**** BEGIN LOGGING AT Fri May 06 23:59:56 2005
May 07 00:00:07 <rwhitby-away> OK, I've added the anonymiser patch to OpenSlug
May 07 00:00:26 <CIA-1> 03rwhitby 07 * r1.3232 10openembedded/packages/linux/ (2 files in 2 dirs): OpenSlug: added anonymiser patch
May 07 00:02:02 <rwhitby-away> sets the build identification to "openslug@downloads.openslug.org"
May 07 00:05:38 <jbowler> Anyone know what the BitKeeper 'BitKeeper allows a limited number of offline commits' message is about?
May 07 00:08:33 <ByronT> and the initial OpenSlug binary README is there as well
May 07 00:11:31 <rwhitby-away> jbowler: I just got that too
May 07 00:11:51 <ByronT> ditto - when I committed
May 07 00:12:10 <jbowler> I've been getting it for a while now, it started at '8' now I'm down to '1'.  I don't like this...
May 07 00:12:29 <rwhitby-away> I'm down to 1
May 07 00:12:49 <jbowler> I figured it was because of my massive set of changes in 'jonslug' so I got a bk repo (jonslug.bkbits.net) and pushed all my changes to that.  No help.
May 07 00:14:14 <jbowler> bk log -d is incomprehensible, the HTTP post fails with 'remote: ERROR-listkey failed' (that's not an HTTP error), then the 
May 07 00:14:36 <jbowler> 'direct connection' fails with 'drainErrorMsg: Unexpected response:' and a valid HTML first line.
May 07 00:16:23 <jacques> lie down with dogs...
May 07 00:17:36 <ByronT> it looks like I sent a borked message to the changesets list as well
May 07 00:17:50 <ByronT> that should have said something about a README checkin
May 07 00:27:05 <rwhitby-away> maybe we'll be using the SVN on berlios for more than just snapshot storage ....
May 07 00:27:52 <rwhitby-away> wonder if anyone on #oe is experiencing the same thing ....
May 07 00:29:23 <jbowler> I'm going to try a local openslug commit (of the 2.6.11-mm4 test kernel), that should take my count to 0.  I'll tell you what happens, unless, of course, McVoy pops out of the monitor with an AK47
May 07 00:31:50 <ByronT> good night all
May 07 00:42:57 <jbowler> No, weird, I've done another commit and the count is still at '1' (but I still get the message.)
May 07 00:51:47 <macsat_zzz> Is anyone here running xmail ?
May 07 00:57:11 <jbowler> bk commit: helptool for 'log' observes: bk commit will fail if there are more than 40 pending changesets that are older than one week.
May 07 00:57:58 <jbowler> But it seems that the pending log count is per primary repo, so I have 184 on nslu2-linux, 195 on jonslug (just created) and 1 on nslu2-package-symlinks
May 07 00:58:48 <jbowler> I know I've made several commits to nslu2-package-symlinks, so some are, or were, being logged.
May 07 01:04:14 <jbowler> I can explicitly 'bk log' the changesets in nslu2-package-symlinks and my new oe-scripts repo, my count is down to 0 there.  I can't for jonslug or openembedded/nslu2-linux, so I think there's a break in the OE repo.
May 07 01:18:00 <jbowler> If I reclone bk://jonslug.bkbits.net/jonslug I get a clone with 12649 logs pending (!), clearly the output of bk log -p is not very useful.
May 07 01:31:37 <rwhitby-away> I'm thinking of adding a postinst to the openslug kernel module which will flash the new kernel ....
May 07 01:32:30 <rwhitby-away> or maybe just add a /usr/sbin/update-kernel script ....
May 07 01:57:31 <jf-away> rwhitby-away: eggdrop is my first attempt to script the adduser package (in rc script plus postinst). May you have a look at those scripts and if they look corrects, I could update the BestPractice page
May 07 01:58:31 <macsat_zzz> do you have eggdrop running now? And added to the package stream ?
May 07 01:59:03 <jf-away> macsat_zzz : eggdrop is compiling fine. pushing it to the feeds
May 07 01:59:25 <macsat_zzz> okidoki, nice
May 07 01:59:52 <macsat_zzz> isnt it kindda memory intensive? I remmeber older versions where (some 3-5 years ago)
May 07 02:12:04 <rwhitby-away> jf-away: do you need to check for a ":" at the end of "eggdrop" in the greps?  Just in case someone creates an "eggdrophahafooledyou" user ....
May 07 02:12:47 <rwhitby-away> other than that, it looks good.
May 07 02:13:22 <rwhitby-away> Do we have a working "su" on both wileyware and unslung ?
May 07 02:14:10 <rwhitby-away> and you might want to kill any running eggdrop process in a prerm script.
May 07 02:14:48 <rwhitby-away> ah, jp30-work has added alternatives handling - nice.
May 07 02:14:48 <macsat_zzz> I dont see a "su" on wiley
May 07 02:15:17 <rwhitby-away> I wonder if update-alternatives exists on wiley ....
May 07 03:26:10 <jf-away> rwhitby-away: ok for the eggdrop:. As for the su, it lacks in the default busybox. is there an alternative ?
May 07 03:26:50 <dyoung> jf-away, did that slug slither its way to you yet?
May 07 03:27:02 <jf-away> dyoung: no :/
May 07 03:27:16 <dyoung> Bah.  "1 week" my a$$.
May 07 03:27:29 <dyoung> I guess 1 week only works for australia.
May 07 03:27:31 <jf-away> thursday was off in France, saturday is often like a strike day
May 07 03:27:37 <dyoung> hahaha okay.
May 07 03:28:57 <jf-away> I only find 4.5v adapter, 800mA
May 07 03:29:05 <jf-away> or 6v
May 07 03:29:33 <dyoung> lemme look inside.
May 07 03:31:54 <dyoung> Hmm.  the 4.5 MIGHT work... 
May 07 03:33:01 <dyoung> I'm not sure if the HC74 is being powered directly or if it goes through that PWM controller first.
May 07 03:33:28 <dyoung> or you could use the 6V with a diode or 3.  thats pretty UBE though.
May 07 03:35:12 <dyoung> ~ube
May 07 03:35:13 <jbot> methinks ube is Ugly But Effective.  Or sometimes Unorthrodox But Effective
May 07 04:30:08 <rwhitby> BK is officially broken, not just for nslu2-linux, but also for #oe.
May 07 04:48:36 <dyoung> bummers
May 07 04:48:50 <dyoung> so we exceeded the number of changesets?
May 07 04:49:08 <rwhitby> I think "we" includes everyone in #oe too.
May 07 04:49:16 <rwhitby> I think it's per-repo.
May 07 04:49:18 <dyoung> yeah, tahts kinda what I meant too...
May 07 04:49:27 <dyoung> Major bummers!
May 07 04:50:08 <rwhitby> I'm just installing cogito on the slug and the wl500gx now ....
May 07 04:51:45 <rwhitby> now, which of our packages includes /usr/bin/env ... ?  None :-(
May 07 04:53:03 <jacques> should be in coreutils
May 07 04:53:23 <jacques> and busybox
May 07 04:53:49 <rwhitby> ok, that rules out the wiley then, cause /usr/bin is readonly.
May 07 04:58:29 <rwhitby> so do I add a "ln -s /opt/bin/env /usr/bin/env" to coreutils ?
May 07 05:00:21 <jacques> is something explicitly looking in /usr/bin ?
May 07 05:00:22 <dyoung> ... or patch cogito (or whatever program) to obey path rules
May 07 05:00:29 <jacques> yeah
May 07 05:00:55 <jacques> on gentoo /usr/bin/env -> /bin/env
May 07 05:01:18 <rwhitby> heh - the main use of /usr/bin/env is to allow a script to obey path rules for finding bash ...
May 07 05:01:28 <jacques> lol
May 07 05:02:22 <dyoung> Hmm
May 07 05:05:28 <CIA-1> 03rwhitby * 10unslung/make/libcurl.mk: Now creates staging/bin before trying to copy something into it.
May 07 05:08:25 <rwhitby> so our only real choices are svn (on berlios), cvs (on sf.net) or cogito (on nudi).
May 07 05:11:22 <dyoung> I'm sure with a crisis like this the #oe folk will come up with a preferred solution soon.
May 07 05:11:59 <rwhitby> dunno, I was talking with mickeyl, and it sounded like they were waiting until June
May 07 05:11:59 <dyoung> but I agree its a PITA to have to halt things.
May 07 05:12:28 <dyoung> In the same breath though he figured it was a temporary problem.
May 07 05:12:40 <rwhitby> temporary bk problem
May 07 05:12:54 <dyoung> right.
May 07 05:12:58 <jacques> I'm not so confident
May 07 05:13:09 <dyoung> tempoary until june.
May 07 05:13:13 <jacques> I think bk will only get worse between now and shutdown
May 07 05:13:27 <jacques> the sooner we dont depend on it the better
May 07 05:13:43 <rwhitby> the question is whether we fork from #oe now ...
May 07 05:13:50 <rwhitby> or wait for them to decide
May 07 05:14:00 <jacques> yes, that is a tough one
May 07 05:14:27 <jacques> I guess they will be in just as much pain as we
May 07 05:15:05 <rwhitby> they may be too large to move quickly though ....
May 07 05:15:15 <jacques> another good point
May 07 05:15:20 <dyoung> Using the OE build system is rather convenient.
May 07 05:15:58 <dyoung> we could in theory retain the use of that data and migrate then sync later when they've come to a decision.
May 07 05:16:32 <dyoung> in which case SVN sounds like a good plan.
May 07 05:16:52 <rwhitby> then again, the change from oemake to bitbake was pretty swift ...
May 07 05:17:12 <rwhitby> anyone used SVN before?
May 07 05:17:30 <dyoung> not as a "core" developer.
May 07 05:17:48 <rwhitby> (as an aside, where would discussion of the bk problem be happening right now ?)
May 07 05:18:11 <dyoung> #bk-can-bite-me ?
May 07 05:18:19 <dyoung> j/k.
May 07 05:18:51 <dyoung> I would imagine that the kernel folk are probably buzzing bout it now.
May 07 05:19:43 <dyoung> I'll idle there and see what if anything comes up
May 07 05:22:25 <dyoung> does linus have a "official" home page someplace?
May 07 05:23:49 <dyoung> wow, kernel.org has 2 4-way opteron servers now.
May 07 05:23:56 <dyoung> I'm jealous.
May 07 05:37:58 * jacques starts an unslung cvs build for the first time in a long time.
May 07 05:40:34 * rwhitby installs the unslung-devel package so that he can compile the latest cogito on the slug
May 07 05:43:21 <rwhitby> dyoung: any traffic
May 07 05:45:52 <perlguru> hello
May 07 05:46:01 <rwhitby> g'day
May 07 05:46:23 <perlguru> i've got my wl-hdd, now waiting for the hdd
May 07 05:50:14 <rwhitby> all these people buying Asus gear now ...
May 07 05:52:30 <jacques> yeah, it's kinda scary
May 07 05:53:04 <rwhitby> perlguru: you know that the wl-hdd's disk to network performance is abysmal, right?
May 07 05:53:49 <rwhitby> Looks like jbowler got the last openembedded Changeset (1.3236).
May 07 05:56:28 <jacques> but they are so damn cute
May 07 05:56:44 <jacques> maybe it's just a driver issue - pio mode or something
May 07 05:56:51 <rwhitby> the wl-530g is the cute one
May 07 05:57:19 <jacques> yeah that is even cuter - but I haven't seen any hacking activity - no forum for it - have you?
May 07 05:57:33 <rwhitby> nope, but I have compiled the firmware for it
May 07 05:57:44 <jacques> what are the specs ?
May 07 05:57:51 <rwhitby> Marvell chip
May 07 05:57:58 <rwhitby> 16MB RAM
May 07 05:58:01 <jacques> for the proc ?
May 07 05:58:08 <jacques> 4MB flash ?
May 07 05:58:17 <rwhitby> yeah, I think so.
May 07 05:58:36 <jacques> hmm all these 16MB RAM devices.... 32 would be so much better
May 07 05:58:53 <rwhitby> yeah.  I'm not really considering hacking the wl530g much
May 07 05:59:19 <rwhitby> apart from adding the South Australian timezone ...
May 07 05:59:54 <rwhitby> the wl500gx is the Asus device for sluggers
May 07 06:00:02 <jacques> yeah
May 07 06:08:35 <jacques> hmm unslung cvs build doesn't get very far
May 07 06:08:50 <rwhitby> ?
May 07 06:09:06 <jacques> dies while configuring openldap-2.2.26
May 07 06:09:24 <rwhitby> got all the RequiredSoftware installed?
May 07 06:09:24 <jacques> configure: error: Berkeley DB version mismatch
May 07 06:09:54 <rwhitby> look above it for an autoconf error
May 07 06:10:47 <jacques> i dunno what to make of it
May 07 06:10:54 <jacques> checking for db.h... yes
May 07 06:10:55 <jacques> checking for Berkeley DB link (default)... no
May 07 06:10:55 <jacques> checking for Berkeley DB link (-ldb43)... no
May 07 06:10:55 <jacques> checking for Berkeley DB link (-ldb-43)... no
May 07 06:10:55 <jacques> checking for Berkeley DB link (-ldb-4.3)... no
May 07 06:10:57 <jacques> checking for Berkeley DB link (-ldb-4-3)... no
May 07 06:11:00 <jacques> checking for Berkeley DB link (-ldb42)... no
May 07 06:11:02 <jacques> checking for Berkeley DB link (-ldb-42)... no
May 07 06:11:04 <jacques> checking for Berkeley DB link (-ldb-4.2)... yes
May 07 06:11:06 <jacques> checking for Berkeley DB version match... no
May 07 06:11:22 <rwhitby> do you have autoconf and automake installed?
May 07 06:11:41 <jacques> heh
May 07 06:11:50 <jacques> this is a gentoo box (like all my others)
May 07 06:12:00 <dyoung> no traffic related to the issue at hand.
May 07 06:12:21 <jacques> so yes
May 07 06:13:12 <dyoung> openldap.  Wasnt there a issue/problem with that brought up in the last 24 hours?
May 07 06:13:23 <jacques> oh well, that's what make -k is for
May 07 06:13:37 <dyoung> something about the new samba.
May 07 06:13:46 <perlguru> jacques, did you read the gentoo newsletter?
May 07 06:14:10 <jacques> perlguru, no
May 07 06:14:22 <perlguru> has a topic on openldap 2.2
May 07 06:15:04 <jacques> but why would that matter for me cross-compiling it?
May 07 06:15:31 <dyoung> was that directed toward me?
May 07 06:16:09 <jacques> no toward perlguru 
May 07 06:16:12 <dyoung> Not sure if someone "did something" to it recently to facilitate and/or hack the mk to make it work with new samba.
May 07 06:16:14 <dyoung> oh.
May 07 06:16:15 <dyoung> :-)
May 07 06:16:39 <jacques> whatever problems gentoo is having with openldap, I don't see why they should affect unslung cvs
May 07 06:18:35 <perlguru> it is just an announcement. expect config troubles. but it shouldn't affert compilling...
May 07 06:20:46 <caplink811-away> jacques: openldap couln't compile cross
May 07 06:20:59 <caplink811-away> s/couln't/couldn't
May 07 06:21:01 <jacques> caplink811-away, yeah that's the problem I'm having
May 07 06:22:09 <caplink811-away> refers to the MainMakefile -> # openldap runs its own binaries at compile-time and expects them to have same byte-order as target
May 07 06:22:35 <caplink811-away> so that's the reason while it coulnd't compile cross
May 07 06:23:50 <jacques> caplink811-away, mine doesn't even get past the configure stage 
May 07 06:24:56 <caplink811-away> Jacmet: hm, 2.23 had configured and compiled, but 2.26 i haven't try yet
May 07 06:25:09 <rwhitby> jacques: just configured for me, and is compiling now ...
May 07 06:25:17 <jacques> okay
May 07 06:25:20 <jacques> so it's broken
May 07 06:25:39 <rwhitby> yep, not robust enough
May 07 06:25:53 <jacques> like all the other packages that were allowing themselves to be contaminated by native versions installed
May 07 06:26:29 <rwhitby> yep
May 07 06:26:36 <jacques> it looks like it's being compiled cross because it's a dep of something - because it's in NATIVE_PACKAGES list only
May 07 06:27:05 <rwhitby> php-ldap probably
May 07 06:27:21 <rwhitby> (php.mk)
May 07 06:27:27 <jacques> oh crap
May 07 06:27:31 <jacques> I just thought of something
May 07 06:27:49 <jacques> this is tha machine I setup to run armeb binaries automatically
May 07 06:27:57 <jacques> this might be... interesting
May 07 06:28:36 <jacques> some configures might try to run a binary and if they can't, make the (correct) assumption that they are cross-compiling and do the right thing
May 07 06:28:38 <caplink811-away> if i remember, jp30 had compiled openldap libraries cross
May 07 06:28:54 <jacques> well what would happen if they actually could run the armeb binaries
May 07 06:29:02 <jacques> probably nothing good
May 07 06:31:09 <jacques> going to start a build on another machine
May 07 06:31:21 <caplink811-away> btw, configure (cross) runs through, but compile dies at
May 07 06:32:16 <rwhitby> mine compiles fine, and just dies in -ipk when running install -s (wrong strip)
May 07 06:33:05 <caplink811-away> oops, compile doesn't die... rwhitby: yes, so it is, when i remember right
May 07 06:33:31 <jacques> belnet.dl.sourceforge.net is not responding
May 07 06:33:44 <jacques> I thought we were getting away from explicit sf mirrors
May 07 06:34:44 <jacques> dl.sourceforge.net worked fine
May 07 06:37:10 <rwhitby> ok, I have a 17MB tarball of the latest openembedded source tree (everything).  Should I check it into SVN on berlios, or CVS on sf.net?
May 07 06:37:28 <rwhitby> (so we can continue development while #oe is offline)
May 07 06:37:38 <rwhitby> s/#oe/bkbits.net/
May 07 06:40:01 <jacques> hmm, pros and cons, svn is better than cvs but do we need its features? I like it personally. is berlios as fast (or rather slow) as sf.net ?
May 07 06:40:11 <rwhitby> dunno
May 07 06:40:29 <rwhitby> everything outside of AU is slow for me ...
May 07 06:41:07 <jacques> the number of packages in unslung cvs now is mind-boggling
May 07 06:41:41 <jacques> >320
May 07 06:41:53 <rwhitby> Optware :-)
May 07 06:43:13 <jacques> do we need a new channel?  :-)
May 07 06:43:27 <dyoung> are we smart enough to transfer keys and dev status along with said source tree?
May 07 06:44:06 <rwhitby> individuals would have to upload keys to berlios, and we add them as developers there
May 07 06:45:12 <jacques> my key is working great on berlios now :-) (after however many hours)
May 07 06:45:19 <rwhitby> sf.net for packages, berlios for firmware.  At least we can keep the two sets of developers as is.
May 07 06:47:00 <jacques> hmm we might wanna update dropbear to 0.45
May 07 06:48:21 <dyoung> how many bk sla.. err, devs do we  have?
May 07 06:50:08 <rwhitby> 16
May 07 06:53:04 <rwhitby> ok, importing to SVN on berlios now
May 07 06:53:30 <dyoung> part of the openslug project?
May 07 06:53:48 <rwhitby> yep - openslug/trunk
May 07 06:54:04 <rwhitby> serves as our first snapshot anyway
May 07 06:57:31 <rwhitby> I took the openembedded dir from nudi, removed all the SCCS dirs, removed the BitKeeper dir, removed the ChangeSets file.  That's what's currently uploading to svn on berlios.
May 07 06:58:40 <jacques> cool
May 07 07:01:24 <rwhitby> Committed revision 1. :-)
May 07 07:07:46 <jacques> heh yeah, looks like my openldap problem was specific to that machine - probably the qemu stuff
May 07 07:08:08 <rwhitby> ok, you can get our berlios repo with "svn checkout svn+ssh://svn.berlios.de/svnroot/repos/openslug/trunk"
May 07 07:09:10 <rwhitby> I probably should have put it another level down (../openslug/firmware/trunk or something)
May 07 07:12:29 <rwhitby> 306ms ping to svn.berlios.de - not too bad
May 07 07:12:35 <rwhitby> s/306/360/
May 07 07:12:59 <jacques> about 200ms for me
May 07 07:13:30 <jf-away> 80 from here ;)
May 07 07:13:37 <rwhitby> heh, it was faster to check it out here in AU than it was on nudi.
May 07 07:14:07 <dyoung> 230 from here, but I'm using a lot of local bw right now.
May 07 07:15:30 <jacques> rwhitby, what are your pings to nudi?
May 07 07:16:35 <rwhitby> 230 to 300, ave 260.
May 07 07:19:28 <rwhitby> 22 hops to svn.berlios.de from here
May 07 07:20:18 <rwhitby> and 20 to nudi :-)
May 07 07:21:02 <ka6sox> 1 hop to nudi from here....24 hops to berlios.de
May 07 07:21:11 <dyoung> um.....
May 07 07:21:39 <ka6sox> oh...wait..that was the tunnel.
May 07 07:22:03 <rwhitby> ok, building openslug-packages from the svn repo checkout ....
May 07 07:22:19 <jacques> what's openslug-packages?
May 07 07:22:29 <rwhitby> image plus known good packages
May 07 07:23:04 <jacques> oh cool
May 07 07:23:55 <rwhitby> ok, the berlios repo is good enough for bitbake to realise it doesn't have to do much.
May 07 07:25:40 <rwhitby> Hmm - need an svn mode for emacs ....
May 07 07:28:43 <rwhitby> (require 'vc-svn)
May 07 07:28:51 <[cc]smart> moving to svn ? is there sth. that should be declard on homepage ?
May 07 07:29:22 <rwhitby> not moving yet, just using it as a place to continue while bk is down
May 07 07:31:46 <NAiL> rwhitby: "known good packages" include mysql, which doesn't compile because of linuxthreads :(
May 07 07:33:40 <rwhitby> NAiL: that's why the default is glibc :-)
May 07 07:33:53 <jacques> ah heh
May 07 07:33:57 <NAiL> rwhitby: This is the glibc build.
May 07 07:34:15 <rwhitby> I have it built here.
May 07 07:34:15 <jacques> hmm
May 07 07:34:17 <jacques> me too
May 07 07:34:30 <NAiL> rwhitby: It *has* built earlier. But not now :(
May 07 07:34:48 <rwhitby> 4.1.10a?
May 07 07:34:50 <NAiL> It won't build at all with uClibc then?
May 07 07:34:56 <rwhitby> dunno
May 07 07:35:20 <NAiL> 4.1.10a-r3
May 07 07:35:38 <rwhitby> let me rebuild it now to confirm
May 07 07:36:11 <rwhitby> Hmm - we need to get the berlios repo talking to CIA
May 07 07:39:18 <NAiL> rwhitby: Just to clarify, my build is glibc, and it's completely from scratch following the StepByStep guide. I set up another tree so I could do both glibc and uClibc builds easily.
May 07 07:41:40 <rwhitby> [cc]smart: I have updated the homepage
May 07 07:50:00 <rwhitby> and sent a message to the mailing lst
May 07 08:02:31 <rwhitby> NAiL: mysql-native completed, and mysql is still compiling ....
May 07 08:05:30 <NAiL> odd
May 07 08:05:55 <rwhitby> NAiL: mysql completed successfully
May 07 08:06:07 <rwhitby> I only rebuilt from scratch, nothing else
May 07 08:06:20 <rwhitby> rebuilt mysql from scratch
May 07 08:06:43 <NAiL> strange
May 07 08:07:03 <rwhitby> I'll start a complete rebuild everything from scratch now (to test the berlios repo properly) and it will run overnight
May 07 08:07:30 <NAiL> ok, we'll see if it fails then ;)
May 07 08:07:52 <NAiL> iirc, it has built before
May 07 08:09:02 <rwhitby-asleep> night all
May 07 08:09:14 <jacques> `night rwhitby-asleep 
May 07 08:09:36 <NAiL> nite
May 07 08:53:19 <NAiL> jbowler: My slug hung this morning. I assume it was NTP :(
May 07 08:53:56 <NAiL> (but I'm still running 4.1.2)
May 07 08:55:16 <jbowler> NAiL: ah, 4.1.2...  If it's causing a hang it will be the same in 4.2.0
May 07 08:55:31 <NAiL> *sigh*
May 07 09:15:46 <NAiL> rwhitby-asleep: When you come back... MySQL failing on LinuxThreads wouldn't have anything to do with my gentoo-box being set up using nptl?
May 07 10:23:26 <NAiL> rwhitby-asleep: Yes. nptl is the problem. It's fixed in MySQL 4.1.2.
May 07 10:23:46 <NAiL> err. 4.1.12
May 07 10:30:46 <Christopher-bbl> jp30-work: phpmyadmin also conplains about there not being an mbstring extension in php. Are there plans to compile this in?
May 07 10:32:06 <NAiL> (but apparently, 4.1.12 hasn't been released yet)
May 07 10:55:17 <NAiL> NOTE: package mysql-native-4.1.10a: completed
May 07 10:55:23 <NAiL> Voilá!
May 07 10:56:44 <ka6sox> sweet
May 07 11:13:28 <NAiL> NOTE: package mysql-4.1.10a: completed
May 07 11:13:35 <NAiL> Both works. Finally :)
May 07 11:14:55 <ka6sox> cross and native?
May 07 11:15:04 <NAiL> yeah
May 07 11:15:20 <ka6sox> now all we need is for it to work under uclibc for Wiley!
May 07 11:15:40 <NAiL> Doesn't it work? Mine's compiled for uclibc
May 07 11:15:46 <ka6sox> :)
May 07 11:15:58 <NAiL> Haven't installed and tested yet
May 07 11:16:00 <ka6sox> oh..I thought it was for Unslung
May 07 11:16:11 <NAiL> ah, no. openslug
May 07 11:16:18 <ka6sox> k
May 07 11:21:44 <jacques> jeez
May 07 11:21:52 <jacques> the compressed samba source is 15MB
May 07 11:21:57 <jacques> what is it doing?
May 07 11:23:09 <NAiL> heh
May 07 11:32:00 <ka6sox> morning jp30
May 07 11:32:26 <jp30> morning, ka6sox
May 07 11:33:26 <jp30> i got a reply from dan kegel about the fix-embedded-paths issue. he said to send him exact details of our crosstool config and he would look at it
May 07 11:34:40 <ka6sox> very good.
May 07 11:57:22 <jstueve_away> samba failed cross for me.
May 07 11:58:21 <eno> jstueve_away: error message?
May 07 11:58:35 <jstueve_away> looking in the log now...
May 07 11:59:37 <jstueve_away> /home/slug/build/tmp/staging/armeb-linux-uclibc/include/linux/posix_types.h:44:
May 07 11:59:37 <jstueve_away> error: redefinition of typedef '__kernel_key_t'
May 07 11:59:37 <jstueve_away> /home/slug/build/tmp/staging/armeb-linux-uclibc/include/bits/ipc.h:39: error: previous declaration of '__kernel_key_t' was here
May 07 11:59:37 <jstueve_away> make: *** [client/smbmount.o] Error 1
May 07 11:59:37 <jstueve_away> FATAL: oe_runmake failed
May 07 12:00:49 <eno> oe and uclibc
May 07 12:01:16 <eno> how about hosting uname?
May 07 12:01:25 <jstueve> hm?
May 07 12:01:36 <jstueve> yes on the first name?
May 07 12:02:01 <jstueve> hosting uname is: Linux
May 07 12:02:55 <eno> ok, i only tried under unslung and glibc
May 07 12:05:41 <jstueve> thnks...
May 07 12:08:59 <eno> could be similar to this: http://64.233.179.104/search?q=cache:BG1sDDHN580J:ibot.rikers.org/uclibc/20020904.html.gz+__kernel_key_t+uclibc&hl=en&start=1
May 07 12:22:49 <jbowler> jstueve: check the definition of __kernel_key_t on line 39 of bits/ipc.h, it should be typedef int
May 07 12:23:01 <jstueve> checking
May 07 12:25:47 <jstueve> it is int in ipc.h
May 07 12:27:10 <jbowler> Hum...  What is the one in linux/posix_types.h then (on all my builds the posix_types definition is consistently 'int' too).
May 07 12:29:24 <jstueve> same int
May 07 12:29:34 <jstueve> might not have cleared the directory 
May 07 12:30:21 <jbowler> Ah.  'redefintion' - it's complaining that the same thing was defined twice, even though it's identical.  Bet there is a gcc warning option for that.
May 07 12:31:51 <jstueve> ahh... talking to NAiL he said it is trying to compile the smbmount and we've defined --without-smbmount
May 07 12:32:43 <jstueve> said to delete the build/tmp/work/samba directory... but I didn't define the samba-3.0.10-r0
May 07 12:32:54 <jstueve> for the want of a tag, a samba build failed...
May 07 12:33:17 <jbowler> The .bb file explicitly specifies --with-smbmount
May 07 12:33:46 <NAiL> which doesn't necessarily compile :(
May 07 12:33:59 <NAiL> removing smbmount (Should be enough to remove --with-smbmount probably)
May 07 12:34:06 <NAiL> allows samba to compile
May 07 12:34:22 <jstueve> hrm.. failed again
May 07 12:35:55 <jbowler> Doesn't matter, is a uclibc bug because the glibc (etc) kernel headers simply don't include the typedef.
May 07 12:40:53 <NAiL> jstueve: My samba build (uclibc) is up at my repository though ;)
May 07 12:41:34 <jstueve> good enough... :
May 07 12:43:21 <NAiL> jbowler: I don't have to do anything at all after installing the ntp ipkg, do I? It set tickadj and started ntpd itself...
May 07 12:43:28 <NAiL> the 4.2.0 one
May 07 12:43:59 <jbowler> NAiL: that's correct.  I believe you will find that the ntp drift is a lot less than 140 this time...
May 07 12:44:47 <NAiL> jbowler: Possible. I ended up at 129 before the slug crashed this morning
May 07 12:45:49 <jbowler> I have netconsole operating now, so there's a possibility of debugging this kind of thing without serial...  (The changes are ready to push, but I can't commit them.)
May 07 12:46:05 <NAiL> heh, yeah...
May 07 12:46:51 <jbowler> That extra setting in linux/posix_types.h is coming from linux-libc-headers (uclibc actually installs these, glibc doesn't)
May 07 12:52:00 <jstueve> so is that something that can be fixed?
May 07 12:53:16 <jbowler> I'm guessing that the fact this is missing from glibc means someone did the fix elsewhere.  Certainly the linux header package can be patched, but it's a core package so it's difficult to do it safely.
May 07 12:53:34 * jstueve nods
May 07 12:53:41 <ka6sox> jbowler, could get dicey
May 07 12:54:56 <jbowler> ka6sox: oh, I don't know: PATCHposix-types_openslug = 'fix-posix-types.patch;patch=1' would be safe...
May 07 12:57:16 <jbowler> ;-) anyway, I got the wrong file.  It's bits/ipc.h which differs...
May 07 12:59:47 <ka6sox> ah.
May 07 13:00:05 <copperbeech> jp30:ping
May 07 13:00:14 <jp30> hi, copperbeech
May 07 13:00:40 <copperbeech> jp30: native build of mysql completed - but fell over during configure of php
May 07 13:01:05 <CIA-1> 03jeanfabrice * 10unslung/make/syslog-ng.mk: New version : 1.6.7
May 07 13:01:46 <jp30> copperbeech, what was the problem with php?
May 07 13:01:49 <copperbeech> checking for LDAP support... yes, shared
May 07 13:01:50 <copperbeech> checking for LDAP Cyrus SASL support... /share/hdd/data/public/cvs/unslung/staging/opt
May 07 13:01:52 <copperbeech> checking for 3 arg ldap_set_rebind_proc... yes
May 07 13:01:53 <copperbeech> checking for ldap_parse_reference... no
May 07 13:01:55 <copperbeech> checking for ldap_start_tls_s... no
May 07 13:01:56 <copperbeech> checking for sasl_version in -lldap... no
May 07 13:01:58 <copperbeech> configure: error: LDAP SASL check failed. Please check config.log for more information.
May 07 13:01:59 <copperbeech> make[1]: *** [/share/hdd/data/public/cvs/unslung/builds/php/.configured] Error 1
May 07 13:02:01 <copperbeech> make[1]: Leaving directory `/share/hdd/data/public/cvs/unslung'
May 07 13:02:14 <copperbeech> that's the last few lines of the build output...
May 07 13:02:37 <copperbeech> it seemed to fial on the LDAP SASL (whatever that is?)
May 07 13:02:58 <jp30> hmm. well LDAP certainly builds correctly native
May 07 13:04:50 <jp30> it might be an openldap version mismatch, openldap has been upgraded recently. you might need to disclean to make sure the openldap in staging is consistent. i'm sure that's not what you want to hear, though
May 07 13:04:56 <copperbeech> jp30: LDAP seemed to have built ok - but php seems not to find it... 
May 07 13:05:15 <copperbeech> disclean? 
May 07 13:06:56 <jp30> sorry, distclean. i suspect that fixing this might not be worth the effort. do you really not have access to a PC that could run the cross build env?
May 07 13:07:28 <copperbeech> I take it distclean would remove everything that has already built?
May 07 13:07:38 <jbowler> SASL: just pass the '--without' option to configure, or ipkg remove sasl before building it.
May 07 13:08:00 <copperbeech> what is LDAP SASL ?
May 07 13:08:07 <jbowler> it == openldap (i.e. if SASL isn't there openldap won't try to build support in)
May 07 13:08:31 <jbowler> Secure Access Sockets Layer.  It's enormous, but it does built...
May 07 13:08:53 <jp30> jbowler, openldap stages cyrus-sasl
May 07 13:09:17 <copperbeech> I think that openldap built SASL fine... it's to do with the version checking of LDAP SASL and php if I understand jp30 correctly. 
May 07 13:09:45 <jp30> copperbeech, you would need to look at config.log and see how the failed test works, and why it is failing
May 07 13:09:56 <copperbeech> hold on a sec...
May 07 13:10:59 <jbowler> From OpenLDAP configure (the 2.2.24 version in bitbake/oe): --with-cyrus-sasl       with Cyrus SASL support auto
May 07 13:13:01 <jbowler> Hum, and the bitbake build isn't explicitly turning it off either - it just fails at present because it isn't built earlier (this is openslug-packages, but it doesn't matter.)
May 07 13:14:07 <copperbeech> ok the bit of configure.log which seems relevant is as follows:
May 07 13:14:16 <copperbeech> configure:59697: /opt/armeb/bin/armv5b-softfloat-linux-gcc -o conftest -I/opt/include -O2    -I/share/hdd/data/public/cvs/unslung/staging/opt/include -I/share/hdd/data/public/cvs/unslung/staging/opt/include/libxml2 -I/share/hdd/data/public/cvs/unslung/staging/opt/include/libxslt -I/share/hdd/data/public/cvs/unslung/staging/opt/include/libexslt  -L/opt/lib -L/share/hdd/data/public/cvs/unslung/stagi
May 07 13:14:18 <copperbeech> ng/opt/lib -Wl,-rpath,/opt/lib -Wl,-rpath-link,/share/hdd/data/public/cvs/unslung/staging/opt/lib -L/share/hdd/data/public/cvs/unslung/staging/opt/lib/mysql -Wl,-rpath=/opt/lib/mysql -ldl -pthread -I/opt/include -O2    -I/share/hdd/data/public/cvs/unslung/staging/opt/include -I/share/hdd/data/public/cvs/unslung/staging/opt/include/libxml2 -I/share/hdd/data/public/cvs/unslung/staging/opt/include/li
May 07 13:14:19 <copperbeech> bxslt -I/share/hdd/data/public/cvs/unslung/staging/opt/include/libexslt  -D_REENTRANT -Wl,-rpath,/share/hdd/data/public/cvs/unslung/staging/opt/lib -L/share/hdd/data/public/cvs/unslung/staging/opt/lib -Wl,-rpath,/share/hdd/data/public/cvs/unslung/staging/opt/lib -L/share/hdd/data/public/cvs/unslung/staging/opt/lib -Wl,-rpath,/share/hdd/data/public/cvs/unslung/staging/opt/lib -L/share/hdd/data/publ
May 07 13:14:22 <copperbeech> ic/cvs/unslung/staging/opt/lib -L/share/hdd/data/public/cvs/unslung/staging/opt/lib -L/opt/lib -L/share/hdd/data/public/cvs/unslung/staging/opt/lib -Wl,-rpath,/opt/lib -Wl,-rpath-link,/share/hdd/data/public/cvs/unslung/staging/opt/lib -L/share/hdd/data/public/cvs/unslung/staging/opt/lib/mysql -Wl,-rpath=/opt/lib/mysql -ldl -Wl,-rpath,/opt/lib -L/opt/lib -lldap -llber -lsasl2 conftest.c...
May 07 13:14:27 <copperbeech> configure:59667: checking for sasl_version in -lldap
May 07 13:14:29 <copperbeech> ...-lldap  -ldb-4.2 -lresolv -lm -lnsl  -lxml2 -lz -lm >&5
May 07 13:14:31 <copperbeech> /share/hdd/data/public/cvs/unslung/staging/opt/lib/libldap.so: undefined reference to `SSL_CTX_set_tmp_rsa_callback'
May 07 13:14:34 <copperbeech> /share/hdd/data/public/cvs/unslung/staging/opt/lib/libldap.so: undefined reference to `ssl3_send_alert'
May 07 13:14:36 <copperbeech> /share/hdd/data/public/cvs/unslung/staging/opt/lib/libldap.so: undefined reference to `sk_new_null'
May 07 13:14:38 <copperbeech> ---[snip]---
May 07 13:14:40 <copperbeech> /share/hdd/data/public/cvs/unslung/staging/opt/lib/libldap.so: undefined reference to `SSL_get_verify_result'
May 07 13:14:43 <copperbeech> /share/hdd/data/public/cvs/unslung/staging/opt/lib/libldap.so: undefined reference to `SSL_CTX_set_client_CA_list'
May 07 13:14:46 <copperbeech> /share/hdd/data/public/cvs/unslung/staging/opt/lib/libldap.so: undefined reference to `SSL_write'
May 07 13:14:48 <copperbeech> collect2: ld returned 1 exit status
May 07 13:14:50 <copperbeech> configure:59703: $? = 1
May 07 13:14:52 <copperbeech> configure: failed program was:
May 07 13:14:54 <copperbeech> | /* confdefs.h.  */
May 07 13:14:56 <copperbeech> | 
May 07 13:14:58 <copperbeech> | #define PACKAGE_NAME ""
May 07 13:15:00 <copperbeech> | #define PACKAGE_TARNAME ""
May 07 13:15:02 <copperbeech> | #define PACKAGE_VERSION ""
May 07 13:15:04 <copperbeech> ---[snip]---
May 07 13:15:06 <copperbeech> | #define HAVE_3ARG_SETREBINDPROC 1
May 07 13:15:08 <copperbeech> | #define HAVE_LDAP_SASL_SASL_H 1
May 07 13:15:10 <copperbeech> | /* end confdefs.h.  */
May 07 13:15:12 <copperbeech> | 
May 07 13:15:14 <copperbeech> | /* Override any gcc2 internal prototype to avoid an error.  */
May 07 13:15:16 <copperbeech> | #ifdef __cplusplus
May 07 13:15:18 <copperbeech> | extern "C"
May 07 13:15:20 <copperbeech> | #endif
May 07 13:15:22 <copperbeech> | /* We use char because int might match the return type of a gcc2
May 07 13:15:24 <copperbeech> |    builtin and then its argument prototype would still apply.  */
May 07 13:15:26 <copperbeech> | char sasl_version ();
May 07 13:15:28 <copperbeech> | int
May 07 13:15:30 <copperbeech> | main ()
May 07 13:15:32 <copperbeech> | {
May 07 13:15:34 <copperbeech> | sasl_version ();
May 07 13:15:36 <copperbeech> |   ;
May 07 13:15:38 <copperbeech> |   return 0;
May 07 13:15:40 <copperbeech> | }
May 07 13:15:42 <copperbeech> configure:59729: result: no
May 07 13:15:44 <copperbeech> configure:59876: error: LDAP SASL check failed. Please check config.log for more information.
May 07 13:16:08 <jp30> jbowler, there is no need to disable cyrus-sasl, unslung has a working cyrus-sasl and openldap. it's just that copperbeech is trying to build in a non-standard way
May 07 13:16:11 <copperbeech> seems to be a load of lines (snipped out) talking about undefined references....
May 07 13:16:38 <jbowler> That configure.log means there is a -lsasl missing I believe.
May 07 13:17:16 <jp30> and -lssl, it looks like
May 07 13:17:41 <jp30> i wonder why it works when cross-compiling?
May 07 13:18:07 <jp30> this could be an -rpath-link issue
May 07 13:18:28 <jbowler> jp30: yes, -lssl, not -lsasl
May 07 13:20:40 <copperbeech> sooo.. what does that mean in terms of tracing this problem ? Most of this it a bit outside my experience at the mo...
May 07 13:21:25 <jp30> copperbeech, we are configuring php here, correct?
May 07 13:21:48 <jp30> (ie. openldap built correctly already?)
May 07 13:22:18 <copperbeech> I'm trying to build php-stage - so yes the problem comes when configuring php - openldap is built and staged...
May 07 13:27:52 <jp30> ok, copperbeech, i have something for you to check.
May 07 13:28:15 <copperbeech> jp30: go for it...
May 07 13:29:27 <jp30> i'm going to explain why first... the problem appears to be that symbols starting with SSL_ are missing, and as jbowler pointed out the -lssl flag wasn't given to the failing link command...
May 07 13:29:48 <copperbeech> just for the record - I have built the following packages natively:
May 07 13:29:50 <copperbeech> openldap
May 07 13:29:52 <copperbeech> cyrus-sasl
May 07 13:29:53 <copperbeech> imap
May 07 13:29:55 <copperbeech> mysql
May 07 13:29:56 <copperbeech> readline
May 07 13:29:58 <copperbeech> ncurses
May 07 13:29:59 <copperbeech> libxslt
May 07 13:30:01 <copperbeech> libxml2
May 07 13:30:02 <copperbeech> libgd
May 07 13:30:04 <copperbeech> fontconfig
May 07 13:30:06 <copperbeech> expat
May 07 13:30:07 <copperbeech> freetype
May 07 13:30:09 <copperbeech> libjpeg
May 07 13:30:10 <copperbeech> libpng
May 07 13:30:12 <copperbeech> libdb
May 07 13:30:14 <copperbeech> gdbm
May 07 13:30:15 <copperbeech> bzip2
May 07 13:30:17 <copperbeech> libstdc++
May 07 13:30:18 <copperbeech> openssl
May 07 13:30:20 <copperbeech> zlib
May 07 13:30:21 <copperbeech> - which just goes to show the great work the team did when creating the native build system!
May 07 13:30:25 <copperbeech> jp30: ok yes...
May 07 13:30:40 <jp30> that shouldn't be a problem, because the ldap library should contain a record that tells the linker that it requires libssl... clear so far?
May 07 13:30:48 <copperbeech> yep
May 07 13:30:59 <jp30> so what we are going to do is look to see if your libldap contains that record
May 07 13:31:05 <copperbeech> ok 
May 07 13:32:23 <jp30> to do this, you need to invoke "objdump -p staging/opt/lib/libldap.so", and look to see if there's a line that says "NEEDED      libssl.so.0.9.7"
May 07 13:32:32 <copperbeech> for the record: openssl and zlib were built april 16th
May 07 13:32:40 <copperbeech> jp30: ok hold on a sec
May 07 13:33:59 <copperbeech> which ipkg has objdump :-)
May 07 13:34:04 * copperbeech grins
May 07 13:34:14 <jp30> it is part of crosstools-native
May 07 13:35:38 <jp30> ah, on a slug, try /opt/armeb/bin/armv5b-softfloat-linux-objdump 
May 07 13:38:44 <copperbeech> unslung/staging/opt/lib/libldap.so:     file format elf32-bigarm
May 07 13:38:46 <copperbeech> Program Header:
May 07 13:38:47 <copperbeech>     LOAD off    0x00000000 vaddr 0x00000000 paddr 0x00000000 align 2**15
May 07 13:38:49 <copperbeech>          filesz 0x0002f5fc memsz 0x0002f5fc flags r-x
May 07 13:38:50 <copperbeech>     LOAD off    0x0002f5fc vaddr 0x000375fc paddr 0x000375fc align 2**15
May 07 13:38:52 <copperbeech>          filesz 0x00001942 memsz 0x0000197c flags rw-
May 07 13:38:53 <copperbeech>  DYNAMIC off    0x0002f864 vaddr 0x00037864 paddr 0x00037864 align 2**2
May 07 13:38:55 <copperbeech>          filesz 0x000000d8 memsz 0x000000d8 flags rw-
May 07 13:38:57 <copperbeech> Dynamic Section:
May 07 13:38:58 <copperbeech>   NEEDED      libc.so.6
May 07 13:39:00 <copperbeech>   SONAME      libldap-2.2.so.7
May 07 13:39:01 <copperbeech>   RPATH       /opt/lib
May 07 13:39:03 <copperbeech>   INIT        0x88f8
May 07 13:39:04 <copperbeech>   FINI        0x2b444
May 07 13:39:06 <copperbeech>   HASH        0x94
May 07 13:39:07 <copperbeech>   STRTAB      0x3548
May 07 13:39:09 <copperbeech>   SYMTAB      0x11a8
May 07 13:39:10 <copperbeech>   STRSZ       0x2504
May 07 13:39:12 <copperbeech>   SYMENT      0x10
May 07 13:39:14 <copperbeech>   PLTGOT      0x3793c
May 07 13:39:15 <copperbeech>   PLTRELSZ    0xbe8
May 07 13:39:17 <copperbeech>   PLTREL      0x11
May 07 13:39:19 <copperbeech>   JMPREL      0x7d10
May 07 13:39:21 <copperbeech>   REL         0x5f00
May 07 13:39:23 <copperbeech>   RELSZ       0x1e10
May 07 13:39:25 <copperbeech>   RELENT      0x8
May 07 13:39:27 <copperbeech>   TEXTREL     0x0
May 07 13:39:29 <copperbeech>   VERNEED     0x5ec0
May 07 13:39:31 <copperbeech>   VERNEEDNUM  0x1
May 07 13:39:33 <copperbeech>   VERSYM      0x5a4c
May 07 13:39:35 <copperbeech>   RELCOUNT    0x3aa
May 07 13:39:37 <copperbeech> Version References:
May 07 13:39:39 <copperbeech>   required from libc.so.6:
May 07 13:39:41 <copperbeech>     0x09691f73 0x00 04 GLIBC_2.1.3
May 07 13:39:43 <copperbeech>     0x0d696911 0x00 03 GLIBC_2.1
May 07 13:39:45 <copperbeech>     0x0d696910 0x00 02 GLIBC_2.0
May 07 13:39:47 <copperbeech> private flags = 2: [APCS-32] [FPA float format] [has entry point]
May 07 13:39:49 <copperbeech> not there!
May 07 13:41:07 <copperbeech> jp30: wonder why that's not got the reference?
May 07 13:44:03 <jp30> the official libldap.so doesn't either... so it's not your system which is at fault
May 07 13:44:30 <jp30> next step is to look at how libldap.so was built
May 07 13:44:50 <jp30> cd into builds/openldap/libraries/libldap
May 07 13:47:28 <jp30> ...and use objdump in the same way on .libs/libldap.so. don't copy the whole output of objdump to the channel this time, just look to see if there's a NEEDED line for libssl
May 07 13:55:02 <copperbeech> ok...
May 07 13:57:27 <copperbeech>   NEEDED      liblber-2.2.so.7
May 07 13:57:29 <copperbeech>   NEEDED      libc.so.6
May 07 13:57:31 <copperbeech> is all that is in there
May 07 13:59:06 <jp30> that's interesting; that's changed since the last time i built libldap natively
May 07 13:59:56 <jp30> ok, well, the quick symptomatic relief for your problem is to add -lssl to PHP_LDFLAGS
May 07 14:00:23 <jp30> ...do you understand why?
May 07 14:00:25 <jbowler> jstueve, NAiL: I have a fix for the uclibc ipc.h problem, need to do a full build to verify it (I made it NSLU2 only for safety).  samba builds from OE without problem now.
May 07 14:02:07 <copperbeech> yes I think so - a library normally lists what other libraries it depends on, and normally the ssl library would be listed in one of the NEEDED lines for the linked libraries
May 07 14:02:23 <copperbeech> in this case is isn't called up by the dependant libraries
May 07 14:03:58 <copperbeech> so we force it by adding it to the LD flags in the make file
May 07 14:07:15 <jp30> ok, good. so i suggest you try that and see how much further you get.
May 07 14:10:26 <copperbeech> ok - building php-stage
May 07 14:17:53 <dyoung> Does anyone have a outstanding bk changeset?
May 07 14:37:00 <copperbeech> jp30: that seemed to do it... now got past the ldap check..
May 07 15:08:25 <rwhitby-asleep> BitKeeper is still broken :-(
May 07 15:08:53 <rwhitby> dyoung: yeah, me and jbowler at least
May 07 15:10:44 <jp30> morning, rwhitby
May 07 15:11:37 <rwhitby> morning
May 07 15:14:09 <jstueve> morning
May 07 15:15:09 <NAiL> rwhitby: Morning (or, evening)
May 07 15:16:12 <NAiL> jbowler: Ah, uclibc build? Full build, not just samba? Guess I can do that ;)
May 07 15:17:03 <jstueve> jbowler: cool, what do I need to do?
May 07 15:19:23 <dyoung> ok thats basically what I was asking.
May 07 15:19:52 <dyoung> because it seems some projects are still pushing.
May 07 15:20:49 <NAiL> rwhitby: Fixed the linux-threads-problem with mysql. It was because I use NPTL. Details are on the ml.
May 07 15:22:56 <jp30> NAiL, isn't it dangerous that mysql detects the threading behaviour of the host system. i would have thought that oe would do something like what i did with unslung to prevent this
May 07 15:25:39 <NAiL> jp30: There doesn't appear to be any actual differences, but it should detect the target threading, yes.
May 07 15:26:13 <NAiL> I guess I'll find out in a new minutes, when I install mysql
May 07 15:31:19 <rwhitby> dyoung: seems each repo gets 40 extra after it breaks
May 07 15:31:47 <rwhitby> and it only affects free licenses
May 07 15:32:56 <dyoung> i see.
May 07 15:33:07 <dyoung> I guess we were kind of blind-sided then eh.
May 07 15:34:34 <rwhitby> yep.  But my full build from berlios worked
May 07 15:37:44 <rwhitby> (image completed, doing packages now)
May 07 15:38:12 <rwhitby> I find it hard to believe that openlogging.org is unintentionally broken for this long ...
May 07 15:42:14 <NAiL> hnmm
May 07 15:42:41 <NAiL> mysql's playing hard to set up :(
May 07 15:44:38 <NAiL> 050507 22:43:48 [ERROR] Fatal error: Can't open privilege tables: Table 'mysql.host' doesn't exist
May 07 15:47:11 <jp30> did you run mysql_install_db
May 07 15:47:32 <jp30> ?
May 07 15:48:04 <rwhitby> rwhitby@nudi:/home/openslug$ ping www.google.com
May 07 15:48:05 <rwhitby> ping: unknown host www.google.com
May 07 15:49:16 <NAiL> jp30: Yes
May 07 15:49:31 <NAiL> rwhitby: I get google.com, but not google.no ;)
May 07 15:49:46 <NAiL> jp30: But mysql.host is nowhere to be seen
May 07 15:50:50 <NAiL> ah... it's a table.not a file. But.. hm..
May 07 15:51:14 <ka6sox> rwhitby: let me see if its a resolv.conf problem.
May 07 15:51:46 <NAiL> fs ipk # ping www.google.com
May 07 15:51:46 <NAiL> ping: unknown host www.google.com
May 07 15:52:06 <rwhitby> ka6sox: I think it's wider scope than just nudi
May 07 15:52:11 <NAiL> root@LKG7F654F:/var/mysql# ping www.google.com
May 07 15:52:11 <NAiL> ping: www.google.com: Unknown host
May 07 15:52:17 <NAiL> Yeah ;)
May 07 15:52:28 <rwhitby> OH NO! Google was running from a BitKeeper repository !!!!!!!
May 07 15:52:33 <NAiL> heh
May 07 15:52:53 <ka6sox> lol!
May 07 15:53:07 <ka6sox> maybe it *was* the bitkeeper repository?
May 07 15:53:38 <NAiL> rwhitby: Drop www
May 07 15:53:40 <NAiL> root@LKG7F654F:/var/mysql# ping google.com
May 07 15:53:40 <NAiL> PING google.com (216.239.39.99): 56 data bytes
May 07 15:54:12 <ka6sox> I thought google was like...Akamai...everywhere.
May 07 15:55:49 <jp30> google are just transitioning to their rightful place as a top level domain; .com's too small for them
May 07 15:56:06 <NAiL> heh... http://google/
May 07 15:56:41 <ka6sox>  .google
May 07 15:57:09 <ka6sox> their load balancing router must have died.
May 07 15:57:19 <ka6sox> (bad Cisco IOS upgrade)
May 07 15:58:22 <NAiL> Nah, it was a linksys.. WRT54G v1.0 ;)
May 07 15:58:36 <NAiL> (ok, cisco subsidiary)
May 07 15:58:59 <rwhitby> wl530g :-) - lots of them stacked tightly
May 07 15:59:04 <NAiL> hwehehe
May 07 16:01:20 <ka6sox> a bunch of NSLU2's running Openslug?
May 07 16:02:10 <rwhitby> and php/mysql/apache
May 07 16:03:14 <ka6sox> built with a native toolchain.
May 07 16:03:22 <NAiL> I can get to google.com, and google.no, but not www.*.com/no
May 07 16:04:19 <ka6sox> "works for me!"
May 07 16:04:25 <ka6sox> www.google.com
May 07 16:05:47 <NAiL> oh well... It'll get back up soon enough
May 07 16:05:56 * NAiL goes googling for "google under attack"
May 07 16:06:17 <ka6sox> news.google.com
May 07 16:06:41 <NAiL> I can't get to news.google.com
May 07 16:07:03 <ka6sox> me neither.
May 07 16:07:14 <ka6sox> google be limping along
May 07 16:07:30 <NAiL> If you have the ip cached, apparently ;)
May 07 16:09:36 <NAiL> heh... brilliant. The links on googles page adjust to what hostname you use...
May 07 16:09:59 <NAiL> http://news.216.239.39.104/ <-- if you click on the "news" link when going to google by ip ;)
May 07 16:10:20 <rwhitby> gmail down too
May 07 16:10:24 <ka6sox> maybe dns poisoning.
May 07 16:11:07 <NAiL> seems up again now
May 07 16:11:18 <NAiL> GAAAH
May 07 16:11:23 * NAiL hates google
May 07 16:11:47 <NAiL> It insists on going to google.no, but google.no doesn't have "news"
May 07 16:15:55 <ka6sox> images is down too.
May 07 16:16:41 <NAiL> Strange.. /. was down earlier today. Someone trying to take out major sites? ;)
May 07 16:20:39 <rwhitby> in protest about bitkeeper
May 07 16:21:37 <dyoung> Oh well, since the internet is broken, I guess I should go outside.
May 07 16:23:07 <NAiL> outside?
May 07 16:23:15 <ka6sox> I'm watching browser requests and there is a LOT of redirection going on...
May 07 16:23:33 <dyoung> Outside, you know that stuff with the sky?
May 07 16:23:53 <ka6sox> wear Sunglasses and Sunscreen
May 07 16:23:55 <NAiL> dyoung: Aaah... That side. 
May 07 16:24:36 <NAiL> I might go there as well. Time for a smoke. It's a tad bright outside though...
May 07 16:24:52 * NAiL points out the fact that he's in GMT+1
May 07 16:25:58 <ka6sox> but north of 51deg north.
May 07 16:26:21 <NAiL> No idea how far north that is, but quite possibly ;)
May 07 16:26:46 <NAiL> The sun will be up in an hour or so I guess
May 07 16:27:07 <ka6sox> where are you in .no?
May 07 16:27:35 <NAiL> Trondheim ( 63.4170°N, 10.4170°E)
May 07 16:29:23 <NAiL> Ah, according to some astronomy-thingymabobs, the sunwon't be up until 4:30
May 07 16:29:29 * rwhitby just commited his first set of changes to the SVN repo on berlios
May 07 16:29:45 <ka6sox> ah...I was in Fairbanks, AK on the longest day of the year...in a Jacuzzi at 2am...the sun was still shining brightly.
May 07 16:30:17 <ka6sox> was that off of nudi?
May 07 16:30:29 <rwhitby> no, from here
May 07 16:30:31 <ka6sox> k
May 07 16:30:41 <ka6sox> I upgraded svn on nudi...so its currentl
May 07 16:30:42 <rwhitby> nudi would work though
May 07 16:30:50 <ka6sox> current.
May 07 16:33:15 <NAiL> Is there a openslug-devtools target (like openslug-packages, but with gcc, binutils etc?)
May 07 16:38:17 <dyoung> Guess its time to RTF SVN M.
May 07 16:39:08 <rwhitby> svn co, svn stat, (require 'vc-svn), svn commit.
May 07 16:39:13 <rwhitby> svn update
May 07 16:39:17 <rwhitby> very much like CVS
May 07 16:39:27 <dyoung> svn ci foo.c ?
May 07 16:40:05 <rwhitby> that's commit
May 07 16:40:13 <rwhitby> single step, like CVS
May 07 16:41:32 <rwhitby> http://svn.collab.net/repos/svn/trunk/README <- quickstart guide
May 07 16:41:53 <rwhitby> that and svn help are the limit of my knowledge at the moment
May 07 16:42:04 <jstueve> any hints on getting samba working?
May 07 16:42:18 <jstueve> pdb_generate_sam_sid: Failed to store generated machine SID.
May 07 16:42:18 <jstueve> PANIC: Could not generate a machine SID
May 07 16:42:41 <jstueve> openslug - uclibc
May 07 16:43:05 <NAiL> But where is it trying to store that SID?
May 07 16:43:06 <rwhitby> dunno
May 07 16:44:23 <ka6sox> sounds like a samba.org question.
May 07 16:44:31 <jstueve> okay...
May 07 16:44:55 <NAiL> jstueve: Increase the debug level, and see where it tries to write the machine SID
May 07 17:18:43 <jbowler> NAiL, jstueve: sorry about the delay, feeding the neighbour's menagerie.  No one can do a BitKeeper commit on the OE repository, or anything derived from it, at present.
May 07 17:20:19 <rwhitby-away> jbowler: we can commit to the new SVN repo on berlios
May 07 17:20:26 <rwhitby-away> talk to ka6sox for details.
May 07 17:20:37 <rwhitby-away> then we can transfer from there back to bk when it comes back on line
May 07 17:20:41 <jbowler> NAiL, jstueve: the fix, however, can be accomplished by building uclibc then editing uClibc-0.9.27/libc/sysdeps/linux/common/bits/ipc.h to before remove the __kernel_key_t typedef and change the use of it in the struct to __key_t
May 07 17:22:01 <NAiL> jbowler: Ok, so basically, cat ipc.h|grep -v __kernel_key_t|sed s/__kernel_key_t/__key_t/g ?
May 07 17:22:10 <rwhitby-away> or we can wait ...
May 07 17:22:13 <NAiL> (more or less)
May 07 17:22:15 <rwhitby-away> back later
May 07 17:22:31 <rwhitby-away> NAiL: openslug-packages does include dev stuff.
May 07 17:22:46 <NAiL> odd
May 07 17:23:36 <NAiL> not binutils
May 07 17:24:09 <NAiL> (I added that later though)
May 07 17:26:01 <jbowler> rwhitby: openlogging isn't broken.  The problem seems to be that some commits to OE were not logged.  On Apr26 there were 0 pending logs, but before and after there were >0!
May 07 17:29:10 <jbowler> NAiL: no, that'll give a duplicate typedef of __kernel_dev_t.  Delete the first instance, change the second one as in your sed.  or '/^typedef int __kernel_key_t/d;s/__kernel_key_t/__key_t/'
May 07 17:31:04 <NAiL> jbowler: Ah, ok. Then I get it ;)
May 07 17:32:11 <jbowler> should be 'duplicate typedef of __key_t', sorry.
May 07 17:33:19 <jbowler> Even easier in fact is to go into staging/arm*/include/bits and hack the ipc.h there.
May 07 17:36:20 <jstueve> hrm...
May 07 17:37:30 <jstueve> okay.. ipkg gcc doesn't install a gcc, but an armeb-linux-uclibc-gcc
May 07 17:37:43 <jstueve> so I symlinked gcc to that fike
May 07 17:37:59 <NAiL> jstueve: Install gcc-symlinks
May 07 17:38:07 <jstueve> oh.
May 07 17:38:40 <jstueve> okay.. samy error
May 07 17:38:43 <jstueve> same error
May 07 17:38:44 <jstueve> checking for C compiler default output... configure: error: C compiler cannot create executables
May 07 17:39:06 <NAiL> check config.log, and see what it really complains about
May 07 17:43:28 <NAiL> hah! In a little while, I have one glibc and one uclibc openslug! :)
May 07 17:44:03 <NAiL> (as soon as it's done verifying, atleast)
May 07 18:04:25 <jbowler> ~seen dyoung
May 07 18:04:27 <jbot> dyoung is currently on #openjtag (2d 4h 12m 55s) #nslu2-linux (2d 4h 12m 55s) #oe (2d 4h 12m 55s).  Has said a total of 171 messages.  Is idling for 1h 24m 59s
May 07 18:04:47 <jbowler> ping dyoung: is x1205_validate_tm one of your functions?
May 07 18:08:08 <ka6sox> jbowler: he went "outside"
May 07 18:08:26 <ka6sox> I told him to wear his sunglasses and Sunscreen.
May 07 18:08:39 <ka6sox> :)
May 07 18:09:07 <jbowler> Oh, ok,,, there was a big hole in the ozone layer.
May 07 18:10:06 <jbowler> I think I've got a handle on the panic which causes a machine running NTP to die after a while (I've seen it NAiL has, I think dyoung has).
May 07 18:11:26 <NAiL> oh?
May 07 18:11:36 <ka6sox> I've seen reference to that....maybe a 1's complement at Sercomm when they were doing this.
May 07 18:12:24 <jbowler> I've got a kernel panic out of netconsole (neat, no serial needed...)  NAiL: you said your slug died yesterday with NTP, it takes several hours...
May 07 18:12:50 <NAiL> yeah, It took... probably 14-15 hours or so?
May 07 18:13:37 <NAiL> Can't remember how many hours, but it took a few atleast ;)
May 07 18:18:58 <NAiL> If there anything that can be fixed without massive recompiling? ;)
May 07 18:19:49 <ka6sox> this is where distcc would be helpful?
May 07 18:22:19 <ka6sox> Hi [g2] 
May 07 18:22:25 <NAiL> g2 :)
May 07 18:23:10 <[g2]> hey guys!
May 07 18:23:28 <[g2]> I heard there was a SVN on berlios
May 07 18:23:37 <NAiL> <-- just got slug #2 up. One's glibc and one's uclibc :)
May 07 18:24:03 <[g2]> NAiL, excellent
May 07 18:24:09 <jbowler> Hi [g2] - I'm just trying to access the berlios repo...
May 07 18:24:10 * bbf kicks his wl500gx
May 07 18:24:15 <[g2]> and I noticed you've got Mysql too ?
May 07 18:24:36 <NAiL> yeah, but I'm having problems starting it on uclibc
May 07 18:25:53 <[g2]> does it run on glibc ?
May 07 18:26:31 <bbf> hmm has anyone sucefully used bluez's bt stack ? cross compiling it was simple, but I think something is broken :P
May 07 18:26:38 <[g2]> jbowler, Hey!
May 07 18:26:52 <[g2]> I noticed your info on NTP on the ml
May 07 18:26:55 <NAiL> [g2]: Checking
May 07 18:27:39 <[g2]> bff on OpenSlug rwhitby's been using that as a bluetooth access point to his treo for I think a month or two
May 07 18:29:24 <bbf> I have traced all the way back to a ioctl() call, but as I'm not very familiar with it, I dunno what's wrong :P
May 07 18:29:53 <[g2]> bbf I think there a BT howto on the wiki
May 07 18:30:19 <bbf> on chupa.nl there's a BT howto but using the affix stack
May 07 18:30:38 <[g2]> so you are not using the bluez stack ?
May 07 18:30:45 <bbf> and the binaries posted there for it are old, and broken also :P
May 07 18:30:49 <bbf> i'm trying to :D
May 07 18:31:03 <[g2]> do you need the affix stack ?
May 07 18:31:14 <bbf> anyone that works
May 07 18:31:34 <[g2]> well bluez has worked on the nslu2 with openslug
May 07 18:31:42 <[g2]> are you running unslung or openslug
May 07 18:31:54 <bbf> neither ... (/me hides)
May 07 18:32:06 <bbf> wl500gx
May 07 18:32:31 <NAiL> [g2]: Get this error: 050508  1:31:35 [ERROR] Fatal error: Can't open privilege tables: Table 'mysql.host' doesn't exist
May 07 18:32:56 <[g2]> NAiL, Ummm did you create the table ? :)
May 07 18:33:01 <NAiL> mysql_install_db
May 07 18:33:02 <NAiL> yes
May 07 18:33:13 <NAiL> get the same err on both glibc and uclibc
May 07 18:33:16 <[g2]> NAiL, is taht uClibc or glibc ?
May 07 18:33:21 <[g2]> good
May 07 18:33:33 <[g2]> that's the way it should fail
May 07 18:33:37 <NAiL> heh
May 07 18:33:41 <NAiL> it should fail?
May 07 18:33:42 <[g2]> bbf welcome to the crowd
May 07 18:33:55 <bbf> :)
May 07 18:34:12 <[g2]> it *shouldnl't* fail, but if it does then it should fail consistently
May 07 18:34:22 <NAiL> ah, yes.
May 07 18:34:42 <[g2]> bbf, the wl500g... looks nice. I don't have one, but there are several ppl here running one
May 07 18:34:43 <NAiL> But I still can't figure out what the problem is. It *looks* like the tables are created as they should.
May 07 18:35:14 <[g2]> I just started running my first LAMP server last week
May 07 18:35:39 <[g2]> So I'll be of no help as I barely got that running
May 07 18:35:41 <NAiL> [g2]: Hasn't anyone gotten mysql to run?
May 07 18:35:54 <[g2]> I think the unslung guys have
May 07 18:36:20 <[g2]> these are the reasons I think the community needs to get its stuff in one sock
May 07 18:36:53 <[g2]> to make the shole job of running all this sw much more frictionless
May 07 18:37:01 <ka6sox> ipkgfind.nslu2-linux.org has all the packages that are known to work.
May 07 18:37:02 <[g2]> s/shole/whole/
May 07 18:37:12 <ka6sox> and what feeds they are in.
May 07 18:38:18 <[g2]> ka6sox, I have to check it out some more.  I think that's a start, but it's a bit off from what I'm talking about
May 07 18:38:31 <ka6sox> k
May 07 18:38:37 <NAiL> so there *is* a openslug feed ;)
May 07 18:39:14 <[g2]> I wouldn't really call it a feed
May 07 18:39:35 <NAiL> well.. it contains a few ipkgs, right?
May 07 18:39:45 <[g2]> to my knowledge (and I may be totally wrong) but the openslug stuff is just packages built from OE
May 07 18:40:18 <[g2]> so if you're not using the nslu2-symlinks then it just bb frotz
May 07 18:40:23 <[g2]> and frotz builds
May 07 18:40:51 <[g2]> or add the symlinks for frotz and maybe its dependencies and then it builds
May 07 18:41:19 <NAiL> frotz: An interjection of mildest disgust. <-- aha
May 07 18:41:23 <[g2]> this ties into the site I'm about to launch
May 07 18:41:45 <NAiL> site? 
May 07 18:41:51 <[g2]> I've run Zork on my slug
May 07 18:42:01 <[g2]> yeah site
May 07 18:43:51 <NAiL> one-sock-community.org? ;)
May 07 18:44:12 <[g2]> no....
May 07 18:44:20 <[g2]> three-sock-community :)
May 07 18:45:16 <NAiL> Anyways, what site? :)
May 07 18:45:38 <[g2]> coming soon
May 07 18:45:59 <NAiL> mmkay
May 07 18:47:20 <[g2]> I just need to straighten out my firewalling/logging and how much access new users get and then I'm ready to launch
May 07 18:47:42 <[g2]> and bw may be an issue
May 07 18:51:27 <NAiL> ./mysqld: ready for connections.
May 07 18:51:27 <NAiL> Version: '4.1.10a'  socket: '/tmp/mysql.sock'  port: 3306  Source distribution
May 07 18:55:23 <ka6sox> bbiaw
May 07 18:55:48 <NAiL> MySQL is running, but all the databases are... not there
May 07 18:56:53 <NAiL> me triggered some fancy bug in busybox vi as well
May 07 18:57:07 <[g2]> the lockup bug ?
May 07 18:57:42 <NAiL> no.. got bucketloads of ^@, and trying to quit was... interesting. It didn't want me to quit. After trying a few times it segfaulted :)
May 07 18:58:04 <NAiL> The ^@'s came and went, in the middle of "known good" text
May 07 19:04:35 <NAiL> [g2]: To get mysql up and running I had to specify data-dir, and create the grant tables manually (dump them on a working box and insert them on the slug)
May 07 19:05:45 <[g2]> NAiL, excellent. I guess we need to add that stuff to the startup proceedure
May 07 19:06:27 <NAiL> [g2]: And, mysql is compiled with innodb support. If you start mysql with innodb, it takes 14mb mem and starts 10 processes
May 07 19:06:39 <NAiL> without innodb, 3.5mb and 3 processes
May 07 19:06:42 <NAiL> (glibc)
May 07 19:08:06 <[g2]> NAiL, you rock!
May 07 19:08:16 <NAiL> I can send you the sql dump, and hopefully the instructions on how to get things up and running
May 07 19:09:45 <NAiL> but first, I have to go outside and say hello to the rising sun ;)
May 07 19:12:27 <bbf> you can control how well innodb behaves thru my.cnf tho
May 07 19:12:45 <NAiL> Quite possible, but it is nevertheless a big memory hog on the slug
May 07 19:13:05 <bbf> what do you intend to use it for ?
May 07 19:13:33 <NAiL> Simple databases, nothing big. 
May 07 19:13:36 <NAiL> php-stuff basically
May 07 19:14:14 <bbf> sqlite is really nice, if you haven't tried it out
May 07 19:14:47 <NAiL> Can't say I have. Might be worth looking into
May 07 19:15:07 <bbf> I dunno if php4 has support for it, or if it's only 5
May 07 19:16:30 <NAiL> 5 has it included, atleast
May 07 19:16:35 <jbowler> I ported SleepyCat DB4.3 a week or so ago for OpenLDAP.  Doesn't seem excessive in size.
May 07 19:17:09 <NAiL> 4 supports it
May 07 19:17:27 <NAiL> sqlite is < 300kb compiled
May 07 19:18:10 <bbf> in x86, it's less than 200 kb
May 07 19:18:38 <bbf> maybe there's a overhead integrating it to php
May 07 19:19:10 <NAiL> Don't think it'd be that big, it being a lib
May 07 19:19:33 <bbf> that's a lib, and the server itself
May 07 19:19:39 <bbf> it's embedded
May 07 19:19:56 <bbf> how big is your mysql bin ? ;)
May 07 19:20:54 <NAiL> mysqld is 4088324, mysql is only 167940
May 07 19:21:04 <NAiL> but then you have libmysqlclient as well
May 07 19:21:28 <bbf> mysqld + libmysqlclient + mysql = sqlite
May 07 19:21:42 <bbf> so, do you still find it big ?
May 07 19:21:56 <NAiL> I never found it big ;)
May 07 19:23:20 <bbf> here I go flashing my router again... may please bluez work now
May 07 19:24:49 <NAiL> how do I specify parameters for configure in oe?
May 07 19:38:37 <NAiL> Anyone want to test installing mysql? ;)
May 07 19:39:31 <rwhitby-away> bbf: I could never get the bluetooth stuff to work on the wl500gx.
May 07 19:39:42 <rwhitby-away> Doh' - just missed him
May 07 19:56:20 <jstueve> NAiL: I will once my bb openslug-packages completes
May 07 19:56:39 <NAiL> hehe, ok :)
May 07 19:56:49 <NAiL> jstueve: How far along is it?
May 07 19:56:57 <jstueve> samba compile
May 07 19:57:04 <NAiL> mmkay
May 07 19:57:08 <jstueve> mysql took this time.. *shrug*
May 07 20:55:42 <bbf> based on the delay it took me to crawl back to here, you could say my firmware was broken :D ... but after 4 builds, bluez's stack is finally working
May 07 20:57:18 <bbf> except for sdp :P
**** ENDING LOGGING AT Sat May 07 23:59:56 2005