**** BEGIN LOGGING AT Tue Jul 12 23:59:57 2005
Jul 13 00:01:27 <CIA-8> 03jp30 * 10unslung/make/libstdc++.mk: strip libstdc++
Jul 13 00:03:56 <CIA-8> 03jp30 * 10unslung/make/libtiff.mk: remove libtool archives from staging, static libs from package
Jul 13 00:12:10 <CIA-8> 03jbowler * r77 10/trunk/openslug/nslu2-linux/packages/install: Add new install package to fix install problems
Jul 13 00:15:07 <CIA-8> 03jp30 * 10unslung/make/libtiff.mk: libtiff: upstream upgrade, fix download location
Jul 13 00:17:10 <CIA-8> 03jp30 * 10unslung/make/libpng.mk: libpng: don't stage libtool archives
Jul 13 00:28:30 <CIA-8> 03jp30 * 10unslung/make/gtk.mk: don't stage libtool archives
Jul 13 00:38:40 <CIA-8> 03jbowler 07org.openembedded.nslu2-linux * rf4eaaacf... 10/packages/meta/openslug-packages.bb: 
Jul 13 00:38:40 <CIA-8> Move various packages which only compile with glibc to the 'linux only'
Jul 13 00:38:40 <CIA-8> list.
Jul 13 00:40:33 <CIA-8> 03jp30 * 10unslung/make/ncurses.mk: ncurses: strip, don't package static libs
Jul 13 00:43:46 <CIA-8> 03jp30 * 10unslung/make/ncurses.mk: ncurses: unbork symlink tweak
Jul 13 00:43:47 <CIA-8> 03jbowler 07org.openembedded.nslu2-linux * r43007d43... 10/ (8 files in 8 dirs): 
Jul 13 00:43:47 <CIA-8> Add install-native package and remove the install-sh from those packages
Jul 13 00:43:47 <CIA-8> which used it to get round the install -s problem - now simply add
Jul 13 00:43:47 <CIA-8> install-native to the DEPENDS for packages which attempt to use the host
Jul 13 00:43:47 <CIA-8> install with '-s'
Jul 13 03:14:05 <kared> !release
Jul 13 04:57:52 <CIA-8> 03repvik * r78 10/trunk/openslug/nslu2-linux/packages/xinetd: Added xinetd
Jul 13 04:58:45 <DaKa2> :)
Jul 13 05:36:07 <VoodooZ_work> can somebody explain to me why I'm getting those annoying monotone 3-way merges on files I never even looked at. (openzaurus-3.5.4.conf!)
Jul 13 05:36:41 <VoodooZ_work> I wouldn't expect to have to do merges unless I made changes to my working copy right?
Jul 13 05:38:42 <VoodooZ_work> Actually, the only difference between the files was the CVSDATE header. Still annoying.
Jul 13 05:39:43 <VoodooZ_work> Unless I shouldn't be using the Makefile's 'make update' target to pull in the latest changes??
Jul 13 05:40:45 <NAiL> VoodooZ_work: Nothing wrong with using make update. There was some oddity last night, when some strange files showed up in a 3-way merge
Jul 13 05:41:35 <VoodooZ_work> crap I hate those!
Jul 13 05:41:56 <VoodooZ_work> so make update and doing it manually are the same right?
Jul 13 05:42:02 <NAiL> yeah
Jul 13 05:42:15 <VoodooZ_work> ok, I was just starting to believe I was doing something wrong.
Jul 13 05:42:55 <VoodooZ_work> After all I would have expected monotone to do its own auto-merge on a file I never even opened.
Jul 13 05:43:08 <VoodooZ_work> Which file was it?
Jul 13 05:44:51 <VoodooZ_work> did you get the openzaurus-....conf one too? How did you fix it?
Jul 13 05:45:51 <NAiL> I didn't. Other people did...
Jul 13 05:46:02 <NAiL> some libopie stuff
Jul 13 05:46:20 <VoodooZ_work> Crap. I have another one now! some vixie cron stuff.
Jul 13 05:46:32 <VoodooZ_work> I miss bk suddenly.
Jul 13 05:47:08 <VoodooZ_work> at least it gave me take the lazy way out by just using the local/remote copy.
Jul 13 05:47:12 <rwhitby-away> For USD$5K per year per developer you can have it back again ....
Jul 13 05:47:19 <VoodooZ_work> yeah. funny.
Jul 13 05:48:17 <VoodooZ_work> If somebody could explain to me what's going on that would help though.
Jul 13 05:48:30 <VoodooZ_work> I can't be the only using it.
Jul 13 05:50:39 <rwhitby-away> Using what?
Jul 13 05:50:45 <VoodooZ_work> now it's  packages/cron/cron_3.0pl1.bb. which I never modified.
Jul 13 05:51:05 <VoodooZ_work> monotone. It's forcing me to do 3-way merges of all sort of files I never touched.
Jul 13 05:51:34 <VoodooZ_work> And it's not too user-friendly as it opens up 3 files in vim with no names/labels to tell me what's what.
Jul 13 05:51:49 <rwhitby-away> yup
Jul 13 05:52:33 <VoodooZ_work> that's why I said bk's merge was better.
Jul 13 05:52:58 <VoodooZ_work> not that I want to go back to it though. It had it's own problems.
Jul 13 05:53:05 <rwhitby-away> yup
Jul 13 05:53:31 <VoodooZ_work> Once I get used to it it should be fine but right now it's a pain.
Jul 13 05:53:35 <rwhitby-away> yup
Jul 13 05:54:22 <rwhitby-away> don't mind me, I'm just waiting for monotone to verify the new revisions (which may take a while .....)
Jul 13 05:54:45 * VoodooZ_work resists the temptation to start from scratch! arrrggh! :)
Jul 13 05:55:09 <VoodooZ_work> Are you doing it via 'make update'?
Jul 13 05:55:22 <rwhitby-away> yep
Jul 13 05:55:27 <VoodooZ_work> ok
Jul 13 05:56:52 <VoodooZ_work> A nice spinning progress indicator would be nice for that rather long process. 
Jul 13 05:57:14 <VoodooZ_work> I would implement it but I'm not that good yet.
Jul 13 05:57:33 <rwhitby-away> night all
Jul 13 05:58:30 <NAiL> nite
Jul 13 06:02:08 <NAiL> strange... 
Jul 13 06:02:15 <DaKa2> ?
Jul 13 06:02:17 <VoodooZ_work> what's that. 
Jul 13 06:02:26 <VoodooZ_work> merging problems? :)
Jul 13 06:02:35 <NAiL> *now* I'm seeing the dropbear disconnects
Jul 13 06:02:50 <jacques> ;-)
Jul 13 06:03:02 <NAiL> but only on unstable
Jul 13 06:03:18 <NAiL> built 1st of july
Jul 13 06:07:54 <rwhitby-asleep> make update succeeded without any merges ...
Jul 13 06:08:34 <VoodooZ_work> damn! I didn't even touch the thing.
Jul 13 06:08:37 <NAiL> I wonder if people who got a specific revision gets into trouble when they do a new update
Jul 13 06:08:45 <VoodooZ_work> What monotone do you have? 
Jul 13 06:09:03 <rwhitby-asleep> 0.19 from debian etch
Jul 13 06:09:29 <VoodooZ_work> 0.19 FC4
Jul 13 06:09:36 <NAiL> 0.19 etch
Jul 13 06:09:37 <rwhitby-asleep> merge succeeded on nudi too
Jul 13 06:10:00 <VoodooZ_work> I just completed my changes too.
Jul 13 06:10:43 <VoodooZ_work> I still don't see why it's asking me to merge zaurus, cron stuff I never used/touch. odd.
Jul 13 06:10:58 <VoodooZ_work> rwhitby-asleep, thanks for confirming.
Jul 13 06:11:15 <VoodooZ_work> did the Makefile change recently?
Jul 13 06:11:24 <rwhitby-asleep> nope
Jul 13 06:12:09 <VoodooZ_work> wow. I don't know what to do now.
Jul 13 06:12:36 <rwhitby-asleep> Sorry VoodooZ_work, but I have no idea why it's asking you to do merges.
Jul 13 06:13:17 <VoodooZ_work> It's ok. I'll just rip apart and start all over again.
Jul 13 06:13:36 <VoodooZ_work> Or maybe I'll take an extended break :(
Jul 13 06:17:31 <VoodooZ_work> I wonder if there's a way to "extract" my changes though.
Jul 13 06:19:01 <NAiL> VoodooZ_work: you could do a recursive diff against a "new" tree, and just cut out the parts you didn't touch
Jul 13 06:19:21 <VoodooZ_work> yeah. I guess.
Jul 13 06:19:49 <VoodooZ_work> I'm paranoid it might happen again though.
Jul 13 06:19:57 <NAiL> if you have uncommitted changes, mt diff lists them
Jul 13 06:20:11 <VoodooZ_work> they're all committed.
Jul 13 06:20:38 <VoodooZ_work> There must be a way to list all my commits and extract them. I'll have to check the mt docs.
Jul 13 06:21:02 <NAiL> ~seen ka6sox
Jul 13 06:21:02 <jbot> ka6sox <~ka6sox@ka6sox.nslu2-linux> was last seen on IRC in channel #nslu2-linux, 37d 11h 49m 15s ago, saying: 'my thoughts exactly'.
Jul 13 06:21:19 <NAiL> ka6sox-office: ping
Jul 13 06:21:41 <NAiL> heh... idle 14 days
Jul 13 06:27:01 <CIA-8> 03rwhitby * r79 10/trunk/openslug/nslu2-linux/packages/ (install install): Fixed the install symlink
Jul 13 06:51:33 <kolla_> so.. time to check what errors I get today :)
Jul 13 06:54:20 <kolla_> NAiL: debian, right? 2.6.12 kernel?
Jul 13 06:56:34 <NAiL> nah, older than that
Jul 13 06:56:53 <NAiL> 2.6.8-2-k7
Jul 13 06:57:22 <NAiL> 2.6.11.2 on the slug
Jul 13 07:01:02 <kolla_> hm, so .. retro
Jul 13 07:01:31 <kolla_> top acted strangely on 2.6.12 on this debian test box we have
Jul 13 07:03:08 <kolla_> from what I understand the -d option of top is a delay in seconds, only with 2.6.12 it acted as if it was 1/100th of a second
Jul 13 07:03:30 <kolla_> so with the default value, it refreshed constantly
Jul 13 07:04:06 <kolla_> went back to 2.6.10 and it was normal again
Jul 13 07:04:27 <kolla_> I dont have this problem on any of my own 2.6.12-machines though
Jul 13 07:15:53 <jbowler> VoodooZ_work: the merges happen because for some reason the 'common ancestor' is back from June 27.
Jul 13 07:16:13 <jbowler> I stared at it for about an hour and couldn't work out what was going on.
Jul 13 07:16:52 <jbowler> All the OE files (opie, zaurus) are actually unchanged on nslu2-linux and all the stuff like cron only has changes on nslu2-linux.
Jul 13 07:17:59 <jbowler> I did a diff -r --exclude=MT --brief org.openembedded.dev org.openembedded.nslu2-linux after the merge to check the result.
Jul 13 07:31:25 <VoodooZ_work> jbowler, really?
Jul 13 07:31:36 <VoodooZ_work> how come the others didn't get it though?
Jul 13 07:32:38 <jbowler> You got it, [cc]smart got it, NAiL got it, I got it - so far as I could see we all encountered the same merge or something very similar...
Jul 13 07:32:41 <VoodooZ_work> jbotwler, so what do you suggest I do to recover?
Jul 13 07:33:02 <VoodooZ_work> ok. rwhitby-away didn't get it so I thought I was the freak! :)
Jul 13 07:33:42 <NAiL> uhm, no. I didn't get it
Jul 13 07:34:19 <jbowler> Oh, ok...  there's something I don't understand going on.
Jul 13 07:34:51 <jbowler> My fix was to install 'meld' - it's far from perfect but it's much easier to use than vimdiff.
Jul 13 07:35:11 <VoodooZ_work> ok. I'll do that too then.
Jul 13 07:35:19 <jbowler> The best solution I know is to use kdiff3, but that means rewriting the LUA hook (kdiff3 is the default in monotone 0.20)
Jul 13 07:35:43 <jbowler> For all those OE merges it is just a matter of taking the OE file.
Jul 13 07:36:26 <jbowler> But meld doesn't make that very easy - with kdiff3 it can be done by simply accepting all the changes from the correct side.
Jul 13 07:36:31 <VoodooZ_work> yeah, that's the problem with vim, I couldn't even tell what was what.
Jul 13 07:36:47 <VoodooZ_work> How hard is kdiff3 install?
Jul 13 07:37:24 <jbowler> It needs kde...  If you have kde it's one of the development tools, if you don't it is a big deal.
Jul 13 07:37:45 <VoodooZ_work> yeah, I used KDE under FC4.
Jul 13 07:38:07 <jbowler> Anyway, with 0.19 it needs a LUA hook change too.
Jul 13 07:38:47 <jbowler> (The LUA hook could actually be made to do the same thing as bk - i.e. give a three way choice of left file, right file or merge and only invoke the merge tool on the latter).
Jul 13 07:39:16 <VoodooZ_work> yeah. that would be sweet!
Jul 13 07:40:10 <VoodooZ_work> meld installed.
Jul 13 07:40:53 <VoodooZ_work> Now I wonder if it's not too late for me as I got pissed and started saving the left file (???) for all merges.
Jul 13 07:41:02 <VoodooZ_work> And they kept coming...
Jul 13 07:41:24 <jbowler> Kill the update before it completes, start again.
Jul 13 07:41:48 <jbowler> I had about five OE files to merge.
Jul 13 07:42:23 <VoodooZ_work> ok, then what? using meld, what should I do?
Jul 13 07:42:51 <jbowler> Just run it again, it should use meld in preference to vim.
Jul 13 07:43:20 <jbowler> And now you will be able to see the messages which say which file is which :)
Jul 13 07:44:14 <VoodooZ_work> It says I should merge all changes to center.
Jul 13 07:44:50 <VoodooZ_work> Already looks much much better than vi!
Jul 13 07:45:16 <jbowler> It's fairly neat but I found it has bugs - sometimes clicking on the arrows to transfer the changes doesn't work.
Jul 13 07:45:31 <VoodooZ_work> ok. I'll keep an eye open for that.
Jul 13 07:46:16 <VoodooZ_work> so for the openzaurus merge for example, I just make sure the center copy has the latest changes right?
Jul 13 07:47:31 <jbowler> I think the left one is 'incoming' and the right is, effectively, your local version, if so left and center should end up identical (with no changes to the left - only change the center).
Jul 13 07:50:53 <VoodooZ_work> strange, the undo didn't seem to work.
Jul 13 07:52:41 <jbowler> I was able to ^Z all the way back to no changes.
Jul 13 07:53:12 <jbowler> I found I could also use 'cp' on the temporary file names (simply overwrite the merge file with the correct version).
Jul 13 07:53:13 <VoodooZ_work> must be me.
Jul 13 07:53:27 <VoodooZ_work> good
Jul 13 07:53:41 <jbowler> This is, of course, horribly wrong - it shouldn't be this hard.
Jul 13 07:53:56 <VoodooZ_work> should the CVSDATE line of openszaurus.conf be commented out now?
Jul 13 07:54:21 <jbowler> You get that too...  Yes, and it should be in 200507
Jul 13 07:54:32 <VoodooZ_work> ok. Just wanted to make sure.
Jul 13 07:55:00 <VoodooZ_work> Once I saved I just quit? What's the STOP button for?
Jul 13 07:55:04 <jbowler> This doesn't make sense because I hit precisely that merge 8 hours ago - and the result is in the nslu2-linux branch and dev.
Jul 13 07:55:31 <jbowler> The process has to exit, apparently, so quit after the save.
Jul 13 07:55:49 <VoodooZ_work> uhoh. If you're confused then I'm definitely lost! ;)
**** BEGIN LOGGING AT Wed Jul 13 07:58:57 2005
Jul 13 08:02:08 <VoodooZ_work> wow! how many merges did you get? It won't stop! 
Jul 13 08:02:30 <jbowler> 5 from .dev 5 from the nslu2-linux side
Jul 13 08:03:03 <VoodooZ_work> ok
Jul 13 08:04:26 <jbowler> But afterward the org.openembedded.dev and org.openembedded.nslu2-linux heads should differ in only the following files:
Jul 13 08:04:27 <VoodooZ_work> did you get the xinetd.init one? 
Jul 13 08:04:31 <jbowler> Files org.openembedded.dev/.mt-attrs and org.openembedded.nslu2-linux/.mt-attrs differ
Jul 13 08:04:31 <jbowler> Only in org.openembedded.dev/packages/cron/cron-3.0pl1: install-sh
Jul 13 08:04:31 <jbowler> Files org.openembedded.dev/packages/cron/cron_3.0pl1.bb and org.openembedded.nslu2-linux/packages/cron/cron_3.0pl1.bb differ
Jul 13 08:04:32 <jbowler> Only in org.openembedded.dev/packages/cyrus-imapd/cyrus-imapd-2.2.12: install-sh
Jul 13 08:04:32 <jbowler> Files org.openembedded.dev/packages/cyrus-imapd/cyrus-imapd_2.2.12.bb and org.openembedded.nslu2-linux/packages/cyrus-imapd/cyrus-imapd_2.2.12.bb differ
Jul 13 08:04:34 <jbowler> Only in org.openembedded.nslu2-linux/packages: install
Jul 13 08:04:36 <jbowler> Files org.openembedded.dev/packages/meta/openslug-packages.bb and org.openembedded.nslu2-linux/packages/meta/openslug-packages.bb differ
Jul 13 08:04:39 <jbowler> Files org.openembedded.dev/packages/mgetty/mgetty_1.1.30.bb and org.openembedded.nslu2-linux/packages/mgetty/mgetty_1.1.30.bb differ
Jul 13 08:05:31 <jbowler> xinetd.init is an nslu2-linux change, not sure whether I got it.
Jul 13 08:05:59 <jbowler> BTW what command did you run - 'sync' or 'pull' (or is this an explicit merge because you got two heads?)
Jul 13 08:06:14 <VoodooZ_work> I use the Makefile's 'make update'.
Jul 13 08:06:27 <VoodooZ_work> sync is bi-directional so I wouldn't use it.
Jul 13 08:07:15 <VoodooZ_work> it seems to have stopped. But it's still thinking...
Jul 13 08:07:35 <jbowler> It's waiting for you to blink ;-)
Jul 13 08:07:49 <VoodooZ_work> hehehe.
Jul 13 08:08:03 <VoodooZ_work> Actually, I got this: enter passphrase for key ID [stephane641@rogers.com]:
Jul 13 08:08:03 <VoodooZ_work> monotone: misuse: empty passphrase not allowed
Jul 13 08:08:03 <VoodooZ_work> make: *** [update-openembedded] Error 1
Jul 13 08:08:15 <VoodooZ_work> Darn! I must have pressed enter too early.
Jul 13 08:08:25 <VoodooZ_work> so I have to redo the whole thing.
Jul 13 08:11:39 <jbowler> The Makefile (at least the current version) does 'monotone update' followed, sometimes, by 'monotone merge' - were you in 'update' or 'merge'?
Jul 13 08:12:25 <VoodooZ_work> euh? I only do make update to pull and update everything. rwhitby-away said it was the right way.
Jul 13 08:12:59 <VoodooZ_work> Seems to have worked. It's updating now...
Jul 13 08:13:01 <jbowler> Yes, but the Makefile executes various monotone commands - I'm interested in knowing when you hit the merge.
Jul 13 08:13:22 <VoodooZ_work> I hit the merge prior to the update. Probably during the pull.
Jul 13 08:14:07 <jbowler> Pull is non-interactive.  I got the wrong line in the Makefile - you must be in the explicit 'merge'
Jul 13 08:14:25 <jbowler> monotone merge -b org.openembedded.nslu2-linux
Jul 13 08:15:02 <VoodooZ_work> yeah. it looks like it. there's an if statement that looks for different heads or seomthign
Jul 13 08:15:35 <jbowler> So you have local changes commited to the nslu2-linux branch which are not pushed back to the server, that creates the multiple-heads situation.
Jul 13 08:16:31 <VoodooZ_work> yeah, most of my changes are specific to my apps so I'll never push them. But none of them were even related to xinetd and zaurus...
Jul 13 08:17:10 <jbowler> Then the spurious files happen because the common ancestor is a common ancestor of all the files - a point where the branch (remote and local) was identical.
Jul 13 08:17:30 <VoodooZ_work> I see. Still pretty scary though.
Jul 13 08:18:01 <VoodooZ_work> It just happened suddenly although I've been committing and pulling on an hourly basis almost
Jul 13 08:18:43 <jbowler> Yes.  That's one thing I still don't understand - why it suddenly happened...
Jul 13 08:19:48 <VoodooZ_work> Probably somebody messing with it at the other end?? OE?
Jul 13 08:20:26 <jbowler> I couldn't see any evidence - I thought maybe no newline at the ends of files, or extra blank lines, but all the files are identical (where they should be identical).
Jul 13 08:20:38 <VoodooZ_work> yep.
Jul 13 08:24:39 <jbowler> Oh well, bbl - more thought required ;-)
Jul 13 08:25:23 <VoodooZ_work> yep. Thanks for the great help. Much appreciated.
Jul 13 08:27:42 <kolla_> hm, when I try to use bitbake, it says "ERROR: Unable to open conf/bitbake.conf"
Jul 13 08:35:52 <DaKa2> kolla_: did you source setup-env?
Jul 13 08:38:52 <kolla_> ah, openslug/setup-env
Jul 13 08:38:55 <kolla_> nope, thanks
Jul 13 08:39:00 <DaKa2> :)
Jul 13 08:40:07 <kolla_> seems like the latest updates fixed the bdb-problems with postfix?
Jul 13 08:48:06 <jf-logger> 
ACTION 

Jul 13 09:29:44 <DaKa2> finally one of my slugs have ForcePowerAlwaysOn and ASerialPort
Jul 13 09:40:06 <CIA-8> 03jp30 * 10unslung/make/abook.mk: strip
Jul 13 09:41:23 <Tiersten> DaKa2: cool
Jul 13 09:43:05 <DaKa2> now there is just playing with i2c left
Jul 13 09:43:21 <Tiersten> You got the I2C bus extender chips?
Jul 13 09:43:26 <DaKa2> yup
Jul 13 09:43:48 <DaKa2> and some other fun stuff, like i2c temp meters
Jul 13 09:44:09 <DaKa2> thermal glue and one on the cpu maybe? :-)
Jul 13 09:44:09 <Tiersten> You done the speed mod?
Jul 13 09:44:13 <DaKa2> yup
Jul 13 09:44:29 <DaKa2> on 2 of my 3 slugs
Jul 13 09:46:21 <DaKa2> oh my god, Im going to melt in here.. need something to drink, brb
Jul 13 09:47:44 <CIA-8> 03jp30 * 10unslung/make/adns.mk: strip
Jul 13 10:01:05 <VoodooZ_work> DaKa2, Which method did you use for ForcePOwerOn? I've been waiting to do it on my RoboSlug.
Jul 13 10:01:21 <DaKa2> I used a pic12f675
Jul 13 10:01:28 <DaKa2> with a small program
Jul 13 10:01:44 <DaKa2> ie, method my own :-)
Jul 13 10:01:47 <VoodooZ_work> i see.
Jul 13 10:01:50 <Tiersten> I was going to do something similar
Jul 13 10:01:59 <Tiersten> And then hook it up to the I2C bus as well
Jul 13 10:02:23 <VoodooZ_work> I use I2C to read a sensor array of 12 Sharp  rangers and a motor controller. All custom made with AVRs.
Jul 13 10:03:18 <DaKa2> I have 8 range sensors and a homebuilt motorcontroller with homebuilt PICs on i2c
Jul 13 10:03:24 <DaKa2> on my robot
Jul 13 10:03:34 <DaKa2> using an mini-itx right now
Jul 13 10:04:08 <VoodooZ_work> cool! I didn't know you were building a roboslug too!
Jul 13 10:04:13 <VoodooZ_work> Do you have a web page?
Jul 13 10:04:24 <DaKa2> nope.. it still cannot move :-)
Jul 13 10:04:32 <VoodooZ_work> mine is at http://robotics.no-ip.org but there's not much on my roboslug yet.
Jul 13 10:04:42 <Tiersten> Put it online
Jul 13 10:04:44 <DaKa2> havn't really finished the motorcontroller, just a bit left
Jul 13 10:04:51 <Tiersten> And mount a supersoaker
Jul 13 10:04:56 <Tiersten> Then let us drive it around
Jul 13 10:05:06 <DaKa2> well, its loosly based on oap.sf.net
Jul 13 10:05:10 <DaKa2> :-)
Jul 13 10:05:17 <VoodooZ_work> I see.
Jul 13 10:05:36 <DaKa2> I just keep making things differently
Jul 13 10:05:37 <VoodooZ_work> My old robot had a paintball gun and was fully autonomous.
Jul 13 10:05:45 <DaKa2> haha
Jul 13 10:05:51 <VoodooZ_work> our robotic club called it SplatBot!
Jul 13 10:06:09 <DaKa2> :)
Jul 13 10:06:12 <kolla_> :)
Jul 13 10:06:18 <VoodooZ_work> Quite a mess to clean after the fights!
Jul 13 10:06:31 <VoodooZ_work> www.ottawarobotics.org
Jul 13 10:06:57 <DaKa2> http://david.thg.se/bilder/robot/040806/IMG_1198.JPG <-- old picture of my prototype
Jul 13 10:07:24 <DaKa2> <-- not a proffesional
Jul 13 10:08:03 <VoodooZ_work> Me neither! Looks good! I love the prototype look. You'll never see one of my robots with cutesy eyes and stuff.
Jul 13 10:08:34 <VoodooZ_work> My latest looks more professional because I took a long time to design it and had a friend cut the lexan sheets.
Jul 13 10:08:49 <DaKa2> haha, well, I like my robots big, smart, and ugly
Jul 13 10:08:53 <VoodooZ_work> hehehe
Jul 13 10:08:55 <kolla_> hehe.. biltema-batterier :)
Jul 13 10:09:04 <DaKa2> yup :-)
Jul 13 10:09:04 <Tiersten> You trying to give me neck ache or something?
Jul 13 10:09:16 <VoodooZ_work> Mine is kind small because my goal is return to the Trinity firefighting robot competition next april
Jul 13 10:09:21 <VoodooZ_work> hehe
Jul 13 10:09:23 <DaKa2> VoodooZ_work: I really like SRX2
Jul 13 10:09:27 <VoodooZ_work> thanks.
Jul 13 10:09:45 <VoodooZ_work> I put a lot more thought into it this time around.
Jul 13 10:09:54 <VoodooZ_work> You learn something with every iteration
Jul 13 10:10:11 <VoodooZ_work> SRX2 is the one using the slug.
Jul 13 10:10:22 <VoodooZ_work> I should really update my web page.
Jul 13 10:10:57 <VoodooZ_work> btw, how can you do image capture of a window in linux?
Jul 13 10:11:07 <DaKa2> if I ever get this one finished I will make a much better one, I have 200m2 with machines for manifacturing stuff like this
Jul 13 10:11:13 <kolla_> VoodooZ_work: in X?
Jul 13 10:11:20 <DaKa2> hm.. ksnapshot, or scrot
Jul 13 10:11:26 <kolla_> VoodooZ_work: or import
Jul 13 10:11:32 <DaKa2> yup..
Jul 13 10:11:45 <VoodooZ_work> cool.
Jul 13 10:11:53 <VoodooZ_work> kolla_, thanks.
Jul 13 10:11:55 <kolla_> seems like it's time to go home...
Jul 13 10:11:57 <Tiersten> gimp does it as well I think
Jul 13 10:11:58 <kolla_> bbl :)
Jul 13 10:12:12 <VoodooZ_work> cool. I guess I'll updat my web page with my vision GUI now then...
Jul 13 10:12:39 <VoodooZ_work> brb
Jul 13 10:16:19 <DaKa2> wtf.. does openslug use dhcp even if you set an ip?
Jul 13 10:17:02 <DaKa2> not strange it gets permission denied for mounting the root nfs
Jul 13 10:18:14 <DaKa2> but it comes up with the correct ip later..
Jul 13 10:18:29 <CIA-8> 03jp30 * 10unslung/make/expat.mk: stage libexpat.la, but remove host paths first
Jul 13 10:18:32 <DaKa2> so.. it runs dhcp, tries to mount nfs, and then sets my selected ip..
Jul 13 10:33:20 <kolla__> *tap tap*
Jul 13 10:34:24 <Tiersten> ?
Jul 13 10:34:38 <kolla_> --disable-docdir didnt do much for libpam, removing openjade did
Jul 13 10:35:15 <kolla_> Tiersten: I just pulled my laptop with me from work, wasnt sure if konversation had reconnected to the servers properly :)
Jul 13 10:35:24 <Tiersten> ah
Jul 13 10:43:54 <DaKa2> anybody know why openslug uses dhcp even when you have setup a static ip?
Jul 13 10:44:21 <VoodooZ_work> there's a wiki page that explains this i believe.
Jul 13 10:44:50 <DaKa2> hm..
Jul 13 10:44:53 <VoodooZ_work> IIRC there's an invisible file somewhere that can be removed/added to override the auto-config.
Jul 13 10:45:01 <DaKa2> I did search..
Jul 13 10:45:03 <VoodooZ_work> search for sysconf
Jul 13 10:48:31 <DaKa2> I can only find how to use dhcp.. not the other way
Jul 13 10:48:54 <[g2]> you're using openslug right ?
Jul 13 10:49:05 <[g2]> look in /etc/network/interfaces
Jul 13 10:49:09 <DaKa2> yup
Jul 13 10:49:34 <DaKa2> yes?
Jul 13 10:49:50 <DaKa2> "iface eth0 inet static"
Jul 13 10:49:54 <DaKa2> no dhcp there..
Jul 13 10:50:01 <VoodooZ_work> odd
Jul 13 10:50:03 <[g2]> with and address
Jul 13 10:50:07 <DaKa2> yup
Jul 13 10:50:07 <[g2]> an
Jul 13 10:50:13 <DaKa2> it sets the ip later in the boot
Jul 13 10:50:29 <DaKa2> just uses dhcp, tries to mount my nfs root, and then sets the ip
Jul 13 10:50:50 <DaKa2> and since my nfs server is set to the slugs ip it fails
Jul 13 10:50:52 <[g2]> ahh.. you're booting via nfs
Jul 13 10:50:56 <DaKa2> yup..
Jul 13 10:51:08 <DaKa2> it works without dhcp server..
Jul 13 10:51:31 <DaKa2> but why the hell does it even run udhcpc?
Jul 13 10:52:13 <[g2]> dunno, I think mine only runs with dhcp
Jul 13 10:52:48 <[g2]> I haven't played with nfs booting, doesn't that use bootp (and dhcp) ?
Jul 13 10:53:03 <DaKa2> no.. not at home atleast
Jul 13 10:53:08 <DaKa2> where I have no dhcp server
Jul 13 10:53:27 <jbowler-away> DaKa2: run turnup init
Jul 13 10:53:52 <DaKa2> I have..
Jul 13 10:53:59 <DaKa2> first thing I did
Jul 13 10:54:04 <DaKa2> and then turnup nfs
Jul 13 10:54:08 <DaKa2> same as I did at home
Jul 13 10:54:15 <DaKa2> where it works :-)
Jul 13 10:54:43 <[g2]> oh what do the kernel messages say ?
Jul 13 10:54:57 <DaKa2> well, the interesting part:
Jul 13 10:55:02 <DaKa2> udhcpc[211]: Lease of 10.39.21.32 obtained, lease time 604800
Jul 13 10:55:02 <DaKa2> boot: rootfs: mount -t nfs -o nolock,noatime,hard,intr,rsize=1024,wsize=1024 10.39.19.22:/data/slug3
Jul 13 10:55:05 <DaKa2> mount: 10.39.19.22:/data/slug3 failed, reason given by server: Permission denied
Jul 13 10:55:23 <DaKa2> when I loging it has the right ip 10.39.19.81
Jul 13 10:55:30 <jbowler-away> Ah, ok - nfs.  So /boot/nfs before the rootfs is there.
Jul 13 10:56:26 <DaKa2> but why even run udhcpd? /etc/default/functions line 207?
Jul 13 10:56:39 <jbowler-away> So how does it get a lease if there is no DHCP server?
Jul 13 10:56:49 <DaKa2> there is one here
Jul 13 10:57:00 <DaKa2> Im at work, but at home there is none, and it works
Jul 13 10:57:25 <[g2]> it's guess it times out
Jul 13 10:57:31 <[g2]> i'd 
Jul 13 10:58:00 <DaKa2> I could probably ssh home to my slug there and check what it gets
Jul 13 10:58:09 <jbowler-away> It tries once for the DHCP server, if that fails it falls back to the static ip
Jul 13 10:58:29 <DaKa2> but if I have set a static ip I want that...
Jul 13 10:59:58 <jbowler-away> Yes, that's a bug in the bootstrap approach... slugbug
Jul 13 11:01:35 <jbowler-away> It was like that before because bootproto was unreliable, but with turnup init, and given that /boot/network is only run after turnup init has been done, it should be safe to follow it now.
Jul 13 11:01:58 <DaKa2> is there a bug for it, or should I file one?
Jul 13 11:02:58 <jbowler-away> There isn't one - yes, please enter it.
Jul 13 11:09:52 <CIA-8> 03jbowler 07org.openembedded.nslu2-linux * r2117dd5b... 10/ (14 files in 11 dirs): 
Jul 13 11:09:53 <CIA-8> propagate from branch 'org.openembedded.dev' (head 026dfb76437b8ee365af6ed7b1a122bd595d202f)
Jul 13 11:09:53 <CIA-8>  to branch 'org.openembedded.nslu2-linux' (head 43007d439c7ed96133926d0748cef53fa1b321d0)
Jul 13 11:13:23 <jbowler-away> [cc]smart: good changes in turnup, looks like memstick is now fully supported!  I found a couple of problems, slugbugs 192 and 193.  191 is also about the same change but I think I'm wrong (see the bug).  193 is the really significant problem.
Jul 13 11:17:04 <[cc]smart> ok, just read 193. i will be off for 10 minutes. can we discuss this in 10 minutes ?
Jul 13 11:17:42 <[cc]smart> i assume this means yes :)
Jul 13 11:18:56 <jbowler-away> yes
Jul 13 11:35:54 <[cc]smart> back
Jul 13 11:36:03 <[cc]smart> jbowler-away: ping
Jul 13 11:36:31 <jbowler-away> pong
Jul 13 11:37:40 <[cc]smart> when i'm going to mess with populate-var i'd try to change it structurally as to make it work with a config file
Jul 13 11:37:53 <[cc]smart> sounds reasonable/worthwhile ?
Jul 13 11:38:19 <[cc]smart> so that populate-var become like populate-volatile-fs
Jul 13 11:38:22 <jbowler-away> yes - that sounds like a good approach
Jul 13 11:39:14 <[cc]smart> by this we would get populate-var out of this bb file too which i think is hoorible :)
Jul 13 11:39:32 <[cc]smart> now, to check if i understood your suggestion coirrectly
Jul 13 11:39:50 <[cc]smart> populate-volatile should - read from a list of configs
Jul 13 11:39:54 <[cc]smart> check for eixtence
Jul 13 11:39:58 <jbowler-away> It's somewhat odd.  I can't see what OE would have any problems with using a config file, just so long as the intelligence in the .bb isn't lost.
Jul 13 11:40:13 <[cc]smart> if not existant it should create what is defined in the configfile
Jul 13 11:40:25 <[cc]smart> if a target is found it should check if it's a symbolic link
Jul 13 11:40:34 <jbowler-away> (I haven't looked at what the bb file actually does, because seeing that it did something fairly gross...)
Jul 13 11:40:41 <jbowler-away> s/because/beyond/
Jul 13 11:40:43 <[cc]smart> if it's a symbolic link it should create what is defined in the configfile in the target position of the link
Jul 13 11:40:45 <[cc]smart> taht correct ?
Jul 13 11:41:07 <jbowler-away> I think so, a bit like makedevs and /etc/device_table.txt
Jul 13 11:41:38 <[cc]smart> yes i think that's what o understood from device_table.txt
Jul 13 11:41:46 <[cc]smart> s/o/i/
Jul 13 11:41:54 <[cc]smart> will try that
Jul 13 11:42:03 <jbowler-away> So is the config file just one file, with say a list of directories, or something like /etc/volatile.d/*.conf
Jul 13 11:42:27 <jbowler-away> (The latter approach has the advantage that packages can add directories to /var)
Jul 13 11:42:31 <[cc]smart> now, since it's a change that hits all OE somewhere in a central area, who/how would it be introduced ?
Jul 13 11:42:57 <[cc]smart> i think it should be in /etc/default or such
Jul 13 11:43:05 <[cc]smart> could be a file like:
Jul 13 11:43:29 <[cc]smart> <type of target> <owner> <group> <mode> <target>
Jul 13 11:43:31 <jbowler-away> Write it, put it somewhere (pastebin for example) and then discuss it with mickeyl/koen (i.e. at least one OE core dev) on #oe
Jul 13 11:43:58 <jbowler-away> I don't think they will have any problem with it, given the current approach ;-)
Jul 13 11:43:59 <[cc]smart> one target per line
Jul 13 11:44:24 * [g2] notes mickeyl is hanging out in #openslug btw
Jul 13 11:44:34 <jbowler-away> Right, that format is very close to device_table.txt.  Another approach would be something like:
Jul 13 11:44:40 <[cc]smart> does he copy nslu2-linux ?
Jul 13 11:44:57 <jbowler-away>  /var/tmp volatile <owner> <group>...
Jul 13 11:45:11 <jbowler-away>  /var/lib/nfs persistent <owner> <group>...
Jul 13 11:45:16 <[g2]> I'd guess not as he gave up another group to join #nopenslug
Jul 13 11:45:25 <[cc]smart> what does volatile/persistent in this case mean ?
Jul 13 11:45:51 <jbowler-away> Wether it goes in tmpfs or in some form of persistent storage.
Jul 13 11:46:05 <[cc]smart> ah, i think this is not necessary
Jul 13 11:46:11 <jbowler-away> So if it's volatile and /var is on disk then a symlink to /media/ram is needed
Jul 13 11:46:26 <[cc]smart> if it's persistent, it exists alread or, is newly created and becomes persistent
Jul 13 11:46:29 <jbowler-away> Otherwise you have to check for a symlink - which might be more difficult.
Jul 13 11:46:41 <[cc]smart> if it doesn' exist, it's because it hasn't been created or it is volatile
Jul 13 11:47:01 <[cc]smart> i'd like to check for symlink
Jul 13 11:47:03 <jbowler-away> Yes.  I think that works.
Jul 13 11:47:24 <[cc]smart> makes it more easy to use i think
Jul 13 11:47:27 <[cc]smart> and more flexible
Jul 13 11:47:55 <jbowler-away> The symlink check is quite easy too.
Jul 13 11:48:00 <[cc]smart> ok. i'll try that
Jul 13 11:48:18 <[cc]smart> do you have a specific suggestion for the check ?
Jul 13 11:48:26 <jbowler-away> Take a look at what I had to do in /etc/init.d/nfsserver - it goes and creates /var/lib/nfs every time (i.e. just like populate-vars.sh)
Jul 13 11:48:56 <jbowler-away> The problem with the symlink check is that not all OE distros have 'readlink' built in.
Jul 13 11:49:33 <jbowler-away> So the easy one is MD(){; if test -h "$1"; then MD "$(readlink "$1")"; else mkdir -p "$1"; }
Jul 13 11:50:16 <jbowler-away> But if readlink isn't there it becomes a sed of the output of ls -l
Jul 13 11:50:48 <jbowler-away> rl(){; readlink 
Jul 13 11:51:22 <jbowler-away> "$@" 2>/dev/null || ls -l "$@" | sed -e 's/^.*-> \(.*\)$/\1/'; }
Jul 13 11:51:54 <[cc]smart> ok... i'l have a try...
Jul 13 11:52:15 <jbowler-away> That latter can be made safer...  for l in "$@"; do if test -h "$l" etc
Jul 13 11:53:09 <jbowler-away> How about using /etc/default/*.tmpdir?  That way packages can add their own definition.
Jul 13 11:53:59 <jbowler-away> Also turnup can more safely overwrite the default file - /etc/default/var.tmpdir?
Jul 13 11:57:07 <[g2]> mgetty is still borked for me
Jul 13 11:57:17 <[g2]> cp: cannot create regular file `/usr/share/info/_inst.20278_': Permission denied
Jul 13 11:57:45 <jbowler-away> Why is it trying to copy to /?
Jul 13 11:57:58 <[g2]> hence the borkage
Jul 13 11:58:31 <jbowler-away> It looks like a ${D} has got lost, but it's certainly not a general problem.
Jul 13 11:59:42 <[g2]> you mean my mgetty is specail ?
Jul 13 11:59:56 <jbowler-away> Well, why is it trying to write to /?
Jul 13 12:00:14 <[g2]> it's a stright pull from the repo
Jul 13 12:00:18 <[g2]> straight
Jul 13 12:00:30 <jbowler-away> That was a real question...
Jul 13 12:00:56 <VoodooZ_work> DaKa2, Please let me know how your I2C testing goes with the slug as I keep getting issues with mine.
Jul 13 12:01:15 <DaKa2> ok, when I get time :)
Jul 13 12:01:37 <[g2]> md5sum openembedded/packages/mgetty/mgetty_1.1.30.bb
Jul 13 12:01:37 <[g2]> b8b1dea9e4b7542fa41917a0d23fe0b7  openembedded/packages/mgetty/mgetty_1.1.30.bb
Jul 13 12:01:42 <VoodooZ_work> DaKa2, I think it might be a race condition that leaves my I2C modules irresponsive. I'll try booting the Modules/Slug in different order to see if it helps.
Jul 13 12:02:55 <DaKa2> ugh...
Jul 13 12:06:31 <jbowler-away> So far as I can see in my mgetty builds nothing uses 'infodir'.  It's in run.* but not used, so that's probably your problem.
Jul 13 12:06:50 <jbowler-away> bbl
Jul 13 12:06:58 <[g2]> do you have the same md5 ?
Jul 13 12:07:01 <[g2]> see you
Jul 13 12:08:27 <jbowler-away> I think so (same md5)  I'm at 2117dd5b726765202daffe67318297ef78df5a4f, no changes
Jul 13 12:09:07 <[g2]> thx
Jul 13 12:17:40 <[cc]smart> jbowler-away: i thought about the recursion and currently i'm not sure if this is really needed
Jul 13 12:18:29 <[cc]smart> do you see an actual situation where a link should follow behind another one ?
Jul 13 12:19:20 <[cc]smart> i feel the first link in a hierarchy will perform a switch into a tmpfs or wherever, and what is beyond there is either volatile, and the whole stuff below with it and this can be created there
Jul 13 12:19:42 <[cc]smart> or static and already there
Jul 13 12:23:34 <[cc]smart> for now i assume this recursion is not needed.
Jul 13 12:23:53 <[cc]smart> i'm relatively sure about that,
Jul 13 12:42:13 <[g2]> NAiL, heh :)
Jul 13 12:42:16 <[g2]> http://groups.yahoo.com/group/nslu2-linux/message/7416
Jul 13 12:42:21 <NAiL> yes
Jul 13 12:42:25 <NAiL> push them symlinks?
Jul 13 12:42:25 <[g2]> audio should be audiofile
Jul 13 12:42:37 <[g2]> yes and add mpd
Jul 13 12:42:42 <[g2]> to openslug packages
Jul 13 12:42:51 <[g2]> it builds
Jul 13 12:42:55 <NAiL> neat
Jul 13 12:43:34 <VoodooZ_Log> cool
Jul 13 12:43:41 <VoodooZ_Log> What do you use as sound card?
Jul 13 12:44:06 <CIA-8> 03repvik * r80 10/trunk/openslug/nslu2-linux/packages/ (audiofile flac libao libmad libmikmod libogg libvorbis mpd): Added symlinks for mpd
Jul 13 12:45:16 <[g2]> NAiL, you just needed mpd
Jul 13 12:45:29 <VoodooZ_Log> I might want to add a USB sound card to my roboslug later for speech synt or simply to scare my cat. Any cheap one to recommend?
Jul 13 12:45:43 <[g2]> dunno
Jul 13 12:46:18 <VoodooZ_Log> so that mpd thing is only used for streaming then right?
Jul 13 12:46:19 <NAiL> [g2]: heh, ok. 
Jul 13 12:46:51 <[g2]> I think playback
Jul 13 12:47:49 <CIA-8> 03repvik * r81 10/trunk/openslug/nslu2-linux/packages/ (audiofile flac libao libmad libmikmod libogg libvorbis): Back out all but mpd ;-)
Jul 13 12:48:00 <DaKa2> anyone working on bug 196? That should be why my whole rebuild fails on postfix..
Jul 13 12:48:49 <CIA-8> 03nail 07org.openembedded.nslu2-linux * r2913dd5a... 10/packages/meta/openslug-packages.bb: Added mpd
Jul 13 12:49:45 <[g2]> DaKa2, hmmm postfix builds for me
Jul 13 12:49:58 <[g2]> but mgetty doesn't :)
Jul 13 12:50:11 <DaKa2> mine doesnt, because pcre dosnt really build correctly in some cases
Jul 13 12:50:21 <DaKa2> :)
Jul 13 12:50:25 <DaKa2> my mgetty works...
Jul 13 12:50:28 <DaKa2> or, builds
Jul 13 12:54:39 <NAiL> [g2]: ah, all the symlinks actually were there for a reason
Jul 13 12:54:54 <[g2]> yes :)
Jul 13 12:54:55 <NAiL> It has a *bunch* of deps 
Jul 13 12:55:11 * NAiL tries again
Jul 13 12:55:21 <[g2]> what did you miss ?
Jul 13 12:56:47 <NAiL> You said I just needed mpd, so I thought the rest of the links weren't needed
Jul 13 12:57:52 <[g2]> ah.. no.. all the links and just mpd in the packages
Jul 13 12:58:04 <NAiL> yeah
Jul 13 12:58:46 <NAiL> ;-)
Jul 13 12:58:48 <CIA-8> 03repvik * r82 10/trunk/openslug/nslu2-linux/packages/ (audiofile flac libao libmad libmikmod libogg libvorbis): Confusion. Now cleared. Links really *are* needed for mpd
Jul 13 13:01:25 <[g2]> do you want to send the e-mail that the package as been added to openslug-packages or should I
Jul 13 13:01:38 <NAiL> I'll do it. It'll be a laugh ;)
Jul 13 13:01:57 <[g2]> and just note that the package audio is really audiofile
Jul 13 13:02:05 <[g2]> THX for all the help
Jul 13 13:03:12 <DaKa2> [g2]: that script for reading the buildlog and create wiki?
Jul 13 13:03:17 <NAiL> what am I missing here?
Jul 13 13:03:42 <DaKa2> probably something important..
Jul 13 13:03:44 <[g2]> DaKa2, one sec
Jul 13 13:03:58 <NAiL> [g2]: You say package "audio" is "audiofile". But the symlink is audiofile. 
Jul 13 13:04:49 <NAiL> What am I missing here?
Jul 13 13:04:55 * NAiL repeats himself
Jul 13 13:05:03 <NAiL> *then* notes he has repeated himself
Jul 13 13:05:41 * DaKa2 notes that he has no idea
Jul 13 13:06:16 <DaKa2> since the original message clearly said audiofile...
Jul 13 13:07:01 <NAiL> yeah
Jul 13 13:07:32 <NAiL> building audiofile right now
Jul 13 13:08:15 <[g2]> DaKa2, see the bottom of http://www.nslu2-linux.org/wiki/OpenSlug/PackageStatus
Jul 13 13:09:10 <NAiL> Ok, it compiles all right
Jul 13 13:09:17 <NAiL> Now I'm off to see a movie. bbl
Jul 13 13:09:27 <DaKa2> have fun :-)
Jul 13 13:09:46 <DaKa2> now I only need a complete build log..
Jul 13 13:10:25 <[g2]> Ok... look again
Jul 13 13:10:37 <[g2]> except for the indent issue I think that's right
Jul 13 13:10:48 <DaKa2> heh, I copied from edit anyway
Jul 13 13:11:26 <DaKa2> all my builds the last days have failed on something
Jul 13 13:11:31 <[g2]> you can run that on shortened output and you get the results
Jul 13 13:11:44 <[g2]> abbreviated thought they may be
Jul 13 13:12:53 <[g2]> Is the monotone head messed up ?
Jul 13 13:13:19 <DaKa2> not now, except me not getting postfix to build
Jul 13 13:14:07 <jbowler-away> VoodooZ_Log: http://www.dcs-online.biz/shop/product_info.php/products_id/4875 (Yoga AD-100)
Jul 13 13:15:26 <jbowler-away> [cc]smart: I agree using mkdir -p directly should always work, I just find the recursive form easier to program and more natural.
Jul 13 13:18:06 <jbowler-away> Daka2: the pcre problem is slugbug #196.  I believe it's a make time stamp issue (but I may be wrong), it's intermittent - fails about 1 in 3 for me.
Jul 13 13:18:33 <DaKa2> hm.. has failed 5/5 times for me, on two build hosts
Jul 13 13:18:45 <jbowler-away> Ah, ok - you can reproduce it ;-)
Jul 13 13:19:02 <DaKa2> 30 mins ago: 21:48 < DaKa2> anyone working on bug 196? That should be why my whole rebuild fails on postfix..
Jul 13 13:19:27 <DaKa2> well, if that is the problem
Jul 13 13:19:31 <jbowler-away> Try putting a sleep 1 before and after the BUILD_CC lines in do_compile.
Jul 13 13:20:11 <DaKa2> will do.. just tried a make update..
Jul 13 13:20:19 <jbowler-away> dftables.o gets build, then it gets built again inside the makefile.  The one outside is an attempt to fix the problem (the -DLINKS=2)
Jul 13 13:20:19 <DaKa2> monotone: trying 3-way merge
Jul 13 13:20:19 <DaKa2> monotone: misuse: no unique private key for cert construction
Jul 13 13:20:19 <DaKa2> make: *** [update-openembedded] Error 1
Jul 13 13:20:30 <[g2]> DaKa2, If got the same thing
Jul 13 13:20:36 <[g2]> DaKa2, I've got the same thing
Jul 13 13:20:39 <DaKa2> :(
Jul 13 13:20:49 <jbowler-away> Need to find out what is causing this...
Jul 13 13:20:59 <DaKa2> oh well, testingt he sleep 1 thing
Jul 13 13:27:45 <DaKa2> jbowler-away: are you sure postfix fails because of the pcre issue? sleep 1 didn't help, but the pcre build doesnt complain in any way
Jul 13 13:29:19 <DaKa2> or, its prce-native I should rebuild?
Jul 13 13:29:29 <DaKa2> pcre*
Jul 13 13:29:45 <[g2]> DaKa2, rm downloads/pcre...*
Jul 13 13:29:59 <[g2]> whatever the sourceball is and the .md5
Jul 13 13:30:10 <[g2]> then try to rebuild
Jul 13 13:30:12 <DaKa2> ok
Jul 13 13:30:33 <DaKa2> bb -cclean pcre-native pcre && bb pcre-native pcre postfix
Jul 13 13:30:40 <DaKa2> should work?
Jul 13 13:31:02 <[g2]> that good too, but it doesn't remove the downloaded file
Jul 13 13:31:11 <DaKa2> I removed that too
Jul 13 13:31:16 <[g2]> OK
Jul 13 13:32:16 <DaKa2> still fails
Jul 13 13:32:26 <DaKa2> "/home/slug/openslug/tmp/staging/i686-linux/bin/postconf: error while loading shared libraries: libpcre.so.0: cannot open shared object file: No such file or directory"
Jul 13 13:32:43 <DaKa2> but.. libpcre.so.0 exists..
Jul 13 13:36:56 <DaKa2> dunno where it tries to load in from though..
Jul 13 13:41:29 <[g2]> Do we need this: 
Jul 13 13:41:32 <[g2]> CONFIG_LEGACY_PTYS=y
Jul 13 13:41:32 <[g2]> CONFIG_LEGACY_PTY_COUNT=256
Jul 13 13:46:54 <CIA-8> 03marceln * 10unslung/sources/ipkg-web/package.cgi: Bug fix for package without comment
Jul 13 13:53:19 <senneth> monotone: verifying new revisions (this may take a while)
Jul 13 13:53:28 <senneth> how log does it usually take?
Jul 13 13:53:46 <dyoung> "a while" is like 15min+ for me.
Jul 13 13:54:02 <senneth> heh, I was running it for about 5hrs
Jul 13 13:54:08 <senneth> and it didn't finished
Jul 13 13:54:21 <ByronT-Desktop> dyoung!  morning
Jul 13 13:54:21 <dyoung> bummers!
Jul 13 14:03:46 <eno> senneth, r u running monotone on slug?
Jul 13 14:04:52 <jbowler-away> Daka2: that may be something else.  I had problems with pcre-native which was, I believe, built for postfix.  The shared library problem is LD_LIBRARY_PATH related - my build environment does an unset of LD_LIBRARY_PATH but it looks like setup-env no longer does this.
Jul 13 14:05:07 <jacques> or on a Pentium200 ?  :-)
Jul 13 14:05:27 <eno> i had one native session with "a while" == 2 days, although it eventually finished
Jul 13 14:05:31 <DaKa2> jbowler-away: o...k... any quick fix?
Jul 13 14:05:46 <jbowler-away> Try 'unset LD_LIBRARY_PATH' then make
Jul 13 14:06:20 <jbowler-away> It's trying to run something from staging, and the prce (native) .so is presumably there too.
Jul 13 14:06:32 <DaKa2> yup, it is
Jul 13 14:06:58 <senneth> eno: not really, it's a centrino laptop with 1,4GHz CPU and 512MB ram
Jul 13 14:07:08 <DaKa2> well, my LD_LIBRARY_PATH was empty, but iys unset now
Jul 13 14:07:09 <[cc]smart> jbowler-away: think i have it
Jul 13 14:07:10 <jbowler-away> Problem is on my system it's installed in /usr/lib too...
Jul 13 14:07:26 <DaKa2> ahh, not on mine, I think..
Jul 13 14:07:51 <jbowler-away> senneth: an initial setup of the database can take a very long time at the moment
Jul 13 14:08:04 <senneth> ic
Jul 13 14:08:26 <senneth> so i'll leave it running over the night
Jul 13 14:09:02 <jbowler> Hum: temp/log.do_compile.19374:/bin/sh: ./postconf: cannot execute binary file
Jul 13 14:09:57 <DaKa2> still no go..
Jul 13 14:10:00 <jbowler> But it built for me...
Jul 13 14:10:53 <jbowler> Yeah, well - the exec of postconf fails, but it is in a pipe so make doesn't care...
Jul 13 14:10:53 <DaKa2> actually, missed something, in the bottom of the do_compile log: "./postfix-install: Error: "" should be "no" or an absolute path name.", thats probably whats making it fail
Jul 13 14:12:08 <jbowler> Yes, I believe it is broken in the postconf step - it probably needs LD_LIBRARY_PATH on the command line - but it's not causing bb to fail.
Jul 13 14:13:54 <jbowler> What is the argument to postfix-install?  (You can see it in work/postfix*/temp/run.do_install*)
Jul 13 14:14:39 <DaKa2> " sh ./postfix-install 'install_root=/home/slug/openslug/tmp/work/postfix-2.0.20-r0/image' -non-interactive"
Jul 13 14:14:59 <jbowler> Hum - same as mine, and it apparently works.
Jul 13 14:15:14 <DaKa2> it seems to want some enviorment vars
Jul 13 14:15:44 <DaKa2> README_DIRECTORY seems to be the one failing
Jul 13 14:15:58 <DaKa2> (guessing)
Jul 13 14:19:01 <jbowler> Well, I can't see a way of getting the 'Error...' above with -noninteractive set, unless there is something really weird about sh on your system.
Jul 13 14:19:25 <DaKa2> on my systems..
Jul 13 14:19:31 <DaKa2> both debian unstable
Jul 13 14:19:41 <jbowler> Sorry, looking at the wrong line in the script...
Jul 13 14:19:53 <DaKa2> one really newly updated, one.. long time ago update
Jul 13 14:22:02 <DaKa2> where the hell does $readme_directory get set in tmp/work/postfix-2.0.20-r0/postfix-2.0.20/postfix-install ?
Jul 13 14:22:37 <jbowler> Line 417
Jul 13 14:23:06 <DaKa2> oh..
Jul 13 14:23:36 <jbowler> eval : ${readme_directory=`<postconf> -d -h readme_directory`}
Jul 13 14:23:47 <jbowler> So it is postconf failing on your system I think.
Jul 13 14:24:26 <DaKa2> could it be everybody else having libpcre in /usr/lib?
Jul 13 14:24:33 <[cc]smart> late enough now.... n8 all
Jul 13 14:24:39 <DaKa2> nite nite
Jul 13 14:25:51 <jbowler>  /home/work-tmp/jbowler/nslu2/openslug/staging/i686-linux/bin/postconf --help
Jul 13 14:26:06 <jbowler> That works for me and gives me a ('fatal') usage message.
Jul 13 14:26:36 <kolla_> woohoo.. the build went fine.. finally :)
Jul 13 14:26:43 <jbowler> ldd on that file tells me:
Jul 13 14:26:51 <jbowler>         libpcre.so.0 => /home/work-tmp/jbowler/nslu2/openslug/staging/i686-linux/lib/libpcre.so.0 (0xb7f17000)
Jul 13 14:27:13 <DaKa2> libpcre.so.0 => not found
Jul 13 14:27:33 <jbowler> So I think you have a libtool problem maybe...
Jul 13 14:27:45 <DaKa2> $ ls tmp/staging/i686-linux/lib/libpcre.so.0 -la
Jul 13 14:27:46 <DaKa2> lrwxrwxrwx  1 slug users 16 Jul 13 23:18 tmp/staging/i686-linux/lib/libpcre.so.0 -> libpcre.so.0.0.1
Jul 13 14:28:24 <jbowler> Same here.
Jul 13 14:28:26 <DaKa2> dammit...
Jul 13 14:28:37 <DaKa2> WHY! OH WHY!!!
Jul 13 14:30:10 <DaKa2> "LD_LIBRARY_PATH=/home/slug/openslug/tmp/staging/i686-linux/lib/ tmp/staging/i686-linux/bin/postconf --help" works
Jul 13 14:30:39 <kolla_> ah, that problem
Jul 13 14:30:41 <jbowler> Yes, but why on earth does it work on my system?
Jul 13 14:30:54 * kolla_ cheated
Jul 13 14:30:55 <DaKa2> yes, I would want to know that too
Jul 13 14:31:10 <DaKa2> kolla_?
Jul 13 14:31:33 <jbowler> I wonder if it still works with my new package/install - because do_stage for postfix-native just uses 'install'
Jul 13 14:31:48 <kolla_> I..uhm... put /home/slug/openslug/tmp/staging/i686-linux/lib in my /etc/ld.so.conf.. 
Jul 13 14:31:53 <NAiL> [g2]: ping?
Jul 13 14:31:59 <[g2]> NAiL, pong
Jul 13 14:32:10 <NAiL> Last year I've been running without legacy PTS
Jul 13 14:32:27 <NAiL> So I think it's safe to drop that support
Jul 13 14:32:52 <NAiL> IIRC, it'll save some 30-40k mem
Jul 13 14:32:53 <[g2]> ok... NAiL can you bop over to #openjtag ?
Jul 13 14:33:08 <[g2]> and a bunch of pid on boot from what I read in lak
Jul 13 14:33:15 <NAiL> sure
Jul 13 14:33:51 <jbowler> monotone: note: branch 'org.openembedded.nslu2-linux' has multiple heads
Jul 13 14:34:04 <DaKa2> kolla_: can that really be good?
Jul 13 14:34:11 <DaKa2> it works, but...
Jul 13 14:34:30 <jbowler> Wouldn't work for me - my tmp directory moves around...
Jul 13 14:35:15 <jbowler> DaKa2: do you have packages/install in your working tree?
Jul 13 14:35:32 <DaKa2> yes
Jul 13 14:36:13 <DaKa2> I pulled it brand new like 7h ago
Jul 13 14:37:31 <DaKa2> and make update still fails..
Jul 13 14:38:11 <kolla_> DaKa2: I just noticed that the postfix bild complained about not finding libdb-3, I dont have that myself on my system, so I tried setting export a LD_LIBRARY_PAT, but that didnt work either.. finally I just stuffed that path into my ld.so.conf, temporarly
Jul 13 14:38:21 <kolla_> s/bil/build
Jul 13 14:38:23 <kolla_> d..
Jul 13 14:39:14 <kolla_> not like I'll install postfix anyways :)
Jul 13 14:39:19 <DaKa2> it works, and I guess that dir will be searched last for libs, so you cannot take over any standard system libs, but still seems wrong..
Jul 13 14:39:34 <kolla_> yes, definetly wrong :)
Jul 13 14:39:35 <jbowler> I tried moving my /usr/lib/libpcre.so* but postfix still worked - so I assume ldd isn't lying about where the .so comes from
Jul 13 14:39:40 <DaKa2> Im not using it either, I just want a complete build to work out of the box
Jul 13 14:39:59 <DaKa2> this is just so strange..
Jul 13 14:40:02 <kolla_> hm, no screen.bb
Jul 13 14:40:08 <kolla_> *sob*
Jul 13 14:40:12 <DaKa2> make one? :-)
Jul 13 14:40:16 <kolla_> yes, 
Jul 13 14:40:27 <DaKa2> screen is a must
Jul 13 14:40:28 <kolla_> there's normally one in oe
Jul 13 14:41:02 <kolla_> or.. perhaps not, dont remember
Jul 13 14:41:03 <DaKa2> in nonworking..
Jul 13 14:41:20 <DaKa2> openembedded/packages/nonworking/screen
Jul 13 14:41:23 <kolla_> odd
Jul 13 14:42:07 <DaKa2> this postfix stuff is so odd, just cannot understand why it fails for me
Jul 13 14:42:30 <kolla_> DaKa2: well, did the same here
Jul 13 14:42:53 <DaKa2> what dist are you running?
Jul 13 14:42:56 <kolla_> gentoo
Jul 13 14:43:00 <kolla_> ~x86
Jul 13 14:43:01 <jbot> extra, extra, read all about it, x86 is this frankinsteins monster of a system compaired to arm.
Jul 13 14:43:02 <DaKa2> hm.. well, no that then..
Jul 13 14:43:11 <DaKa2> not*
Jul 13 14:43:31 <DaKa2> :))
Jul 13 14:43:34 <DaKa2> ~arm
Jul 13 14:43:35 <jbot> i heard arm is both big-endian and little-endian. hh-linux is little-endian
Jul 13 14:44:20 <CIA-8> 03marceln * 10unslung/ (Makefile make/ipkg-web.mk): 
Jul 13 14:44:20 <CIA-8> ipkg-web is a small shell script as webfront-end for ipkg.
Jul 13 14:44:20 <CIA-8> Lets all have a critical look at it and decide if it's a
Jul 13 14:44:20 <CIA-8> good addision for unslung.
Jul 13 14:47:18 <jbowler> DaKa2: the path of the libpcre.so is built in to the postconf executable when it is linked, same with libdb-3.2.so
Jul 13 14:47:44 <DaKa2> ok... why does it build in the wrong for me?
Jul 13 14:48:00 <jbowler> See what ldd -v outputs - it should give the full path
Jul 13 14:48:31 <jbowler> (But I think it will just say 'not found')
Jul 13 14:48:47 <DaKa2> yup, not found
Jul 13 14:49:22 <kolla_> and setting LD_LIBRARY_PATH doesnt work either
Jul 13 14:49:40 <kolla_> I guess it is re-set inside the build somewhere
Jul 13 14:50:32 <jbowler> <DaKa2> "LD_LIBRARY_PATH=/home/slug/openslug/tmp/staging/i686-linux/lib/ tmp/staging/i686-linux/bin/postconf --help" works
Jul 13 14:50:46 <DaKa2> jbowler: looking in my postconf I just have libpcre.so.0, no path
Jul 13 14:51:13 <DaKa2> but not if you set LD_LIBRARY_PATH and then run make
Jul 13 14:51:25 <kolla_> jbowler: I meant that you cannot "LD_LIBRARY_PATH=.... make openslug" :)
Jul 13 14:51:32 <jbowler> From the postfix-native build:
Jul 13 14:51:35 <jbowler> ccache gcc  -DHAS_PCRE -I/home/work-tmp/jbowler/nslu2/openslug/staging/i686-linux/include -O2 -g -O -I. -I../../include -DLINUX2 -o postconf postconf.o ../../lib/libglobal.a ../../lib/libutil.a -lpcre -ldb -lnsl -lresolv -L/home/work-tmp/jbowler/nslu2/openslug/staging/i686-linux/lib -Wl,-rpath-link,/home/work-tmp/jbowler/nslu2/openslug/staging/i686-linux/lib -Wl,-rpath,/home/work-tmp/jbowler/nslu2/openslug/staging/i686-linux/lib -Wl,-O1
Jul 13 14:52:20 <jbowler> I believe the 'Wl,rpathlink,' stuff is what gets it to work.
Jul 13 14:52:43 <jbowler> And my gcc is: gcc (GCC) 3.3.5-20050130 
Jul 13 14:53:20 <DaKa2> gcc  -DHAS_PCRE -I/home/slug/openslug/tmp/staging/i686-linux/include -O2 -g -O -I. -I../../include -DLINUX2 -o postconf postconf.o ../../lib/libglobal.a ../../lib/libutil.a -L/home/slug/openslug/tmp/staging/i686-linux/lib -lpcre -ldb -lnsl -lresolv
Jul 13 14:53:52 <DaKa2> gcc version 3.3.4 (Debian 1:3.3.4-7)
Jul 13 14:55:01 <jbowler> No -Wl !
Jul 13 14:55:15 <DaKa2> exactly
Jul 13 14:55:30 <DaKa2> and different order of other commands
Jul 13 14:55:51 <DaKa2> kindof strange..
Jul 13 14:56:49 <DaKa2> WHY!! OH WHY! now it worked.. I did exactly the same as before
Jul 13 14:57:23 <jbowler> And you got the -Wl's added on?
Jul 13 14:57:36 <DaKa2> yep
Jul 13 14:57:58 <DaKa2> I think someone is playing with me...
Jul 13 14:58:01 <jbowler> You still have all the evidence, unless you wiped the temp directory - i.e. run.do_compile and log.do_compile
Jul 13 14:58:17 <DaKa2> well... -cclean does... :-(
Jul 13 14:58:40 <DaKa2> well, I still have them on my home build host
Jul 13 14:58:42 <DaKa2> I think
Jul 13 14:59:04 <DaKa2> but I should probably clean tmp on here and start again, It'll probably fail again
Jul 13 14:59:13 <jbowler> Hum.  That's a pity.  The -Wl comes from LDFLAGS.  The LDFLAGS override the settings in the makefile because of the "-e" in EXTRA_OEMAKE
Jul 13 14:59:40 <jbowler> The other problem in the pcre build (dftables) was because of a -e issue too.
Jul 13 15:01:21 <jbowler> I suspect this is the same - I suspect your old run.do_compile would have shown the correct LDFLAGS.
Jul 13 15:01:36 <DaKa2> Ill check at my home build
Jul 13 15:02:03 <kolla_> er, same here
Jul 13 15:02:12 <kolla_> I rebuildt postfix, and now it worked
Jul 13 15:02:28 <DaKa2> dammit.. I didn't have it at home
Jul 13 15:02:31 <jbowler> in src/postconf/Makefile the command which links postconf (line 124) corresponds, with the Makefile settings to your failing command.
Jul 13 15:02:51 <kolla_> and I also now have -Wl, which wasnt there the last time
Jul 13 15:03:23 <jbowler> But run.do_compile overwrites "SYSLIBS" in a way which will only work if -e is on the make command line...
Jul 13 15:03:49 <jbowler> (I.e. it is SYSLIBS in the environment, not LDFLAGS, which actually causes the change.)
Jul 13 15:04:31 <DaKa2> well, its going to take some time before I can find out.. rebuilding..
Jul 13 15:05:23 <kolla_> after having done a complete build.. I guess reflashing is next?
Jul 13 15:06:16 <kolla_> hm, what decides which of the packages in openembedded/packages are available?
Jul 13 15:06:30 <DaKa2> oe-symlinks
Jul 13 15:06:38 <jbowler> openslug-packages.bb
Jul 13 15:06:49 <jbowler> If it gets built it is available.
Jul 13 15:06:49 <DaKa2> and what gets built is openslug-packages.bb
Jul 13 15:07:30 <kolla_> yes, that I figured out
Jul 13 15:07:30 <DaKa2> well, you can bb stuff in oe-symlinks, but ok openslug-packages.bb
Jul 13 15:09:55 <DaKa2> dammit.. now it worked at home too
Jul 13 15:10:22 <DaKa2> but that was with bb postfix, not make openslug
Jul 13 15:10:25 <kolla_> hm, so I put twin into openslug-packages.bb.. should be able to "bb twin"? I still got "ERROR: Nothing provides twin"
Jul 13 15:10:40 <DaKa2> is it in oe-symlinks?
Jul 13 15:10:48 <kolla_> DaKa2: ah yes, same here.. bb postfix worked
Jul 13 15:11:29 <kolla_> DaKa2: ah, thanks
Jul 13 15:11:54 <DaKa2> kolla_: if you put it in oe-symlinks you can bb it, if you put it in oe-symlinks and openslug-packages.bb it gets built by make openslug
Jul 13 15:12:04 <jbowler> Dang.  Use 'make -e' not 'make' on the master makefile.
Jul 13 15:12:19 <DaKa2> you use another master makefile?
Jul 13 15:12:34 <jbowler> No, I use bitbake directly.
Jul 13 15:12:39 <DaKa2> :)
Jul 13 15:12:56 <DaKa2> that explains alot
Jul 13 15:13:04 <kolla_> DaKa2: what I thought
Jul 13 15:13:58 <DaKa2> jbowler: ok, so, this should probably get fixed somehow...
Jul 13 15:14:44 <jbowler> If it does work with -e and doesn't, reliably, without then the fix is to set MFLAGS, or whatever complexity FSF have come up with.
Jul 13 15:15:08 <DaKa2> but why are nobody else getting this, just me and kolla_?
Jul 13 15:15:25 <DaKa2> not everybody runs bb directly?
Jul 13 15:15:42 <[g2]> I do
Jul 13 15:15:46 <DaKa2> oh my god, that was a bad sentence..
Jul 13 15:15:50 <[g2]> 95% of the time
Jul 13 15:15:53 <jbowler> Some people haven't pulled in the postfix changes, some people don't use the makefile, some people don't care enough to take any notice.
Jul 13 15:16:21 <DaKa2> well, make -e is going to take a while
Jul 13 15:16:52 <kolla_> screen didnt compile, allright :)
Jul 13 15:16:58 <jbowler> It's like mgetty - it had a similar problem, but running a second time (without cleaning) it worked...
Jul 13 15:17:38 <DaKa2> thats well, strange..
Jul 13 15:18:44 <DaKa2> I like the debian way, everything should build from source in a clean chroot with just build-essentials, but it takes alot of work..
Jul 13 15:18:45 <jbowler> The first time it made the top level recursively as a side effect of something which had already built.  The second time the something had already built, it didn't build it again and made the top level directly - something like that...
Jul 13 15:19:13 <kolla_> sad thing is that I cannot just install the ipkgs I have buildt now, can I?
Jul 13 15:19:42 <DaKa2> kolla_: well, have you flashed you slug with a recent openslug firmware?
Jul 13 15:19:52 <kolla_> define "recent"
Jul 13 15:20:14 <kolla_> Jun 21 00:46:18 PDT 2005
Jul 13 15:20:17 <kolla_> is the kernel
Jul 13 15:20:29 <DaKa2> well, I dont think the 2.0 beta works because it has an old glibc, but Im not really sure on that
Jul 13 15:20:37 <DaKa2> anybody?
Jul 13 15:20:39 <kolla_> yes, exactly.. 
Jul 13 15:20:48 <jbowler> I think you will get a new glibc automagically, but no guarantees
Jul 13 15:21:04 <kolla_> :)
Jul 13 15:21:10 <jbowler> However, the ipkg upgrade would kill all your kernel modules and the system would no longer boot.
Jul 13 15:21:19 <jbowler> Use once and reflash.
Jul 13 15:21:23 <kolla_> I know I cannot use the kernel modules, since different gcc is used
Jul 13 15:21:24 <DaKa2> :))
Jul 13 15:21:38 <DaKa2> I would suggest a reflash :-)
Jul 13 15:21:51 <kolla_> jbowler: which is why I never let any silly package manager deal with my kernels :)
Jul 13 15:22:13 <DaKa2> hm.. debian does it nicly
Jul 13 15:22:21 <kolla_> well.. "nicely" :)
Jul 13 15:22:32 <DaKa2> hm..
Jul 13 15:22:38 <kolla_> I've had my share with debians mkinitrd :)
Jul 13 15:23:04 <DaKa2> ? never had any problems..
Jul 13 15:23:16 <DaKa2> oh well..
Jul 13 15:23:22 <kolla_> when SATA went from being IDE to being "SCSI", it was lots of problems :)
Jul 13 15:23:39 <DaKa2> ow, well. that probably hurts a bit
Jul 13 15:23:41 <kolla_> I prefer doing it manually, not like it's any work :)
Jul 13 15:24:23 <kolla_> hm, gotta read up on how to flash without the webinterface :)
Jul 13 15:24:53 <DaKa2> should probably get going.. last subway leaves in 20 minutes
Jul 13 15:26:10 <DaKa2> feels kindof bad staying at work untill 0:29 in the middle of the week when you are on vacation, so, I didn't work , still feels wrong
Jul 13 15:27:03 <[g2]> DaKa2, need broadband at home :)
Jul 13 15:27:43 <DaKa2> yeah, 24 mbit down is just to bad to have at home
Jul 13 15:29:04 <[g2]> I hear NAiL has 8
Jul 13 15:29:29 <DaKa2> what I really need at home is a bigger home, so I can have all the nice toys there instead of at work
Jul 13 15:29:46 <DaKa2> I have 24mbit down, but only 1 mbit up
Jul 13 15:30:11 <DaKa2> Im thinging of getting 10/10
Jul 13 15:30:24 <kolla_> heh, I've only got 1Mb down and 256k up :|
Jul 13 15:30:35 <kolla_> *sniff*
Jul 13 15:30:35 <DaKa2> that would cost about 33USD/month
Jul 13 15:30:47 <kolla_> 256k up is a pita
Jul 13 15:31:05 <DaKa2> yup.. I remember, had that 3 years ago..
Jul 13 15:32:39 <DaKa2> 100/100 would be 50USB/month, but I really dont need that
Jul 13 15:33:58 <Tiersten> Please form an orderly queue to lynch DaKa2
Jul 13 15:34:06 <DaKa2> what? :-)
Jul 13 15:36:08 <DaKa2> jbowler: make -e makes gcc-cross-initial fail..
Jul 13 15:36:30 <DaKa2> but I have to run, bbi45m
Jul 13 15:39:13 <jbowler> DaKa2: ok, that was a bad idea then :-(
Jul 13 15:39:53 <jbowler> The problem just seems to revolve around the files which don't use autotools - mgetty, pcre, postfix
Jul 13 15:41:39 <kolla_> is it flashing via tftp from redboot that is the norm for openslug users?
Jul 13 15:42:27 <ByronT-Desktop> I've used Sercomm from an XP machine
Jul 13 15:43:09 <kolla_> dont have serial access :)
Jul 13 15:43:34 <ByronT-Desktop> Sercomm only needs a windows machine on the same subnet
Jul 13 15:43:45 <rwhitby> kolla_: if you have a working OpenSlug 2.0 or later, use reflash. If not. get it into upgrade mode (either by dd or the button on the back) and use SerComm
Jul 13 15:44:00 <ByronT-Desktop> but I would say that a good number of openslug probably use upslug
Jul 13 15:47:14 <kolla_> aha, a command "reflash"
Jul 13 15:47:51 <kolla_> nice
Jul 13 15:48:45 <kolla_> twin compiled nicely, good
Jul 13 15:48:56 <kolla_> then I can have a desktop on the slug :)
Jul 13 15:59:02 <kolla_> think I want to alter some stuff in the kernel before I flash
Jul 13 16:05:50 <CIA-8> 03repvik * r83 10/releases/OpenSlug-2.0-beta/openembedded/site/armeb-linux: Forgot to update site/armeb-linux so samba doesn't break
Jul 13 16:12:51 <CIA-8> 03repvik * r84 10/releases/OpenSlug-2.0-beta/openembedded/packages/nfs-utils/files/ (. nfsserver): Added the initscript
Jul 13 16:27:13 <DaKa2> hm..
Jul 13 16:32:10 <[g2]> DaKa2, around ?
Jul 13 16:32:13 <DaKa2> yup
Jul 13 16:32:28 <[g2]> you're not going to miss the train are you ?
Jul 13 16:32:33 <DaKa2> always, except when not
Jul 13 16:32:34 <DaKa2> Im home
Jul 13 16:32:55 <[g2]> on the postfix issue.. did you do a bb -D > zzz
Jul 13 16:32:57 <DaKa2> the last one left work 25 mins ago..
Jul 13 16:33:09 <DaKa2> uhm, no?
Jul 13 16:33:20 <[g2]> you see bb debugging output
Jul 13 16:33:22 <DaKa2> but bb postfix works
Jul 13 16:33:34 <[g2]> pcre builds now ?
Jul 13 16:33:41 <DaKa2> its when you run it from the main makefile..
Jul 13 16:33:51 <DaKa2> it was never the problem of pcre for me
Jul 13 16:34:12 <kolla_> same here
Jul 13 16:34:18 <DaKa2> but a complete rebuild with make fails
Jul 13 16:34:32 <DaKa2> on postfix, but bb postfix works
Jul 13 16:34:37 <[g2]> ahh
Jul 13 16:34:38 <DaKa2> even in a clean tmp
Jul 13 16:34:50 <[g2]> so this is the makefile issue
Jul 13 16:35:07 <DaKa2> "the" makefile issue? only one :-)
Jul 13 16:35:15 <[g2]> :)
Jul 13 16:35:28 <[g2]> "a" makefile issue :)
Jul 13 16:35:33 <DaKa2> :))
Jul 13 16:35:35 <[g2]> "a known"
Jul 13 16:36:11 <DaKa2> well, any idea what to do? does bb openslug-packages openslug-image do the same as man openslug?
Jul 13 16:36:16 <DaKa2> make openslug*
Jul 13 16:36:18 <[g2]> Ok... I think it's getting clearer
Jul 13 16:36:34 <[g2]> well openslug-update works right ?
Jul 13 16:36:56 <[g2]> I think ipkg-cl has issues
Jul 13 16:36:59 <DaKa2> make: *** No rule to make target `openslug-update'.  Stop.
Jul 13 16:37:15 <[g2]> update-openslug :)
Jul 13 16:37:19 <DaKa2> :)
Jul 13 16:37:40 <DaKa2> make: *** No rule to make target `update-openslug'.  Stop.
Jul 13 16:37:52 <DaKa2> you mean update-openembedded?
Jul 13 16:39:52 <DaKa2> because that doesn't work..
Jul 13 16:40:41 <[g2]> so it's the openslug-build that fails
Jul 13 16:41:03 <[g2]> "Don't make me look at that Makefile!"
Jul 13 16:41:07 <[g2]> :)
Jul 13 16:41:11 <DaKa2> :)
Jul 13 16:41:24 <DaKa2> openslug-build is the one failing
Jul 13 16:41:37 <DaKa2> I think
Jul 13 16:41:59 <DaKa2> build-openslug...
Jul 13 16:44:02 <DaKa2> hm.. when is make update going to work again? *big smile* :)
Jul 13 16:50:25 <jbowler> DaKa2: is it failing because of the merge (no private key?)
Jul 13 16:50:32 <DaKa2> yup...
Jul 13 16:51:06 <jbowler> Can you tell me the ancestor revision and the two heads (I'm assuming, since you have no private key, you have no local commits).
Jul 13 16:51:24 <DaKa2> monotone: [source] 2117dd5b726765202daffe67318297ef78df5a4f
Jul 13 16:51:24 <DaKa2> monotone: [source] 2913dd5a4db27b7b121a13b703b88a41e73464aa
Jul 13 16:51:24 <DaKa2> monotone: common ancestor 43007d439c7ed96133926d0748cef53fa1b321d0 found
Jul 13 16:51:24 <DaKa2> monotone: trying 3-way merge
Jul 13 16:51:24 <DaKa2> monotone: misuse: no unique private key for cert construction
Jul 13 16:51:34 <DaKa2> nope, nothing local
Jul 13 16:52:20 <jbowler> thx - no saying I can fix it, but to my mind it definately shouldn't be happening and you have the simplest case...
Jul 13 16:52:38 <DaKa2> :)
Jul 13 16:57:46 <jbowler> Stupid.  There's no overlap at all between the different files ancestor->left and ancestor->right.
Jul 13 16:58:10 <DaKa2> ?
Jul 13 16:58:19 * DaKa2 understands 0
Jul 13 16:59:20 <jbowler> The only change in the second revision is to openslug-packages.bb, the only changes in the right are poodle/zaurus/opie/mplayer.
Jul 13 17:00:22 <jbowler> So a merge is trivial, but it still apparently needs to create a new revision on your tree.
Jul 13 17:00:37 <DaKa2> o.k..
Jul 13 17:00:52 <jbowler> NAiL: I'm going to merge those two heads on nslu2-linux, they might be the cause of this problem.
Jul 13 17:01:16 <NAiL> jbowler: uhm.. go ahead. Not sure why you're telling me 
Jul 13 17:01:34 <jbowler> Yep.  I get the same 3-way as DaKa2
Jul 13 17:01:51 <NAiL> ugh.. that was my doing?
Jul 13 17:02:11 <jbowler> Well, I don't know - I think we both did it, since the heads are:
Jul 13 17:02:14 <DaKa2> 2913dd5a4db27b7b121a13b703b88a41e73464aa is yours... I think...
Jul 13 17:02:26 <jbowler> right, and the other is mine.
Jul 13 17:02:57 <jbowler> DakA2 - what is the contents of MT/revision (in the top level of openembedded?)
Jul 13 17:03:06 <kolla_> there.. new kernel, new image
Jul 13 17:03:09 <DaKa2> But I have got 3-way merges before that worked..
Jul 13 17:03:21 <DaKa2> 3c79bddb5bae39f208c2121db3a94b142ec9b5af
Jul 13 17:03:31 <NAiL> monotone: branch 'org.nslu2-linux.dev' is currently merged:
Jul 13 17:03:31 <NAiL> 3c79bddb5bae39f208c2121db3a94b142ec9b5af rwhitby@nslu2-linux.org 2005-07-11T19:54:07
Jul 13 17:03:50 <jbowler> Sorry - org.openembedded.nslu2-linux
Jul 13 17:03:56 <NAiL> aah
Jul 13 17:03:58 <kolla_> hm, with reflash I dont need the .img right?
Jul 13 17:04:08 <DaKa2> 43007d439c7ed96133926d0748cef53fa1b321d0
Jul 13 17:04:12 <kolla_> just the zimage and jffs2 is enough?
Jul 13 17:04:16 <DaKa2> mised the openembedded
Jul 13 17:04:27 <jbowler> kolla_ - you need either the image or both the kernel and the rootfs, or you need to know that you can just change one safely
Jul 13 17:05:52 <jbowler> If you are changing a kernel be very careful - you must ipkg upgrade along with the reflash or your rootfs will be unuseable.
Jul 13 17:06:15 <jbowler> ('changing a kernel' means changing System.map, so the modules change.)
Jul 13 17:06:43 <jbowler> DaKa2: try make update now.
Jul 13 17:07:04 <DaKa2> ok
Jul 13 17:07:08 <kolla_> jbowler: right, I suddenly saw the -i
Jul 13 17:07:27 <DaKa2> worked
Jul 13 17:07:54 <jbowler> Ok, that was just because there were two heads in there - different problem.
Jul 13 17:07:55 <DaKa2> but why couldn't I do that merge locally? I've been able to do that before
Jul 13 17:08:17 <jbowler> Because it actually had to create a new revision in order to make a manifest for the branch which has both sets of changes.
Jul 13 17:08:38 <DaKa2> hm.. I really dont understand this monotone..
Jul 13 17:08:38 <jbowler> You should never see that happen in make update unless there are two heads.
Jul 13 17:09:16 <jbowler> The branch is just a symbolic label for a file (the manifest) which lists all the other files - name and revision.
Jul 13 17:09:41 <jbowler> A branch can name multiple manifests - multiple heads - then it's ambiguous.
Jul 13 17:10:14 <jbowler> To fix that the two (or more) manifests are combined (merged), to create a unique manifest with both sets of changes.
Jul 13 17:10:49 <jbowler> But that requires commiting a change to the database, so you need write access to do it...
Jul 13 17:10:49 <DaKa2> oh, I actually think I understand
Jul 13 17:11:08 <DaKa2> but not how this happened
Jul 13 17:11:54 <jbowler> Both I and NAiL commited a change to our local db (creating two different heads), then we synced - pushing out our new head and getting the others.
Jul 13 17:11:57 <jbowler> Two heads.
Jul 13 17:12:06 <DaKa2> ahh
Jul 13 17:12:11 <kolla_> uhm.. the part about "restoring saved configuration files" .. which files are those?
Jul 13 17:12:18 <kolla_> not /linuxrc I hope?
Jul 13 17:12:24 <kolla_> :)
Jul 13 17:12:48 <jbowler> Yes, it does restore /linuxrc.
Jul 13 17:13:01 <DaKa2> someone didn't update to the head in monotone before they pushed their changes to monotone?
Jul 13 17:13:04 <jbowler> Run turnup flash if you want to put the one in the image back.
Jul 13 17:13:43 <jbowler> DaKa2: it's a race, there is no locking and if you keep a local commit (necessary sometimes) it's difficult to avoid.
Jul 13 17:14:00 <kolla_> jbowler: it's just that if it tries to load kernel modules from the external disk, those wont work :)
Jul 13 17:14:26 <jbowler> Exactly: do ipkg upgrade before you reboot!
Jul 13 17:14:33 <kolla_> right :)
Jul 13 17:14:37 <DaKa2> ahh, but still, kindof stupid system if something like that can happen, but still, Im really really happy to get away from BK
Jul 13 17:14:46 <kolla_> that also involves setting up my repository
Jul 13 17:14:52 <jbowler> It's much easier to do it before, indeed, I don't know how to do upgrade a rootfs which is not on /
Jul 13 17:15:20 <jbowler> kolla_: well, someone needs to work out how to do an ipkg upgrade on /mnt ;-)
Jul 13 17:15:25 <DaKa2> chroot it?
Jul 13 17:15:45 <kolla_> should work fine
Jul 13 17:15:58 <jbowler> Hum...  I think ipkg takes a destdir arg, but it delays the preinst and postinst scripts.
Jul 13 17:16:02 <kolla_> for now I just copy over the modules on the jffs2 over to my disk
Jul 13 17:16:05 <jbowler> chroot sounds better if it works.
Jul 13 17:16:24 <DaKa2> it should, if the packages are avaliable under the chroot
Jul 13 17:16:39 <kolla_> indeed
Jul 13 17:16:58 <jbowler> Yes, copy the modules should work in this case - because I don't think anything has changed in the ABI so all the user mode stuff (like depmod/insmod) should be ok.
Jul 13 17:17:56 <kolla_> :)
Jul 13 17:18:27 <jbowler> chroot might have problems if you have anything mounted tmpfs in /var
Jul 13 17:18:57 <CIA-8> 03jbowler 07org.openembedded.nslu2-linux * r479eca03... 10/ (13 files in 10 dirs): 
Jul 13 17:18:57 <CIA-8> propagate from branch 'org.openembedded.dev' (head e7c80380016749ce0426efb89bd3a65e04ea1bd9)
Jul 13 17:18:57 <CIA-8>  to branch 'org.openembedded.nslu2-linux' (head 479c5da4a656d8b1d5f4da2bc403d06731d2b9db)
Jul 13 17:18:59 <jbowler> Still, sounds like the best solution, I hadn't thought of it, great idea 
Jul 13 17:19:21 <DaKa2> well, only placing files that are going to be hidden by the mount later
Jul 13 17:19:30 <DaKa2> not a big problem I think
Jul 13 17:19:48 <DaKa2> you can place files in /var/tmp, and then mount tmpfs over that
Jul 13 17:20:02 <jbowler>  /var needs enough in it for ipkg to work.  /var/lib/ipkg, /var/tmp
Jul 13 17:20:04 <kolla_> does ipkg understand ftp?
Jul 13 17:20:11 <jbowler> yes
Jul 13 17:20:11 <kolla_> uhm, it just uses wget, right?
Jul 13 17:20:18 <jbowler> I believe so
Jul 13 17:20:19 <kolla_> so then ftp is just as good
Jul 13 17:20:29 <DaKa2> hm.. yeah, on a 128mb usb memstick that could be a problem
Jul 13 17:20:38 <jbowler> I use ftp (I've actually deleted the official feeds too.)
Jul 13 17:21:01 <DaKa2> I use a mounted nfs-dir
Jul 13 17:21:14 <DaKa2> for some reason
Jul 13 17:21:24 <jbowler> We could run selected /etc/init.d/ scripts - they should work in a chroot environment.
Jul 13 17:22:10 <jbowler> Only the file system needs to be correct.  The kernel devices and modules are already there.
Jul 13 17:22:11 <DaKa2> hm.. yes.. or mount everything in fstab not /
Jul 13 17:22:31 <jbowler> Right.
Jul 13 17:22:47 <DaKa2> can you mount procfs in two places?
Jul 13 17:22:57 <jbowler> I believe so.
Jul 13 17:23:06 <DaKa2> then it should work
Jul 13 17:23:18 <kolla_> you can mount procfs as many plases as you like
Jul 13 17:23:24 <jbowler> The interesting thing would be NFS mounts, because /var/lib/nfs would be outside the chroot (since it's accessed from kernel threads).
Jul 13 17:23:49 <kolla_> I have a complete debian installed inside chroot on my gentoo box at work, for building .deb packages :)
Jul 13 17:24:10 <DaKa2> :)
Jul 13 17:24:38 <kolla_> did anyone look at nbd?
Jul 13 17:24:40 <jbowler> YES!  My monotone pull from vanille.de completed, sucessfully, about 18 hours!
Jul 13 17:24:56 <NAiL> hahaha
Jul 13 17:25:06 <DaKa2> 18h? how much did you pull??
Jul 13 17:25:09 <jbowler> Hey
Jul 13 17:25:14 <NAiL> now, pull all the revs that have been done since you started :-P
Jul 13 17:25:55 <jbowler> From scratch, all 1222 certs, 459 revisions - 22.0MByte, 99% CPU, 50% memory (17MByte of the 32MByte).
Jul 13 17:26:11 <jbowler> NAiL: too late, already started that ;-)
Jul 13 17:26:21 <jbowler> 15 revisions in this time.
Jul 13 17:26:38 <NAiL> Shouldn't take more than an hour or two :-P
Jul 13 17:27:09 <jbowler> good thing it's turbo'ed
Jul 13 17:27:15 <kolla_> ah, need to generate Packages.gz
Jul 13 17:27:25 <DaKa2> bb package-index
Jul 13 17:27:34 <kolla_> ah :)
Jul 13 17:28:10 <DaKa2> jbowler: ah, you did it on the slug :-))
Jul 13 17:28:31 <kolla_> :)
Jul 13 17:28:43 <jbowler> Yes, now where do I get the native development environment.
Jul 13 17:29:00 <kolla_> when I run gentoo on such things, I nfs mount the portage tree from a pc
Jul 13 17:29:00 <DaKa2> why not copy a working MT tree from a bit faster PC?
Jul 13 17:29:56 <jbowler> It's not the tree - it's the database which is the problem.  I believe the working tree construction will be pretty fast.
Jul 13 17:30:22 <DaKa2> monotone/nslu2-linux.db?
Jul 13 17:30:31 <DaKa2> then just copy that? :-)
Jul 13 17:30:32 <jbowler> Copying the database is a viable option, but I'm doing this to test it...
Jul 13 17:30:46 <DaKa2> well, a good test
Jul 13 17:30:56 <DaKa2> but still, 18h...
Jul 13 17:31:38 <kolla_> hehe.. funny slashdot header.. "Study Shows One Third of All Studies Are Nonsense"
Jul 13 17:31:58 <DaKa2> :)
Jul 13 17:35:12 <NAiL> 90% of all statistics are crud
Jul 13 17:37:56 <DaKa2> 175% of all 100%+ of something statistics are flawed
Jul 13 17:38:14 <kolla_> now, another try.. with time with a hole punctured through the access list on the router 
Jul 13 17:38:15 <kolla_> :P
Jul 13 17:38:23 <kolla_> s/with/this
Jul 13 17:40:20 <kolla_> now.. upgrade, this will be interresting
Jul 13 17:40:35 <DaKa2> :)
Jul 13 17:40:51 <kolla_> here goes libc... wee
Jul 13 17:41:37 <kolla_> er.. wonder what "installing kernel-image" does
Jul 13 17:41:51 <kolla_> put stuff in /boot ?
Jul 13 17:42:09 <DaKa2> probably... maybe..
Jul 13 17:42:14 <kolla_> :)
Jul 13 17:42:49 <jbowler> put stuff in boot, and leave it there.
Jul 13 17:43:27 <DaKa2> not remove when new version comes?
Jul 13 17:44:08 <jbowler> I think it does remove the old one, ipkg is like that.
Jul 13 17:44:09 <kolla_> yes, it did both
Jul 13 17:44:25 <kolla_> replaced 2.6.11 with 2.6.12 in /boot
Jul 13 17:44:27 <jbowler> (But turnup doesn't copy the original /boot/zImage...)
Jul 13 17:44:29 <DaKa2> thats good
Jul 13 17:44:42 <jbowler> 2.6.12 or 2.6.12.2?
Jul 13 17:45:53 <kolla_> the latter
Jul 13 17:47:15 <jbowler> Good, then there should be a /lib/modules/2.6.12.2/... and /lib/modules/2.6.11.2/*/... should all be empty (just the directories)
Jul 13 17:47:37 <kolla_> yes
Jul 13 17:47:47 <kolla_> hm, busybox gets in the way as usual
Jul 13 17:51:51 <kolla_> ipkg should have an oppsite of -force-overwrite.. -ignore-overwrite, so that it would drop installing symlinks for things that are covered by other packages
Jul 13 17:52:15 <DaKa2> nah, better usage of alternatives
Jul 13 17:52:23 <kolla_> in case of busybox that is
Jul 13 17:53:18 <DaKa2> that is, busybox provides /usr/bin/awk.busybox and gawk provides /usr/bin/awk.gawk, and /etc/alternatives links the one with highest priority
Jul 13 17:53:33 <DaKa2> (thats the file Im having problems with)
Jul 13 17:53:44 <kolla_> mm
Jul 13 18:12:33 <kolla_> woohoo
Jul 13 18:13:11 <DaKa2> working?
Jul 13 18:13:34 <kolla_> Linux babysmurf 2.6.12.2 #1 Thu Jul 14 01:38:09 CEST 2005 armv5teb unknown unknown GNU/Linux
Jul 13 18:13:35 <kolla_> :)
Jul 13 18:13:38 <kolla_> yep
Jul 13 18:13:41 <NAiL> kolla_: Now's the time to try as many packages as you can ;-)
Jul 13 18:13:45 <kolla_> and ipv6 too
Jul 13 18:14:20 <kolla_> cool
Jul 13 18:15:21 <jbowler> Were you able to reboot through to the disk?  (I.e. without going back to flash and fixing something up?)
Jul 13 18:15:22 <DaKa2> I was going to set up apache+mysql+php on a slug, but Im thinking sleep instead
Jul 13 18:16:53 <kolla_> yes
Jul 13 18:17:12 <Tiersten> Sleep?  Whats that?
Jul 13 18:17:21 <Tiersten> If you find out, please tell NAiL
Jul 13 18:17:26 <kolla_> flased from disk, ipkg update/upgrade, reboot
Jul 13 18:17:30 <kolla_> flashed
Jul 13 18:17:45 <kolla_> Tiersten: I'm going to work in 4 hours :)
Jul 13 18:17:59 <jbowler> Woohoo, it works (I think you are the first person to actually do this :)
Jul 13 18:18:04 <kolla_> tend to replace sleep with beer, good thing I dont have a drivers license :)
Jul 13 18:18:16 <DaKa2> kolla_: lol..
Jul 13 18:18:19 <Tiersten> :)
Jul 13 18:18:30 <kolla_> jbowler: amazing, I'm usually never first at anything these days :)
Jul 13 18:18:59 <DaKa2> oh well, its 3 am, really time to sleep, bbl
Jul 13 18:19:09 <NAiL> hmm
Jul 13 18:19:15 <NAiL> It's 3:18 am
Jul 13 18:19:21 <NAiL> No time to sleep :-P
Jul 13 18:19:22 <DaKa2> well, close
Jul 13 18:19:26 <jbowler> Well, that kernel's only been out a few days...
Jul 13 18:19:28 <DaKa2> not?
Jul 13 18:19:42 <NAiL> naah
Jul 13 18:19:45 <kolla_> :)
Jul 13 18:19:57 <NAiL> (Well, ok, as soon as I finish this beer, I'm going to bed myself)
Jul 13 18:20:09 <DaKa2> It'll soon get too light outside to sleep
Jul 13 18:20:19 <DaKa2> well, it already is
Jul 13 18:20:21 <NAiL> I'd say it is
Jul 13 18:20:24 <NAiL> yeah ;)
Jul 13 18:20:36 <DaKa2> damned summer
Jul 13 18:20:51 <NAiL> Don't complain. This is what we long for all winter
Jul 13 18:20:58 <DaKa2> uhm.. it is?
Jul 13 18:21:05 <DaKa2> who could ever want this?
Jul 13 18:21:11 * DaKa2 loves with winter
Jul 13 18:21:12 <NAiL> Well... Me :)
Jul 13 18:21:15 <DaKa2> with?
Jul 13 18:21:18 <DaKa2> the*
Jul 13 18:21:24 <DaKa2> Im tired...
Jul 13 18:21:55 * DaKa2 zzZzZzZ
Jul 13 18:22:01 <NAiL> nite
Jul 13 18:22:43 <[g2]> Nite
Jul 13 18:22:43 <NAiL> Hopefully I'll be able to work a lot tomorrow
Jul 13 18:22:51 <NAiL> Been a few days with almost no work now
Jul 13 18:23:03 <NAiL> annoys me :)
Jul 13 18:23:09 <kolla_> where do you work?
Jul 13 18:23:36 <NAiL> ah, not that kind of work. openslug work
Jul 13 18:23:42 <kolla_> oh, right :)
Jul 13 18:23:45 <NAiL> Other than that, I do consulting 
Jul 13 18:24:09 <kolla_> ok :)
Jul 13 18:25:31 <kolla_> there, now both slugs have ipv6 addresses set in dns as well
Jul 13 18:25:56 <[g2]> kolla_, you running ipv6 on openslug ?
Jul 13 18:26:32 <kolla_> yes
Jul 13 18:26:38 <kolla_> running ipv6 everywhere :)
Jul 13 18:28:57 <[g2]> did you setup your own defconfig ?
Jul 13 18:29:12 <[g2]> or is everything already in openslug ?
Jul 13 18:31:00 <kolla_> my own defconfig yes
Jul 13 18:31:30 <kolla_> I also added ikconfig and ikconfig_proc
Jul 13 18:31:52 <kolla_> so I will keep the kernel config around :)
Jul 13 18:32:02 * NAiL is interested, but I've gotta sleep first :-P
Jul 13 18:32:04 <[g2]> can you pls add a openslug howto for ipv6 ?
Jul 13 18:32:13 <[g2]> on the wiki ?
Jul 13 18:32:29 <[g2]> several ppl would like to do taht
Jul 13 18:32:46 <kolla_> sure
Jul 13 18:32:51 <[g2]> excellent
Jul 13 18:33:04 <kolla_> ping6 was already available in openslug
Jul 13 18:33:23 <[g2]> if the kernel mods aren't too extensive maybe we could just add it to the default config
Jul 13 18:33:47 <kolla_> the busybox netstat doesnt cut it though :)
Jul 13 18:34:01 <[g2]> right.... 
Jul 13 18:34:18 <[g2]> that kind of stuff is very useful too
Jul 13 18:34:21 <kolla_> addin support in the kernel is no problem, the problem is apps :)
Jul 13 18:34:54 <[g2]> with lots of ppl interested I'm sure that would get figured out 
Jul 13 18:35:11 <[g2]> jbowler ping
Jul 13 18:40:34 <jbowler> pong
Jul 13 18:41:57 <[g2]> I'm trying to understand ipkg-cl and the 2.6.12.x dependency
Jul 13 18:42:20 <[g2]> can you peek at http://pastebin.ca/17665
Jul 13 18:43:14 <jbowler> You have to install the kernel before you can install any of it's modules - ipkg will do that for you, automagically
Jul 13 18:44:01 <[g2]> that's what is getting created from the bb openslug-image
Jul 13 18:44:18 <[g2]> that's what is getting created from the bb -D openslug-image 
Jul 13 18:44:25 <jbowler> rm -rf tmp
Jul 13 18:45:08 <[g2]> that may work, but won't tell me where the issue is with OE
Jul 13 18:45:12 <kolla_> heh, busybox is ipv6 enabled
Jul 13 18:45:17 <kolla_> route -Ainet6 works fine
Jul 13 18:46:03 <[g2]> it's a 90 minute penatly for rm -rf tmp and 3hours + for the packages
Jul 13 18:46:03 <jbowler> [g2]: there isn't an issue with OE, you have apparently not rebuilt the ixp modules after rebuilding the kernel
Jul 13 18:46:16 <[g2]> I sure have
Jul 13 18:46:50 <[g2]>  bb ixp4xx-csr ixp425-eth
Jul 13 18:46:51 <[g2]> NOTE: Using cache in '/home/tom/dev/1july_mono/openslug/tmp/cache'
Jul 13 18:46:51 <[g2]> NOTE: Parsing finished. 612 cached, 0 parsed, 25 skipped, 0 masked.
Jul 13 18:46:51 <[g2]> NOTE: build 200507132130: started
Jul 13 18:46:51 <[g2]> OE Build Configuration:
Jul 13 18:46:53 <[g2]> TARGET_ARCH   = "armeb"
Jul 13 18:46:55 <[g2]> TARGET_OS     = "linux"
Jul 13 18:46:57 <[g2]> MACHINE       = "nslu2"
Jul 13 18:46:59 <[g2]> DISTRO        = "openslug"
Jul 13 18:47:01 <[g2]> TARGET_FPU    = "soft"
Jul 13 18:47:03 <[g2]> NOTE: build 200507132130: completed
Jul 13 18:47:05 <[g2]> Build statistics:
Jul 13 18:47:07 <[g2]>   Attempted builds: 0
Jul 13 18:47:24 <jbowler> If you have rebuilt them, from a current tree, why were they built against a set of stuff in staging which is from the 2.6.11.2 kernel?
Jul 13 18:48:00 <[g2]> PREFERRED_VERSION_openslug-kernel = "2.6.11.2"
Jul 13 18:48:02 <jbowler> That log you just posted did not rebuilt the ixp modules.
Jul 13 18:48:22 <[g2]> no it's trying to build the image
Jul 13 18:48:41 <[g2]> it builds with the 2.6.12.2 kernel modules
Jul 13 18:48:49 <jbowler> <[g2]>  bb ixp4xx-csr ixp425-eth
Jul 13 18:48:58 <jbowler> <[g2]>   Attempted builds: 0
Jul 13 18:49:12 <[g2]> right they are already built
Jul 13 18:49:48 <jbowler> I don't understand what you have changed.  This isn't a clean system if you have that "PREFERRED_VERSION" line
Jul 13 18:50:05 <jbowler> Did you clean (rm -rf tmp) after adding that line (in openslug.conf?)
Jul 13 18:50:44 <[g2]> no. I bb -c clean openslug-kernel nslu2-kernel openslug-image ixp4xx-csr ixp425-eth
Jul 13 18:51:17 <[g2]> bb reparses all the files after touch the .conf
Jul 13 18:52:09 <[g2]> the -c clean *dosn't* remove the deploy/ipks
Jul 13 18:52:25 <jbowler> Ah, ok, that is what I was trying to find out.
Jul 13 18:52:32 <[g2]> I needed to remove the 2.6.12.2's via find
Jul 13 18:52:39 <jbowler> ipkg does not allow you to go backward.
Jul 13 18:52:59 <[g2]> so I should just nuke tmp/ipk
Jul 13 18:53:19 <jbowler> I do not know that fixing it by hand will work.  I never do that - when it gets too confusing I just start again.
Jul 13 18:53:58 <jbowler> In any case, once you have done an ipkg upgrade on a system the only way out is explicity ipkg remove then ipkg install.
Jul 13 18:54:20 <jbowler> Trying to do that to anything in the rootfs can be very difficult - one error and the system will be unbootable.
Jul 13 18:54:41 <[g2]> which is already is :)
Jul 13 18:54:50 <jbowler> (I.e. there is no viable ipkg downgrade - I guess one could be written, but I don't think anyone regards it as important).
Jul 13 18:55:05 <jbowler> Can you boot to flash?
Jul 13 18:55:21 <jbowler> (I.e. disconnect the disk and network if necessary then reboot)
Jul 13 18:55:43 <[g2]> it does, but the only kernel modules are the ixp ones all the others are 2.6.12.2 and it's a 2.6.11.2 kernel :)
Jul 13 18:56:34 <jbowler> Eh.  You ran ipkg upgrade on the flash?
Jul 13 18:56:51 <[g2]> No.. that's the way the rootfs got built
Jul 13 18:57:03 <[g2]> hence why I'm looking into this issue
Jul 13 18:57:41 <jbowler> Oh, ok.  Then upslug it with a valid rootfs, and run turnup disk -i -f <whatever>
Jul 13 18:58:00 <[g2]> a rm -rf tmp will likely solve it, but if one has built with a later rev pack then tmp is borked 
Jul 13 18:58:02 <jbowler> If it's 2.6.11.2 you want the OpenSlug-2.0 binary image will be fine.
Jul 13 18:58:43 <[g2]> yes thx, I'm well aware of how to fix the problem, I've been searching for it's root
Jul 13 18:59:14 <[g2]> the root appears as you've pointed out that the packages don't go backwards :)
Jul 13 18:59:36 <jbowler> ?  incremental builds don't work with bitbake.
Jul 13 18:59:43 <kolla_> hm, tv-shop is on, that's large message saying I'm up too late
Jul 13 19:00:17 <[g2]> jbowler, having built the 2.6.12.2 in the tree, trying to go back to 2.6.11.2 is not easy
Jul 13 19:00:31 <jbowler> bitbake thinks things are built when they are out of date, or, worse, inconsistent.
Jul 13 19:01:08 <[g2]> right and doing the -c clean purges bitbakes knowledge of those items
Jul 13 19:01:25 <jbowler> The bottom line is to rebuild from scratch whenever possible - because you never know what someone else changed in the last sync...
Jul 13 19:01:28 <[g2]> *however* ipk is still aware of them
Jul 13 19:02:00 <[g2]> we have different ideas there
Jul 13 19:02:04 <jbowler> Yes.  That's why the feeds have to be kept separate - one error, one too-new-ipk and everyones system is broken.
Jul 13 19:02:34 <jbowler> bbl
Jul 13 19:02:42 <[g2]> ciao
Jul 13 19:05:14 <kolla_> ssh works fine over ipv6 at least
Jul 13 19:06:14 <[g2]> cool
Jul 13 21:26:29 <jtucker> My slug no longer wants to boot!  The top green light flashes endlessly.  I was troubleshooting an Apache/PHP problem, had to reboot, and now it won't start!
Jul 13 21:28:12 <jtucker> Can anyone please offer help?  I tried a simple paperclip reset with no success, then I removed power for a few seconds and tried starting up; still no joy.
Jul 13 21:30:52 <jimtucker65> Sorry, I was disconncted for a minute by accident.  If anyone replied, I didn't hear.
**** ENDING LOGGING AT Wed Jul 13 23:59:56 2005