**** BEGIN LOGGING AT Fri Jul 29 23:59:56 2005
Jul 30 02:02:54 <koen> good morning all
Jul 30 02:05:27 <dyoung> morning.
Jul 30 04:53:01 <rwhitby> DaKa2: where are we now with respect to openslug native building ?
Jul 30 04:54:52 <DaKa2> well, I think its mostly done, I needs testing, probably our own openslug-native.conf if native.conf is not good (and someone else uses it as it is and doesn't want changes)
Jul 30 04:56:06 <rwhitby> ok, let's copy native.conf to openslug-native.conf and start from there.
Jul 30 04:56:19 <DaKa2> oh right, perl-modules meta package..
Jul 30 04:58:07 <DaKa2> I have looked a bit at that, option 1: package with hardcoded depends, but that has to be maintained, 2: package created from perl.bb with recommends: RRECOMMENDS = ${PACKAGES}, but that gives warnings when installed because not all of perls 944 packages gets built
Jul 30 04:58:22 <DaKa2> rwhitby: what do you think? any other options?
Jul 30 05:02:01 <rwhitby> can we tell perl to build empty packages
Jul 30 05:02:14 <rwhitby> I think we work on the second option
Jul 30 05:02:57 <rwhitby> put "ALLOW_EMPTY = 1" in the perl bb
Jul 30 05:03:37 <DaKa2> hm.. ok.. then perl build close to 1000 packages :)
Jul 30 05:04:33 <rwhitby> 850 - 1000 - not much difference :-)
Jul 30 05:04:53 <DaKa2> it could be better to fix how perl generates its list of packages to build
Jul 30 05:05:16 <DaKa2> dunno what OE thinks of that though...
Jul 30 05:05:32 <rwhitby> let's ask them
Jul 30 05:06:34 <DaKa2> if you remove the last do_split_package in perl.inc it will generate a few more packages, and some modules will get split in arch indep and arch dependant packages
Jul 30 05:07:07 <DaKa2> but the list of packages to create will be correct, no empty packages
Jul 30 05:07:07 <rwhitby> why would we want to do that?
Jul 30 05:08:22 <DaKa2> that would be the only way (probably) to fix it, without rewriting package.bbclass or adding a couple of hundred python lines somewhere..
Jul 30 06:27:38 <wrichter> has anyone managed to bitbake cups for openslug-2.3?
Jul 30 06:29:45 <NAiL> not that I know
Jul 30 06:29:57 <NAiL> Can't say I know if anyone has tried either ;)
Jul 30 06:30:58 <NAiL> I can see if it compiles after trying your patch
Jul 30 06:31:04 <rwhitby> wrichter: people are supposed to add things to openslug-packages.bb when they get them to work.  if it's not there, then either someone is not sharing or it hasn't been done.  either way, the effective result is the same (people who don't share simply don't exist in my nslu2-linux world view)
Jul 30 06:32:21 <NAiL> rwhitby: wrichter got the 3CRSHEW696 to work. So he exists :-P
Jul 30 06:32:56 <wrichter> *g* that saves me a few hours in philosophy class...
Jul 30 06:33:56 <NAiL> wrichter: I'm currently updating my tree. I'll check your patch then, and I can see if cups compiles as well 
Jul 30 06:34:03 <rwhitby> NAiL: wrichter shared on the mailing list :-)
Jul 30 06:34:14 <NAiL> yes 
Jul 30 06:34:56 <rwhitby> and if he keeps it up we will have to punish him with write access to the repo :-)
Jul 30 06:34:57 <wrichter> anyway, I'm playing with cups and am hitting a brick wall - something doesn't seem to work right wrt to automake/autoheader/whatever - the compile fails with some #defines missing when run through bitbake
Jul 30 06:35:13 <NAiL> pastebin it
Jul 30 06:35:22 <NAiL> pastebin.ca
Jul 30 06:37:19 <wrichter> NAiL: what should I pastebin?
Jul 30 06:37:27 <NAiL> the error log
Jul 30 06:41:08 <wrichter> http://pastebin.ca/18851
Jul 30 06:42:08 <wrichter> it works fine when compiling for the native platform (./configure && make)
Jul 30 06:42:56 <wrichter> the pristine cups source code 
Jul 30 06:47:11 <wrichter> pastebin is nice - you learn something new every day... here's the 3Com patch w/out the line wrapping all messed up: http://pastebin.ca/18852
Jul 30 07:03:35 <NAiL> wrichter: the patch is in. It'll show up in #nslu2-linux soon
Jul 30 07:44:27 <malfi> There seems to be one patch missing for unionfs-modules, see: http://pastebin.ca/18855, this patch is mentioned in the bb-file, but does not exist in the files subdir.
Jul 30 07:45:23 <koen> malfi: http://www.handhelds.org/hypermail/oe-commits/current/0465.html
Jul 30 07:46:02 <malfi> oh, slowly me, thanks alot
Jul 30 07:48:14 <koen> the webpage (and rss feeds) can be seen at http://cia.navi.cx/stats/project/openembedded/
Jul 30 07:49:28 <rwhitby> mickeyl: ping - do you have a make.patch lying around for the unionfs version you checked in yesterday ?
Jul 30 07:49:54 <koen> rwhitby: mickeyl is away for the next few days
Jul 30 07:50:23 <rwhitby> oh, he must have forgot to change his nick.
Jul 30 08:08:03 <malfi> Browsing the openslug-packages.conf, I wonder about gtk-docs included. Why would anybody want this package on a slug?
Jul 30 08:08:50 <koen> to generate docs?
Jul 30 08:08:52 <rwhitby> it's probably a build dependency for something else
Jul 30 08:09:08 <rwhitby> openslug-packages.conf is not the same as openslug-packages.bb
Jul 30 08:09:50 <rwhitby> the former restricts the bb files that bitbake sees.  the latter determines what is available in the openslug feed for installation on a slug.
Jul 30 08:10:01 <malfi> aaah
Jul 30 08:37:51 <malfi> I need latex on the crosscompile host in order to build libpam :-/
Jul 30 10:06:12 <jbowler-away> Anyone have any opinions, one way or the other, about whether the 'turnup memstick' problem (slugbug 216) should be fixed in OpenSlug-2.3-beta?
Jul 30 10:11:35 <jbowler-away> malfi: isn't this slugbug 186?
Jul 30 10:11:55 <malfi> oh, let me check 
Jul 30 10:13:22 <malfi> this libpam thing? No, I just installed tetex and everything went fine. latex is needed for some doc-stuff.
Jul 30 10:14:20 <jbowler-away> Yeah, but the doc stuff doesn't get built if there is no jade suppport - so I think  not having jade on the build system hides all doc problems.
Jul 30 10:14:27 <malfi> ether slugbug is a PITA or I don't find any full text search...
Jul 30 10:15:23 <jbowler-away>  http://slugbug.nslu2-linux.org/bug.php?op=show&bugid=186&pos=0
Jul 30 10:16:52 <malfi> I've never heart of jade before, it directly said, that latex: no such file or directory
Jul 30 10:19:10 <malfi> I make a pastebin...
Jul 30 10:19:51 <jbowler-away> latex is not used in my (0722) build of libpam
Jul 30 10:20:16 <jbowler-away> And I apparently have no latex installed, no TeX for that matter.
Jul 30 10:20:20 <malfi> http://pastebin.ca/18879
Jul 30 10:20:30 <jbowler-away> I do, however, see this:
Jul 30 10:20:31 <jbowler-away> log.do_configure.14681:checking for sgml2latex... no
Jul 30 10:21:45 <malfi> sgml2latex is part of the sgml-tools
Jul 30 10:21:58 <malfi> I had this installed but no latex
Jul 30 10:23:29 <jbowler-away> Presumably the configure script is incomplete - it should probably be checking for both sgml2latex and latex
Jul 30 10:23:37 <malfi> checking for sgml2latex... yes, probably it thinks that there must latex too, if there is sgml2latex. So it's probably an unresolved dependancy of setup...
Jul 30 10:23:55 <jbowler-away> What OS?
Jul 30 10:24:02 <malfi> slackware 10.1
Jul 30 10:24:12 <malfi> there is no dependency-checking ;)
Jul 30 10:24:32 <jbowler-away> Can you update 186 with that information?  (It's a separate bug really, but the fix is the same...)
Jul 30 10:25:52 <malfi> I can do this, but I think it's not a bug but rather my mistake.
Jul 30 10:26:47 <jbowler-away> I think someone else will hit the same problem (actually I think it is a bug as well, but it's more important that someone else will encounter the same issue).
Jul 30 10:27:29 <jbowler-away> And I will fix it to stop 'doc' being in the list of THINGSTOMAKE - but it will still be an issue in OpenSlug-2.3-beta
Jul 30 10:27:43 <malfi> I've read once again this ticket, I don't get it. It seems like a completly different problem...
Jul 30 10:28:59 <jbowler-away> Ok, I agree - enter it as a separate bug.
Jul 30 10:29:17 <jbowler-away> I was just saying that it's not a big deal either way.
Jul 30 10:29:31 <malfi> I agree with you on that. :-) 
Jul 30 10:30:46 <malfi> by now I'm just trying to find a server which still hosts ypserv-2.17.tar.bz2 (#226)
Jul 30 10:31:18 <malfi> if the build-process is up and running again, I will file that ticket for libpam
Jul 30 10:32:08 <jbowler-away> malfi: it's on the named server, but it's in a sub-directory
Jul 30 10:36:27 <jbowler-away> malfi: it's fixed in OpenSlug-2.3 (but not in the monotone head)
Jul 30 10:36:58 <malfi> see #228 for the pam inconvenience
Jul 30 10:37:49 <malfi> jbowler-away: oh, I checked http://cia.navi.cx/stats/project/openembedded/ as I was told to, and didn't find any change regarding this yellowpages stuff..
Jul 30 10:38:28 <malfi> jbowler-away: so it's okay to file that bug, isn't it?
Jul 30 10:39:11 <jbowler-away> Yes - it's a bug in the head.
Jul 30 10:39:45 <jbowler-away> What I'm saying is that it is not a bug in the release (because I fixed it there locally), so if you get the .bb files from the release it will work.
Jul 30 10:39:59 <jbowler-away> And I just fixed it in revision e8a8ea202059d2f8a7a8f6d316818ee71b449d4a too
Jul 30 10:41:08 <malfi> unfortunatly it's not possible to file a bug for head, so I took 2.4-beta, as this seems the working-tree.
Jul 30 10:46:49 <malfi> Which device has the broadest user base of all these openembedded targets? Openzaurus?
Jul 30 10:48:19 <NAiL> malfi: that's the correct place to file the bug
Jul 30 10:48:35 <landslide> so once i have the master make file and called make build-openslug.... what do I actually get?
Jul 30 10:49:09 <NAiL> You'll get the firmware image, and heaps of packages
Jul 30 10:49:29 <landslide> apache?
Jul 30 10:49:34 <NAiL> not yet
Jul 30 10:49:44 <landslide> nano?
Jul 30 10:50:13 <NAiL> no, but it's probably not that hard to build
Jul 30 10:50:18 <landslide> how do i get access to the arm build environment?
Jul 30 10:50:25 <landslide> i want to compile apex for me 128mb slug.
Jul 30 10:50:42 <malfi> the crosscompiler is built by this mastermakefile
Jul 30 10:50:50 <landslide> sweeet
Jul 30 10:51:48 <landslide> so how far are we from having apache compiling?
Jul 30 10:51:57 <landslide> is it just something someone needs to sit down and work out?
Jul 30 10:52:56 <malfi> landslide: http://pastebin.ca/18886 that are the current openslug packages
Jul 30 10:54:23 <malfi> landslide: as there isn't even an openembedded package for apache, probably nobody tried it or got it working
Jul 30 10:54:41 <landslide> unslug has it though?
Jul 30 10:55:09 <malfi> well, maybe it's somewhere in the optware...
Jul 30 10:55:45 <koen> nonworking/apache/apache_2.0.47.bb 
Jul 30 10:55:57 <koen> it was moved to nonworking as something broke
Jul 30 10:57:12 <landslide> k
Jul 30 11:07:15 <perlguru> apache in unslung works fine
Jul 30 11:07:40 <NAiL> landslide: nano should be in the feed within half an hour ;-)
Jul 30 11:07:45 <perlguru> so check that makefile and patches....
Jul 30 11:10:28 <NAiL> landslide: if you do make update now, you'll get the revision that builds nano as well.
Jul 30 11:12:21 <landslide> wow thanks guys
Jul 30 11:12:26 <landslide> so what are you all actually working on?
Jul 30 11:12:35 <landslide> like is there a particular goal?
Jul 30 11:13:56 <NAiL> My goal is to get my slug to do what I want ;-)
Jul 30 11:15:48 <landslide> so do you guys have like a mini version of x working?
Jul 30 11:15:49 <koen> "roll over" "play dead"
Jul 30 11:15:51 <landslide> like tinyx?
Jul 30 11:16:13 <landslide> vncserver.. etc...
Jul 30 11:16:36 <landslide> php4/5?
Jul 30 11:19:16 <NAiL> landslide: people are working on getting apache/php working (by compiling natively). 
Jul 30 11:19:31 <NAiL> dunno if anyone has tried any xserver stuff
Jul 30 11:19:40 <NAiL> kdrive might be the most suitable... dunno
Jul 30 11:20:12 <landslide> ok...
Jul 30 11:22:04 <landslide> and openslug is not using busybox at all?
Jul 30 11:22:13 <NAiL> feel free to try to compile any xserver ;-)
Jul 30 11:22:24 <NAiL> openslug uses busybox extensively
Jul 30 11:22:32 <landslide> ok.
Jul 30 11:22:50 <landslide> it is taking AGES to do build-openslug.
Jul 30 11:22:57 <landslide> i think it'll be still going tomorrow morning.
Jul 30 11:23:09 <NAiL> The first time take a very long time
Jul 30 11:23:12 <NAiL> takes
Jul 30 11:23:26 <NAiL> it has to set up the cross compile environment
Jul 30 11:23:33 <NAiL> and gcc/glibc is a monster to compile
Jul 30 11:23:44 <landslide> hehe
Jul 30 11:23:49 <landslide> where does it put it?
Jul 30 11:24:00 <NAiL> in openslug/tmp/staging, I guess
Jul 30 11:24:14 <NAiL> image in openslug/tmp/deploy/image
Jul 30 11:24:15 <landslide> ok
Jul 30 11:24:20 <malfi> http://www.nslu2-linux.org/wiki/HowTo/UseTheMasterMakefile
Jul 30 11:24:20 <NAiL> packages in deploy/ipk
Jul 30 11:24:58 <malfi> just updated that wiki entry about the directory structure, please correct it, if I got anything wrong
Jul 30 11:24:58 <landslide> ahh.
Jul 30 11:25:19 <landslide> very helpful
Jul 30 11:26:39 <landslide> well it looks like the compiler is built
Jul 30 11:26:43 <landslide> it's up to ncurses package.
Jul 30 11:26:46 <landslide> shouldn't be long.
Jul 30 11:27:09 <malfi> a make build-openslug takes many hours
Jul 30 11:28:40 <landslide> if i wanted to compile apex... is there somewhere i should put the source?
Jul 30 11:28:47 <landslide> an environment that needs to be setup?
Jul 30 11:28:52 <landslide> as it needs the arm assembler?
Jul 30 11:29:35 <malfi> you might ask in #nslu2-linux, apex is beyond openslug, I guess
Jul 30 11:30:28 <landslide> ok no worries.
Jul 30 11:33:19 <jbowler-away> Ask in #openjtag
Jul 30 11:35:14 <NAiL> malfi: nice work on the wiki
Jul 30 11:36:29 <malfi> thanks :)
Jul 30 11:36:30 <landslide> ok
Jul 30 11:36:43 <jbowler-away> Yes, great to see some documentation!
Jul 30 14:05:40 <malfi> NOTE: build 200507302149: completed
Jul 30 14:05:43 * malfi cheers.
Jul 30 14:09:54 <malfi> Probably a stupid question, but it seems as the md5sums are generated after the tarballs of this and that are downloaded. Shouldn't be the md5sum in the bitbake-file or somewhere else in the distribution to assure data integrity?
Jul 30 14:10:24 <jbowler-away> Yes, it should be.
Jul 30 14:10:47 * koen|tv asks "which md5sums?"
Jul 30 14:11:17 <malfi> those in ./downloads
Jul 30 14:11:54 <koen|tv> those aren't md5sums :)
Jul 30 14:12:19 <koen|tv> they are the result of 'touch bla.tar.gz.md5'
Jul 30 14:12:43 <malfi> cat downloads/busybox-1.00.tar.gz.md5 -> fa62459e098fc00b22772aaf2e75bc98
Jul 30 14:13:22 <koen|tv> ah, zecke added real md5 support
Jul 30 14:14:00 <koen|tv> you can set the md5 in the .bb, but I forgot which syntax we agreed on 
Jul 30 14:14:28 <malfi> ah, i see. But how it's now it doesn't  help a thing.
Jul 30 14:15:01 <koen|tv> ../qte/qte_2.3.10.bb:SRC_URI = "ftp://ftp.trolltech.com/pub/qt/source/qt-embedded-${PV}-free.tar.gz;md5=1f7ad30113afc500cab7f5b2f4dec0d7 \
Jul 30 14:15:29 <malfi> ah okay, I've got the idea.
Jul 30 14:15:33 <koen|tv> bitbake will (not implemented yet) match that against the .md5 in ${DL_DIR} 
Jul 30 14:16:43 * malfi nods understanding.
Jul 30 14:18:00 <malfi> understandingly (it should be an adverb)
Jul 30 16:13:43 <malfi> you guys broke ixp425 once again? g
Jul 30 16:16:01 <malfi> http://pastebin.ca/18903
Jul 30 16:52:49 <jbowler-away> malfi: I think your 'rm -r' command didn't work
Jul 30 16:56:26 <dyoung> or hes got the patch with the added lf cached someplace.
Jul 30 16:57:06 <dyoung> ah, just got to the bottom; jbowler is right.
Jul 30 16:58:11 <jbowler-away> There are some bogus permissions: tar: ixp425-eth-1.1-r6/.pc/makefile.patch/Makefile: Cannot open: Permission denied
Jul 30 17:07:29 <rwhitby> jbowler-away: is it "uCSlugC" or "uCSlugc" or "UcSlugC" ?
Jul 30 17:09:54 <jbowler-away> uCslugc (see www.uclibc.org)
Jul 30 17:10:43 <jbowler-away> There are, in fact, only about 6 'openslug' directories in the openembedded source, so it's quite manageable to create a new distro.
Jul 30 17:10:44 <rwhitby> Not uCSlugc ? ("Slug" being a proper name?)
Jul 30 17:11:04 <jbowler-away> Umm... :-)
Jul 30 17:11:50 <rwhitby> how's this: DISTRO=ucslugc, proper name is "uCSlugc" ..
Jul 30 17:11:54 <jbowler-away> Well, the directories will all be ucslugc and the bb variables UCSLUGC, so I guess we could make it cool by calling it uCsLuGc, or have I got this all wrong.
Jul 30 17:12:17 <jbowler-away> Yep, that works for me
Jul 30 17:12:50 <jbowler-away> uCSlugc it is
Jul 30 17:13:41 <rwhitby> ok, feel free to add it to the master makefile (ucslugc, sibling to unslung and openslug), and we can start officially supporting it if you like.
Jul 30 17:14:02 <rwhitby> and we'll see if you can do it without interrupting the autobuilds :-)
Jul 30 17:14:50 <rwhitby> all the infrastructure should be in place today to simply add a new distro to our autobuilds
Jul 30 17:14:50 <jbowler-away> I think that would be good - there seem to be a few people around who want configurable flash only or flash mainly systems.
Jul 30 17:16:01 <jbowler-away> It's also good to prove that adding a new distro is easy (or at least prove how it can be done) - then no one has an excuse for distributing modified openslug firmware as 'openslug'
Jul 30 17:16:51 <dyoung> There seem to be a bunch of application specific fooslug distros that people are talking about...
Jul 30 17:17:15 <dyoung> AudioSlug RoboSlug FrogLogginSlug CameraSlug, etc
Jul 30 17:17:55 <rwhitby> yeah, but we should only need the openslug and ucslugc feeds ...
Jul 30 17:18:59 <dyoung> Right, and actually half those I just mentioned are for "inslung" type application anyways.
Jul 30 17:21:05 <jbowler-away> My idea is to insert the (literal) 'openslug' override into OVERRIDES after ${DISTRO} in osuclibc.conf, then openslug provides a default definition 
Jul 30 17:21:38 <rwhitby> osuclibc.conf?
Jul 30 17:22:00 <rwhitby> not ucslugc.conf?
Jul 30 17:22:06 <jbowler-away> Oops, wrong name.
Jul 30 17:22:19 <jbowler-away> ucslugc.conf
Jul 30 17:22:28 <rwhitby> but yeah, that's a great idea
Jul 30 17:23:24 <malfi> sorry, i've been disconnected, what do you mean with the rm failure?
Jul 30 17:24:45 <jbowler-away> Your rm -r command has to fail, because there is a file with permissions 000 in there.
Jul 30 17:25:10 <rwhitby> jbowler-away: I'm wondering if the wiki will require the name to be UcSlugc ..
Jul 30 17:25:29 <malfi> actually I invoked this rm with sudo, and it worked
Jul 30 17:25:58 <jbowler-away> We seem to have settled on using [[page|name]] so I think it will work, but that might be a problem.
Jul 30 17:26:34 <rwhitby> All our other distros have an initial capital
Jul 30 17:27:32 <jbowler-away> Well, that works too - I suspect technically the 'u' in uClibc is greek (i.e. micro) but that doesn't matter a whole lot...
Jul 30 17:27:59 <jbowler-away> malfi: the log you posted did not have a sudo command
Jul 30 17:28:15 <rwhitby> MicroSlugC ?
Jul 30 17:28:21 <rwhitby> MuSlugC ?
Jul 30 17:28:52 <malfi> the user who invoked it, has id 0...
Jul 30 17:28:54 <rwhitby> UcSlugC ?
Jul 30 17:28:58 <landslide> problem with microcom.port5.com.... m102.tar.gz.
Jul 30 17:29:02 <landslide> anyone know why?
Jul 30 17:29:05 <jbowler-away> Prefer UcSlugc, or UcSlugC, it just sounds better.
Jul 30 17:29:41 <rwhitby> UcSlugC then?
Jul 30 17:29:43 <malfi> slµg ?
Jul 30 17:30:03 <rwhitby> ooh, very subtle.
Jul 30 17:30:07 <jbowler-away> malfi: I would recommend not building as root.  That's untested and,  in the past, builds have installed stuff on the build system when run as root.
Jul 30 17:30:45 <rwhitby> indeed, you will soon end up the /bin/bash being an arm executable on your x86 host :-)
Jul 30 17:31:04 <malfi> jbowler-away: I know that doing these things as root is evil, but I did that for nfs-convenience...
Jul 30 17:31:32 <malfi> jbowler-away: but you're right, I will change that tomorrow, but do you really think that this getuid is related to the patch-problem?
Jul 30 17:32:07 <malfi> landslide: why? because microcom.port5.com is down :)
Jul 30 17:32:50 <jbowler-away> The error you have is a result of either the patch already being applied (as it said) or the preceding patch not being applied.
Jul 30 17:33:08 <landslide> ah-huh... :)
Jul 30 17:33:18 <jbowler-away> What I don't understand is where you got the -r6 directory from in the first place - what happened in the original build?
Jul 30 17:34:04 <malfi> jbowler-away: the same, but I wanted to be sure, that this is reproducable, so I deleted it, as it should be build again from the openembedded bitbake-files, and the problem reoccured.
Jul 30 17:35:45 <jbowler-away> Maybe one of the other patch files is suffering from a similar problem.  I've definately seen the 'patch already applied' issue before, but it was a long time ago.
Jul 30 17:36:58 <jbowler-away> Here's my log:
Jul 30 17:37:02 <jbowler-away> NOTE: made /home/work-tmp/jbowler/nslu2/openslug/work/ixp425-eth-1.1-r6 to projectdir by creating .patches/
Jul 30 17:37:02 <jbowler-away> NOTE: import patch 'ixp425_eth_1_1_update_nf_bridge.patch' as 'ixp425_eth_1_1_update_nfbridge'
Jul 30 17:37:02 <jbowler-away> NOTE: import patch 'ixp425_eth.c.patch' as 'ixp425eth.c'
Jul 30 17:37:03 <jbowler-away> NOTE: please refresh this patch, line-numbers have changed
Jul 30 17:37:04 <jbowler-away> NOTE: please check ixp425_eth.c.rej
Jul 30 17:38:20 <rwhitby> jbowler-away: http://www.nslu2-linux.org/wiki/UcSlugC/Homepage
Jul 30 17:39:24 <malfi> I will delete all my files in downloads/ixp*, in openembedded/packages/ixp425_eth/ and in openslug/work/ixp425_eth once again, monotone again and see, if anything was corrupted
Jul 30 17:39:45 <rwhitby> when it's ready we can add it to the sidebar too
Jul 30 17:40:25 <rwhitby> we really need to reorganise the sidebar ...
Jul 30 17:41:19 * malfi nods to rwhitby.
Jul 30 18:02:21 <landslide> can i get m102.tar.gz from a mirror host instead of microcom.port5.com?
Jul 30 18:09:21 <malfi> yes
Jul 30 18:09:33 <rwhitby> http://sources.nslu2-linux.org/sources/m102.tar.gz
Jul 30 18:09:36 <malfi> just edit the url in the .bb file
Jul 30 18:10:08 <rwhitby> and if you're using the lastest monotone head, then it should pick that one up automatically as a last resort mirror
Jul 30 18:10:30 <rwhitby> (that was added in the last 12 hours)
Jul 30 18:11:55 <landslide> where is the bb file?
Jul 30 18:12:44 <rwhitby> landslide: you don't need to edit the bb file.  we have sources mirrors now
Jul 30 18:13:37 <landslide> ok, do i need to just run 'make build-openslug' again? to get the new head?
Jul 30 18:13:54 <rwhitby> You'll want to "make update" first
Jul 30 18:14:55 <rwhitby> (be prepared for a monotone wait if you're not running "make update" regularly)
Jul 30 18:15:09 <landslide> ok
Jul 30 18:15:47 <malfi> argh
Jul 30 18:15:48 <landslide> it says 'misuse: no server given and no default server set'
Jul 30 18:16:12 <malfi> cat openslug/.patches/applied -> ixp425_eth_1_1_update_nfbridge
Jul 30 18:16:29 <malfi> someday I will understand all those hidden whistle-and-bell-files
Jul 30 18:19:13 <jbowler-away> I think I should have commented out the 'do_ixp425_c_patch_unpack' lines too
Jul 30 18:20:15 <rwhitby> landslide: do a make setup
Jul 30 18:21:47 <rwhitby> you must not have run make setup since moving to monotone 0.21
Jul 30 18:22:11 <rwhitby> (there was a rough transitional period there - things should be smoother from now on ...)
Jul 30 18:29:18 <dyoung> So how do you say UcSlugC ?  You-see-slug-see?  or Uck-slug-see ?
Jul 30 18:29:34 <malfi> the linksys nslu2 link on the wiki startpage leads to a 404
Jul 30 18:29:41 <jbowler-away> dyoung: the first
Jul 30 18:29:57 <malfi> sounds like some gangster rap... :)
Jul 30 18:31:01 <rwhitby> you-see-slug-see is how I say it
Jul 30 18:34:22 <rwhitby> malfi: startpage updated.  Removed the link :-)
Jul 30 18:34:37 * malfi grins.
Jul 30 18:34:39 <rwhitby> malfi: MCSlugC :-)
Jul 30 18:34:54 <rwhitby> em-see-slug-see - a well know slug rapper
Jul 30 18:35:02 <malfi> Well, do I get it right, that you plan to migrate these opt-packages to openembedded?
Jul 30 18:35:49 <rwhitby> optware will continue as is, and people will continue to add packages to it.  Anyone is free (and in fact, encourage) to add the same packages to OpenEmbedded.
Jul 30 18:36:11 <jbowler-away> Or fix the ones which are already there but don't quite work...
Jul 30 18:36:14 <rwhitby> The trouble is that no-one has been able to get OpenEmbedded to emit packages which work on Unslug or the WL500g firmware.
Jul 30 18:36:35 <rwhitby> and the OpenEmbedded learning curve is just too high for a lot of developers.
Jul 30 18:37:09 <malfi> these bb files are not more complicated then these makefiles, are they?
Jul 30 18:38:13 <malfi> rwhitby: Why isn't OE able to produce unslung packages?
Jul 30 18:38:13 <rwhitby> malfi: when they are easy, no. :-)  but as the complexity of the changes required for cross compilation increases, I for one find that a complex .bb is harder than a complex .mk, but that's probably cause my python-foo is not that good.
Jul 30 18:38:35 <rwhitby> http://www.nslu2-linux.org/wiki/Unslung/OEPackagesForUnslung
Jul 30 18:39:02 <rwhitby> hmm - the ldconfig reason on that page may have gone away now ...
Jul 30 18:39:05 * malfi goes reading and stops asking stupid questions for now. :)
Jul 30 18:40:03 <rwhitby> malfi: in general, we can get executables that work without any problems, but turning them into user-installable packages that work and integrate properly is the hard bit.
Jul 30 18:40:30 <malfi> rwhitby: okay I see
Jul 30 18:41:53 <malfi> rwhitby: how many ppl use unslung in contrast to openslug?
Jul 30 18:42:41 <dyoung> malfi, when I checked last week the download stats for unslung binary image, there were 3500+ downloads of unslung-5.5 
Jul 30 18:47:12 <malfi> there isn't even a officially linked openslug-2.3-beta binary firmware image. the source code is available at berlios but the OpenSlug/UsingTheBinary stills says "The current version of the OpenSlug binary image is 2.0"
Jul 30 18:47:25 <malfi> are odd numbers unstable?
Jul 30 18:52:39 <rwhitby> malfi: all numbers are unstable :-)
Jul 30 18:53:01 <dyoung> malfi thats because there isnt a officially released binary of 2.3-beta .
Jul 30 18:53:16 <dyoung> I'm not sure why that is actually....
Jul 30 18:53:27 <dyoung> lighting probably.
Jul 30 18:58:19 <rwhitby> malfi: 1:7 openslug:unslung is the current ratio
Jul 30 18:59:05 <rwhitby> dyoung: yes, a lightning strike on the house of our trusted third party, combined with our trusted third party's helper being in Albania on business
Jul 30 19:02:35 <jbowler-away> No one is going to believe that story.
Jul 30 19:02:55 <jbowler-away> I mean, the fact that it's true really doesn't count for anything.
Jul 30 19:03:19 <dyoung> "we were just too lazy to do it"
Jul 30 19:03:41 <dyoung> although I dont know if anyone else has access to the FRS ?
Jul 30 19:05:00 <rwhitby> I don't.
Jul 30 19:05:18 <landslide> rwhitby: i ran make setup, then tried make update (didn't work same error) so i tried make build-openslug again and it didn't fallback to the mirror for m102.tar.gz
Jul 30 19:05:22 <rwhitby> jbowler-away: yeah, I probably wouldn't believe that if someone said it to me.  But yes, it is true.
Jul 30 19:05:59 <rwhitby> landslide: did make setup do the "monotone unset database  <foo>" lines?
Jul 30 19:06:34 <landslide> lemme check
Jul 30 19:07:15 <landslide> rwhitby: couldn't say, it's too far back in my screen buffer.
Jul 30 19:09:47 <rwhitby> landslide: check the Makefile for those unset commands, run them manually, and run the command that's after them in make setup too
Jul 30 19:12:24 <landslide> ok i ran the unset commands.
Jul 30 19:12:54 <landslide> the command after them does not work.
Jul 30 19:13:04 <rwhitby> then run the monotone pull from monotone.nslu2-linux.org with
Jul 30 19:13:17 <rwhitby> what's the error?
Jul 30 19:13:28 <landslide> just brings up the usage help
Jul 30 19:13:42 <rwhitby> monotone -d monotone/nslu2-linux.db pull monotone.nslu2-linux.org org.openembedded.* org.nslu2-linux.*
Jul 30 19:13:56 <landslide> yep
Jul 30 19:14:12 <rwhitby> sure you typed it right?
Jul 30 19:14:13 <landslide> my monotone is 0.19
Jul 30 19:14:19 <rwhitby> oh, you need 0.21
Jul 30 19:14:29 <landslide> ah-huh.
Jul 30 19:14:49 * rwhitby wonders how we can check that automatically in the makefile ...
Jul 30 19:15:06 <malfi> parse monotone --version ?
Jul 30 19:15:24 <rwhitby> patches welcome :-)
Jul 30 19:15:49 <landslide> hehe.
Jul 30 19:16:01 <landslide> so.. i need to get monotone and build it myself?
Jul 30 19:16:07 <landslide> debian testing is only at 0.19
Jul 30 19:16:21 <rwhitby> landslide: venge.net has a debian pkg for 0.21
Jul 30 19:16:53 <landslide> changing over to linux.. brb.
Jul 30 19:18:57 <malfi> test  `monotone --version | awk '{print $2;}' | cut -d. -f2` -ge 21  && echo everythings fine
Jul 30 19:20:26 <malfi> if monotone is 0.21 or above, then this condition is true
Jul 30 19:21:08 <landslide> ah.
Jul 30 19:21:14 <landslide> what was the link to 0.21 again?
Jul 30 19:22:39 <rwhitby> ~google monotone 0.21
Jul 30 19:22:47 <rwhitby> oh, no jbot here
Jul 30 19:22:58 <rwhitby> http://venge.net/monotone/downloads/
Jul 30 19:23:04 <landslide> ah... got it.
Jul 30 19:46:00 <landslide> what is monotone doing with the rull, and verifying new revisions?
Jul 30 19:46:08 <landslide> is it comparing the delta changes i received?
Jul 30 19:46:20 <landslide> rull=pull
Jul 30 19:52:28 <landslide> there are 160 revs in.. and i'm only up to 34 it's taken about 20 mins
Jul 30 19:52:37 <landslide> is this normal behaviour?
Jul 30 19:52:40 <rwhitby-away> yep
Jul 30 19:53:13 <malfi> night
Jul 30 19:53:16 <rwhitby-away> you'd be better to save your private key (if you have one), and then delete the monotone/nslu2-linux.db and let make setup download a new up-to-date copy
Jul 30 20:02:57 <landslide> it just finished.
Jul 30 20:03:04 <landslide> should i still get a new copy.
Jul 30 20:17:39 <rwhitby-away> no
Jul 30 20:17:44 <rwhitby-away> you've got the latest now
Jul 30 20:24:44 <landslide> boy, this whole process took me 2 days!
Jul 30 20:24:50 <landslide> to get the build environment setup.
Jul 30 20:24:51 <landslide> hehe
Jul 30 20:26:32 <rwhitby-away> yeah, you picked exactly the time of most flux to start it :-)
Jul 30 20:26:54 <rwhitby-away> it would be much quicker if you started again today, right ? :-)
Jul 30 20:32:19 <landslide> ain't that always the way ;P
Jul 30 20:43:44 <rwhitby-away> back later ..
Jul 30 20:51:53 <rwhitby-away> OK, so who is working on the OE distro conf file for openslug native builds?
Jul 30 20:52:17 <rwhitby-away> DaKa2: do you have an initial version that we can start from?
Jul 30 21:46:51 <landslide> i've got monotone latest.
Jul 30 21:47:06 <landslide> ran the pull command and the unsets from the make file manually
Jul 30 21:47:14 <landslide> then ran make update
Jul 30 21:47:22 <landslide> then ran make build-openslug
Jul 30 21:47:46 <landslide> it still doesn't revert to another mirror for m102.tar.gz
Jul 30 21:47:51 <landslide> it's up to attempt 5 atm... 
Jul 30 21:50:33 <dyoung> you *may* h ave to hit control-c to make it go to next mirror... I dont remember how that works
Jul 30 21:50:41 <landslide> ok
Jul 30 21:55:01 <landslide> can i just fudge it by changing the bb file? if so where is the file ;)
Jul 30 22:05:53 <dyoung> the Q&D way is to wget http://sources.nslu2-linux.org/sources/m102.tar.gz -O downloads/m102.tar.gz; md5sum downloads/m102.tar.gz > downloads/m102.tar.gz.md5
Jul 30 22:13:02 <dyoung> if you missed it; I recently said;
Jul 30 22:13:05 <dyoung> the Q&D way is to wget http://sources.nslu2-linux.org/sources/m102.tar.gz -O downloads/m102.tar.gz; md5sum downloads/m102.tar.gz > downloads/m102.tar.gz.md5
Jul 30 22:13:08 <landslide> yep..
Jul 30 22:13:17 <landslide> machine rebooted for no reason :(
Jul 30 22:17:51 <landslide> yes, that worked great.
Jul 30 23:37:01 <DaKa2> rwhitby-away: only thing I have is on http://www.nslu2-linux.org/wiki/HowTo/OpenSlugNativeCompileEnvironment
Jul 30 23:37:13 <DaKa2> and now I have to run, bbl
**** ENDING LOGGING AT Sat Jul 30 23:59:56 2005