=== daker_ is now known as daker === chriadam|away is now known as chriadam [01:35] since everytime I install the SDK it breaks my machine I've decided to install it in a VM [01:35] lol [01:55] is it possible to make my application generate a prompt dialog, that is system wide? so it can block me from using the browser? [01:55] what call is used for a system wide prompt? [02:17] xtalmath: I suspect not, at least for an app with standard permissions [02:17] Not sure, though [02:19] perhaps its just a matter of going fullscreen and not forwarding keyboard input, handling it all? === kalikiana_ is now known as kalikiana [06:21] good morning === chriadam is now known as chriadam|away [08:22] popey, mzanetti: Hey guys so did you get my pings from last night :( So the accounts issue I think is definitely an apparmor issue based off of the name I would assume. [08:34] davmor2, right. I saw it, but didn't look into it any more yday [08:53] davmor2, can't reproduce this reliably [08:53] I've seen it once, but not any more [09:30] dholbach: click-review complains when a click package contains two desktop files. Yet we do allow two desktop files in a click (as I understand it from conversations between aq and ted last week). Should we lift that check, or maybe turn it into a warning? === zbenjamin_ is now known as zbenjamin [10:49] popey, sounds like it should be a warning, let me ask jdstrand [10:49] jdstrand, should or do we allow two desktop files in a click? [10:52] dholbach: thanks. === faenil_ is now known as faenil === faenil is now known as faenil_ === faenil_ is now known as faenil [11:46] Is there any way I can get my hands on the new qmake app template that supports desktop file translation? [11:46] I hear one exists alerady, but may not have been released yet. [11:47] I've tried on 15.04 with ppa installed, new project pro file does not yet include the desktop file for translation. [11:48] popey: ping, wrt issues with the date (or time?) picker in some app(s) you were seeing [11:49] yo [11:49] kalikiana: davmor2 saw it first, in reminders (notes) app on rc-proposed. http://people.canonical.com/~alan/reminders/com.ubuntu.reminders_0.5.491_multi.click is a good test, create a local note then press the clock icon in the note to make a reminder [11:49] a bird told me you saw some issues with 1.3 versus 1.2 in an app using a date or time picker [11:50] popey: okay, will give it a try [11:50] is that a special branch? [11:50] of reminders [11:53] kalikiana: no, its trunk [11:53] kalikiana: we wanted to push to the store, but this issue came up in qa [11:54] can i get GPS coordinates from CLI? [11:55] popey: just checking to get an idea where the relevant changes are. ie if this is a very recent uitk change or potentially with the code in the app (or both) [11:56] kalikiana: version in store is bzr rev 472, this one is 491 (obviously from the filename) :D - mostly bug fixes. [11:56] kalikiana: some of us can't reproduce this issue, I have only seen it on rc-proposed [11:57] so between ota-6 and now [12:08] hmmm on stable it takes ages to startup, but it looks fine, I'll switch to proposed and try there [12:21] popey: I see the border and the highlight of where the selection would be but no number, I assume that's the bug [12:21] kalikiana: correct [12:22] hmm and the last change in trunk moved everything to 1.3 [12:22] I'll investigate that a bit === faenil is now known as faenil_ [12:39] kalikiana: it happens with r490 (before we moved everything), r491 was a (failed) attempt to fix this issue (or at least be consistent) - r490 click can be found in same place [13:00] dholbach: re 2 desktop files> there is no security or policy reason to not allow it. I'm not sure the apps scope will handle it properly (for adding or deleting) [13:00] popey, ^ [13:00] popey, do you have the example click somewhere? [13:05] dpm: kyleN: I need to support plurals in C++, is ngettext the way to go? [13:06] karni, for plural translations yes [13:06] Don't we have a unified way for translating both from C++ and QML? (C++ equivalent of i18n.tr ) [13:07] in c++ we use gettext [13:07] at least in our projects [13:08] karni: We use gettext for clock app as well. Example http://bazaar.launchpad.net/~ubuntu-clock-dev/ubuntu-clock-app/trunk/view/head:/backend/modules/WorldClock/statictimezonemodel.cpp [13:08] kyleN: I know, and I think in general the story is kinda poor. Compare that to translation experience on Android, where you drop translated strings in a directory named by language code, and you're basically set. [13:08] nik90: thanks, looking [13:08] karni: make sure to set the locale properly, otherwise it won't work [13:09] nik90: right, that is what we're doing. I just happen to have a more interesting case of a required plural form :) thanks for reference though [13:09] ah ok [13:10] dholbach: i can get one [13:10] nik90: inteersting, you're using dgettext [13:11] nik90: if you bind the translation domain, I suppose plain gettext has the same behavior as dgettext (?) [13:11] karni: I believe I looked at the app scope if I vaguely remember. It is the only c++ code we were translating, so used a reference implementation. [13:11] :D [13:12] karni yes. dgettext allows you to specify the domain for that gettext call [13:16] popey: ah, thanks, I don't need to biset it then (was in a meeting so didn't get far yet, investigating more in a moment) [13:16] kalikiana: thanks for taking the time to look at it, appreciate it [13:26] dholbach: jdstrand http://people.canonical.com/~alan/multi-desktop-click.sil_0.1_all.click === JMulholland_ is now known as JMulholland [13:28] popey: dholbach asked me about that (I think) earlier. I see this: [13:28] - lint_hooks_multiple_apps [13:28] more than one desktop app specified in hooks [13:28] /tmp/multi-desktop-click.sil_0.1_all.click: FAIL [13:29] jdstrand: yeah, i asked dholbach, he asked you, then he asked for a click :) [13:29] jdstrand: I think it should be a warning, not an error [13:29] but, I think that is only there because of (old?) limitations on the device [13:29] if num_desktops != 1 then warning [13:29] sure [13:29] do you want me to file a bug to track this? [13:30] I'm happy to change that, but someone should confirm that will work correctly [13:30] ok, i can do that [13:30] I can't recall, possibly b euno's team? there might've been a store component, but I think that might have been fixed? so, the store and the scope (install and remove) I think are what were the issues [13:31] ahhh, good point, will test that too === JMulholland_ is now known as JMulholland === javiercrowsoft1 is now known as javiercrowsoft [13:36] popey: weather app spotted an niggle, open it, allow location, let it load, scroll to the last day, and tap on it. If it is the first one you click on you have no idea that the days details are off the bottom of the screen, would be nice if we could make it scroll up slightly [13:36] davmor2: will file a bug for that, thanks, agree, a niggle [13:37] popey: should we be able to still access the 10 day forecast also? [13:37] not with TWC [13:37] TWC limits us to 5 days, OWM API allows 10 days [13:37] we deliberately didn't add links to the TWC site [13:38] ah that's fine then [13:42] popey: and another, swipe up to add a location, tap on the plus, add a location, it moves back and shows you the location only now the back arrow is surrounded by black [13:42] davmor2: sounds like a toolkit bug [13:42] you on rc-proposed? [13:42] davmor2: that is a recent SDK bug [13:42] (I have seen this in another app) [13:43] * nik90 was just passing by ;) [13:43] only happens on that one page which is odd [13:43] yeah, i saw it in reminders only once [13:47] jdstrand: filed https://bugs.launchpad.net/click-reviewers-tools/+bug/1496402 to track it [13:47] Ubuntu bug 1496402 in Canonical Click Reviewers tools "Number of .desktop files check should issue warning, not error" [Undecided,New] [13:47] JamesTait: ^ see that bug, do you forsee any issues with having clicks containing multiple .desktop files from the store point of view? (click scope and dash seem to cope fine) [13:48] * JamesTait looks [13:48] ta [13:50] popey, off the top of my head, the only scanning we do of the .desktop file is in click-reviewers-tools on the server. So I don't think it'll cause a problem. [13:51] popey: ack, commented with a question that might simply be a paste of JamesTait's answer [13:53] heh [13:53] On a less technical level, it might be an idea to have such packages trigger a manual review just because, IIUC, that would allow the two apps in the package to share data. [13:54] JamesTait: yes, but that was always intended in the security design [13:54] yeah, and in this instance the developer wants two apps to share data [13:54] i think its a relatively rare use case [13:55] desktop apps should be able to share user data-- the user is driving that interaction. the sharing becomes an issues with behind the scenes stuff, like push helpers [13:56] issue* [13:56] anyhoo, ok, I'll fix this in the tools [13:56] thanks [13:56] Right. But the alternative to a manual review might be something like a warning on the upload UI saying "are you sure you want these apps to share data?" and Joe Q Nastyhacker isn't going to say "Oops, no." which puts the onus on the end user. [13:57] Which is fine, and is how we deal with other permissions, as long as it's clear to the end user what it actually means. === faenil_ is now known as faenil === faenil is now known as faenil_ [14:17] popey: Ensure icon looks like a cloud with a 12 spiked sun:- http://people.canonical.com/~alan/screenshots/device-2015-03-05-131345.png I assume that is the wrong screenshot right [14:17] correct [14:17] will fix [14:18] fixed davmor2 http://people.canonical.com/~alan/screenshots/device-2015-09-11-142604.png [14:19] popey: thanks I figured you had the image there somewhere :) [14:20] popey: question how do I know what the wind speed is for now? [14:21] davmor2: go outside and put a finger in the air? [14:22] davmor2: we discussed that but I dont think we ever got a design for it, will make a note [14:23] popey: Today gives me, haze, current of 15°c, minimum 7°c and max 15°c and that is it, if I tap on tomorrow I actually get way more info than I do for today which seems a little backwards [14:23] yeah [14:25] davmor2: there's nothing in the design for it, I'll raise it as a design bug. I know we talked about it in the past, I think we just haven't had the design, thats all. [14:25] popey: also not a fan of the scattered cloud icon, it would be too easy to mistake for rain for a new user I think [14:26] But I love the rest :) [14:26] I dont mind it [14:26] just scrolled forward and see lightning forecast for me next week [14:27] oh, not next week, friday [14:29] popey: you threw me then I'm thinking how is he scrolling to next week it only goes to sunday :) [14:31] heh [14:32] popey: I think that's it. So 1 design bug, 1 niggle, and one sdk bug https://launchpad.net/bugs/1495408 do you want me to write up bugs for those or did you say you were? [14:32] Ubuntu bug 1495408 in ubuntu-ui-toolkit (Ubuntu) "Black square in header" [Critical,In progress] [14:32] I haven't if you could, that would be super [14:33] popey: will do, do you want to let it in as is and fix it up, or fix the design issue first and then land it? [14:34] davmor2: I'll speak to the devs, if you +1 it, I'll hold it back till I speak to them and you again [14:34] okay? [14:36] popey: that's fine, I'm happy for it to go in as is, but then I would provisionally put a block on ota7 on the design bug. I would always expect more info today than any other day :) [14:36] yeah, agreed [14:36] thanks davmor2 [14:41] kalikiana, zsombi: what would be the best way to set the z-value of an Item to be the maximum value for z? [14:41] ^include float.h and set the value to FLT_MAX? [14:42] (for a qml component written in cpp) [14:42] they use FLT_MAX in qquickitem.cpp for widths and heights [14:43] loicm: ^ [14:47] t1mp: there is Number.MAX_VALUE but afaik nowhere in QML is it documented what the maximum value for z is or what behavior to expect - in practise there is no error regardless of how high the value is [14:48] try it and see what doesn't happen [14:49] kalikiana: my code is in cpp [14:50] t1mp: even then you might want to know what happens if QML in the app does use a higher z value [14:50] because whatever you expect might turn out to be false [14:50] for that matter, what is your goal? [14:50] kalikiana: I meant in CPP I don't have the JS Number [14:51] kalikiana: I want the Header to go on top of other components by default [14:51] t1mp: so in CPP you have a max value, but in QML you don't. seems like you'd want to at least verify what happens in practise [14:51] popey: https://bugs.launchpad.net/canonical-devices-system-image/+bug/1496422 https://bugs.launchpad.net/ubuntu-weather-app/+bug/1496425 [14:51] Ubuntu bug 1496422 in Ubuntu Weather App "Info for today is not full enough" [Undecided,New] [14:51] Ubuntu bug 1496425 in Ubuntu Weather App "On the main page if you scroll to the last day in the list and tap on it there is no indication there is more info" [Undecided,New] [14:52] thanks davmor2 [14:53] t1mp: undefined behavior and things like that... I'd assume the worst if you can't find a guaranteed bheavior [14:53] kalikiana: how come you don't have a max value in QML? There is a maximum value for a real [14:54] t1mp: okay, so there is a maximum for real. yet QML does not throw any errors. so... you're free to make a guess what the end result is. maybe you have 2 items one of which by chance is on top? or maybe the order of declaration matters? [14:55] if they have the same z-value, then the order of declaration matters [14:56] t1mp: and note that the text handlers need to go higher than the header, you can't be obscuring them [14:56] ohh [14:57] kalikiana: alternatively, I could keep the z-value at 0 and document that the Header should be declared last or its z-value must be set [14:58] t1mp: how come this hasn't been a problem until now? [14:58] kalikiana: AppHeader is declared last inside the MainView. App developers couldn't add their own Header component [15:00] kalikiana: maybe we can do the same when we add Header to Page, but I'm not sure yet if that would be a Component (with a Loader), or the Item instance [15:01] kalikiana: still, if you have Item { Header { }; Rectangle { } }, would you expect Header to go on top of the Rectangle? [15:03] t1mp: hmm that is an odd example. if you just declare the header like that I'd expect it to do whatever the rectangle does, depending on anchoring [15:04] but I assume even if it goes to the Page it will be assigned in a place like header: Header [15:04] and then you'd expect it to behave like a header [15:05] kalikiana: why is it odd? I expect it can be used like this http://bazaar.launchpad.net/~tpeeters/ubuntu-ui-toolkit/cppHeader/view/head:/tests/unit_x11/tst_components/tst_header.qml (ignore the UbuntuTestCase) [15:06] kalikiana: right, we'll probably do something 'special' for putting the Header in a Page, but still we will also have a Header component that can be used anywhere [15:06] kalikiana: it is an option of course that we don't deal with the z-value at all in the Header, only in the Page [15:07] so then the z-value is the same as for any Item [15:08] ah.. in my example, the flickable contents goes OVER the header, instead of UNDER it [15:08] I didn't even realize that because I set opacity of the Rectangle to 0.5 [15:10] t1mp: well, yeah, I said odd because we're discussing special properties of the header and the example to me says completely up to the developer, no magic [15:11] I would not expect anything about its z if it's just declared like that [15:11] kalikiana: I could do something like header.z: header.flickable.z [15:11] kalikiana: ^but that approach results in different behavior for z when you set a flickable or when you don't set a flickable (locked header) [15:11] kalikiana: so in my example, the flickable contents scrolls on top of the header, unless if I set header.z: 1 === JMulholland_ is now known as JMulholland [15:12] t1mp: I don't see anything in the example defining the size or position of the header [15:12] I have a feeling you're assuming something that I'm not taking for granted [15:15] kalikiana: right. for vertical position, the default value of y=0 works fine [15:15] kalikiana: but implicitWidth of the Header is parent.width [15:15] (as is the case with the ListItem as well) [15:16] t1mp: lemme re-word. if the Header is simply in an Item or maybe lower in the hierarchy. what I do expect it to do? attach itself to the root item? or the parent? [15:16] kalikiana: of course you can anchor the Header somewhere else [15:16] kalikiana: to the parent [15:16] hmmm [15:16] kalikiana: it is a subclass of QQuickItem [15:17] but I do set its implicitWidth to parent.width [15:17] okay [15:17] and implicitHeight units.gu(6) [15:17] so setting the z would only make sense if there is a flickable - otherwise you have no idea what's in there [15:17] you can have an Item in which you place a Header, even without a Flickable [15:18] hmm.. but if there is no Flickable, probably you will anchor the rest of the contents to the bottom of the Header [15:18] otherwise stuff gets occluded (assuming header.z >0) [15:22] t1mp: yes, so you either never set z or you have a special use case that can't be automatic [15:22] setZ(FLT_MAX) can be the default [15:22] the only case where it makes sense to have a default is if flickable is set [15:22] well [15:23] that would make the header on top of potential unrelated things, though [15:23] *put [15:23] say I want my app to have a kittycat border in which there is a header, enclosed by the border like every nother part of the UI [15:24] and maybe the border slightly blends over with transluceny [15:24] and maybe there is a panel coming in from the left side [15:24] right [15:24] I'm making stuff up, but it seems like there's a number of things you would expect to have a higher z than the header [15:26] kalikiana: the z-order only matters for siblings inside the same Item [15:27] so you can still do Item { id: root; Item { Header{}; Rectangle{} }; KittyCatBorder{} } [15:27] on the other hand, [15:27] if I don't touch z, and apps have contents going on top of the header, the appdevs easily detect and fix that issue [15:30] t1mp: the example still has to assume I'm setting something like anchors.top: header.bottom on the Rectangle, otherwise it's just partially hidden [15:31] in that case it's obvious anyway regardless of z being special [15:32] ok, then: Item { id: root; Item { Header{}; Flickable{} }; KittyCatBorder{} } [15:32] that doesn't need additional anchors to the header (alhthough Flickable should fill its parent) [15:32] -h [15:33] yeah, that one works out of the box [15:33] (almost) [15:34] and if you wanted to "lock" the header you might still set the flickable to anchor to the header's bottom [15:39] right [15:43] rschroll, heya. On the "you don't get icons" front, do you get an icon if you try to addtodash https://jakearchibald.github.io/svgomg/ ? [16:10] aquarius: That works. Let me try to find one that doesn't. [16:14] rschroll, cool. I haven't done any comprehensive tests on lots of sites yet, so I'm expecting there will be bugs, certainly :) [16:24] aquarious: en.wikipedia.org has a favicon set, but this isn't picked up [16:24] (Should it be?) [16:25] Also an apple-touch-icon [16:31] aquarius: Twitter has a favicon, a mask-icon, and a , none of which are found [16:31] rschroll, cool -- thank you! [16:36] aqarius: Washingtonpost.com just has a favicon at /favicon.ico (old school!), but also have an open graph image, which we may or may not consider using [16:37] The wonderful thing about standards is, there are so many to choose from. [16:47] fixing bugs [16:49] For the favicon selector, use rel~=icon. (In case you hadn't figured that one out yet_ [16:59] aquarius: This may be useful in hunting down the various places icons might be hiding: https://css-tricks.com/favicon-quiz/ [16:59] https://mathiasbynens.be/notes/touch-icons too [17:25] rschroll, hrm. In two minds what to do about tile images [17:26] rschroll, the MS tile images provide an image (almost always white on transparent background) and a tile colour (the background). But we need just an image... so we could use just the tile image, or we could create an image by compositing but that's work we'd have to do, or we can ignore it. [17:26] I'll go with just using the tile image for now, but it's not ideal. [19:49] Qt Creator has forgotten the run configuration for my app. (It's offering only "custom executable".) How do I restore the old value, so it launches the app on my device. [19:50] I've deleted the associated .user file, to no avail === faenil_ is now known as faenil [20:03] rschroll, is that for addtodash? I think it gets confused because we have multiple desktop files :( [20:32] aquarius: Yeah. I disabled the other desktop files, to work on other things, and Qt Creator is still confused. [20:32] bah [20:33] I have a vague memory of this happening before, and needing to delete some obscure Qt Creator file to get it to reset. [20:33] Because storing this info in the .user file would make way too much sense. [21:01] popey, ping? [21:01] hi kivi [21:01] popey, is the contest still on? [21:03] kivi: https://developer.ubuntu.com/en/blog/2015/07/23/announcing-the-ubucontest-2015/ [21:03] sept 18th apparently :) [21:03] ty [21:03] ah nice [21:08] Wasn't sturmflut involved in the contest? Is he OK? Haven't seen him around here for ages === faenil is now known as faenil_ [21:37] mcphail, he is very busy at work i think [21:45] ogra_: aah. I'm missing his insights [21:53] he is on holiday i think [21:54] oh, maybe not [21:54] yeah, just busy [22:31] I think he was on holiday, but supposedly back from that, so maybe busy with other things?