**** BEGIN LOGGING AT Mon May 02 23:59:56 2005 May 03 00:01:23 morning all May 03 00:23:43 morning. May 03 00:25:23 morning May 03 01:00:26 test May 03 01:54:33 Quagga (http://www.quagga.net/) package has been promoted to wiley. Quagga package has been created by llagendiijk, a member of Chupa Forum. May 03 02:05:40 I cant seem to load that page. May 03 02:06:09 oh, nevermind. May 03 02:06:13 I was trying .com May 03 02:14:14 package needs some improvment but it installs quite correctly May 03 02:14:27 rwhitby around? May 03 02:23:48 I would like to have opinions on running daemons whith unprivileged account on slug/wiley. We allready discussed this a bit with jp-30 yesterday but I still not have a clear view on this point. Should packages have to create their needed user ? should end-user have to create them ? should package have to try running as root ? ... May 03 02:25:17 jp30-work : as eggdrop refuse to launch under root, even when suid nobody May 03 02:27:56 dyoung: pong May 03 02:28:35 jf-work: if we want to use new users, then we need to create and deploy useradd and userdel working for both Unslung and WileyWare May 03 02:28:58 they would need to be in a package, and then packages that needed to use them would depend upon that package May 03 02:29:20 agree May 03 02:29:56 rwhitby, which is the right PHP package to install for future compatibility with the Neuston? May 03 02:29:58 once that infrastructure is in place, then we could convert packages over to using it (and put some policies and username registry in place). May 03 02:30:37 dyoung: I think Apache (threaded) is safest, as a lot of PHP packages seem to depend on Apacheisms in the CGI variables. May 03 02:31:02 now that jp30 has the sweet threading working, then that's a viable and preferred solution I think. May 03 02:31:40 now, permanently adding a user on wiley require some more step, like slug I guess May 03 02:31:54 okay, so php-apache after apache knows to do The Right Thing (tm) ? May 03 02:32:54 given that, I was wondering yesterday how we can have conditionnal in postinst to add a user in wiley / in slug as the procedure may be different May 03 02:34:01 we would need to make sure that the procedure is the same, by creating the same useradd/userdel package which does the same thing (perhaps in a different way) on Slug and Wiley May 03 02:34:27 we want the postinsts to be compatible, calling programs that do The Right Thing on each platform May 03 02:34:44 we will need to coordinate with Oleg on this too May 03 02:40:12 user* are in busybox May 03 02:40:49 yah I thought so too May 03 02:40:58 I mean, I agree - I think they are May 03 02:41:18 yep, but we certainly don't want packages to depend on busybox May 03 02:41:26 cause it causes no end of problem May 03 02:41:29 s May 03 02:41:30 they just have to be compiled in and linked May 03 02:41:52 isn't busybox essential? May 03 02:42:19 not anymore May 03 02:42:40 for example the native devel env requires removing it and installing coreutils May 03 02:42:44 depends whether you're talking about the linksys busybox (which we don't touch, cause we don't have source), or slingbox (which is the busybox that we add to the firmware), or the busybox package (which we don't like using) May 03 02:43:10 we don't like using the busybox package for the reasons that jacques states May 03 02:43:30 OK May 03 02:43:33 if you have a hard disk attached, then there is no reason to use busybox when you could be using the real full-featured programs May 03 02:43:38 and slingbox = busybox? May 03 02:43:55 adduser is perl based May 03 02:43:57 no ? May 03 02:44:11 jf-work: no way can we have an adduser that is perl based May 03 02:44:14 maybe one somewhere is, but not normally May 03 02:44:22 the debian one is. May 03 02:44:42 that has always made my head spin. useradd vs adduser. May 03 02:44:47 we need separate, targetted, adduser and deluser, so that packages can depend on it, and it will be installed without filling anyone's flash disk. May 03 02:45:11 on my gentoo system, adduser is a symlink to useradd May 03 02:45:21 Youre lucky. May 03 02:45:25 and useradd comes from the shadow package (on my system) May 03 02:45:29 (when I say "adduser" or "useradd", I have no idea which one of them I am talking about - so I'm using them both as a general concept) May 03 02:45:33 it seems to be distro dependendant. May 03 02:45:34 afaik adduser is depricated May 03 02:45:40 dyoung agreed May 03 02:45:49 either way, it makes my head spin. May 03 02:45:51 D|dge : yep May 03 02:46:02 useradd is binary May 03 02:46:08 bbiab May 03 02:46:26 ok May 03 02:47:12 useradd is part of shadow password suite. Does we have shadow password on boths boxes ? ;) May 03 02:47:56 bbl May 03 02:50:35 jf-work, I believe useradd can come from multiple packages - if one is not using shadow passwords, it will not come from shadow May 03 02:52:10 no /etc/shadow on either WileyWare or Unslung May 03 02:54:02 i think useradd can come from shadow, coreutils, login, ... May 03 02:55:57 and we can always build a package from any one of those that *only* contains useradd and userdel May 03 02:56:03 (and groupadd and groupdel) May 03 02:56:13 yep May 03 02:57:15 so we just need to find the one which builds on both Unslung/Slug and Unslung/Wiley May 03 02:58:14 looking at shadow now May 03 02:59:23 ok maybe coreutils is not one of the packages May 03 03:00:28 Okay, I think that was all of the stuff thaht was on prodslug1; now moved to prodslug2. May 03 03:00:36 tinylogin is one I was thinking of May 03 03:00:40 jacques: having useradd ? May 03 03:00:42 http://tinylogin.busybox.net/TinyLogin.html#commands May 03 03:00:47 jf-work, yeah May 03 03:04:34 I think util-linux too May 03 03:04:37 ok not util-linux :-\ sorry May 03 03:04:43 but the pam stuff also has its own versions May 03 03:05:23 jacques: maybe shadow ? ;) May 03 03:05:33 jf-work, shadow does for sure May 03 03:06:29 but I think there has to be some coordination between the login and useradd/etc versions, because if you use the shadow useradd, and don't have a shadow-aware login... May 03 03:13:40 PAM can Bite Me. May 03 03:13:47 (sorry I had to get that in) May 03 03:14:57 yeah, I just don't see the need for pam's complexity on the slug May 03 03:15:36 now shadow I could probably be talked into, tho it would probably require a different login, which might not be as optimal as tinylogin resource-usage-wise May 03 03:17:56 vim still does not work for me. (console based) May 03 03:18:34 what's the error? May 03 03:19:11 missing X lib May 03 03:19:33 --without-x may not be enough May 03 03:19:53 i'll try: --enable-gui=no May 03 03:20:14 yeah, it's picking up X from somewhere May 03 03:23:20 strange. i built vim myself. that one works for me May 03 03:23:30 no changes May 03 03:24:02 (no X installed) May 03 03:24:54 the --without-x should be enough when X is installed (but it seems to be broken), when no X is installed naturally it doesn't autoconfig it in :-) May 03 03:26:35 ~lart dyoung-logger May 03 03:26:35 * jbot breaks out the Hoover and sucks up dyoung-logger May 03 04:34:36 03rwhitby * 10unslung/make/puppy.mk: Updated to puppy 1.9 May 03 04:43:14 03rwhitby 07 * r1.3203.1.1 10openembedded/packages/puppy/puppy_1.9.bb: Updated puppy to version 1.9 May 03 04:44:53 jp30-work: can you please push that new version of puppy? May 03 04:49:59 jf-work: can you please push that new version of puppy for wl500g May 03 04:50:05 I'm pushing it for openslug May 03 04:54:47 rwhitby: okay May 03 04:55:15 jf-work: any luck finding a useradd candidate source? May 03 04:56:39 rwhitby: hmm. looking at shadow but also have *real* work to do :( May 03 04:57:27 yeah, I know the feeling ... May 03 04:57:50 rwhitby: you don't know my boss... May 03 04:58:20 03rwhitby * 10unslung/ (4 files in 2 dirs): Added startup scripts for ftpd-topfield May 03 04:58:20 sounds like I don't want to either .... May 03 05:20:21 jf-work: re the xml2-config path May 03 05:20:31 it was in staging/bin - it should be in staging/opt/bin May 03 05:23:50 ok, I see that you've fixed that in php-thttpd.mk - cool. May 03 06:00:25 jp30-work: we need to decide whether the foo-config scripts go in staging/bin or staging/opt/bin, and make all the packages consistent. May 03 06:00:49 We had originally had stuff in staging/bin, but then changed everything to staging/opt/... to make it easy to write the staging target. May 03 06:01:15 But if there is a good reason why these scripts should be in staging/bin rather than staging/opt/bin, then I can be convinced otherwise. May 03 06:30:10 <[g2]> rwhitby, hey! May 03 06:30:18 hey May 03 06:44:06 dyoung: ping May 03 06:44:17 (or anyone else who has nfs mounting working) May 03 06:48:34 yes? May 03 06:49:01 ever seen this error? May 03 06:49:03 <44>May 3 22:28:13 klogd: nfs warning: mount version older than kernel May 03 06:49:03 <44>May 3 22:28:13 klogd: NFS: mount program didn't pass remote address! May 03 06:49:10 yes May 03 06:49:23 on openslug May 03 06:49:34 Running "mount -t nfs 192.168.1.1:/tmp/harddisk /mnt" (mounting /tmp/harddisk from a wl500g to an nslu2) May 03 06:49:34 no problem on unslung May 03 06:49:57 that was running on unslung, mounting from a server on wl500g May 03 06:53:39 does exportfs work? May 03 06:53:52 on the server? May 03 06:53:56 yes May 03 06:54:09 does it show the mount points? May 03 06:54:10 [root@wl500gx xinetd.d]$ exportfs May 03 06:54:22 -> /tmp/harddisk 192.168.1.0/24 May 03 06:54:35 (omit the arrow) May 03 06:55:02 . /share/hdd/data 192.168.1.0/255.255.0.0(rw,sync,no_root_squash) May 03 06:55:04 (/tmp/harddisk is the wl500g equivalent of /share/hdd) May 03 06:56:29 does statd run, or mountd or nfsd? May 03 06:56:37 or -> and May 03 06:58:07 on server, we have [nfsd], [lockd], [rpciod] May 03 06:58:09 no statd May 03 06:58:43 same on client May 03 06:59:25 installed nfs-utils? May 03 06:59:42 oh, mountd is running on client May 03 06:59:52 and server too May 03 07:00:02 and statd and portmap on client May 03 07:00:20 and portmap on server, but no statd on server May 03 07:04:09 ok, worked it out in one direction - can now mount from slug server to wl500g client (had to insmod nfs on client( May 03 07:10:23 but now puppy has errors on the wl500gx :-( May 03 07:10:35 back to running puppy on Unslung on the slug May 03 07:11:47 ... where it runs flawlessly. May 03 07:13:28 got tinylogin READY_FOR_TESTING on slug and promoted on wiley (Jacques was right ;)) May 03 07:13:57 what else does it include? May 03 07:18:22 su (whic conflicts with coreutils), addgroup, adduser, delgroup, deluser, login May 03 07:18:31 passwd May 03 07:18:53 and getty, sulogin under sbin/ May 03 07:19:04 which of those will conflict with existing programs on nslu2 and wl500g ? May 03 07:20:28 login and passwd are linked to original busybox on wiley May 03 07:21:29 I would suggest we rename the package, and only include {add|del}{user|group} May 03 07:21:51 I think we could decide not to install all the symlink of tinylogin but just *group and *user May 03 07:22:28 Hmm - there is a /bin/tinylogin on nslu2 May 03 07:22:45 ok. we can rename it tho. May 03 07:23:23 are {add|del}{user|group} all just symlinks to the one exe? May 03 07:23:24 waht name ? May 03 07:23:31 rwhitby: yes May 03 07:23:49 Let's call it adduser, and then link the others to that. May 03 07:23:57 what name do you preconize for package/exe ? May 03 07:23:58 ok May 03 07:24:14 yeah, "adduser" for package name too May 03 07:25:26 Of course we need to work out how to actually make the changes stick for both the nslu2 and the wl500g. May 03 07:25:35 yep May 03 07:30:16 ByronT: how's the click-through plans going? May 03 07:31:36 maybe today? I need to check with Tom to see if the webpage is ready for the legaleese to be dropped into it May 03 07:31:58 did you see who joined the yahoo group yesterday? May 03 07:32:27 who? May 03 07:32:37 gelloglue - RR May 03 07:33:12 morning, can anybody tell me, which behaviour is recomment, when a package need new users... May 03 07:33:41 caplink811: you will need to depend on the "adduser" package May 03 07:33:47 should the package postinst add the needed users, or shall i leave the user alone, giving a hint in postint, to add users manually May 03 07:33:50 rwhitby: yeah... May 03 07:36:28 caplink811: jf-work is working on an "adduser" package which will provide the {add|del}{user|group} commands May 03 07:37:16 i've read it, but, whats next, package adds the user, should the package save the /etc/passwd, or give the user a hint to do it? May 03 07:38:50 or adds a new user silently and leave the user alone? dunno, whats the right way is.. may be philosophic ;-) May 03 07:38:56 on package startup (i.e. rc script), the package should check for the user, and re-add it if it has disappeared May 03 07:39:47 rwhitby-away: can't we have a third script adding permanently the user ? May 03 07:40:45 yes, that is another option. I'm not sure how we can do that robustly to support Unslung 1.x, 2.x, 3.x and 4.x and also WileyWare. May 03 07:43:12 adding it each time on boot at least is known to work on all versions of both streams of firmware May 03 07:43:25 03jeanfabrice * 10unslung/ (4 files in 3 dirs): Demote tinylogin and promote adduser for wiley, ready_for_testing for slug May 03 07:43:34 ty CIA May 03 07:44:02 should you take a look at it May 03 07:44:25 building for nslu2 now ... May 03 07:45:09 now we just need some guidance in the BestPractises page for how to use it ... May 03 07:48:57 03rwhitby * 10unslung/Makefile: Promoted adduser for nslu2 cross. May 03 07:49:04 (loud silence) May 03 07:50:45 jf-work: can you remove the obsolete tinylogin.mk ? May 03 07:50:47 morning May 03 07:51:01 may be it reasonable to have only one init-script which checks and adds (if needed) user for all packages which have created users in the past? May 03 07:52:26 rwhitby-away; sure May 03 07:52:47 caplink811: since any package that needs a user is very likely to be a daemon, then it will need an rc script anyway, and it's just a one line addition to check for and add the user. May 03 07:53:19 caplink811: how would the single init script know which users it needs to add ? May 03 07:54:06 you would need some sort of /opt/etc/adduser.d directory in which packages drop flag files to request users be created. May 03 07:54:23 rwhitby-away: may be a file e.g. needed_users.conf, where all packages adds users, groups, and the "super"addscript goes along and read from this file and check, and add users May 03 07:54:38 can't have multiple packages writing to a single file May 03 07:54:56 03jeanfabrice * 10unslung/make/tinylogin.mk: Bye bye tinylogin. Replaced by adduser May 03 07:55:03 (that's why we have xinetd.d and cron.d now ...) May 03 07:55:38 nope, meaning via postinst, where the package adds the needed users/groups to one conf file May 03 07:55:50 and what about a neededuser.d ? May 03 07:56:25 seems like a lot of work to save a single line in an rc script ... May 03 07:56:33 :-) May 03 07:58:14 and if you put it in the rc script, then you don't need to put it in the postinst script :-) May 03 07:58:51 thats right May 03 07:59:32 that mean that even package not needing a rc script should have one ? May 03 07:59:53 so, that is my understanding May 03 08:00:03 if it wants it's own user, yes. Can you think of a non-daemon package which needs a user? I can't. May 03 08:00:15 e.g. clamav May 03 08:01:15 screen May 03 08:01:50 in fact, for the wl500g, it will need to be in the rc script, cause /etc/passwd is not in jffs2, and there is no way that Unslung packages are going to mess with flashfs ... May 03 08:02:19 the /etc/passwd file is generated on each boot for wl500g from the web interface nvram settings May 03 08:04:36 (wiley can have /etc/passwd symlink to a permanent file on FS, symlink recreated at boot time) May 03 08:05:04 anyway, you're right rwhitby-away. May 03 08:06:09 ~emulate rwhitby May 03 08:06:10 Well, my work here is done. May 03 08:06:15 night all May 03 08:07:49 jf-work: hm, adduser calls passwd... thats bad if i want automate user creation :-( May 03 08:09:53 03rwhitby * 10unslung/Makefile: Promoted quagga for nslu2 cross. May 03 08:12:12 perhaps we can tweak the binary ? ;) May 03 08:21:26 jf-work: yeah, a shell script which returns 1 as /opt/sbin/passwd let adduser create a user with x passwordbut PATH must point to /opt/sbin first May 03 08:24:07 so that's a non standard solution May 03 08:28:49 caplink811: it seems easy to change the binary not to call an interactive 'passwd'. will check this May 03 08:32:04 jf-work: k, but if a user wants to use adduser for its primarily task, there should be two versions in the package, one which calls passwd, and one which do not?!? May 03 08:33:40 caplink811-away: he can use passwd, no ? May 03 08:53:08 jf-work: sure, but he must realize that he to do it, so, may be a way, that we have an original adduser, and a tweak one, called e.g. rcadduser May 03 10:07:37 03jp30 * 10unslung/make/php-apache.mk: fix PHP_LIBXML_DIR setting May 03 10:36:35 hi, just a question about unslung 4.x alpha: May 03 10:36:49 is 4.15 the current version? May 03 10:37:09 this is the latest I can se in tehe downloads section May 03 10:37:36 s/se/see/ + s/tehe/the/ May 03 11:28:09 hello! May 03 11:28:27 one short question (I'll update the wiki afterwards, really): May 03 11:28:55 is it possible to use the power-button to "shutdown -h" the device (I use OpenSlug)? May 03 11:37:08 03jeanfabrice * 10unslung/ (make/adduser.mk sources/adduser/adduser.c.patch): Remove interactive passwd call from adduser, cleanup makefile and bump IPK_VERSION May 03 11:42:19 jp30-work: ping May 03 11:43:21 hi, pTweety May 03 11:43:37 Hi May 03 11:43:51 today my DevelSlug arrived :-) May 03 11:44:31 good. what do you plan to devel with it? May 03 11:44:34 But before instaling al the required software I was thinking about a meta-package that has al the devel-packages of the native buld machine as dependencie May 03 11:44:45 hmm... that's a good idea May 03 11:45:06 ipkg install unslung-devel should be al you need as a developer May 03 11:46:06 If you want me to create such a thing I#l do it right now May 03 11:46:40 (grr, my keyboard reacts quite slow these days) May 03 11:46:43 anyone in here knows how to convert nslu2 packages to run on the Asus WL-500g ? May 03 11:46:53 (I know it can be done, but dunno how :) ) May 03 11:47:26 pTweety, here's one consideration: most problems of mismatch between the official build slug and individual developer are because developers have too many packages installed, rather than too few. May 03 11:48:39 macsat, cross-compiled unslung packages should require no general "conversion" May 03 11:48:48 I was thinking of exaactly the packages mentioned on http://www.nslu2-linux.org/wiki/Info/NativeBuildMachine May 03 11:49:14 macsat, if there's a specific package you are struggling with, the changes needed are likely to be specific to that package May 03 11:49:31 pTweety, it can't do any harm, so why not go ahead and create it May 03 11:49:44 :-) May 03 11:50:12 you could also put a note on the wiki May 03 11:50:18 jp30-work: are you saying that in general I should be able to install any ipkg package created for the NSLU2, directly on my WL 500g ? May 03 11:50:24 jp30-work, is unslung 4.15 the current one? May 03 11:51:15 As far as I know, the NSLU2 is "big endian" while the WL-500g is "small endian" (spelling might suck?!)....wouldnt that mean that some kind of "conversion" is needed ? May 03 11:52:21 pTweety, you'll need to ask someone more closely involved with firmware development than i am May 03 11:52:55 macsat, you can't install binary ipks intended for the nslu2 on the wl500g May 03 11:53:02 oh sorry. I thought that you are also involved in 4.x May 03 11:53:33 well, i would be if i had time May 03 11:53:49 ok jp30 :) May 03 11:54:29 macsat, you have asked about mailservers for the slug some days ago May 03 11:54:38 I do know though, that some of the binaries that are available to the wl500g are kindda "made from" the NSLU2 packages, since they are filled with "NSLU2" references in readme files, config files and so on :-) May 03 11:54:44 yeah pTweety, correct May 03 11:54:55 still interested in a mailserver :) May 03 11:55:21 have you seen cyrus-imapd, postfix and fetchmail? May 03 11:55:40 they are running well on my prodslug May 03 11:56:43 sounds like MTA's and a pop3 "client" :-) May 03 11:56:53 What I would like is a pop3 server May 03 11:57:13 hmm...come to think of it, doesnt postfix include pop3 ? :-) May 03 11:58:23 macsat, you can find wl500g packages at http://ipkg.nslu2-linux.org/feeds/unslung/wl500g/ May 03 11:58:47 I know jp30 :) May 03 11:58:58 Just a LOT fewer packages than for the NSLU2 :-) May 03 11:59:59 macsat, cyrus-imapd ist a pop3 server as well May 03 12:00:07 pTweety: 4.15 is the latest binary "release", maybe newer sources in the repo May 03 12:01:03 ok pTweety, nice May 03 12:02:05 are they "kindda easy" to setup ?.....I am not a linux novice, but I do have respect for setting up mailservers....have always been amongst the hardest service to setup May 03 12:04:28 well, it's not to hard to set it up. I'll create some wiki-page about it. May 03 12:04:34 give me a moment May 03 12:08:48 great pTweety :-) May 03 12:11:28 hmmm....I cant seem to see postfix nor cyrus-imapd in the "wl500g repository" May 03 12:15:08 ah, you are not talking about nslu2 May 03 12:16:43 well, then you should wait for xmail May 03 12:17:17 this package is mentioned as WL500G_PACKAGES_THAT_NEED_FIXING, but maybe someone takes care May 03 12:18:23 yeah, I am on wl500g :) May 03 12:18:45 Yes, I remember the xmail package was once available but didnt work :) May 03 12:19:04 Would be cool if someone (JeanFabrice ;-) ) could fix it :) May 03 12:19:37 postfix doesn't build right now. I do not remember why May 03 12:20:06 jp30-work: seen our latest dependent user creator package ? May 03 12:20:42 hmmm ok, nice to know May 03 12:20:52 jeanfabrice, have you tried to build postfix the last days? May 03 12:21:15 pTweety: last time I tryed it it seems you were working hard on it May 03 12:21:21 Trying now May 03 12:21:25 thanks May 03 12:21:35 as I wrote on the chupa forums, I am planning to try to give you a hand with new packages, so nice to know that postfix might not be worth the efford ;-) May 03 12:21:42 but then cyrus-imapd is stil mising May 03 12:22:13 jeanfabrice, have you build perl for wl500g? May 03 12:22:23 Did anyone ever try to build qmail for the NSLU2 or the wl500g ? May 03 12:22:29 microperl works fine pTweety May 03 12:22:40 pTweety: found the error. pcre does not compile as far as I remember May 03 12:22:40 ..and JF has build that one May 03 12:23:24 macsat: about qmail : this could be your first package ;) May 03 12:24:11 hehe.....might be a little complex....as far as I remember it is not that simple to compile on a i386 :) May 03 12:24:23 But I could try I guess :) May 03 12:24:45 What about mysql, any particular reason why that is not done (and postgresql is) ? :-)) May 03 12:25:03 why not ? and ask here for help. there're bunch of wizards there (not me) May 03 12:25:32 ok May 03 12:25:36 cool May 03 12:25:53 I hope to get some time to look into it Thursday/Friday or this weekend May 03 12:26:29 Cant really fit it in on working days...:-/ May 03 12:28:29 pTweety: while building pcre : /opt/brcm/hndtools-mipsel-uclibc-3.2.3-full/lib/crt1.o(.text+0x40): undefined reference to `__uClibc_start_main' May 03 12:29:51 hm, I fear that I can'T help you with this. Never seen it before, sorry May 03 12:30:03 pTweety: np May 03 12:31:03 pTweety: pcre has conditionnal for native/cross compilation ? May 03 12:33:04 yes, I remeber having problems with libtool May 03 12:33:51 so do I May 03 12:37:52 I think that my problem was with --mode=link May 03 13:11:37 jeanfabrice, pTweety, looking at the pcre error you quote - it looks like the result of trying to link a standalone executable from a bunch of objects that should have been made into a shared lib May 03 13:11:51 ...it would be interesting to see the command line that produced it May 03 13:12:19 ...assuming you haven't fixed the problem already May 03 13:13:45 jp30-work: I'm far from having fixed this May 03 13:14:03 do you want the libtool command line or the gcc command line ? May 03 13:14:31 the gcc command line to start with May 03 13:14:59 the biggest so.. May 03 13:15:25 you can leave out the objects May 03 13:15:36 i'd like to see the flags May 03 13:50:22 rwhitby-asleep: adduser seems broken as it doesn't create correctly the group name in /etc/group May 03 14:59:20 jp30-work: ping May 03 15:02:03 jp30-work: saw your email - very good summary of the staging/bin situation May 03 15:04:58 is adduser even supposed to create a group? May 03 15:05:09 I thought that had to be done separately May 03 15:05:54 for example, the useradd from shadow man page says: May 03 15:06:01 The group name or number of the user's initial login group. The group name must exist. A group number must refer to an already existing group. May 03 15:06:10 hi, rwhitby. pleased you liked the email. any views on how we should proceed? May 03 15:07:06 yep, 1) moving to 3). May 03 15:07:19 So I will stop breaking all the package :-) May 03 15:07:27 packages May 03 15:08:32 Let's make staging/opt things that are meant to be executed on the target (like target libraries), and staging/bin to be scripts that are meant to be executed on the host during the build process. May 03 15:12:32 rwhitby, I think that's a reasonable way to start. May 03 15:15:16 do you think it would make sense to add "PATH=staging/bin:$$PATH" to TARGET_CONFIGURE_OPTS? May 03 15:15:50 yeah, we should do that May 03 15:16:06 ...that would help to discourage developers from manipulating PATH themselves and doing it wrong (i have seen many errors here) May 03 15:16:42 good stuff. thanks for clarifying the issue, and sorry about stuffing around with things I didn't fully understand. May 03 15:19:31 well, i don't think i fully understood the history of staging/bin either. all i knew was that its contents were all de facto executable on the host, so i started putting config scripts in there May 03 15:19:50 ...following the example of xml2-config May 03 15:20:21 actually, it looks like in most case, xml2-config doesn't need to be in the path to configure packages which use it May 03 15:22:40 ok, can you write this up on the best practices page so we don't have the same thing happening again in another six months :-) May 03 15:27:38 will do May 03 15:28:17 ok, do I need to undo any of the incorrect changes I made? May 03 15:30:28 bbiab May 03 15:32:45 rwhitby, "make distclean && make" works OK at the mo, if more cleaning up is needed, I'll do it as and when. Actually there was already a BestPractices entry. It May 03 15:33:10 ...It's the last rule under "cross compilation". I'll think about whether it should be made clearer. May 03 15:42:42 unslung-4.17-alpha.bin is in downloads. Beware that this is a *real* alpha, in that we have completely reworked the boot process, so be prepared to have to reflash if you try it. May 03 15:43:02 glc has tested it, and I will be testing it tonight probably. May 03 15:59:11 ~seen dyoung May 03 15:59:17 dyoung <~dyoung@dyoung.nslu2-linux> was last seen on IRC in channel #openjtag, 9h 47m 21s ago, saying: 'hi'. May 03 16:02:14 New pictures of Slug-Central are up at http://nslu2-linux.thestuffguy.com/gallery/slug-central May 03 16:03:43 Nothing really new - just a reorg of existing hardware. May 03 16:04:05 looks good. May 03 16:04:15 ByronT-Away, ping May 03 16:04:20 But it now has 100Gb on-line storage between the slug and the wiley. May 03 16:06:42 whoa...lots of Toppy programs eh? May 03 16:07:06 dyoung-logger, when your master arrives please give him a message that I am looking for him. May 03 16:07:13 Not yet, haven't started transferring anything yet. May 03 16:07:48 But it's now set up so that dyoung will be able to sftp Pingu Episodes :-) May 03 16:08:17 ah...Pingu! Justin likes that. May 03 16:08:23 (and we don't have it. May 03 16:08:45 bbiab May 03 16:08:49 k May 03 16:09:05 actually, that's a back later - need to head off to work. May 03 16:09:53 cya May 03 16:46:34 anyone here? May 03 16:46:52 i am having a problem with my new nslu2, i can't find the ip of it... May 03 16:47:19 and i scanned my entire network segment May 03 16:47:21 krod, it is probably 192.168.1.77 May 03 16:48:25 .. May 03 16:48:25 it is May 03 16:48:33 why does it default to that? May 03 16:48:47 i got scared for a second..i just loaded the firmware on it.. May 03 16:49:23 hm, i wounder how it got changed May 03 16:49:36 its using the fixed ip of 192.168.1.77 May 03 16:49:52 that is the address that every new nslu2 has May 03 16:50:19 i believe this is documented in the linksys documentation that comes with the nslu2 May 03 16:50:36 i didn't get the docs, i bought it on ebay May 03 16:50:48 and it had the ip of 192.168.1.102 before... May 03 16:51:44 one more question, how do i keep telnet enabled, because every time it gets rebooted i have to enable it again via web interface May 03 16:54:12 don't use telnet, install openssh or dropbear instead May 03 16:54:36 al'right May 03 16:54:54 and if i installed a package, how would i uninstall that package? May 03 16:56:02 ipkg remove packagename May 03 16:56:29 if you run ipkg without any arguments it will give you more help May 03 16:59:21 thanks May 03 17:01:18 what does busybox use, to load all the services at boot...it dosn't seem to use inetd May 03 17:04:51 krod, have you looked at the wiki? May 03 17:06:20 sorta :) May 03 17:06:26 i will read it fully now.. May 03 17:06:33 thanks for your help May 03 17:11:09 krod, do feel free to ask questions, but bear in mind rules #1 and #2 - they are explained on the wiki front page May 03 17:16:19 :) May 03 17:19:26 do we have a way of asking jbot to politely remind visitors of rules 1 and 2? May 03 17:22:41 ~nslu2-rules May 03 17:22:42 rumour has it, nslu2-rules is 1) Those who ask the questions, update the wiki, 2) Those who complain about the doco, update the wiki, 3) Friends don't let friends flash custom firmware without confirmed RedBoot access, 4) Friends don't let friends flash custom bootloaders without confirmed JTAG access. May 03 18:10:41 there should be a search on the wiki May 03 18:11:02 there is May 03 18:11:33 yeah, i say it just as i pressed enter... May 03 18:11:46 s/say/saw May 03 18:11:50 :) May 03 18:39:27 # ipkg install openssh May 03 18:39:27 ipkg: Cannot create directory `/usr/lib/ipkg': File exists May 03 18:39:52 any ideas? May 03 18:50:22 nvm May 03 18:50:57 # ipkg update May 03 18:50:57 Downloading http://ipkg.nslu2-linux.org/feeds/unslung/unstable/Packages.gz May 03 18:50:57 ipkg_download: ERROR: Command failed with return value 1: `wget --passive-ftp -q -P /tmp/ipkg-3J0vCg http://ipkg.nslu2-linuxAn error ocurred, return value: 1. May 03 19:02:02 so is anyone useing the OptWare DHCP package? May 03 22:57:18 jacques: tinylogin adduser is creating group. Maybe we should patch it not to, as it was patched not co call for passwd ? May 03 22:57:57 hmm May 03 22:58:44 morning all, sorry, I'm late. going to the office May 03 22:59:21 jacques: this is just a one line comment in the code. Just have to take care for future version if ever May 03 23:29:47 ~nslu2-camps May 03 23:29:49 extra, extra, read all about it, nslu2-camps is 1: want full linksys compatibility (including future releases), can load image through web and telnet in to run ipkg, have no clue about redboot, 2: can telnet into redboot and use linux, 3: power user, but no jtag and no custom redboots, 4: bleeding edge - custom redboots, no need or desire for linksys compatibility **** ENDING LOGGING AT Tue May 03 23:59:56 2005