:47 why is the first monotone pull is so slow? Jan 30 12:49:22 hi RP Jan 30 12:49:32 only the first? Jan 30 12:49:51 hi pb Jan 30 12:50:18 don't know, it's not finished yet... Jan 30 12:53:21 hi woglinde Jan 30 12:55:27 Im going to do some testing with the kernel 2.6 for the poodle, anything I should know? I haven't touch OE since Feb 2005... Jan 30 12:55:29 cedric: it downloads it then goes over each revision to build the whole thing IIRC Jan 30 12:56:30 build what? the local repository? Jan 30 12:56:35 yes Jan 30 12:56:54 at least that was my understanding, however flawed it may be Jan 30 12:57:21 very slow in comparison of the previous SCM... Jan 30 12:58:56 yea not sure why they picked it but they must have a reason Jan 30 12:59:05 its alot faster later Jan 30 12:59:29 and i like the idea of a local repsitory, makes rechecking it out if you need to alot simpler and faster Jan 30 13:00:16 the way it works i think is, the revisions are JUST the changes, and it ends up applying like 2700 patches or something around there Jan 30 13:01:35 yep normal stuff, I'm used to subversion/cvs Jan 30 13:02:00 anyway... have to wait! Jan 30 13:04:23 03rw 07org.oe.dev * r909b438e... 10/classes/package.bbclass: package.bbclass: fix stripping logic. work around strip returning 0 even on failure. Jan 30 13:05:02 cedric__: You could also download a snapshot from ewi and then just pull the latest changes into that Jan 30 13:05:04 ~ewi Jan 30 13:05:06 from memory, ewi is ewi546.ewi.utwente.nl - main backup server for OE monotone. Hosts database snapshot: http://ewi546.ewi.utwente.nl/OE/OE.db.bz2 Jan 30 13:09:25 how u doing RP? Jan 30 13:10:19 ok so I do a ctrl+c without any issue? Jan 30 13:12:59 should be able to Jan 30 13:45:48 cedric__: Yes, I think so - You're probably going to replace the database anyway Jan 30 13:48:10 03jbowler 07org.oe.dev * rdd25f6e4... 10/classes/module-base.bbclass: module-base: unbork module builds in bbclass Jan 30 13:50:53 Ouch. montone just hit a persmissions error whilst updating. The result - it loses track of all the files it was working on :-/ Jan 30 13:51:27 kernel/module compilation is back the way it was before renoo's change, but TARGET_CC_KERNEL_ARCH can be set to get the required arch values Jan 30 13:53:42 jbowler-zzz: One bug down then... Jan 30 13:55:14 jbowler-zzz: very good Jan 30 14:02:11 I should have moved the TARGET_CC_KERNEL_ARCH stuff into .bbclass months ago, but I didn't have any way to test it... Jan 30 14:08:55 RP, pb__: any feelings against having a runstrip() shell function from package.bbclass? I.e. to avoid having to put all the code into python strings? Jan 30 14:09:43 sounds good to me Jan 30 14:12:28 everything in the stanza protected by INHIBIT_PACKAGE_STRIP, so that stripfunc is just "runstrip %s" Jan 30 14:14:59 jbowler-away: fine with me Jan 30 14:17:21 jbowler-away: what happens if you call package install more than once? Jan 30 14:17:30 jbowler-away: is it true that runstrip will fail? Jan 30 14:17:43 shouldn't do Jan 30 14:31:30 RP: should the akita usb be enough to power small devices as a low powered gps? Jan 30 14:31:48 guillermo_: I think the limit is about 100mA Jan 30 14:32:17 you can read the limit in the /proc i think... Jan 30 14:33:22 i have a old ethernet usb, that does not work (because of power), the new small one is perfect Jan 30 14:36:23 any rumors about mograting opie to qt4? Jan 30 14:36:48 greentux: I suspect there's no manpower to do ti Jan 30 14:36:58 RP right Jan 30 14:37:32 zecke told me, when it will be a complete rewrite Jan 30 14:37:59 oha, thats a man-year Jan 30 14:39:05 greentux: you see the problem :-/ Jan 30 14:39:26 greentux: yeah, it is a rewrite :} Jan 30 14:39:50 greentux: the only thing we might benefit from simply porting the code to Qt4 is BiDi, RTL and other scripting Jan 30 14:40:02 RP: but its good to know the pending work Jan 30 14:40:14 Any objections to me committing this irda-utils patch? http://bugs.openembedded.org/show_bug.cgi?id=381 Jan 30 14:42:10 today lala tried links2 browser... textmode was perfekt, but graphical wants gpm... and there is no touchpad driver i think Jan 30 14:43:01 greentux: I did send a patch which might help (i remembered seeing it on oz-devel) Jan 30 14:43:25 RP: aaah lala talked about? Jan 30 14:43:59 greentux: yes, it was mentioned Jan 30 14:44:30 RP: fine, we share same room at work... Jan 30 14:45:31 Seeed :) Jan 30 14:48:10 anyone tried usb keyboard or mouse in akita/spitz? Jan 30 14:48:31 greentux: Any news on the whitescreening? Jan 30 14:49:48 RP: we had a 3h meeting afternoon. So i hope the kernel has built and lala will boot it in the morning Jan 30 14:50:03 greentux: ok Jan 30 14:50:04 guillermo_: tried usb keyboad an borzoi, but no keymap Jan 30 14:50:14 too bad Jan 30 14:51:53 hard to find the keys... Jan 30 14:54:21 greentux: any help about where to find in proc the maximum mA available? Jan 30 14:54:43 mom Jan 30 14:56:02 lsusb -v ? Jan 30 14:56:44 no lsusb Jan 30 14:57:11 will "cat /proc/bus/usb/devices" do it? Jan 30 14:58:08 my notebook shows there maxpwr = 0mA ... Jan 30 14:58:29 because it shows MxPwr=100mA, but I don't know if it's the maximum provided or the required by the device Jan 30 14:59:16 you see "product: " Jan 30 14:59:35 after the controller that is max provided, t device it should be required Jan 30 15:02:37 ok, this is weird, if I do "cat /dev/ttyUSB0" I get data, but after 20 seconds, and then it freezes, then 10 more seconds and I get info again Jan 30 15:03:13 could this be a usb-serial problem, cpu problem, or just a gps problem? Jan 30 15:03:13 re Jan 30 15:06:03 8-o Jan 30 15:07:28 I get a screen full of data every 36 seconds, weird! maybe the serial port misconfigured? Jan 30 15:09:19 maybe... but i think it can be a power save mode of the gps... Jan 30 15:09:27 ok... sleeping Jan 30 15:09:33 guillermo_: Perhaps the gps only sends data every 36 seconds? Jan 30 15:09:36 'night greentux Jan 30 15:09:44 night Jan 30 15:09:52 RP: in my desktop I get non stop data Jan 30 15:10:40 as greentux said, perhaps its power saving... Jan 30 15:10:46 the gps shows a blinking light that means normal operation, but I don't kno if that means something Jan 30 15:18:23 I think it's not the gps, because it's like if I get all the info that should be showed by huge chunks, instead of line by line Jan 30 15:18:32 but I'm not missing anything, I think Jan 30 15:37:40 03mreimer 07org.oe.dev * ra14d4eb7... 10/packages/irda-utils/ (files/init irda-utils_0.9.16.bb): Jan 30 15:37:41 irda-utils: instead of assuming that /dev/ttyS1 is the IrDA port, get the Jan 30 15:37:41 port and attendant arguments from /etc/sysconfig/irda. Jan 30 15:59:17 hey Jan 30 16:09:16 um guys? I've pushed a changeset to ewi before it went down this morning. Well, the commit didn't make it into ewi but my database still thinks it has been pushed. How do I fix that? Jan 30 16:09:36 I'd rather _not_ mt update and lose all changes made Jan 30 16:10:06 CoreDump|home: ask in #monotone on the other net Jan 30 16:10:33 =\ Jan 30 16:10:42 * CoreDump|home has an idea Jan 30 16:10:56 CoreDump|home: never had that situation Jan 30 16:11:06 CoreDump|home: how can it think it pushed that change? Jan 30 16:11:50 well, mt diff doesn't show any changes but against current ewi there should be _tons_ of them Jan 30 16:12:10 CoreDump|home: ah well diff is just the working directory Jan 30 16:12:14 CoreDump|home: and you did mnt ci? Jan 30 16:12:19 ci? Jan 30 16:12:26 CoreDump|home: commit Jan 30 16:12:32 yes Jan 30 16:12:38 commit + push Jan 30 16:12:50 after a commit monotone diff will show no changes Jan 30 16:12:57 what happens if you try to push again? Jan 30 16:13:08 I'll try pushing again =) Jan 30 16:13:50 looking good Jan 30 16:13:54 ;) Jan 30 16:14:02 =D thanks zecke Jan 30 16:14:04 ci is a disconnected operation... :) Jan 30 16:14:17 I know, but I _did_ push the change Jan 30 16:14:27 right before ewi went down Jan 30 16:14:47 CoreDump|home: as long as the certs are not in the other db they are not pushed Jan 30 16:15:01 CoreDump|home: and as it uses sqlite with transactions everything should be fine Jan 30 16:15:08 * CoreDump|home dunnos Jan 30 16:15:37 could you do me a favor and mt pull / update? Jan 30 16:16:12 CoreDump|home: oh this will take some hours Jan 30 16:16:23 CoreDump|home: if you are uncertain of your changes use viewmtn Jan 30 16:16:40 viemtn lags behind Jan 30 16:16:51 zecke: if he was unlucky and pushed immediately before ewi crashed, there certainly is a chance that his changes would be lost. Jan 30 16:16:54 anyways, found a seconds buildroot to pull =) Jan 30 16:17:28 2 revs in, looking good Jan 30 16:17:29 pb__: ewi didn't crash, just the monotone server ;) Jan 30 16:17:42 heh phear my push of death Jan 30 16:17:47 pb__: but should ext3 cope with that... Jan 30 16:18:03 zecke: oh, right. in that case, yeah, you're probably right. Jan 30 16:18:27 me pulling on a celeron sucks Jan 30 16:18:29 zecke: ext3's journalling will ensure that the filesystem metadata is left in a consistent state, but it makes no promises about your data. Jan 30 16:18:42 pb__: ah right Jan 30 16:19:02 then store the data as filesystem metadata Jan 30 16:19:10 one would think Jan 30 16:19:54 the kernel doesn't really have the context it would need to ensure that your data was always left intact after a crash. It doesn't have visibility of anything at a higher level than a write() operation: there is no concept of transactions at the filesystem interface. Jan 30 16:19:57 zecke: that would be of no use since the FS uses caching. Can't write something back which was in the cache Jan 30 16:20:23 pb__: hehe doesn't reiser4 introduce transactions? Jan 30 16:20:36 oh, does it? crazy reiser. Jan 30 16:20:38 ~lart reiser Jan 30 16:20:39 * ibot drops a truckload of VAXen on reiser Jan 30 16:21:03 filesystem transactions would certainly be cute, but, well, I fear reiser too much. Jan 30 16:21:09 pb__: and superior filesystems by superior companies have it as well (like nTFS) Jan 30 16:21:20 yeah, indeed Jan 30 16:21:24 http://www.namesys.com/txn-doc.html Jan 30 16:21:46 it's certainly not a (conceptually) insoluble problem, just one that isn't yet solved by the sqlite+ext3 combination Jan 30 16:22:25 "big" databases like oracle and db2 avoid the issue by doing their own I/O on raw devices and implementing transactions internally. Jan 30 16:22:30 * CoreDump|home lost way to much data due to ReiserFS to _ever_ look at it again Jan 30 16:22:47 pb__: now I'm even more wondering why my last fs crash is so long ago... Jan 30 16:23:45 pb__: and superior editors bootstrap themselves in lisp to emulate a better OS (like GNU/Hurd) Jan 30 16:23:50 * zecke should move to bed Jan 30 16:24:02 :-} Jan 30 16:24:32 :-) Jan 30 16:25:16 damn, looks like my push doesn't show up on ewi Jan 30 16:25:23 03coredump 07org.oe.dev * rab7da42d... 10/packages/altboot/ (4 files in 4 dirs): Jan 30 16:25:24 altboot: Jan 30 16:25:24 - First changes to support Kernel 2.6 Jan 30 16:25:24 - As of this push altboot should be considered *unstable* until work on 2.6 has finished! Jan 30 16:25:24 - Add IGNORE_STRIP_ERRORS = "1" to allow for snapshots Jan 30 16:25:25 CoreDump|home: wait a second Jan 30 16:25:27 03coredump 07org.oe.dev * rcd257fc3... 10/packages/altboot/ (10 files in 6 dirs): Jan 30 16:25:27 altboot: More 2.6 changes Jan 30 16:25:28 ok nm Jan 30 16:25:28 - Use altboot-KERNEL_VERSION.cfg instead of altboot.cfg Jan 30 16:25:31 CoreDump|home: well Jan 30 16:25:32 03mreimer 07org.oe.dev * r0bbfeca4... 10/packages/irda-utils/irda-utils_0.9.16.bb: irda-utils: use the correct port for IrDA on hx4700. Jan 30 16:25:49 CoreDump|home: looks like your push was not accepted by ewi' AI. Jan 30 16:25:55 =) Jan 30 16:32:41 RP: damn, your 2.6 is so nice, I don't wanna go back to the 2.4 crap Jan 30 16:35:49 CoreDump|home: Don't then ;-) Jan 30 16:36:21 Have to =( At least for a few days Jan 30 16:39:17 I just refuse to touch 2.4 now. I've probably earned that right :) Jan 30 16:40:11 indeed Jan 30 16:40:24 suspend / resume is a breeze with 2.6 Jan 30 16:40:35 and just plain annoying in 2.4 Jan 30 16:40:48 You have no idea the amount of pain that was involved in that... Jan 30 16:40:50 the whole system "feels" faster Jan 30 16:41:11 RP: More fans :) Jan 30 16:41:14 RP: actually yes =) *thinks in hatred about his notebook* Jan 30 16:41:20 The zaurus pm code took about a year to develop :-/ Jan 30 16:41:25 ouch Jan 30 16:41:53 Cwiiis: Anyone who's used 2.6 becomes a fan ;-) Jan 30 16:42:19 damn right Jan 30 16:42:46 Anyhow, I think I'll try and have an "early" night - 'night all Jan 30 16:42:51 tomorrow I'll see how to add Alsa configs Jan 30 16:43:00 n8 RP Jan 30 16:43:05 RP: cya! Jan 30 16:43:15 Night RP Jan 30 16:49:59 03coredump 07org.oe.dev * r49c83710... 10/packages/keymap-locales/ (6 files in 4 dirs): Jan 30 16:49:59 keymap-locales: Jan 30 16:49:59 - Add kernel 2.6 DE keymap for Akita Jan 30 16:49:59 - Add kernel 2.4 and 2.6 DE keymap for Borzoi and Spitz Jan 30 16:50:15 umlauts ftw Jan 30 16:50:23 ~ewi Jan 30 16:50:25 from memory, ewi is ewi546.ewi.utwente.nl - main backup server for OE monotone. Hosts database snapshot: http://ewi546.ewi.utwente.nl/OE/OE.db.bz2 Jan 30 16:50:57 good nite Jan 30 16:51:13 zecke: night night Jan 30 16:57:16 hmm, should update bash in oe. 3.1 is available, and there appear to be a number of bugfix patches for it on the gnu site that should probably be reviewed Jan 30 16:57:45 good plan Jan 30 16:57:58 man cannot live on busybox sh alone, after all Jan 30 16:58:53 well, some men can. i still almost never use bashisms. code susv3 shell out of habit Jan 30 16:58:56 heh Jan 30 16:59:08 heh Jan 30 16:59:33 although someo f the fancy interactive stuff is nice, like ctrl+R Jan 30 17:00:07 yeah, I couldn't live without ctrl-r Jan 30 17:00:26 ctrl-r stands for ROCKS :) Jan 30 17:00:41 hehe Jan 30 17:06:28 yeah finally ... I may go to bed Jan 30 17:48:00 oh Jan 30 17:48:07 i was going to ask zecke for the bug #'s Jan 30 17:58:53 * CoreDump|home yawns Jan 30 18:50:34 ok, i've got a .bb file i'm making that depends on fltk, and uses auto tools, how do i tell configure where to find fltk? Jan 30 18:57:01 simcop2387: depends. run ./configure --help in its source tree and look for an option related to fltk Jan 30 18:57:40 kergoth: theres an option, i just don't know how to get the directory i need to point it at Jan 30 18:58:22 --with-ftlk-libs=${STAGING_LIBDIR}, --with-fltk-includes=${STAGING_INCDIR}, or --with-fltk=${STAGING_LIBDIR}/.., or whatever Jan 30 18:58:53 ah i see Jan 30 18:59:08 so bitbake takes care of getting those in the right places then? cool Jan 30 18:59:40 the fltk package has a do_stage that populates staging with its stuff Jan 30 18:59:45 k Jan 30 19:00:23 i'm still quite new to this, but once it works, we'll have a bitbake file for the Palm pilot emulator from palm Jan 30 19:01:28 simcop2387: the options get stuffed into there which tell it to look at the staging area, where presumably the library or headers are already installed. i haven't written any bb's, but i did take a look at the do_compile output when hunting a gcc bug. i'm not sure about what exactly you'd want to put there for the value. Jan 30 19:07:05 hmm Jan 30 19:09:26 well well well, what do we have here, configure is trying to use my host ld to test for things? sounds like i need to make a patch Jan 30 19:10:11 or i'm doing it wrong, but i doubt that, right now i'm just inheriting from auto tools Jan 30 19:13:03 ah no i'm reading the wrong file Jan 30 19:13:45 i've got undefined references in libfltk.so, looks like the staging is missing something? Jan 30 19:14:48 against xlib? thats not right Jan 30 19:17:10 ah thats what i need i think Jan 30 19:21:31 hah hah ha Jan 30 19:21:52 nope Jan 30 19:22:06 simcop learns the way of the embed-di Jan 30 19:22:14 use the source, j00! Jan 30 19:22:15 embed-di? Jan 30 19:22:30 not sure whats going on exactly, configure keeps telling me things like this "/tmp/OE/build//staging/arm-linux/lib/libfltk.so: undefined reference to `XOpenIM' Jan 30 19:22:30 " Jan 30 19:22:37 err Jan 30 19:22:41 googling Jan 30 19:22:53 xlib Jan 30 19:23:05 make sure that xlib is linked in Jan 30 19:23:10 says its part of xlib, so i tell configure where xlib and everything is --x-includes=... etc, same thing Jan 30 19:26:45 xlib would be special though, i think it is not in /usr/lib/ Jan 30 19:26:55 it would be something special for X11 no? Jan 30 19:29:41 not sure Jan 30 19:29:57 i used $STAGING_LIBDIR and $STAGING_INCDIR Jan 30 19:30:10 the library appears to be there Jan 30 19:37:10 is there a bug report for the strip errors? Jan 30 19:38:31 strip errors? Jan 30 19:38:49 yes in current OE Jan 30 19:39:07 there's trouble with changes of strip's text output Jan 30 19:39:21 a check used to be against output string, now that is not correct Jan 30 19:40:16 dunno, think it might affect my problem? (i'm still looking into what could be causing it and what i need to do to fix it) Jan 30 19:43:20 i doubt it Jan 30 19:43:36 well i'm taking a break, gonna go watch some tv then come back to figure out whats going wrong Jan 30 19:44:37 oh Jan 30 19:44:39 okay Jan 30 19:44:43 sounds kosher :) Jan 30 19:44:49 now there is a bug for the strip error Jan 30 19:44:51 http://bugs.openembedded.org/show_bug.cgi?id=644 Jan 30 19:45:12 i really think it'd be nifty to use a zaurus to emulate a palm pilot :) Jan 30 19:45:59 its also because i've got a few programs on the palm i want to use still and i don't want to carry around two expensive devices Jan 30 19:50:32 right Jan 30 19:50:39 it is a wonderful idea Jan 30 20:16:00 is there already a way to make initrd images in OE? Jan 30 20:18:07 myopiate: the pivot-boot target Jan 30 20:18:17 myopiate: depreciated for 2.6 kernels Jan 30 20:19:02 at least it is depreciated on spitz and other cxx00 targets Jan 30 20:19:26 http://bugs.openembedded.org/show_bug.cgi?id=421 Jan 30 20:19:41 RP: kernel-2.6.15-r1 bug ^^ Jan 30 20:20:18 RP: hopefully it is not due to bad gcc output Jan 30 20:25:24 I can get to the openembedded domain... is anyone else having this problem? Jan 30 20:26:41 openembedded.org domain that is Jan 30 20:30:20 myopiate: DNS is down Jan 30 20:30:30 for bugs.openembedded.org, see topic Jan 30 21:12:51 wow Jan 30 21:13:00 my zaurus dies at exactly 2 minutes after suspend, i think Jan 30 21:13:20 i'm testing now to see which second it dies at after 2 min Jan 30 21:15:20 shadows: What kernel is that? Jan 30 21:15:26 johnX: 2.6.15-r1 Jan 30 21:15:31 the linux-openzaurus kernel Jan 30 21:15:42 strange Jan 30 21:15:53 I'm running that too, on an Akita Jan 30 21:16:08 kconfig modified only with module-izing the MTD and JFFS2 parts (not needed for hard drive boot) Jan 30 21:16:21 also of note is it was built on amd64 host Jan 30 21:16:39 if gcc is the cause of this issue, then there are serious ramifications Jan 30 21:17:24 did anyone ever figure out the AMD64-gcc-makes-arm-binaries-slightly-bigger issue? Jan 30 21:18:04 johnX: no, i did file a bug report and have documented two work-around methods Jan 30 21:18:18 okayyyy it dies at 181 seconds Jan 30 21:18:22 it works at 180 seconds Jan 30 21:18:36 so there is something at precisely two minutes which fries Jan 30 21:18:37 heh Jan 30 21:18:47 that's pretty weird Jan 30 21:19:08 freaky-deaky Jan 30 21:19:32 i can't even reset the device Jan 30 21:19:36 i have to pull the power from it Jan 30 21:19:41 then reset Jan 30 21:19:53 and put battery back in and it will then boot Jan 30 21:20:43 http://bugs.openembedded.org/show_bug.cgi?id=641 amd64 host build metabug Jan 30 21:30:40 johnX: my concern is that gcc spits out object code from amd64 build host that is not equivilent to that of object code generated by gcc on ia32 build host Jan 30 21:31:03 ugh Jan 30 21:31:56 yeah, I remember doing a whole bunch of test-builds for you the other night... Jan 30 21:32:14 johnX: they helped tremendously Jan 30 21:32:18 thank you again for those Jan 30 21:32:21 no prob Jan 30 21:32:33 we narrowed it down to gcc spitting out different object code, depending on the host arch Jan 30 21:32:50 that's not so good Jan 30 21:32:50 i talked to some gcc devs, they said that it was fixed in the current gcc HEAD Jan 30 21:33:09 which is to say, it's unlikely to be worth fixing or backporting a fix to gcc 3.3.4 Jan 30 21:33:21 well it's kind of a serious bug Jan 30 21:33:26 yeah Jan 30 21:33:42 RP said that so long as the code was equivilent, it is not very serious Jan 30 21:34:10 and he suggested the kernel modules that might be modularized as a workaround, to make a kernel within the size constraints Jan 30 21:34:52 well then let's hope that the suspend issue you have right now is just a simple kernel bug Jan 30 21:35:09 and not a manifestation of anything AMD64 related Jan 30 21:35:28 heh Jan 30 21:35:33 i'm hoping it is amd64 related Jan 30 21:35:42 a kernel bug would mean headaches for RP Jan 30 21:35:44 ;) Jan 30 21:35:59 unless it's a really simple one Jan 30 21:36:13 something seriously bothers me about a compiler that does not build the same code due to the host it is on Jan 30 21:36:49 i am not so bothered by a bug in the kernel. Jan 30 21:37:29 johnX: you have ia32 yes? Jan 30 21:37:42 my build box is IA32 Jan 30 21:37:47 desktop is AMD64 Jan 30 21:37:54 ah okay Jan 30 21:38:10 weird. i am doing a 2nd trial of tests for suspend Jan 30 21:38:17 now it is not behaving consistently Jan 30 21:38:31 that's the worst kind of kernel bug Jan 30 21:38:43 well, it is still failing to resume Jan 30 21:38:51 a suspend or hard crash problem that can't be reproduced reliably... Jan 30 21:39:37 i maintain that god, she is having a vacation and i am her amusement or torture toy Jan 30 21:40:39 possibly a slightly pessimistic outlook, but I really have no way to disprove it Jan 30 21:41:19 bug #569856 "deity.ko causes kernel Oops, system fails to resume" Jan 30 21:41:33 yep Jan 30 21:41:56 i'm on a build where GPE-image fonts are completely random Jan 30 21:42:06 either huge, or tiny, or normal and everything else is huge Jan 30 21:42:10 what a comical bug Jan 30 21:42:24 I think I might have run into that Jan 30 21:42:55 fixed with latest somethingorother panel Jan 30 21:43:25 well that just fixes menu fonts in the panel application menu Jan 30 21:43:28 I thought Jan 30 21:43:38 oh Jan 30 21:43:42 I ran into an issue once when I didn't delete my home dir between reflashes Jan 30 21:43:59 there's also one that caused sqlite to be bollocks'd on amd64 build host generated images Jan 30 21:44:10 which in turn makes failures for looking up font sizes Jan 30 21:44:18 interesting Jan 30 21:44:19 and defaults to what are known to be sane Jan 30 21:44:34 yeah Jan 30 21:44:42 trying to collect and correct all amd64 build host issues Jan 30 21:45:22 at some point GPE started understanding how to deal with a 200dpi screen Jan 30 21:46:08 I used to use a hacked up theme that worked around the fact that it didn't understand what to do with a 200dpi screen Jan 30 21:46:23 ah Jan 30 21:46:25 that may be it Jan 30 21:46:36 200dpi, wowzer that's what it is huh Jan 30 21:46:42 almost medical-grade Jan 30 21:46:51 it's more than 200dpi actually I think Jan 30 21:47:23 i hardly notice the single dead pixel on my screen ;) Jan 30 21:47:38 I think mine are all alive Jan 30 21:48:03 which only happened because I payed extra to buy it from someplace that had a really good return policy Jan 30 21:48:09 ah Jan 30 21:48:12 yeah i got mine used Jan 30 21:48:27 i look at it like, this thing was out of my price range anyways Jan 30 21:48:33 i'm glad to have it dead pixel or not Jan 30 21:48:48 I look at mine as a very inexpensive laptop :) Jan 30 21:49:11 it does everything half-assed Jan 30 21:49:22 heh kinda Jan 30 21:49:26 that's not all software issues mind ya IMO Jan 30 21:49:28 it's the ultimate compromise Jan 30 21:49:38 no I know what you're saying Jan 30 21:49:49 i mean, the wifi? the usb port for cxx00 is in the wrong spot Jan 30 21:50:23 okay for leaving out wifi, but bluetooth Jan 30 21:50:43 that's just senseless to leave out bluetooth. everything has it, and it would make sense to integrate Jan 30 21:50:56 I read the things on oesf.org/forums about people modding their Zaurii and I must say the thought of internal bluetooth sounds better and better... Jan 30 21:51:07 yeah Jan 30 21:51:56 but the thought of me soldering in my C1000 is like a dog doing brain surgery with a sledgehammer Jan 30 21:52:15 that smooth, huh? Jan 30 21:52:36 not the right tool, not the right user Jan 30 21:53:00 dog passes scalpel to the operating dog under the table using hind toes, like in the poker cardgame paintings Jan 30 21:53:13 Surgeon Woofmeister Jan 30 21:53:26 lol Jan 30 21:54:04 I need to get my hands on a socket bluetooth card Jan 30 21:54:26 really, if the antenna did not stick out too far Jan 30 21:54:33 it might be acceptable in the external slot Jan 30 21:54:39 neverminding the wifi Jan 30 21:54:45 I'm thinking of the socket one that doesn't stick out... Jan 30 21:54:56 then I guess when I need wifi I'll do it via usb Jan 30 21:55:16 if you had one that did not stick out, it might replace the built-in hard drive on the c3100 Jan 30 21:55:34 * johnX has a C1000 Jan 30 21:55:47 oh, internal stuff isn't soldered Jan 30 21:55:56 it's lacking some chips even Jan 30 21:56:59 I could theoretically do bluetooth over serial or usb and wedge it somewhere in the empty space Jan 30 21:57:43 it's a pretty low performance protocol Jan 30 21:58:33 this is what I want in reality: http://www.socketcom.com/product/BL4538-641.asp Jan 30 21:58:51 they want $100USD for it though Jan 30 21:59:53 I would really rather avoid screwing around inside my Zaurus... Jan 30 22:00:17 you know what though Jan 30 22:00:25 where do you use your Z? Jan 30 22:00:38 just about everywhere Jan 30 22:00:46 I carry it around in my pants pocket Jan 30 22:00:53 heh Jan 30 22:00:57 in a case, or just as-is Jan 30 22:01:01 just closed Jan 30 22:01:07 that's kind of neat Jan 30 22:01:15 can't do that with a laptop Jan 30 22:01:20 exactly Jan 30 22:01:31 what kind of runtime do you get from it Jan 30 22:01:38 does it last all day for moderate use? Jan 30 22:01:45 yeah, no problem Jan 30 22:01:52 with wifi on quite a lot Jan 30 22:02:02 would be hard to do that with a modern laptop Jan 30 22:02:36 which is the whole reason I got the Z over a laptop: I knew I wouldn't drag a laptop everywhere, so it wouldn't be nearly as useful in the long run Jan 30 22:03:02 but on my Z I keep notes, can ssh, have music, email, web, and some games Jan 30 22:03:27 on the flash Jan 30 22:03:31 or on another storage device Jan 30 22:03:35 SD Jan 30 22:03:54 * shadows reboots Z for the 20th time) Jan 30 22:04:12 ugh. i've spent hours on this. i wonder if it even matters how long i can go without the Z crapping out Jan 30 22:04:14 remember to fsck your cards...doesn't seem to fsck SD automagically Jan 30 22:04:26 or if RP will be like "oh yeah, that bug. hmm. i should probably fix that sometime" Jan 30 22:04:40 hey, I built a a 2.6.15-r1 last night on IA32 Jan 30 22:04:47 works good? Jan 30 22:04:47 for Spitz... Jan 30 22:04:58 oh Jan 30 22:05:02 i could try that huh Jan 30 22:05:05 yeah Jan 30 22:05:09 that's what I was thinking Jan 30 22:05:40 it would need root=/dev/hda6 Jan 30 22:05:58 and JFFS2 set to m instead of y, and MTD set to m instead of y Jan 30 22:06:07 to be comparable Jan 30 22:06:15 also I think mine has preempt off Jan 30 22:06:48 i just had a funny thought Jan 30 22:07:03 if this were a WinCE device, the fs would have been trashed about 10 reboots ago Jan 30 22:07:10 ext3 keeps on clicking without trouble Jan 30 22:07:54 even ext2 manages to be pretty robust on my SD Jan 30 22:08:24 is jffs2 appropriate for SD ? Jan 30 22:08:44 err...I don't think so...but I couldn't defend that position Jan 30 22:11:06 I kind of feel like trying some more esoteric FS on my SD Jan 30 22:11:43 jffs2 no good for bigger than 32mb. uses lots of RAM Jan 30 22:11:57 MB Jan 30 22:14:20 ah okay Jan 30 22:14:42 johnX: results of my tests are at http://bugs.openembedded.org/show_bug.cgi?id=421 Jan 30 22:17:25 that is so odd Jan 30 22:17:38 RP will know what is causing this Jan 30 22:18:09 it must be some registers of the hardware that are inconsistent Jan 30 22:22:41 if you want you can send me your defconfig-cxx00 I'll build a 2.6.15 to test, to try and figure out whether it might be an AMD64 issue Jan 30 22:23:38 okay Jan 30 22:23:53 the other change was in the linux-openzaurus.inc Jan 30 22:24:02 i changed hda1 to hda6 for the root= thing Jan 30 22:24:16 needed since i use a derivation from the normal partition mapping Jan 30 22:24:53 * shadows pokes web server Jan 30 22:26:03 http://jnc.pimpcat.org/files/oz/defconfig-cxx00~jnc-spitz Jan 30 22:29:14 build started Jan 30 22:36:08 okay Jan 30 22:36:28 meanwhile, i'm building a kernel on amd64 build host that is using the optimize-for-size setting in defconfig Jan 30 22:36:36 so that i have a wide variety of test cases Jan 30 22:36:55 i.e. considering the possibility that modularized MTD driver is the cause Jan 30 22:36:58 n Jan 30 22:36:59 `~.~. Jan 30 22:53:50 what the hell, it won't flash Jan 30 22:53:53 the error message in japanese Jan 30 22:53:57 brilliant. Jan 30 22:55:05 maybe file a bug with Sharp Japan Jan 30 22:55:13 oh shut up Jan 30 22:55:18 * shadows =P Jan 30 22:55:21 sorry :D Jan 30 22:55:29 anyways I have another image for you to test Jan 30 22:55:38 super! Jan 30 22:56:24 the kernel here built as "r2"...I wonder if there is any significance Jan 30 22:56:30 hmm Jan 30 22:56:35 that means i need to update Jan 30 22:56:51 if the bug is fixed, though, it means it's fixed Jan 30 22:56:52 if not Jan 30 22:57:00 then we know on ia32 r2, it hasn't been fixed Jan 30 22:57:03 thus worth trying IMO Jan 30 22:57:43 oh haha Jan 30 22:57:57 i bet that the flasher wants me to plug in AC power Jan 30 22:57:57 duh Jan 30 22:58:25 now it works Jan 30 22:58:45 * shadows flashes size optimize kernel, built from amd64 build host Jan 30 22:58:57 woo woo! look at it flash! Jan 30 22:59:02 yeah Jan 30 22:59:04 sexayyy Jan 30 22:59:57 shadows: ftp://bloom.jads.com/pub/spitz-kernel-2.6.15-r2.tar.bz2 Jan 30 23:00:10 awesomethanks Jan 30 23:03:22 alright Jan 30 23:03:25 now we wait 2 minutes Jan 30 23:06:19 okayyyy Jan 30 23:06:26 no love? Jan 30 23:06:30 size optimized kernel not lovin' it Jan 30 23:06:47 no big Zaurk with extra cheese Jan 30 23:07:11 bring in the next contestant Jan 30 23:07:26 *drumroll* Jan 30 23:09:23 please hold your breath for, the flashing "oooh, ahhh" Jan 30 23:11:46 okay booted and stuff Jan 30 23:11:54 i didn't install the modules, which is an important note Jan 30 23:12:00 hmm Jan 30 23:12:03 way too lazy Jan 30 23:12:09 it's probably not an issue Jan 30 23:12:29 since the config is the same, as is the kernel version Jan 30 23:12:44 the object output would be different though Jan 30 23:12:49 ah Jan 30 23:12:55 duh Jan 30 23:12:58 hehe Jan 30 23:13:05 now we wait Jan 30 23:13:12 * johnX flips brain back into "on" position Jan 30 23:13:26 hah hahaha. oh man. i like working on this project Jan 30 23:13:59 you don't have issue with akita and leaving it suspended for many minutes? Jan 30 23:14:10 it's been on and off all day Jan 30 23:14:15 okay Jan 30 23:14:27 to confirm, you have left it suspended for a time greater than 5 min Jan 30 23:14:34 yes Jan 30 23:14:34 and it resumed and survives? Jan 30 23:14:38 greater than 5 hours Jan 30 23:14:49 good Jan 30 23:14:52 so it works, somewhere Jan 30 23:14:58 resumes, while playing the same music it was playing when it was turned off Jan 30 23:15:14 akita should be identical to spitz, with the exception of CF configuration Jan 30 23:15:48 errr...mostly Jan 30 23:15:48 next move would be for me to get a build of OZ gpe-image on ia32 host Jan 30 23:15:59 mostly? Jan 30 23:16:03 next move if this fails is to install those modules :P Jan 30 23:16:09 oh yeah, thanks Jan 30 23:16:19 let my brain join yours in the on position Jan 30 23:16:57 I recall something about the Akita having an "I2C extender" or some such magic that the other Cxx00s don't have Jan 30 23:17:23 * shadows pokes "on" button Jan 30 23:17:25 nothing. Jan 30 23:17:34 damn Jan 30 23:18:04 yeaaah. Jan 30 23:18:22 any suggestions on how best to do the ipk installing of modules? Jan 30 23:18:26 just, erm, do them all? Jan 30 23:18:47 suuure Jan 30 23:18:55 yeah honestly just go: Jan 30 23:19:02 ipkg install *spitz*ipk Jan 30 23:23:08 action initiated. awaiting results Jan 30 23:23:26 probably good that i'm going r1 to r2 Jan 30 23:23:30 means the upgrade thing goes Jan 30 23:23:44 yeah Jan 30 23:24:07 i worry about shock to the hard disk Jan 30 23:24:36 or internal flash going bad Jan 30 23:24:57 you mean physical shock? or killing it from too many writes? Jan 30 23:25:03 either Jan 30 23:25:11 i'm not going around dropping my Z Jan 30 23:25:31 I dropped my Z once a couple feet onto blacktop Jan 30 23:25:36 the sound of the hard drive makes me nervous, compared to something like my Palm IIIxe Jan 30 23:25:49 which i regularly drop kick, smash, sit on, or toss around Jan 30 23:26:02 not a fair comparison for CPU or capabilities Jan 30 23:26:19 PalmOS had nice PIMs and good handwriting recognition Jan 30 23:26:25 but really, i've come to expect my daily use PDA (the Palm IIIxe) to withstand a smacking Jan 30 23:26:28 yeah Jan 30 23:26:38 I must say I'm a big fan of solid state electronics... Jan 30 23:26:45 i upgraded that thing to Palm OS 4.1 Jan 30 23:26:54 it's a really risky thing to do, so worth it Jan 30 23:27:10 i would buy one off eBay and flash to OS 4.1 again, if mine broke Jan 30 23:27:33 though, have you seen flash USB memory thumbdrives die? Jan 30 23:27:43 there's been two that i saw break Jan 30 23:27:53 flash memory went kaput Jan 30 23:28:45 Fry's keeps having Palm IIIxe's cheap...Almost makes me want to get one to play with, but my Z is doing very nicely Jan 30 23:28:50 buy two Jan 30 23:28:59 they're cheap enough Jan 30 23:29:21 then there's two things you'd want to do FYI, one is modify the base so that it has the "no drain" hack Jan 30 23:29:30 otherwise just sitting in the base, the Palm drains battery Jan 30 23:29:46 the second is upgrading the OS to Palm OS 4.1 Jan 30 23:29:52 you'll need a windows xp box for that Jan 30 23:30:10 i have files to do that, if you ever need them Jan 30 23:30:24 Palm IIIxe could never do the things zaurii could do Jan 30 23:30:28 maybe the other way is also true Jan 30 23:30:36 I think you're right Jan 30 23:30:38 i use my Palm IIIxe to keep my work schedule Jan 30 23:30:42 I used to have a Visor Prism Jan 30 23:30:57 ah Jan 30 23:31:18 yeah it's like, if some nutcase stole my Z, i would be very disturbed Jan 30 23:31:32 if my Palm went missing? everything important is blowfish encrypted Jan 30 23:31:40 the device itself isn't hard to replace Jan 30 23:32:00 i had a newer Palm, the Tungsten 'E' Jan 30 23:32:01 the thing I realized with my Prism: I don't actually need a PDA Jan 30 23:32:05 mmm Jan 30 23:32:06 yep Jan 30 23:32:29 w/o a keyboard, the arm based things are overkill and undercapable Jan 30 23:32:39 agreed Jan 30 23:33:00 any kind of keyboard is way better than handwriting recognition... Jan 30 23:33:07 The only PDA function I use is to write notes Jan 30 23:33:26 no addressbook? Jan 30 23:33:32 or is that your cell phone Jan 30 23:33:45 good morning all Jan 30 23:33:50 koen|gprs: salu Jan 30 23:33:55 t$ Jan 30 23:33:58 mornin' Jan 30 23:34:13 shadows: nope, no addressbook, no need right now Jan 30 23:34:35 I think GPE will be quite up to the task by the time I need one Jan 30 23:34:41 haha Jan 30 23:34:45 by that time, it will be sentient Jan 30 23:34:50 the schedule app already looks a lot nicer than it used to Jan 30 23:34:54 and it could tell you why you need a PDA Jan 30 23:35:09 * Geo_KM is away: Away at the moment Jan 30 23:35:11 all I want is that "memo" program from Palm Jan 30 23:35:16 it was *perfect* Jan 30 23:35:19 we're at the letter 'p' btw for installing module ipks Jan 30 23:35:32 is that onto the HD then? Jan 30 23:35:36 yes Jan 30 23:36:06 i reconfigured my partition layout. Sharp ROM has it factory set to hda (hda1 hda2 hda3) Jan 30 23:36:11 everytime I wake up from sleep more bugs get added Jan 30 23:36:11 maybe I should give up sleeping ;) Jan 30 23:36:19 with like, the ROM, usr programs and junk, and the vfat fs Jan 30 23:36:34 i slimmed down the vfat fs, and stuck an extended partition at the end Jan 30 23:36:38 koen|gprs: sleep is for the weak, you can sleep when you're dead :) Jan 30 23:36:48 inside of the extended partition i stuffed a swapfs and rootfs Jan 30 23:37:02 heh...I swap to SD Jan 30 23:37:07 ew? Jan 30 23:37:15 better than "Out of memory" Jan 30 23:37:20 perhaps Jan 30 23:37:28 plus I have an SD card with a lifetime warranty Jan 30 23:37:34 * shadows snickers Jan 30 23:37:36 they've already given me one new one Jan 30 23:37:41 i bet the mfr didn't have any clue what was coming Jan 30 23:38:09 "blame John, he's the reason Corsair doesn't have a lifetime warranty on flash memory stuff any more." Jan 30 23:38:43 but even sandisk has a 5 year warranty... Jan 30 23:39:22 i'll think of a way to put the christian bible stories into context of manufacturer and rebates, warranties, service level agreements Jan 30 23:39:45 it will be comical. Jan 30 23:40:04 * shadows notes "while i am waiting for the damned ipk files to install" Jan 30 23:40:13 too much time on your hands? or maybe you need a "creative outlet"? Jan 30 23:40:20 but then would billy be a deamon or a profit? Jan 30 23:41:07 emte: in the business of collecting soul licensing, i'd imagine Jan 30 23:41:42 read those ELUA's or... ye be damned Jan 30 23:41:45 not religious monopoly and inqusitions? Jan 30 23:41:45 quite literally Jan 30 23:42:17 don't get me started on Hasbro Jan 30 23:42:20 yeesh Jan 30 23:42:23 luckily for me most eulas are illegal over here Jan 30 23:42:28 lol Jan 30 23:45:53 Configuring... Jan 30 23:46:15 the openwrt uses ipkg and is way faster Jan 30 23:46:16 why? Jan 30 23:46:44 the smaller the "Packages" file you're dealing with the faster ipkg seems to act Jan 30 23:46:51 ah Jan 30 23:47:17 on a fresh flash try manually installing a .ipk Jan 30 23:47:22 is there serious overhead in using dpkg? Jan 30 23:47:58 i thought on something like the zaurus, it was silly not to use debian packaging format Jan 30 23:48:09 but that makes it complicated i guess Jan 30 23:48:11 hmmm...I don't know, but keep in mind that ipkg does more than dpkg Jan 30 23:48:24 You'd need to throw in apt-* and anything it depended on Jan 30 23:48:25 oh, it does hmm Jan 30 23:48:53 ipkg handles fetching and installing, dpkg just installs things it's told to Jan 30 23:49:42 ah okay Jan 30 23:50:10 though IIRC the Nokia 770 has dpkg Jan 30 23:59:54 modules installed, c3000 rebooted, suspended Jan 30 23:59:58 waiting 2 min.. Jan 31 00:03:14 oh crap, it worked. Jan 31 00:03:21 trying again and waiting 3 min Jan 31 00:04:16 johnX: can you tell what changed between r1 and r2? Jan 31 00:04:33 once I remember my monotone invocations Jan 31 00:05:01 * shadows casts "focus, mighty wizard" enchantment +3 on John Jan 31 00:05:30 oh right, monotone log Jan 31 00:06:13 yeah this is definitely working with your kernel Jan 31 00:06:16 and modules Jan 31 00:06:40 which at this immediate juncture doesn't tell me a whole hell of a lot Jan 31 00:07:18 next step might be for me to upgrade source and then build another image on my host Jan 31 00:07:31 that's what I would suggest Jan 31 00:08:24 http://pastebin.com/531761 Jan 31 00:08:43 that's the changelog between r1 and r2 *I think* Jan 31 00:09:09 "# Jan 31 00:09:11 Rework backlight class to handle a number of bugs to do with suspend/resume Jan 31 00:09:11 # Jan 31 00:09:11 making state assumptions " Jan 31 00:09:12 maybe Jan 31 00:09:24 could be Jan 31 00:10:36 okay, doing a monotone pull Jan 31 00:10:49 and all the lovely bugs that come with it Jan 31 00:11:06 I really wish orbit2-native would build :( Jan 31 00:11:20 I should try making it build again Jan 31 00:13:30 updated bug http://bugs.openembedded.org/show_bug.cgi?id=421 Jan 31 00:14:12 i don't understand what an ORB is Jan 31 00:14:41 i'm more of a user than anything too technical. i just do technical things when user stuff is broken Jan 31 00:14:58 same here Jan 31 00:20:23 goodmorning Jan 31 00:43:44 johnX: hey when you are able, could you try out optimizing for size on your akita ? Jan 31 00:44:28 johnX: i'd like to suggest making optimize for size the default Jan 31 00:47:20 ok, orbit2-native just failed to build again so I'll rebuild the kernel while I'm digging through it Jan 31 00:51:34 oh what the hell Jan 31 00:51:37 my Z won't resume Jan 31 00:51:42 using your kernel and modules Jan 31 00:51:51 but the light is on Jan 31 00:52:08 the "charge" light? Jan 31 00:52:24 yes Jan 31 00:52:35 charge light responds when i plug/unplug the ac dc adapter Jan 31 00:52:43 which is *different* than using my kernel from before Jan 31 00:53:24 well, once I build this r2 kernel for Akita and get the same problem as you we'll both be able to complain Jan 31 00:54:02 next thing is for me to build an akita kernel Jan 31 00:54:12 and have you do that Jan 31 00:54:37 first i'd like to verify that your machine works with the ordinary ia32 build host made kernel and the optimize as default Jan 31 00:54:54 'cause that is something which, either way, is going to be the default for 2.6 kernels soon Jan 31 00:55:05 like 2.6.17 as soon as that Jan 31 01:55:10 morning Jan 31 02:10:47 03koen 07org.oe.oz354fam083 * r028d0a96... 10/packages/meta/meta-gpe.bb: meta-gpe: ship startup-monitor by default, fixes hh.org bug #1496 Jan 31 02:11:12 shadows: I was just reading back about problem :-/ Jan 31 02:11:21 yeah Jan 31 02:11:32 it actually happens on all kernels and build hosts i've tried Jan 31 02:11:52 shadows: Does the device resume and just not the lcd/backlight? Jan 31 02:12:02 that is the strange thing Jan 31 02:12:05 eg. can you still ssh into it? Jan 31 02:12:25 when build host is ia32, the charge light will change if i plug/unplug the ac dc adapter Jan 31 02:12:37 when build host is amd64, the charge light is off and does not change Jan 31 02:12:51 i do not have networking functional on here yet Jan 31 02:13:20 wpa_supplicant is not communicating to my AP like it does in OZ 3.5.4 testing Jan 31 02:13:22 All I can add is that suspend/resume works fine here :-/ Jan 31 02:13:38 yeah, johnX says his Akita works fine also Jan 31 02:14:04 any idea why it would be a time period? Jan 31 02:14:33 it is almost exactly 2 minutes (120 seconds). if i resume in less time, it works. if i wait too long, it does not Jan 31 02:14:52 The Z periodically wakes up when suspended and charging. I wonder if we do anything after 2 minutes... Jan 31 02:16:00 interesting Jan 31 02:16:19 morning all Jan 31 02:16:34 That happens every 10 minutes though... Jan 31 02:16:36 morning ade|desk Jan 31 02:18:13 RP: also, it is with linux-2.6.15-r1 i did my tests Jan 31 02:18:24 RP: on saturday i couldn't get bitbake to work for an image, failing every time due to missing packages, you asked me to do a build with triple debug messages, now the sodding thing works every time ... most unfair! Jan 31 02:18:36 now i am running amd64 build host compiled linux-2.6.15-r2 and it seems to behave more appropriately Jan 31 02:19:07 ade|desk: sourceforge being down have anything to do with that? Jan 31 02:19:39 not originally, sunday and most of monday that was the stopping factor Jan 31 02:20:04 grr gaim being hosted at sf.net Jan 31 02:20:18 it's not a problem for gaim devs Jan 31 02:20:29 only people crazy enough to try anoncvs at sourceforge ;) Jan 31 02:21:29 are they moving to svn on sf.net ? Jan 31 02:23:08 ade|desk: If the problem does reappear, you know what to do. Strange its just stopped happening... Jan 31 02:25:44 RP: just checked the build box at work (small bloke with abacus), no error there either Jan 31 02:25:46 grr Jan 31 02:26:05 RP: er, ssid not being seen by wpa_supplicant, is that a problem with the hostap driver? Jan 31 02:26:21 they're all blank (wpa_supplicant debug info shows ssid='') Jan 31 02:26:22 shadows: Is the version of wpa_supplicant up to date? Jan 31 02:26:25 hmm Jan 31 02:26:29 good question Jan 31 02:26:39 I suspect we need a newer one in OE Jan 31 02:26:58 v0.2.6 Jan 31 02:27:22 newest on authors website is... wowzer Jan 31 02:27:26 0.4.7 Jan 31 02:27:44 morning all Jan 31 02:27:58 * shadows pokes build system Jan 31 02:28:29 hi dirk Jan 31 02:28:48 RP: i think the new one has a bb, it's just not used by default for openzaurus-unstable and gpe-image Jan 31 02:28:59 do13_: btw, I made some changes in the 2.6.15 kernel backlight code - I have to update the patch to make it operate with tosa to match Jan 31 02:29:00 building and installing to see if that helps Jan 31 02:29:40 Hi Richard. Which parts did you changed? Jan 31 02:30:14 do13_: I created http://www.rpsys.net/openzaurus/patches/corgi_bl_cleanup-r2.patch Jan 31 02:30:35 do13_: and then updated your patch for generic support (for tosa) to: http://www.rpsys.net/openzaurus/patches/corgi_bl_generic-r2.patch Jan 31 02:31:23 do13_: I also moved led_tosa ahead of the other tosa patches in the patch order (so I can publish the complete patch series to LKML soon) Jan 31 02:32:26 shadows: I just tested 2.6.15-r2, left it suspended for more than 3 minutes and it came back fine Jan 31 02:32:34 shadows: I'm going to try again though Jan 31 02:32:45 johnX: okay, this time try more than 10 min Jan 31 02:32:45 RP: Do you plan to include led/tosa.c? Jan 31 02:32:56 shadows: can do Jan 31 02:33:07 do13_: Its up to you :) Jan 31 02:34:02 RP: I'll check with my working copy. At least "charger" needs to be changed into "sharpsl-charge" Jan 31 02:34:30 johnX: Are you trying with AC power connected or disconnected? (and shadows, what about you?) Jan 31 02:34:37 do13_: My copy has that fixed Jan 31 02:34:44 do13_: http://www.rpsys.net/openzaurus/patches/led_tosa-r2.patch Jan 31 02:34:51 I have AC connected Jan 31 02:35:05 RP: good question, i have ac connected now. when i was testing, it was disconnected. Jan 31 02:35:12 so.. maybe i need to do more testing Jan 31 02:35:14 but I had R1 installed for just about as long as it's been out and I've been using it day-to-day Jan 31 02:35:17 morning Jan 31 02:35:22 hi hrw|work Jan 31 02:35:24 Hi Marcin Jan 31 02:35:37 RP: .default_trigger = "charge" is wrong Jan 31 02:36:01 do13_: Ah, yes - I'll change that Jan 31 02:36:52 03jbowler 07org.oe.dev * r98ec75a9... 10/packages/e2fsprogs/e2fsprogs_1.38.bb: Jan 31 02:36:53 e2fsprogs: split out blkid for use in initrd in 1.38 Jan 31 02:36:53 - blkid (and uuidgen) are useful for non-ext2/3 file systems Jan 31 02:36:53 (e.g. reiserfs) and can be used at boot to reliably find the Jan 31 02:36:53 root file system, therefore it is useful to have them in Jan 31 02:36:53 separate packages. e2fsprogs (${PN}) now RDEPENDS on the new Jan 31 02:36:55 packages to ensure nothing gets broken. Jan 31 02:36:59 03jbowler 07org.oe.dev * r7515657d... 10/packages/meta/slugos-image.bb: slugos-image: add e2fsprogs-blkid (fs id finding program) to meta Jan 31 02:37:02 03jbowler 07org.oe.dev * r95642207... 10/packages/slugos-init/ (4 files in 3 dirs): Jan 31 02:37:04 slugos-init: make /linuxrc use disk uuids, remove .recovery stuff in 0.10 Jan 31 02:37:06 - the /boot/disk bootstrap now mounts -U uuid if one is available from Jan 31 02:37:06 do13_: http://www.rpsys.net/openzaurus/patches/led_tosa-r3.patch :) Jan 31 02:37:08 turnup. turnup checks disk for blkids and no longer complains about Jan 31 02:37:10 .recovery, /boo/* no longer create .recovery or check for it on boot; Jan 31 02:37:12 disconnect the disk or the network to boot to flash on a semi broken Jan 31 02:37:14 system. Jan 31 02:37:18 03jbowler 07org.oe.dev * rfe1ec31f... 10/packages/uclibc/uclibc-cvs/thumb-mov-pc-bx.patch: uclibc: EABI thumb patch in cvs Jan 31 02:46:34 RP: Thanks. Now you can publish this:) Jan 31 02:46:35 hey hrw|work Jan 31 02:46:52 RP: Did you get comments from rmk? Jan 31 02:47:43 do13_: Just that he had recieved the emails but didn't deam them important enough to have read yet... Jan 31 02:48:05 RP: isnt that an improvement? Jan 31 02:48:43 shadows: on AC or off AC my Akita will come back from suspend no problem Jan 31 02:49:23 XorA: If he decides to read them at some point, perhaps... Jan 31 02:50:14 crappit, Ive no idea what I am doing budget wise Jan 31 02:51:56 johnX: hm Jan 31 02:51:58 well okay Jan 31 02:52:00 i have network up Jan 31 02:52:08 * XorA hopes tech market picks up this year Jan 31 02:52:10 using wifi and updated wpa-supplicant Jan 31 02:52:53 shadows: I'll be using it tomorrow and I'll be on here tomorrow night, so I'll tell you if it crashes after being suspended off AC for a couple hours Jan 31 02:53:12 but for now I'm going to sleep Jan 31 02:53:22 thanks Jan 31 02:53:28 'night Jan 31 02:53:29 have a good rest sir =) Jan 31 02:54:49 RP: wpa-supplicant 0.4.7 is needed with DCF-660w and openzaurus-unstable from a few days ago, 2.6 kernel etc. Jan 31 02:55:17 the stuff that gpe-image target brings in (0.2.6) is not compatible like you say Jan 31 02:55:22 shadows: ok. I'll adjust that in OE Jan 31 02:55:27 excellent Jan 31 02:55:40 it does add a dependency on ssl though Jan 31 02:55:48 err, i don't know if it really does Jan 31 02:56:23 i'm just going by the coincidence that i installed wpa_supplicant and -not wpa_supplicant-nossl- Jan 31 02:57:16 okay now for tests with power cord plugged in **** ENDING LOGGING AT Tue Jan 31 02:59:56 2006