**** BEGIN LOGGING AT Sat Jun 25 23:59:56 2005 Jun 26 00:00:20 I can try a build without *any* settings later, and see what happens Jun 26 00:00:57 Yes, I thought about porting openntpd (when I saw it while finding openldap). Jun 26 00:01:22 It's a pita to set up. There's no docs :P Jun 26 00:01:39 But I've figured out the right way (With help from the linux maintainer) Jun 26 00:01:40 I think if without works then that is best - though it might be less than optimal. Jun 26 00:02:22 setting the *64 ones might be nice though, if the rest works out the way it should Jun 26 00:02:27 Can it do tickadj? That's one thing which is essential (well, we should fix it in the kernel...) Jun 26 00:02:55 jbowler: Yeah, there's a patch for it. I've got it around here somewhere Jun 26 00:02:59 Ah yes - good point, because of the 2GByte file problem (and the fact that it comes up so much with unslung). Jun 26 00:03:14 exactly Jun 26 00:04:13 Now, I'm getting output from x1205 without the tickadj patch, I'm not sure if it'll go away or not.. Jun 26 00:04:13 DQBLK might be related (they only appear in cross). Jun 26 00:04:49 Isn't it just whining that it had to adjust the clock? Jun 26 00:05:19 might be, I can't remember seeing anything like that with ntpd Jun 26 00:05:20 Jun 26 06:25:42 LKG7F654F daemon.info ntpd[12720]: adjusting local clock by 54.577980s Jun 26 00:05:23 Jun 26 06:26:07 LKG7F654F user.debug kernel: x1205_sync_rtc exit (change seconds 2) Jun 26 00:05:46 I set my clock back a bit more than a minute to test Jun 26 00:06:01 That happens with NTP too after a shutdown (there's some drift or error). Jun 26 00:06:09 ah, ok Jun 26 00:06:31 it doesn't check ntp.drift afaik... Jun 26 00:06:32 I left it in because I wanted to know when the clock drifted - but it's just a debug log. Jun 26 00:07:23 But that won't matter much, will it? It'll sync with the servers once in a while anyways, right? Jun 26 00:08:18 * NAiL knows little to nothing about ntp Jun 26 00:08:36 One thing I know though.. I need food! :D Jun 26 00:08:44 With ntp the drift file is just a cache of information it builds up over about an hour about the drift in the local clock. Jun 26 00:09:03 It writes the file so that it has the information immediately on reboot. Jun 26 00:10:15 I though about reading the file and using it to adjust the tickadj input, but my value has never gone above about 6... Jun 26 00:14:27 root@LKG7F654F:/# cat /etc/ntp.drift Jun 26 00:14:27 10.398 Jun 26 00:14:49 Not critical, in other words. Jun 26 00:14:59 Which means that the tickadj value is 10101.01 - which is the expected value. Jun 26 00:15:42 It's 100/99 Jun 26 00:16:06 yeah Jun 26 00:16:43 Hmm... I'm kinda getting the hang of OE now. Jun 26 00:17:02 Need to learn how to make new .bb's though Jun 26 00:18:17 Look for the shortest. Jun 26 00:18:32 (Always assuming it doesn't include anything else ;-) Jun 26 00:18:40 heheh, exactly Jun 26 00:19:10 The minimum is two lines - SRC_URI and inherit autotools. Jun 26 00:19:50 Nah, I'm kinda thinking of making a native-development-package for openslug. Basically it includes those kernel headers needed, makes the symlinks, fixes the two .so files and depends on all the required ipks (binutils, gcc etc) Jun 26 00:20:23 DEPENDS="bucketloads of ipkgs" Jun 26 00:20:47 hmm, I don't think I have to manually make any symlinks Jun 26 00:20:55 just install the correct -symlinks ipkgs Jun 26 00:21:09 the .so files won't be a problem when we move to glibc 2.3.5 Jun 26 00:21:27 jacques: It's the symlink in /usr/include. asm-arm -> asm Jun 26 00:21:44 the kernel headers should be packaged by the same .bb that builds and packages the kernel Jun 26 00:21:57 NAiL, that symlink already exists in the kernel includes Jun 26 00:22:23 I usually do tar czvf kernel-includes.tgz asm asm-arm asm-generic linux Jun 26 00:22:28 jacques: Yeah, I thought about that, but how do I make that .bb make an ipkg that depends on "lotsastuff", without the kernel depending on the lot? Jun 26 00:22:38 jacques: Ah, yes. Jun 26 00:22:52 I'm not talking about all the deps right now Jun 26 00:23:09 just the kernel includes - should be part of the packaging phase of the kernel .bb Jun 26 00:23:19 yeah Jun 26 00:23:28 of course Jun 26 00:23:32 * NAiL slaps his forehead Jun 26 00:23:49 then you could make a really simple ipkg that just depends on stuff Jun 26 00:23:54 yeah Jun 26 00:24:05 a -task ipkg like debian has Jun 26 00:24:12 or rather task- Jun 26 00:24:18 A meta package. Jun 26 00:24:21 yeah Jun 26 00:24:47 Except it isn't like the current meta/*.bb because it has to generate an ipk Jun 26 00:26:05 However I believe this is trivial to do - in fact I think the empty .bb may do it (a dummy file is required to get force the package to be built). Jun 26 00:29:36 btw, it's still safe to remove the /initrd/boot/zImage, right? Jun 26 00:30:15 Yes, unless you use APEX Jun 26 00:30:23 Anyway, that's a symlink Jun 26 00:30:29 which I don't, since I don't have jtag Jun 26 00:30:39 yah, I know, but I didn't bother writing the version number :P Jun 26 00:30:47 :-) Jun 26 00:32:06 ok, my build is going to take a while, so I'm considering going to bed (!) Jun 26 00:32:44 better ask what questions you need of me Jun 26 00:32:51 before I head that way as well Jun 26 00:33:09 I'm not very fit to think of good questions now, unfortunately Jun 26 00:33:22 I've got FrankenSlug, his head, and the GB switch in my arms righ now Jun 26 00:33:46 well Unslung 5.5 cksum = 35485092 8388608 Jun 26 00:33:59 nice, updated wiki? Jun 26 00:34:24 heck, I *just* got away from the wife's iMac.... Jun 26 00:34:30 haha, ok Jun 26 00:34:35 http load did not work Jun 26 00:34:43 oh.. :( Jun 26 00:34:45 it crapped out for some unknown reason Jun 26 00:35:05 so I had to setup tftp server - which can be non-trivial for OS X users Jun 26 00:35:27 (they are going to be getting their hands dirty in *nix before they even get to play with unslung) Jun 26 00:35:35 I'd be a very happy camper if you'd provide the details of that later Jun 26 00:35:40 That's what happened to me too, (http crapped out). Jun 26 00:35:48 tftp is the way to go Jun 26 00:36:03 Right now I'm gonna snuggle up with maimedslug (almost) under my pillow Jun 26 00:36:14 in fact the load from tftp was so fast, that I was sorta concerned that it didn't take Jun 26 00:37:11 Good night/morning/evening(etc), people. Jun 26 00:37:12 but I would be not a single OS X user is going to be able to catch the redboot window unless they change the NIC from auto-negotiation Jun 26 00:37:24 err... would "bet" Jun 26 00:37:56 yeah, I know. Gotta note that on the wiki. I guess most (all?) macs have mdi-x Jun 26 00:38:04 ALL Jun 26 00:38:05 anyway, zzzz Jun 26 00:38:13 hehe, not surprised :P Jun 26 00:38:31 I was thinking of some curse words to use Jun 26 00:38:48 while trying to catch that @#$%(&@^%*&#$% redboot window from OS X Jun 26 00:39:12 and I don't curse... Jun 26 00:40:11 Thanks for the good work Jun 26 00:40:34 jbowler, did you attempt to connect from OS X? Jun 26 00:41:16 No - I don't even have a Mac, I was trying from RedBoot to a Linux machine (http/ftp/tftp). Jun 26 00:42:02 ah... it seems that iMacs do autonegotiation at start of any thread that accesses a net interface Jun 26 00:43:16 so with nslu2, you're trying to catch a 2.5 sec (at most) window, and it takes the iMac 3 secs (maybe longer) to do the autonegotiation Jun 26 00:44:15 for the record, I did catch the redboot window ONCE with autonegotiation set.... it said I had .14 secs before it booted Jun 26 00:44:54 I started the telnet way before I hit the power button on the slug Jun 26 01:19:45 You shouldn't plug it in directly if you want to catch redboot. Use a hub Jun 26 01:45:37 re Jun 26 02:24:18 Wow. ByronT is not just Away he's Away-Away! Jun 26 02:35:59 NAiL: we're missing the init file for cron Jun 26 02:36:16 can you check it in and push it? Jun 26 02:36:27 Does Allan still have the PCB? Jun 26 02:36:36 I think so Jun 26 02:37:02 Wondering if he ever got around to doing better scans of it or if he can post it Jun 26 02:37:20 Going blind tracing ETH1 on the scans he posted Jun 26 02:49:50 rwhitby-away: you told something about the database yesterday, should i repull the whole structure? Jun 26 02:51:11 rwhitby-away: or is there a new Makefile available? Jun 26 02:52:45 there is a new Makefile available, and you should start from it. Do not copy an old database. Jun 26 02:53:03 ok, then i will restart from scratch now Jun 26 02:58:12 rwhitby-away: Does the WL-500g have a plain switch or is it capable of VLAN tagging? Jun 26 02:58:40 good question Jun 26 02:58:48 not sure if it has the ADM chip or not Jun 26 02:59:20 hmm, monotone is hanging or very slow Jun 26 02:59:26 Wonder if anybody has taken one apart and got pics... Jun 26 02:59:30 ok, not hanging, only slow Jun 26 02:59:54 mr_claus: not monotone, your network connection (or that of the source, or somewhere between) Jun 26 02:59:58 Found pics Jun 26 03:00:11 Tiersten: wl500g.info ... Jun 26 03:00:53 Uses a Broadcom BCM5325 and it says it's got VLAN Jun 26 03:01:07 rwhitby-away: hmm, 3mbit down, 512kb up? one line was 1min there Jun 26 03:02:09 mr_claus: as I said "or that of the source, or somewhere between". you are on a different continent from at least one of the sources. Why do you blame the tool for network speed? Jun 26 03:02:52 Did you ever pull an openembedded repository from bk? what are you comparing against when you say it is slow? Jun 26 03:04:57 rwhitby-away: thats what i see, i dont know if it's hanging or not and slow is relative, if a line is there (the bytes which were displayed are not very much) for more then 1 min then it's slow, i dont see if it's working Jun 26 03:05:43 rwhitby-away: slow means, nothing happens on the screen, the transferred bytes are not very much and i dont see that anything will be done by the tool Jun 26 03:06:02 mr_claus: there is 20MB to transfer on the first pull. Jun 26 03:07:00 My bad. It uses a BCM5365 but it still does have VLAN support Jun 26 03:07:00 rwhitby-away: yes i know and if i see that the lines are going then i see that the tool is working, if nothing happens i dont know anything, that was the reason why i asked it's hanging or slow? Jun 26 03:07:40 mr_claus: *shrug* Jun 26 03:08:16 The first pull is quite slow because it has a lot to process. Subsequent ones should be much faster. Jun 26 03:08:25 At least that's what I've found Jun 26 03:09:08 mr_claus: you didn't ask. you said "monotone is hanging or very slow". without a question mark. you made a statement which was not true. monotone was not hanging, and it is not slow (relative to BK or any other CM system for the amount of data transferred). If you're asking a question instead of making a statement, please use the correct punctuation. Jun 26 03:10:11 rwhitby-away: ok, you're right, i forgot the question mark, i wanted to ask :) Jun 26 03:10:29 mr_claus: which step was it. if it's not doing anything for more than a minute then perhaps we can turn on the verbose option and fill your screen with warm and fuzzy useless information for that time :-) Jun 26 03:10:59 rwhitby-away: the lines i can scroll back are to less, so i dont know anymore Jun 26 03:11:12 rwhitby-away: it was one of the steps before files were be transferred Jun 26 03:11:37 ok. I put the --quiet flag on all setup steps, but could take them out again if that's the consensus. Jun 26 03:11:47 rwhitby-away: a lot of lines which look be same, but a byte counter goes up every line Jun 26 03:12:43 uh? Never had that problem Jun 26 03:13:01 rwhitby-away: next time if i do an initial pull i will increase my lines which i can scroll back, perhaps that was a one time thing Jun 26 03:15:51 mr_claus: are you doing this on a terminal which does not understand how to do a carriage return without also doing a line feed? Jun 26 03:16:38 that "ticker" ends up as multiple overwrites of a single line on a proper terminal Jun 26 03:16:53 see the --ticker option to monotone if you need to change it Jun 26 03:17:17 (it defaults to "count") Jun 26 03:17:47 * rwhitby-away wonders whether he can put the --quiet switch in .monotonerc somehow, like he does for CVS .... Jun 26 03:20:55 monotone: [bytes in: 10001389] [bytes out: 166526] [certs in: 165] [revs in: 45]monotone: [bytes in: 10005484] [bytes out: 166526] [certs in: 165] [revs in: 45] Jun 26 03:21:07 from that line new lines will be created Jun 26 03:21:28 until that line all goes into the same line Jun 26 03:21:45 monotone: [bytes in: 21866854] [bytes out: 166526] [certs in: 165] [revs in: 45] Jun 26 03:21:49 yep, your terminal should be making them all appear on a single line. Jun 26 03:21:55 at that point it needs a while to continue Jun 26 03:21:58 you need to fix your terminal emulation. Jun 26 03:22:17 (or change the ticker to "dots") Jun 26 03:22:21 the terminal does but only until the line i postet Jun 26 03:22:41 perhaps because the size of the number Jun 26 03:23:05 until 10000000 is goes into a single line, above it goes in new lines Jun 26 03:23:33 oh, weird. Jun 26 03:23:38 it's still at the line 21866854 Jun 26 03:24:10 i increased the columns now and will restart the process Jun 26 03:26:31 ok, increasing the columns to 81 is working, i see that i can switch off auto-wrap too, so i will decrease to 80 and switch auto-wrap off Jun 26 03:45:03 mr_claus: is the Makefile otherwise working ? I'm more interested in testing the functionality instead of the ticker output :-) Jun 26 03:45:59 BTW, I have removed the --quiet options from the Makefile. I will work out how to turn it off just for me :-) Jun 26 03:47:05 rwhitby-away: yes, it's working nice, yesterday i have done a complete build out of the box, but i put in "make openslug-build" and it made unslung and openslug Jun 26 03:47:29 rwhitby-away: now i will try again, the "make setup" works nice Jun 26 03:49:43 rwhitby-away: now the "make openslug-build" work as expected it seems :), it start with openslug, yesterday it started with unslung Jun 26 04:14:48 <[cc]smart> hi. Jun 26 04:14:55 g'day Jun 26 04:15:26 <[cc]smart> i just had my successfull monotone first contact. Jun 26 04:15:45 <[cc]smart> now i'm about to commit a few lines locally Jun 26 04:16:01 <[cc]smart> afaik i nned a key ofr that which i'm about to generate Jun 26 04:16:14 <[cc]smart> checking existing entries people use email adresses as key names Jun 26 04:16:37 <[cc]smart> are those on @nslu2-linux.org forwards that are registered somewhere ? Jun 26 04:16:57 <[cc]smart> if such was available, i'd ot have to put my email address on sth. accessible... Jun 26 04:20:31 [cc]smart: yes, the @nslu2-linux.org addresses are core team dev addresses. You will need to choose an email address that you own, and use that. Jun 26 04:21:01 Setup a gmail one or something? You can set those to forward to where ever you want or just read off the web Jun 26 04:23:02 [cc]smart: one thing we do ask is that your key id is somehow recognisably connected with your registered IRC nick or user name. Jun 26 04:24:55 pssst ... why does openslug kernel not include anything ipv6'ish? Jun 26 04:25:21 Fuzzel: pssst ... cause no-one has contributed it yet? Jun 26 04:25:55 <[cc]smart> rwhitby-away: so ccsmart would be acceptable as keyname ? Jun 26 04:26:01 as in ... if I have the diff it can go in ? :) Jun 26 04:26:19 [cc]smart: yep. ccsmart@somewhere would be fine Jun 26 04:26:47 <[cc]smart> where is actually "somewhere", aka unspecified... ? Jun 26 04:27:04 Fuzzel: I expect so, but you should talk to [g2] and jbowler about it. Jun 26 04:27:40 [cc]smart: no, we require a valid email address. Specifically so someone in #oe knows who to contact if you stuff something up :-) Jun 26 04:28:08 <[cc]smart> kk, will have to think about how i'll do this without using my real backend address Jun 26 04:28:17 yep, understood. Jun 26 04:28:28 <[cc]smart> out for WE :) Jun 26 04:28:59 rwhitby: okey... I'll first make it nicely work with most of the stuff on my slug Jun 26 04:29:10 [cc]smart: ccsmart@gmail taken already then? Jun 26 04:29:11 it pings and ssh's already though Jun 26 04:29:40 <[cc]smart> Tiersten: dunno, but i'll try to keep my set of addresses lean, too. Jun 26 04:30:29 [cc]smart: I use a domain, zoneedit.com and redirection.net. Jun 26 04:30:49 redirect the ones that get spammed to bounce@emample.com Jun 26 04:30:57 bounce@example.com Jun 26 04:56:32 8<-------- Jun 26 04:56:35 NOTE: package cron-3.0pl1: started Jun 26 04:56:35 NOTE: package cron-3.0pl1-r2: task do_fetch: started Jun 26 04:56:35 NOTE: package cron-3.0pl1-r2: task do_fetch: completed Jun 26 04:56:35 NOTE: package cron-3.0pl1-r2: task do_unpack: started Jun 26 04:56:35 NOTE: Unpacking /mnt.vevey/downloads/cron3.0pl1.tar.gz to /mnt.vevey/openslug/tmp/work/cron-3.0pl1-r2/ Jun 26 04:56:38 NOTE: Unpacking /mnt.vevey/openembedded/packages/cron/cron-3.0pl1/init to /mnt.vevey/openslug/tmp/work/cron-3.0pl1-r2/ Jun 26 04:56:41 cp: cannot stat `/mnt.vevey/openembedded/packages/cron/cron-3.0pl1/init': No such file or directory Jun 26 04:56:44 NOTE: Task failed: Jun 26 04:56:47 NOTE: package cron-3.0pl1-r2: task do_unpack: failed Jun 26 04:56:50 -------->8 Jun 26 04:56:52 hmmm Jun 26 04:57:03 either mtn misses something or I do ;) Jun 26 05:34:00 no, NAiL didn't push the init file with his cron change (or it got lost somewhere) Jun 26 05:38:08 hm Jun 26 05:38:26 NAiL: not blaming you - quite likely it got lost Jun 26 05:38:33 hehe :P Jun 26 05:38:41 if you can put it back and push it would be very much appreciated :-) Jun 26 05:39:51 My nslu2-world build croaked on cron-3.0pl1-r2 . Jun 26 05:40:01 yep, same problem Jun 26 05:40:08 missing init file Jun 26 05:40:23 Oh, I see that right above in the log. Jun 26 05:40:27 doh! Jun 26 05:41:32 That was the sleepdeprived me :P Jun 26 05:42:39 now there's two heads Jun 26 05:42:49 atleast on my side Jun 26 05:43:07 NAiL: do you have that cron init file? Jun 26 05:43:14 it's pushed Jun 26 05:44:33 now, how do I add a symlink to oe-symlinks/packages? The build will fail on that later for you other people Jun 26 05:44:35 did you sync it back to mtn.nslu2-linux.org/org.openembedded.nslu2-linux ? Jun 26 05:45:15 NAiL: which package? Jun 26 05:45:23 openvpn Jun 26 05:46:26 what about openembedded/packages/cron/cron-3.0pl1/init ? Jun 26 05:46:48 I must not be licking it right. Jun 26 05:46:58 dyoung-zzzz: I'm waiting for work from NAiL as to whether it has been synced back to our public repo .... Jun 26 05:47:01 dyoung-zzzz: that one is pushed... but where it went, nobody knows Jun 26 05:47:12 rwhitby-away: there's two heads... Jun 26 05:47:22 NAiL: monotone merge Jun 26 05:48:27 hmm.. try now... when I synced now it sent something Jun 26 05:49:12 <-- woke up and saw the screen. Not *quite* awake yet Jun 26 05:49:32 NAiL: I ask again - exactly where did you push it to. Jun 26 05:50:21 I didn't specify. I assumed mt figured out that part... Jun 26 05:50:29 now openvpn fails :) Jun 26 05:51:09 Fuzzel: cd oe-links/packages && ln -s ../../openembedded/packages/openvpn Jun 26 05:51:21 ah thx Jun 26 05:51:41 oe-symlinks you mean I assume Jun 26 05:51:54 yeah, that'd be right ;) Jun 26 05:52:06 seems to work Jun 26 05:52:34 fortunately this box is well suited for speed-compiling ;) Jun 26 05:53:21 hmm lzo Jun 26 05:53:31 duh Jun 26 05:53:35 same problem Jun 26 05:54:15 * Fuzzel also needs to wake up indeed Jun 26 05:54:52 I happen to know that because it's my fault. You'd have to read logs to figure it out :P Jun 26 05:55:33 it seems to work Jun 26 05:55:43 the old bitkeeper checkout did go in one go Jun 26 05:56:03 btw... samba4 will also be very interresting for nslu because of the low memory footprint Jun 26 05:56:26 Fuzzel: you're working with the *bleeding* edge new repo here. Jun 26 05:56:28 I'll try to get that and ipv6 going during the week Jun 26 05:56:38 rwhitby: that is not a bad thing :) Jun 26 05:56:53 I'm running out of band-aids. Jun 26 05:57:03 03rwhitby * r54 10/trunk/openslug/nslu2-linux/packages/openvpn: Added openvpn Jun 26 05:57:03 dyoung-zzzz: haha. What's up now? Jun 26 05:57:18 working with the bleeding edge. Jun 26 05:57:38 is there a equivilent of "bk parent" ? Jun 26 05:58:12 rwhitby: actually if it is bleeding edge then it should have samba4 soonish, let's see if I can make that work :) Jun 26 05:58:18 NAiL: what is the current head of your org.openembedded.nslu2-linux branch? Jun 26 05:58:43 Fuzzel: the repo structure is bleeding edge, we're right in the middle of converting from bk to monotone Jun 26 05:58:53 and you're working with the in-transition repo Jun 26 05:59:06 rwhitby: well you need testers for that too I guess :) Jun 26 05:59:20 and everything is way better than that bloody silly bitkeeper crap, that is the license of it Jun 26 05:59:22 Fuzzel: appreciate you testing it. just letting you know the status. Jun 26 05:59:26 rwhitby-away: is that in ./ or in ./openembedded? Jun 26 05:59:36 NAiL: in openembedded Jun 26 06:00:04 monotone heads -b org.openembedded.nslu2-linux Jun 26 06:00:21 rwhitby: I already love the whole openslug/unslung stuff from the second I saw how many people where working on it :) Jun 26 06:00:24 monotone: branch 'org.openembedded.nslu2-linux' is currently merged: Jun 26 06:00:25 6ae9841c3fedc501b700c85f5a0aba57c0fad556 oyvind@repvik.org 2005-06-26T04:34:50 Jun 26 06:00:55 thats the same for me too. Jun 26 06:01:01 and my joeblow is at the same head. But it's missing the init file Jun 26 06:01:10 did joeblow pull? Jun 26 06:01:13 NAiL: monotone list unknown Jun 26 06:01:31 packages/cron/cron-3.0pl1/init Jun 26 06:02:05 * NAiL doesn't get it Jun 26 06:02:35 NAiL: you forgot to monotone add it Jun 26 06:02:59 list unknown is your friend :-) Jun 26 06:03:04 I thought I did that before committing... Jun 26 06:03:16 now pushed Jun 26 06:03:25 monotone: branch 'org.openembedded.nslu2-linux' is currently merged: Jun 26 06:03:25 2965086892d73ebee6bfdf2329f4b2abcab6ef4f oyvind@repvik.org 2005-06-26T13:01:51 Jun 26 06:04:07 joeblow synced and at same rev Jun 26 06:04:07 I broked it. Jun 26 06:04:18 dyoung-zzzz: Whatcha break? Jun 26 06:04:21 NAiL; that did it Jun 26 06:04:33 monotone list unknown Jun 26 06:04:34 monotone: fatal: std::exception: boost::filesystem::is_directory: "/home/dereky/nslu2-world/openslug/tmp/work/bluez-libs-2.16-r0/bluez-libs-2.16/include/bluetooth/bluetooth/bluetooth [...] Jun 26 06:04:45 monotone: this is almost certainly a bug in monotone. Jun 26 06:05:20 dyoung-zzzz: is that a recursive symlink? Jun 26 06:05:28 rwhitby: Ok, it was me sleepwalking. I pushed to the wrong repo Jun 26 06:05:43 ~lart NAiL Jun 26 06:05:43 * jbot chops NAiL in half with a free AOL CD Jun 26 06:05:47 ouch Jun 26 06:06:26 Those AOL CD's are sharp ... Jun 26 06:06:31 haha :P Jun 26 06:06:57 anyway, there's now a cron init file in org.nslu2-linux.dev Jun 26 06:08:06 not anymore though :P Jun 26 06:09:12 http://ewi546.ewi.utwente.nl/tmp/viewmtn/revision.psp?id=2965086892d73ebee6bfdf2329f4b2abcab6ef4f Jun 26 06:09:45 what are you doing pushing to org.nslu2-linux.dev ? Jun 26 06:10:03 apparently sleeping Jun 26 06:10:32 the Makefile sets up the default servers and collections specifically. Do *not* override them unless you are awake. Jun 26 06:10:42 rwhitby: uh... I haven't Jun 26 06:11:01 well how did an openembedded thing get into org.nslu2-linux.dev ? Jun 26 06:11:29 rwhitby: the init script is simple. I was in ./ when I did mt add;commit;sync Jun 26 06:11:41 *all* other changes I've done was in ./openembedded Jun 26 06:12:00 what was the init script doing in the top-level directory? Jun 26 06:12:05 (which syncs to org.openembedded.nslu2-linux from what I can see) Jun 26 06:12:42 rwhitby: uh.. mt add openembedded/packages/cron/cron/init Jun 26 06:12:55 instead of cd openembedded; add packages/cron/cron/init Jun 26 06:13:13 oh, ok. hmm - that's an easy thing to do. we need to prevent that. Jun 26 06:13:31 Good idea. I think I'll remember that one from now on though :P Jun 26 06:15:04 http://ewi546.ewi.utwente.nl/tmp/viewmtn/getfile.py?id=9e8c63bdb9b708fb7bcddc1988c30f2433a698b1&path=conf/machine/nslu2.conf Jun 26 06:15:07 TARGET_CC_ARCH = "-march=armv5te -mtune=xscale" Jun 26 06:15:08 PACKAGE_ARCH = "armeb" Jun 26 06:15:15 why is the first one amrv5te ? Jun 26 06:15:21 which is Little Endian afaik Jun 26 06:16:09 dunno, but it works - I was manually compiling dhrystone code with it last night Jun 26 06:16:12 Fuzzel: you'll need to ask jbowler that one. Jun 26 06:16:28 either it gets ignored or overwritten Jun 26 06:16:33 otherwise it would not work I guess ;) Jun 26 06:16:35 I expect the xscale tune flips the endianness Jun 26 06:16:43 could well be it indeed Jun 26 06:16:51 xscale does not imply endian Jun 26 06:17:09 try -march=i386 -mtune=xscale, and see if it still builds :P Jun 26 06:17:11 and if it did, it certainly wouldn't be BE Jun 26 06:18:05 doesnt armv5te describe the instruction set? Jun 26 06:18:27 not the particulars of the endian-ness? Jun 26 06:18:33 for this type of device, doing a lot of networking stuff, big endian is better anyway :) Jun 26 06:18:54 -march=name Jun 26 06:18:54 This specifies the name of the target ARM architecture. GCC uses this name to determine what kind of instructions it can emit when gener- Jun 26 06:18:57 ating assembly code. This option can be used in conjunction with or instead of the -mcpu= option. Permissible names are: armv2, armv2a, Jun 26 06:19:00 dyoung-zzzz, that wouldn't surprise me Jun 26 06:19:01 armv3, armv3m, armv4, armv4t, armv5, armv5t, armv5te. Jun 26 06:19:24 there you have it. it determins the instruction set. Jun 26 06:19:38 yeah, none of those are endian-specific Jun 26 06:19:44 Fuzzel: build not freaking out on anything else yet? Jun 26 06:19:55 nail: not yet Jun 26 06:20:04 it's fetching samba now Jun 26 06:20:10 my head hurts. Jun 26 06:20:14 time for sleep. Jun 26 06:20:19 when I have all those packages cached then all would be much faster to rebuild Jun 26 06:20:32 network io (samba.org only doing 100kb/s) is the bottle Jun 26 06:20:39 that samba build will be effectively broken. It'll build, but it's got a rather crippling bug :P Jun 26 06:20:55 read about it Jun 26 06:21:02 but I want to do samba4 anyway :) Jun 26 06:21:21 Where's the info on footprint? Jun 26 06:21:32 I know one of the devel guys Jun 26 06:21:37 he said something about 12k per client Jun 26 06:21:57 23k now as far as I can see Jun 26 06:22:53 ooh... sounds rather nice Jun 26 06:23:05 uhuh.. and full AD PDC support etc Jun 26 06:23:12 internal ldap stuff Jun 26 06:23:15 lots of cool stuff Jun 26 06:23:26 but not all features from v3 are in there yet, though it will remain small Jun 26 06:24:05 Sounds interesting Jun 26 06:24:09 jups Jun 26 06:24:21 Hmm - monotone should not allow you to add a file that is inside a directory which is in .mt-ignore Jun 26 06:25:19 Ok. /me goes back to sleep. If anything else breaks, holler, and I'll probably wake up again :P Jun 26 06:25:28 hehe :) Jun 26 06:25:30 nighters Jun 26 06:26:09 I wonder why I woke up exactly 5 mins after rwhitby mentioned me, "cron" and "init" :P Jun 26 06:27:44 rwhitby: Powerbutton in openslug, does it do insta-off, gracefully shut down or do nothing? Jun 26 06:28:04 mine does insta-off, but I don't know if that's because of the udev changes Jun 26 06:28:05 I would guess nothing Jun 26 06:28:18 since AFAIK nothing's listening for it Jun 26 06:28:29 but, that's just a guess Jun 26 06:28:39 hmm mine shutted down nicely Jun 26 06:28:52 Fuzzel: huh? What firmware? Jun 26 06:28:59 ah no ignore me :) Jun 26 06:29:04 did shutdown -h now as usual Jun 26 06:29:07 not the powerbutton Jun 26 06:29:09 heh Jun 26 06:29:10 aah :P Jun 26 06:29:26 Fuzzel, shutdown -h works now, but reboot doesn't Jun 26 06:29:34 when you are booted to a disk Jun 26 06:29:59 but a "powerbuttond" is basically a util that opens the filedescriptor and polls it once in a while to see if it's "1"? Jun 26 06:30:20 hmm.. no Jun 26 06:30:20 normally apm stuff implements it Jun 26 06:30:27 at least for normal pc things Jun 26 06:30:38 you could check the binary from linksys Jun 26 06:30:41 yeah, but the slug is kinda.. differently abled :P Jun 26 06:31:37 no need to turn it off Jun 26 06:31:48 the linksys firmware runs a daemon that listens for the button Jun 26 06:31:52 you've got a point there :P Jun 26 06:31:53 already soldered it to always be on Jun 26 06:32:09 though the powerbutton still works to turn it off if wanted Jun 26 06:32:56 eeeh Jun 26 06:33:00 cool Jun 26 06:33:11 the buzzer can do different frequencies ;) Jun 26 06:33:23 oh god... can you make it sing? :) Jun 26 06:33:41 lol, don't tell jbowler-zzz Jun 26 06:33:48 or sing the national anthem every morning Jun 26 06:33:50 * NAiL did "cat /dev/rbuttons" and pressed the reset button. The more times I pressed it, the further down the frequency sank ;) Jun 26 06:33:56 hmmm NSLU2-AlarmClock? :) Jun 26 06:34:09 Fuzzel: Bright idea! ;) Jun 26 06:34:27 too bad the powersupply is a bit chunky Jun 26 06:34:30 (but nothing showed up in /dev/rbuttons though) Jun 26 06:34:53 I guess I'll fetch another slug as a travel mate... file/stream/*-server everywhere you go Jun 26 06:35:51 Maybe I'll implement streaming in AeroSlug... It'd be really nice for the plane to stream music to people it flys over (Via bluetooth!) Jun 26 06:37:58 mpd + usb speakers ;) Jun 26 06:38:26 Doubt we'd hear much over the engine whine Jun 26 06:39:53 in a plane not indeed..... and if you can install capable speakers that size then you can also install a bigbox(tm) Jun 26 06:40:11 hehe Jun 26 06:41:16 Fuzzel: Which auto power on method did you use? Jun 26 06:41:35 the solder the three/four wires trick + the chip Jun 26 06:41:38 Weight is an issue. The two heaviest part so far is the batteries (~120g) and the slug itself, which I don't know how much weighs Jun 26 06:41:54