[08:23] <MacSlow> Greetings everybody!
[08:24] <juliux> hi MacSlow
[08:31] <MacSlow> hi juliux
[13:11] <zul> @schedule montreal
[13:11] <ubotu> Schedule for America/Montreal: 22 Nov 09:00: Desktop Team Development | 23 Nov 07:00: MOTU meeting | 29 Nov 09:00: Desktop Team Development
[13:20] <Hobbsee> @now
[13:20] <ubotu> Current time in Etc/UTC: November 22 2007, 13:20:57 - Next meeting: Desktop Team Development in 39 minutes
[13:58] <Keybuk> GOOD AFTERNOON!
[13:59] <seb128> hey hey hey
[13:59] <kwwii> howdy
[13:59]  * Hobbsee waves
[14:00] <Keybuk> (we've lost a pitti)
[14:00] <Hobbsee> uh oh
[14:00]  * Hobbsee sends out a search party
[14:01] <Keybuk> Apologies from Ted today
[14:01] <kwwii> that is such a pitti :-)
[14:01] <seb128> yeah for fixed merges.ubuntu.com index ;-)
[14:01] <Keybuk> something to do with turkey apparently
[14:01] <Riddell> hi
[14:01] <kwwii> hooray for turkey day
[14:02] <Keybuk> first thing on the agenda
[14:02] <Keybuk> today is spec approval day
[14:02] <Keybuk> I realise that quite a few of these are blocked on me reviewing them (hi, mvo! :p)
[14:02] <Keybuk> but we'll run down the list anyway
[14:02] <Keybuk> Riddell:
[14:03] <Keybuk>   kubuntu-hardy-kde4 ... still drafting
[14:03]  * mvo waves
[14:03] <Keybuk> I assume that this is pending our current discussion in <-- that window
[14:03] <Keybuk> ?
[14:03] <Riddell> Keybuk: mm, yes, I'll got onto it after the meeting
[14:04] <Keybuk> and kubuntu-hardy-catchup is approved
[14:04] <Keybuk> are there any other kubuntu specs that should be goals for hardy?
[14:05] <Riddell> only KDE parts of other specs
[14:05] <Keybuk> could you mail me a list of those parts, and who is going to do them
[14:05] <Keybuk> just one liners for each is fine
[14:05] <Riddell> ok
[14:05] <Keybuk>   "some-random-spec: boyrabbit has to implement a Qt frontend"
[14:05] <Keybuk> or something
[14:06] <Keybuk> ok, Ken
[14:06] <Keybuk> art-team we discussed -- you're going to work on getting that to review today?
[14:06] <kwwii> Keybuk: yes, I've already started...should be done in a bit
[14:06] <Hobbsee> Keybuk: the one about world domination, of course.
[14:07] <Keybuk> and hardy-theme and hardy-icon-theme are awaiting the outcome of decision by Mark etc.
[14:07] <kwwii> Keybuk: right, I guess that for now they are as final as we can get without a definite decision
[14:07] <Keybuk> still no pitti
[14:07] <kwwii> but they are structured such that we could probalby go ahead with them
[14:08] <Keybuk> and mvo and I have yet to bash out his spec list
[14:08] <mvo> Keybuk: I can try to call him
[14:08] <mvo> yeah, my spec list got a bit out of hand
[14:08] <Keybuk> mvo: please do
[14:09] <Keybuk> MacSlow: you have gdm-face-browser at review
[14:09] <Keybuk> (waiting on me, but I can give you immediate feedback -- still needs the > 100 mockup :p)
[14:10] <seb128> new gdm doesn't look good for information
[14:10] <Keybuk> seb128: how do you mean?
[14:10] <StevenK> TBH, I'd prefer to sleep, but I'm here. :-)
[14:10] <seb128> Keybuk: no way to select the language or session, no gdmflexiserver, no gdmsetup and SVN moves moderately quickly
[14:10] <mvo> pitti will be here soon
[14:11] <MacSlow> Keybuk, hm
[14:11] <pitti> hi! I'm terribly sorry
[14:11] <Keybuk> MacSlow: also you have hardy-desktop-effects to complete drafting of
[14:11] <Keybuk> and one I snuck in to your list today, so I won't nag about that
[14:11] <seb128> hey pitti
[14:11] <Keybuk> pitti: hey
[14:12]  * pitti mixed it up with 1600, sorry
[14:12] <Keybuk> StevenK: you've been drafting about-this-computer, and it's now ready for review again?
[14:12] <StevenK> Keybuk: Yes.
[14:12] <MacSlow> Keybuk, 100 is the threshold were I want to disable the face-browser... for numerous reasons... one is performance on low-end hardware with very little ram... the other is images become too small to be cleanly recognizable... especially on low resolution screens
[14:12]  * mvo hugs pitti
[14:12] <Keybuk> MacSlow: right, so that's the missing mock-up -- a username box and example flow for it
[14:13] <StevenK> Keybuk: Well, I think I've addressed all of your concerns
[14:13] <MacSlow> Keybuk, ok
[14:13] <Keybuk> StevenK: ok, cool; Ted will own that spec, though I appreciate you'll want to help out as your -mobile time permits
[14:14] <Keybuk> (but I'm not going to ask for any time, since I've only just had a rant at david for stealing people on my team :p)
[14:14] <StevenK> Haha
[14:14] <Keybuk> mpt: you're the drafter for exit-strategy -- that spec needs a little more tidying up before review, how's that going?
[14:14] <StevenK> Keybuk: TBH, I'm perferctly happy owning it, I was planning on doing it outside core hours
[14:15] <MacSlow> seb128, I don't really like to have these gdm-setup things exposed directly in the greeter itself
[14:15] <mpt> Keybuk, I'm finishing off the design right now (it'll be another hour or three), but I don't know who can write the Implementation section
[14:15] <Keybuk> StevenK: but if it's on my roadmap (which it is), then I can't rely on you being able to have enough free time :-/
[14:15] <Keybuk> mpt: Ted should be able to help with that
[14:16] <mpt> ok
[14:16] <MacSlow> seb128, after all that admin work... and a admin can log in and use the tool from the system menu
[14:16] <StevenK> Keybuk: Sure.
[14:16] <Keybuk> pitti: we discussed yours earlier
[14:16] <pitti> yep; any questions from other folks?
[14:16] <Keybuk> restricted-manager-rewrite and partition-management are approved
[14:17] <Keybuk> hardy-reducing-duplication is awaiting cjwatson approval?
[14:17] <pitti> right, doko gave his thumbs-up
[14:17] <Keybuk> policykit-integration awaiting comments from kees and cjwatson?
[14:17] <pitti> correct
[14:17] <Keybuk> ok
[14:17] <pitti> I was planning on working on this, so I hope I won't get it implemented before approval :-P
[14:18] <Keybuk> :-)
[14:18] <Keybuk> but it's so much fun when that happens
[14:18] <StevenK> Hah. A few mobile specs are in danger of that
[14:18] <Keybuk> I enjoy it when the specification process is subverted
[14:18] <StevenK> In that case, I've just uploaded about-window, enjoy.
[14:18]  * StevenK ducks
[14:18] <Keybuk> and for the record, on my plate are hardy-hardware-detection which is a spec that involves several pieces from several people across several teams
[14:18] <Keybuk> so I'll be chasing it
[14:19] <Keybuk> and I'll also be keeping an eye on prefetch, which is hopefully just a matter of nudging upstream for it and nudging the kernel team
[14:19] <Keybuk> (and running lots of tests in vmware to compare with readahead)
[14:19] <Keybuk> are there any other desktopish specs we should go over?
[14:19] <mpt> packaging-tools-usability
[14:20] <mpt> I completed that in a hurry and perhaps mvo would like to give it a sanity check?
[14:20] <mvo> thanks for drafting that mpt
[14:20] <mvo> mpt: I looked over it a couple of minutes ago and it looked good, I need to compare it to my notes but IIRC it covers the session pretty well
[14:20] <Keybuk> mvo: ok, update the status whiteboard once you're happy
[14:21] <Keybuk> I think that's in my approval queue
[14:21] <mvo> ok
[14:21] <mpt> I made some arbitrary guesses about stuff that's too big for Hardy
[14:21] <MacSlow> When can I throw in the "display names with user-photos in the face-browser"-question?
[14:22] <Keybuk> MacSlow: presently; we'll clear off the spec list, that's a topic on the agenda :)
[14:23] <Keybuk> ok, no other specs ;)
[14:23] <Keybuk> MacSlow: right, your topic :-)  explain and we'll debate
[14:23] <MacSlow> will I look stupid when I ask for the place to read up on the agenda?
[14:23] <MacSlow> ah... ok
[14:24] <MacSlow> so Kees had the objection that displaying user-names with the users photo in the face-browser is a security-hole
[14:24] <MacSlow> http://people.ubuntu.com/~mmueller/face-browser-3.png
[14:24] <MacSlow> see that mockup to get an idae
[14:24] <Keybuk> you're not actually showing usernames there though?
[14:24] <Keybuk> you're showing users names
[14:24] <MacSlow> correct
[14:24] <MacSlow> it's showing the real names...
[14:24] <MacSlow> for two reasons...
[14:25] <MacSlow> one is to not expose the actualy login-id of a user
[14:25] <MacSlow> the second is to be more real world like
[14:25]  * pitti likes that better, too
[14:25] <Keybuk> to me, this seems to alleviate kees' concerns
[14:26] <MacSlow> third is to help with the possibility of two or more people having chosen the same image from the stock image set provided
[14:26] <pitti> not hard to guess the usernames then, of course
[14:26] <pitti> (OTOH kdm has done this for ages)
[14:26] <Keybuk> http://www.guidebookgallery.org/pics/gui/startupshutdown/login/macosx103-1-1.png
[14:26] <StevenK> pitti: Maybe. "Martin Pitt" -> "pitti" isn't an obvious mapping, for example
[14:26] <pitti> StevenK: installer defaults to a lower-case variant of the name
[14:26] <StevenK> kdm shows both the real name and username
[14:26] <MacSlow> well I think it not necessary easy to guess login-id from a real name... very often login-ids are totally abstract jibberish
[14:27] <MacSlow> especially on larger installations
[14:27] <pitti> right (where it matters more, too)
[14:27] <mvo> and it does not help with the password (as long its not a joe password)
[14:27] <Keybuk> I think that not displaying some form of name is a usability problem
[14:27] <pitti> abstract jibberish like "MacSlow" for example :-P
[14:27] <Keybuk> especially for abstract pictures
[14:27] <MacSlow> pitti, but you're right... sometimes it can be easy to map back
[14:27] <kwwii> we had a long discussion about this same issues years ago at suse when we switched to a face browser, in the end those who it worries can turn it off
[14:27] <pitti> MacSlow: I'm not overly concerned about this TBH
[14:28] <MacSlow> pitti, right again... I think It's hard to map MacSlow to "Mirco Müller"
[14:28] <Keybuk> http://www.winsupersite.com/images/showcase/winxp_login.jpg
[14:28] <pitti> meh, WinXP doesn't even assign passwords by default :)
[14:28] <MacSlow> yeah... and MacOS does show real names too
[14:28]  * pitti was shocked when he installed his very fist XP a few days ago (my last install was 3.11, duh)
[14:28] <Keybuk> pitti: what is the concern with revealing the username?
[14:28] <MacSlow> pitti, what... you're kidding?
[14:29] <Keybuk> what extra thing does it actually give the attacker?
[14:29] <pitti> Keybuk: one half (the easy one, admittedly) of getting an account
[14:29] <Keybuk> pitti: but what does that allow them to do?
[14:29] <MacSlow> Keybuk, I have no idea... I'm not a network/security person
[14:29]  * mpt is reminded of the idea for a login system that doesn't use login names at all, just two-word passwords
[14:29] <Keybuk> they can get that by *clicking on the picture* :)
[14:29] <pitti> Keybuk: right, but they can't do that right now in gdm
[14:30] <Keybuk> sure, so the problem isn't "names visible on the face browser"
[14:30] <Keybuk> it's actually "face browser"
[14:30] <pitti> but as I said, it's only a minor point, and I'm not overly concerned about it
[14:30] <Keybuk> since the face browser inherently reduces the security to just a password because you can click on the picture to get the username half entered for you
[14:30] <mpt> OS X makes it a preference, probably for that reason
[14:30] <Keybuk> and security-concious can always turn off the face browser
[14:30] <pitti> I just remember people squeaking when ssh allowed you to detect the validity of an username
[14:30] <pitti> (timing attack)
[14:31] <pitti> but that's a different use case class
[14:31] <pitti> with local access like gdm I don't think it's an important problem
[14:31] <Keybuk> ok
[14:31] <Keybuk> anyone else have any thoughts?
[14:32] <MacSlow> So I take from that that "real names" are fine as a default, with the option available via gdm-setup to disable them
[14:32] <pitti> are there some defaults for the faces? like picking the next free picture that's shipped by default?
[14:32]  * ogra has lots of thoughts ... but wont tell Keybuk :P
[14:32] <pitti> MacSlow: that will look very friendly
[14:32] <pitti> (to avoid all users having the same default pic, which would be confusing)
[14:33] <MacSlow> pitti, I'm getting my head together with kwwii and andrean to provide a set of 100 256x256 stock-images for this case
[14:33] <pitti> 100! wow
[14:33] <StevenK> Keybuk: So, can I get paroled? It's 1:30am here... :-)
[14:33] <MacSlow> kwwii, said email went out about 30 min. ago in case you're wondering
[14:33] <Keybuk> StevenK: go to bed :)
[14:33] <pitti> MacSlow: I expected something like 5 :)
[14:33] <StevenK> Keybuk: Thanks :-)
[14:33] <kwwii> not sure if we can find/make 100 different pics but we need a certain amount to make it look nice
[14:34] <kwwii> we definitely need to make the process of adding your own pic as easy as possible
[14:34] <mpt> Users & Groups should have a "Take My Photo" button
[14:34] <mpt> if a camera is detected :-)
[14:34] <MacSlow> pitti, yes... 100 is currently the threshold value for the face-browser... but I picked that a bit randomly I admit since I've not tested performance of this case on a low-spec GPU (lowest thing I have available is a I915)
[14:34] <kwwii> mpt: good idea
[14:35] <ogra> mpt, shouldnt that be rather "about me" ?
[14:35] <mpt> ogra, I'm looking forward to the day when they're one and the same
[14:35] <ogra> yeah
[14:35] <Keybuk> ok, we're drifting off the original topic here
[14:35] <MacSlow> kwwii, I've added MeMaker and cheese as potential "means" for image-aquisation in addition to the stock image-set
[14:35] <Keybuk> so appears we have a consensus that displaying the names is fine as long as the security concious can turn it off (even if that means losing the face browser altogether)
[14:35] <ogra> mpt, that will be after the day where screenlocking, fusa and gdm are one app :)
[14:35] <kwwii> MacSlow: cool
[14:36] <MacSlow> kwwii, MeMaker creates abstract (south-park-ish) iamges from SVG graphics... cheese is a webcam-app like PhotoBooth
[14:36] <pitti> will gdm, fusa, and gnome-screensaver be merged to something that provides all three use cases eventually?
[14:36]  * ogra was planning to pull MeMaker into main as game for edubuntu anyway
[14:36] <Keybuk> mvo: you have the next item - syncs and moms
[14:36] <kwwii> MacSlow: yeah, I've heard of cheese before, never MeMaker though
[14:36] <ogra> pitti, wouldnt that be cool ?
[14:36] <MacSlow> ogra, cool be my guest :)
[14:36] <pitti> absolutely
[14:36] <mvo> Keybuk: yeah, it got discussed on the mailinglist already a bit
[14:37] <mvo> Sometimes it take some time for a sync-from-debian request to get
[14:37] <mvo>    processed. The relevant package still appears in MoM even when the
[14:37] <mvo>    bug is filed and its possible that some work gets duplicated
[14:37] <mvo>    when people do not check the bugpage first and look at the open
[14:37] <mvo>    merge. Maybe we could teach MoM about requested syncs? Or get
[14:37] <mvo>    some limited archive-admin power to do syncs? The more patches we
[14:37] <mvo>    forward the more often we will want syncs instead of merges.
[14:37]  * pitti points out that he has a script to sync packages yourself
[14:37] <MacSlow> kwwii, ttps://edge.launchpad.net/memaker
[14:37] <pitti> when handled with care, it should be absolutely ok
[14:37] <mvo> it seems like a good option is to teach MoM about pending sync requests
[14:37] <pitti> http://people.ubuntu.com/~pitti/scripts/syncpackage
[14:38] <mvo> pitti: oh? isn't that something that requires archive-admin power?
[14:38] <pitti> my main concerns are:
[14:38] <pitti> - we lose the papertrail for bugs
[14:38] <Keybuk> mvo: is there any easy way to tell?
[14:38] <pitti> - we lose peer review
[14:38] <pitti> mvo: no, it's just a normal upload
[14:38] <Keybuk> mvo: do you have a url listing sync requests?
[14:38] <Hobbsee> mvo: only that it signs with your key, rather than the installer key
[14:38] <pitti> https://edge.launchpad.net/~ubuntu-archive/+subscribedbugs?field.searchtext=sync
[14:38] <mvo> Keybuk: just the heuristics, open bug with sync and debian + archive-admin is subscribed
[14:38] <pitti> ^ list
[14:39] <pitti> mvo: that's in fact how a sync bug needs to look like
[14:39] <pitti> mvo: you just need archive admin power to use sync-source.py on drescher
[14:39] <pitti> but in the end this does nothing else than my script, with just some additional sanity checks
[14:39] <mvo> aha, I see.
[14:40] <Keybuk> hmm
[14:40] <Keybuk> how do I turn that into a feeds URL?
[14:40] <pitti> so it's not so much a technical issue, more a question of which kind of documentation we want for syncs we did in the past
[14:40] <pitti> Keybuk: isn't that a LP feature they are already working on?
[14:40] <mpt> Keybuk, the answer is "not yet"
[14:40] <Keybuk> oh
[14:41] <mvo> my only concern is duplication of work when people look at MoM output and think that this is a open merge when in facts its already dealt with
[14:41] <Keybuk> can probably beautiful soup that table
[14:41] <mpt> The eventual answer will be "click the orange button in the URL bar"
[14:41] <seb128> mvo: merges have a name next to them
[14:41] <Keybuk> I'll add it to MoM about the time I fix the most recent fuckage :p
[14:41] <pitti> mvo: I dont' think that's a big problem, though; if you grab someone else's merge, you should contact him first anyway
[14:41] <mvo> seb128: and that name is often e.g. Ian Jackson (so unclaimed by default)
[14:42] <Hobbsee> mpt: \o/
[14:42] <seb128> mvo: usually it's good practice to tell to whoever is assigned that you will do it ;-)
[14:42] <pitti> mvo: good point about that
[14:42] <seb128> right
[14:42] <mvo> seb128: see above, there are some people where it does matter
[14:42] <pitti> mvo: so you're going to merge dpkg? *hug*
[14:42] <seb128> I've no conflicted yet with anybody
[14:42] <seb128> I don't think it happens that often
[14:42] <mvo> pitti: I did everything else from ian list ...
[14:42] <Keybuk> ok
[14:42] <Keybuk> did I miss any other agenda items?
[14:43] <pitti> mvo: *more hugs*
[14:43] <mvo> if everybody thinks its not a issue, I will shut up :)
[14:43] <seb128> mvo: well, I don't say it never happens but I don't think we got many conflicts
[14:43] <seb128> who here duplicated work on merges this cycle?
[14:43] <Keybuk> mvo: even if not, it's not that difficult to add
[14:43] <seb128> (just curious)
[14:43] <MacSlow> not me
[14:43] <Hobbsee> seb128: it's more common in universe.
[14:43] <MacSlow> :)
[14:43] <Keybuk> I can make it hide table rows, or have a different table
[14:43] <Hobbsee> so universe would appreciate it, at least.
[14:43] <seb128> Hobbsee: for universe you reinvented the wheel anyway and have a comment thingy ;-)
[14:44] <mvo> Keybuk: maybe just a small comment (just like currently the uploaders) and a differnt color?
[14:44] <seb128> but right
[14:44] <mvo> possible with link?
[14:44] <seb128> having a such feature would be nice
[14:44] <Hobbsee> seb128: some of them did, but i tend to prefer mom :P
[14:44] <Keybuk> mvo: I have mad ajax idea for submitting comments <g>
[14:44] <mvo> heh :)
[14:44] <mvo> web 2.0!
[14:44] <seb128> web 3!
[14:45] <Hobbsee> web 42!
[14:45]  * MacSlow eyes bleed
[14:46] <pitti> mine too, I just looked at our dpkg delta
[14:46] <Keybuk> ok
[14:46] <Keybuk> any other agenda items?
[14:46] <seb128> pitti: dpkg is clearly a task for Keybuk ;-)
[14:46] <seb128> not an item
[14:46] <pitti> I'm sure he'll jump for joy
[14:47] <Keybuk> hmm? :)
[14:47] <seb128> bug I want to point https://wiki.ubuntu.com/UbuntuDevelopment/PatchTaggingGuidelines
[14:47] <seb128> from now desktop packages expect proper tagging
[14:47] <seb128> and patches must be sent upstream
[14:48] <seb128> s/bug/but
[14:49] <Keybuk> seb128: could you add an example of a patch-with-a-comment to that?
[14:49] <Keybuk> do you mean just placing the text above the patch?
[14:49] <Keybuk> e.g.
[14:49] <Keybuk> Ubuntu: ...
[14:49] <Keybuk> diff -r ...
[14:49] <MacSlow> seb128, does that include also the work from last cycle?
[14:49] <seb128> Keybuk: could you make merges.ubuntu.com/extracted list the debian/ubuntu-applied-patches directory also if there is one?
[14:49] <seb128> Keybuk: ok, will do
[14:50] <MacSlow> seb128, I know vuntz waits for my cleaned up patches anyway
[14:50] <pitti> seb128: maybe we should generally use # Tag:
[14:50] <pitti> to use the exact same syntax for dpatches
[14:50] <seb128> MacSlow: no need to modify everything right now, but tagging patches on the way when you do a merge or an update would be nice
[14:50] <seb128> and new patches should be tagged
[14:50] <Keybuk> seb128: what's in that directory?
[14:50] <pitti> Keybuk: it's only used for packages without a patch system; the broken-out patches
[14:50] <seb128> Keybuk: a copy of the diff we apply when the package use no patch system
[14:50] <Keybuk> ah, cool
[14:51] <Keybuk> yes, added to todo
[14:51] <seb128> thanks
[14:51] <Keybuk> so people should
[14:51] <Keybuk>  * add tags for all new patches
[14:51] <Keybuk>  * add tags when merging or updating a source package for any existing patches
[14:51] <pitti>  * submit them upstream along the way
[14:51] <seb128> right
[14:51] <Keybuk> oh, and yes, submit them upstream
[14:51] <Keybuk> forgot the most important bit :-)
[14:51] <Keybuk> and to debian, if relevant
[14:52]  * pitti is happy that all our remaining sudo patches except the ~/.sudo-successful tag one were accepted upstream
[14:52] <Keybuk> when is sudo going to use policykit? :p
[14:52] <seb128> Keybuk: if we want to make list of patches to forward, etc, would it make sense to do that in patches.ubuntu.com? It's already dealing with the datas
[14:52] <pitti> just at the moment when we switch everything to PK :)
[14:53] <pitti> Keybuk: pk_0wn_me_priv to allow everything :)
[14:53] <pitti> "get me a root shell over dbus", yay
[14:53] <Keybuk> org.2600.h4x0r.RootShell ()
[14:54] <Keybuk> seb128: yeah, if tags are added consistently in patches, then I can extract that information and generate html from it
[14:54] <Keybuk> which is why I asked for an example on the wiki page, so I can make regexp :p
[14:54] <seb128> Keybuk: excellent ;-)
[14:54] <Keybuk> ok
[14:54] <Keybuk> any other business?
[14:55] <MacSlow> not from my side
[14:55] <Keybuk> ok, then; end of meeting (with 3 min to go)
[14:56]  * mvo waves
[14:56] <seb128> Keybuk: thanks
[14:56] <kwwii> thanks everyone
[14:56] <pitti> thanks everyone
[14:58]  * MacSlow goes back to mockup-drawing
[14:59]  * kwwii goes back to spec'ing