**** BEGIN LOGGING AT Wed Jun 22 23:59:57 2005
Jun 23 00:11:00 <phedny> salut
Jun 23 00:35:17 <jeanfabrice> 
ACTION 

Jun 23 01:56:09 <jf-work> nail: ping
Jun 23 02:01:23 * dyoung-zzzz twitches
Jun 23 03:14:07 <NAiL> jf-work: pong
Jun 23 03:14:35 <dyoung-zzzz> nail, we were wondering about the ownership of the mirror files on your mirror.
Jun 23 03:14:47 <dyoung-zzzz> the files seem to be owned by unslung.
Jun 23 03:16:00 <NAiL> yeah
Jun 23 03:16:29 <dyoung-zzzz> it would be ideal if you can chown -R them to nslu2
Jun 23 03:16:44 <dyoung-zzzz> since thats how the scripts are structured to login
Jun 23 03:17:02 <dyoung-zzzz> login/rsync.
Jun 23 03:17:11 <NAiL> done
Jun 23 03:17:18 <jf-work> thx
Jun 23 03:17:51 <dyoung-zzzz> thansk nail
Jun 23 03:18:05 <dyoung-zzzz> (thats what jf-work was pinging you about)
Jun 23 03:18:37 <jf-work> NAiL mirror refresh in progress...
Jun 23 03:19:24 <jf-work> done without errors :)
Jun 23 03:19:24 <dyoung-zzzz> ~emulate rwhitby
Jun 23 03:19:25 <jbot> Well, my work here is done.
Jun 23 03:19:59 <NAiL> :)
Jun 23 03:20:42 <dyoung-zzzz> nail, did samba work out for you?
Jun 23 03:20:47 <NAiL> yeah
Jun 23 03:20:55 <dyoung-zzzz> Cool.
Jun 23 03:23:48 <dyoung-zzzz> nail, does your SlugAir project have tilt meters or accelerometers in it?
Jun 23 03:24:15 <NAiL> accelerometers and gyro
Jun 23 03:24:29 <dyoung-zzzz> are you reading those into Slug?
Jun 23 03:24:40 <dyoung-zzzz> or are those gonna be into a AVR?
Jun 23 03:25:20 <NAiL> they're most likely going to be read by an AVR, then fed to the slug
Jun 23 03:25:29 <dyoung-zzzz> I see.
Jun 23 03:26:11 <dyoung-zzzz> I find myself in a quandry about how best to read accelerometers.
Jun 23 03:26:44 <dyoung-zzzz> (and tilt meters)
Jun 23 03:27:20 <dyoung-zzzz> UBE is in the cards....
Jun 23 03:27:20 <dyoung-zzzz> ~ube
Jun 23 03:27:21 <jbot> ube is, like, Ugly But Effective.  Or sometimes Unorthrodox But Effective
Jun 23 03:27:40 <NAiL> yeah, I'ev been wondering a bit about that myself
Jun 23 03:28:17 <NAiL> they output an analog signal, IIRC?
Jun 23 03:28:35 <NAiL> I'm not that much into that part yet. The Atmel guy is fiddling with that right now
Jun 23 03:33:28 <NAiL> I can't think of an easy way to read the sensors into the slug directly
Jun 23 03:35:28 <NAiL> hmm... what's the difference between gyro and tilt?
Jun 23 03:38:28 <dyoung-zzzz> Through I2C is all I can think of.
Jun 23 03:42:45 <NAiL> dyoung-zzzz: btw, you won't need the unslung user login, do you?
Jun 23 03:45:27 <jf-work> NAiL: if your question is relative to the mirror sync, the answer is no
Jun 23 03:45:48 <dyoung-zzzz> yeah, what jf-work said. :-)
Jun 23 03:46:03 <NAiL> ok, then I'll remove it 
Jun 23 03:46:19 <dyoung-zzzz> the script is now hard wired to use the user nslu2 as the login.
Jun 23 03:46:56 <dyoung-zzzz> which is convenient because the other mirrors generally have the same nslu2 user.
Jun 23 03:47:18 <NAiL> yeah
Jun 23 03:49:42 <jf-work> ok, it is now time for lunch. let's go.
Jun 23 04:04:39 <NAiL> second slug successfully openslug2'ed
Jun 23 04:49:30 <NAiL> hmmyay
Jun 23 04:49:31 <NAiL> cc: Internal error: Killed (program cc1)
Jun 23 04:49:52 <NAiL> Methinks I ran out of swap..
Jun 23 05:02:16 <[g2]> jacques, still around ?
Jun 23 05:03:29 <rwhitby-away> hey [g2] - congrats on the OpenSlug 2.0 release
Jun 23 05:03:47 <[g2]> rwhitby-away, thx but it was really jbowler-zzz and ByronT-Away 
Jun 23 05:04:11 <[g2]> the Release notes page looked really nice I thought
Jun 23 05:04:27 <[g2]> and most importantly the release is really shaping up
Jun 23 05:06:09 <rwhitby-away> I'm putting together a complete nslu2-linux development directory structure.  Top-level Makefile, subdirs: bitbake, openembedded, unslung, openslug, apex, optware, monotone.  Build everything to do with this project with a single "make all".
Jun 23 05:06:47 <[g2]> rwhitby-away, is the MAN!
Jun 23 05:07:28 <[g2]> are you just going to change TMPDIR and the BBFILES on the fly ?
Jun 23 05:07:32 <rwhitby-away> the repo transition is the perfect opportunity.  everyone is going to need to re-pull everything anyway, so we might as well clean up a bit.
Jun 23 05:07:56 <[g2]> it think freezing changes now has been a good thing
Jun 23 05:07:59 <rwhitby-away> no, that stuff will still be below the openslug and unslung directories as before
Jun 23 05:08:19 <rwhitby-away> the downloads, bitbake, and openembedded directories will be shared
Jun 23 05:08:29 <[g2]> so you'll just re-run the setup-env equiv?
Jun 23 05:08:44 <rwhitby-away> it will be run independently in unslung and openslug subdirs
Jun 23 05:08:52 <[g2]> ok
Jun 23 05:09:12 <[g2]> makes sense, just "down" a subdir
Jun 23 05:09:18 <rwhitby-away> yep
Jun 23 05:09:44 <rwhitby-away> so if you only want to do stuff in one of the different areas, you just cd into there and do like you do today
Jun 23 05:09:56 <rwhitby-away> it will all live under /home/slug/...
Jun 23 05:10:04 <NAiL> [g2]: Btw, a few other things with the dev env install script.. gnu-config, libperl & m4 are pulled in as deps to automake. They are also installed "manually" later. Is there a reason? (Ie, doesn't ipkg update deps?)
Jun 23 05:10:26 <NAiL> Started with a "clean" slug not long ago
Jun 23 05:10:28 <[g2]> nod you could even add an extra couple directories for OpenSlug-2.0 and Unslung-5.5
Jun 23 05:11:47 <[g2]> NAiL, dunno for sure perl is compiles but doesn't run in the cross build, but I didn't pull all perl things as I'm sure it would be nicer to build cross 
Jun 23 05:12:06 <[g2]> s/is //
Jun 23 05:13:12 <NAiL> any reason ncurses-dev isn't installed? 
Jun 23 05:13:33 <[g2]> not really
Jun 23 05:13:37 <rwhitby-away> NAiL: this script in a repo anywhere?
Jun 23 05:14:15 <NAiL> rwhitby-away: Not that I know. It's on the wiki. 
Jun 23 05:14:38 <rwhitby-away> [g2]: where should it go?
Jun 23 05:15:01 <rwhitby-away> (much easier to update and run from a repo rather than a wiki page ...)
Jun 23 05:15:35 <rwhitby-away> jacques is setting up banana slug now.  he and you should be using the same script and collaborating
Jun 23 05:15:36 <[g2]> rwhitby-away, well it's mostly baked but I didn't know if it was fully baked yet so that's why I didn't make it a script
Jun 23 05:16:01 <rwhitby-away> [g2]: you know that's no excuse.  release early, release often.
Jun 23 05:16:11 <[g2]> I did, just to the wiki
Jun 23 05:16:22 <NAiL> [g2]: btw, samba initscript+conf-file issue is basically fixed, but not pushed afaik.
Jun 23 05:16:39 <[g2]> NAiL, cool and thx
Jun 23 05:17:06 <rwhitby-away> [g2]: seriously, jacques should be using this to set up banana, and editing it on the fly as he goes.  can you sync up with him on this?
Jun 23 05:17:45 <[g2]> I ask jacques what he's doing, certainly that makes total sense to me
Jun 23 05:18:03 <[g2]> I think the base is all there
Jun 23 05:18:17 <[g2]> I hadn't built perl natively, but NAiL did the other day
Jun 23 05:18:28 <rwhitby-away> banana is the openslug native build host, so it should be set up so that other developers can replicate it's installation.  the only way to do that is with a script in a repo.
Jun 23 05:18:31 <NAiL> yup. Works like a charm 
Jun 23 05:18:45 <[g2]> is Test 35 the only issue ?
Jun 23 05:18:57 <NAiL> yeah, fixed with procps ps
Jun 23 05:19:29 <[g2]> Ok.. so it's 100% perl, 98% Python (only 3 or 4 tests fail out of 249)
Jun 23 05:19:59 <[g2]> the leaves the glibc tests and LTP suite
Jun 23 05:20:07 <[g2]> and your make of MySQL 
Jun 23 05:20:31 <NAiL> Haven't stared with that yet, as soon as I get this slug up proper I will
Jun 23 05:20:39 <[g2]> nod.
Jun 23 05:22:07 <NAiL> hmm.. isn't there a procps ipkg?
Jun 23 05:26:19 <[g2]> NAiL, I need to find out where that comes from
Jun 23 05:26:38 <[g2]> we also need to figure out what we want to replace from busybox
Jun 23 05:33:30 <NAiL> gotta run, look at stuff later.
Jun 23 05:33:42 <[g2]> cheers
Jun 23 05:35:26 * jf-work entering logger mode
Jun 23 07:11:20 * jf-work entering logger mode
Jun 23 07:48:59 <Sutak> Has anyone tested the hybrid-ircd package? I try to start ircd and get: Could not load modules from /opt/modules/autoload/: No such file or directory 
Jun 23 07:49:27 <Sutak> And finally: Error loading core module m_die.so: terminating ircd
Jun 23 07:50:18 <ByronT> I know some of the devs are using it, you can ask (politely) in #nslu2-linux
Jun 23 07:51:15 <rwhitby-asleep> ByronT: I'm not aware of anyone using it.  A number of people are using miau as a bouncer ...
Jun 23 07:51:29 <Sutak> Byront: I am trying. :)
Jun 23 07:51:36 <ByronT> ah
Jun 23 07:52:08 <Sutak> The package consists only of 2 files: the binary and simple.conf
Jun 23 07:52:22 <ByronT> hmm.. Sutak was already in #nslu2-linux..... evidence that I have not had the requisite amount of coffee to be answering questions this morning
Jun 23 07:52:32 <Sutak> Maybe module support has to be disabled? But how?
Jun 23 07:53:11 <Sutak> ByronT: No problem.
Jun 23 07:53:28 <ByronT> Sutak, I was mistaken, as rwhitby said, some are using miau as a bouncer... but feel free to try to fix ircd, and become part of the development!
Jun 23 07:55:57 <Sutak> ByronT: I would be happy if I could fix it...if I do so (by pure luck) I will contribute my findings.
Jun 23 08:08:01 <Sutak> Thank you anyway..and good bye.
Jun 23 09:19:47 * jf-away entering logger mode
Jun 23 09:21:44 <NAiL> [g2]: MySQL is still building. Hasn't started compiling sql_yacc yet, but I guess that one's gonna take a *long* time...
Jun 23 09:23:34 <[g2]> NAiL, thx
Jun 23 09:23:43 <[g2]> how long has it been going ?
Jun 23 09:25:49 <NAiL> [g2]: four-five hours
Jun 23 09:26:13 <[g2]> it'd be interesting to compare build times on the 32 versus 64 MB units
Jun 23 09:26:54 <NAiL> It'll be interesting to see whether or not building mysql wears out the red hdd-access light on my enclosure by swapping ;)
Jun 23 09:27:08 <NAiL> It's more or less continous atm
Jun 23 09:28:27 <NAiL> The build was interrupted earlier, since I forgot that I didn't have a working perl on this slug
Jun 23 09:28:34 <[g2]> If I can get my loaner going 425@533 with 256MB that wold be very interesting too
Jun 23 09:28:51 <NAiL> I can build if from scratch again when I'm finished building everything I need
Jun 23 09:32:53 <[g2]> do you have your config parameters ? or are they all stock
Jun 23 09:35:49 <NAiL> Removed innoDB support
Jun 23 09:36:02 <NAiL> Might have done some other changes, can't remember. Can check later.
Jun 23 09:36:17 <NAiL> innodb support is a really big memory hog
Jun 23 09:37:16 <NAiL> ah, yes
Jun 23 09:37:17 <NAiL> ./configure --prefix=/usr --disable-dependency-tracking --with-low-memory --with-mysql-user=mysql --without-debug --without-innodb --libexecdir=/usr/sbin
Jun 23 09:37:57 <NAiL> --disable-dep-tracking is if it's a one-shot build. I'm not hacking the code, so I need no deps. 
Jun 23 09:38:08 <NAiL> (which speeds up the build a bit)
Jun 23 09:41:05 <NAiL> The reason I'm changing the libexecdir is that mysqld resides there. I like having it in /usr/sbin (where safe_mysql, which I don't use, goes)
Jun 23 09:41:48 <NAiL> For a standard build, don't change libexecdir ;)
Jun 23 10:38:17 <[g2]> jbowler-away, what was the /proc/cpu/alignment number to fix alignment errors ?
Jun 23 10:38:30 <[g2]> Alignment trap: try (4415) PC=0x000085e4 Instr=0xe79e1000 Address=0xbefffde9 FSR 0x013
Jun 23 10:38:30 <[g2]> Alignment trap: try (4415) PC=0x000085e4 Instr=0xe79e1000 Address=0xbefffde9 FSR 0x013
Jun 23 10:38:47 <[g2]> I got that during a perl 5.8.7 config
Jun 23 10:38:56 <[g2]> actually configure.gnu
Jun 23 10:39:37 <[g2]> cat /proc/cpu/alignment
Jun 23 10:39:37 <[g2]> User:           4132
Jun 23 10:39:37 <[g2]> System:         6
Jun 23 10:39:37 <[g2]> Skipped:        0
Jun 23 10:39:37 <[g2]> Half:           6
Jun 23 10:39:38 <[g2]> Word:           0
Jun 23 10:39:40 <[g2]> Multi:          0
Jun 23 10:39:42 <[g2]> User faults:    2 (fixup)
Jun 23 10:39:45 <[g2]> I guess 2 is your friend :)
Jun 23 10:41:32 <jbowler-away> You want 5 - see /etc/init.d/alignment.sh
Jun 23 10:43:00 <[g2]> 2s not enough ?
Jun 23 10:43:44 <[g2]> User faults:    5 (signal+warn)
Jun 23 10:44:36 <jbowler-away> 2 just rotates the value back to what the program apparently expected.  5 causes the program to crash, so gdb can debug it.
Jun 23 10:46:22 <jbowler-away> It's also possible, of course, that configure is deliberately testing unaligned accesses - in which case 4 or 5 will give the right result whereas the default will cause bad things to happen.
Jun 23 10:47:06 <[g2]> I think perl 5.8.4 configured properly
Jun 23 10:47:16 <[g2]> but 5.8.7 was unhappy
Jun 23 10:47:58 <jbowler-away> ... well, what's causing the alignment trap?
Jun 23 10:49:02 <[g2]> tzname[] found.
Jun 23 10:49:02 <[g2]> (Testing for character data alignment may crash the test.  That's okay.)
Jun 23 10:49:31 <[g2]> so 5 is the magic number then 
Jun 23 10:49:43 <jbowler-away> So the default '0' is probably a bad thing.  Also this is a big red flag for cross compilation problems.
Jun 23 10:50:25 <[g2]> well currently the default is 1 (warn) so errors were just spewing off the screen
Jun 23 10:53:07 <jbowler-away> Yep, alignbytes all over the place...
Jun 23 11:01:01 <jbowler-away> In the cross compile code byteorder is wrong I believe.  That might explain a lot of things...
Jun 23 11:01:18 <[g2]> tzname[] found.
Jun 23 11:01:18 <[g2]> (Testing for character data alignment may crash the test.  That's okay.)
Jun 23 11:01:18 <[g2]> Alignment trap: try (12267) PC=0x000085e4 Instr=0xe79e1000 Address=0xbefffde9 FSR 0x013
Jun 23 11:01:18 <[g2]> It seems that you must access character data in an aligned manner.
Jun 23 11:03:26 <jbowler-away> It looks like congif.h has byteorder correct and config.sh has it wrong so long as USE_CROSS_COMPILE is being defined.
Jun 23 11:04:26 <jbowler-away> bbl
Jun 23 11:19:16 <NAiL> Hmm... my samba build is logging like crazy. It's logged more than a mb since I installed it earlier today.
Jun 23 11:20:27 * NAiL slaps himself 
Jun 23 11:20:34 <NAiL> Helps setting the proper loglevel
Jun 23 12:07:59 <[g2]> JoeLlama, you'll want to check out our wiki http://www.nslu2-linux.org
Jun 23 12:08:52 <JoeLlama> okee doke :)
Jun 23 12:09:23 <[g2]> I'm an openslug dev so I'm really interested in OpenSlug :)
Jun 23 12:10:09 <JoeLlama> I dunno what openslug is
Jun 23 12:10:20 <JoeLlama> I don't do much with linux at this time.
Jun 23 12:15:10 <NAiL> [g2]: g++ -DMYSQL_SERVER -DDEFAULT_MYSQL_HOME="\"/usr\"" -DDATADIR="\"/usr/var\"" -DSHAREDIR="\"/usr/share/mysql\"" -DHAVE_CONFIG_H -I. -I. -I.. -I../zlib -I../include -I../regex -I.     -O3 -DDBUG_OFF    -fno-implicit-templates -fno-exceptions -fno-rtti -fno-inline -c sql_yacc.cc
Jun 23 12:15:18 <[g2]> it depends on what kind of processors you want to play with
Jun 23 12:15:22 <NAiL> It didn't even take a long time ;)
Jun 23 12:16:08 <[g2]> NAiL, that built ?
Jun 23 12:16:13 <[g2]> or that died :(
Jun 23 12:16:41 * [g2] checks perl build
Jun 23 12:17:04 * [g2] 's perl build still going
Jun 23 12:17:13 <NAiL> It built, and it's still happily chugging along
Jun 23 12:37:09 <NAiL> hmm
Jun 23 12:37:32 <NAiL> I'm trying to add a patch to busybox, and I'm failing miserably. What kind of magic am I missing?
Jun 23 12:42:16 <NAiL> I missed the magic of correct typing actually...
Jun 23 13:21:48 <[g2]> NAiL, how long did the perl tests take ?
Jun 23 13:23:16 <NAiL> Don't know, didn't time them. Took between one to two hours I guess
Jun 23 13:23:38 <NAiL> stupid busybox
Jun 23 13:24:46 <NAiL> Trying to silence start-stop-daemon... but it doesn't listen to "--quiet" at all. So the checks I added in the sourcefile are never false..
Jun 23 13:29:51 <jeanfabrice> llagendijk: bind should be in the wiley feed since yesterday. great work !
Jun 23 13:41:46 <llagendijk> jeanfabrice, thanks!
Jun 23 13:42:30 <llagendijk> jeanfabrice, do you want me to step the version of quagga? I am not sure it makes (much) sense.
Jun 23 13:43:25 <jacques> [g2], perl tests take around 75 minutes
Jun 23 13:43:38 <jeanfabrice> well, you know quagga.mk better than I do :) I was just looking at the diff between cvs 1.5 and 1.6
Jun 23 13:44:24 <llagendijk> jeanfabrice, well lets leave it for now. I will update it when a new quagga is released anyhow
Jun 23 13:45:05 <jeanfabrice> since you don't change the IPK version, quagga is not rebuild by the build machine.
Jun 23 13:56:50 <CIA-3> 03llagendijk * 10unslung/make/quagga.mk: 
Jun 23 13:56:50 <CIA-3> Bumped version to make sure it gets rebuilt, so feeds contain latest details like maintainer and description.
Jun 23 13:56:50 <CIA-3> No functional changes
Jun 23 13:57:09 <llagendijk> jeanfabrice, ok, I bumped the version, so ipkg header contains latest details
Jun 23 13:57:30 <jeanfabrice> ok. thank you llagendijk.
Jun 23 14:22:16 <llagendijk> is there somebody here that knows why there is this option TAR_OPTIONS=--format=ustar for IPKG_BUILD for the ipkg utils?
Jun 23 14:22:32 <llagendijk> My Gnu tar does not understand that option.....
Jun 23 14:57:59 <[g2]> jacques, what was this one ? t/op/magic................................FAILED at test 35
Jun 23 15:03:35 <jeanfabrice> 
ACTION 

Jun 23 15:03:38 <jacques> bad ps
Jun 23 15:09:24 <[g2]> jacques, thx that sounds familar
Jun 23 15:09:48 <ilg_> I checked out trunk from svn, but the nslu2-linux packages dir has no links. is there any way to fix it? I checked 'svn checkout' for a command line switch, but found none. 
Jun 23 17:10:12 <NAiL> [g2]-away: Yay. Mysql build finished. Around 13 hours ;)
Jun 23 17:11:55 <jacques> that's insane
Jun 23 17:12:12 <jacques> haven't seen build times like than since I emerged gentoo on my 32MB ipaq handheld
Jun 23 17:13:33 <NAiL> haha
Jun 23 17:14:16 <NAiL> Ok... anyone familiar with how oe builds busybox? I've applied a patch, but the binary produced is identical to the unpatched one...
Jun 23 17:15:02 <NAiL> The patch is applied, I see that in the patched file in tmp/work/
Jun 23 17:30:20 * NAiL goes insane
Jun 23 17:31:39 <NAiL> argument handling in busybox is ... weird. And with comments like these, who can go wrong?
Jun 23 17:31:42 <NAiL> If sequential founded option haved from this string
Jun 23 17:31:44 <NAiL> then your incongruously pairs unsets and complementaly make add sets.
Jun 23 17:32:17 <dyoung-web> You are speaking The English!
Jun 23 17:33:17 <NAiL> Oh yes. I wish other people were too :P
Jun 23 17:33:54 <jacques> lol that's ... strange
Jun 23 17:34:07 <NAiL> heheh, yeah
Jun 23 17:34:18 <dyoung-web> The Engrish so Fun!
Jun 23 17:34:30 <NAiL> For this result you must set bb_opt_complementaly = "s-d:d-s".
Jun 23 17:34:31 <NAiL> Result have last option flag only from called arguments.
Jun 23 17:34:54 <dyoung-web> When you are utilizing the argument, the string of incongrous pairs make complinetaly!
Jun 23 17:35:02 <NAiL> rofl
Jun 23 17:36:07 <NAiL> Trying to make sense so I can add a parameter. But ... I don't speak engrish. 
Jun 23 17:36:24 <NAiL> I thought the comments were supposed to help, not to confuse
Jun 23 17:36:48 <NAiL> (It probably doesn't help that I've got a fever)
Jun 23 17:42:37 <NAiL> [g2]: Anything else I should build natively? My slug is bored :P
Jun 23 17:42:53 <[g2]> NAiL, you over-acheiver!
Jun 23 17:43:17 <NAiL> haha ;)
Jun 23 17:43:31 <[g2]> I do want to run the LTP
Jun 23 17:43:47 <NAiL> ~google LTP
Jun 23 17:43:48 <[g2]> it's like 2K tests or something like that
Jun 23 17:43:54 <NAiL> ooh
Jun 23 17:44:04 <[g2]> probably more like 3K
Jun 23 17:44:07 <[g2]> now
Jun 23 17:44:29 <[g2]> http://ltp.sourceforge.net/
Jun 23 17:44:33 <NAiL> hahaha, first hit on google: Liturgy Training Publications (LTP)
Jun 23 17:44:38 <NAiL> Yeah, found that as well ;)
Jun 23 17:44:39 <[g2]> 2900+
Jun 23 17:46:04 <[g2]> the uClibc guys had some arm patches for the suite
Jun 23 17:46:14 <[g2]> I took a look at it many months ago
Jun 23 17:46:30 <[g2]> but the native tool environ wasn't up for it yet
Jun 23 17:47:19 <NAiL> 17mb tarball. Joy! ;)
Jun 23 17:47:56 <NAiL> I guess we'll know sometime tomorrow 
Jun 23 17:48:19 <dyoung-web> I've been wondering, how long does it take for perl to build native on slug?
Jun 23 17:48:39 <jacques> ./Configure -de
Jun 23 17:48:39 <jacques> real    18m53.431s
Jun 23 17:48:39 <jacques> make
Jun 23 17:48:39 <jacques> real    89m2.520s
Jun 23 17:48:39 <jacques> make test
Jun 23 17:48:40 <jacques> ~ 75 minutes
Jun 23 17:49:33 <[g2]> if you fix up procps first then you'll score 100%
Jun 23 17:49:53 <[g2]> Failed 1 test script out of 878, 99.89% okay.
Jun 23 17:50:02 <jacques> yep
Jun 23 17:50:14 <jacques> did you have to do anything with /etc/hosts ?
Jun 23 17:50:21 <[g2]> jacques, so lets talk native environment
Jun 23 17:50:24 <jacques> I used to have to, and now I do just in case
Jun 23 17:51:27 <[g2]> rwhitby-web (and I agree with him) wants to have a simple script that sets up the native environment we use
Jun 23 17:51:34 <[g2]> I think we are really close
Jun 23 17:52:26 <NAiL> [g2]: I took the liberty of adding ncurses-dev and a strip-symlink, that ok by you?
Jun 23 17:52:48 <[g2]> NAiL, fine by me
Jun 23 17:52:50 <jacques> so what still needs to be done ?
Jun 23 17:53:19 <dyoung-web> 75min isnt too bad.
Jun 23 17:53:27 <[g2]> I think the script just needs to be added to the repo and a couple of sed commands (maybe not anymore)
Jun 23 17:53:37 <jacques> that's just make test
Jun 23 17:54:01 <jacques> configure 19 minutes, make 90 minutes, make test 75 minutes
Jun 23 17:54:01 <dyoung-web> oh, 90m   75m  total
Jun 23 17:54:08 <dyoung-web> plus.
Jun 23 17:54:21 <jacques> gotta configure first so add 19
Jun 23 17:54:23 <[g2]> we can also change the default in /proc/cpu/alignment to 5
Jun 23 17:54:24 <dyoung-web> stupid web interface doest let me type the plus  cha
Jun 23 17:54:34 <jacques> [g2], why?
Jun 23 17:54:35 <NAiL> [g2]: tail -n 1 libc.so and the other .so is still needed in 2.0
Jun 23 17:54:53 <[g2]> NAiL, good idea
Jun 23 17:55:39 <[g2]> Ok so we got the two tails
Jun 23 17:55:43 <jacques> I'm uncomfortable with alignment being other than 0 or 1
Jun 23 17:56:17 <[g2]> well perl 5.8.7 configure.gnu fails misserably without it
Jun 23 17:56:32 <NAiL> [g2]: It shouldn't be very hard to add a ipkg that includes the necessary kernel headers? 
Jun 23 17:56:38 <jacques> [g2], fails how?
Jun 23 17:56:44 <[g2]> it checks for alignment and will loop printing error messages if its not setup that way
Jun 23 17:57:03 <jacques> have you tried with 0 ?
Jun 23 17:57:22 <[g2]> no but 5 worked beautifully
Jun 23 17:57:37 <jacques> every time that number goes up, performance goes down
Jun 23 17:58:00 <[g2]> 5 is signal and warn
Jun 23 17:58:16 <jacques> ok
Jun 23 17:58:22 <[g2]> we don't generate but 6 of those during the startup
Jun 23 17:58:39 <jacques> yeah but the checking itself isn't free
Jun 23 17:58:49 <jacques> never mind any fixup
Jun 23 17:58:50 <NAiL> hrrm..
Jun 23 17:58:55 <NAiL> [: ==: unknown operand
Jun 23 17:58:55 <NAiL> make: *** [install] Error 1
Jun 23 17:59:15 <[g2]> well 0 is total ignore right ?
Jun 23 17:59:21 <[g2]> for user space
Jun 23 17:59:41 <jacques> right
Jun 23 17:59:55 <jacques> maybe 5 isn't any worse than 1
Jun 23 18:00:05 <[g2]> that's not an acceptable default in mind
Jun 23 18:00:09 <[g2]> 0 that is
Jun 23 18:00:15 <jacques> why not?
Jun 23 18:00:21 <jacques> for me default should be 0
Jun 23 18:00:23 <[g2]> because it's silent
Jun 23 18:00:56 <jacques> if someone wants to lower their performance and get notified, they should have the skill to echo whatever > /proc/cpu/alignment
Jun 23 18:01:06 <[g2]> well I then you may have a bug when perl config does it's alignment check
Jun 23 18:01:30 <[g2]> because 1 printed out like 4K+ messages for me today
Jun 23 18:01:52 <[g2]> till ctrl-c it
Jun 23 18:01:56 <jacques> I need to verify some things on that
Jun 23 18:02:15 <[g2]> wget perl-5.8.7 ./configure.gnu
Jun 23 18:02:26 <jacques> for me, on glibc 2.3.2 I have that problem and I think I actually changed alignment to 0 during the test and then it continued on
Jun 23 18:02:43 <jacques> I do ./Configure -de
Jun 23 18:02:59 <jacques> on glibc 2.3.5 I think it printed about 2 errors and that was it
Jun 23 18:03:09 <[g2]> this is the test that fails
Jun 23 18:03:17 <[g2]> tzname[] found.
Jun 23 18:03:17 <[g2]> (Testing for character data alignment may crash the test.  That's okay.)
Jun 23 18:03:23 <jacques> yes I know
Jun 23 18:03:27 <[g2]> ok
Jun 23 18:03:36 <jacques> I've pasted same thing in like three channels
Jun 23 18:03:38 <[g2]> others may probably don't know
Jun 23 18:04:15 <[g2]> ok. so there are the reasons we should standardize or at least communicate what we are doing
Jun 23 18:04:31 <jacques> anyway I guess it's not worth arguing about alignment setting because it's trivial for me to change it for myself
Jun 23 18:04:33 <[g2]> you did all the hard work on the original native compile
Jun 23 18:04:38 <rwhitby-web> amen.  put it in a script or Makefile :-)
Jun 23 18:04:56 <rwhitby-web> if it's done manually, then it will be left out
Jun 23 18:05:00 <jacques> well I had several issues setting up banana
Jun 23 18:05:13 <jacques> one is that not all the devel apps I like to install are even in symlinks
Jun 23 18:05:32 <jacques> I was reporting on this as I went along in some channel
Jun 23 18:06:19 <jacques> from my notes:
Jun 23 18:06:20 <rwhitby-web> nod
Jun 23 18:06:22 <jacques> currently missing from openslug-packages
Jun 23 18:06:22 <jacques> binutils gnu-config patch pkgconfig
Jun 23 18:06:34 <jacques> and I think patch isn't even in symlinks
Jun 23 18:07:04 <rwhitby-web> nod
Jun 23 18:07:42 <jacques> [g2], you have bison in your script, but it's broken when oe-built
Jun 23 18:08:13 <rwhitby-web> BTW, I think we are ready to start using www.nslu2-linux.org/Makefile as the reference development environment for all nslu2-linux sub-projects
Jun 23 18:08:43 <jacques> would be nice if oe packaged the kernel include files needed for devel
Jun 23 18:08:54 <jacques> might not be that hard to do
Jun 23 18:12:09 <NAiL> jacques: I think that's a 5 min job
Jun 23 18:13:16 <NAiL> hmm.. anyone mind adding me as a dev to slugbug? I'm starting to get the hang of this stuff ;)
Jun 23 18:16:18 <[g2]> jacques, right the kernel includes needs to be added to OE
Jun 23 18:16:19 <ka6sox-office> you have to sign up
Jun 23 18:16:37 <NAiL> ka6sox-office: sign up?
Jun 23 18:17:02 <ka6sox-office> go to slugbug.nslu2-linux.org and you can register yourself.
Jun 23 18:17:09 <NAiL> I'm registered as 'repvik'
Jun 23 18:17:11 <ka6sox-office> after that I can change your status.
Jun 23 18:17:21 <ka6sox-office> okay I'll upgrade it.
Jun 23 18:17:28 <NAiL> thanks :)
Jun 23 18:18:02 <[g2]> NAiL, didn't you use bison today with the native MySQLl build ?
Jun 23 18:18:20 <[g2]> MySQL
Jun 23 18:18:22 <NAiL> [g2]: No idea, but I wouldn't exactly be surprised. 
Jun 23 18:18:25 <ka6sox-office> NAiL, done
Jun 23 18:18:36 <NAiL> I had to remove the sources, because I was low on space
Jun 23 18:18:44 <jacques> I suppose it's possible they fixed it in oe in the last week, but I doubt it
Jun 23 18:18:58 <NAiL> Any simple way I can test bison?
Jun 23 18:18:59 <[g2]> NAiL, you didn't build bison yourself right ?
Jun 23 18:19:04 <NAiL> [g2]: nope
Jun 23 18:19:12 <jacques> bison is usually only needed if one changes the grammar files
Jun 23 18:20:29 <jacques> trying to remember what I was building that needed bison
Jun 23 18:48:59 <NAiL> [g2]: You have to do the alignment stuff for configure.gnu to work? I didn't do that... 
Jun 23 18:49:33 <[g2]> NAiL, this was for a naitve perl 5.8.7
Jun 23 18:49:46 <[g2]> I didn't have to do it for 5.8.4
Jun 23 18:50:07 <[g2]> or I didn't see it in the log
Jun 23 18:50:20 <NAiL> I built 5.8.7
Jun 23 18:50:35 <NAiL> but I've deleted the build dir now though :(
Jun 23 18:50:53 <[g2]> well it's a test in ./configure.gnu
Jun 23 18:51:33 <NAiL> hmm... why didn't that bite me?
Jun 23 18:51:45 <[g2]> dunno
Jun 23 18:51:53 <[g2]> maybe it errors out
Jun 23 18:51:58 <[g2]> after a long time
Jun 23 18:52:17 <[g2]> I saw the messages on the serial console scrolling by
Jun 23 18:52:42 <NAiL> ah, I didn't watch it that closely
Jun 23 18:53:06 <NAiL> intriguing... Gonna try again
Jun 23 19:04:21 <NAiL> tzname[] found.
Jun 23 19:04:22 <NAiL> (Testing for character data alignment may crash the test.  That's okay.)
Jun 23 19:04:22 <NAiL> It seems that you must access character data in an aligned manner.
Jun 23 19:04:32 <NAiL> and it continues on
Jun 23 19:04:48 <NAiL> one alignment trap
Jun 23 19:06:13 <NAiL> [g2]: ping?
Jun 23 19:06:30 <[g2]> NAiL, Pong
Jun 23 19:06:47 <NAiL> No bucketloads of messages here
Jun 23 19:07:20 <[g2]> what is /proc/cpu/alignment set at ?
Jun 23 19:07:33 <NAiL> 1
Jun 23 19:07:48 <NAiL> if it's the last line you want, that is
Jun 23 19:07:56 <[g2]> nod.
Jun 23 19:08:27 <[g2]> and you ran the default ./configure.gnu for perl 5.8.7 ?
Jun 23 19:08:48 <NAiL> tar zxvf tarball && cd perl-5.8.7 && ./configure.gnu
Jun 23 19:09:00 <[g2]> which glibc are running ?
Jun 23 19:09:20 <NAiL> 2.3.3
Jun 23 19:09:49 <[g2]> well that make absolutely no sense to me
Jun 23 19:10:03 <NAiL> heh, glad to be of help :P
Jun 23 19:10:22 <[g2]> the test clearly check for alignment errors and with the warning level of 1 they *should* be generated
Jun 23 19:12:37 <NAiL> hmm.. should I build perl again? See if it stuffs up anywhere?
Jun 23 19:21:43 * NAiL starts a build and goes to bed again
Jun 23 19:36:25 <ov2k> hello
Jun 23 19:36:38 <ov2k> anybody around?
Jun 23 19:37:33 <NAiL> kinda
Jun 23 19:37:37 <ov2k> cool
Jun 23 19:37:56 <ov2k> I'm interested in getting openvpn working
Jun 23 19:38:06 <ov2k> there's an ipkg
Jun 23 19:38:38 <ov2k> but it doesn't pass the self-tests
Jun 23 19:38:47 <ov2k> which is the same as what happens when I build it myself
Jun 23 19:39:42 <ov2k> when built from source, unmodified, I get an assertion failure in a static inline function
Jun 23 19:39:58 <ov2k> the inputs fail some bounds checking
Jun 23 19:40:49 <ov2k> it's all integer conditionals
Jun 23 19:40:57 <NAiL> hmm
Jun 23 19:41:30 <NAiL> I don't think anyone has successfully gotten openvpn to run yet
Jun 23 19:41:35 <NAiL> you using openslug?
Jun 23 19:41:45 <ov2k> when I insert an fprintf(stderr, ... line
Jun 23 19:41:47 <ov2k> it passes
Jun 23 19:42:06 <ov2k> if I insert an if() fprintf
Jun 23 19:42:25 <ov2k> and test on the assert conditional
Jun 23 19:42:27 <ov2k> it passes
Jun 23 19:42:40 <ov2k> I'm just a little worried that this implies some bad things somewhere deep in the system
Jun 23 19:42:55 <ov2k> particularly because nothing I do should fix the assert failure
Jun 23 19:43:05 <ov2k> unslung 5.5-beta
Jun 23 19:43:11 <NAiL> mmh
Jun 23 19:43:32 <ov2k> as far as we know, the integer part of the slug is correct, right?
Jun 23 19:43:44 <ov2k> no weird errors should be coming out of that?
Jun 23 19:44:14 <NAiL> It really should be, haven't noticed anything crazy yet.
Jun 23 19:44:33 <NAiL> ssh/dropbear would probably choke on that
Jun 23 19:44:49 <ov2k> true enough
Jun 23 19:44:53 <ov2k> and I use them all the time
Jun 23 19:45:12 <NAiL> exactly..
Jun 23 19:45:26 <ov2k> I should ask on the openvpn lists too
Jun 23 19:45:46 <ov2k> I just wanted to check that there was nothing with the slug that might cause this
Jun 23 19:46:10 <NAiL> I'm gonna build it on openslug and see if it fails
Jun 23 19:46:44 <ov2k> the test lines are
Jun 23 19:46:45 <ov2k> ./openvpn --genkey --secret key
Jun 23 19:46:51 <ov2k> ./openvpn --test-crypto --secret key
Jun 23 19:46:55 <ov2k> it fails on the second
Jun 23 19:47:09 <ov2k> around line 70-something
Jun 23 19:48:20 <ov2k> the thing that bothers me is that it goes away when I insert the if fprintf line
Jun 23 19:48:25 <NAiL> this may take a little while. Building natively.
Jun 23 19:48:32 <ov2k> the ipkg has the same problem
Jun 23 19:48:47 <ov2k> I think it's the straight ./configure; make setup
Jun 23 19:49:12 <ov2k> I've heard some hints that it might be related to threading
Jun 23 19:49:35 <ov2k> from one google search
Jun 23 19:49:43 <ov2k> but the error is slightly different there
Jun 23 19:52:29 <NAiL> I'll see what happens when my build is finished. Could you file a bug and assign it to me? I've gotta go to bed, getting up in ~4 hours
Jun 23 19:52:37 <NAiL> me == repvik
Jun 23 19:52:38 <ov2k> sure
Jun 23 19:52:41 <ov2k> thanks for the help
Jun 23 19:52:50 <NAiL> np
Jun 23 19:53:03 <[g2]> NAiL, is on slug high :)
Jun 23 19:53:21 <[g2]> you've just been on fire that last week!
Jun 23 19:53:30 <ov2k> although how do I assign it?
Jun 23 19:53:39 <ov2k> I'm not too familiar with SlugBug
Jun 23 19:54:08 <[g2]> samba, aeroslug, conf files, perl, mysql, openvpn, .....
Jun 23 19:54:45 <NAiL> ov2k: just create it, then assign it afterwards
Jun 23 19:54:57 <NAiL> [g2]: not to mention init scripts ;)
Jun 23 19:55:22 <[g2]> the list just goes on 
Jun 23 19:55:23 <NAiL> as soon as I fix that darn busybox bug I'm gonna chrun out init scripts till my arm falls off :P
Jun 23 19:55:31 <NAiL> s/chrun/churn/
Jun 23 19:55:48 <[g2]> you've got "run" on the brain :)
Jun 23 19:55:49 <ov2k> go openslug go!
Jun 23 19:55:55 <NAiL> hahaha
Jun 23 19:56:11 <[g2]> openslug rocks!
Jun 23 19:56:33 <NAiL> yup
Jun 23 19:57:02 <ov2k> will transition someday
Jun 23 19:57:28 <ov2k> perhaps when I'm less worried about not having serial access
Jun 23 19:57:59 <NAiL> not having serial access?
Jun 23 19:59:13 <ov2k> I recall there being warnings about not trying openslug without serial access
Jun 23 19:59:19 <ov2k> perhaps I'm a little behind the times
Jun 23 19:59:45 <NAiL> I wouldn't know. I haven't been serial-enabled yet
Jun 23 19:59:50 <[g2]> well being that the openslug co-lead doesn't have serial access I'd say so :)
Jun 23 20:00:12 <[g2]> NAiL, do you run netconsole
Jun 23 20:00:47 <[g2]> NAiL, NOW I know why you don't get those error messages on the alignment traps :)
Jun 23 20:00:48 <NAiL> *cough*no*cough*
Jun 23 20:01:02 <[g2]> no serial, no net-console
Jun 23 20:01:03 <NAiL> they don't show up in dmesg?
Jun 23 20:01:13 <[g2]> have you looked :)
Jun 23 20:01:22 <NAiL> yes. There was exactly one
Jun 23 20:01:31 <NAiL> [<c0109930>]
Jun 23 20:01:31 <NAiL> Alignment trap: try (8212) PC=0x000085e0 Instr=0xe79e1000 Address=0xbefffdd9 FSR 0x013
Jun 23 20:01:44 <[g2]> that was looping for me
Jun 23 20:01:56 <[g2]> what's the count 
Jun 23 20:02:04 <NAiL> no count
Jun 23 20:02:05 <[g2]> in /proc/cpu/alignment ?
Jun 23 20:02:08 <NAiL> oh
Jun 23 20:02:32 <NAiL> User 1, System 6, Half 6
Jun 23 20:02:45 <[g2]> ok my User was like 4K+
Jun 23 20:02:51 <[g2]> almost 5K iirc
Jun 23 20:04:22 <ov2k> sheesh, I should lurk on here more often
Jun 23 20:04:27 <[g2]> cat /proc/cpu/alignment
Jun 23 20:04:27 <[g2]> User:           4133
Jun 23 20:04:27 <[g2]> System:         6
Jun 23 20:04:33 <ov2k> this is a good way to learn
Jun 23 20:04:39 <NAiL> [g2]: My build is 20050621181404, if that would matter
Jun 23 20:04:51 <NAiL> ov2k: Agreed :D
Jun 23 20:05:06 <[g2]> ov2k, what do you want to learn ?
Jun 23 20:06:34 <NAiL> gah, openvpn requires lzo 
Jun 23 20:06:36 <NAiL> *building*
Jun 23 20:07:21 <ov2k> I'm not very familiar with linux
Jun 23 20:07:31 <ov2k> I know my way around the system from a user's perspective
Jun 23 20:07:44 <ov2k> but not from administering or really understanding how the heck things work
Jun 23 20:07:53 <ov2k> there's just a lot to learn about
Jun 23 20:07:56 <NAiL> [g2]: Hmm... the build stamp... it's not a fixed TZ. Wouldn't that kinda be a good idea?
Jun 23 20:08:39 <ov2k> the build stamp doesn't assume UTC?
Jun 23 20:08:50 <[g2]> I think it does
Jun 23 20:08:50 <NAiL> nope
Jun 23 20:08:55 <ov2k> haha
Jun 23 20:08:55 <NAiL> hmm? 
Jun 23 20:09:00 <[g2]> :)
Jun 23 20:09:03 * NAiL checks tz on the box
Jun 23 20:09:36 <NAiL> [g2]: I'm GMT+2, and the stamp was local time.
Jun 23 20:09:51 <[g2]> I'm teasing you
Jun 23 20:10:02 <[g2]> cause I know you're missing sleep
Jun 23 20:10:10 <NAiL> bastard :P
Jun 23 20:10:15 <[g2]> LOL
Jun 23 20:10:20 <ov2k> aren't addictions great?
Jun 23 20:10:29 <[g2]> what addication
Jun 23 20:10:35 <NAiL> mine
Jun 23 20:10:38 <[g2]> OCD
Jun 23 20:10:52 <NAiL> ocd?
Jun 23 20:11:05 <NAiL> of course, duh?
Jun 23 20:11:10 <[g2]> :)
Jun 23 20:12:13 <[g2]> ov2k, gl on your voyage of discovery
Jun 23 20:12:56 <ov2k> thanks
Jun 23 20:13:03 <ov2k> it's been an adventure already
Jun 23 20:13:31 <NAiL> It's great fun :D
Jun 23 20:13:42 * NAiL stumbles to bed 
Jun 23 20:13:50 <[g2]> sweet dreams
Jun 23 20:13:58 <[g2]> in UTC of course
Jun 23 20:14:04 <NAiL> (actually, /me tilts slightly to the left and pulls up his duvet)
Jun 23 20:14:30 <NAiL> Ready to sleep-hack openslug :P
Jun 23 20:14:33 <NAiL> g'nite
Jun 23 20:14:55 <[g2]> they say hackers dream in assember :)
Jun 23 20:15:32 <ov2k> haha
Jun 23 20:15:35 <ov2k> good night
Jun 23 20:15:38 <NAiL> Too many years since I programmed in asm. I probably will soon though, when I get the brand spanking new avr :D
Jun 23 20:16:07 <NAiL> multimedia-capable with usb2. Can't wait to start toying with it
Jun 23 20:17:10 * [g2] builds away
Jun 23 20:24:16 <NAiL> *pling* You've got mail. One new bug ;)
Jun 23 20:24:23 * NAiL really goes to sleep now
Jun 23 20:27:27 <ov2k> haha
Jun 23 20:27:29 <ov2k> good night
Jun 23 21:11:19 <rwhitby-web> unslung oe feed has been updated
Jun 23 22:58:13 <CIA-3> 03bzhou * 10unslung/ (3 files in 2 dirs): upstream upgrade from R10B5 to R10B6
**** ENDING LOGGING AT Thu Jun 23 23:59:56 2005