**** BEGIN LOGGING AT Sat Feb 17 02:59:57 2007 Feb 17 03:32:56 Good Evening! For SlubOS/BE, where on the bootup disk is the ip# saved? I am bringing up a second slug, with the image of the Feb 17 03:33:44 another disk running SlugOS/BE. Even though the second slug had a different ip#, after bootup it assumes the ip# of the Feb 17 03:33:58 other slug from where the disk copy was made. Feb 17 03:35:31 it should be /etc/network/interfaces Feb 17 03:37:24 osas, thank you very much! I see the number there. Feb 17 03:37:31 k Feb 17 03:38:28 how did you instructed the second slug to boot up from the mirrored disk? Feb 17 03:42:16 chacko ? Feb 17 03:42:37 Hi Feb 17 03:42:51 see my previous question Feb 17 03:42:54 how did you instructed the second slug to boot up from the mirrored disk? Feb 17 03:43:21 osas, I did it with turnup disk without the -i, since I had a fully made image of the working disk Feb 17 03:43:31 oh ... I see Feb 17 03:43:33 k Feb 17 03:43:50 did you used dd to mirror the disk? Feb 17 03:44:33 osas, everything worked --- everything worked, it was unreal, I cannot tell any difference between the two slug!! Feb 17 03:44:44 did you used dd to mirror the disk? Feb 17 03:44:52 I used rsync to make the copy on the first slug Feb 17 03:45:03 ah ... I see Feb 17 03:45:04 nice Feb 17 03:45:23 now I know how to back-up my flash stick ;-) Feb 17 03:45:25 thx Feb 17 03:45:47 osas, I can help you with that if you need, I routinely do that Feb 17 03:46:10 I know where to find you ;-) Feb 17 03:46:37 For my harddisk, I have 4 identical 10GB partitions, and I usually backup my entire images to one of these in a rotating basis Feb 17 03:46:39 maybe you can post it on the wiki Feb 17 03:46:59 I see Feb 17 03:47:09 my goal is to ackup the flash Feb 17 03:47:17 I'm running asterisk from flash Feb 17 03:47:29 osas, I did it based on the stuff in the Wiki, but I will add some stuff to confirm what I did Feb 17 03:47:43 k Feb 17 03:48:00 I can backup from the flash just as well, using rsync --- the same script Feb 17 03:48:20 I was thinking about dd ... Feb 17 03:48:32 I have actually gone back and forth between flash and harddisk Feb 17 03:48:57 chacko: your rsync made a perfect copy of the disk, but the sysconf partition in flash overwrote some of the critical files, like the IP address. If you were to reset the sysconf partition prior to booting the "cloneslug" up, it might not require the extra step to restore the IP address. Feb 17 03:49:12 I have two linuxrc files under / (internal flash) Feb 17 03:49:20 one for flash and the other one for hd Feb 17 03:49:33 I just need to use the right one and reboot Feb 17 03:50:40 mwester, in my case nothing was overwritten, that why I had the same ip# !! I did not use turnp disk -i, but without -i Feb 17 03:51:52 without the new disk, I had a differnt --- correct --- ip# for the second slug in the flash, but the boot from the disk used everything from Feb 17 03:52:04 the other --- as to be expected. Feb 17 03:53:44 osas, the big advantage of rsync over dd is that after the first copy, rsync does delta copy by default Feb 17 03:54:04 I see Feb 17 03:54:32 osas, for me the first copy takes 15 minutes to 30 min, depending on the data, but the subsequent just a min or less Feb 17 03:54:49 nice Feb 17 03:55:15 * osas will need to start playing with rsync Feb 17 03:55:56 mwester, is there a way to determine what other data like the ip# may be different between two different slug need to operate on the same net Feb 17 03:56:19 just the IP Feb 17 03:56:37 Great!! Feb 17 04:23:23 03bzhou * r5515 10optware/trunk/make/py-formencode.mk: py-formencode: 0.6 -> 0.7 Feb 17 04:24:57 In case anyone has some insights into this, I've been stuck on this all afternoon: http://www.pastebin.ca/359980 Feb 17 04:25:57 check in the #oe channel Feb 17 04:26:11 you did :-) Feb 17 04:31:50 I've been in both all day Feb 17 04:31:57 I've been learning a lot Feb 17 04:32:08 an it seems I'm close to getting this to work Feb 17 04:32:32 I know I haven't defined a target package, but one thing at a time ;) Feb 17 04:32:54 as far as I can tell, I'm doing exactly as the docs indicate Feb 17 04:33:11 I'm onviously missing something though Feb 17 04:33:36 it seems that CFLAGS just isn't getting picked up Feb 17 04:34:07 which of course, is the item I really need. The others are in there just for sanity checking Feb 17 04:34:31 osas, hi Feb 17 04:35:02 osas, I need to change the hostname too! Where/how? Feb 17 04:39:06 chacko: just run turnup init, or check the turnup script to see where it writes the hostname if you prefer. Feb 17 04:43:33 mwester, thank you! Feb 17 08:13:47 how do I go about building a toolchain to include a specific version of libgcc1 ? Feb 17 12:21:00 Hi all Feb 17 12:23:18 I've just installed unslung 6.8 on my nslu2 and when I connect by telnet or ssh on the shell I can't get auto-completion Feb 17 12:24:23 alain__: the default shell in unslung 6.8 is very basic, it doesn't have support for auto-completion Feb 17 12:24:43 you need to install bash or some other shell to get that Feb 17 12:24:50 ok Feb 17 12:25:13 alain__: and questions like this are better suited for #nslu2-general ;) Feb 17 12:25:31 ok sorry Feb 17 12:25:48 I asked this on nslu2-genral but got no response Feb 17 12:26:16 ohh, I missed that. Sorry :) Feb 17 12:27:02 how much memory does bash use ? Feb 17 12:28:22 not very much :) less than sshd Feb 17 12:56:52 I'm trying to compile glftpd on debian sarge, but I miss something : /usr/bin/ld: ERROR: /usr/lib/libc_nonshared.a(elf-init.oS) uses hardware FP, whereas /glftpd/bin/ansi2gl uses software FP Feb 17 12:56:59 any ideas what's wrong ? Feb 17 13:03:40 morning Feb 17 13:07:11 mornin Feb 17 13:42:14 I guess I need to use nwfpe Feb 17 13:45:42 should I ? :) Feb 17 14:04:40 does anyone here know off-hand if commits to monotone.openembedded.org are synced back into the nslu2 monotone server? Feb 17 14:05:59 yes Feb 17 14:06:25 hi NAiL and thanks. Makes kinda sense but asking doesn't hurt heh Feb 17 14:07:05 W_SHaRK: arm debian uses hardfloat (which is really silly since there are hardly any ARM devices with FP) Feb 17 14:07:37 the kernel already has nwfpe. It handles all FP through exceptions. Feb 17 14:08:33 hm Feb 17 14:09:01 should I compile this nwfpe module or is it already running ? Feb 17 14:09:18 It's already running Feb 17 14:09:24 look at the dmesg output Feb 17 14:09:39 AFAIK, it's not possible to compile an ARM kernel without either nwfpe or fastfpe Feb 17 14:10:16 checking Feb 17 14:10:27 so you need to compile glftpd with hardware fp Feb 17 14:11:26 how do i do that ? with cross compile ? Feb 17 14:12:16 are you compiling natively? Feb 17 14:12:26 or cross? Feb 17 14:12:27 yes Feb 17 14:12:29 native Feb 17 14:12:55 odd Feb 17 14:13:07 thanks a lot for your infos, it's much appreciated ! :) Feb 17 14:13:38 ok so I'll set a cross compile env Feb 17 14:14:03 You might want to check #debian-arm on irc.debian.org Feb 17 14:14:14 thx Feb 17 14:14:25 If you're compiling natively and you get that error, I believe it's a bug with the toolchain Feb 17 14:18:43 are you a fan of NiN btw ? :) Feb 17 14:19:20 heh, I was a bit more NiN fan in the old days. I don't think it's actually related to my nick though :-P Feb 17 14:19:37 :) Feb 17 14:20:32 cbpage ping Feb 17 15:02:14 is there a set of virtual packages for building a toolchain that will cross-compile for 3.10-beta ? Feb 17 15:03:34 when I got my recipie going for logrotate, I found it ended up with a dependency on libgcc1 (>= 4.1.1). How do I sddress this? Feb 17 15:03:57 erthe resultant logrorate package did, I mean Feb 17 15:04:05 how do I address this? Feb 17 15:07:39 http://www.nslu2-linux.org/wiki/DebianSlug/CrossCompiling Feb 17 15:12:25 if only I was running debian Feb 17 15:15:00 you can set a virtual machine on your pc or something like that ? Feb 17 15:16:48 I suppose Feb 17 15:19:21 I'm running FC6. I had no trouble with the current toolchain via 'bb virtual/armeb-linux-gcc virtual/libc'. I'm just suprised the virtual packages aren't versioned such that you can built toolchains suitable for each release Feb 17 15:19:48 My setup is as described here: http://www.nslu2-linux.org/wiki/Optware/SlugosbeDevel Feb 17 15:20:18 I guess the other option is to wait for my second slug to arrive (next week) and use it as a dev enviroment Feb 17 15:20:32 installing the bleeding edge code there Feb 17 15:20:55 or, I suppose I could try a native build... Feb 17 15:21:18 that too, would take a great deal of setup Feb 17 15:21:51 I've already spent days at this and now I find I can't even test the damn package Feb 17 15:21:58 quite frustrating Feb 17 15:23:27 make setup-slugos-3.10-beta will check out the proper sources for 3.10-beta. Feb 17 15:23:53 cd releases/slugos-3.10-beta/ && make DISTRO=openslug Feb 17 15:24:07 after a while it'll have built you a toolchain that is correct for 3.10 Feb 17 15:24:10 I'm still a N00b with the arm slug... anyway, I thought native compilation will be easier, but it definitlvy not working on my nslu Feb 17 15:24:17 oh Feb 17 15:24:37 geez. I wish I'd known that days ago :) Feb 17 15:24:50 will update the wiki as soon as i get this working Feb 17 15:28:04 NAiL: is there a way to avoid building the whole distro in that case, as there is with the current minimum toolchain? Feb 17 15:30:39 NAiL: before I heard from you, I had attempted 'bb virtual/armeb-linux-gcc-3.4.4 virtual/libc' Feb 17 15:31:06 well, building the kernel should build about the minimum Feb 17 15:31:10 more or less Feb 17 15:31:36 it does seem to be building the right versions, but it's still building. Should I wait for it to finish or would it just run up against other dependencies? Feb 17 15:32:27 perhaps I should stop and start again with your approach since it's known to work Feb 17 15:35:45 hi, not sure if this is the right place to ask about the wl500g, somehow I can't find the prebuild toolchain from ASUS, any ideas where to get it? the GPL package from ASUS did not have it too Feb 17 15:38:13 sorry can't help you :/ Feb 17 15:38:35 is there a dedicated optware channel? Feb 17 15:39:55 well ok, I'll try a different approach then :) eno created optware packages of the software that I am working on, but that only worked out for non uclibc targets Feb 17 15:40:04 for uclibc we had a problem with a dependency - with id3lib Feb 17 15:40:15 I'd like to take a look at it, maybe I can fix it Feb 17 15:40:46 so basically I just want to build optware for some uclibc target, but I am new to optware and I do not have any hardware supporting it Feb 17 15:41:34 which target is the "easiest" in terms of installation? Feb 17 15:41:50 i.e. - without any external vendor downlaods or something Feb 17 15:41:54 ? Feb 17 15:43:19 did you look at : http://www.nslu2-linux.org/wiki/Optware/AddAPackageToOptware ? Feb 17 15:44:23 I did.. but I have to admit that I was looking for wl500g specific info :) let me check again Feb 17 15:45:01 well, it's a bit nslu2 specific here Feb 17 15:45:04 I guess Feb 17 15:45:58 well, thats the thing - I did a test with optware and could build for nslu2 Feb 17 15:46:07 but as far as I know nslu2 is glibc based? Feb 17 15:46:25 I need some uclibc target Feb 17 15:46:28 03osas * r5516 10optware/trunk/ (make/asterisk14.mk sources/asterisk14/nv.patch): asterisk14: r55097 -> 55196 + adding three extra nv_app files Feb 17 15:47:28 what are you running on your wl500g ? Feb 17 15:47:48 I do not have a wl500g :) I just have a nasty user who keeps pestering me about a port of my upnp server to the wl500g Feb 17 15:47:55 I think he uses Oleg's firmware Feb 17 15:48:08 ha sorry :) Feb 17 15:48:11 :) Feb 17 15:48:35 so right now I am looking for an easy way to build something for this target Feb 17 15:49:01 I understand now :) Feb 17 15:49:09 but I failed at... finding the prebuild toolchain, which is really stupid :) Feb 17 15:49:22 the links in the forums turn out dead, and the asus.zip did not have it Feb 17 15:49:25 :) Feb 17 15:50:42 are you talking about this forum ? http://wl500g.info/ Feb 17 15:50:49 yes Feb 17 15:51:23 maybe I missed something.. but so far the links I found did not work for me; I got oleg's update to the toolchain Feb 17 15:51:29 but afaik the original from ASUS is still required Feb 17 15:51:38 Jin^eLD: You can make the toolchain by using the master makefile AFAIK Feb 17 15:52:16 NAiL: you mean just set OPTWARE_TARGET to oleg and it will build the toolchain? Feb 17 15:52:56 I don't remember the specific target, but it should be in the list at the start of the optware makefile Feb 17 15:52:58 Jin^eLD: try "cd optware; make oleg-target; cd oleg; make directories; make toolchain" Feb 17 15:53:12 hey eno :) Feb 17 15:53:23 thx Feb 17 15:53:56 still building mine on my poor dual PII-300 lol Feb 17 15:54:15 crap.. Resolving nas.kynisk.com... failed: Name or service not known - it can't download the linux-libc-headers-TS101.tar.bz2 file Feb 17 15:55:07 google keeps silent.. are there no mirrors for this one? Feb 17 15:55:19 no, and you shouldn't need it Feb 17 15:55:27 infact, I should remove that part from the makefile Feb 17 15:56:25 ok where do I disable it? Feb 17 15:57:33 optware/make/buildroot.mk Feb 17 15:57:44 yep just found it.. Feb 17 15:58:05 or you can try touch downloads/linux-libc-headers-TS101.tar.bz2 Feb 17 15:58:17 if you are maintaining the MakeFile NAiL, you should correct the setup-host-ubuntu to use sudo instead of su, because ubuntu disables root by default Feb 17 15:58:58 eno: thx, it seems to be building now Feb 17 15:59:23 W_SHaRK: I'm not, but I'll do it ;) Feb 17 15:59:46 oops sorry, I forgot the wiki powa :) Feb 17 16:01:47 03nail * r5517 10optware/trunk/make/buildroot.mk: buildroot.mk: Remove traces of the ts101 target Feb 17 17:01:04 any1 knows "ERROR: You must set MACHINE to nslu2le or nslu2be. Feb 17 17:01:07 ? Feb 17 17:01:23 i was trying to "make slugosbe-image" Feb 17 17:02:54 eno: your OE env isn't setup right. Dunno about the makefile thingy I'm afraid Feb 17 17:14:11 ah, slugosbe/conf/auto.conf is where it's set Feb 17 17:14:36 hi eno Feb 17 17:14:37 i probably setup slugosbe a couple of days ago, and it was using nslu2 Feb 17 17:14:41 hi osas Feb 17 17:14:50 can you check this: http://bugs.digium.com/view.php?id=8637 Feb 17 17:15:02 tha latest post from digium Feb 17 17:15:15 about the pg_conf Feb 17 17:15:37 I was trying to get postgres support for asterisk14 Feb 17 17:17:40 besides checking the existence and trying to execute pg_config, what info does configure need? Feb 17 17:17:53 dunno exactly ... Feb 17 17:18:04 I will try to find out Feb 17 17:18:52 do you think that we can build two pg_conf? Feb 17 17:19:04 one against the host and the other one cross? Feb 17 17:20:08 install the host one in the stage and the other on in tha package Feb 17 17:20:37 sometimes we do that, we even have a host/{builds,staging} for that Feb 17 17:20:54 it really depends on what info configure needs Feb 17 17:21:18 I will check and let you know ... Feb 17 17:21:34 can we fix postgres? Feb 17 17:21:42 if we have these info (such as -I or -L -l), then we might just well give it Feb 17 17:22:25 right now the buil will fail b/c is checking the executable ... Feb 17 17:22:35 so we will need to patch the config ... Feb 17 17:22:36 i prefer not changing postgres unless we have to Feb 17 17:22:49 and the whole purpose of the bug was to get rid of the patches Feb 17 17:23:16 well, to me, patching configure to make it cross compilable is common Feb 17 17:23:18 I will dig more and let you know ... Feb 17 17:23:55 yeah ... but some good S/W out there will compile ok w/o a patch :-) Feb 17 17:24:04 and this is my goal Feb 17 17:24:22 I got openser to fix some issues :-) Feb 17 17:25:17 in this case, fixing postgresql will be more work Feb 17 17:25:40 I see Feb 17 17:26:40 maybe we can live without it since we have odbc :-) Feb 17 17:28:29 checking and running pg_config made the assumption that compilation is not cross Feb 17 17:29:32 there are other packages depending on postgresql that can get -I -L -l info in different ways Feb 17 17:29:48 although i wish postgresql were using pkgconfig Feb 17 17:30:09 I will see what I can do ... Feb 17 17:42:53 eno: so the oleg toolchain built now, how do I make a specific package? Feb 17 17:43:49 build that is Feb 17 17:48:46 I tried like, make mediatomb but it starts building binutils and stuff like that, allthough the toolchain is already built, so I cam kind of confused Feb 17 17:49:32 s/cam/am/ Feb 17 17:57:39 try "cd oleg; make mediatomb" Feb 17 17:58:28 if you type "make mediatomb" while at optware, it will try to build nslu2 toolchain Feb 17 17:58:59 oh ok, that was my mistake, thanks Feb 17 17:59:46 you probably need "make ipkg-utils" Feb 17 18:00:21 uhm Feb 17 18:00:56 this is so that you have ipkg-utils for packaging Feb 17 18:00:59 it checked out mediatomb and is doing rm -rf on oleg/buildroot Feb 17 18:01:11 so it seems it wiped out the toolchain once again Feb 17 18:01:14 :) Feb 17 18:01:23 that would be the 3rd time I am building it today hehe Feb 17 18:02:33 that could be because NAiL changed make/buildroot.mk Feb 17 18:02:46 I did not check out that yet though Feb 17 18:03:04 that's wierd, i cannot explain Feb 17 18:03:38 well, maybe I have better luck in an hour :) Feb 17 18:04:10 this wl500g thing, does it have an MMU? Feb 17 18:27:01 looks like I can use the same minimum toolchain build in openslug-3.10-beta release that can be used for the non-release toolchain Feb 17 18:27:05 good news Feb 17 19:48:13 03marceln * r5518 10optware/trunk/ (make/scponly.mk sources/scponly/): Updated to version 4.6 and the new template. Feb 17 19:51:06 marceln: any idea why scponly is marked broken? Feb 17 19:52:41 No. Feb 17 19:53:06 I was able to build it. Feb 17 19:53:30 But after the update i will mark it ready for testing. Feb 17 19:54:29 The only reason i haven't done this yet is that i also want to test the package. Feb 17 19:54:51 alright, sounds good Feb 17 19:55:02 So far i have build a chrooted environment on the slug and it works. Feb 17 19:55:14 Although "rsync" didn't. Feb 17 19:55:53 No i gone test it on the wl-hdd (mipsel uclibc) Feb 17 19:57:50 could you also update tftp-hpa when you get a chance? Feb 17 19:57:58 Ofcourse. Feb 17 20:03:28 * CoreDump|home needs a seconds slug Feb 17 20:07:10 * mwester snorts, remembering that such thoughts were the slippery slope that led to him now having four slugs, and two dsmg600s... Feb 17 20:08:00 hehe Feb 17 20:08:12 4 slugs? nice Feb 17 20:08:29 Do they reproduce? Feb 17 20:08:41 marceln: yeah Feb 17 20:08:48 and they are kinda addictive Feb 17 20:09:32 yeah, two in production (netslug -> DHCP + pxeboot + Fedora install server + ftp + nfs + etc, and slugger -> dual 250GB drives for home network storage). Feb 17 20:10:23 And "unslung" which is my dev slug, and refslug which runs the stock linksys fw as a reference system (and occasionally doubles as a testslug. Feb 17 20:10:29 well, I _do_ need a stable NAS box and developing for the slug makes that kinda impossible. Feb 17 20:10:41 mwester: =D Feb 17 20:10:48 eno: tftp-hpa upgraded to 0.4.8 but i didn't see any important changes. Feb 17 20:10:53 03marceln * r5519 10optware/trunk/make/tftp-hpa.mk: Updated to version 0.4.8 Feb 17 20:10:53 The dsmg600s are both doing SlugOS 4.3-beta development and testing right now. Feb 17 20:10:56 I could see me using 3 of them heh Feb 17 20:11:08 They're cheap -- cheaper than the disk drives you plug into them! Feb 17 20:11:17 yeah Feb 17 20:11:23 my thoughts exactly Feb 17 20:11:37 and they don't suck lots of juice Feb 17 20:11:50 And no noise Feb 17 20:12:00 * CoreDump|home googles for dsmg600 Feb 17 20:12:13 (plus, they fit in a suitcase for when I travel on business trips) Feb 17 20:12:56 DSMG600 - D-Link NAS - the old Rev A units are IXP-based. The current Rev B units are MIPS (I think). Feb 17 20:13:01 hey the sdmg600 is are kinda pricey Feb 17 20:13:19 I picked up the Rev A's for <100USD each Feb 17 20:13:34 the new one is ~200 ¤ Feb 17 20:13:55 Are you perhaps looking at the -323 unit? Feb 17 20:14:22 "D-Link DSM-G600 Network Media Storage" Feb 17 20:14:30 200 euros Feb 17 20:14:53 Yikes! quite the markup. Retail at the CompUSA store across the street here is $129 USD Feb 17 20:15:05 (no disks included) Feb 17 20:15:55 ouch Feb 17 20:16:36 well, I guess I'll get two more slugs in the next month or so. Should last me some time heh Feb 17 21:26:53 hope there's linux on the mips one too Feb 17 21:27:06 if there's one thing i know about dlink -- they can't write drivers/firmware worth a damn ;) Feb 17 21:27:35 dammit.. I only changed the id3lib.tar.gz package in downloads, make id3lib - and it wiped out my toolchain and starts to build the toolchain all over again.. the 4th time today Feb 17 21:27:41 any ideas what could be causing this? Feb 17 21:54:45 03marceln * r5520 10optware/trunk/Makefile: scponly works for the slug (wl-hdd builds but doesn't work) Feb 17 21:56:15 03marceln * r5521 10optware/trunk/Makefile: Scponly ready for testing on the slug Feb 17 22:21:34 03bzhou * r5522 10optware/trunk/Makefile: scponly promoted Feb 17 22:36:58 yay! My second slug will be ariving next week Feb 17 22:37:30 cool Feb 17 22:37:35 the more, the merrier ) Feb 17 22:37:36 ;) Feb 17 22:37:39 slugs 3 and 4 will stay at a friend's house Feb 17 22:37:57 connected via vpn Feb 17 22:38:07 nice Feb 17 22:38:36 what exactly does it take to put together a native package? Is there an automated native build system out there? Feb 17 22:38:56 I'm not sure the native builder is active anymore Feb 17 22:39:06 There was a slug dedicated to doing native builds Feb 17 22:39:24 AFAIK, it hasn't been updated for a long time Feb 17 22:39:26 since I seem to be aquiring so many of the damn things, I figure I'd put together a distributed native build mechanism Feb 17 22:39:41 distcc Feb 17 22:39:50 saw that mentioned Feb 17 22:39:55 how comprehensive it it? Feb 17 22:40:25 does it just work from a list of packages and allocate to a list of slugs? Feb 17 22:40:40 there is someone running gentooslug, build with distcc Feb 17 22:41:22 I may play with it a bit just as soon as I get done with logrotate and (next) ntpclient Feb 17 22:41:44 hillct: btw, you might want to send the .bb to the mailinglist so it's added to the repo Feb 17 22:41:52 isn't ntpclient already in the repo? Feb 17 22:42:04 not in OE Feb 17 22:42:08 it's in optware Feb 17 22:42:13 same with logrotate Feb 17 22:42:36 oe only has logrotate-script which is IMHO rather weak Feb 17 22:43:07 isn't ntpdate enough? ;) Feb 17 22:43:19 I'll send the .bb files off to the list just as soon as I've tested them Feb 17 22:43:45 there's rdate that's part of busybox Feb 17 22:43:47 sometimes i build optware packages natively, with or without distcc Feb 17 22:46:40 I didn't see ntpdate in the repos Feb 17 22:47:18 ntpdate is part of the xntpd package which is bloated Feb 17 22:47:26 ntpclient is nice and lean Feb 17 22:48:08 it operates as a demon rather than being fired off via cron Feb 17 22:48:18 ah Feb 17 22:48:22 tried openntpd? Feb 17 22:48:31 IIRC it's way smaller than xntpd Feb 17 22:48:59 hasn't looked at openntpd Feb 17 22:49:18 guess I'll compare with ntpclient once I have it going Feb 17 22:49:20 I spent a couple of weeks tearing my hair out over openntpd ;) Feb 17 22:49:48 do you feel strongly that it's the way to go? Feb 17 22:49:57 I'm not interested in reinventing the wheel Feb 17 22:50:40 Well, it's small, fast and works very nice. Except if your clock is *way* off, or is drifting too fast for it to cope (which used to be the case with earlier firmware, fixed now) Feb 17 22:52:38 I'll see about building ntpclient just because I need the practice with bitbake. If it gets too hairy, I'll just go with openntpd Feb 17 22:53:23 cool Feb 17 22:53:34 let me know what comes out of it Feb 17 22:53:44 I'll push it to the stable feed Feb 17 22:54:54 can I safely remove all the build directories (in slugosbe/tmp) other than cross, after I build the toolchain? Feb 17 22:55:05 I could use the space Feb 17 22:56:02 you can remove "work" Feb 17 22:56:18 You won't save that much space with the rest of the dirs Feb 17 22:56:29 (and I don't know which of the other dirs are safe to remove) Feb 17 22:56:42 do not remove the stamps dir ;) Feb 17 22:56:52 or staging Feb 17 22:57:08 deploy is safe Feb 17 22:57:25 rootfs should be safe too Feb 17 23:17:23 03bzhou * r5523 10optware/trunk/ (make/asterisk14.mk sources/asterisk14/nv.patch): Feb 17 23:17:23 asterisk14: worked around use_ast_mutex_t_instead_of_pthread_mutex_t compilation error on uclibc Feb 17 23:17:23 by changing include order, see http://bugs.digium.com/view.php?id=5558 Feb 17 23:17:23 exclude .svn when building tarball next time Feb 17 23:18:10 what are your thoughts on the Fedora/Redhat style cron run-parts mechanism? Feb 17 23:18:49 seems it makes for far more convenient crod job addition by other packages Feb 17 23:43:55 NAiL: I'm going to update some of the build enviroment docs. Is there a bitbake command that will effectively clean out the work directory only? or will we be better off just saying to delete the work directory after the toolchain build in order to safely recover 1.7GB on the build host ? Feb 18 02:05:46 oleo: ping? Feb 18 02:06:40 I've finally managed to get a complete toolchain to work for the ts101. But it looks like it'll be a fair bit of work to do the changes to buildroot.mk to get it to work. Feb 18 02:06:48 Mind helping me out a bit? Feb 18 02:07:12 morning rwhitby Feb 18 02:14:07 03nail * r5524 10optware/trunk/sources/buildroot/ (3 files): buildroot: Remove old ts-101 stuff **** ENDING LOGGING AT Sun Feb 18 02:59:57 2007