**** BEGIN LOGGING AT Tue Jul 26 23:59:57 2005 Jul 27 01:01:42 is the noirqdebug bootoption something I want? :) Jul 27 01:27:31 kolla_: I run meld over 1M/256k cable :) Jul 27 01:27:38 that's painfull too Jul 27 01:33:58 koen, whats the policy for sources that provide semi-broken tar files? ctrlproxy is available from a central source, but the tar.gz file they provide has a non-standard gz trailer for some reason causing tar -z to think it failed to decompress it even though it worked. Jul 27 01:34:35 I figure the choices are re tar it myself and put it someplace, or do something ugly in the bb to make it ignore the error. Jul 27 01:35:37 we used to go for the .bb route, but since we have to mirror the sources anyway we sare starting to host stuff like that ourselves Jul 27 01:36:42 Okay, which is preferred? Jul 27 01:37:26 hosting a repaired source Jul 27 01:38:00 Okay, and would you prefer to have that hosted at oesources.org? or.. ? Jul 27 01:39:00 It's all the same to me Jul 27 01:39:38 Okay cool. I'll mail you a link later in case you want to double mirror it at oesources.org . :-) Jul 27 01:39:58 koen: hehe, ok :) Jul 27 01:40:14 thanks koen . Jul 27 01:40:16 greg@treke.net is the address for that (to cut down on forwarding :)) Jul 27 01:40:17 but now I'm at work and can enjoy the "54Mbit" wireless Jul 27 01:40:48 so.. noirqdebug.. I think I'll bring it along Jul 27 01:43:09 * koen unpacks the old 10/100M switch Jul 27 01:43:17 * koen connect it to the USB200M Jul 27 01:43:38 * koen ifup eth2, and pings slug.kabel.utwente.nl Jul 27 01:43:52 and another slug hits the net Jul 27 01:44:10 cool. Jul 27 01:44:20 I dont have one on the open net yet. Jul 27 01:44:34 I was waiting for my fat one to do that. Jul 27 01:45:00 I have to sneak into the lab to get to a good soldering station Jul 27 01:50:18 * kolla_ fights with kdiff3 Jul 27 01:51:02 darn thing says that there's one more conflict, it keeps bringing up the same line as a conflict even though I tell it to use option C Jul 27 01:52:24 when I use meld I have to merge to left or center for it to work Jul 27 01:52:31 does kdiff3 suffer from the same problem? Jul 27 01:53:13 seems so Jul 27 01:53:42 I want the right option since it's the right for me :) Jul 27 01:54:45 nah, same bug when I select the middle option Jul 27 01:55:54 why are there 3 kernel configs anyways? Jul 27 02:05:44 the ancester it picks has multiple commits to the file Jul 27 02:06:16 try looking at the graph and pick a sane ancestor using monotone explicit_mergse r1 r2 ancestor Jul 27 02:27:58 will the slug read/write ntfs? Jul 27 02:31:52 are there kerneldrivers that write to ntfs? Jul 27 02:32:15 last time I only mounted ntfs rw it blew up Jul 27 02:32:40 ntfs writes are not supported in Linux last time I checked. Jul 27 02:33:56 last time I tried it for S/G it rendered my test NTFS (winxp) partition totally unusable. Jul 27 02:34:15 that was > 1.5 years tho Jul 27 02:38:25 ok. Jul 27 02:38:30 might stick with fat32 then ;) Jul 27 02:39:48 ntfs is still unusable on linux. It's "ok" on x86 with the driver that uses the windows dll Jul 27 02:40:00 i had a memstick mounted as rootfs... but for some reason it reverted back to booting the ramfs. how do i reinstate an existing root filesystem? Jul 27 02:40:04 but on arm, you might be able to read :P Jul 27 02:40:08 i've tried modifying ever file I could think of. Jul 27 02:40:19 nail: ok.. Jul 27 02:40:25 NAiL: arm*EB* :P Jul 27 02:40:56 koen: any diff between be and le? Jul 27 02:41:43 NAiL: probably Jul 27 02:43:55 anyone hacked the WRV54G aswell? Jul 27 02:44:03 slug and snail? Jul 27 02:48:10 I do have a WRT54GS. Can't remember what the V did Jul 27 02:49:40 VPN Jul 27 02:50:03 http://www.linuxjournal.com/article/7941 Jul 27 02:50:22 c'mon nail, I was waiting for your stellar commit rate! Jul 27 02:50:32 Need to test that mirror stuff! Jul 27 02:50:43 ;-) Jul 27 02:58:32 heh Jul 27 02:58:37 Don't have anything to commit Jul 27 02:58:44 * dyoung gasp Jul 27 04:12:11 so the WRT54GS doesn't have a usb port at all? Jul 27 04:12:30 i was looking at the links from openwrt.org they all point to the slug wiki Jul 27 04:13:18 it doesn't have usb port Jul 27 04:14:15 i had a memstick mounted as rootfs... but for some reason it reverted back to booting the ramfs. how do i reinstate an existing memstick root filesystem? Jul 27 04:14:26 i tried turnup memstick -i /dev/sda1 Jul 27 04:14:39 try without the -i Jul 27 04:14:41 it then asked to force with -f which I did, but it still didn't boot it. Jul 27 04:14:46 ok. Jul 27 04:21:47 I had a WRT54G, then a WRT54GS, got rid of them both, and got a WL500g, then replaced that with a WL500gx. Jul 27 04:22:49 is the wl500gx better as the wrts? Jul 27 04:23:48 I think it is. It's got USB, and it runs Optware packages :-) Jul 27 04:24:14 sweet Jul 27 04:30:21 and it has room to mount a usb->serial PCB inside and usb type-b socket on the back, to give you a usb client serial console on it. Jul 27 05:54:43 I just stumbled across http://www.myfmstation.com/html/linex_usb.html, using a nslu2 as a web-stream2FM-radio adapter would be very nifty :) Jul 27 06:10:57 DaKa2: how's the cross-built perl going? Jul 27 06:11:06 well, think its working Jul 27 06:11:30 if I build openslug now I'll get it? Jul 27 06:11:40 the only thing left is packaging Jul 27 06:11:40 yes Jul 27 06:12:28 there needs to be talks with OE about the granularity.. Jul 27 06:13:40 how many does it emit? Jul 27 06:13:43 858 Jul 27 06:14:13 what's our preferred outcome? Jul 27 06:14:49 I dont know.. I like libperl, perl, perl-modules, but I understand thats not good for lots of OE uses Jul 27 06:15:27 DaKa2: how big are those parts? Jul 27 06:15:58 koen: http://ipkg.nslu2-linux.org/feeds/openslug/native/2.4-beta/ Jul 27 06:16:27 I dont know about unpacked Jul 27 06:17:11 damn power outage! :( Jul 27 06:17:31 and my slug is not modified yet for ForcePowerAlwaysOn Jul 27 06:30:49 Have we determined whether udev will work for OpenSlug yet? Jul 27 06:31:14 I heard someone was using udev 0.63 Jul 27 06:31:37 0.63-r0 did kill my hx4700, though Jul 27 06:31:37 I was just about to add /dev/video* to openslug ... Jul 27 06:31:43 udev would be nice, since devfs is actually dropped Jul 27 06:40:35 mapping hotplug Jul 27 06:40:38 script grep Jul 27 06:40:41 map eth2 Jul 27 06:40:48 and eth2 comes up when I plug in the usb stick Jul 27 06:40:56 dyoung-zzzz: looks like pwc have accepted the endian-fix upstream Jul 27 07:11:32 rwhitby, udev works fine for me. Jul 27 07:11:48 The only proble is with hotplug-ng not registering cold-plug events on bootup Jul 27 07:12:09 well, that seems like a bit of a show-stopper ... Jul 27 07:12:20 kind of. Jul 27 07:12:42 can it be fixed? Jul 27 07:12:46 Then again, It could be so I suggest you add udev to your package list and give it a try with your h/w Jul 27 07:13:19 probably. Is there an updated hotplug-ng version? Jul 27 07:13:27 Or is this project orphaned already? Jul 27 07:15:12 anyone know were our current openslug device_table comes from in the repo? Jul 27 07:17:12 rwhitby, a quick google reports that there's a newer rev 002 of hotplug-ng that could replace our aging 001. Jul 27 07:18:24 Actually, the author says it's already obsolete: http://lkml.org/lkml/2005/5/6/101 Jul 27 07:22:44 It sounds like the newer kernels and module-init-tools make it all better. Not too sure though. Jul 27 07:28:14 ok, /dev contents comes from /etc/device_table on boot Jul 27 08:50:21 jbowler-away, I was told the oe-symlinks folder was obsolete. How should I configure my BBFILES in my local.conf then? Jul 27 09:01:22 VoodooZ_Work: if you don't want to use the default (i.e. for private packages) make conf/openslug-packages.conf and set it in there. Jul 27 09:02:52 Or if you want to track the default, make conf/openslug-bbfiles.conf instead and do BBFILES += in there. Jul 27 09:03:23 Note that 'bb freeze' will overwrite both these files if you build it... Jul 27 09:57:29 jbowler-away, thanks. I'll try it. Jul 27 10:05:40 jbowler-away, when you say make conf/openslug-packages.conf do you mean 'make' literaly. And are you talking about conf/distro/openslug-packages or a new file? Jul 27 10:05:55 sorry for the ignorance but i'm not clear. Jul 27 10:12:55 jbowler-away, Thanks. I figured it out. Jul 27 10:14:42 I simply created a openslug-packages.conf file in my local conf/ and it worked fine. Jul 27 10:16:53 actually, now only my package works and nothing else! I did do BBFILES. I guess it depends on the order. Jul 27 10:48:31 hey [g2]! Jul 27 10:48:49 <[g2]> hey VoodooZ_Work Jul 27 10:49:17 so what's the scoop on that new oe-symlink stuff? what's built by the Makefile and what's not? Jul 27 10:49:44 <[g2]> new as in last couple days ? Jul 27 10:50:10 if I add a conf/openslug-packages file with BBFILES += will it append it to the default list? Jul 27 10:50:12 yes. Jul 27 10:50:42 ANd why do I still have a oe-symlinks folder when I deleted it earlier (rwhitby told me it was obsolete) Jul 27 10:50:43 <[g2]> dunno I've been playing with postfix, mysql and web server CSS Jul 27 10:51:11 there was no announcements and I can't find a wiki page on it. Jul 27 10:51:12 <[g2]> I know jbowler-away was going to make some changes but I don't know how it works yet Jul 27 10:51:22 ok. thanks. Jul 27 10:51:36 hopefully I'll be around when he does explain it to you. Jul 27 10:52:03 <[g2]> I'd imagine once the kinks are worked out it'll just start working or there will be an announcement Jul 27 10:52:11 I see. Jul 27 10:56:13 * koen just realizes his slug in on the net with the default password Jul 27 10:56:19 * koen fixes Jul 27 10:58:55 too late! *snicker* Jul 27 11:00:32 <[g2]> koen, IPv6 right ? It takes a while to touch all those addrs :) Jul 27 11:00:38 :) Jul 27 12:13:17 <[g2]> [cc]smart, hey.... Can I bug you with at postfix config question ? Jul 27 12:22:04 <[cc]smart> [g2]: try, maybe i even know an answer Jul 27 12:26:54 <[g2]> [cc]smart, I don't think my local delivery is setup quite right Jul 27 12:27:15 <[g2]> when I try sendmail from the command line I get Jul 27 12:27:52 <[g2]> postdrop: warning: unable to look up public/pickup: No such file or directory Jul 27 12:28:52 <[g2]> postfix check doesn't return any errors and I don't see anything in /var/log/ Jul 27 12:39:43 <[cc]smart> postconf | grep mail_spool_directory Jul 27 12:40:39 <[g2]> postconf | grep mail_spool_directory Jul 27 12:40:39 <[g2]> mail_spool_directory = /var/mail Jul 27 12:40:58 <[g2]> heh.. no such directory Jul 27 12:45:36 <[g2]> hmmm... same thing when changed to /var/spool/mail which is there Jul 27 12:46:01 try stracing it for other missing patchs Jul 27 12:46:05 paths* Jul 27 12:46:25 <[g2]> funny you should mention that :) Jul 27 12:46:37 strace still busted? Jul 27 12:46:57 <[g2]> I think strace got better, but was removed from the packages Jul 27 12:47:03 <[g2]> lsof is broke Jul 27 12:47:16 <[g2]> or just doesn't work with all the things installed Jul 27 12:48:11 <[g2]> hmm.. looks like strace should be there Jul 27 12:49:41 <[cc]smart> restart postfix after the change Jul 27 12:49:50 <[g2]> I did Jul 27 12:50:17 dunno why strace was removed... It worked the last time I checked ;) Jul 27 12:51:05 <[cc]smart> "public" is your receiver address ? Jul 27 12:51:46 <[g2]> dunno I don't think it should be Jul 27 12:53:02 <[cc]smart> postconf | grep alias_maps Jul 27 12:53:41 <[g2]> postconf | grep alias_maps Jul 27 12:53:41 <[g2]> alias_maps = hash:/etc/postfix/aliases Jul 27 12:53:41 <[g2]> local_recipient_maps = unix:passwd.byname $alias_maps Jul 27 12:53:41 <[g2]> proxy_read_maps = $local_recipient_maps $mydestination $virtual_alias_maps $virtual_alias_domains $virtual_mailbox_maps $virtual_mailbox_domains $relay_recipient_maps $relay_domains $canonical_maps $sender_canonical_maps $recipient_canonical_maps $relocated_maps $transport_maps $mynetworks Jul 27 12:53:42 <[g2]> virtual_alias_domains = $virtual_alias_maps Jul 27 12:53:46 <[g2]> virtual_alias_maps = $virtual_maps Jul 27 12:54:03 <[cc]smart> ls -la /etc/postfix/aliases Jul 27 12:54:28 <[g2]> ls -la /etc/postfix/aliases Jul 27 12:54:28 <[g2]> -rw-r--r-- 1 root root 7100 Jul 22 13:16 /etc/postfix/aliases Jul 27 12:59:10 <[cc]smart> ps -ef | grep pickup Jul 27 12:59:48 <[g2]> none of the postfix daemons are there Jul 27 13:00:37 <[cc]smart> hm... you need pickup running Jul 27 13:01:03 <[g2]> when I do a postfix start is says started but I don't see any tasks Jul 27 13:02:35 <[cc]smart> something with postfix -v start ? Jul 27 13:02:36 isn't iptables in any feed? Jul 27 13:04:22 <[cc]smart> is pickup installed at all ? find /usr -name pickup Jul 27 13:05:09 <[g2]> find /usr -name pickup Jul 27 13:05:09 <[g2]> /usr/libexec/postfix/pickup Jul 27 13:07:03 <[cc]smart> and does postfix -v start give more info ? Jul 27 13:07:13 <[cc]smart> or -vv Jul 27 13:07:51 <[g2]> -vv does :) Jul 27 13:08:09 <[cc]smart> sth. clueful ? Jul 27 13:17:31 <[g2]> that's going to take some sorting through Jul 27 13:17:53 <[cc]smart> sounds good :) Jul 27 13:18:01 <[cc]smart> will take a nap then Jul 27 13:18:06 <[cc]smart> n8 all Jul 27 13:18:09 <[g2]> thx Jul 27 13:18:10 <[g2]> cheers Jul 27 13:24:08 <[g2]> senneth, that's pkgconfig not pkg-config right ? Jul 27 13:48:00 hi all, i was wondering if anyone could tell me if there is a package for lighttpd available for openslug? Jul 27 13:48:38 i'm asking because if one does not exist, i will attempt to make one. i just want to see if anyone is working on one first. Jul 27 13:57:03 you could check if a .bb file exists Jul 27 13:57:15 if not, then we would welcome one Jul 27 13:57:46 mickeyl, how's the life in slug world these days? Jul 27 13:58:54 ByronT-Desktop: pretty good. I'm waiting for a bit more time to make it do more things, i.e. i really would like to attach an usb sound card an an usb vga card Jul 27 13:59:11 i downloaded the 2.3 beta and didn't see one Jul 27 13:59:35 MrLebowski: what's in the feed is not necessarily all that's there Jul 27 13:59:37 check http://monotone.vanille.de/viewmtn/manifest.psp?id=c0c44670a7289bfe19eecd324760261e100769b9& Jul 27 13:59:52 if you don't see anything with lighttpd there, then it's not in :) Jul 27 14:01:09 mickeyl, I seem to recall a conversation where some ppl were looking to get this device to work - http://www.griffintechnology.com/products/imic/ Jul 27 14:01:32 ok, thank you. that is waht i was looking for.. and i don't see it there. Jul 27 14:05:03 ByronT-Desktop: that looks interesting. Jul 27 14:11:01 * dyoung hides Jul 27 14:12:54 hehe Jul 27 14:13:14 dyoung is always in the middle of the *interesting* conversations Jul 27 14:13:55 hehe Jul 27 15:43:07 DaKa2: ping Jul 27 15:43:17 there's a problem with the update-alternative Jul 27 15:43:29 i think there are ./ missing all over the place Jul 27 15:43:49 otherwise the links don't work Jul 27 15:43:58 i.e. /bin/pidof -> pidof.sysvinit Jul 27 15:44:00 doesn't work Jul 27 15:44:10 but /bin/pidof -> ./pidof.sysvinit does Jul 27 15:44:19 how come? Jul 27 15:44:28 what? that cannot be right Jul 27 15:44:35 no idea _why_ Jul 27 15:44:39 it's so, here Jul 27 15:44:41 though Jul 27 15:44:41 :) Jul 27 15:44:57 because /bin/pidof -> pidof.sysvinit is the right way.. Jul 27 15:45:11 thats the way its supposed to work... Jul 27 15:45:16 interesting Jul 27 15:45:36 and it does work on openslug.. Jul 27 15:46:11 do you have . in your PATH? Jul 27 15:46:25 shouldn't matter Jul 27 15:46:31 hope not.. Jul 27 15:46:42 PATH should not affect symlink resolution Jul 27 15:47:02 exactly, and: PATH='/usr/local/bin:/usr/bin:/bin:/usr/local/sbin:/usr/sbin:/sbin' Jul 27 15:47:12 perhaps it's aproblem of it needing to resolve twice Jul 27 15:47:24 since pidof.sysvinit is yet another symlink that doesn't seem to work Jul 27 15:47:28 the original busybox links have no ./ Jul 27 15:48:08 interestingly even the tab completion doesn't seem to think that pidof.sysvinit is there Jul 27 15:48:18 ok, so perhaps the problem is pidof.sysvinit, not pidof Jul 27 15:48:42 that seems more likley.. Jul 27 15:48:58 lrwxrwxrwx 1 root root 16 Jul 27 21:42 pidof.sysvinit -> ../sbin/killall5 Jul 27 15:48:58 root@spitz:/bin# ls -l ../sbin/killall5 Jul 27 15:48:58 ls: ../sbin/killall5: No such file or directory Jul 27 15:49:07 although Jul 27 15:49:09 root@spitz:/bin# which killall5 Jul 27 15:49:12 | /sbin/killall5 Jul 27 15:49:17 * mickeyl puzzled Jul 27 15:49:25 ../sbin doesn't work Jul 27 15:49:26 killall5? what? Jul 27 15:49:31 bytsaving busybox bastards Jul 27 15:49:40 +e somewhere Jul 27 15:49:43 +e ? Jul 27 15:49:48 ah Jul 27 15:50:03 so when i link pidof to /sbin/killall5 it works Jul 27 15:50:15 it doesn't work when i link it to ../sbin/killall5 Jul 27 15:50:29 aah Jul 27 15:50:51 root@spitz:/bin# mount Jul 27 15:50:51 proc on /proc type proc (rw) Jul 27 15:50:51 ramfs on /dev type ramfs (rw) Jul 27 15:51:05 ~lart / as command Jul 27 15:51:19 it is the pivot_root Jul 27 15:51:28 we have /dev/mtdblock2 on / Jul 27 15:51:35 and mounted /dev/hda on / above Jul 27 15:51:58 somehow the system gets confused when using ../ Jul 27 15:52:03 perhaps it resolves the wrong / Jul 27 15:52:37 but, uhm... the link from pidof.sysvinit is not created by u-a, thats in the original package Jul 27 15:52:44 yea Jul 27 15:53:08 the original package tries to be relocatable Jul 27 15:53:20 that's why they use ../sbin instead of /sbin/ Jul 27 15:53:27 it's not an u-a problem, i take that back Jul 27 15:53:31 :) Jul 27 15:53:32 false alarm Jul 27 15:53:42 i need to figure out how to fix that anyway Jul 27 15:53:45 that should have caused problems earlier too.. Jul 27 15:54:21 let me see how the package was built previously Jul 27 15:54:52 oh, maybe I moved pidof.. Jul 27 15:55:04 hm.. no.. Jul 27 15:56:10 * mickeyl waits ages for openzaurus.org being resolved Jul 27 15:56:31 hail sf.net Jul 27 15:56:35 mickey@r2d2:/tmp$ ipkgfiles sysvinit-pidof_2.86-r12_arm.ipk Jul 27 15:56:36 drwxr-xr-x root/root 0 2005-04-07 13:59:50 ./ Jul 27 15:56:36 drwxr-xr-x root/root 0 2005-04-07 13:59:47 ./bin/ Jul 27 15:56:36 lrwxrwxrwx root/root 0 2005-04-07 13:59:47 ./bin/pidof -> ../sbin/killall5 Jul 27 15:56:39 hmm Jul 27 15:56:42 same way Jul 27 15:56:47 i wonder why it worked back then Jul 27 15:57:01 are you guys are pivot_rooting as well? Jul 27 15:57:05 -are Jul 27 15:57:16 think so.. Jul 27 15:57:21 how do you manage to unmount the old root? Jul 27 15:57:27 we have problems doing that Jul 27 15:58:02 uhm.. hope someone else can answer that.. Jul 27 15:58:22 aah-ha Jul 27 15:58:26 umount /dev/mtdblock2 Jul 27 15:58:35 after that it works Jul 27 15:58:43 so it's definitly the "double root" Jul 27 15:59:02 strange.. should that really affect symlinks? Jul 27 15:59:09 no it affects the ../ Jul 27 15:59:12 when going over the / Jul 27 15:59:35 cd bin && ls -l ../sbin does work when there is a single / Jul 27 15:59:43 it doesn't, when there is a double / Jul 27 15:59:51 s/double/additional old/ Jul 27 16:00:25 you really shouldn't have double / then :-) Jul 27 16:00:51 yah, but we can't umount it within the init process because the init process is started by the old / Jul 27 16:01:09 i wonder how other people are doing that Jul 27 16:01:14 that's why i asked Jul 27 16:01:59 pivot root, umount proc, umount old fs, exec /sbin/init Jul 27 16:02:02 (iirc) Jul 27 16:02:12 or umount it later Jul 27 16:02:38 since the oldroot should be in /mnt/oldroot or something Jul 27 16:10:17 hmm it's funny, the original root doesn't have the sysvinit.pidof - that's why ../sbin doesn't work Jul 27 16:10:20 * mickeyl shakes head and sighs Jul 27 16:21:04 mickeyl: the path name resolution of symlinks happens entirely within the kernel, what you are seeing is typically a symptom of having a pre-pivot root process Jul 27 16:21:25 i see Jul 27 16:21:34 pivot_root only works if done from process 1 (so far as I can see) Jul 27 16:22:04 and it may even be necessary for all other processes created before the pivot to die Jul 27 16:22:27 yeah. that would also explain why we can't unmount the old root Jul 27 16:22:28 <[g2]> iirc I think there's a -i option to reset the process id Jul 27 16:22:37 because there's still a process having a lock on it Jul 27 16:22:41 the code in openslug in /etc/default/functions (the 'swivel' function) gets it right, when executed from process 1 Jul 27 16:22:50 excellent. I'll have a look at that Jul 27 16:22:53 * mickeyl eyes are falling Jul 27 16:22:57 thanks. g'night Jul 27 16:23:14 <[g2]> nite Jul 27 16:23:20 g'nite Jul 27 16:25:57 mickey|zzZZzz (you still mounted be able to umount /initrd though - busybox has file descriptor 10 open on /linuxrc, until that's closed by init you can't umount it - that's why I added /etc/init.d/umountinitrd.sh) Jul 27 16:27:01 .... or something like that .... **** ENDING LOGGING AT Wed Jul 27 23:59:56 2005