[09:43] <CIA-3> ubiquity: evand * r3850 ubiquity/ (debian/changelog ubiquity/frontend/kde_components/PartMan.py):
[09:43] <CIA-3> ubiquity: Fix references to create_label and allow_change_step in the KDE
[09:43] <CIA-3> ubiquity: frontend's PartMan module (LP: #527932).
[10:17] <CIA-3> ubiquity: evand * r3851 ubiquity/ (debian/changelog ubiquity/components/ubi-console-setup.py):
[10:17] <CIA-3> ubiquity: str.split returns a list of one empty string when splitting on an
[10:17] <CIA-3> ubiquity: empty string.
[10:23] <CIA-3> partman-base: cjwatson * r183 ubuntu/ (52 files in 5 dirs): merge from Debian 138
[10:54] <cjwatson> ev: is anyone working on converting ubiquity-slideshow-ubuntu to po4a?  translate-toolkit no longer ships po2html (the changelog says it was too badly broken to use), so u-s-u won't build, but I started on the change and it seems fairly complex
[11:07] <ev> cjwatson: ah, lovely.  I'll take care of it.
[11:22] <cjwatson> ok, cool
[11:23] <cjwatson> I mostly noticed because the current slides/index.html has been mangled onto a single line, including JavaScript comments, which extend to the end of the line :-)
[11:23] <ev> ah, yikes
[11:26] <ev> dear po4a, thanks for not using return codes.  No love, Evan.
[11:27] <cjwatson> you mean for whether something was translated?  you probably just want to let it decide that itself ...?
[11:28] <ev> yeah, good call
[11:28] <cjwatson> it has a --keep option which is sometimes useful for tuning things
[11:29] <ev> indeed, though I'm not sure what a reasonable value for that would be.  I'm inclined to keep it at 0 for at least the time being, so there's not confusion as to why certain translations are not showing up.
[11:30] <cjwatson> yeah, or maybe 1, which would mirror the current behaviour (any translations at all => keep it)
[11:31] <ev> ah, right
[11:31] <ev> thanks
[11:31] <CIA-3> ubiquity: cjwatson * r3849 parted-2.1/ (4 files in 2 dirs):
[11:31] <CIA-3> ubiquity: Automatic update of included source packages: partconf
[11:31] <CIA-3> ubiquity: 1.32ubuntu1~ppa1, partman-base 135ubuntu5~ppa1.
[11:31] <cjwatson> anyway I should stop back-seat-driving :-)
[11:34] <ev> I'm always appreciative of the advice, but I'll interpret that as "I've got other things to do and I'm being polite" ;)
[11:35] <cjwatson> no, I just remembered how it's really annoying to be trying to do something when people are going "but you could paint it RED"
[11:36] <ev> haha, fair enough
[11:36] <CIA-3> ubiquity: cjwatson * r3850 parted-2.1/debian/changelog: releasing version 2.1.29~ppa1
[11:37] <ev> mmm optimal address alignment
[11:38] <cjwatson> I'm still not sure whether partman needs to do anything special to make use of it
[11:38] <cjwatson> do you have any hardware that cares?
[11:39] <ev> not to my knowledge
[11:39] <ev> Scott might, he was asking for it the other week
[11:40] <cjwatson> I think the default is probably cylinder-alignment
[11:40] <cjwatson> probably have to hit it with ped_device_get_optimal_aligned_constraint
[11:40] <cjwatson> or something
[11:42] <cjwatson> hmm, it would help if we built parted against blkid
[11:42] <ev> indeed
[11:57] <cjwatson> ev: what's the status of bug 336751?
[11:59] <ev> if memory serves, michaelforrest doesn't believe it's a bug, and definitely didn't did not like the solution I came up with of providing a transition between pages with a GtkSpinner embedded in the window.
[12:00] <ev> michaelforrest: is that accurate?
[12:01] <michaelforrest> I think that's  a different bug
[12:03] <michaelforrest> having the full page spinner for normal transitions (that are also accompanied by a spinning mouse cursor) is definitely a no-no. This bug you're talking about now is to do with avoiding any new windows being spawned
[12:03] <michaelforrest> I'm not sure what the solution is
[12:03] <michaelforrest> do we need to say 'starting up the partitioner' at all? can't we just spin the mouse cursor like everywhere else and wait for it to appear? (is it in a new window?)
[12:04] <michaelforrest> (by the way, we do have a much nicer idea for this sort of thing in future, but it's not quite designed-up yet!)
[12:05] <michaelforrest> I'll add my comments to the bug, shall I
[12:05] <CIA-3> ubiquity-slideshow-ubuntu: evand * r215 ubiquity-slideshow-ubuntu/ (debian/changelog slideshows/kubuntu/slides/chat.html): Fix typo in Kubuntu chat slide.
[12:05] <michaelforrest> ev remind me - how long does it take for that window to appear normally?
[12:06] <michaelforrest> or is it proportional to hard drive size / cpu?
[12:06] <ev> a few seconds
[12:06] <ev> at most
[12:07] <michaelforrest> do you have a screenshot handy?
[12:07] <michaelforrest> (I can get one - not a big deal if not)
[12:07] <ev> of what it looks like with the window up?
[12:07] <cjwatson> partitioner time taken is roughly proportional to the number of partitions
[12:07] <michaelforrest> ok so I've selected "Specify partitions manually" and clicked "Forward"
[12:08] <cjwatson> it's roughly five times faster in lucid now than in karmic
[12:08] <michaelforrest> I would like the button itself to become disabled and contain the progress bar...
[12:08] <cjwatson> but I still have a (not completely unreasonable) test case where it takes 15 seconds or so to start up
[12:08] <cjwatson> obviously when that was a minute plus, it was more of a big deal
[12:10] <cjwatson> maybe now we could just spin
[12:11] <cjwatson> I think part of what annoyed mpt is that you get a dialog popping up after some operations in the manual partitioner, not just at startup
[12:11] <cjwatson> well, maybe not, that's not what his bug says, looking at it
[12:11] <cjwatson> it annoys me :)
[12:12] <michaelforrest> this came up in the dallas design sprint basically - it's just about avoiding new windows
[12:12] <cjwatson> michaelforrest: would it make a difference to have the full-page spinner *just* for partitioner startup, not for other page transitions?
[12:12] <cjwatson> I agree that it is jarring to have it for all page transitions
[12:13] <michaelforrest> cjwatson: that full page thing is just awful and should be forgotten about as soon as possible
[12:13] <cjwatson> ok
[12:13] <michaelforrest> I have responded to the bug - I hope you think my suggestion makes sense
[12:13] <cjwatson> I think it came from mpt's comment "So, I suggest that while waiting for the partitioner to launch, the spinner should be centred in the pane"
[12:13] <michaelforrest> yeah .. no :)
[12:14] <mpt> no?
[12:14] <michaelforrest> so I'm saying we change the text of 'Forward' to 'starting partitioner'
[12:14] <cjwatson> the only query I have there is that it would cause the Forward button to expand very significantly, and push everything else to the left
[12:14] <cjwatson> wouldn't that look odd?
[12:14] <mpt> michaelforrest, so "Back" should temporarily jump to the left?
[12:14] <michaelforrest> so that it's not too disconcerting that it's taking longer than usual
[12:14] <michaelforrest> mpt: I don't see why not. The decision has been made now- it's not like the user is aiming for the button
[12:14] <cjwatson> how about instead we show the partitioner page, but put "Starting partitioner..." in its contents?
[12:15] <michaelforrest> cjwatson: I would be happy with that solution too
[12:15] <cjwatson> and then fill in the proper contents of the page once we know what they should be
[12:15] <michaelforrest> yeah that works for me
[12:15] <cjwatson> (and spin the mouse cursor, of course)
[12:15] <mpt> michaelforrest, because it would look ugly
[12:16] <michaelforrest> but a full screen blank screen with a spinner in the center is beautiful?
[12:16] <mpt> michaelforrest, no, but it would be less ugly. It's the same as (for example) System Preferences panes when they're slow to load.
[12:17] <michaelforrest> I think we have a solution here now - colin's suggestion is informative and responsive in a way that staying on the same page while waiting won't be
[12:17] <michaelforrest> so it's moot
[12:18] <cjwatson> ev: what do you think about my temporary-page-contents option?  I forget how straightforward that is to implement
[12:18] <cjwatson> for operations in the manual partitioner, I think now we can probably just ditch the progress dialog there and use a spinner
[12:19] <ev> cjwatson: manual partitioner operations>  we already do
[12:19] <cjwatson> ok
[12:19] <cjwatson> I'm obviously behind
[12:20] <ev> cjwatson: I'm a bit confused, are you saying we should put this "starting partitioner..." message on just the advanced partition page listview as it starts, or would we also be placing it somewhere on the automatic page as it starts?
[12:21] <cjwatson> I think I mean on the advanced partitioner page as it starts; for the automatic partitioning page, it seems to me that we'd probably be better off with just a spinner
[12:21] <ev> gotcha
[12:21] <cjwatson> particularly now that os-prober output will have been cached from earlier on
[12:21] <cjwatson> (clock-setup runs it)
[12:21] <ev> okay, so would be it be okay to use the existing spinner and progress message location on the advanced partitioning page for this?
[12:22] <ev> indeed
[12:22] <cjwatson> do we have a progress message on the advanced partitioning page, other than in the progress dialog?
[12:22] <mpt> michaelforrest, sure, that sounds fine
[12:24] <ev> cjwatson: yes, http://people.canonical.com/~evand/tmp/transitions.html - ignore the page transitions, skip straight to the advanced partitioning page
[12:24] <ev> assuming I correctly understand your question
[12:26] <mpt> ev, maybe you want to suppress the spinner until/unless the next step has taken more than, say, 2 seconds to display -- otherwise the installer might seem slower than it is
[12:26] <mpt> Otherwise, that looks pretty cool
[12:26] <cjwatson> OK, I'd forgotten that had landed.  Modulo mpt's comments, yeah, I think that's fine
[12:26] <ev> mpt: just to be clear, by spinner we're talking about the spinning mouse cursor, correct?
[12:27] <cjwatson> so we can display the full list box and all the furniture of that page, and just have that in-page spinner and progress message
[12:27] <ev> yarp
[12:27] <mpt> ev, no, the GtkSpinner as shown in that Flash
[12:28] <ev> on the advanced partitioning page?
[12:29] <cjwatson> commented on the bug; does this match everyone's understanding?
[12:31] <ev> yes
[12:38] <CIA-3> ubiquity-slideshow-ubuntu: evand * r216 ubiquity-slideshow-ubuntu/ (4 files in 2 dirs): Move from po2html, which is no longer included in Debian, to po4a.
[12:45] <ev> speaking of the slideshow, Dylan became an Ubuntu member last night.  An attempt to gain per-package upload rights will hopefully follow :)
[12:49] <cjwatson> cool
[13:17] <CIA-3> ubiquity-slideshow-ubuntu: evand * r217 ubiquity-slideshow-ubuntu/debian/ (5 files):
[13:17] <CIA-3> ubiquity-slideshow-ubuntu: Make each slideshow package Replace and Conflict the ubiquity-
[13:17] <CIA-3> ubiquity-slideshow-ubuntu: slideshow virtual package, to ensure that only one slideshow can be
[13:17] <CIA-3> ubiquity-slideshow-ubuntu: installed at a time and so we can put all of the files in a single
[13:17] <CIA-3> ubiquity-slideshow-ubuntu: location (LP: #526483). The upgrade slideshows do the same with the
[13:17] <CIA-3> ubiquity-slideshow-ubuntu: ubiquity-slideshow-upgrade virtual package.
[13:34] <CIA-3> ubiquity-slideshow-ubuntu: evand * r218 ubiquity-slideshow-ubuntu/ (debian/changelog update-launchpad-translations.sh):
[13:34] <CIA-3> ubiquity-slideshow-ubuntu: Update update-launchpad-translations.sh to reflect recent directory
[13:34] <CIA-3> ubiquity-slideshow-ubuntu: layout changes.
[14:01] <CIA-3> ubiquity-slideshow-ubuntu: evand * r219 ubiquity-slideshow-ubuntu/ (134 files in 3 dirs): Update translations from Launchpad.
[14:14] <CIA-3> ubiquity-slideshow-ubuntu: evand * r220 ubiquity-slideshow-ubuntu/ (debian/changelog po/ubuntu/br.po): Remove possibly inappropriate URL from br translation (LP #528465).
[14:14] <CIA-3> ubiquity-slideshow-ubuntu: evand * r221 ubiquity-slideshow-ubuntu/debian/changelog: fix LP format on previous commit
[14:26] <cjwatson> cool, parted 2.1 at least seems to minimally work with the installer
[14:26] <ev> awesome
[14:38] <CIA-3> ubiquity-slideshow-ubuntu: evand * r222 ubiquity-slideshow-ubuntu/debian/changelog: releasing version 15
[14:43] <CIA-3> partman-base: cjwatson * r184 ubuntu/debian/ (changelog control):
[14:43] <CIA-3> partman-base: Build against parted 2.1. We don't use its improved alignment features
[14:43] <CIA-3> partman-base: yet, but we plan to do so by Lucid.
[14:43] <CIA-3> ubiquity: evand * r3852 ubiquity/ (debian/changelog ubiquity/frontend/kde_ui.py):
[14:43] <CIA-3> ubiquity: Update the KDE frontend with the new slideshow location (as of
[14:43] <CIA-3> ubiquity: ubiquity-slideshow 15).
[14:56] <ev> michaelforrest: would I be correct in assuming that you'll provide the text for the restricted extras page?  Do you know if this is something that Amanda needs to approve before we push it into the archive?
[14:57] <CIA-3> partman-base: cjwatson * r185 ubuntu/debian/changelog: releasing version 138ubuntu1
[14:58] <michaelforrest> ev: I am going to brief Roz about it and then yeah, I imaging Amanda will need to approve it too
[14:58] <michaelforrest> ev: however, before anything else, I want a go-ahead from Mark!
[14:58] <ev> yeah
[14:59] <michaelforrest> I think I'll get it started actually
[14:59] <michaelforrest> we'll never get it done otherwise
[14:59] <ev> michaelforrest: by the way, the jockey installer integration didn't make it for lucid
[14:59] <michaelforrest> jockey?
[14:59] <ev> hardware drivers
[14:59] <michaelforrest> that was the thing we needed for my second checkbox to work?
[14:59] <ev> yes
[14:59] <michaelforrest> wah
[14:59] <michaelforrest> ok
[14:59] <ev> sorry, it's entirely my fault
[15:09] <CIA-3> partconf: cjwatson * r835 ubuntu/debian/control: set Vcs-Bzr for Ubuntu
[15:10] <CIA-3> partconf: cjwatson * r836 ubuntu/debian/changelog: releasing version 1.32ubuntu1
[15:13] <michaelforrest> cjwatson: is there any way I can see the work that cody sent you?
[15:14] <michaelforrest> cjwatson: the pre-greeter stuff?
[15:14] <cjwatson> boot a current desktop image, press F6, add " maybe-ubiquity" (no quotes) to the end of the kernel parameters
[15:14] <michaelforrest> ooh ok cool
[15:15] <cjwatson> that's not quite what Cody sent me since I did a bit of polishing, but what he sent just had the left half of the screen red and the right half blue ;-)
[15:15] <michaelforrest> will yesterday's work?
[15:15] <cjwatson> so I figured anything was an improvement
[15:15] <cjwatson> yeah, should do
[15:15] <cjwatson> I'm not sure both actions are hooked up properly, but you'll get the idea
[15:15] <cjwatson> I hate the language selector the way it is right now; I don't think a giant drop-down work
[15:15] <cjwatson> works
[15:17] <cjwatson> TBH, I had wanted DX to take care of making it look pretty :-/
[15:17] <michaelforrest> yeah I would think we would be doing that
[15:17] <michaelforrest> do I have to select 'install' or 'try' ubuntu on gfxboot?
[15:18] <michaelforrest> ooh mental.
[15:18] <michaelforrest> (I can see it now)
[15:18] <michaelforrest> yeah the actual spec is to have a language selector first, then the try/install screen
[15:18] <michaelforrest> certainly no giant dropdown
[15:18] <michaelforrest> ok there's no way this is landing in lucid
[15:19] <cjwatson> oh, it is?  https://wiki.ubuntu.com/Lucid/GfxbootUpdate is not that specific
[15:19] <cjwatson> just says "will include language selection"
[15:19] <cjwatson> do you have a better spec somewhere else with mockups?  I've been asking for visual design bits for this for months :-/
[15:19] <michaelforrest> (hold on - getting something)
[15:20] <michaelforrest> cjwatson: go here: http://use-case-mapper.canonical.com/specifications/0AU5sFuLRpCpBZGZra2pqY2pfMTAxZ25rcnBnNXY
[15:20] <michaelforrest> that's my spec-in-progress
[15:20] <cjwatson> can I link to that from the public wiki (even if not everyone can actually see it)?
[15:20] <michaelforrest> you have to have access to the google doc
[15:21] <michaelforrest> I could make the google doc public
[15:21] <michaelforrest> but I don't think it's all public yet
[15:21] <cjwatson> I don't mind, I just want somewhere to keep the link
[15:21] <michaelforrest> ok that's fine
[15:21] <ev> I was wondering about this...
[15:21] <ev> do we really want a greeter
[15:21] <ev> I mean, if the spec is to provide the option after language selection, why not just start ubiquity, and have it provide that interface
[15:21] <michaelforrest> ev: I would be fine with taht
[15:22] <michaelforrest> *that
[15:22] <cjwatson> that's fine for installation but doesn't work for the live session
[15:22] <ev> exiting back to the live session like it does on crash when the "try ubuntu" option is selected
[15:22] <michaelforrest> yeah I  thought there was some technical reason we can't
[15:22] <cjwatson> going through ubiquity just for language selection for the live session seems really weird to me
[15:23] <michaelforrest> cjwatson: maybe weird from a technical perspective— would it make any difference to a normal user though?
[15:23] <michaelforrest> cjwatson: and it sounds like it might make life easier
[15:23] <michaelforrest> i.e. not having to involve dx
[15:23] <cjwatson> I actually thought I meant that it would look weird
[15:23] <michaelforrest> because it would say "Install" at the top?
[15:24] <cjwatson> well, the whole interface is oriented towards installation
[15:24] <cjwatson> and anyone who's done it before (or seen it in one of the many books that has installer screenshots) might recognise that as the installer and panic
[15:24] <cjwatson> or so it seems to me ...
[15:25] <michaelforrest> I don't think that would necessarily be a problem
[15:25] <michaelforrest> especially if we called the window "Ubuntu"
[15:25] <michaelforrest> instead of "Installer"
[15:25] <michaelforrest> it would just be a language selector and then a 'try or install' screen
[15:25] <cjwatson> (insert comment about derivative branding here, but ...)
[15:26] <michaelforrest> It's a CD user's first experience
[15:26] <michaelforrest> ok so I think evan is right
[15:26] <ev> my only concern with what I'm suggesting is going from the live CD desktop back to the installer
[15:26] <michaelforrest> what happens then?
[15:27] <ev> well, I'm wondering how we approach that.  Do we simply restart the session?
[15:27] <cjwatson> we could still make it controlled by that boot option, so that the "Try Ubuntu" button only appears if you're in the boot flow
[15:27] <cjwatson> and then going from the live CD desktop to the installer is just starting it in a window, as before
[15:27] <cjwatson> do you mean going from the installer to the live CD desktop?
[15:27] <michaelforrest> oh right I see
[15:27] <ev> live CD desktop to the installer, how do we get rid of gnome panel and friends
[15:28] <cjwatson> don't?
[15:28] <michaelforrest> yeah don't
[15:28] <ev> oh, cool
[15:28] <michaelforrest> it's cool that you can still do stuff while it's installing
[15:28] <ev> I like simple solutions
[15:28] <michaelforrest> first thing I liked about ubuntu :)
[15:28] <ev> hooray
[15:28] <cjwatson> Mark explicitly asked for that, way back in Montreal
[15:28] <cjwatson> (er - the summit for DappeR)
[15:28] <cjwatson> Dapper
[15:29] <michaelforrest> so we're adding the language selector / try+install to ubiquity, and then filtering those screens unless we're in the 'boot flow'?
[15:29] <cjwatson> the language selector's already there, which saves some work
[15:29] <cjwatson> I think what ev is suggesting is to add a "Try Ubuntu without installing" button to the language selection screen?
[15:29] <cjwatson> it wouldn't offer keyboard selection, but that's probably OK since it's accessible in System -> Preferences
[15:30] <michaelforrest> no we'll have a language selection screen on its own, then a try/install screen
[15:30] <cjwatson> oh
[15:30] <ev> actually, I'm not suggesting any changes to this particular part of the design spec
[15:30] <cjwatson> ok, I suppose that's workable, as long as it's conditional so that it only happens when we're in this workflow
[15:30] <ev> was just suggesting that its implementation did not need a custom gdm greeter
[15:30] <michaelforrest> the design will be the same - we're talking about the implementation
[15:31] <michaelforrest> so happy about that 20px padding, by the way :)
[15:31] <cjwatson> ev: so you're saying that ubiquity's language page could be the implementation of the language selector in this specification, right?
[15:31] <cjwatson> michaelforrest: I bet netbook users aren't ;-)
[15:31] <ev> yes
[15:31] <michaelforrest> yeah I've been trying to test it on a netbook
[15:31] <michaelforrest> with no joy :(
[15:31] <cjwatson> we were already at the very edge of the screen limits on a number of systems
[15:31] <ev> that the live CD boots straight into the ubiquity-only session with the new design
[15:32] <cjwatson> ok, and then an optional plugin on the page after the language selector?
[15:32] <ev> bingo
[15:33] <michaelforrest> hmm - okay I'm talking to otto again - he is saying 'I hope they didn't make the window bigger instead of making the content smaller'
[15:33] <michaelforrest> (re: padding)
[15:33] <ev> ...I don't recall the request being phrased in that way
[15:33] <cjwatson> michaelforrest: BTW (UX spec) how is the user supposed to do advanced partitioning when they haven't set up their keyboard yet?
[15:33] <michaelforrest> no it wasn't
[15:34] <michaelforrest> cjwatson: this is why I asked you for feedback before xmas
[15:34] <ev> cjwatson: surely you can do it with the mouse
[15:34] <michaelforrest> I am glad you are now looking at it
[15:34] <ev> unless you have some odd mountpoitns
[15:34] <ev> mountpoints*
[15:34] <cjwatson> michaelforrest: before xmas, I was in a deathmarch customer project which was occupying about 200% of my work time - I was literally working 16-hour days for weeks at a stretch
[15:34] <michaelforrest> cjwatson: have a look at http://junction.local/specifications/0AU5sFuLRpCpBZGZra2pqY2pfMTAxZ25rcnBnNXY/use_cases/6_3
[15:34] <michaelforrest> sure, not a problem
[15:34] <cjwatson> ev: certainly in general it accepts keyboard input
[15:35] <ev> fair point
[15:35] <michaelforrest> I have thought about it though, trust me - have a look at use case 6.3
[15:35] <cjwatson> michaelforrest: I can't see junction.local :)
[15:35] <ev> s/junction.local/use-case-mapper.canonical.com/, I think
[15:35] <michaelforrest> oops
[15:35] <michaelforrest> sorry
[15:36] <michaelforrest> I didn't include a keyboard step in the variation (2.b) but I imagined if they REALLY needed to set up the keyboard, then it's in the dropdown
[15:36] <cjwatson> ah, I see what you mean now
[15:36] <michaelforrest> by the way - I have an interactive mockup of the partition tool now
[15:36] <ev> michaelforrest: me too :)
[15:36] <michaelforrest> I just need to get it packaged (Adobe Air packaged - not proper packaged ;))
[15:37] <ev> very keen to see the differences between ours, as I'm curious if you used any kind of standard selection hints
[15:37] <michaelforrest> and also I think we decided not to throw up gparted after all - unless this clarification changes that?
[15:37] <cjwatson> I'm also a bit concerned about deferring full locale setup to first boot
[15:37] <cjwatson> gparted die die die
[15:37] <ev> the default gnome way of putting a brown background on the item is quite ugly with a green/orange/red box in front of it
[15:37] <michaelforrest> ok if we hate gparted, I can accomodate that :)
[15:37] <cjwatson> we used to use gparted in the installer, in dapper
[15:37] <cjwatson> it was a nightmare
[15:38] <cjwatson> likewise qtparted
[15:38] <ev> we agreed to replace gparted with throwing up the advanced partitioning page in a window
[15:38] <ev> at the sprint, if memory serves
[15:38] <cjwatson> yeah, I think so
[15:38] <michaelforrest> ok — I was just trying to help you only have to maintain one partitioning tool
[15:38] <ev> though my question was, if the user runs through that, does it skip past the automatic partitioner
[15:38] <michaelforrest> yeah that's right - I need to update the spec
[15:38] <cjwatson> we only have to maintain one partitioning tool right now :)
[15:38] <ev> I imagine it would have to
[15:38] <ev> lol
[15:38] <cjwatson> gparted maintains itself and we don't involve it in the installer workflow
[15:39] <michaelforrest> so shouldn't gparted just go in the bin and we give people the 'advanced partitioner' then?
[15:39] <michaelforrest> or should I just shut up and not try to understand..
[15:39] <michaelforrest> (I mean on the desktop)
[15:40] <cjwatson> so mostly that would sort of be my vote except that the advanced partitioner isn't really set up to be a standalone desktop app
[15:40] <ev> partman isn't in general
[15:40] <cjwatson> it has some expectations which hold well during installation, but not so much outside that
[15:41] <michaelforrest> ok so ev - about the window size - if it's already to the limit without the padding, then we need to make the content smaller to accommodate the padding
[15:41] <cjwatson> gparted and partman both use the same core partitioning backend, but they have different business logic on top of it which is optimised for different purposes
[15:43] <ev> michaelforrest: uhm, easier said than done.
[15:43] <michaelforrest> ev: sorry that wasn't really clear before  - I didn't really think about it - but it was pretty obvious to otto, and he knows about that sort of thing..
[15:43] <michaelforrest> yeah I imagine so
[15:43] <michaelforrest> but we need to do it
[15:43] <michaelforrest> run a script on the gtk xml?
[15:43] <ev> sure, I'll see what I can do, but it's going to be difficult to shave 20px off each side, so I'm not making any promises
[15:43] <michaelforrest> to avoid fiddling with glade ?
[15:43] <cjwatson> there's no absolute positioning or sizing in there
[15:44] <ev> it's a matter of having too much content
[15:44] <ev> the text on the user setup page, for example, takes up a lot of space
[15:44] <michaelforrest> I'm sure you can fit it in
[15:45] <cjwatson> this would be the second or third attempt to squash things down, I think
[15:45] <michaelforrest> from a design perspective, it's a major difference in how professional it looks
[15:45] <michaelforrest> ev there is more than enough space on the user setup page judging from the screenshot I'm looking at
[15:46] <cjwatson> michaelforrest: the thing is that the notebook format (and general sanity - we don't want to be growing and shrinking the window in the middle) means that the installer grows to the size of the largest page
[15:46] <ev> hrm, indeed
[15:46] <ev> I could have sworn that was where it grows though
[15:47] <michaelforrest> cjwatson: yeah I'm aware of that
[15:47] <ev> I'll look into it and see what I can shave off
[15:47] <cjwatson> might be horizontal growth on one page and vertical growth on another?  I'm not sure how smart gtk is ...
[15:47] <ev> gtk, smart, hah
[15:47] <cjwatson> could be worse, could be qtdesigner *shudder*
[15:47] <ev> but indeed, I'll watch out for that
[15:47] <ev> thanks
[15:47] <ev> hahahahaha
[15:47] <ev> oh god
[15:47] <ev> NEVER AGAIN
[15:48] <michaelforrest> I am currently writing a spec for a glade-style tool..
[15:48] <michaelforrest> if I have to build it myself, I will...
[15:48] <ev> michaelforrest: while I have you here...
[15:48] <ev> What are your thoughts on the following for the "unable to install restricted extras (dvd, mp3, flash) during install because we didn't have a network connection" scenario.  On first boot you get the following window:
[15:48] <ev> http://people.canonical.com/~evand/tmp/update-notifier-restricted-extras.png
[15:48] <ev> and if you click run this action now, you get:
[15:48] <ev> http://people.canonical.com/~evand/tmp/software-center-restricted-extras.png
[15:48] <ev> This is similar to how incomplete language support is handled.  That is, when you do an install with a language that we don't include langpacks for on the CD (because they take up way too much space), you see an update-manager notification that tells you:
[15:48] <ev> The language support files for your selected language seem to be incomplete.
[15:48] <ev>  You can install the missing components by clicking on "Run this action now"
[15:48] <ev>  and follow the instructions. An active internet connection is required.
[15:48] <ev>  If you would like to do this at a later time, please use "System ->
[15:48] <ev>  Administration -> Language Support" instead.
[15:49] <ev> cjwatson: if you have any thoughts as well, I'm keen to hear them
[15:49] <michaelforrest> ok the language needs to change a lot
[15:49] <ev> sure
[15:49] <ev> it was a quick mock up
[15:49] <cjwatson> ev: technically I think it's fine, I'll defer to Michael on the design obv.
[15:49] <ev> I'm more concerned about the workflow
[15:50] <michaelforrest> I would like to unify any post-installation steps into one place if at all possible
[15:50] <cjwatson> if we don't have a network connection during install, then we need to ask later, and *right now, this release* we only really have one way to do that
[15:50] <michaelforrest> but I would want to talk to mpt / johnlea about how to do this
[15:50] <cjwatson> independent of future first-boot work
[15:50] <cjwatson> assuming you're talking about lucid here ...
[15:50] <ev> yeah, lucid
[15:51] <michaelforrest> ev: I would kinda like to just go ahead and install it in the background silently (ducks)
[15:51] <michaelforrest> (*ducks*)
[15:51] <ev> I'm keen to get something in place, so Steve Langasek doesn't laugh hysterically at me when I ask to land this.
[15:51] <ev> well, this is for the case where we don't have a network connection
[15:51] <ev> and thus cannot do it until post-install, when a network connection is established
[15:51] <michaelforrest> yeah so it's no help popping it up before an internet connection is available
[15:52] <ev> and as far as installing it in the background, I think Amanda might take issue with that :)
[15:52] <michaelforrest> chrome installed a load of new shit in the background without asking me! crazy. whatever will they do next.
[15:52] <michaelforrest> I can't say I was that impressed ;)
[15:52] <ev> michaelforrest: ah, indeed, though our current stack makes that difficult unless we're trying to wget http://www.ubuntu.com/am_I_online in a cron job
[15:53] <michaelforrest> so on first network connection, it's gonna check for updates and there's gonna be that {!} icon
[15:53] <michaelforrest> ?
[15:53] <ev> on first boot, regardless of network connection, the first mentioned window will show up
[15:53] <michaelforrest> is that how it works? or is the update check only on a cron job?
[15:53] <michaelforrest> yeah I don't want anything popping up like that
[15:53] <michaelforrest> but can we roll it into the first update process?
[15:53] <cjwatson> I think it's a cron job
[15:54] <cjwatson> I would love to have a way to queue something for the next time apt runs
[15:54] <michaelforrest> We're always going to want to install updates on the first install right?
[15:54] <cjwatson> I've been wishing for that for about 5 years
[15:54] <michaelforrest> so why can't we make it check for updates on first network connection?
[15:54] <cjwatson> it would be very handy for language packs
[15:54] <ev> indeed, I think this very much mirrors the langpack case
[15:55] <cjwatson> so first time you connect to the network, it automatically starts downloading stuff to make sure that you can't use the network straight away? :-)
[15:55] <michaelforrest> I'd like to just roll everything into that process - language packs, updates, restricted extras
[15:55] <michaelforrest> no - it says 'updates are available - it is recommended that you install these now'
[15:55] <cjwatson> the "do I have any updates" download is quite big in itself
[15:55] <michaelforrest> really? why?
[15:56] <cjwatson> downloads all the Packages files ...
[15:56] <cjwatson> why> because we have a lot of packages in Ubuntu
[15:56] <ev> which would be very unfortunate if you're on a 3G dongle and it runs
[15:56] <michaelforrest> anyway I think it's a safe assumption that on first boot, the user needs to install updates, so we don't necessarily need to check
[15:56] <michaelforrest> am I wrong?
[15:56] <michaelforrest> no :)
[15:57] <cjwatson> my concern is basically that we're putting more and more cute things in the way of the user playing with their new toy
[15:57] <michaelforrest> we're not!
[15:57] <cjwatson> I think updates should be later
[15:57] <michaelforrest> we're making sure the toy is ready before they start trying to play with ti!
[15:57] <michaelforrest> *it!
[15:57] <ev> lets not confuse updates with what we're trying to do here
[15:57] <cjwatson> the toy is ready upon installation
[15:57] <ev> I think it's the wrong language
[15:57] <michaelforrest> not to watch youtube videos
[15:57] <michaelforrest> not to listen to mp3s
[15:58] <michaelforrest> these are REALLY important use cases
[15:58] <ev> to be clear, we're talking about a set of packages that need to be installed, not updated
[15:58] <cjwatson> right, installation is a different thing
[15:58] <michaelforrest> sure - so we don't have to call it 'updates'
[15:58] <ev> ubuntu-restricted-extras, language-pack-$LL, ...
[15:58] <cjwatson> installing updates takes HOURS
[15:58] <cjwatson> or can easily do
[15:59] <michaelforrest> this is a problem. it also affects software center
[15:59] <michaelforrest> (I am going to correct that to *centre, as we probably all agree on that spelling)
[16:00] <ev> haha
[16:00] <michaelforrest> I think I need to work with people to get the first-use use cases formalised
[16:01] <michaelforrest> ev - don't do this 'restricted extras not yet available' thing - just don't do anything if there wasn't network for now
[16:01] <ev> michaelforrest: isn't that more confusing?  We've asked the user if they want to add this package, they've said yes, and we haven't installed it
[16:02] <michaelforrest> I mean - don't ask if there's no network during install
[16:02] <ev> or are you suggesting that we make showing this page conditional on wget http://www.ubuntu.com/am_I_online working?
[16:02] <cjwatson> I really want to get pdiffs landed so that incremental checks for updates are quicker, but the last time I tried, it got derailed into a bluesky discussion of how to redesign the entire archive, which was a bit unhelpful
[16:02] <michaelforrest> yeah basically. for now.
[16:02] <ev> ah, okay
[16:02] <cjwatson> but I suppose incrementals aren't so important here
[16:05] <mpt> ev, cjwatson: michaelforrest and I had a chat about the partitioner feedback stuff. It was partly a miscommunication and partly me just being clingy. :-) What cjwatson summed up in the bug report we're both happy with.
[16:06] <ev> software-center could make the assumption that anyone passing a list of packages to be installed knew they would be in the cache after an update, and do both an update and install if and only if the user presses yes on the "there are more packages that need to be installed to complete the Ubuntu set up." question
[16:06] <ev> mpt: lovely
[16:08] <cjwatson> mpt: great, thanks
[16:08] <michaelforrest> how big are security updates?
[16:09] <michaelforrest> usually? if there is a usually?
[16:10] <mpt> michaelforrest, it increases steadily based on how long it's been since that version of Ubuntu was released
[16:11] <mpt> e.g. if you installed 9.10 now, it would be ... a couple of hundred MB, I think
[16:11] <mpt> (iirc, from doing it a couple of weeks ago)
[16:11] <michaelforrest> right ok
[16:11] <mpt> whereas if you did it back in November, it would be just a few MB
[16:11] <michaelforrest> we're wondering about having a rationalised updates screen with a list like
[16:11] <michaelforrest> - proprietary drivers / restricted extras  [29Mb]
[16:11] <michaelforrest>     - security [100MB]
[16:11] <michaelforrest>     - language [290KB]
[16:11] <michaelforrest>     - system updates [140MB] 
[16:11] <michaelforrest>     - application updates [200MB]
[16:12] <michaelforrest> (with selective updates , and this list, being the second choice)
[16:12] <michaelforrest> (made up numbers, of course !)
[16:12] <mpt> yeah, it would be nice to say "oh, sheesh, I don't have enough quota to download all that now, let me get just the security updates"
[16:13] <michaelforrest> I would like to put something much nicer than 'proprietary drivers / restricted extras' :)
[16:14] <michaelforrest> I think if we're prepared to flash up 'updates are available' on a regular basis, then we should be happy to pop up a dialog box like this when the internet first becomes available
[16:15] <michaelforrest> I'm going to write some use cases for first-login and we can discuss them in our tuesday desktop design team meeting
[16:16] <michaelforrest> cjwatson: what is the first thing you do after  you install ubuntu ?
[16:16] <michaelforrest> personally
[16:19] <cjwatson> start a terminal window and ssh :-)
[16:19] <cjwatson> probably install a few packages I know I want
[16:19] <michaelforrest> ok
[16:20] <cjwatson> and I guess I start a web browser, though think blogs rather than youtube in my case
[16:22] <michaelforrest> cool.
[16:22] <michaelforrest> you're use case #1.
[16:22] <michaelforrest> (well - 5.1)
[16:22] <ev> michaelforrest: by the way, this is what I've come up with so far for the automatic partitioner working off the design spec: <http://people.canonical.com/~evand/tmp/automatic-paritioner.png>.  Could I perhaps have access to that AIR package, so I can bring it more in line with what you're looking for over the weekend, should I find some spare time?
[16:23] <michaelforrest> ok lemme have a crack at that now
[16:23] <michaelforrest> exciting to see it already though!
[16:23] <cjwatson> michaelforrest: so the one thing left unresolved by the above discussion of the installation flow is how we represent that you can press a key to get at the full boot menu
[16:24] <ev> the redesign excites me, so I worked on it on the plane to/from Pycon.
[16:24] <michaelforrest> cjwatson: yes that is true.
[16:24] <michaelforrest> ev :)
[16:24] <cjwatson> I wonder if there are any stock icons showing a keyboard key
[16:24] <michaelforrest> cjwatson: ok - so I am wondering if we should show anything. it's a very complex piece of information to attempt to convey graphically
[16:25] <michaelforrest> cjwatson: apple don't mention anything about specialised boot options - you find out about them on forums. also, we could put something in the CD inlay
[16:26] <michaelforrest> we've had a brainstorm, including contributions from our resident anthropologist
[16:26] <cjwatson> I don't mind that so much, but I would like to have some indication of when you need to press the key (this is a major flaw with Macs)
[16:26] <michaelforrest> the tricky bit is to say 'you can press a key here' without saying 'please press a key'
[16:26] <cjwatson> even if that's just an Ubuntu logo or something
[16:27] <cjwatson> the Mac problem is that by the time you see anything other than a grey screen, it's too late :)
[16:28] <michaelforrest> so I think this is information that is of interest to technical users and people with particular accessibility issues
[16:28] <michaelforrest> and hopefully nobody else will ever need to know
[16:28] <michaelforrest> I would like to think that technical users will know how to get help
[16:28] <michaelforrest> so the best suggestion we had was to put the accessibility icon there
[16:29] <michaelforrest> I think that makes the right trade off of 'normal people won't press it', without excluding people who need it for accessibility reasons
[16:29] <michaelforrest> what do you think?
[16:30] <cjwatson> hmm, I prefer the Ubuntu logo I think
[16:31] <cjwatson> it doesn't invite people to press a key
[16:32] <cjwatson> something about using the accessibility icon for this doesn't sit right with me, I'm not quite sure what
[16:32] <michaelforrest> it's that it's not 100% accurate, probably :)
[16:32] <cjwatson> hmm, no
[16:32] <michaelforrest> however, I think it works perfectly if one allows artistic license..
[16:33] <cjwatson> hmm, the accessibility icon isn't the wheelchair logo any more is it?
[16:33] <michaelforrest> no it's that arms-out man
[16:34] <cjwatson> do you mean the one that's next to Universal Access?
[16:34] <cjwatson> right, that one
[16:34] <michaelforrest> which actually feels more apt
[16:34] <michaelforrest> yeah
[16:34] <cjwatson> that feels better, I think there was something about Ubuntu displaying an icon normally used on toilets as it booted that I didn't like :)
[16:34] <michaelforrest> indeed.
[16:35] <cjwatson> centre of the screen, or off in a corner?
[16:35] <cjwatson> might be worth greyscaling it?
[16:35] <michaelforrest> something like in this first pic http://use-case-mapper.canonical.com/specifications/0AU5sFuLRpCpBZGZra2pqY2pfMTAxZ25rcnBnNXY/use_cases/6_1
[16:35] <michaelforrest> the icon instead of the text
[16:35] <michaelforrest> so in a grey band along the bottom
[16:36] <michaelforrest> or a lower-contrast band
[16:36] <michaelforrest> if we're on black
[16:36] <cjwatson> remember that by the time the logo and the boot progress bar appear, it's too late
[16:36] <cjwatson> though I could certainly display the logo
[16:36] <cjwatson> that looks like a spinner in the middle though?
[16:36] <michaelforrest> can't we put the logo on that screen ?
[16:36] <cjwatson> we're on black, yeah
[16:37] <cjwatson> I can display the logo certainly, it was just the thing that looks like a spinner that confused me
[16:37] <michaelforrest> yeah sorry - I guess that's the next screen then
[16:37] <michaelforrest> or whatever we have. it's meant to be abstracted from the design
[16:37]  * cjwatson can't afford too much abstraction at the moment :)
[16:38] <cjwatson> ok, I think I can do that
[16:38] <cjwatson> not sure what a lower-contrast band on a black background means though
[16:38] <michaelforrest> yeah I have a new photoshop document open to help de-abstractify.
[16:58] <michaelforrest> cjwatson: where do you suppose I might find an icon with a picture of a keyboard on it?
[17:01] <cjwatson> System -> Preferences?
[17:02] <ogra> locate keyboard |grep png
[17:02] <ogra> should give you plenty
[17:02] <cjwatson> not sure they're very *good* icons, but ...
[17:24] <michaelforrest> ok cjwatson- basically, otto is looking at this next week
[17:24] <michaelforrest> we have a guy in to help with icons
[17:24] <michaelforrest> my efforts, sadly, are rubbish.
[17:25] <michaelforrest> so I'm going to focus on getting evan his installer air demo
[17:26] <cjwatson> how about I go ahead with the accessibility icon thing
[17:26] <cjwatson> it seems workable enough
[17:26] <cjwatson> I can't put the icon in place until I have something I can land as the default workflow for lucid anyway, of course
[21:06] <michaelforrest> cjwatson:  go ahead with the accessibility icon thing. we'll see how it turns out.
[21:07] <cjwatson> michaelforrest: ok, cool.  thanks.