[09:36] <jerom1> Hi,
[09:37] <jerom1> i use kickstart but i cannot use bash interpreter, can you confirm %post --interpreter doesn't work
[10:04] <cjwatson> you cannot use bash in kickstart; that is correct
[10:04] <cjwatson> it's always possible to avoid bashisms
[10:05] <cjwatson> it's usually a fairly trivial change
[10:07] <jerom1> is it correct if i comment test "if [ "$2" != /bin/sh ] ; then" in lib/kickseed/handlers/post.sh in initrd ?
[10:08] <jerom1> i must use bash, my datacenter sysadmin must use bash (in standard post install ...)
[10:13] <cjwatson> you cannot use bash. it is not there.
[10:13] <cjwatson> you only have busybox sh in the installer environment.
[10:13] <cjwatson> show me your %post script and I'll be entirely happy to convert it to ordinary sh for you; this is a trivial operation
[10:14] <cjwatson> in fact there's often nothing to convert
[10:14] <jerom1> oki I will convince my sysadmin
[10:15] <cjwatson> your sysadmin should recognise that there is a very large subset of bash which can be used in plain POSIX /bin/sh
[10:15] <cjwatson> (this isn't Solaris /bin/sh, which is considerably less capable)
[10:16] <jerom1> oki i demand to my sysadmin to show his bash script
[10:17] <cjwatson> hmm, well
[10:17] <cjwatson> actually
[10:17] <cjwatson> strictly speaking if you aren't using --nochroot then you ought to be allowed to use --interpreter /bin/bash
[10:17] <cjwatson> (sorry, I've only just woken up ...)
[10:17] <cjwatson> I had forgotten that --nochroot wasn't the default
[10:18] <cjwatson> I'll make that change in kickseed for feisty. As I say, though, it shouldn't block you, because all bash scripts can be rewritten to work in sh with at most minor modifications.
[10:18] <jerom1> oki good :-)
[10:19] <cjwatson> it's a bit more than just removing that check in post.sh, because /bin/sh is hardcoded elsewhere
[10:20] <cjwatson> not a huge change, but hard to describe over IRC
[10:20] <jerom1> oki
[10:20] <jerom1> thanks
[11:09] <cjwatson> ok, kickseed thus improved in feisty
[11:09] <cjwatson> thanks for the report; sorry I incorrectly dismissed it at first
[11:09] <jerom1> no problem, great thanks
[01:48] <Riddell> cjwatson: seems like qt4 ubiquity does the guided partitioning immediately after you click Next without warning, is that something I missed out or something that hasn't been put back in yet?
[01:51] <cjwatson> it should calculate the new partitions but not actually commit them until after the summary page. AFAIK that's working in gtk+qt3
[01:54] <Riddell> yes, it was, I wonder what i missed out
[01:59] <cjwatson> the summary page implementation maybe?
[02:00] <cjwatson> that code path can be hairy ...
[02:07] <Riddell> is there an extra mainloop nesting there?  the way I've coded it the flag will only know about 1 nesting
[02:29] <cjwatson> Riddell: there used to be, but isn't any more
[02:30] <Riddell> cjwatson: what changed?
[02:31] <cjwatson> it was made a whole lot less crackful
[02:31] <cjwatson> I rewrote the partitioning commit step
[02:34] <Riddell> so it should just work without coding anything special, instead of wiping my hard disk at the wrong time
[02:34] <cjwatson> do you see the summary page at all?
[02:34] <Riddell> choose Guided paritioning, click Next -> progress box saying hard disk being wiped
[02:35] <cjwatson> I'm not convinced by the on_steps_switch_page change you made
[02:36] <cjwatson> is there nothing like the aboutToShow signal any more?
[02:36] <cjwatson> IMO it should be done with signals to make sure that we handle page changes properly event if they aren't done via set_current_page
[02:38] <cjwatson> -        if step == "stepWelcome":
[02:38] <cjwatson> +        if step == str("stepWelcome"):
[02:38] <cjwatson> >>> "stepWelcome".__class__
[02:38] <cjwatson> <type 'str'>
[02:38] <cjwatson> why that change?
[02:41] <cjwatson> other than that, unclear - trying it myself now
[02:41] <Riddell> it wasn't matching, at least on one of the places in that if bunch, I didn't look into why
[02:42] <cjwatson> brb, parents
[03:18] <cjwatson> have fun :) I'll see what I can fix up
[04:16] <cjwatson> Riddell: can't reproduce here
[04:17] <cjwatson> I had to fix a partman problem to stop parted_server segfaulting and see the content of the autopartitioning page at all, but that's not your problem ...
[05:36] <allmanj> hey - i'm looking for packages on the dapper cd that aren't installed by ubuntu-standard or ubuntu-desktop so i can remove them as i'm after creeping a little over burnable size...
[05:37] <allmanj> that, or i'd like to modify what counts as "ubuntu-desktop"
[05:48] <cjwatson> allmanj: look at the Task lines in the Packages file and/or study https://wiki.ubuntu.com/SeedManagement
[05:49] <allmanj> dapper appears to have an ubuntu-desktop package... is this not used?
[05:53] <cjwatson> sure, it's used
[05:54] <cjwatson> but as a component of the task ...
[05:54] <cjwatson> language packs are a normal-ish thing to strip out
[05:54] <allmanj> i'm thinking the samba server is a candidate
[05:54] <cjwatson> sounds reasonable
[05:55] <allmanj> aye - it's big and i'm pretty sure i'll never need it
[10:12] <macogw> what does ubiquity show up as in the system monitor's processes list?  it froze installing feisty -.-
[10:16] <macogw> the X button in the top right doesnt do anything. minimize and maximize work just dandy