[00:51] wallyworld: Mind reviewing a mostly-mechnical MP? https://code.launchpad.net/~stevenk/launchpad/force-bug-enums-into-line/+merge/97123 [00:51] StevenK: sure [00:53] StevenK: have you pushed a subsequent change, mp says diff still updating [00:54] I have not, that's it [00:54] and it just disappeared [00:54] so it seems to hang around for a bit [00:54] Reload the page, longpoll is dumb [00:54] i have been [00:54] and yet it stilled stayed on the page [00:55] Sigh. This is why we can't have nice things. [00:55] no, IE is why we can't have nice things [00:56] StevenK: r=me [00:59] * StevenK lp-lands it, it's very benign [00:59] WCPGW [00:59] Haha [00:59] :-) [03:19] lifeless: Are you still AFK? [03:33] StevenK: He's sick AFAIK. [03:34] Well, we both knew that already. [03:34] :-P [03:34] Hah [03:34] But okay, noted. [03:57] StevenK: future ref - staff calendar will show you [03:57] StevenK: and yes, throat full of razor blades, cold sweats, muscle pain, sinus. ye whole works. [03:58] that said, [03:58] !ask | StevenK [03:58] Also, backscroll. [03:59] lifeless: It is not urgent, and for further win, is not work-related. [03:59] lifeless: So don't worry about it. [04:00] StevenK: well, I'm here for a few minutes, so please do ask [04:03] lifeless: You need to be lucid enough to remember your GPG passphrase. :-) [04:03] go one [04:03] lifeless: http://keyring.debian.org/replacing_keys.html [04:04] I'd like to get my 1024D key out, and my 4096R key in. [04:04] heh [04:04] I should too [04:05] lifeless: I can toss you the full fingerprints in privmsg [04:06] mail, signed by your old key; isn't that part of the protocol ? [04:06] Doesn't have to be [04:06] If it still valid, I can choose to sign it with the old key [04:07] is your new key in the keyservers? [04:07] Yeah [04:07] 588A553F, ? [04:07] That's it [04:11] hmm, I don't have a trust path to it [04:12] skype ? [04:12] StevenK: ^ [04:13] Let me get that out [04:14] You don't have a trust path to C87FFC2F either? [04:14] don't seem to [04:14] of course, gpg may have lost local trust settings or whatever [04:14] whatever; will take 30 seconds to directly verify [04:16] lifeless: I'm Skype-enabled if you're ready. [04:21] * StevenK prods Skype [04:22] * StevenK waits for lifeless to ping timeout from IRC ... [05:16] huwshimi: hello, do you have any exisiting/preferred code to use for a tri-state checkbox? or a preferred implementation technique? eg css or? [05:16] tri-state checkbox? [05:16] one which can show three states [05:17] yes, no, maybe [05:17] or similar [05:17] all, nothing, some [05:17] Unchecked, checked, or greyed out check. If it is played with, it collapses to checked or unchecked only. [05:17] a checkbox with three states... is that one where you need to visit the parallel universe to see the third state? [05:18] no, it's represented typically as a grey tick [05:18] wallyworld: What's the situation? [05:18] sorry, juju has made me flippant this afternoon [05:18] huwshimi: for the sharing ui. the user needs to select to share "all", 'Some' or 'Nothing' [05:18] bigjools: fuck off [05:18] huwshimi: 'some' would be grey [05:19] 'all' would be ticked [05:19] wallyworld: Would checkboxes with those labels work? [05:19] huwshimi: tri-state is more compact. the other solution is radio buttons, but yuk [05:19] wallyworld: sorry I can't hear you over the sound of my own awesomeness [05:20] huwshimi: radio buttons would take up a bit of space [05:20] and not scale that well [05:20] bigjools: Must be pretty quiet, then. [05:20] wallyworld: The standard toolkits don't even SUPPORT tri-state checkboxes [05:20] StevenK: speak up, i can't hear you over the whinging pom [05:20] StevenK: good comeback, I must remember that one [05:20] wallyworld: :-D [05:21] huwshimi: yeah, there's several css/js based solutions people have used [05:21] wallyworld: I'd really prefer we not use a tri-state [05:21] :-( [05:21] it's a very common paradigm [05:21] bigjools: Can't tell if trolling, or ... [05:21] huwshimi: what don't you like? [05:23] huwshimi: if i resort to radio buttons, the sharing picker is sure going to look fugly [05:23] wallyworld: I'd prefer us to be much more explicit, especially as sharing the wrong thing has high ramifications. A tri-state will make the user have to think/guess about what each state might be. [05:24] wallyworld: I'm not convinced it will be ugly. If it is, we may be doing something else worng [05:24] wallyworld: And no, it's not a common paradigm in web applications (it's not even common elsewhere). [05:24] huwshimi: hmmm. ok. i'll rework the ui and see how it looks [05:24] The states are ALL, NONE and SOME. If the user changes it, it collapses to ALL or NONE. Sounds pretty clear to me. [05:24] huwshimi: we must look at different apps then [05:25] it's not that common on the web [05:25] I remember tri-state from *shudder* my time doing Visual Basic. [05:25] but what about installers as one example [05:25] where you get to choose all packages, non, or some [05:27] I don't recall seeing many tristate checkboxes outside horrifyingly awful Microsoft Office options dialogs. [05:27] wgrant: i'm sure i've used a linux installer at some point with them [05:27] maybe suse [05:27] Red Hat doesn't exist [05:28] If only [05:28] it's a very compact and convenient way to represent all/some/none [05:28] And it makes it rather challenging to recover. [05:28] Indeed. You can't go back to Some if you click the wrong checkbox [05:29] I hate the idea of comboboxes for each too, since that's just disgusting [05:29] no, it's tri-state [05:29] wallyworld: The most common use for a tri-state is to represent the state of other checkboxes on the page. [05:29] as you click, it cycles through the states [05:29] wallyworld: I don't think I've ever seen one of those before. [05:30] AFAICR they just collapse into one of the two extremes. [05:30] huwshimi: yes, that's true [05:30] You can never return to some. [05:30] That's what I recall. [05:30] i'm clearly outvoted :-) [05:30] i'll do it as radio buttons [05:30] Ewww [05:31] what alternatives do i have? drop down combo? [05:31] i guess i could do that [05:31] New widget, but ... [05:31] That would be even worse, I think. [05:31] sooo, we don't like tri-state checkeboxes, nor radio buttons, nor drop down combo; there ain't much kleft [05:32] |-- ALL / -|- SOME / --| NONE [05:32] If you excuse my horrid ASCII-art, that's my idea, but I don't like it either. [05:32] I was thinking maybe a lozenge radio button. [05:32] Which is slightly less obese than full radio buttons. [05:32] wallyworld: Can you show me the situation somehow? [05:32] maybe [05:33] huwshimi: https://qastaging.launchpad.net/launchpad/+sharing [05:33] wgrant: Can haz example? I've never heard of those [05:33] huwshimi: click on the +Share with someone link [05:34] huwshimi: choose a person and then see the 2nd step where you get to chosse the sharing permissions [05:34] huwshimi: there are currently checkboxes but they need to be tri-state [05:34] StevenK: The [ Test | One | Two ] on http://arstechnica.com/apple/reviews/2011/07/mac-os-x-10-7.ars/3 [05:35] huwshimi: lozenge buttons seem the best suggestion so far i think [05:35] wgrant, wallyworld: That was what I meant with my ASCII-art [05:35] sinzui will cry, as they have bad history in Launchpad. [05:35] But for unrelated uses. [05:35] StevenK: Oh, I assumed your thing was more like the horrible GTK3 toggle things. [05:36] StevenK: ah of course. you also printed out naked women on dot matrix tractor printers and stood 100 feet away too didn't you? [05:36] With the line that's either on the left or right. [05:37] wallyworld: No ... it was the fastest way to get my point across. [05:37] Sadly, wgrant did it better, like usual. [05:37] StevenK: just giving you shit [05:37] I know. ;-) [05:37] wallyworld: Since you're full of it, it has to go somewhere ... [05:37] lol [05:38] huwshimi: you find what i mean ok? [05:38] wallyworld: I did [05:38] huwshimi: i await your verdict :-) [05:38] * StevenK votes for lozenge radio buttons. [05:40] * wallyworld so does wallyworld [05:41] do we have any code in lp for those? [05:41] Curious how to do them in YUI [05:41] surely there's a widget someone has donr [05:42] Do we use the lozenge anywhere else in Launchpad? [05:42] We used to have lozenge menus, but they were abolished in 3.0. [05:43] And sinzui probably danced on their grave. [05:43] We all did. [05:43] It was a glorious day. [05:44] If we don't use the elsewhere then I don't think now is the right time to be introducing new UI concepts [05:44] wallyworld: What you might have to do is produce a few really quick mockups so we can see the options [05:44] Sure, but what else can we use that doesn't look horrible? [05:44] But the concept is the correct way to show a tri-state ... [05:45] I'm sure we can do it with any of the options and not make it horrible [05:45] but introducing unusual UI concepts probably isn't the smartest idea right now [05:46] huwshimi: lozenge maybe new for lp but not outside of lp. and lp is very primative so i wouldn't use it's current widget set as a guide to the current best practice [05:46] wallyworld: Right, but no-one has enough time to work on a new UI concept to the point where we get it right [05:47] if not lozenge, then the only choices are radio button and drop down. i can see what they both look like [05:47] And they'll both be terrible, I fear [05:47] me too :-( [05:52] wallyworld, wgrant, StevenK: Tell me why all the rest of the options are terrible. [05:52] I want to be convinced [05:53] huwshimi: radio buttons are too verbose and not that well suited to being displayed for each item in a list [05:53] huwshimi: drop downs don't allow the user to easily see all of the choices available without clocking [05:54] Yes, does a grid of 21 radio buttons make for good UI? [05:54] wallyworld: How are radio buttons more verbose than the lozenge [05:54] ? [05:54] StevenK: There will be a maximum of nine [05:54] StevenK: Usually six [05:55] I doubt it, for large projects. [05:55] huwshimi: they are more compact, radio buttons have the clickable widget and text separate [05:55] wallyworld: We have plenty of space [05:56] There are *better* ways to display tri-stated information than an arry of radio buttons. [05:56] Just thinking about it makes me want to stab my eyes. [05:56] huwshimi: i think guis are better when they don't misuse space [05:57] StevenK: I'm not convinced that we can do a better job of building them though [05:58] I think I'd rather us do something that we can't go wrong with than introduce a new UI concept that will most likely introduce a support and maintenance cost [05:59] I think we have to mock up the three choices [05:59] I'm not sure a dropdown is a choice. [05:59] Neither. [05:59] i guess it's hard sometimes for people who understand these ui concepts to see how it would introduce a support cost [06:00] it's not like lp is aimed at people who don't know computers [06:00] and if you are using the sharing ui and you can't use a lozenge or whatever then you have no business in the ui [06:01] wallyworld: Sure, but given our track record of UI design, creating a lozenge widget for example has a high risk of failure. It's not just that the user won't be able to figure it out, it's that we won't do a good enough job at creating something that works perfectly. [06:02] oh, the pain, stabbed in the heart [06:05] drama queen. [06:05] * nigelb ducks and runs [06:06] wallyworld: We have to consider these things when making decisions. Especially as I'm not sure we'd gain enough to warrant the extra development time (forgetting adding another inconsistent UI concept etc.) [06:08] huwshimi: np. there's a line between "adding inconsistent ui" and "using the right tool for the job" and i guess that's what needs to be looked at [06:09] wallyworld: I think if we do some mockups we can make a decision pretty quickly [06:10] huwshimi: ok. thanks for the input, appreciated [06:10] nigelb: i wouldn't stick my head up after what happened in the cricket :-P [06:11] dammit [06:11] :P [06:40] wallyworld: Oh, nice, a dict marshaller for lazr.restful. How does that affect the WADL? [07:37] wgrant: not sure about what you mean by the effect on the wadl. it generates fine and looks ok but i'm no expert. all the web services tests that invoke that method pass [07:41] wallyworld: Right, just wondering how easily we can convince launchpadlib to interpret it well. [07:43] service.sharePillarInformation(pillar=ws_pillar, sharee=ws_grantee, [07:43] permissions={InformationType.USERDATA.title: SharingPermission.ALL.title}) [07:43] wgrant: you just pass in a dict as expected [07:43] wgrant: the keys are values are marshalled [07:44] wallyworld: Ah, right, it's only going the other way that it'll need special code. [07:44] Methods that *return* dicts === almaisan-away is now known as al-maisan [07:45] wgrant: there's an unmarshall method which is properly implemented (generates json data). will that be used for return types? [07:45] wallyworld: The server side has always worked, AFAIK [07:45] I wrote a couple of methods which returned dicts [07:46] someone filed a bug related to returning dicts from memory [07:46] lazr.restfulclient doesn't know what to do with them, so it just returns them as dicts of strs. [07:46] bug 481090 [07:46] <_mup_> Bug #481090: Cannot define a method that returns a dict < https://launchpad.net/bugs/481090 > [07:47] ah, ok, lazr.restfulclient [08:14] StevenK: signed and pushed [08:14] lifeless: Thanks. I've submitted an RT (two in fact, because I'm a complete blockhead); we shall see what keyring-maint says. [08:26] goood morning [08:43] czajkowski: Proof required. :-P [08:46] I have tea and oddles of toast with peanut butter nothign going to upset me :) [09:05] g'morning launchpadders === gmb changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: gmb | Firefighting: - | Critical bugtasks: 4*10 [10:33] Does anyone else see this whilst trying to use the bzr beta PPA?: http://pastebin.ubuntu.com/881655/ [10:34] gmb: they are not publishing precise packages yet [10:34] Ah. [10:34] Bottom. === al-maisan is now known as almaisan-away [10:35] is precise not beta enough for you? :) === almaisan-away is now known as al-maisan [10:35] I want my bleeding edge to look like it came out of an accident with a bandsaw. [10:36] You want ppa:bzr/daily :) [10:36] :). [10:37] christ, my desktop PC has gone 3 weeks with no net connection and now dist-upgrade has taken an hour so far (50 mins of that actually installing) [10:37] wgrant, You may know the answer to this: Is it wise or unwise to use setup.py from within the Makefile of a debian package? [10:38] joys of precise :) [10:42] gmb: If Makefile == debian/rules, then how else are you going to do it? [10:43] gmb: debian package ... Makefile? you mean debian/rules? [10:45] also re bzr ppa, at the moment bzr/beta == bzr/proposed [10:45] er [10:46] also re bzr ppa, at the moment bzr/beta == bzr/ppa [10:46] maxb, StevenK: Well, yes (AFAICT that's calling the Makefile in the top level of the source, which is that I'm referring to). [10:46] gmb: Oh, right, your build step calls $(MAKE) [10:46] urgh, you have a horrid package with both a setup.py and a Makefile [10:47] maxb, It mixes python and shell stuff. [10:47] I'm considering doing the python bit manually though [10:47] Because setup.py just vomits. [10:47] ... when the package is being built, anyway. [10:47] Actually, maybe it would be simpler to have a separate package for the python stuff. [10:47] Doing the python bit in Make will cause you to put your keyboard cord around your neck and yank hard. [10:49] StevenK, Ah, I see. Separate package is looking very tempting, then. [10:49] Screw it, the original package can depend on the new one. [10:49] WCPGW [10:49] Hah. [10:50] StevenK, maxb: Thanks chaps. [10:50] At least I know i'm not entirely mad. [10:51] But you don't have that far to travel. [10:52] StevenK, Well, no, I still work on Launchpad. "Madness" is within spitting distance. [10:57] gmb: What's the package? [10:58] wgrant, charm-tools. [10:58] * StevenK reads a conversation on another channel, and reaches for his keyboard so he can wrap the cord around his neck, and then realises it's a wireless keyboard ... [10:58] Hm, so -server wrote it? [10:58] wgrant, Yes, I think so. [10:59] Odd that it is so awkward, then. [10:59] wgrant, Well, what's awkward is that I'm adding Python stuff to it. It worked perfectly until I started pratting about. [10:59] Oh [10:59] Anyway, I'm going to bug clint about it this afternoon :) [10:59] That would do it. [10:59] Separate package :) [10:59] Yeah :) [11:00] Ooh. [11:00] Parallel test suite bugs. [11:00] You have something working, then? :) [11:02] wgrant, "Working" is relative. But yes, we do, now that we've solved a lot of the weirdnesses with lxc and -start-ephemeral. [11:04] Heh. [11:04] Good good. [11:17] StevenK: #firstworldproblems [11:20] Holy mother of dog and all her wacky nephews. Why are my comments on merge proposals posted with all the spaces stripped out? What's this nonsense? === al-maisan is now known as almaisan-away === matsubara-afk is now known as matsubara === danhg_ is now known as danhg === almaisan-away is now known as al-maisan === Ursinha` is now known as Ursinha [14:48] benji: have you looked at wallyworld's lsazr.restful branch yet? [14:48] i was about to write a review for it [14:48] but you are a pending reviewers [14:48] fjlacoste: I wasn't aware of it. I would be happy to look and equally happy for you to do so. [14:49] benji: i'll take care of it [14:49] I wonder why I wasn't aware of it. I'll try to figure that out. [14:54] benji: i ifnd when i'm working off the web UI i sometimes miss things b/c i was looking at /launchpad/+activereviews instead of /launchpad-project/+activereviews? [14:54] ... that shouldn't have ended in a question mark. [14:55] jcsackett: ooh, that may be it; I hope firefox auto-complete learns the better one quickly ;) [14:56] benji: i hope it learns it faster than chromium's did for me. :-) [14:56] heh [14:58] benji: well, i agree that it's weird unless he explicitely asked a review from you [14:58] because i would have expected that to be assigned to the team [15:00] fjlacoste: he did ask for a review from me (at least that's what it says now), and since the email looks so much like all the non-personal review requests I see I passed over it. I'm in the process of making messages like that highlighted so they won't slip through again. [15:21] gary_poster: I bet you're already doing this, but it dawned on me that using -vvvvvv with the test runner would show me the actual order in which the tests were run [15:21] gary_poster: you also know that I'm in the wrong channel [15:22] benji yes :-) [15:22] I'm using -vv, which is giving me the order [15:22] I'm switching channels! [15:23] yeah, never remember how many Vs do what, so I just hold down V until I get tired === al-maisan is now known as almaisan-away === beuno is now known as beuno-lunch === beuno-lunch is now known as beuno === salgado is now known as salgado-lunch [16:12] jcsackett, I reopened bug 933794 and updated it. +audit-sharing still needs a summary of users, policies and level of sharing [16:12] <_mup_> Bug #933794: +audit-sharing does not show a block summarising who is shared with

< https://launchpad.net/bugs/933794 > === salgado-lunch is now known as salgado [17:00] sinzui: dig. [17:01] sinzui: thought that might be a new bug; i'm good with just changing the existing one though. [17:15] fjlacoste: any idea why on lp all the footers with Contact Launchpad Support end up https://answers.launchpad.net/launchpad [17:25] czajkowski, I know [17:26] sinzui: oh do tell :) [17:26] dont leave me hanging :) [17:27] czajkowski, We want users to ask question in a place we support rather then sending email to a place to be ignored [17:27] sinzui: so where are people finding the feedback mail address [17:27] and how am I getting questions via RT then [17:28] czajkowski, The link is not static. anonymous users see a different link because only logged in users can ask a question..and anonymous users are probably trying to login which is not something we at Lp can help with [17:28] czajkowski, anonymous users send those [17:28] hmm ok [17:28] thanks sinzui [17:29] czajkowski, many rt emails are in reply to emails generated from Lp, such as commercial inquiries and translation issues [17:30] the commercial folks are using feedback as well [17:30] czajkowski, anonymous users see a link to https://help.launchpad.net/Feedback [17:30] just curious as spring cleaning rt [17:30] sinzui: cheers for the ifo [17:31] czajkowski, I hope that changes I making these week will reduce the emails. I am replacing the generic license issues message with ones specific to the case. This will land in a few days [17:31] The generic email is confusing [17:31] excellent === fjlacoste is now known as flacoste [18:49] I see we have some BugTasks which have a productseries bug no product, and I also see IBugTask.target can be an IProductSeries. however, I cannot seem to figure out how to create (in a test) a bugtask with a productseries but no product... the factory method won't do that and if I try to transitionToTarget(series) I get an error: [18:49] IllegalTarget('Series tasks may only be created by approving nominations.',) [18:49] heh, now I think I understand... maybe I need to create *and* approve a nomination? [18:50] conjoined masters [18:53] lifeless, you mean this is a conjoined master or I need to figure out what conjoined masters are and create one? [19:07] conjoined masters - any series only task gets a context task added if none exists [19:08] there should be no series only task bugs around [19:08] search for conjoined should give you some context [19:12] lifeless, oh, ok, that makes sense. but I don't actually want a bug with series-only tasks... all I want is a series-only task which AFAICT is created via nominations? [19:30] salgado: thats my point, there isn't such a thing atm [19:30] salgado: can you point to a bug that demonstrates what you want in prod / qas / s ? [19:30] bah [19:30] iwlwifi has a lot to answer for [19:31] lifeless, a series-only task? there seems to be 14177 of them on staging [19:31] salgado: how are you determining that ? [19:31] select count(*) from bugtask where product is null and productseries is not null; [19:32] ah terminology [19:32] thats a series task, not series only [19:32] tasks are always one-of product|productseries|distro|distroseries|distropackage|distroseriespackage [19:33] the conjoined master means there is always a matching product task whenever there is a product series task, and likewise for the distro+distroseries and distropackage+distroseriespackage [19:33] salgado: you can make a series task by nomination or targeting; if this is for the test suite, makeBugTask(target=someseries) should do it [19:34] lifeless, right, I understood those tasks with series but no product will have a matching task for the product. but what I wanted was just to know how to create such a task with a productseries but no product [19:34] salgado: via the UI, API or python ? [19:36] lifeless, I couldn't get makeBug() to do that. it seems to always set the product on the task as well. I've now managed to do it via makeBugNomination(), though [19:37] I think I was expecing makeBug() to create a single task bug it actually creates two in that case [19:38] for product and productseries [19:38] salgado: uhm, the db won't let that happen [19:38] Check constraints: [19:38] "bugtask_assignment_checks" CHECK ( [19:38] CASE [19:38] WHEN product IS NOT NULL THEN productseries IS NULL AND distribution IS NULL AND distroseries IS NULL AND sourcepackagename IS NULL [19:38] salgado: so yes, it was making two tasks [19:38] and you then need to grab the task you want [19:40] right, yeah, thanks for the help, lifeless [19:40] no worries [19:40] this is fugly [19:40] I want to fix, but one thing at a time [20:08] sinzui: did the recent 404 changes drop showing the oopsid ? [20:09] sinzui: (It is fine if they did, just seeking confirmation it was deliberate, or whether I need to investigate something being bokred) [20:10] sinzui: the breadcrumbs on https://launchpad.net/~ubuntu-security/+archive/ppa/+build/3244524 include links that will 404/403 [20:10] lifeless, oops rules should not change [20:11] ah indeed [20:11] it is a 403 which does not oops [20:11] even though it matches the same bad-internal-link rule when I think about it as a human [20:12] sinzui: (the ppa is private, the build is unembargoed, but the breadcrumbs still link to the ppa) [20:12] lifeless, I had not consider that case. I think the user who can see that page gets Lp.LimitedView thus 200 [20:12] A 404/403 page does not include breadcrumbs [20:12] sinzui: the build page is 200 [20:12] well my test show that error pages do not have breadcrumbs [20:12] sinzui: the build page includes a link to the ppa page [20:12] sinzui: the ppa page is 403 [20:13] sinzui: the link to the ppa page is in the breadcrumbs of the build page [20:13] ah. I see. [20:14] That is a whole new case. Someone should have written a custom breadcrumb adapter for build that exclude 404/403 situations [20:15] lifeless, maybe that build page should not be public since Lp think the parent object is private... [20:15] * sinzui looks for bug [20:17] sinzui: the build is public because it is a binary-copied security update [20:18] sinzui: https://launchpad.net/ubuntu/+source/postgresql-8.4 [20:18] This is another example of custom visibility rules deviating from the pattern and making other assumptions unsafe [20:18] one way to address it would be to expose the same raw data under each archive its been copied too, or something [20:19] I'd need to check code to be totally sure what its doing [20:19] I will file a bug about the symptoms [20:19] okay [20:21] bug 954411 [20:21] <_mup_> Bug #954411: build breadcrumbs of unembargoed builds in private ppa link to the ppa which is inaccessible <403> < https://launchpad.net/bugs/954411 > [20:24] sinzui: i have to drop my son to school a bit later today since he is going to camp so will miss the standup [20:25] wallyworld, thank you [20:26] sinzui: i'll be packing the lazr.restful change, landing the other branches which depend on it, and prtotyping a tri-state ui for the sharing picker [20:27] sinzui: we need a tri-state widget for All/Some/Nothing. tri-state checkbox was rejected on irc. so it's down to radio buttons or a lozenge [20:27] Yuck [20:28] sinzui: discuss with the guys and i'll ctach up with them when i get back [20:28] if some means 'do not change', tri-state could work, but if some means some, surely they should be shown? [20:29] lifeless: some does mean some [20:29] so we need something that looks good and is compact when rendered as part of a

[20:33] wallyworld, I read the scrollback this morning. wgrant is correct hat I do not like the lozenge because Lp users cannot see them. Lp 2.0 required extra support to explain to user the link is already on the page..just click the think that does not look like anything you have see on your operating system before [20:34] sinzui: interesting. i like lozenge widgets do i guess it does come down to personal preference. i guess radio buttons or drop down is what we have to choose from then [20:35] and drop downs were not preferred [20:35] so i'll see what can be done with radio buttons [20:35] wallyworld, I agree with compact is more most important and the user must see the options [20:35]