=== udsbotu changed the topic of #ubuntu-uds-appdev-1 to: Currently no events are active in this room - http://summit.ubuntu.com/uds-1308/appdev-1/ - http://irclogs.ubuntu.com/2013/08/28/%23ubuntu-uds-appdev-1.html === Kay is now known as Guest57912 === Guest57912 is now known as KayC === fecub is now known as [fecub] === udsbotu changed the topic of #ubuntu-uds-appdev-1 to: Track: App Development | AppDev Roundtable - Wed | Url: http://summit.ubuntu.com/uds-1308/meeting/21926/appdev-roundtable-wed/ [13:57] dpm: and anybody else who wants to join the roundtable: https://plus.google.com/hangouts/_/158f0544c1a2389ef0a0fd77846d67028b645e4b?authuser=1&hl=en [13:58] thanks mhall119, setting up another device to try to join in [14:01] popey: want to join the roundtable hangout? [14:03] your on [14:04] yes, able to see [14:04] only michael? [14:04] https://plus.google.com/hangouts/_/158f0544c1a2389ef0a0fd77846d67028b645e4b?authuser=1&hl=en [14:10] dpm: popey: either of you joining? [14:10] mhall119, trying now [14:11] Followup on file access from yesterday [14:11] I asked on ubuntu-security, and was pointed to this page: https://wiki.ubuntu.com/SecurityTeam/Specifications/ApplicationConfinement/Manifest [14:11] This looks like what mhall was suggesting [14:11] not what rickspencer was talking about [14:12] So, what's the deal here? [14:13] Is it currently possible to download files to disk? how will this be implemented? [14:14] This is still in development, though? [14:14] For the purposes of the app showdown, do I just worry about the manifest? [14:15] ok, thanks! [14:15] mhall119: on the weather app design call [14:15] sounds cool [14:15] kenvandine [14:15] haven't seen it yet [14:15] ill look for the session [14:16] can't seem to find it [14:17] http://summit.ubuntu.com/uds-1308/meeting/21909/foundations-s-touch-download-service/ [14:17] Maybe some sessions should only be half an hour long, this way you can still have you and your colleuge on the same session and there will be more questions :D [14:18] I have a bunch of sorta specific questions; don't know if they're appropriate here [14:18] rschroll: may as well ask [14:18] oh, yeah, it's not in appdev [14:18] bookmarked it though [14:19] How do you use Arguments with qmlscene? I get both args for the app and qmlscene mixed together. [14:19] http://askubuntu.com/questions/336083/how-to-use-arguments-in-qml-without-getting-qmlscene-arguments [14:20] rschroll: can't you just ignore them? [14:21] yes, but I have to decide whether an input is for me or for qmlscene [14:21] It works, but I thought I was missing a better way [14:24] Another specific question: Is it possible to style the options in a ValueSelector? I can copy the code for it and apply the styling I want, but that seems ugly. [14:25] Is there a way to modify the styling of an existing selector, or to subclass the ValueSelector? [14:25] rschroll: might be better to ask in #ubuntu-touch where the sdk developers reside [14:25] ok [14:37] Will there be an API that enables you to do an action every x minutes while your app isn't running? Like checking if you have a mail or a message, or... [14:37] I was wondering if we have any contacts related API [14:39] ok, thanks [14:40] I mean dialog box for choosing contact [14:41] ok [14:42] the nexus 4 dropped in price today, will this have any effect on the prizes of the showdown? ^^ [14:42] mhall119: your fan that is reflected in the picture frame is hypnotizing me :) [14:42] better? [14:43] not sure lag ... [14:44] yes much better :) [14:45] I don't want my shitty gitHub client on everybody's phone! [14:46] hahaha [14:46] lol [14:50] It's available immediately actually [14:50] immediately? It used to take youtube a while to process [14:51] I look forward to going back and re-living mhall119's grandma joke [14:51] :P [14:51] mhall119: Thanks for the help! [14:59] will some people of the design team be joining us now? === udsbotu changed the topic of #ubuntu-uds-appdev-1 to: Track: App Development | Visual design guide for Apps | Url: http://summit.ubuntu.com/uds-1308/meeting/21924/appdev-1308-visual-design-guide/ [15:05] https://plus.google.com/hangouts/_/a420dc3ec7461b0ef0d414fd55625834325d12ad?authuser=1&hl=en [15:05] up [15:05] we can see you [15:06] have you guys seen the sidebar component? What are your thoughts on that? [15:06] Yep... [15:06] the sidebar component is visible in the file manager [15:07] GuidoPallemans: only in desktop mode [15:08] indeed, but I always include a button to view the contents of the sidebar, is this ok from a design point of view? [15:17] I missed the first part of this talk :( [15:18] aquarius, they are recorded, so you can watch later on youtube ;) [15:18] do we have iconography ? how icons should look like for toolbar and else where [15:19] There was suggestion from katie at the end about how the gradient-background approach is really intended for the ritual apps. Is there guidance on what non-ritual apps (i.e., most of them) should be doing? I saw something about "the medium theme" -- is that avaialble? [15:19] also, I agree with kunal_ -- having a standard set of icons for common toolbar actions -- refresh, about, restart, new game, that sort of thing -- would be really nice [15:20] QUESTION: I would love to see apps use realistic components (e.g. a notepad app look like a real notepad)...Apple does this well - does the team have any guidance on this? [15:20] more specifically, can we get some guidelines for how to use realistic looking visual elements [15:21] agreed with mhall119 -- there are lots of examples of how the design team have chosen the visual design for specific apps, but not *why* those were chosen. :) [15:21] jonobacon, just say no to skeuomorphism ;) [15:21] on the phone, the sidebar is hidden, but accessible through a menu button (at least in my app) [15:22] QUESTION: are they any guide lines for when a user should use a Coverflow or a listview or a gridView or a Path/SVGView ect [15:23] example taking the listview in music app and having in coverflow ect [15:24] dynamically changing. [15:24] ahem. "Do, well, whatever you want!" is not guidance, katie ;-) [15:25] not much engineers use pinterest though [15:25] GuidoPallemans, should do more, though. Pinterest is an *excellent* source of design inspiration. [15:26] mhall119: I think what we could do with help from the design team on is a set of guidelines for how people can add visual flourish without dumping a gradient in there [15:26] can we ask the team to commit to making some guidelines for this [15:26] agreed with jonobacon. [15:26] otherwise I worry we will end up with a tonne of gradient apps :-) [15:27] What happened with the "Locally Integrated Menu - LID" idea for desktop? LID is dead for convergence apps? How the desktop menu will correspond to the phone/tablet menu in convergence app? [15:27] we will. Because currently it's easy to have (a) a grey app or (b) a gradient, and that's it. We are not at all free to lay out the app how we want -- the header has to be at the top, with a stock font size, etc -- so colour is one of the big ways to differentiate. [15:29] QUESTION I don't know if this is the right place to ask, but I saw the locker-type lock (where you highlight numbers by sliding around a circle), and thought it was really nice, will this be implemented? [15:29] mhall119: I think the challenge app devs face is that they need a set of more prescriptive guidelines from the design team - general approaches design are hard for people who don't have design experience - can we ask the design team to commit to expanding the set of app guidelines that cover the topics outlines in this session? [15:30] An exemple of expected guideline : "how to handle 'settings' panel ?" (a) as a tab (--> non modal) (b) as a special page (--> called from the toolbar, modal) [15:30] Perhaps something that would be useful is the design team's interpretation of an app which is *not* a ritual app. At the moment all the visual design we've seen is for the ritual apps, so that's all that everyone has to copy. [15:30] aquarius: +1 [15:30] * WebbyIT think that is the most interesting session until now [15:31] much as it pains me to agree with you :-) [15:31] LOL [15:31] * aquarius grins [15:31] ahahaha [15:33] The design blog post about visual exploration is http://design.canonical.com/2013/07/shorts-visual-exploration/ [15:34] Question: are there any icons / anything that you guys(designers) that is in QML ? example using your examples as real world examples. some developers have trouble cuttin up pictures ect. [15:34] ohh that's a shame [15:34] so like when designing a page keep icons and other things in there own file [15:34] make a branch ect [15:35] maybe there should be a weekly hangout where an app, where the design guys talk about an app, how it can be done better, and/or a developer can ask for advice and defend his design solutions etc. [15:35] mhall119: what is the hangout URL? [15:35] would like to join [15:36] whoops brain fart [15:36] https://plus.google.com/hangouts/_/a420dc3ec7461b0ef0d414fd55625834325d12ad?authuser=1&hl=en [15:36] QUESTION: (also asked by kunal) will there be a standard set of icons for common toolbar actions -- refresh, about, restart, new game, that sort of thing, or should everyone be just choosing their own? [15:38] Any news on the "back" button ? days ago, Design Team said they were working on a new implementation requiring only one tap (instead of 2 now) [15:38] justForQuestions: can you clarify your question about icons? I don't understand [15:40] mhall119: Like the examples that are on the design page and for core apps. If the icons and what not where in seperate files then the developer's life becomes easier. and also for developers to play with the design that canonical has worked so hard on. Would help with training I think [15:49] a weekly clinic where the design guidelines get applied to an app will be great, even greater would be if developers can show their own designs and ask questions around that [15:53] ooh, cool, there's a toolbar action icon set? making those available would be excellent! [15:54] /usr/share/icons/ubuntu-mobile [15:54] is that it? [15:55] GuidoPallemans: I think it is... Or at least I hope, since it's what I'm using :D [15:56] yeah me too [15:56] also on desktop [15:57] mhall119: katie I was referring to these: http://imgur.com/qGZlaeC [15:58] thanks, katie et al. [16:00] katie, one other question: you have a high-res image of a "generic" mobile phone, I think (the one used in all the pictures at http://design.canonical.com/2013/07/shorts-visual-exploration/). Is that asset available to everyone else so we can make nice looking mockups too? :-) /cc mhall119 === udsbotu changed the topic of #ubuntu-uds-appdev-1 to: Track: App Development | Core Apps Review | Url: http://summit.ubuntu.com/uds-1308/meeting/21953/appdev-1308-coreapps-review/ [16:00] aquarius: I also asked victorp if he had any of the Edge [16:00] no reply yet though [16:01] mhall119, yeah -- having some would be nice for our mockup-y thing :) [16:01] agreed === james is now known as Guest81785 [16:04] * popey waves to mhall119 [16:04] popey: getting it setup [16:04] https://plus.google.com/hangouts/_/9fe4244b49578122ad1761a9b2734a41a3f03ea4?authuser=1&hl=en for anybody who wants to join the fishbowl [16:08] hehe [16:09] popey: https://plus.google.com/hangouts/_/731ba92244eff4fa4e63dea732e20d3b55ae333b?authuser=1&hl=en [16:09] I just like to watch :) but I don't see the video yet [16:09] I'll be trying to listen to a couple sessions [16:09] dpm: dude, you gotta tell me when you move sessions, I started the wrong one [16:10] ok the video is starting [16:10] ah no it says starting soon [16:10] * WebbyIT is interesting to listening [16:10] http://pad.ubuntu.com/uds-1308-appdev-1308-coreapps-review [16:11] mhall119, ah, sorry, I generally do, we had to change it so that popey could participate. Sorry Mike, I was doing too many things at the same time [16:11] timp: should be live now [16:11] yes, I can see you [16:11] mhall119: now you're live [16:11] mhall119: ahh, you're all alone? :( [16:11] mhall119: oh you are not alone, but I don't see the thumbnails at the bottom [16:12] yeah, the broadcast only shows one person at a time [16:12] popey: way to go. You look like a shampoo commercial with the wind in your hair :) [16:14] I don't see an etherpad in firefox [16:15] timp: try opening http://pad.ubuntu.com in a separate tab and logging in, then go back to the summit page [16:16] mhall119: that works. thanks. [16:32] that weather app looks great [16:32] mojo706: it really does, doesn't it? [16:34] yes [16:38] can you show that again? it wasnt on the big screen [16:38] could you show that again [16:38] yeah we couldn't see it properly [16:39] it isnt on the big screen, you should say something [16:39] very nice [16:39] its ok [16:40] will be there any time soon Ubuntu Touch UI Kit for Gimp, Inkscape, Photoshop or Illustrator. and prototyping tool like -http://jaunesarmiento.me/fries/ [16:40] michael do you have the source? can you demo it? [16:41] GuidoPallemans: https://launchpad.net/ubuntu-rssreader-app [16:44] markdown for doc viewer? and zsh? [16:45] that behaviour should be added to the keyboard [16:46] GuidoPallemans: what behavior? [16:46] the arrow behaviour in the console app === alex_abreu is now known as alex-abreu [16:53] ping mhall119 [16:53] bzoltan: pong [16:53] mhall119: what time the SDK porting session starts? [16:54] bzoltan: later today, it was moved [16:54] mhall119: is this schedule up to date? http://summit.ubuntu.com/uds-1308/2013-08-28/display? [16:55] bzoltan: should be, refresh to make sure you have the latest [16:55] mhall119: OK, so it starts in an hour [16:56] for some reason I thought it would start now [16:56] nice one dpm mhall119 [16:56] mhall119: I try to make it, but 9pm here might be tricky :) I just put the kids to sleep [16:56] Lunchtime! [16:56] :D [16:56] popey: LOL :) 8pm, late dinner time [16:57] thanks popey, mhall119 [16:57] or early breakfast [16:57] shouldn't every second vUDS be scheduled according to a European timezone? :) === magicluriv is now known as luriv [16:58] timp, yeah, it kind of slipped off, but as someone living in Europe, I'll make sure to raise it next time we plan UDS :) === udsbotu changed the topic of #ubuntu-uds-appdev-1 to: Currently no events are active in this room - http://summit.ubuntu.com/uds-1308/appdev-1/ - http://irclogs.ubuntu.com/2013/08/28/%23ubuntu-uds-appdev-1.html [17:04] timp, dpm: regardless how much I love my own TZ, I think it is the audience and target group what counts ... if the contributors are mostly in the US timezones then we should adopt to them. [17:13] There is no calendar since Evolution is not installed [17:21] hi [17:21] hello [17:21] Mickael [17:22] Hello any 1 there ? [17:23] hi VJ [17:24] Hey webbyIT , i am newbe in Linux world , really exited so can you like give me some advise for startups ? [17:24] VJ, for support go to #ubuntu === udsbotu changed the topic of #ubuntu-uds-appdev-1 to: Track: App Development | Porting the Ubuntu SDK to Windows and Mac | Url: http://summit.ubuntu.com/uds-1308/meeting/21917/appdev-1308-sdk-porting/ [18:01] hi === zyga_ is now known as zyga [18:01] hello [18:02] mhall119, o/ [18:05] https://plus.google.com/hangouts/_/fd996734764b0bc880a1eb44f2066f60ccb38848?authuser=1&hl=en please join the fishbowl if you're interested in this session [18:06] it's streaming [18:06] it's up [18:06] you're live [18:07] ok [18:11] QUESTION: what about the APIs which will be on Ubuntu, such as online-accounts? would they be ported too? [18:14] Depends if the libs are on machine or allowed on the machine (qt creator stuff) [18:15] any linked lib that is not on machine is going to make qtcreator not running [18:15] it's not that simple [18:16] Hi - I have USB bridging working OK (with a couple of minor convenience bugs) on OSX host with a virtualbox guest, and have been working on touch apps that way. [18:16] should build one that uses installbuilder just like qt does and that could all be one with links and what not as long as they run on doz [18:17] that is how sailfish does it also [18:17] I just heard mhall ask if qtcreator and adb works over USB from a guest - this works OK for me from virtualbox [18:17] thanks mmcc [18:18] yes, flashing is a bit of a dance [18:18] another thing is some of the scripts that are in the sdk are wrote in bash and not sure about windows and bash (have used in years ) [18:18] hove not * [18:19] i would stay away from xcode or visual studio [18:19] avoid avoid xcode [18:19] you'd basically have to embed qtcreator itself, to be useful [18:19] (this from a longtime xcode user) [18:19] otherwise you won't have any UI design tool really [18:19] xcode would be very hostile to that kind of thing [18:20] at which point you might as well just use emacs or vim [18:20] Things like anything like dbus and metadata stuff would also be hard to port [18:21] dbus is ported. but making everything work right natively is hard [18:22] what are the pro files for the qtcreator [18:22] what is package called ? [18:22] mhall119: i'd just run everything in a vm [18:22] LinkedLibMess: the .pro files are qmake description files [18:23] I agree, if you have to use a VM to run, might as well use it to develop, since this stuff is supposed to have a quick edit/run cycle [18:23] i would just qemu everything [18:23] a touch image with the sdk included. [18:25] having a full desktop image to run the app is really helpful for debugging when you want to tail debug logs while you're testing [18:25] and it's probably eaiser to just install ubuntu in virtualbox or vmware and just develop on ubuntu in a vm [18:26] (like some developers i know do)I [18:26] I do that. It works* now. (* except for some 3d accel issues when running QML) [18:27] i think they would be fine [18:27] you could, if you use vmware fusion or something [18:28] and cutting/pasting between vm and host is easy as well [18:28] virtualbox has a rootless thing too [18:28] and host<->vm disk access is pretty easy [18:28] But what about the plugins that Ubuntu is using one can launch a vm from qtcreator real easy but having the plugins and what not that are only used by qtcreator be crossplatform ie the linked libs [18:28] anyone writing android apps is already doing it all in a vm anyway ;) [18:28] coding in qtcreator inside the VM and using a mac email client is not bad. I do that [18:28] only for windows and mac ^^ tgat is with install builder [18:29] coding on mac and using vm to test apps is the way to go [18:30] premade VM images is most convenient [18:30] correct but other company's do it so why not canonical ? [18:30] other companies do what? [18:31] as long as the libs are on windows qtcreator should run [18:31] dobey: qt is one themselfs [18:32] LinkedLibMess: the question will be how much work it is to get all of the libs and eternal executables on windows and osx [18:32] it all depends on what is used to make the qtcreator plugins for Ubuntu if them libs are there on doz or osx then it will work [18:32] mhall119: none at all make installbuilder [18:32] installbuilder? [18:32] the qml components are an issue as well as the "plug-ins" for qtcreator itself [18:33] installbuilder packages for different platforms it is what sailfish and also qt themselfs [18:33] mhall119: ^^ [18:33] LinkedLibMess: it's not jus packaging that's a concern [18:33] I was just saying, with VM's, we'd probably have to offer images for x86 and amd64 as well? [18:33] but it would all be slow [18:34] if everything is in a vm things will be slow [18:34] not if you have a core i7 3770 [18:34] dpm: I think we can be okay with just x86 [18:34] and 16GB RAM [18:34] dobey, hehehe [18:34] all my VMs are fast :) [18:34] even less beefy systems are OK as long as the 3d is working [18:35] there is no reason why one can not use doz with qtcreator and then launch vm from qtcreator on the run button that pushs (app) to vmmachine [18:35] but it again is all about if the libs can be compiled on doz and osx [18:35] there are plenty of reasons you can't do that [18:36] extending eclipse seems better [18:36] doesn't qtcreator actually load QML components when in design mode? which means the components need to load their compiled plugins? [18:36] dobey: like what ? [18:36] qtcreator needs the qml components to exist as well [18:36] mmcc: yes [18:36] not if they are on the VM dobey [18:36] mmcc: also, when doing autocomplete [18:36] and have an emulator [18:36] dobey: autocomplete might be handled by that txt .qmltypes file [18:37] though there would be some issues with highlighting and what not but not that bad [18:37] mmcc: it also does error checking, so that if it can't load it, then it complains about it [18:37] mmcc: which would be very annoying [18:37] dobey: right [18:37] someone give me windows licence and I will do it [18:37] lol [18:38] folks, I have to run early, but I will volunteer to test out VM images on OS X. I'm mike.mccracken@canonical. [18:38] i run qtcreator in a vm without video accel and it was fast [18:38] qmlscene runs slower natively on my laptop than it did in my vm without accel :( [18:38] I have also run qtcreator for years I know for sure that it is slow with things like 3d and gl and stuff like that [18:38] dobey: it's surely due to some combination of my hardware and vbox running on OS X host :) [18:39] ('it' being the problems I mentioned with 3d) [18:39] mmcc: your issues, yes :) [18:39] also state the specs that need to run the vms [18:40] yeah, we ship python on win/mac [18:40] If sailfish can do Ubuntu can do it [18:40] sailfish != ubuntu [18:40] apple != orange [18:40] look for yourselfs they have linux mac and windows [18:40] the point is it is crossplatform [18:41] yes, qt itself is cross platform [18:41] yes it uses vm on all and uses installbuilder to build the sdk [18:42] it is shipped with the sdk [18:42] it is the same for windows and osx [18:42] it can DL other images like armel7 and what not also [18:43] via script [18:43] ugh. google really needs to run that gigabit fiber to their data centers, rather than kansas city [18:43] install builder .. http://installbuilder.bitrock.com/ [18:44] packaging isn't an issue [18:44] that's to build the installer, not the sdk [18:44] mhall119: they package the libs with it if needed [18:44] for each one [18:44] we use that to build the installer for ubuntu one too [18:46] dobey: it was nice chatting with you :) [18:46] just wish I knew more about the linked libs that are linking making things that are not crossplatform [18:46] I'm afraid I didn't follow this session [18:47] SDK folks in the hangout? [18:47] like LIBS+= -L /usr/lib/something -lsomething [18:47] lool: mhall119 and zyga and dpm are yes [18:47] we're lesser mortals [18:47] not part of the SDK team [18:47] https://plus.google.com/hangouts/_/fd996734764b0bc880a1eb44f2066f60ccb38848?authuser=1&hl=en [18:47] https://plus.google.com/hangouts/_/fd996734764b0bc880a1eb44f2066f60ccb38848?authuser=1&hl=en [18:49] i can't join, because google+ seems to hate me today. i tried to join one earlier and got no video or audio once i joined :-/ [18:49] it's ok. people have to run osx to develop on ios :) [18:50] QUESTION: what is package name for the SDK not ubuntu-ui-toolkit but qtcreator build ? [18:50] dobey, no its still wrong [18:50] LinkedLibMess: ubuntu-sdk [18:50] LinkedLibMess: ubuntu-sdk is the metapackage [18:50] that is package name for that metapackage ? [18:51] thanks [18:54] qtcreator-plugin-ubuntuu << that is what I am looking for that metapackage I guess :) [18:55] or offline installer .... === udsbotu changed the topic of #ubuntu-uds-appdev-1 to: Track: App Development | App Developer Mode for image-based Ubuntu Touch | Url: http://summit.ubuntu.com/uds-1308/meeting/21914/appdev-1308-app-developer-mode/ [19:00] the bot lies [19:01] the bot is right, it's us that are late [19:02] thanks, interesting session! [19:02] bye [19:05] who is opening the next hangout? [19:05] mhall119 is the track lead, right? [19:05] yes [19:06] https://plus.google.com/hangouts/_/775b5f9d919a2624105bafd80c1846c84238b409?authuser=1&hl=en for joining the fish bowl, the more the merrier [19:06] thanks! [19:06] I'll just hang out here, my internet connection is not good enough today it seems :-( [19:07] did the hangout start already? [19:07] no [19:07] dholbach: it's running but not live yet [19:08] go go go! :) [19:08] it's on air now [19:08] supposedly [19:08] o/ [19:08] https://plus.google.com/hangouts/_/775b5f9d919a2624105bafd80c1846c84238b409 [19:08] * tedg got video [19:09] i can see you [19:09] you're broadcating [19:09] You are broadcasting. [19:09] you seem to do [19:09] lool, can I join the hangout so I can ask my questions easier? [19:09] rickspencer3: https://plus.google.com/hangouts/_/775b5f9d919a2624105bafd80c1846c84238b409?authuser=1&hl=en [19:14] who's speaking second from the left? or can you turn on your lower third? :) [19:15] That was David Jordan, I believe [19:15] I don't mean to be dismissive about the complex-dependencies case, but it's not one we can handle yet and I think it's important to get the things we can currently handle working well [19:16] dmj_nova1: (you?) [19:16] yes [19:16] sergiusens, you look like you're in the witness protection programme [19:16] my G+ plugin died [19:16] beuno: lol [19:16] beuno: light? [19:17] beuno, HAHA [19:17] beuno: heheh [19:17] sergiusens, yes :) [19:18] cjwatson: I'm fine with taking things one step at a time [19:18] someone from SDK team joining? [19:19] bzoltan: around? [19:19] * ogra_ was surprised he isnt [19:19] since he owns the spec [19:20] dmj_nova1: the click spec does have a slot for supporting additional "frameworks" which is probably going to form part of a later solution to this [19:20] we deliberately excluded having complex dependencies though as then it's going to turn into .deb real quick [19:20] I can join [19:20] what's the link? [19:20] beuno: https://plus.google.com/hangouts/_/775b5f9d919a2624105bafd80c1846c84238b409?authuser=1&hl=en [19:20] beuno: https://plus.google.com/hangouts/_/775b5f9d919a2624105bafd80c1846c84238b409?authuser=1&hl=en [19:20] as a dev advocating for more complex apps (which will be necessary for having truly killer apps) and someone involved with a more complex application, it will be helpful to know where things are going, that it will be possible to extend the platform [19:22] if it's complexity contained within a single application then bundling is viable; it's only a difficulty if you want to do lots of library sharing [19:22] and bundling complex dependencies is certainly not ideal [19:23] for instance with dmedia, bundling might work if we knew the user would only ever use a single dmedia-based application, but 15 dmedia instances running on the same system would be a nightmare [19:23] we've also talked about having a way to share code among click packages from a single origin [19:24] which would require some extensions to apparmor profiles [19:24] I'm reasonably confident we can do it :) [19:27] i'm not sure if this is 100% on-topic for this discussion but if an app requires a new online-accounts plugin how would a developer install it? would it have to go into the Ubuntu image as a .deb or will there be a click mechanism for it? [19:27] ssweeny: good question [19:27] cjwatson: We should definitely have a conversation about how to enable app ecosystems to extend the platform [19:28] Yep. But after 13.10 :-) [19:28] * cjwatson <- a bit slammed [19:29] ssweeny: for 13.10, there are a couple of answers [19:29] It would be cool to have SSH on the phone :-) [19:29] ssweeny: if it's something you can bundle in your app, do so [19:29] cjwatson: I'm not sure it is [19:29] one moment, typing :) [19:29] ssweeny: if it's an add-on to a system interface, and *if it can be apparmor-confined*, then you can have the system interface in question define a click hook that click packages can attach to [19:30] rickspencer3, people wanting to really deeply hack into the system can use the flipped image... they are always there since they are an interim product we need for system images [19:30] cjwatson: ssweeny: that's probably something we should have a separate discussion about [19:30] cjwatson, right now those plugins are implemented as xml files dropped into a specific location on the filesystem [19:30] mhall119: (sorry I have a hard time breaking the notes into questions/ideas/concerns) [19:30] ogra_, or just go enter Read/Write mode, right? [19:31] lool: no worries, I can clean it up later [19:31] rickspencer3, thats still fairly restricted [19:31] ssweeny: if it can't be apparmor-confined, then you can't do it for 13.10 [19:31] ok [19:31] thanks ogra_ [19:31] cjwatson, ok [19:31] ssweeny: if they can't execute arbitrary code (pure data), then you don't need confinement, just a click hook [19:32] cjwatson, all they do is fill in values that online-accounts uses (API endpoints, keys, etc) [19:32] ssweeny: OK, then that can be done with a click hook. Contact me out of band if doc/hooks.rst in the click source package isn't sufficient [19:32] cjwatson, ok, will do. thanks! [19:32] what happened when we are supposed to talk about technical details at UDS? ;) [19:32] :-) [19:32] (I'm assuming you're a developer of the online-accounts package that would define this hook) [19:33] cjwatson: The pure containment is a good thing for reducing the amount of verification time (which is the reason for this stuff I believe), but there definitely should be a mechanism for distributing apps outside basic scope [19:34] dmj_nova1: we do have some slots in the code to build out this kind of thing in future [19:34] Do we need a setting in system settings to enable this stuff? [19:35] tedg: I would only say to enable ssh [19:35] I think what we reached was a setting to enable ssh until the next reboot [19:35] tedg: but not the _system developer mode_ [19:35] (backend: "service ssh start") [19:35] To enable adb as well, no? [19:35] Or whatever [19:35] Ah, okay. yes [19:36] also, do you believe that end users will be able to access this read-write mode/switch from image-based updates on their purchased devices from an OEM/carrier? [19:37] rickspencer3: ^- is that something you can speak to? [19:38] cjwatson, yeah, but in another discussion right now [19:41] Won't it be the same as a PIN if you set it? [19:41] If it's swipe, it should be null.... [19:41] If you set a PIN, it's the PIN. [19:41] novacut has some useful peer techniques [19:42] Perhaps only enable SSH if you have a PIN set? [19:42] tedg, with swipe you indeed type in "swipe" [19:42] ogra_, Hah, and then I get a phone from a German OEM and I won't be able to log in without the character map ;-) [19:43] :D [19:44] tedg: If you have an open SSH server then you can brute-force the entire PIN space probably without the user noticing [19:44] So I cannot recommend using the PIN [19:45] cjwatson, Hmm, but I'd imagine that'll be the user account's password, no? [19:45] I like this QR code idea [19:45] tedg: The SSH server doesn't necessarily have to have passwords enabled, especially not if we use this clever QR code idea suggested on the hangout [19:45] cjwatson, ++ [19:45] Sorry, just got the QR code thing :-) [19:45] that's probably the first time I've heard someone smart say something good about QR codes [19:45] Okay, so disable passwords, then scan. [19:46] :-) [19:46] www.youtube.com/watch?v=Awz_sGj0Zhw [19:47] lool, I think "no password" isn't good [19:47] I think if you're using a PIN to login, it should be the PIN. [19:47] tedg, "locked password" [19:47] not "no password" [19:47] tedg: disabled password [19:47] "!" as password [19:47] A PIN should be a different PAM authentication mechanism [19:48] It should not be hooked up such that things that do remote auth might accidentally use it [19:48] Why wouldn't you use the PIN with PAM? [19:48] But only for auth when logging in on screen [19:48] btw, the pin in the video is a one-time pin, just to get the key across securely [19:48] tedg: You would, just not via PAM password auth [19:48] Keep it separate because it's not the same thing :) [19:49] I guess I kinda do see it as the same thing... "this is how I login to my device" -- PIN, password or swipe. [19:49] If I plug in my phone and I'm using it as a desktop... how do I get through the greeter? [19:49] It should be kept separate at the PAM layer so that different session-creating services can use different policies [19:49] That's all I'm saying [19:50] A PIN does not make sense to expose for network login methods where you can attack the entire PIN space at your leisure [19:50] But it's fine for local login [19:50] Or, to put it another way, why conflate PINs with passwords when we don't have to, given that they're conceptually different [19:51] And I've just exhibited a case where distinguishing them is useful :) [19:51] cjwatson, I see what you're saying, but I guess I worry that we're separating things that users don't see as different. [19:51] yeah, let's not mix unlock pin with complete device access [19:51] lool, do you want me to join the hangout to try to answer dmj_nova1's question? [19:51] tedg: This wouldn't generally need to be exposed to users, so I don't see that as a concern, really [19:52] rickspencer3: if you like, we kind of switched topic I'm afraid [19:52] Not many users will need remote access to their phone, but let's not open security holes for those who do :) [19:52] dmj_nova1: including link to your peering setup in pad, thanks [19:52] cjwatson, I'm just guessing that those who do probably aren't happy with a PIN anywhere ;-) [19:52] lool: feel free to ask me or jderose about how it works [19:53] thanks [19:53] or look at the peering code in dmedia [19:53] tedg: Some won't be, but I think some will be fine - attacking an entire PIN space on a phone screen is maybe possible but would take a long time [19:53] And be kind of obvious [19:53] lool: you can find us in #novacut [19:53] cjwatson, Don't we delay on failed attempts with network services? [19:53] lool, I don't much care to join the hangout, I can just answer here [19:53] dmj_nova1, did you get an answer to your question? [19:54] tedg: Yeah, but if you leave your phone connected overnight (etc.) that's still plenty of time [19:54] tedg: Somebody actually getting hold of your phone is quite a different matter and a lot more obvious ... [19:55] Interesting. Perhaps we need the delay to increase exponentially ;-) Anyway, topic for beer. [19:56] .wc [19:56] sorry, that was window close in irssi ;d [19:56] mhall119, Assign mpt to update the design first [19:56] tedg: design of what? [19:56] mhall119, System settings [19:57] tedg: done [19:59] thanks a bunch! [19:59] thanks === udsbotu changed the topic of #ubuntu-uds-appdev-1 to: Currently no events are active in this room - http://summit.ubuntu.com/uds-1308/appdev-1/ - http://irclogs.ubuntu.com/2013/08/28/%23ubuntu-uds-appdev-1.html === udsbotu changed the topic of #ubuntu-uds-appdev-1 to: Currently no events are active in this room - http://summit.ubuntu.com/uds-1308/appdev-1/ - http://irclogs.ubuntu.com/2013/08/29/%23ubuntu-uds-appdev-1.html