[10:18] <xivulon> hi, trying to help mint devs to use wubi, apparently they use remastersys and the ISO is a bit different, I am not too familiar with that, what would be a good way for them to a livecd close enough to the ubuntu one?
[10:20] <evand> Is it not working as is?
[10:20] <evand> I'd be surprised if the installer worked and Wubi did not
[10:21] <evand> short of the contents of disk/info
[10:21] <xivulon> yeah already suggested them to fix that, but he mentioned that the ISO was "different" not sure what he meant, I am waiting for the logs
[10:27] <cjwatson> evand: did the USB samples arrive at Millbank?
[10:27] <evand> cjwatson: not yet to my knowledge, but I'll ask around
[10:27] <evand> do you recall who they are being shipped to?
[10:27] <evand> cezz?
[10:27] <cjwatson> evand: one of that lot over there :)
[10:28] <evand> :)
[10:28] <cjwatson> (i.e. don't know)
[10:43] <evand> cjwatson: two were sent, one for a before-9 arrival, one on a regular shipment
[10:43] <evand> neither have arrived yet
[11:03] <xivulon> evand: what was the bug you mentioned yesterday?
[11:12] <evand> xivulon: trying to find it
[11:13] <xivulon> weird bug report (in italian): sudden shutdown during installation, apparently reproducible
[11:18] <evand> xivulon: http://pastebin.ubuntu.com/65124/
[11:18] <evand> I cannot find a bug number
[11:18] <evand> perhaps talk to bdmurray
[11:29] <xivulon> thx evand
[11:33] <xivulon> In any case I think that the best course of action for the future is to get rid of the /boot bindmount alltogether and instead rsync /boot to /host/ubuntu/disks/boot during update-grub
[11:36] <xivulon> that will also solve #243105 and #252900
[11:36] <xivulon> cjwatson, I think you suggested that to begin with some time ago' re 252900
[11:39] <evand> cjwatson: arriving at half two
[11:44] <evand> "
[11:44] <evand>     *
[11:44] <evand> arrr
[11:44] <evand> I know I configured irssi to not do that.
[11:45] <evand> mpt: "Merge the automatic and manual partitioning pages." (https://wiki.ubuntu.com/UbiquityVisualRefresh) - if you have some time, can you mock this up?  I'm perplexed as to how it would look
[11:45] <cjwatson> xivulon: I dunno, I can't say I'm entirely convinced - it feels pretty fragile
[11:46] <evand> though the more I think about it, the more it seems feasible as the resize option is moving into the segmented bar anyway
[11:46] <cjwatson> evand: sounds like it'd present a scary UI to people not prepared for it
[11:46] <evand> indeed, that's why I'm curious to see the mock up
[11:46] <mpt> evand, sure
[11:46] <evand> I can't think of a reasonable way of doing it, but maybe he can
[11:47] <evand> mpt: thakns
[11:51] <evand> mpt: I'm interested to know of any other ideas you may have that I'm not already aware of as I'm sitting at a desk with screenshots of ubiquity spread around trying to decide on what UI changes I need to make in 9.04
[11:51] <evand> cjwatson: you as well
[11:52] <cjwatson> syncing up user-setup comes to mind: the private directory stuff
[11:54] <evand> indeed.  Will we need to add the open ID stuff to that page eventually, or has that idea finally died of old age?
[11:54] <mpt> evand, earlier I mentioned to you the idea of splitting out the migration assistant, so that you can (a) migrate from another OS despite not dual-booting and (b) migrate from another computer even after you've finished installation
[11:55] <cjwatson> evand: will have to talk with online services about current plans there
[11:55] <mpt> evand, I don't know what effects that would have on the installer interaction, that's something that would need designing
[11:56] <evand> mpt: it's already possible to run it separate of ubiquity, it just doesn't have a UI.  What would the advantage be of removing it from the installer?
[11:56] <mpt> evand, (a) and (b) above :-)
[11:56] <evand> cjwatson: noted
[11:57] <evand> a) I'm confused by.  Short of adding the ability to import to a temporary medium like a USB disk or CD, there's no way to import from an OS you plan to delete. b) Already possible, just needs a GUI.
[11:57] <mpt> evand, exactly, it would involve putting stuff on a temporary medium
[11:58] <evand> ah, but that can be done inside ubiquity, no?
[11:59] <evand> (I'm not moved either way, I'm just trying to understand the rationale)
[11:59] <mpt> It could, and that would have pluses and minuses
[11:59] <mpt> The plus would be less complex interaction
[12:00] <mpt> The minus would be that the migration utility would look quite different depending on whether you were running it as part of the installer or standalone
[12:00] <mpt> which could be confusing
[12:00] <evand> hrmm
[12:02] <mpt> I think it needs a bunch of sketching out how it would work if invoked separately, and conversely how it would look if still embedded
[12:02] <evand> So you're proposing removing it from ubiquity and moving it into a separate utility that's seeded on the CDs and included in the default install?
[12:03] <evand> I could remove the UI, but keep the component in place for preseeding (Wubi), and leave it in, or rather finally add it to, the alternate installer.  So it's feasible.
[12:03] <mpt> yes
[12:04] <evand> ok, I need to give that one some thought (perhaps discussing it at length at UDS), but I've made a note to do so.
[12:04] <mpt> great, thanks
[12:05] <evand> mpt: any suggestions for what a good password strength meter would look like?  I'm tempted to use a series of lock images, but I suspect that's poor design.
[12:06] <mpt> In my original installer mockup I just had "☹ Too short" next to the password field
[12:06] <mpt> For something more elaborate, I guess a color-coded gauge
[12:06] <evand> right, I forgot that was in the wiki
[12:06] <mpt> short and red for a bad password, long and green for a good one
[12:07] <mpt> Partly it depends how much space you have for it
[12:08] <evand> assuming we put it next to the password input boxes, plenty.
[13:39] <xivulon> ohh http://mgerards.net/blog/?p=16#more-16 there is loop support in grub2
[13:40]  * xivulon lobbies for grub2
[13:40] <cjwatson> sigh
[13:40] <cjwatson> we don't need lobbying, we need extensive testing
[13:41] <xivulon> davmor2 ^
[13:41] <xivulon> (kidding)
[13:43] <xivulon> well jokes asides, it seems that grub2 could potentially address several issues, hope it proves robust enough for jaunty
[13:44] <CIA-2> ubiquity: cjwatson * r2937 ubiquity/debian/ (changelog rules): Add an intro message noting that we're alpha again.
[14:04] <CIA-2> ubiquity: cjwatson * r2938 ubiquity/ (configure configure.ac debian/changelog):
[14:04] <CIA-2> ubiquity: Autoconf likes to have a version in AC_INIT, but start omitting our
[14:04] <CIA-2> ubiquity: sub-minor version from it so that we don't have to regenerate configure
[14:04] <CIA-2> ubiquity: with every single upload.
[14:04] <xivulon> cjwatson was reading http://wiki.debian.org/Grub/Grub2, do you think it is realistic to expect grub2 in jaunty? as that would imply quite different code paths for me
[14:05] <xivulon> and if not, would be using grub2 withing wubi be acceptable?
[14:05] <cjwatson> I haven't thought grub2 was particularly realistic for a long time
[14:05] <cjwatson> it is always possible that I may be surprised
[14:06] <cjwatson> grub-installer does support grub2 in theory, though nobody's ever stepped up to make sure it contains the necessary Ubuntu changes to grub AfAIK
[14:06] <cjwatson> AFAIK
[14:06] <xivulon> I am not familiar with possible the issues, but would the problems affect the wubi user case
[14:07] <cjwatson> I can't imagine how they wouldn't
[14:07] <cjwatson> the issues are that it is a complete rewrite that hasn't been deployed in any major distribution to my knowledge
[14:08] <cjwatson> people have been talking about it more or less since Ubuntu started, but there's much more talk than action
[14:08] <cjwatson> the last time we looked at it it turned out that the descriptions of it substantially over-represented what it was capable of
[14:08] <cjwatson> so I'm afraid I'm pretty cynical about it now
[14:09] <cjwatson> I'd welcome somebody else taking it on, but I'm not going to
[14:09] <xivulon> well since in my case there are some clear advantages (at least on paper), and since the usage of wubi is fairly segmented, we might use that as a test bench
[14:10] <cjwatson> be wary of what it promises, and test it on a wide variety of machines
[14:10] <xivulon> remember that I am now left with grub4dos which is probably less tested than grub
[14:10] <evand> yay, usb key just arrived
[14:11] <xivulon> well I can only test it on my laptop, my wife's and vms :)
[14:17] <davmor2> I can test on hw but not at the moment
[14:19] <CIA-2> ubiquity: cjwatson * r2939 ubiquity/debian/intro-alpha.txt: update intro message for 9.04
[14:21] <xivulon> davmor2 no rush, I have to check first if fat + ntfs + loop support works well and from within a windows partition. Otherwise there is no much point.
[14:25] <CIA-2> oem-config: cjwatson * r557 oem-config/debian/changelog:
[14:25] <CIA-2> oem-config: Switch to a new versioning scheme and omit the sub-minor version from
[14:25] <CIA-2> oem-config: AC_INIT so that we don't have to regenerate configure with every single
[14:25] <CIA-2> oem-config: upload.
[14:26] <CIA-2> oem-config: cjwatson * r558 oem-config/d-i/sources.list: update for jaunty
[14:26] <CIA-2> ubiquity: cjwatson * r2940 ubiquity/d-i/sources.list: update for jaunty
[14:41] <evand> 10:37:51 < PovAdct_w> if I run wubi.exe from the xubuntu desktop CD, I get a wizard driving me through *Kubuntu*  installation... can someone confirm?
[14:41] <evand> xivulon: ^
[14:46] <xivulon> evand does it mean that the artwork is wrong?
[14:47] <evand> No idea, I'd suggest pinging him
[14:51] <xivulon> no idea the metalink is ok, and the artwork is ok. not sure how he can end up running kubuntu from a xubuntu iso
[14:51] <xivulon> evand can you get more info?
[14:51] <evand> xivulon: I'm rather busy today, can you ping him?
[14:52] <xivulon> which channel?
[14:52] <evand> He was in #ubuntu-release-party, but given the high volume I'd suggest private messaging him
[15:01] <xivulon> evand it's ok he had a kubuntu CD in the tray
[15:10] <evand> ah
[15:10] <evand> ok
[15:42] <evand> WrapLabel
[15:42] <evand> A GtkLabel subclass that can wrap to any width, unlike GtkLabel which has a fixed wrap point. It does this through a couple of tricks in the size_allocate and size_request handlers by using PangoLayout's height-for-width.
[15:42] <evand> cjwatson: ^ from libview.  We'd have to write Python bindings though.
[16:18] <davmor2> xivulon: If I'm not around when it's ready to test send me a mail davmor2@davmor2.co.uk
[17:25] <evand> ah ha! cjwatson: http://medsphere.org/projects/widgets/wiki/FBox
[17:30] <evand> I think I just might rewrite that in Python
[20:10] <superm1> evand, i've got a change for usb-creator i really should have proposed for merging sooner, could you take a look? https://code.edge.launchpad.net/~dell-team/usb-creator/command-line-args
[23:23] <cjwatson> evand: Hmm. I wonder how well that works within a treeview