**** BEGIN LOGGING AT Tue Feb 22 23:59:56 2005 Feb 23 00:00:44 ~seen dyoung Feb 23 00:00:46 dyoung <~dyoung@dyoung.nslu2-linux> was last seen on IRC in channel #nslu2-linux, 8h 15m 4s ago, saying: 'back later.'. Feb 23 00:01:09 well..its been 8 hrs..so another 2 to go! Feb 23 00:18:39 morning Feb 23 00:19:11 hi there Feb 23 00:35:47 http://www.bash.org/?464561 Feb 23 00:44:21 where do they come up with this stuff? Feb 23 00:45:34 being bored makes some ppl creative :) Feb 23 00:57:12 ~seen dyoung-web Feb 23 00:57:14 dyoung-web was last seen on IRC in channel #nslu2-linux, 5h 16m 23s ago, saying: 'back later.'. Feb 23 01:21:21 morning Feb 23 01:23:15 morgan Feb 23 01:34:07 morgan who? Feb 23 01:36:11 good morgan Feb 23 01:36:30 or guten morgan Feb 23 01:37:56 i'll stick with goeiemorgen then Feb 23 01:38:50 works for me Feb 23 01:39:49 still getting used to an activity that never seems to sleep. Feb 23 02:10:30 g'day Feb 23 02:12:07 howdy Feb 23 02:43:19 morning Feb 23 02:45:45 morning Feb 23 02:48:22 for a 1.5Gb openslug cornice disk, how should I partition it? Feb 23 02:48:43 128MB for swap, nnMB for rootfs, and rest for /home ... Feb 23 02:49:32 I wonder what a fully configured openslug (with *all* possible packages) takes up on a disk? Feb 23 02:50:19 Any suggestions on a default openslug partition scheme? (for when opensling is given a blank disk with no paritions on it ....) Feb 23 02:51:16 why even have separate / and /home ? Feb 23 02:51:16 I guess I could follow the unslung scheme for the moment ... then someone could use the linksys firmware to partition the disk, and keep their data intact when moving to openslug ... Feb 23 02:52:07 I'm thinking of being able to move the same disk between stock linksys, unslung and openslug, and only having to blow away one partition of the three (thereby keeping the user's mp3 collection still available on the disk) Feb 23 02:52:26 so what is data now on unslung would be /home on openslug Feb 23 02:52:35 and what is conf now would be / on openslug Feb 23 02:52:40 and swap would remain the same Feb 23 02:53:38 okay so swap is SDX3 Feb 23 02:53:39 this would only be for a default partitioning scheme - we wouldn't touch a partition table that had already been carefully crafted Feb 23 02:54:04 yeah, /home on sda1, / on sda2 and swap on sda3 (all by default) Feb 23 02:54:27 and only sda2 gets touched by unsling and opensling Feb 23 02:55:09 share/hdd/conf is sda1? Feb 23 02:55:39 conf is sda2 isn't it? Feb 23 02:57:02 yeah, it is Feb 23 02:57:10 yeah...its late Feb 23 02:57:28 . /dev/sda1 on /share/hdd/data type ext3 (rw) Feb 23 02:57:29 . /dev/sda2 on /share/hdd/conf type ext3 (rw,sync) Feb 23 03:00:38 okay then /home and / make sense Feb 23 03:06:24 ok, I'm going to make sda2 to be 256MB (same size as my flash key) Feb 23 03:07:14 then 128MB for swap (probably twice what it needs to be) and the rest for /home Feb 23 03:09:03 this is for OpenSlug? Feb 23 03:09:11 yep Feb 23 03:16:09 dyoung: ping (or wake up) Feb 23 03:17:33 ~nefarious Feb 23 03:17:34 well, nefarious is to be infamous by way of nastyness Feb 23 03:17:54 oh? Feb 23 03:18:24 dyoung has an electric shock cable hooked up, and he gets a shock when you say that word ... Feb 23 03:18:35 lol! Feb 23 03:19:05 I've sent him some things to chew on...till I get back tommorrow afternoon. Feb 23 03:21:20 rwhitby: seen my kernel config comments? Feb 23 03:22:35 yeah Feb 23 03:22:41 nice work Feb 23 03:23:06 once I get external disk booting working, I'll be enabling bluetooth modules, so I expect I will need to use it then Feb 23 03:25:21 i've written it on the wiki page but i'll repeat it to be sure: i have not tested any of this, so don't blame me if removing anything which i've listed breaks something ;) Feb 23 03:25:59 yep, that's the basis under which we all work here :-) Feb 23 03:27:53 well, and it's always true with alpha software ofcourse Feb 23 03:28:20 *insert standard disclaimer here* Feb 23 03:28:46 well...I give up for the night... Feb 23 03:28:53 night ka6sox Feb 23 03:28:56 nn Feb 23 04:09:39 ok, running OpenSlug from an external disk. Now to get it to happen automatically .... Feb 23 04:10:08 neat Feb 23 04:14:20 rwhitby: btw, i thought of something else Feb 23 04:14:45 rwhitby: when using pivot_root, you should unmount the dir you pivot initrd to after boot, and do a freeramdisk on /dev/ram0 Feb 23 04:14:56 rwhitby: gets you back about 1mb of mem Feb 23 04:16:45 yep, we've got that covered. Not going to do it under everything is working Feb 23 04:18:47 ok, but it is something to remember... for instance you should put freeramdisk in busybox :) Feb 23 04:19:44 hmm - we put it in unslung and switchbox - guess we need to add it to openslug busybox too :-) Feb 23 04:20:12 you don't really need it in switchbox i guess Feb 23 04:21:01 that's true Feb 23 04:21:38 wonder if you can copy it from /initrd/bin and use it .... Feb 23 04:22:01 use the switchbox freeramdisk in openslug directly :-) Feb 23 04:24:24 hm, you can't run it from initrd/bin Feb 23 04:24:36 'cause you only can run it when initrd is unmounted Feb 23 04:24:45 so you could copy, unmount, run, delete Feb 23 04:24:57 but then i think it's a better solution to just put it in busybox Feb 23 04:25:22 I think we can afford it Feb 23 04:25:26 I agree, I was just being facetious Feb 23 04:25:43 that's what I thought :-) Feb 23 04:26:11 you succeeded :) Feb 23 04:31:57 I just succeeded in an automatic boot to external disk too ... including dropping back to jffs2 if the disk is not connected or is hosed. Feb 23 04:33:17 Preferred root device is /dev/sda2 (disk 1) Feb 23 04:33:17 Sleeping 10sec while waiting for disk... Feb 23 04:33:17 mount: Mounting /dev/sda2 on /mnt/newroot failed: No such device or address Feb 23 04:33:17 umount: /mnt/newroot: Invalid argument Feb 23 04:33:18 Cant find valid rootfs on /dev/sda2, using jffs2 ... Feb 23 04:33:18 Root filesystem will be mounted from /dev/mtdblock4 ... Feb 23 04:35:11 nice! Feb 23 04:35:45 are you checking on the existence of init? Feb 23 04:35:56 yep Feb 23 04:35:58 k Feb 23 04:35:59 nice Feb 23 04:36:28 on my setup at home i have the 10s delay trimmed down to 2s Feb 23 04:37:27 yeah, for unslung 3.16 we went from 10 to 5, but got lots of people complaining Feb 23 04:37:47 stupid ass slow hdd's :) Feb 23 04:42:04 Preferred root device is /dev/sda2 (disk 1) Feb 23 04:42:04 Sleeping 10sec while waiting for disk... Feb 23 04:42:04 kjournald starting. Commit interval 5 seconds Feb 23 04:42:04 EXT3-fs: mounted filesystem with ordered data mode. Feb 23 04:42:05 Root filesystem will be mounted from /dev/sda2 ... Feb 23 04:42:05 INIT: version 2.86 booting Feb 23 04:42:36 yay Feb 23 04:42:56 I've wanted this for a long time Feb 23 04:44:14 maybe we should introduce a .hdd_delay on jffs2 Feb 23 04:44:33 or maybe something like .switchbox.conf Feb 23 04:46:15 03rwhitby 07 * r1.2846.1.1 10openembedded/ (7 files in 7 dirs): OpenSlug-1.3-alpha: Now supports booting from an external disk Feb 23 04:56:02 After a lot of trial and error I got Gallery working on Apache/php on the slu, I like to write an howto (for nlsu2-linux.org) how do I do this? Do I just make a document and send it somehwre or do I make that on the web? Feb 23 04:57:03 Edit http://www.nslu2-linux.org/wiki/HowTo/RunGalleryOnApache Feb 23 04:58:02 Thanks, will make something nice today :0 Feb 23 04:58:13 thank *you* Feb 23 04:59:40 such a wiki page might make it into a BuzBee article .... Feb 23 05:00:09 root Feb 23 05:00:21 oops - wrong window :-) Feb 23 05:00:35 heh Feb 23 05:00:58 good thing openslug doesn't have a password yet :-) Feb 23 05:01:40 Got my Linksys USBBT100 bluetooth dongle today :-) Feb 23 05:01:44 Gallery on the NSLU2 is quite a neat idea. What would be nice is writing a small app to automatically mount and add pics when you plug in your camera Feb 23 05:01:46 cool Feb 23 05:03:31 I'll let you know here when I am done writing, then you can koorekt mij bat engliszh ;) Feb 23 05:03:51 root@nslu2:/# free Feb 23 05:03:51 total used free shared buffers Feb 23 05:03:51 Mem: 30308 11604 18704 0 1704 Feb 23 05:03:51 root@nslu2:/# umount /initrd Feb 23 05:03:51 root@nslu2:/# /sbin/freeramdisk /dev/ram0 Feb 23 05:03:53 root@nslu2:/# free Feb 23 05:03:54 total used free shared buffers Feb 23 05:03:56 Mem: 30308 10180 20128 0 680 Feb 23 05:04:16 giel: is that the right syntax for freeramdisk? Feb 23 05:05:20 peterroubos: Have you worked out the netpbm problem? Feb 23 05:05:36 netpbm problem? Feb 23 05:05:47 pngtojpeg Feb 23 05:06:17 tried building it yesterday: no luck Feb 23 05:06:27 ah Feb 23 05:06:29 need some research Feb 23 05:06:44 strange config script Feb 23 05:06:47 what was wrong with it? Feb 23 05:06:48 ah Feb 23 05:07:31 but Gallery should with ImageMagick also ;) Feb 23 05:07:32 yep worked it out, it's running Feb 23 05:08:01 It works with imagemagic now, but Image magick wants to overwrite some library Feb 23 05:08:17 and if that is allready installed, imagemagick will abort installation Feb 23 05:08:24 -force-overwrite ;) Feb 23 05:09:17 hehe, I just uninstalled the lib first... (eh you know what lib that is?, otherise I have to search for it again, for documentation) Feb 23 05:09:55 no Feb 23 05:10:26 i had no problem installing imagemagick Feb 23 05:10:29 no problem, i'll look into it, will write the howto now and put the -force-overwrite in it Feb 23 05:10:46 cool Feb 23 05:11:38 got these? jhead, jpegtran Feb 23 05:12:14 rwhitby: yep Feb 23 05:12:35 ok, to prepare an external disk, you need to parition it with fdisk, then format it with mke2fs, then use the following command to copy the rootfs over to it: Feb 23 05:13:19 "cd / ; /usr/bin/find . -print0 -mount | /usr/bin/cpio -p -0 -d -m -u /mnt/sda2" Feb 23 05:13:34 (where /mnt/sda2 is where you mounted the partition). Feb 23 05:13:58 and what about device nodes? Feb 23 05:14:07 double check that /mnt/sda2/dev/console exists, then touch /.sda2root (or whichever partition you prepared) in the jffs2. Feb 23 05:14:23 then reboot Feb 23 05:14:54 giel: that's why we're using find and cpio Feb 23 05:15:50 rwhitby: uhm yeah, but no device nodes will be copied this way, will they? Feb 23 05:15:56 did for me Feb 23 05:16:13 hm okay Feb 23 05:16:20 but isn't /dev a mounted ramdisk? Feb 23 05:16:28 not in openslug Feb 23 05:16:33 ah ok Feb 23 05:16:50 i don't know what's standard and what isn't anymore, as i've changed a whole lot in my setup Feb 23 05:17:37 you need to start contributing some of those changes that are generally applicable .... Feb 23 05:17:58 i do mostly by yelling in here :) Feb 23 05:18:39 btw, a quick bk question Feb 23 05:19:00 i did a bk unedit defconfig and a pull and an update, but i didn't get a new version of defconfig Feb 23 05:19:38 * rwhitby shrugs Feb 23 05:19:43 hm well Feb 23 05:19:47 dunno anything about bk unedit Feb 23 05:19:49 don't mind Feb 23 05:19:51 me neither Feb 23 05:20:37 okay, on with the hacking Feb 23 05:26:08 jbot, seen mathieu_xwave? Feb 23 05:26:14 mathieu_xwave <~mathieu@80-218-37-110.dclient.hispeed.ch> was last seen on IRC in channel #nslu2-linux, 20d 20h 49m 6s ago, saying: 'bye'. Feb 23 05:30:45 03rwhitby 07 * r1.2855 10openembedded/packages/openslug-init/openslug-init_0.10.bb: Fixed the concatenations of and and friends. Feb 23 05:31:24 heh - the ${D} and ${sysconfdir} got eaten by the shell Feb 23 05:41:30 heh Feb 23 05:41:40 \$ Feb 23 05:41:51 or '....' Feb 23 05:43:56 anybody care to read/laugh/test http://www.nslu2-linux.org/wiki/HowTo/RunGalleryOnApache Feb 23 05:44:10 * perlguru-work reading.... Feb 23 05:46:02 linked onto the HowTo page .... Feb 23 05:46:35 ok, so who here has configured a kernel for bluetooth before? Feb 23 05:50:17 peterroubos: remove 'share/hdd/conf' from paths, because some people have /opt on the data part (like me) just use /opt/etc/... Feb 23 05:51:06 is there a slugbug on the dependancy of imagemagick on bzip2? Feb 23 05:51:25 ok Feb 23 05:51:54 @perlguru, not tha I know about, I found that out myself Feb 23 05:52:01 and you could add some headers, like: Installation, Configuration, Notes, etc... Feb 23 05:52:27 peterroubos: check slugbug.nslu2-linux.org Feb 23 05:52:52 Because I got errors using Imagemagic, I tried to run it standalone and it gave me an error about a missing lib, I googled, saw that it was part of Bzip2 Feb 23 05:54:43 ok, look at #105 and #106 ;) Feb 23 05:55:22 (we should have a bug-day) Feb 23 05:55:38 #106 is new to me Feb 23 05:55:45 :) Feb 23 05:56:55 hehe :-) I'll works some more on the how-to l8r (this evening) kids just came out of school and I must entertain them. Feb 23 05:57:14 bye Feb 23 05:58:24 morning Feb 23 06:00:11 VoodooZ_Work: got my BT dongle - now I need a copy of your defconfig :-) Feb 23 06:00:32 good. Feb 23 06:00:56 The problem is nothing works anymore so I can't rebuild my darn private changes. Feb 23 06:01:09 but I can give you the major CONFIG_ lines Feb 23 06:01:27 some of them are optional but I didn't get a chance to clean them up. Feb 23 06:07:13 ok, can you email me the defconfig? Feb 23 06:07:26 giel: ping Feb 23 06:09:45 ok, anyone object to the following changes to OpenSlug kernel defconfig ... (22 lines to follow) Feb 23 06:09:58 # CONFIG_PM is not set Feb 23 06:09:58 # CONFIG_APM is not set Feb 23 06:09:58 CONFIG_PACKET=y Feb 23 06:09:58 CONFIG_NET_SCHED=m Feb 23 06:09:59 CONFIG_NET_ESTIMATOR=m Feb 23 06:09:59 CONFIG_NET_CLS=m Feb 23 06:10:01 CONFIG_NET_CLS_ROUTE=m Feb 23 06:10:03 CONFIG_NET_CLS_POLICE=m Feb 23 06:10:30 # CONFIG_INPUT is not set Feb 23 06:10:30 # CONFIG_INPUT_MOUSEDEV is not set Feb 23 06:10:30 # CONFIG_INPUT_EVDEV is not set Feb 23 06:10:30 # CONFIG_SOUND_GAMEPORT is not set Feb 23 06:10:31 # CONFIG_USB_HID is not set Feb 23 06:10:31 # CONFIG_USB_HIDINPUT is not set Feb 23 06:10:33 # CONFIG_HID_FF is not set Feb 23 06:10:35 # CONFIG_HID_PID is not set Feb 23 06:10:37 # CONFIG_LOGITECH_FF is not set Feb 23 06:10:39 # CONFIG_THRUSTMASTER_FF is not set Feb 23 06:10:41 # CONFIG_USB_HIDDEV is not set Feb 23 06:10:51 CONFIG_USB_SERIAL_FTDI_SIO=m Feb 23 06:10:51 CONFIG_USB_SERIAL_VISOR=m Feb 23 06:10:51 CONFIG_USB_SERIAL_PL2303=m Feb 23 06:10:54 (that's it) Feb 23 06:14:38 Why is the visor one made into a module? Feb 23 06:15:17 oh - should it be a =y? Feb 23 06:15:33 Did somebody actually test those changes though? I know I screwed my build pretty badly last time I tried to do blind optimizations. Feb 23 06:15:54 No. Why do we have it at all? Feb 23 06:16:19 cause we have at least two Treo users on the list :-) Feb 23 06:16:24 ah Feb 23 06:16:35 rwhitby: Just in case you missed it: http://www.nslu2-linux.org/wiki/HowTo/NetworkSlugOverBluetooth Feb 23 06:16:37 Only because you're one :) Feb 23 06:17:11 hey, there's gotta be some perks to this job .... Feb 23 06:17:22 heh Feb 23 06:17:46 Just don't start using a PowerMate okay? I think I can justify a Palm driver but a USB knob is pushing it :) Feb 23 06:18:38 didn't you know that the whole nslu2-linux.org project was just a way of getting people to write stuff for my slug ... Feb 23 06:18:47 :) Feb 23 06:20:10 VoodooZ_Work: yep, got that page, and the links off it too Feb 23 06:21:04 good Feb 23 06:21:23 rwhitby: pong, looks ok for me Feb 23 06:21:47 ok, that saved another 23k Feb 23 06:22:44 is it possible openslug's kernel was too fat? all build from yesterday kept crashing at openslug-image. with an error about padding and size . Feb 23 06:23:42 And those were stock unmodified. I don't know what the hell is going on with my builds lately. nothing seems to work. Feb 23 06:24:55 no, kernel was fine here Feb 23 06:25:16 I just flashed one that was even compiled without the -j x tweak and although it built fine this time once flashed it give s me the libgcc_s.so.1 error again Feb 23 06:26:20 Unless you guys are using uclibc? Feb 23 06:26:45 giel: something went wrong. I guess I need to rebuild the rootfs too ? Feb 23 06:26:59 (got an oops and kernel panic) Feb 23 06:27:15 what does it say? Feb 23 06:27:37 rwhitby: hm, weird Feb 23 06:27:43 nslu2 login: Unable to handle kernel paging request at virtual address 853875df Feb 23 06:27:43 pgd = c0004000 Feb 23 06:27:43 [853875df] *pgd=00000000 Feb 23 06:27:43 Internal error: Oops: f3 [#1] Feb 23 06:27:43 Modules linked in: ixp425_eth ixp400 Feb 23 06:28:08 odd. not what I'm getting. Feb 23 06:28:21 I just ran the new kernel with the old rootfs, so any modules in the rootfs are from the old build (not from the new kernel build) Feb 23 06:28:36 shouldn't really matter i think Feb 23 06:29:08 Are you in a position to test small changes to the defconfig there? Feb 23 06:29:30 not now Feb 23 06:29:36 i'm at work Feb 23 06:30:10 ok, I'll just try doing only the input device stuff first. Feb 23 06:38:55 ok, just removing input devices worked. I'll stick with that for the moment (unless we go past 1MB) Feb 23 06:39:36 I'm also curious about the APM and PM junk as I wanted to clean stuff up too. Feb 23 06:39:50 Did you try the PREEMPT one? Is it worth it or? Feb 23 06:40:00 no, dunno. Feb 23 06:40:14 I just do what giel tells me to do :-) Feb 23 06:40:19 :) Feb 23 06:40:33 Maybe I should start pestering him then :) Feb 23 06:45:13 no! don't pester me! :) Feb 23 06:45:28 wasn't someone else trying PREEMPT? Feb 23 06:48:17 i don't think PREEMPT is really needed for the slug... it's mostly useful for responsiveness in desktop/multimedia systems Feb 23 06:48:22 I don't know. It's been in the back of my mind since I started playing with the slug. Feb 23 06:48:38 Well, wouldn't that help me for robotics. Feb 23 06:48:52 real-time and timers? Feb 23 06:48:59 hm... yeah, probably Feb 23 06:49:10 What's the side-effects? Feb 23 06:49:12 you're experiencing sluggish behaviour in userspace now? Feb 23 06:49:25 heh sluggish Feb 23 06:49:27 i meant laggish Feb 23 06:49:34 Not really Feb 23 06:49:45 well, if it ain't broke, don't fix it :) Feb 23 06:49:58 then again, I spend most of my time working on openslug so the robot code is lagging a bit. Feb 23 06:50:08 the downside is that preemption introduces many possible race conditions and locking issues Feb 23 06:50:13 I'm starting to test my computer vision stuff slowly. Feb 23 06:50:17 in the kernel that is Feb 23 06:50:22 ouch Feb 23 06:50:26 so you can try and hope the ixp is preemption safe Feb 23 06:50:35 What about the scheduler options you mentionned in the wiki page? Feb 23 06:50:43 those are i/o schedulers Feb 23 06:50:49 I'm ditching the ixp and csr crap anyways. Feb 23 06:50:57 hm, then you can probably use preemption Feb 23 06:51:03 I use I2C a lot for sensors. wouldn't that help? Feb 23 06:51:45 generally, kernel preemption improves responsiveness Feb 23 06:51:54 are you modifying kernel code yourself? Feb 23 06:51:55 I mean the I/O sched stuff. Feb 23 06:52:01 nope. Feb 23 06:52:14 oh, probl. not, with I/O sched i meant block device I/O Feb 23 06:53:04 i see. Feb 23 06:53:39 time for some coffee Feb 23 06:53:49 any tips on making sure i2c doesn't fail me? It's kind of my achilles heel as all my sensors and motors are controlled via it. Feb 23 06:54:21 don't burn out the GPIO? ;-) Feb 23 06:54:37 heheheh. good one! Feb 23 06:54:44 Nah! been there, done that! :) Feb 23 06:56:39 there, my procfs stub works Feb 23 06:56:58 (oh, you can ignore that remark, it doesn't have anything to do with the slug) Feb 23 07:30:34 root@nslu2:/media/card# hcitool inq Feb 23 07:30:34 Inquiring ... Feb 23 07:30:34 00:10:C6:29:C3:C5 clock offset: 0xe621 class: 0x72010c Feb 23 07:32:36 good sign Feb 23 07:32:57 built everything as modules instead of compiling it in Feb 23 07:33:20 What I do is I setup pand to search for a connection on startup so my robot can connet to my portable which acts as a NAP. Feb 23 07:34:06 and with BNEP on too after they connect the bnep0 interface will come up and configure itself if the interfaces file is setup correctly. Feb 23 07:34:23 Unless you're needs are different of course. Feb 23 07:34:38 IN my case I just want a plain TCP/IP tunnel Feb 23 07:35:10 In my case, the slug will be acting as a bluetooth network gateway to allow the Treo to get to the LAN and internet. Feb 23 07:35:15 to route request to the internet is a bit more complicated though. I've read about bridging but I'm sure there's other ways. Feb 23 07:35:25 makes sense. Feb 23 07:35:44 In my case, the slug is "mobile" so it's the other way around. Feb 23 07:35:53 BTW, I didn't have to comment out the pcmcia-cs stuff, I just allowed it to build Feb 23 07:36:14 But I'd love to use them for allowing my portable to surf around the house too eventually. Feb 23 07:36:36 of course. It's not a problem. but I just wanted to save space Feb 23 07:37:08 did you use the bluez-utils-nodbus one in your openslug.conf? Feb 23 07:37:24 just ipkg installed it manually Feb 23 07:38:19 cool. the nodbus removes all the useless dbus stuff which is nice as when I started I had to use the whole thing and there was errors. Feb 23 07:46:24 morning [g2] Feb 23 07:46:56 [g2]: 1.3-alpha is released Feb 23 07:47:08 morning [g2] Feb 23 07:47:17 <[g2]> morning Feb 23 07:47:26 <[g2]> rwhitby, what's in 1.3 ? Feb 23 07:47:41 switch to external hard disk Feb 23 07:48:15 <[g2]> ah... the find and cpio stuff Feb 23 07:48:18 check logs 2.5 hours ago Feb 23 07:48:22 <[g2]> mke2fs Feb 23 07:48:34 <[g2]> will do Feb 23 07:48:43 also changed switchbox to fallback to jffs2 if the disk is not connected or is hosed Feb 23 07:49:06 <[g2]> did you bump the switchbox rev ? Feb 23 07:49:11 yep Feb 23 07:49:21 <[g2]> ok, I'll pull and build not Feb 23 07:49:22 <[g2]> now Feb 23 07:49:31 everything should be bumped Feb 23 07:49:40 <[g2]> I'm quite excited about OpenSlug Feb 23 07:49:47 have me a good binary for the weekend! Feb 23 07:50:19 FrankenSlug is just pining for it Feb 23 07:50:32 <[g2]> I think the only big issue is working out the ehci interrupt issue Feb 23 07:50:40 [23:42] ok, to prepare an external disk, you need to parition it with fdisk, then format it with mke2fs, then use the following command to copy the rootfs over to it: Feb 23 07:50:41 [23:43] "cd / ; /usr/bin/find . -print0 -mount | /usr/bin/cpio -p -0 -d -m -u /mnt/sda2" Feb 23 07:50:41 [23:43] (where /mnt/sda2 is where you mounted the partition). Feb 23 07:50:41 [23:44] double check that /mnt/sda2/dev/console exists, then touch /.sda2root (or whichever partition you prepared) in the jffs2. Feb 23 07:50:41 [23:44] then reboot Feb 23 07:52:08 [g2]: I decided to partition my cornice drive the same as unslung, so that the data partition (sda1) would carry over unchanged. So the default partition scheme is sda1 is /home, sda2 is /, and sda3 is swap. Obviously OpenSluggers can change that easily .... Feb 23 07:52:58 we could just tell people to format the drive on the linksys firmware before flashing openslug, so we don't have to teach them how to partition a drive Feb 23 07:53:01 <[g2]> so you still want /opt on sda2 ? Feb 23 07:53:24 dunno - maybe sda1 is /opt :-) Feb 23 07:53:24 <[g2]> These are OpenSluggers right ? Feb 23 07:53:29 yeah Feb 23 07:53:33 The 6GB CF microdrives are going to be out soon so hopefully the 4GBs are going down Feb 23 07:53:51 Apple have just released the 6GB iPod mini Feb 23 07:54:57 [g2]: The main thing is putting / on sda2, so that people's data partition can stay intact when they upgrade from linksys or unslung to openslug Feb 23 07:55:22 and then opensling only touches sda2 Feb 23 07:55:58 (by default - of course true opensluggers will partition how they like) Feb 23 07:58:20 <[g2]> rwhitby, we should talk about package management issues for OpenSlug versus Unslung for the long-term (next 6 months) Feb 23 07:58:29 something tells me I'm going to need to print out this log for the OpenSlug doco Feb 23 07:58:48 yeah, but not now - my work here is done for tonight Feb 23 07:59:38 <[g2]> Ok. I'm going to write up a little something on the wiki about thoughts on the three different types of OpenSlug users Feb 23 07:59:46 <[g2]> for the Beta release Feb 23 07:59:49 ok, cool Feb 23 08:00:36 night all Feb 23 08:00:42 have fun with 1.3 Feb 23 08:00:42 <[g2]> nite Feb 23 08:00:53 <[g2]> it's building right now. THX! Feb 23 08:01:09 <[g2]> s/building/built/ Feb 23 08:01:18 np - had to do it to get my bluetooth working ;-) Feb 23 08:02:02 <[g2]> rwhitby-asleep, is BT in 1.3 or as modules Feb 23 08:03:08 I'm right her... oh, wait BT as in BlueTooth.... Feb 23 08:03:19 <[g2]> nod. Feb 23 08:03:23 mods Feb 23 08:03:43 actually, BT is not in 1.3 - maybe tomorrow Feb 23 08:04:34 do you want BT modules in the base image? Feb 23 08:04:46 <[g2]> as modules sure Feb 23 08:05:03 <[g2]> as long as it doens't push us over the 1M base kernel issue Feb 23 08:13:08 Man. My badluck will never end: I just cloned again, fresh build and now base-files craps out with: ln: creating symbolic link `/local/openslug/build/tmp/work/base-files-3.0.14-r23/imagemnt/card' to `/media/card': No such file or directory Feb 23 08:13:58 Openslug is finally moving in beta and I can't build it anymore. f'n great. Maybe I'll take a month off. Feb 23 08:15:57 last change was rwhitby. Do you guys do full clean builds when you test this stuff? Feb 23 08:20:57 anyone around? Feb 23 08:21:38 we're hidin Feb 23 08:21:40 +g Feb 23 08:22:25 They are all hiding from me :) Feb 23 08:22:31 Yep Feb 23 08:22:41 heh Feb 23 08:22:51 I don't blame you. I'm cursed. stay away from me. Feb 23 08:23:00 VoodooZ_Work, they do that to me too Feb 23 08:23:10 yeah. sucks. Feb 23 08:23:20 of course, I bug them with insane newbie-ish questions Feb 23 08:23:23 I'm trying to cross compile ghostscript and one of the steps is to fill out an architecture specific file with object sizes etc. If I post it to the yahoo group can someone help me fill it out? Feb 23 08:27:51 I have uploaded it and it resides in http://groups.yahoo.com/group/nslu2-linux/files. If someone can help me fill out it's contents I can proceed with the compilation. Feb 23 08:28:03 It's called arch.h Feb 23 08:29:08 good. kergoth confirmed it's a bug. Feb 23 08:29:41 * giel pats VoodooZ_Work Feb 23 08:29:50 hang in there, don't let it get to you Feb 23 08:30:38 yeah, I know but it's hard to do work when the floor keeps changing under you. Oh well, It's alpha so... Feb 23 08:32:34 <[g2]> VoodooZ_Work, we do pulls and sometimes full builds Feb 23 08:32:45 thought so. Feb 23 08:32:49 real life intrudes.... Feb 23 08:32:55 <[g2]> I know you are frustrated right now Feb 23 08:33:08 * ByronT-Work is away -( work )- at 10:33a -( P:On / L:On )- Feb 23 08:33:11 are you guys still using nslu2-linux as your parent or the oe-devel one? Feb 23 08:33:20 <[g2]> nslu2-linux Feb 23 08:33:23 ok. Feb 23 08:33:32 <[g2]> is that your parent ? Feb 23 08:33:40 yep. always been. Feb 23 08:33:44 <[g2]> ok Feb 23 08:33:58 I just had to change it for a sec to get kergoth base-files' fix though. Feb 23 08:34:01 <[g2]> I know you've been changing a lot on the build system(s) Feb 23 08:34:24 Not really. Just trying to setup the quad opteron one. Feb 23 08:34:25 <[g2]> A couple days ago you had a fresh clone that worked Feb 23 08:34:34 you dont have to change the parent Feb 23 08:34:40 just specify the parent to bk pull Feb 23 08:34:46 ok. Feb 23 08:34:50 bk pull http://oe-devel.bkbits.net/openembedded; Feb 23 08:35:02 I was going to change it back after anyways. Feb 23 08:35:07 * kergoth nods Feb 23 08:35:07 <[g2]> kergoth, *always* the MAN! Feb 23 08:35:10 heh Feb 23 08:35:12 yep! Feb 23 08:35:20 i wasnt the man when my sed script went wild and broke 33 packages :P Feb 23 08:35:27 :( Feb 23 08:35:35 I drove me crazy! :) Feb 23 08:35:42 heheh Feb 23 08:35:43 s/I/and/ Feb 23 08:35:52 that doesnt take much Feb 23 08:35:53 * kergoth runs Feb 23 08:36:00 <[g2]> kergoth, even *the MAN* makes mistake or you'd be *the computer* Feb 23 08:36:10 * VoodooZ_Work nods Feb 23 08:36:59 <[g2]> So VoodooZ_Work I'd go back to that working setup and start changing things one at a time from there Feb 23 08:37:02 Is it a problem if I use bk://nslu2-linux.bkbits.net/openembedded as my parent? Feb 23 08:37:20 Or is the http version? Feb 23 08:37:30 doesnt matter Feb 23 08:37:33 <[g2]> nod. Feb 23 08:37:41 ok. Just wanted to make sure. Feb 23 08:37:43 though myself for example, i need to use http most of the time. Feb 23 08:37:47 because i'm behind a proxy at work Feb 23 08:37:55 ic Feb 23 08:37:57 http proxy, and i dont know how to get bk:// through it :) Feb 23 08:38:04 bk obeys the http_proxy var for http:// Feb 23 08:38:49 odd. I'm getting conflicts on the pull. Feb 23 08:39:19 unsurprising, given the number of files i've been touching Feb 23 08:39:24 heheh Feb 23 08:39:25 just do a bk resolve, it probably wont even need your input Feb 23 08:39:31 should be all automatic Feb 23 08:39:47 not for me. It left it in RESYNC Feb 23 08:39:58 I guess I should simply take the remote one. Feb 23 08:40:06 (on resolve) Feb 23 08:40:16 hit 'd' Feb 23 08:40:20 look at the differences between them Feb 23 08:40:26 see if there are any local changes you need to keep Feb 23 08:40:46 there shouldn't be as this is a fresh clone. Feb 23 08:40:58 I get a r6 and r10 Feb 23 08:41:17 ? Feb 23 08:41:24 if you cloned an nslu2 tree Feb 23 08:41:31 and now you're pulling an assload of changes from oe-devel Feb 23 08:41:36 of course there are likely to be local changes you'd want to keep Feb 23 08:41:44 the local changes in the nslu2 tree that arent upstream yet. Feb 23 08:42:12 ic. I thought you meant my private changes which I haven't put in yet. Feb 23 08:42:49 locally modified files wouldve aborted the pull completely, never gotten to RESYNC Feb 23 08:42:51 The change is: mkdir -p ${D}/proc ${D}/mnt/jffs2 ${D}/mnt/repair ${D}/mnt/newroot VS mkdir -p ${D}/proc ${D}/mnt/tmpmnt ${D}/mnt/repair. Feb 23 08:42:57 [g2]: Is it one of yours? Feb 23 08:43:24 (switchbox_1.00.bb) Feb 23 08:43:49 <[g2]> VoodooZ_Work, don't think so Feb 23 08:44:00 with a switchbox package, you probably want whatever is in the nslu2 repo, as opposed to whats in oe-devel Feb 23 08:44:11 which is why i recommended not just blindly picking one or the other Feb 23 08:44:11 heh Feb 23 08:44:13 yeah. Feb 23 08:44:22 ok. Feb 23 08:44:43 Maybe I'll wait until the changes get merged in nslu2 and go to lunch before I break my keyboard. Feb 23 08:44:45 :) Feb 23 08:44:50 good plan Feb 23 08:45:13 i can go ahead and do a sync, push changes in both directions, if people want the repositories updated Feb 23 08:48:21 I simply used the local files in both cases. Feb 23 08:51:01 ok, base-files built. Feb 23 08:51:03 thanks Feb 23 08:51:14 but I'm not out of the woods yet. Feb 23 08:53:19 I even wiped my sources folder as I want to make absolutely sure It works. Feb 23 10:13:37 does anybody know what package contains librt.so.1 ? Feb 23 10:13:59 sorry din't change my name Feb 23 10:24:01 now it's module-init-tools that's broken! mv: can't stat source /local/openslug/build/tmp/work/module-init-tools-3.1-r0/imagebin/lsmod Feb 23 10:54:51 Stay! Feb 23 11:16:25 03ptweety * 10unslung/ (5 files in 3 dirs): cyrus-imapd again ready for native only testing Feb 23 11:23:02 03ptweety * 10unslung/ (Makefile make/perl-dbi.mk): perl-dbi ready again for native only testing Feb 23 11:25:02 Anyone willing to discuss the NPE ethernet driver for a minute? Feb 23 12:10:05 hey peeps please can someone explain to me how to use the sercomm upgrade utility? Feb 23 12:18:18 ~nslu-general Feb 23 12:18:41 ~nslu2-general Feb 23 12:18:43 #nslu2-general is the IRC channel specifically set up for end-user questions about the NSLU2 device. Questions that are not related to the nslu2-linux development projects should not be asked in #nslu2-linux, but should be asked in #nslu2-general instead. Feb 23 12:18:46 ah, there Feb 23 12:18:51 <[cc]smart> ~botsnack Feb 23 12:18:51 thanks, [cc]smart Feb 23 12:36:34 giel: Are you trying to tell me to go to another channel? Don't you think you might want to know what it is I want to discuss? Feb 23 12:36:58 beewoolie-away: pretty sure he was talking to Gorx. Feb 23 12:36:59 no, not you Feb 23 12:37:29 K. Guess there is still no one interested in talking about it. :-< Feb 23 12:45:41 talk about what? Feb 23 12:50:42 Hi jp30, thanks for helping me out yesterday (on Gallery) evrything is up and running, and the Howto is posted. Feb 23 12:58:59 peterroubos, yes, i saw it, very nice. hopefully we'll get those problems with imagemagick fixed soon Feb 23 13:04:45 siddy: I'm working to get a new bootloader running on the slug. The final hurdle is the NPE. I was hoping to be able to get away with not having to upload the firmware. Looks like that isn't the case. Feb 23 13:05:01 jp30, Well, with the "workaround" it isn't a problem Feb 23 13:05:01 siddy: I was wondering if anyone knew a little more about the NPE kernel driver Feb 23 13:06:43 NSLU2-IRC: Change your nick Feb 23 13:16:32 Hi! Feb 23 13:16:48 hi Feb 23 13:17:23 i have a name resolving problem with my nslu2 Feb 23 13:17:27 ~nslu2-general Feb 23 13:17:28 #nslu2-general is the IRC channel specifically set up for end-user questions about the NSLU2 device. Questions that are not related to the nslu2-linux development projects should not be asked in #nslu2-linux, but should be asked in #nslu2-general instead. Feb 23 13:17:57 ithanks and bye! Feb 23 13:18:17 bye Feb 23 13:18:27 see ya over there... :) Feb 23 13:18:40 beewoolie-away, sorry, I don't know of anybody having dealt with stuff like that. *shrug* Feb 23 13:19:14 siddy: Fair enough. I've found someone else interested on another channel. Feb 23 13:20:57 hey peep Feb 23 13:20:59 d Feb 23 13:21:10 other channel gorx... Feb 23 13:22:58 nite everyone Feb 23 13:25:18 any complaints and/or issues with altering wiki irc page to drop users into nslu2-general? Feb 23 13:26:09 my thought is that developers can find this channel with ease.... Feb 23 13:26:33 ask ka6sox and rwhitby. if they okay it then tell me and I'll change it over Feb 23 13:45:12 <[1]garpinc> Is someone able to help me with the architecture settings in arch.h file so I can compile ghostscript for nslu2. I posted file to yahoo group. Can someone look at it? Feb 23 13:51:03 <[1]garpinc> rwhitby-treo can you help? Feb 23 13:51:34 Tiersten: please do change the wiki IRC page as suggested Feb 23 13:53:17 k Feb 23 13:53:43 Tiersten- I'm catching the wiki side if you'll get the java side Feb 23 13:53:51 it might be nice to have a footnote mentioning this channel and purpose. Feb 23 13:53:54 I've changed it Feb 23 13:55:23 wiki page change for devs, as well Feb 23 14:11:14 [g2], still no luck. A fresh repo with all the defaults and with none of my changes, no -j x options, on a proven dual xeon box, and it still bails out in the same place. I'm running out of ideas. Feb 23 14:19:41 [g2], I noticed it prints this though before crashing: Preferred root device is jffs2 Feb 23 14:19:58 How do I make it boot my flash disk for example? Feb 23 14:28:54 * ByronT-Work is back -( work )- gone 5 hr 55 min 46 s Feb 23 14:29:55 VoodooZ_movie, do you get that crash when trying to boot while flash disk is attached? Feb 23 14:30:33 either way. I haven't been able to get a single load to work in the last 2-3 days. ANd I've tried everything. Feb 23 14:31:08 ah ok Feb 23 14:31:13 at first I was simply trying to boot directly off jffs2 too but then I tried the default from a fresh clone and it's the same. Feb 23 14:31:17 VoodooZ_Home: our repo is now fixed after kergoths wide-ranging changes (he picked up an old version of two files) Feb 23 14:31:44 Actually, I was there when he fixed it today and that load doesn't work more. Feb 23 14:31:51 VoodooZ_Home, so it crashes for you when trying to boot with nothing attached to the slug usb ports? Feb 23 14:32:01 yep. Feb 23 14:32:31 hold on. I just did a bk pull and I'm getting quite a few changes. Feb 23 14:32:42 VoodooZ_Home: all I can say is "Works for me" - sorry. I have no idea why you're having problems after it's built correctly. Feb 23 14:33:13 heh Feb 23 14:33:42 It's not like I changed something this time. I just cloned and built and flashed. Feb 23 14:33:45 "quite a few" holy smokes... Feb 23 14:34:21 rwhitby-away, I had to unplug my eletric shock cable. Feb 23 14:34:45 Actually it broke... I was getting to many shocks with "nefarious" and "'244". Feb 23 14:34:54 dammit how do I change the source-POS-forge mirror oe uses again ? Feb 23 14:35:09 3.61K/s from ale-f*cking-POS-eron Feb 23 14:35:18 2.56K/s Feb 23 14:35:32 Blazing Speed. Feb 23 14:35:35 well now it's just hung Feb 23 14:35:57 hi, all: it says on the wiki homepage that unslung packages are backwards compatible with earlier versions of the unslung firmware. this appears to be false, however (as Jim Tucker's post to the yahoo list illustrated). should it be changed? Feb 23 14:37:08 708.92B/s Feb 23 14:39:35 jp30: yeah, probably. Feb 23 14:39:36 back later Feb 23 14:50:35 Gorx is trying to turn me into newbie killer Feb 23 14:50:50 yay, killing newbies Feb 23 14:50:59 can i help? Feb 23 14:51:02 ~newbie Feb 23 14:51:03 i heard newbie is someone who is new to linux or debian, and should read the docs (/usr/share/doc/) Feb 23 14:51:04 * kergoth grins Feb 23 14:51:14 I was going to say "...into kergoth" but I didn't want to offend Feb 23 14:51:20 hehe Feb 23 14:52:41 at least now the java irc client shoots him into other channel Feb 23 14:53:01 <[g2]> kergoth, what did you tweak earlier in the repo with the concatenation ? Feb 23 14:53:25 ? Feb 23 14:53:38 ByronT-Work: ah, it defaults there now? Feb 23 14:53:40 thats good Feb 23 14:53:51 just for your benefit! Feb 23 14:55:13 so you don't get botthered by stuff like [16:49] then you need to telnet/ssh in and troubleshoot cups Feb 23 14:55:14 [16:49] gotta url? Feb 23 14:55:22 haha Feb 23 14:55:24 oh my Feb 23 15:59:37 <[1]garpinc> rwhitby-web got a min? Feb 23 15:59:58 impecable timing Feb 23 16:00:21 <[1]garpinc> who... Me or rwhitby? Feb 23 16:00:31 <[1]garpinc> :-) Feb 23 16:00:45 over here it looked like you asked the second that he left... Feb 23 16:00:53 after he left Feb 23 16:01:11 <[1]garpinc> rwhitby-web got a min? Feb 23 16:01:19 depends what for .... Feb 23 16:01:19 so you might want to try the "rwhitby-###: PING" method Feb 23 16:01:41 <[1]garpinc> Did you get mesg about ghostscript? Feb 23 16:02:01 nope. Feb 23 16:03:17 <[1]garpinc> I am trying to cross compile ghostscript and to do so part of the instructions say to create an arch.h file and modify it for architecture. I created it but don't know nslu2 architecture good enough to modify the file. Feb 23 16:03:20 heading home... bbl Feb 23 16:03:35 <[1]garpinc> I uploaded it to yahoo group. Can you modify it for me? Feb 23 16:04:35 <[1]garpinc> rwhitby-web: got the message now? Feb 23 16:06:37 <[1]garpinc> rwhitby-web: ping Feb 23 16:08:58 <[1]garpinc> gotta go home too... rwhitby-web if you can do it then please repost in in yahoo group under arch.h.nslu2. Otherwise I will catch up when I get home. Feb 23 16:44:01 kergoth: ping Feb 23 16:44:30 rwhitby-web: pong Feb 23 16:45:24 kergoth: we've got rootfs on external hard disk support working in OpenSlug now, and I'm looking for your advice on a default partition scheme and mount points Feb 23 16:45:30 here's the background: Feb 23 16:45:48 1/ we pivot_root to the external disk, so / is mounted from there. Feb 23 16:46:24 2/ we'd like to keep the three-partition stock linksys disk formatting scheme as the default partitioning scheme, to make it easy for people to migrate to openslug without loosing all their mp3's on the disk :-) Feb 23 16:46:46 (the linksys scheme is sda1 is data, sda2 is system and sda3 is swap) Feb 23 16:47:09 3/ We'd like for openslug to install packages by default onto the external disk, rather than the jffs2. Feb 23 16:48:02 4/ For the first betas of openslug, we want people to run rootfs from the external disk, rather than jffs2 (unless they know what they are doing). Then recovery is as simple as booting with the disk unplugged and reformatting the disk Feb 23 16:48:48 5/ compatibility of installed locations for packages between unslung and openslug would be nice (to save documenting two paths for config files) but not absolutely required Feb 23 16:49:12 6/ we want to diverge from the normal openembedded practices as little as possible Feb 23 16:49:26 [g2]: can you think of any other background points? Feb 23 16:51:00 we've seen the mount points under /media and /mnt, and they may need extending for large external disk support (e.g. not small flash cards) Feb 23 16:51:15 in oe, /mnt/*/ is deprecated. Feb 23 16:51:20 they're just compatibility symlinks Feb 23 16:51:28 (you could imaging someone having both a flash disk and a large external disk mounted on the slug). Feb 23 16:51:36 ok, so everything is /media now? Feb 23 16:52:18 its FHS 2.0 Feb 23 16:52:22 /mnt is a temporary mountpoint Feb 23 16:52:29 /media is media mountpoints Feb 23 16:52:31 yeah, i was wondering abut that too. I noticed an error earlier this week about htat Feb 23 16:52:35 so no subdirs under /mnt/ Feb 23 16:52:42 my initial thought was to use the linksys paritioning scheme (to allow for easy migration without data loss) and then mount sda2 as / and sda1 as /home and sda3 as swap Feb 23 16:52:59 other option is to mount sda1 as /opt instead Feb 23 16:53:18 but then I saw the /media direcory, and figured we needed your counsel on the matter Feb 23 16:53:59 <[g2]> rwhitby-web, that's pretty much a good summary Feb 23 16:54:15 my personal recommendation would be to mount it under /media/, and use unionroot to overlay the hd on /, and have ipkg's 'dest' be /media/whatever, and make that dest default. Feb 23 16:54:36 installs go directly to the real mountpoint, but the files are accessible from normal locations in / via the union Feb 23 16:56:36 <[g2]> that's sounds very sensible except that we're out of room on the kernel Feb 23 16:56:53 <[g2]> I don't know how big unionfs is, maybe we could fit it in Feb 23 16:57:00 its massive. Feb 23 16:57:11 it used to be one of the biggest kernel modules i've ever seen Feb 23 16:57:12 ok, we weren't intending on adding unionroot to OpenSlug 1.x, so what would be your non-unionroot recommendation (keeping in mind a transitition to unionfs in the next major release)? Feb 23 16:57:16 we got it down quite a bit though Feb 23 16:57:27 03jp30 * 10unslung/ (make/imagemagick.mk sources/imagemagick/control): imagemagick: depend on bzip2, libtool, do not install libltdl Feb 23 16:57:40 <[g2]> last check we had like 10K left :( Feb 23 16:57:43 i'd mount it on /media/ and pivot to it then. Feb 23 16:57:58 i dont see an advantage of putting it on /opt unless you arent pivoting, and want /opt/bin etc paths to function Feb 23 16:58:20 oh - we do the pivoting in switchbox, so OE sees the chosen root parition from external disk as / before OE starts Feb 23 16:58:56 so /dev/sda2 will be mounted as /, and we're deciding what to do with /dev/sda1 .. Feb 23 16:59:47 ahh gotcha Feb 23 16:59:55 (/dev/sda2 is prepared by doing a find | cpio from jffs2 to /dev/sda2 when setting up for external disk) Feb 23 16:59:56 <[g2]> and /dev/sda1 will be mounted as /home in /sda2 Feb 23 17:00:08 that's one option Feb 23 17:00:24 thats what i'd do. in fact, thats what we do do in OZ Feb 23 17:00:29 on pdas with split flash Feb 23 17:01:10 or mount it under /media/ and bind mount it on /home Feb 23 17:01:12 I like the idea of /home, cause then someone's public area from when they were using stock firmware will appear as /home/public ... Feb 23 17:01:44 ah - we could mount under /meda, and then bind mount both /home and /opt from it Feb 23 17:02:32 question for those folks who use stock firmware features - when you create a new user and give them a share, where does their private disk area end up under /dev/sda1 ? Feb 23 17:02:38 subdirs of it? /media/sda1/somethingorother /opt? Feb 23 17:02:47 yeah Feb 23 17:02:55 thatd work well Feb 23 17:03:03 the key with this is, you need to do that bind mount in a hotplug script Feb 23 17:03:07 or udev or whatever Feb 23 17:03:16 because you dont want the bind mount happening if the hd didnt mount Feb 23 17:03:28 agreed Feb 23 17:04:06 can someone with unslung firmware go into the web interface, and create a new user and give them a private share and tell me where that share lives under /share/hdd/data/... ? Feb 23 17:04:50 I know that the public area is under /share/hdd/data/public, and currently when we unsling to hdd-data the /opt directory lives under /share/hdd/data/opt - so that all works well. Feb 23 17:05:45 if people stored all their mp3's in a subdir of /share/hdd/data/public, and we mounted /share/hdd/data/public as /home, then they would end up with a /home/mp3s dir on unslung .... Feb 23 17:06:09 you mean openslug? Feb 23 17:06:26 (of course /dev/sda1 would be mounted under /media/sda1 or /media/hdd or something in openslug, rather than /share/hdd/data in unslung) Feb 23 17:07:09 so /opt would be a bind mount to /media/hdd/opt and /home would be a bind mount to /media/hdd/public, or maybe just /media/hdd ... Feb 23 17:07:18 yeah Feb 23 17:07:39 (depending on where private shares end up in unslung, so that they could be carried across as normal home directories in openslug) Feb 23 17:08:17 Come-on people - no-one has a Linksys web interface where they can create a user with a private share and tell me where the share is placed on disk? Feb 23 17:09:47 kergoth: should we have packages install under /opt like in unslung, or should we use the ipkg-link functionality in oe? Feb 23 17:10:30 and have different dests for jffs2 and hdd ? Feb 23 17:11:17 well, the advantage of ipkg-link is that the ipk doesnt have to change based on dest. Feb 23 17:11:21 the paths remain the same Feb 23 17:11:37 which is really quite nice. Feb 23 17:12:01 * VoodooZ_Home decides to take his frustration out on the game Chronicles of Riddick instead... Feb 23 17:12:52 kergoth: yeah - will that remain the recommended practise in the light of unionfs? Feb 23 17:13:07 s/practise/practice/ Feb 23 17:15:02 i.e. will the dest stuff be deprecated by unionfs? Feb 23 17:15:23 yes Feb 23 17:15:26 it will Feb 23 17:15:30 by what i described earlier Feb 23 17:15:33 you'll still use ipkg dests Feb 23 17:15:37 but unionfs will replace ipkg-link'ing Feb 23 17:17:13 <[g2]> for the initial beta release I was thinking of creating the dest dir on the hd/flash stick that's "Opensling'ed" and *not* having it on the jffs2 partition Feb 23 17:17:26 <[g2]> that way by default nobody can install to the jffs2 partition Feb 23 17:17:27 isnt that what rwhitby-web just said? Feb 23 17:17:38 he said switchbox would pivot to the hd Feb 23 17:17:44 ok, so we should design our layout based on a default dest of hdd for openslug, but with people being able to repoint that to jffs2 Feb 23 17:18:14 we can mount the jffs2 under /media/jffs2, and install packages there :-) Feb 23 17:18:30 <[g2]> kergoth switchbox pivots based on what's in /.xxxxx of the jffs2 partition Feb 23 17:18:51 but seriously, the default dest will just be / and where that ends up will depend upon whether you're booting from jffs2 or hdd Feb 23 17:19:06 then we will have hdd and jffs2 dests where you can override the default one Feb 23 17:19:30 so if you're booted to jffs2, you can install to hdd, and if you're booted to hdd, you can install to jffs2 Feb 23 17:19:41 that might even come in handy for os upgrades ... Feb 23 17:20:04 <[g2]> rwhitby-web, the install to jffs2 is what I'm trying to avoid in the first beta Feb 23 17:20:19 right - you will have to know what you are doing to know how to do that Feb 23 17:20:43 (assuming you've pivoted to hdd already) Feb 23 17:21:24 <[g2]> right. It think we boot to jffs2 and manually allow the user to change root to the hd / flash stick via the opensling Feb 23 17:21:39 there's not a single person in this channel who can bring up a linksys web interface and create a new user with a private share and tell me where the directory ends up ????? Feb 23 17:21:58 rwhitby-web, i can in a couple of hours if you can wait Feb 23 17:22:00 * [g2] doesn't even run Unslung Feb 23 17:22:17 <[g2]> I only run linksys to reflash Openslug :) Feb 23 17:22:19 ... no access to my slug's web interface from here Feb 23 17:22:23 iirc, /share/hdd/data/dyoung Feb 23 17:22:50 if you give me a few minutes I can plug my stock nslu2 in Feb 23 17:22:57 ~praise Tiersten Feb 23 17:22:59 All hail Tiersten! Feb 23 17:23:05 ~praise dyoung Feb 23 17:23:07 All hail dyoung! Feb 23 17:23:20 Yeah, /share/hdd/data/dyoung; but the uid/gid is admin:everyone Feb 23 17:23:21 if it is as dyoung recalls, then we will mount sda1 as /home Feb 23 17:23:38 so thats kind of a pita. Feb 23 17:24:01 <[g2]> Ok I think we may want to forgore that compatibility Feb 23 17:24:17 is it readable/writeable for groupu everyone? Feb 23 17:24:22 <[g2]> and start using the normal user layout Feb 23 17:24:25 its 775 Feb 23 17:24:56 ah - so we can just map the gid of everyone from unslung to a corresponding group on openslug Feb 23 17:25:41 anyway, as long as it is readable to all, then people can chmod -R it after they migrate Feb 23 17:25:57 and chown -R it Feb 23 17:26:06 okay, I'm not sure if I created that as a private share. Feb 23 17:26:21 so maybe tiersten should double check. Feb 23 17:26:22 ok, let's see what a private share does before we make a decision Feb 23 17:26:38 or wait for another volunteer. Feb 23 17:26:41 kergoth: is /media/hdd a good name for the external disk mount point? Feb 23 17:26:47 back later food. Feb 23 17:26:52 or /media/sda1 ? Feb 23 17:27:23 cause I think /media/card and /media/cf don't make as much sense for openslug ... Feb 23 17:27:48 yeah. confirmed what dyoung said Feb 23 17:27:52 hdd sounds fine to me Feb 23 17:28:04 /share/hdd/data/tman if you create a private share on disk 1 Feb 23 17:28:15 the directory is owned by admin.everyone Feb 23 17:28:59 the public area is /share/hdd/data/public so it makes sense Feb 23 17:29:52 <[g2]> we should pick a couple high profile target apps like what ? samba, nfs, and see how they'd transfter from Unslung to OpenSlug with different versions of the sw. Feb 23 17:30:17 mt-daapd :-) Feb 23 17:30:23 hit the buzbee article apps Feb 23 17:30:33 then he can write the migration to openslug article Feb 23 17:30:46 <[g2]> those would be good choices for the uses Feb 23 17:31:05 all the developer packages are already there in OE Feb 23 17:31:07 transfer? Feb 23 17:31:10 oe already has samba Feb 23 17:31:15 i dont get what you're trying to do Feb 23 17:31:40 I think g2 means add any missing apps to OE if they are not there already. Feb 23 17:31:50 and yes, samba, nfs, etc are there already Feb 23 17:31:56 <[g2]> I mean samba in OE is 3.10 Feb 23 17:32:04 <[g2]> what version is Unslung running ? Feb 23 17:32:09 "udevstart is about 30% faster with all these changes now. Feb 23 17:32:10 " Feb 23 17:32:15 Feb 23 17:32:19 :) Feb 23 17:33:00 [g2]: since we're not carrying any user interface over, then we don't need to support migration of samba configs, etc. Feb 23 17:33:14 just let them start from scratch with swat or webmin in openslug Feb 23 17:33:30 or emacs :-) Feb 23 17:33:44 yeah, i dont get the purpose of this 'transfer' that keeps being alluded to Feb 23 17:33:49 if a package isnt in oe, add it to oe Feb 23 17:33:51 <[g2]> Ok then we shouldn't be setting up partitions based on the old layout Feb 23 17:33:54 its that simple Feb 23 17:34:11 the user being able to get to their old data if they want to is a good thing Feb 23 17:34:18 tryingi to use that data automatically in openslug is not Feb 23 17:34:19 [g2]: why not - we want the data to be migrated, but not the outdated app configs Feb 23 17:34:21 since compatibility isnt guaranteed Feb 23 17:34:27 yeah Feb 23 17:34:37 <[g2]> exactly what kergoth said Feb 23 17:34:37 data good, rootfs/appconfig bad Feb 23 17:34:48 exactly Feb 23 17:35:26 Assume that a user is actually using the slug as intended by Linksys, and has loaded up a disk with data,neatly organised by user in shares. Feb 23 17:35:42 take that disk, and put openslug in the conf partition. Feb 23 17:35:50 <[g2]> so being that they are root have total control over the devices I fail to understand how it might be hard to get their data Feb 23 17:36:09 eh? Feb 23 17:36:11 no one said it was hard Feb 23 17:36:22 yep, just look in /home/username Feb 23 17:36:33 (where you'd expect it to be) Feb 23 17:36:47 and /home/public is where all your old public share data was Feb 23 17:37:08 could even have a default openslug samba config which shared /home/public .... Feb 23 17:38:09 <[g2]> Ok, I think I'm giving up on the migration strategy and siding with kergoth Feb 23 17:38:30 <[g2]> meaning, bb app, ipkg install app, and migrate data to said app Feb 23 17:38:45 yep - that's my position too Feb 23 17:38:56 and if app is in unslung but missing from oe, then we add it to oe Feb 23 17:38:59 with openslug, i think we need to keep focused on doing the right thing, at least within reason. Feb 23 17:39:05 fully agree Feb 23 17:39:06 there's no reason for hacks there, since compatibility isnt an issue Feb 23 17:39:14 fully agree Feb 23 17:39:23 :) Feb 23 17:39:43 has someone made a list of things that need adding to oe yet? Feb 23 17:39:43 and when there is a default choice to be made (like the default disk partitioning) we do compatibility if it is clean Feb 23 17:39:56 rwhitby-web: right Feb 23 17:40:05 from what I've seen here, we have a clean data migration strategy Feb 23 17:40:20 <[g2]> ok and the user must install to hd or flashstick initially for the first Beta and then we can focus on getting boatloads of packages in there and not worry about size very much Feb 23 17:40:29 correct Feb 23 17:41:14 but that's a doco thing. if our power users choose not to switch rootfs to hdd, everything will still work properly for them out of jffs2 Feb 23 17:41:16 <[g2]> so the only thing we need to worry about is /swap Feb 23 17:41:29 that's easy - sda3 is swap Feb 23 17:41:47 default size is what linksys formats to Feb 23 17:41:48 <[g2]> actually I think the user can config swap and root Feb 23 17:42:35 yes, but we need to give a default config for camp#2 people so they don't have to think about disk partitioning Feb 23 17:43:15 and have the opensling script create that partioning if the user has not previously partitioned the disk Feb 23 17:43:16 <[g2]> right. the question is whether is a doco thing or that's the default working version of the sw. Feb 23 17:44:11 always remember. users dont usually actually _read_ doco :P Feb 23 17:44:11 <[g2]> Ok .... I'm gonna toss a big ring out here Feb 23 17:44:16 I propose that the sda1:/home, sda2:/, sda3:swap partition scheme be the default scheme that opensling creates, but that openslug works with any scheme and mounts the disks based on labels Feb 23 17:45:07 <[g2]> I propose we might want to make the sda1 be everything and leave sda2 untouched and sda3 is swap Feb 23 17:45:36 <[g2]> if sda2 becomes root, then the users can *never* go back to unslung or linksys Feb 23 17:45:39 that introduces maintenance problems, cause you're mixing rootfs and user data Feb 23 17:46:05 a rootfs vs user split is a very useful thing. you can wipe the rootfs and start fresh on the OS without losing your data Feb 23 17:46:13 exactly Feb 23 17:46:46 [g2]: we can update the unsling script to be able to migrate an openslug disk back to unslung Feb 23 17:47:04 <[g2]> Well if we are wiping out backward compatibility, then a simple script can move the data around to the right spot Feb 23 17:47:29 we are not doing backward compatibility - we are doing data migrating into openslug Feb 23 17:47:42 I don't want to support going from openslug back to anything else Feb 23 17:47:58 I do want to support seamless migration of data from anything else into openslug Feb 23 17:48:33 anything over and above those two premises is icing on the cake Feb 23 17:48:42 <[g2]> We'll I'm missing the point on the data migration to OpenSlug Feb 23 17:49:05 you don't have an unslung disk with 100Gb of mp3's on it .... Feb 23 17:49:16 <[g2]> let's pretend I do Feb 23 17:49:43 yep, so you flash openslug, and run opensling, and your mp3's magically appear under /home/public Feb 23 17:49:44 <[g2]> and there are shared via samba ? Feb 23 17:50:38 does that matter? once they are under /home/ then you can do lots of things with them, including having samba share them by default Feb 23 17:50:55 oooh... this sounds like a user type discussion (I'm well qualified) Feb 23 17:51:07 <[g2]> ByronT, jump on in Feb 23 17:51:14 I would propose that our default openslug samba config *does* automatically share /home/public Feb 23 17:51:49 <[g2]> let's prented it didn't and we needed that Feb 23 17:52:11 hrm Feb 23 17:52:20 <[g2]> mout -o bind /mnt/inthewoods /home/public Feb 23 17:52:20 ?= only works if unset, not if set to "" Feb 23 17:52:20 for one thing, at least make samba stick to pre-defined defaults Feb 23 17:52:20 so in the scenario stated, you flash openslug, run opensling, ipkg install samba, and you're MP3s are shared by samba as they were before .... Feb 23 17:52:22 * kergoth wonders if thats sane or not Feb 23 17:52:22 <[g2]> mount -o bind /mnt/inthewoods /home/public Feb 23 17:53:15 [g2]: don't follow your point Feb 23 17:53:50 <[g2]> my point is that from anywhere that data can be mapped to /home/public Feb 23 17:53:57 what is it about data migration from disks that were previously used with stock or unslung firmware, to becoming available under /home ? Feb 23 17:54:33 yes, I agree. I still don't understand your point or objection ? Feb 23 17:54:36 I don't know, from the doco point of view, I believe that there is going to be a need for a migration doc Feb 23 17:54:45 for unslung to openslug Feb 23 17:54:46 <[g2]> it's a flexibility issue. Feb 23 17:55:14 such that openslug is more of a linux distro rather than stuck with unslung hangovers Feb 23 17:55:15 [g2]: we're discussing a default config here, not a restriction on flexibility Feb 23 17:55:35 most linux distros suggest that you mount /home from a different partition to / Feb 23 17:55:39 <[g2]> If look at the linksys data and figure out where it needs to get mapped, we can just tweak the install scripts for OpenSlug to map from whereever to what's needed Feb 23 17:56:04 I thought that's exactly what we are doing and have solved? Feb 23 17:56:13 g2 seems to be advocating that "migration" is just bind mounts, not actual movement or rearranging of content Feb 23 17:56:19 <[g2]> I'm suggesting that we may want to mount sda1 as /mnt/media/home Feb 23 17:56:36 * kergoth wonders what /mnt/media/home is Feb 23 17:56:59 * rwhitby-web wonders too ... Feb 23 17:57:41 since sda2 will be mounted as /, and sda1 will be mounted as /media/hdd, and also bind mounted to /home, what is the question/issue? Feb 23 17:58:34 [g2]: ? Feb 23 17:58:34 <[g2]> I was saying the opposite sda2 is /, sda1 is /home Feb 23 17:58:55 um, isn't that what I just wrote? Feb 23 17:59:16 * kergoth chuckles and heads home Feb 23 17:59:20 heh Feb 23 17:59:20 "sda2 is /, sda1 is /home", "sda2 will be mounted as /, and sda1 will be mounted as /media/hdd, and also bind mounted to /home" Feb 23 17:59:20 <[g2]> no you wrote sda1 as /media/hdd Feb 23 17:59:40 and also bind mounted to /home Feb 23 18:00:10 the advantage is that sda1 is always in a known path in /media, even if the user chooses to kill the bind mount and do something else Feb 23 18:00:14 as i see it. Feb 23 18:00:21 * kergoth actually heads home now Feb 23 18:00:27 cause kergoth suggested that oe practice is to mount under /media and have the hotplug script do the bind mount, so that /home doesn't point to nowhere if the disk is not mounted Feb 23 18:01:08 but in this case that point is moot, cause if / is mounted from sda2, then you can be pretty sure that sda1 is also mounted Feb 23 18:01:22 (since they are on the same physical disk) Feb 23 18:02:03 [g2]: can you please explain your question/issue - I'm just not getting it at all .... Feb 23 18:02:14 so is the question to me more of what should default smb ipkg share be like Feb 23 18:02:30 <[g2]> rwhitby-web, I think we are pretty much in agreement Feb 23 18:02:37 <[g2]> sda2 is / Feb 23 18:03:01 <[g2]> sda1 is /media/hdd Feb 23 18:03:10 yep, yep Feb 23 18:03:23 <[g2]> we bind mount /media/hdd to /home Feb 23 18:03:32 so what was the last 30mins about? Feb 23 18:03:42 * ByronT scratches his head Feb 23 18:03:43 (been out to blunch) Feb 23 18:04:03 dyoung-web: it was about 30 mins too long :-) Feb 23 18:04:20 they agreed to agree Feb 23 18:04:25 okay cool. Feb 23 18:04:32 Whys it called opensling? Feb 23 18:04:34 I think we're all in violent agreement Feb 23 18:04:47 openslug version of unsling - got a better verb ? Feb 23 18:04:50 for name compatibility? Feb 23 18:05:08 beats "openslug_migrate" I guess. Feb 23 18:05:28 unopen? Feb 23 18:05:39 actually it would be called "openslug_external_rootfs_prep_and_copy" Feb 23 18:05:43 <[g2]> because with out unslung worries it'd be called setup Feb 23 18:06:05 setup.exe? Feb 23 18:06:23 Must... Resist.... Feb 23 18:06:41 Urge...... Feb 23 18:06:42 <[g2]> ok so lets start migrating packages Feb 23 18:06:52 ok, so we have a plan. Feb 23 18:06:57 heck no! that would encourage the newbies who have no business trying openslug to just click and go Feb 23 18:07:08 Okay cool. Feb 23 18:07:17 [g2]: I think we should use the term "add missing packages to oe" rather than "migrating packages" Feb 23 18:07:19 With any luck I'll have the remaining nfsroot bits worked out this evening. Feb 23 18:07:43 cause we're only migrating data, not packages Feb 23 18:07:51 so this means I should have a *good* openslug binary for doco prep this weekend? Feb 23 18:07:59 <[g2]> rwhitby-away, php gcc are in OE, but don't work like the unslung version Feb 23 18:08:00 is mt-daapd a oe package yet? Feb 23 18:08:09 we're providing the latest version of the same packages with a clean configuration in openslug, not migrating from unslung. Feb 23 18:08:10 and I can be in the first 20 still... Feb 23 18:08:44 I don't think we want openslug packages to work like unslung ones ... we want them to work like normal linux packages. Feb 23 18:09:02 <[g2]> Ok I'm *all* for that Feb 23 18:09:16 nods Feb 23 18:09:36 <[g2]> do we want users to have step-by-step instructions to get said apps working ? Feb 23 18:09:43 Doesnt look like mt-daapd is there yet. Feb 23 18:09:50 cool. as part of the clean configuration, we may decide to have a default (but not imposed) setting which picks up the migrated *data* automatically, but we should stop at that. Feb 23 18:09:56 <[g2]> what version of samba is linksys running ? Feb 23 18:10:05 2.x something Feb 23 18:10:16 [g2]: said instructions should come with the packages, since we are standard linux in openslug Feb 23 18:10:29 [g2], instructions should be something like 1) ipkg update 2) ipkg install package 3) configure your personal preferences if applicable Feb 23 18:11:13 4) home website of package Feb 23 18:11:13 <[g2]> Ok so there's no real migration Feb 23 18:11:22 dyoung-web: yeah, and 4) if you get stuck, consult the net for the standard doco and help forums for that software application, don't come asking us. We just packaged it. Feb 23 18:11:35 Right. Feb 23 18:11:44 no migration of apps, some migration of data (where it can be done cleanly) Feb 23 18:11:54 Yeah, this is a good plan. Feb 23 18:12:06 openslug must remain pristine and hack-free Feb 23 18:12:18 * dyoung-web hides Feb 23 18:12:23 we've even gotten rid of the real-root-dev hack in switchbox, so we're good there now too Feb 23 18:12:32 excepting dyoung's hacks? Feb 23 18:12:52 * rwhitby-web - the original binary-sed-the-kernel progenitor, also hides Feb 23 18:13:02 gone is /dev/slug too. that o ne was pretty bad... Feb 23 18:13:07 yep Feb 23 18:13:19 [g2]: the biggest advantage of using a bind mount for sda1 is flexibility. i can easily then change the bind mount to use a subdir of the partition instead of the partition itself, and the original content is always available. Feb 23 18:13:48 note that when you switch to unionfs, we need _every_ component of the union to have its original content mounted somewhere safe and independent from the union (aka, somewhere in /media) Feb 23 18:13:54 * kergoth gets home Feb 23 18:13:59 ok, I think we're good to go. any objections? Feb 23 18:14:13 show of hands? Feb 23 18:14:19 <[g2]> here Feb 23 18:14:20 none here Feb 23 18:14:33 <[g2]> here as in good to go Feb 23 18:14:38 capcom? flight? medical? go for launch? Feb 23 18:14:40 do I count? Feb 23 18:14:44 sure do Feb 23 18:14:46 <[g2]> We are saying OpenSlug is just a normal OE distro Feb 23 18:14:54 [g2] nod Feb 23 18:15:01 no probs here Feb 23 18:15:03 as it should be. Feb 23 18:15:26 I cant wait! Feb 23 18:15:31 <[g2]> Ok so we've just go to get cracking Feb 23 18:15:31 [g2]: thats the idea. OE distros tend to do the "right thing". few hacks. and it leverages the rest of hte oe community, due to the small # of deviations. Feb 23 18:15:33 yep, with a default partition scheme which happens to be the same as Unslung, but is not restricted to be so. Feb 23 18:15:34 <- is excited Feb 23 18:16:01 <[g2]> Ok so here's a big one also. Feb 23 18:16:13 FrankenSlug is getting all warm inside... Feb 23 18:16:24 <[g2]> Let's stop talking about OpenSlug in here and move the discussion to OE Feb 23 18:16:27 dyoung: we need to release an Unslung 4.x beta with nfsswitch.conf in it ... Feb 23 18:16:41 [g2]: yes, i agree with that wholeheartedly (and always recommended that..) Feb 23 18:16:42 :) Feb 23 18:17:18 * [g2] hugs kergoth for spending so much time and energy in nslu2-linux Feb 23 18:17:42 i should use my slug more often Feb 23 18:17:47 need to find a good use for it Feb 23 18:18:17 * rwhitby-web now goes to lunch Feb 23 18:18:19 rwhitby-web: as in do away with .nfsroot and use /mnt/jffs2/etc/nfsswitch.conf instead? Feb 23 18:18:24 hmm, food. good idea Feb 23 18:18:49 thats probably a better idea since the file contents are used. Feb 23 18:19:04 dyoung-web: no, referring to the bug reports on unslung which say that we need to add a standard /etc/nsswitch.conf (or whatever it's called) file Feb 23 18:19:12 nothing to do with switchbox Feb 23 18:19:15 of, you said nfsswitch. Feb 23 18:19:16 heh Feb 23 18:19:25 okay Feb 23 18:19:26 cool Feb 23 18:19:27 sorry - my bad :-) Feb 23 18:19:52 we do need to change the nfsroot stuff to use the drivers from jffs2 and work for unslung and openslug .... Feb 23 18:20:17 we may even decide to dump payload altogether ... Feb 23 18:20:39 rwhitby-lunch, check pm Feb 23 18:24:26 yeah, I was in the middle making all that nfsstuff work. Feb 23 18:24:41 cool Feb 23 18:25:02 back later ... Feb 23 18:26:07 <[g2]> thx all Feb 23 18:26:33 there's a dyoung special openslug binary??! Feb 23 18:27:33 eh? Feb 23 18:27:56 openslug yahoo table Feb 23 18:28:27 bbl - wife wants help with boy Feb 23 18:31:25 <[g2]> dyoung, did beewoolie-away run APEX ? Feb 23 18:31:27 03jp30 * 10unslung/sources/apache/rc.apache: openldap: don't stage libtool archives (.la); apache, apr-util: ldap support, fixup nsswitch.conf if it is broken Feb 23 18:31:37 oh that. ka6sox needed openslug in a hurry to demo at the socal linux expo. Feb 23 18:32:01 He was hesitant to flash it. Feb 23 18:32:08 hes been running it from ram though Feb 23 18:32:48 [g2]: have you fiddled with 1.1.1 yet? Feb 23 18:33:15 <[g2]> no, APEX 1.1.1 ? Feb 23 18:33:22 Yeah, Feb 23 18:33:38 <[g2]> see you in #openJTAG Feb 23 18:33:52 I wonder if I can get there from here.... Feb 23 19:08:22 dyoung-web: ping Feb 23 19:08:52 he might be paying more attention in #openJTAG Feb 23 19:09:55 pong Feb 23 19:10:25 did you get mail? Feb 23 19:13:03 today? Feb 23 19:13:09 3am Feb 23 19:13:21 and there should be some from 9am too Feb 23 19:20:32 I got the ones from lastnight., Feb 23 19:21:34 nothing today. Feb 23 19:24:18 hmmm....let me get back into the machine and send it again. Feb 23 19:24:26 could you open them? Feb 23 19:24:39 Yes Feb 23 19:24:51 though I really wish I had pcb working on the mini. Feb 23 19:25:04 oh....doesn't work? Feb 23 19:29:11 not directly on the mini. Gotta ssh elsewhere. Feb 23 19:29:27 ugh Feb 23 20:14:02 wouldn't a mini running debian run pcb, tho? Feb 23 20:16:43 Probably. Feb 23 20:18:19 cbpage, he still is runnin OSX Feb 23 20:18:35 and I use 2 Apples running Debian with PCB. Feb 23 20:20:47 03jp30 * 10unslung/Makefile: promote perl-dbi, cyrus-imapd Feb 23 20:30:08 is there a openslug feed yet? Feb 23 20:30:31 dyoung-web: we need to start one. Feb 23 20:30:44 there is an unslung-oe feed currently Feb 23 20:30:45 Okay. Feb 23 20:31:27 <[g2]> and that should probably generate the OpenSlug Beta Binary Feb 23 20:32:07 Feed me! Feb 23 20:32:28 I guess we should probably put the proper ipkg config to look for that feed too eh? Feb 23 20:32:36 yes! Feb 23 20:32:55 wheres the unslung-oe feed? Feb 23 20:35:49 nevermind answer that later, I'll use my private one for now. Feb 23 20:45:34 03jp30 * 10unslung/make/openldap.mk: openldap: do not stage server or clients, just libraries Feb 23 20:52:27 03jp30 * 10unslung/make/openldap.mk: openldap: unbork previous commit Feb 23 21:08:05 03jp30 * 10unslung/Makefile: promote gtk, vte Feb 23 21:17:12 03jp30 * 10unslung/make/xvid.mk: xvid: synchronize makefile version number with control file Feb 23 21:45:50 good night Feb 23 22:01:02 03jp30 * 10unslung/make/ (10 files): cosmetic changes to some pre-template.mk makefiles Feb 23 22:28:02 jp30 you around? Feb 23 23:07:06 03jp30 * 10unslung/make/apache.mk: make apache require upgraded apr-util Feb 23 23:09:09 jp30 I'm about to check in cvs based transcode. Feb 23 23:10:18 hi, garpinc, sounds good. Feb 23 23:10:43 also jp30 can you help me with ghostscript... Feb 23 23:11:05 To crosscompile it requires proper values to be in arch.h Feb 23 23:11:13 well, i could take a look... from what you were saying to rwhitby, it sounds hard Feb 23 23:11:23 hi rwhitby-web Feb 23 23:11:25 I uploaded it to yahoo groups. I don't know proper settings. Feb 23 23:11:35 nfsroot is workable. I had user error. Feb 23 23:13:55 garpinc, i haven't used yahoo except for the mailing list... looking now... Feb 23 23:14:24 it's in files section. Feb 23 23:15:53 compile genarch.c native to generate the file? Feb 23 23:16:32 03garpinc * 10unslung/make/transcode.mk: cvs based transcode Feb 23 23:17:07 dyoung: so no changes to switchbox? Feb 23 23:17:07 I don't have native compile env. Can someone else do that and send it to me? Feb 23 23:17:28 jp30 transcode should be all set. Feb 23 23:18:04 garpinc: did you add it to the READY_FOR_TESTING target? Feb 23 23:18:14 garpinc, you need at least ARCH_IS_BIG_ENDIAN 1; some of those other things are very murky, best to take dyoung's advice Feb 23 23:18:50 no changes, but I'm about to obliterate the payload dependancy for nfsroot but leave it for recovery. Feb 23 23:19:05 right. Feb 23 23:19:24 can we do the network and nfs stuff the same for unslung and openslug? Feb 23 23:19:33 do you have a pending switchbox changes? Feb 23 23:19:47 We should be able to; I was about to test it with openslug right now. Feb 23 23:20:12 jp30 do you have native compile env? Feb 23 23:21:18 I have no pending switchbox changes, so you go ahead and release a new version with nfs payload obliterated when it's ready. Feb 23 23:21:46 okay gonna test the method with openslug first then. Feb 23 23:22:01 cool Feb 23 23:22:04 later ... Feb 23 23:22:05 garpinc, sure do... have you set up ghostscript in cvs? Feb 23 23:22:40 not yet. Feb 23 23:24:00 03jcoalson * 10unslung/ (3 files in 2 dirs): add new "flac" package Feb 23 23:24:30 garpinc, why don't you get it in, then i (or some other dev) can start building it native to get the arch.h file for you Feb 23 23:24:54 I was reading cross compile instructions and they said that was the first step to do a make obj/arch.h and then edit it. I suppose you would do something entirely different for native. I could check in what I have. All it does right now is get the source. Feb 23 23:26:01 well, check it in and let's take a look. Feb 23 23:26:22 hang on a sec... Feb 23 23:27:48 03jcoalson * 10unslung/Makefile: add new "flac" package Feb 23 23:29:52 garpinc, still having trouble with transcode: aclocal: configure.in: 26: macro `AM_PROG_AS' not found in library Feb 23 23:35:03 I now the answer to that one. Feb 23 23:35:16 automake1.8 needs to installed. Feb 23 23:35:51 I use colinux and automake1.4 was installed. I haad to remove that and apt-get install automake1.8 Feb 23 23:37:05 garpinc, you need to give an explicit version to aclocal and automake. on the official build machines "automake" is automake 1.4. we all have 1.9 but the binaries are called aclocal-1.9 and automake-1.9. look at how the x11 .mks do this Feb 23 23:37:24 automake 1.9 shoudl be backwards compat with 1.8 Feb 23 23:37:43 s/all/also/ Feb 23 23:38:45 hmm but I don't call it explicitly I call autoreconf -i -f Feb 23 23:39:09 autoreconf will use the values of ACLOCAL and AUTOMAKE envs Feb 23 23:41:43 03garpinc * 10unslung/ (make/ghostscript.mk sources/ghostscript/control): So far uncompilable ghostscript Feb 23 23:44:10 03garpinc * 10unslung/make/transcode.mk: added env variables for automake Feb 23 23:44:24 jp30 all set on both counts... Feb 23 23:47:15 garpinc, what time is it where you are? Feb 23 23:49:51 ...it's nearly midnight here... Feb 23 23:50:22 2:49.... I know... I know.. I should get to sleep.... Feb 23 23:50:54 am.... Feb 23 23:51:28 03jp30 * 10unslung/sources/ghostscript/arch.h: ghostscript: added arch.h for cross-compilation Feb 23 23:51:54 thought so, you made a 2am kind of typo with transcode... don't worry, i will fix it up Feb 23 23:53:10 03jp30 * 10unslung/make/transcode.mk: transcode: pass explicit version to automake Feb 23 23:54:45 and i've checked in the arch.h you need to sources/ghostscript/arch.h Feb 23 23:55:17 Ok thanks... Feb 23 23:56:55 now i am going to get some sleep, will check if transcode worked tomorrow **** ENDING LOGGING AT Wed Feb 23 23:59:56 2005