[10:52] <harrisony> i have the live cd but would prefer not to use the gui too install (as my computer has 256mb of ram and dont have the bandwith for the alternate installer
[01:29] <snoops> the alternate installer is the one you need..
[01:30] <snoops> xubuntu may be more suitable too
[01:31] <harrisony> damn, spend 5 days downloading the installer!
[01:32] <snoops> you should be able to install it from the live cd with 256mb of ram..it'll just take awhile. From the sounds of it not as long as downloading the other edition
[01:41] <cjwatson> 256MB installs took about 45 minutes in tests, I think
[01:42] <cjwatson> we only discovered that particular slowdown (and it's a recent slowdown) quite close to release, and the cause wasn't obvious
[04:56] <tepsipakki> cjwatson: got a minute? I'm merging back to debian the changes in xorg, but jcristau want's to know if the change to x11-common.init ("Only source /etc/default/rcS if it exists, allowing x11-common to be installed by debootstrap.") would break, since the script is 'set -e'?
[05:13] <cjwatson> tepsipakki: he needs to read the documentation of 'set -e' more carefully ;-)
[05:13] <cjwatson> "When this option is on, if a simple command fails for any of the reasons listed in Consequences of Shell Errors or returns an exit status value >0, and is not part of the compound list following a while, until, or if keyword, and is not a part of an AND or OR list, and is not a pipeline preceded by the ! reserved word, then the shell shall immediately exit."
[05:13] <cjwatson> http://www.opengroup.org/onlinepubs/009695399/utilities/set.html
[05:13] <cjwatson> if it breaks then the shell is buggy
[05:14] <cjwatson> if you want to be paranoid, though, then 'if [ -f /etc/default/rcS ] ; then . /etc/default/rcS; fi' is fine too
[05:14] <cjwatson> seems to be a common misunderstanding of set -e ...
[05:17] <tepsipakki> heh, ok. I'll educate him ;)
[05:18] <tepsipakki> actually, he suggested using if, so maybe it's a good compromise :P
[05:26] <cjwatson> fine by me, I'm not religious about using &&
[05:26] <cjwatson> it's certainly not worth the time to persuade anyone about it :)
[05:27] <tepsipakki> actually, he showed a test-case about it..
[05:27] <tepsipakki> 18:19 < jcristau> $ cat foo.sh
[05:27] <tepsipakki> 18:19 < jcristau> #!/bin/sh
[05:27] <tepsipakki> 18:19 < jcristau> set -e
[05:27] <tepsipakki> 18:19 < jcristau> [ -e /nonexistent ]  && echo a
[05:27] <tepsipakki> 18:19 < jcristau> $ ./foo.sh
[05:27] <tepsipakki> 18:19 < jcristau> $ echo $?
[05:27] <tepsipakki> 18:19 < jcristau> 1
[05:28] <cjwatson> that's only if it's the last command in the file
[05:28] <tepsipakki> ok, right
[05:29] <cjwatson> $ cat ./foo.sh
[05:29] <cjwatson> #!/bin/sh
[05:29] <cjwatson> set -e
[05:29] <cjwatson> [ -e /nonexistent ]  && echo a
[05:29] <cjwatson> echo hello
[05:29] <cjwatson> $ ./foo.sh
[05:29] <cjwatson> hello
[05:29] <cjwatson> $ echo $?
[05:29] <cjwatson> 0
[05:29] <cjwatson> paste that back to him ;-)
[05:29] <tepsipakki> already did :)
[05:32] <cjwatson> this is basically a consequence of http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_03_05
[05:34] <tepsipakki> "got a minute?" was obviously a wrong question.. "could you explain this comprehensively?" would've been correct :)
[05:34] <cjwatson> haha
[05:34] <cjwatson> you made the mistake of asking a shell pedant ;-)
[05:35] <tepsipakki> I'll be more careful the next time <g>
[07:47] <phratman> When installing sun-java6-jdk, I have to accept a licensing agreement. Is there any way to tell apt to auto-accept it?
[07:48] <phratman> I want my post installation script to automatically grab that and accept the terms so that the systems are all outfitted with the 1.6 JDK.
[07:48] <phratman> I tried apt-get --force-yes -y install sun-java6-jdk
[07:49] <cjwatson> preseed the debconf question it's asking
[07:49] <cjwatson> neither --force-yes nor -y will help you; those relate only to questions apt asks itself
[07:50] <cjwatson> preseeding the boolean question shared/accepted-sun-dlj-v1-1 to true ought to do it
[07:51] <cjwatson> so 'sun-java6-jdk shared/accepted-sun-dlj-v1-1 boolean true' in the preseed file
[07:52] <phratman> How would I do that through kickstart / kickseed?
[07:52] <phratman> I can't pass d-i commands to it can I?
[07:52] <phratman> cjwatson: By the way, https://bugs.launchpad.net/ubuntu/+bug/109475
[07:53] <cjwatson> you certainly can pass preseeding commands via kickseed; I wrote a kickstart extension to do that
[07:53] <phratman> cjwatson: Is that available as a package or something? Where would I get this extension?
[07:53] <cjwatson> https://help.ubuntu.com/6.10/ubuntu/installation-guide/i386/automatic-install.html
[07:53] <phratman> It's not available in system-config-kickstart from what I've seen.
[07:53] <phratman> Ah.
[07:53] <cjwatson> the bit that starts "A new preseed command ..."
[07:54] <phratman> Yeah, gotcha.
[07:54] <cjwatson> don't worry about the 6.10 there; that command's been available since 5.10
[07:54] <phratman> Does this method also apply to something like krb5-user or anything else that demands user input?
[07:54] <cjwatson> yes, as long as it does so using debconf
[07:54] <phratman> Okay, perfect.
[07:55] <cjwatson> any packages that ask for input using any other method have a serious bug, according to Debian policy for some time, so this is mostly not a problem
[07:55] <phratman> cjwatson: Did you get a chance to take a look at that bug listing by the way?
[07:55] <cjwatson> I'm just doing so now
[07:55] <phratman> Understood.
[07:56] <phratman> cjwatson: Also, do preseed commands apply to post installation scripts as well?
[07:58] <cjwatson> phratman: could you clarify?
[08:02] <phratman> cjwatson: Consider if I install ubuntu-desktop through the kickstart menu option, and then apt-get install krb5-user through a post installation script. Will the preseed command cascade down from the installation of ubuntu-desktop to krb5-user as well?
[08:03] <cjwatson> oh, certainly - in fact it doesn't work that way so there's no cascading to be done
[08:03] <phratman> cjwatson: In other words, if I want to bypass the debconf prompts in both packages that are prompted from ubuntu-desktop and from krb5-user, do I just enter one preseed command?
[08:03] <phratman> Ah, cool.
[08:03] <cjwatson> one preseed command per question to preseed
[08:03] <cjwatson> I would recommend installing krb5-user in %packages though, unless you have to do so conditionally for some reason
[08:04] <cjwatson> that way you'll get a progress bar and such
[08:04] <cjwatson> (well, you probably don't care. but it's still easier.)
[08:04] <phratman> Well, I didn't see a %packages option in system-config-kickstart.
[08:04] <phratman> It exists in Redhat's configuration editor, but not Ubuntu's.
[08:04] <cjwatson> system-config-kickstart is regrettably incomplete in Ubuntu, but you can just edit it into the .ks file
[08:04] <phratman> Is there any documentation on what all I can preseed?
[08:05] <cjwatson> it's in the installation guide
[08:05] <cjwatson> https://help.ubuntu.com/6.10/ubuntu/installation-guide/i386/appendix-preseed.html
[08:05] <cjwatson> there is no authoritative complete list, though, because it's whatever packages choose to provide
[08:06] <cjwatson> maybe one day we will figure out a way to index it all and remove the non-preseedable cruft
[08:07] <phratman> Are you the only person who works on kickstart / kickseed?
[08:07] <phratman> I noticed someone else was recently added to your list of developers, but all the major fixes for the installer thus far have come from you.
[08:09] <cjwatson> I'm the only person at the moment, but Evan is going to be full-time soon and I'm hoping he'll find time to do some work on automated installation
[08:09] <cjwatson> his primary responsibility is ubiquity though
[08:09] <cjwatson> I am looking at your bug BTW, but need to knock together a test rig in order to respond, which I'm working on now
[08:09] <phratman> I replicated it at home too, but I can't confirm my own bugs.
[08:10] <phratman> I'll be back later.
[08:10] <cjwatson> I'm not bothered about the confirmed state there
[08:10] <cjwatson> short answer is it's a known documented omission, but I'll try to find a workaround for you
[08:11] <cjwatson> (if you look at the logs you should see error messages in there noting that --onpart doesn't work)
[09:34] <cjwatson> phratman: I've responded to your bug in (I hope) sufficient detail to let you work around it if you want
[09:36] <cjwatson> (nasty, nasty workaround, but still)