[00:00] <newsages> Zbar?
[00:30] <nik90> zsombi, popey: It seems we got another crash in the clock app, https://bugs.launchpad.net/ubuntu-clock-app/+bug/1431579
[00:31] <nik90> zsombi, popey: I will investigate this tomorrow morning when I find some time. Just letting you know in advance.
[00:49] <newsages> zxing?
[07:39] <dpm> bzoltan_, zbenjamin, morning! Quick question: when running QML unit tests with qmltestrunner. Can they be executed from QtC?
[07:45] <zbenjamin> dpm: you could create yourself a custom runconfiguration to execute whatever you want, but only on the desktop
[07:45] <zbenjamin> dpm: currently we do not automatically create a runconfig for tests
[07:48] <dpm> zbenjamin, thanks! Is this something that you think would make sense to add as an additional runconfig to the QML templates? Is there anything stopping us from doing it (other than the fact that needs to be planned, etc.)?
[07:51] <zbenjamin> dpm: the fact that there is no standard how to define and where to put tests. The project files do not contain them so its guesswork what and where to start the tests from
[07:52] <dpm> zbenjamin, ok, that makes it clear. So I guess that's the first step we need to fix before looking at the templates. Do we have the same situation for Autopilot tests?
[07:52] <zbenjamin> dpm: exactly the same
[07:52] <dpm> zbenjamin, ok, got it. Expect to hear more from us on the subject of testing soon, then :)
[07:54] <zbenjamin> dpm: its also questionable if it really makes sense as a runconfiguration... you want to see the output after all so maybe a testing mode would be better.
[07:55] <dpm> zbenjamin, when you say "mode", what do you exactly mean? As a separate Testing tab?
[07:55] <zbenjamin> yes
[07:56] <dpm> zbenjamin, is this something upstream haven't looked at implementing? I think what you're saying makes sense, but it strikes me that it'd be an Ubuntu-specific tab
[07:56] <zbenjamin> dpm: nop, there is no Testing tab
[07:57] <zbenjamin> dpm: i think tests is usually something you would run from the console
[07:57] <dpm> yeah, that's what you do now, but then it forces you to get out of the IDE
[07:57] <dpm> which also sort of implies that tests are something like an afterthought
[07:58] <dpm> which is probably not what we want to promote
[08:10] <dholbach> good morning
[08:54] <mzanetti> rpadovani, hey ho. look what DanChapman created for us :) https://code.launchpad.net/~dpniel/component-store/liwa-xtra-props/+merge/252839
[09:41] <zyga> hey, how  the "developer namespace" on https://myapps.developer.ubuntu.com/dev/account/ is supposed to look like
[09:41] <zyga> I tried reverse domain names -- that doesn't work
[09:41] <zyga> I tried 'zygoon' but then it rejected zygoon.myapp
[09:41] <zyga> any ideas?
[09:44] <JamesTait> zyga, your namespace should be a single word, no dots, IIRC lower-case alphanumeric.
[09:44] <JamesTait> zyga, and then your pkg name needs to be like $pkg_name.$namespace
[09:45] <JamesTait> zyga, so e.g. calculator.zygoon
[09:47] <zyga> ah
[09:47] <zyga> thanks
[09:47] <zyga> docs are very stale
[09:47] <zyga> they still talk about reverse domain name things
[09:48] <JamesTait> zyga, do you have a URL? We need to get that fixed.
[09:48] <zyga> http://askubuntu.com/questions/343491/what-is-the-naming-convention-for-click-packages + random google for the errors I got
[09:48] <zyga> wait
[09:51] <JamesTait> zyga, we're in the process of transitioning from old-style reverse-DNS namespaces to the new short namespaces; so if a dev has already uploaded a package with the old-style namespace, it'll continue to have that namespace for now, even though that namespace isn't visible any more AFAIK.
[09:52] <JamesTait> But new packages require the new-style namespace.
[09:53] <zyga> JamesTait: I see, thanks
[09:53] <zyga> JamesTait: it would be excellent if there was a short example next to each of the forms on the website
[09:53] <zyga> JamesTait: e.g. myapp.example
[09:53] <zyga> JamesTait: and the same next to the namespace form
[09:54] <JamesTait> zyga, then watch that space - there's work ongoing to improve that right now.
[09:55] <JamesTait> zyga, it's an area that has caused some mild confusion, so we're in the process of cleaning it up.
[09:56] <zyga> JamesTait: cool, thanks a lot!
[09:56]  * zyga just published an app for safari books online
[10:17] <nik90> dpm: Hey, good morning, when you got a moment can you review https://code.launchpad.net/~nik90/ubuntu-clock-app/fix-translation-plural-forms/+merge/252838
[10:22] <dpm> nik90, approved, thanks!
[10:22] <dholbach> dpm, https://code.launchpad.net/~dholbach/help-app/1430735/
[10:22] <nik90> dpm: thnx a lot :)
[10:22] <dholbach> dpm, run 'make web' and have a look at the HTML
[10:23] <dholbach> dpm, ".image" is not implemented yet, but I did the ".textonly" according to your specification :)
[10:23] <dpm> dholbach, oh wow, nice!
[10:24] <nik90> dpm: yup I explained that to the bug reporter, but he offered a explanation in https://bugs.launchpad.net/ubuntu-clock-app/+bug/1431446/comments/2. So I figured I might as well do this
[10:25] <dpm> dholbach, I was thinking something else seeing DS-McGuire's screenshot the other day. It might still be useful to have the python-markdown's attribute list support we dropped a few days ago. For example if we want to give a link class="button", so perhaps it might make sense to re-add it. But step by step, first let's test .textonly
[10:25] <dpm> ok, thanks nik90
[10:26] <dholbach> dpm, I'm not sure I understand
[10:26] <dpm> let me put together an example
[10:27] <dpm> on this screenshot, the buttons at the bottom:
[10:27] <dpm> https://lh5.googleusercontent.com/-u4yjLmcthAY/VQCGoD9Q2WI/AAAAAAAAIjs/AqNShovl6go/w512-h830-no/xzUDQl1.png
[10:27] <dpm> they are links with the "button" class
[10:28] <dholbach> ok...
[10:28] <dpm> i.e. the html markup is something like <a class="button" href=http://example.com>Get in touch<a>
[10:28] <dholbach> right
[10:28] <dpm> and on the .md files we could have:
[10:29] <dpm> [Get in touch](http://example.com) {: .button}
[10:29] <dholbach> sure
[10:29] <dholbach> that's easy to do
[10:30] <dholbach> but what's changing in terms of the Q&A entries?
[10:30] <dpm> nothing
[10:30] <dholbach> ok
[10:30] <dpm> that's just another feature
[10:30] <dpm> we didn't have a use case for it, but perhaps we might have found one :)
[10:31] <dholbach> maybe we should start thinking of shorter markup
[10:31] <dholbach> {: .button}, {: .textonly}, {: .image} get a bit long
[10:31] <dholbach> especially if you type them for every question/answer pair
[10:32] <dholbach> but sure, let me add the button thing now
[10:33] <dpm> dholbach, let's get the .textonly first, though. I didn't want to distract you from that, just thought I'd give you a heads up that we might have a use case for that branch (or was it just a revision) we dropped
[10:34] <dholbach> .textonly is done in the branch
[10:34] <dpm> ah, I thought you were still working on it
[10:35] <dpm> regarding the markup, two things: python-markdown can also use this syntax: {: class="button textonly"}
[10:35] <dpm> that'd be shorter
[10:35] <dpm> however
[10:35] <dpm> I'm not sure we should be mixing the two things
[10:36] <dpm> - attribute lists just add classes to the final markup
[10:36] <dpm> - our textonly extension adds divs to have separate rows in the final markup
[10:36] <dholbach> right
[10:36] <dholbach> so the button use-case will be dealt with differently, right?
[10:37] <dholbach> that'll just be the plain old attr-list thing?
[10:37] <dpm> yes
[10:37] <dholbach> ok
[10:38] <JamesTait> dpm, ping re https://translations.launchpad.net/software-center-agent/trunk/+pots/click-feed when you have a minute or three.
[10:38] <dpm> dholbach, this might be a crazy idea, but, perhaps something like...
[10:38] <dpm> {: rows="eight-col"} for text only columns
[10:39] <dpm> and {: rows= "eight-col four-col"} for text+pictures
[10:39] <dpm> it'd be kind of reusing the attr-lists syntax
[10:39] <dholbach> I was more thinking of something like !!T and !!I
[10:39] <dholbach> or something
[10:39] <dpm> but still keeping it separate
[10:39] <dpm> that could work too
[10:39] <dholbach> yes, less to type, less to memorise
[10:40] <dholbach> and instead of having editors to think of what the html markup might be like, just go with a definition of "text layout" and "image layout"
[10:40] <dpm> yeah, I like that point. The thing I'd like perhaps is to make it more generic, i.e. think in terms of columns and not Text or Image
[10:41] <dholbach> mh
[10:41] <dpm> dholbach, in any case, it's just a question of the syntax, I'll test the branch for now, the important thing is that the markup conversion works
[10:41] <dpm> we can still decide the exact syntax later
[10:41] <dholbach> right, I wasn't getting hooked up on the specific naming
[10:42] <dholbach> it's more that I think it'd make sense (and make it easy for editors) if we nail the default cases and keep that part easy
[10:42] <dpm> It's nothing I feel strongly about, just a suggestion
[10:42] <dholbach> if we need special cases which need a more generic approach I think that's fine
[10:42] <dholbach> and we'll find a way
[10:42] <dpm> ok
[10:42] <dholbach> we might need our own markup for that too, as attr-list and our extension might conflict
[10:43] <dpm> yeah
[10:43] <dholbach> or might both pick up that syntax :)
[10:43] <dpm> hi JamesTait, what's up with it?
[10:43] <JamesTait> dpm, I have a confession.
[10:43] <dholbach> thanks dpm - I'll go with what we talked about - it should be easy to change the exact syntax later on and add a more generic case
[10:43] <JamesTait> dpm, it was very green, it's now very red, and I think that's because we changed the .pot file.
[10:44] <dpm> dholbach, sounds like a plan
[10:44] <dholbach> rock on
[10:44] <dpm> JamesTait, yeah, just noticed the slightly different colour :) what did you change in the .pot file?
[10:45] <dpm> do you have a diff somewhere?
[10:45] <dpm> I'd have expected the translations to stay the same unless the original strings changed
[10:46] <JamesTait> dpm, we have an endpoint in devportal that generates the .pot file based on what's in the DB.  Previously, that wasn't committed to source code control, because it didn't make sense to, since it wasn't changes in the source that caused those strings to update, but changes in the DB.
[10:46] <JamesTait> dpm, that endpoint now adds a msgctxt to each entry, and we have a tool to auto-import the .po files into our db and push to click index.
[10:47] <JamesTait> dpm, I suspect that's the breaking change, since the strings are still the same in the majority of cases.
[10:47] <dpm> JamesTait, that all sounds good so far. Adding a msgctxt shouldn't affect translations
[10:47] <dpm> have you changed the name of the template in LP?
[10:47] <dpm> oh, actually
[10:48] <dpm> yes, adding a msgctxt might affect the translations
[10:48] <dpm> as gettext then thinks it cannot reuse the original translations for a different context
[10:48] <dpm> and expects translators to review it
[10:49] <JamesTait> dpm, now, we have the old translations, and I have poedit and the new .pot - I can load the old translations, update them to the new template and upload them to lp for review, perhaps.
[10:50] <zyga> JamesTait: the "your account" section doesn't display paypal email correctly, it's only visible when clicking on the edit button
[10:50] <dpm> JamesTait, I think there are two options:
[10:50] <JamesTait> dpm, I've done the first part of that for Catalan, but haven't uploaded anything because - well, I'm not on the right translations team, not 100% certain of the process, and didn't want to further break things.
[10:51] <dpm> 1) What you are describing
[10:51] <dpm> 2) Ask translators to retranslate
[10:51] <dpm> TBH, I think 2) might be the safest bet
[10:51] <dpm> It's not a long translation, it's easy to do, and since the messages are generic enough there are global suggestions from other projects
[10:52] <dpm> so in many cases, it'll just be a matter of pointing and clicking
[10:52] <JamesTait> dpm, I thought that as well. :)
[10:52] <dpm> JamesTait, I can send an e-mail to the translators list and we can probably get this fixed in a matter of hours
[10:52] <dpm> JamesTait, however,
[10:52] <dpm> It's not yet clear to me how these translations are uploaded to the server
[10:53] <JamesTait> dpm, that would be fantastic. I'm sorry we broke it, we expected lp to gracefully handle this.
[10:53] <dpm> i.e. on my phone they are still in English
[10:53] <JamesTait> dpm, OK, I can explain that part.
[10:53] <dpm> JamesTait, so I think we need to explain that to translators and ensure that they are uploaded
[10:54] <JamesTait> zyga, in https://myapps.developer.ubuntu.com/dev/account/ ?
[10:54] <dpm> JamesTait, no worries about breaking it, it's not that you did it on purpose, and it makes sense to have context. But I think we need to explain it well to translators and make sure their translations are used
[10:55] <JamesTait> dpm, the basic process is: we periodically (i.e. whenever an admin changes highglights or departments) run a script to export a .pot file based on what's in the devportal db.
[10:55] <JamesTait> dpm, that gets commited to sca trunk, and lp picks up the change.
[10:56] <JamesTait> dpm, our (frankly amazing) translators do their magic.
[10:56] <zyga> JamesTait: no, there it is displayed
[10:56] <zyga> JamesTait: on... https://myapps.developer.ubuntu.com/dev/click-apps/1755/
[10:57] <zyga> JamesTait: (per app page)
[10:57] <JamesTait> dpm, the updated translations end up in a bzr branch.
[10:57] <zyga> JamesTait: there's a "Your account" section at the bottom
[10:57] <JamesTait> dpm, we have a tool that iterates over those files, parses them and updates the devportal db, pushing them out to click index.
[10:58] <JamesTait> dpm, the next time the phone talks to click index (via the store scope), the departments and highlights will have updated translations.
[10:59] <JamesTait> zyga, ah, I see it. Could you file a bug report against software-center-agent and I'll see that it gets picked up.
[11:00] <zyga> JamesTait: sure
[11:00] <JamesTait> Thanks, zyga. :)
[11:01] <dpm> JamesTait, ok, that's clear now. One last question: once the translations end up in the bzr branch, how often are they uploaded to the db?
[11:01] <zyga> https://bugs.launchpad.net/software-center-agent/+bug/1431818
[11:03] <JamesTait> dpm, I'm not quite sure about that. The process is new, so it might still be manual right now, but I know nessita is pushing hard to get this all running smoothly right now, so it might become a daily cron job or something.
[11:03] <JamesTait> Thanks, zyga!
[11:04] <dpm> JamesTait, ok thanks. Could you try to clarify that before I send the e-mail to translators?
[11:04] <JamesTait> dpm, certainly will. :)
[11:05] <dpm> cool, thanks!
[11:06] <JamesTait> dpm, apologies again for the breakage, and thanks for the support.
[11:06] <dpm> no worries :)
[11:06]  * JamesTait adds dpm to the cake list.
[11:06] <dpm> :-)
[11:49] <rpadovani> mzanetti, cool!
[12:50] <AskUbuntu_> Ubuntu touch how to get HTML 5 app template running? | http://askubuntu.com/q/596310
[12:52] <nik90> zsombi: During my clock app test on vivid, I am still getting crashes consistently at different parts of the app like opening/closing the settings page, swiping up the bottom edge etc..I have a feeling that the way I am passing the alarmModel across the app from mainView seems to be the cause.
[12:53] <nik90> zsombi: though this all happens only on vivid :/
[12:54] <zsombi> nik90: because on vivid we have different way on using the model...
[12:57] <nik90> zsombi: hmmm..I will look through the alarm example app that you use in the test suite to see if I can spot the difference.
[13:02] <zsombi> nik90: my sample still uses the single page alarm management, I did not push the changes I made when I tried your page stack way...
[13:03] <zsombi> nik90: however you can check how the things are done
[13:03] <nik90> zsombi: yeah I noticed...the only other difference I see is that I load the alarm model using a QML Loader while you load it normally and pass it on to different part of your sample app. The usage otherwise seems pretty much the same.
[13:04] <zsombi> nik90: eventually I could prepare the page stack mode I made
[13:20] <dobey> does oxide not support javascript dialogs?
[13:21] <zyga> hey, I have a question about webapp-container's option --webappUrlPatterns=URL_PATTERNS -- what is the pattern exactly supposed to be? a glob or a regular expression
[13:21] <zyga> the --help page suggests a glob
[13:21] <zyga> but the example in generated app suggests it's regexp
[13:22] <zyga> and finally it's a weird regexp at that, trailing characters are /*
[13:22] <zyga> so what is it?
[13:23] <popey> have you looked at other webapps?
[13:23] <zyga> popey: no, where can I find them?
[13:23] <zyga> is there a launchpad project group for webapps?
[13:24] <popey> http://bazaar.launchpad.net/~popey/webapps-core/bbcnews/view/head:/bbcnews.desktop
[13:24] <popey> http://bazaar.launchpad.net/~popey/webapps-core/youtube/view/head:/youtube.desktop
[13:24] <popey> two examples of mine
[13:24] <zyga> hmm
[13:24] <zyga> ok, so what does ? mean
[13:24] <zyga> is it regexp or glob ?
[13:24] <zyga> and ditto for .
[13:24] <popey> ^ dbarth_
[13:28] <dbarth_> zyga: there is a slide deck where this is defined, and documentation on developer.ubuntu.com
[13:28] <dbarth_> 1 sec.
[13:28]  * nik90 is pissed at vivid UT..steps out to cool down
[13:28] <dbarth_> https://developer.ubuntu.com/en/web/tutorials/
[13:28] <dbarth_> zyga: ^^
[13:29] <dbarth_> this should help you with the url pattern
[13:29] <dbarth_> specifically, this is a list of regexps, but with some constraints on the extent of the pattern you can use (for security reasons)
[13:31] <zyga> dbarth_: if it's a regexp then how can http://bazaar.launchpad.net/~popey/webapps-core/bbcnews/view/head:/bbcnews.desktop work?
[13:31] <zyga> dbarth_: news* ?
[13:31] <zyga> dbarth_: that will match "new" followed by any number of 's'
[13:32] <zyga> dbarth_: is that true?
[13:32] <popey> it works.
[13:32] <zyga> popey: then it's not a regexp
[13:32] <popey> (I don't care)
[13:32] <zyga> popey: I'm trying to understand what it is exactly
[13:32] <zyga> popey: so that I can write what I want exactly
[13:32] <zyga> popey: without apt-get source'ing the code to check
[13:32] <popey> I made a webapp based on other webapps.
[13:32] <zyga> in any case the --help message is ambigous and could be improved
[13:33] <dbarth_> zyga: calling alexabreu who has the fine details of the regexp engine
[13:33] <zyga> dbarth_: thanks!
[13:34] <dbarth_> zyga: the bbc one should work with any number of symbols after 'new'
[13:34] <zyga> dbarth_: yeah, It looks like some weird syntax that's neither glob nor regexp, or a regexp that's preprocessed to correct some things
[13:36] <dbarth_> it is yes, it's not a true / full regexp
[13:42] <zyga> alexabreu: hey!
[13:42] <alexabreu> zyga, hey :)
[13:43] <zyga> alexabreu: how does the pattern matching on webapps work?
[13:43] <zyga> alexabreu: what's the syntax exactly?
[13:44] <alexabreu> zyga, what do you want to do?
[13:44] <zyga> alexabreu: understand the syntax to know what my patterns mean in practice
[13:44] <alexabreu> zyga, what patterns do you have?
[13:44] <zyga> alexabreu: looking at http://bazaar.launchpad.net/~popey/webapps-core/bbcnews/view/head:/bbcnews.desktop it's niehther a regexp nor a glob
[13:44] <zyga> alexabreu: that's irrelevant, I want to know what the system does
[13:45] <zyga> alexabreu: (and I want this to be documented on the webapps-container launcher for others)
[13:46] <alexabreu> zyga, basically you cannot have patterns for TLD/SLD, the subdomain wildcards are basically replaced by [^/]* and for the path its "free for all", the widlcard really means .*
[13:46] <zyga> alexabreu: hmm, what do you mean by the first part of your statement?
[13:46] <zyga> alexabreu: *.foo.com/* is invalid?
[13:47] <zyga> alexabreu: how about (m|ssl).foo.com/*
[13:47] <alexabreu> zyga, no *.foo.com/* is valid
[13:47] <zyga> alexabreu: can you point me to the source that has all the rules or a document that has those rules please?
[13:47] <alexabreu> zyga, dont see those as full regexps, more like wildards w/ restricted semantics depending on the location
[13:48] <zyga> alexabreu: the problem is that this is not specified anywhere and I don't know what restrictions apply without testing case-by-case
[13:49] <alexabreu> zyga, there are some docs here https://developer.ubuntu.com/en/web/ubuntu-webapps-guide/
[13:49] <alexabreu> zyga, if you see some missing pieces, or some parts that are unclear please tell me
[13:50] <zyga> alexabreu: thanks, looking!
[13:50] <zyga> alexabreu: may I suggest to link to that website from w-c --help?
[13:51] <alexabreu> zyga, and yes, your use case is relevant since I can help you on a case by case basis and see if things can be improved on the pattern side
[13:51] <alexabreu> zyga, yes, very much so :)
[13:51] <zyga> alexabreu: ok, I'm debugging this... https://github.com/zyga/SafariBooksOnline/blob/master/SafariBooks.desktop
[13:51] <zyga> alexabreu: I realized that m. wasn't enough
[13:51] <zyga> alexabreu: added (ssl|m). instead
[13:52] <zyga> alexabreu: now I added *. because desktop version (that you can bump into) is going to endlessly redirect between the container and browser
[13:52] <zyga> alexabreu: so I started reading about how to describe everything and I realized this is not a regexp anymore by looking at the trailing /*
[13:52] <zyga> alexabreu: don't get me wrong, this is okay, it's just hard to discover what it really is at first glance
[13:53] <alexabreu> zyga, right, we wanted it to be simpler than regexps
[13:53] <zyga> alexabreu: after decates of globs and various regexp implementations I was expecting exactly one or the other
[13:53] <zyga> alexabreu: that's fine, this is similar to adblock somewhat
[13:53] <alexabreu> zyga, people can easily mess up regexps (also that are quite simple there), so ...
[13:53] <zyga> alexabreu: it just needs clear and discoverable explanation on what the pattern means
[13:54] <alexabreu> zyga, right, ...
[13:54] <alexabreu> zyga, so now on your side https?://*.safaribooksonline.com/* would do the work?
[13:55] <alexabreu> zyga, one thing that I might do is add a little "url pattern tester" in the docs
[13:55] <alexabreu> but I am not sure how it would help since you have to test it against something
[13:55] <zyga> alexabreu: I'm checking that now
[13:55] <zyga> alexabreu: yeah!
[13:55] <zyga> alexabreu: or a small app that you can run against a list of URLs
[13:56] <alexabreu> zyga, right ... something like that
[13:56] <zyga> alexabreu: I think just improving --help might be good
[13:56] <akiva-thinkpad> Hi all
[13:57] <alexabreu> zyga, yes
[13:58] <zyga> alexabreu: thanks for your help!
[13:59] <zyga> alexabreu: yeah, the updated pattern works ok
[13:59] <alexabreu> zyga, np, I'll take a note to improve the doc
[14:05] <zyga> alexabreu: https://developer.ubuntu.com/en/web/api/ ? 404?
[14:06] <zyga> alexabreu: semi-offtopic, is it possible to navigate from the generic browser to a webapp?
[14:06] <zyga> alexabreu: say, I google for example.org
[14:06] <zyga> alexabreu: click on a link from google
[14:06] <zyga> alexabreu: and instead of following there inside the browser
[14:06] <zyga> alexabreu: switch to example.org webapp
[14:07] <zyga> alexabreu: ?
[14:07] <alexabreu> zyga, unfortunatley no, ... not from the webbrowser (but from other apps yes)
[14:08] <zyga> alexabreu: that would imply that a global (or per user) list of patterns exist and that conflicts have a handler that can resolve them
[14:08] <alexabreu> zyga, mmmh the 404 is a bug indeed
[14:09] <zyga> alexabreu: the same thing happens on the next tab (cookbook)
[14:09] <alexabreu> zyga, yeah
[14:09] <ogra_> we hade too many cooks that broke it ;)
[14:09] <ogra_> *had
[14:10] <alexabreu> zyga, well we have that (list of handlers), the list is not queryable though per se by an external app, which makes it has to know if an app outside the webbrowser can handle the url per se
[14:10] <alexabreu> zyga, it is a good suggestion though
[14:10] <zyga> alexabreu: oh, that's great then
[14:10] <zyga> alexabreu: I think it would be a great way to fix some navigation issues
[14:11] <zyga> alexabreu: like clicking on g+ from gmail
[14:11] <zyga> alexabreu: having to re-authenticate in the basic browser
[14:11] <zyga> alexabreu: and then clicking on gmail
[14:11] <zyga> alexabreu: and ending up with two gmails
[14:12] <alexabreu> zyga, well, it is debatable though, kind of hard to know what the user wants, and might be disturbing to the user all those app changes
[14:12] <zyga> alexabreu: yeah, I'm sure this is not the final perfect design but auth issues are also problematic
[14:12] <alexabreu> zyga, having a ui option to open a tab as a webapp if one exists for that website could be interesting
[14:12] <zyga> alexabreu: double so if you have 2fa
[14:21]  * zyga is totally psyched by insta-reviews
[14:21] <zyga> upload, refresh, published
[14:30] <zyga> I found a bug in unity 8, install app, rename desktop file, push to store, update, tap on sidebar
[14:30] <zyga> it slows down (freezes) heavily
[14:30] <zyga> after unpinning the launcher and re-launching it's okay
[14:59] <dholbach> dpm, lp:~dholbach/help-app/1430735/ now can do class='button' and does the .textonly thing according to what we said in the bug report
[14:59] <dholbach> markup is "{: .button }" (using attr_list)
[14:59] <dholbach> and !!T and !!I
[14:59] <dholbach> ... which we can obviously change to whatever else
[15:00] <dholbach> but it seems to work reliably in the branch now
[15:00] <dholbach> shall I propose it as a partial fix for bug 1430735?
[15:00] <dpm> dholbach, awesome! I won't probably have time to test this today, but I'll have some time Monday morning
[15:00] <dholbach> I know that !!I still needs to be implemented
[15:00] <dpm> dholbach, yeah, makes sense to propose it
[15:00] <dholbach> ok
[15:01] <dholbach> dpm, balloons: https://code.launchpad.net/~dholbach/help-app/1430735/+merge/252910 :)
[15:05] <balloons> dholbach, Text conflict in edit-here/translations.py
[15:05] <dholbach> yep
[15:05] <dholbach> I'll merge from trunk
[15:05] <dholbach> that'll fix it
[15:06] <dholbach> done
[15:16] <dholbach> balloons, I added a commit message with a bit more explanation
[15:17] <balloons> danke
[15:27] <nik90> dpm: Any idea why bug 1431914 happens where it loads the translations of another language than the one chosen in system settings?
[15:46] <doflaherty> I've got an app and I want it to be able to export pictures created in it via the content hub - where can I save the file on the device?
[16:06] <dobey> doflaherty: $XDG_DATA_HOME
[16:07] <dpm> nik90, I don't know, I suspect bad translations? I'll check it out later
[16:08] <dpm> the system should deal with loading the correct translations for the right lang code
[16:08] <dobey> nik90: what version of the clock app is that in?
[16:08] <nik90> dpm: the bug reporter has provided strings which have been translated into pt_BR and pt...and yet it defaults to pt incorrectly sometimes
[16:09] <nik90> dpm: the current version in the store is 3.3.192
[16:09] <dobey> nik90: that string was translated in pt_BR after that revision
[16:10] <nik90> dpm: oh .. so if translations are not present in pt_BR, it then defaults to pt?
[16:10] <nik90> I mean dobey ^^
[16:10] <dpm> that could well be, but I'm not sure
[16:10] <dobey> gettext automatically falls back to xx when the lang is xx_YY and xx_YY doesn't have the translation. if xx has the translation, then it is displayed. if not, then it is displayed untranslated
[16:10] <dpm> I know it's the behaviour for ca and ca@valencia
[16:11] <dpm> ok, then that it is
[16:11] <dobey> nik90: so afaict, simply releasing an update will fix it
[16:11] <dpm> thanks dobey
[16:11] <nik90> dobey, dpm: ack. thnx
[16:11] <doflaherty> dobey - is that available from QML?  I am saving a canvas, which works if I save it as /tmp/foo.png but I'm not sure if that is only working because I'm running from QtC
[16:11] <nik90> dobey: yeah releasing to the store is pending QA feedback atm :)
[16:12] <dobey> nik90: sure. just being clear what the "fix" is for that bug :)
[16:12] <dobey> doflaherty: you need to write c++ to use QStandardPaths afaik
[16:12] <doflaherty> dobey, yeah that's what the camera-app seems to do
[16:17] <doflaherty> I didn't think confined apps could write to temp though
[16:31] <dobey> doflaherty: they can't
[16:31] <dobey> doflaherty: they can write to their confined cache and data directories
[16:31] <doflaherty> dobey: and that's what XDG_DATA_HOME points to?
[16:32] <dobey> doflaherty: i'm pretty sure it is overridden for confined apps to point to their confined directory, yes. as is $XDG_CACHE_HOME
[16:33] <dobey> QStandardPaths::writableLocation(StandardLocation::GenericDataLocation) for example should give you the $XDG_DATA_HOME you can write to
[16:33] <dobey> and ::GenericCacheLocation for the cache dir
[16:34] <dobey> anyway, i need to go get lunch
[16:34] <doflaherty> alright, thanks
[16:36] <ogra_> why oh why is the QT-project forum css so broken ...
[16:36] <rpadovani> DanChapman o/ Do you know a C++ class to create a "QmlMergeModel" where you can merge more than on model? gcollura says you're the right person to ask to :-)
[16:37] <nik90> ogra_: that's their new design :P
[16:37] <ogra_> nik90, one that makes all code look crappy as well ?
[16:37] <nik90> hehehe
[16:38] <nik90> ogra_: although I love what they did with the qt docs though ..for instance http://doc.qt.io/qt-5/qml-qtquick-listview.html
[16:39] <dholbach> dpm, balloons: I updated https://code.launchpad.net/~dholbach/help-app/1430735/+merge/252910 - we're almost there
[16:39] <ogra_> http://forum.qt.io/topic/39756/qml-and-js-reading-json-file-and-change-content-on-the-go/8
[16:39] <ogra_> something like this
[16:39] <dholbach> dpm, balloons: just need to figure out how to parse the image from the markdown, so we can safely put it into its own <div>
[16:39] <ogra_> where all indendation is one and you have parts with scrollbars in the middle of code blocks etc
[16:40] <nik90> ogra_: yeah that looks hideous with no indentation
[16:40] <ogra_> yep
[16:43] <dholbach> and with that....... I call it a day - have a great weekend everyone!
[16:45] <nik90> dholbach: enjoy your weekend :)
[16:46] <dholbach> thanks nik90 - you too!
[16:47] <mzanetti> seb128, look at that... .5 appeared here :)
[16:49] <seb128> mzanetti, go figure ... better later than not though ;-)
[16:49] <mzanetti> indeed :D
[17:01] <JamesTait> dpm, still around?
[17:07] <nik90> rpadovani: hey, if you got time, can you review https://code.launchpad.net/~nik90/ubuntu-clock-app/improve-setting-listitem/+merge/252930
[17:08] <rpadovani> nik90, on Sunday night, sorry, out of home
[17:08] <nik90> rpadovani: np
[17:47] <mzanetti> seb128, confirming that it works again. thanks a lot
[17:48] <dpm> JamesTait, around, but about to call it a day
[17:48] <JamesTait> dpm, nm, sent you mail, otherwise I'll forget. :)
[17:49]  * JamesTait is getting old.
[17:49] <dpm> JamesTait, ok, perfect, thanks! I saw the e-mail on my inbox, haven't had the chance to read it yet
[17:49] <seb128> mzanetti, yw!
[17:50] <dpm> balloons, http://pastebin.ubuntu.com/10592112/ (the output you were asking for on your comment in the tutorial)
[17:51] <balloons> dpm, thanks.. I managed to recreate.. though I don't know why. It'll likely ship like that for now unless I find a fix. Going to try and pages published :-)
[17:51] <mzanetti> balloons, can I help?
[17:51] <mzanetti> (assuming this is about the qmltests)
[17:52] <dpm> balloons, yeah, I think it'd be good to investigate it, but I don't see it as much of an issue, shouldn't block publishing the tutorials
[17:52] <balloons> mzanetti, sure.. Just a weird error being given by qmltestrunner before the testrun
[17:52] <mzanetti> ah, the <unknown file>
[17:52] <balloons> right
[17:52] <mzanetti> weird indeed... let me have a look
[17:52] <balloons> there's a few updates I'm going to push, but they didn't change anything so you should be good
[17:57] <balloons> mzanetti, if you also want to lend an eyeball to the tutorials, I'll share the link
[18:14] <nik90> balloons: what was the website where you can see all the necessary autopilot bug requests?
[18:14] <nik90> for core apps
[18:14] <balloons> nik90, qml tests too! http://people.canonical.com/~nskaggs/core-apps-test-sponsoring.html
[18:14] <nik90> thnx
[18:14] <nik90> balloons: :-)
[18:16] <ahayzen_> balloons, what is that list sorted by?
[18:16] <boghison> Hi! Could someone help me in regards to OAs?
[18:17] <balloons> ahayzen_, should be by priority then age
[18:17] <balloons> clicking a column will resort
[18:17] <ahayzen_> hmmm
[18:17] <ahayzen_> ah
[18:17] <ahayzen_> it was like totally random for me lol
[18:20] <balloons> lol
[19:29] <balloons> so ahayzen_ nik90 about?
[19:29] <ahayzen_> balloons, yup
[19:29] <balloons> ahayzen_, what do you think? https://developer.ubuntu.com/en/apps/quality/
[19:29] <balloons> there's a couple new qml testing tutorials there
[19:29] <ahayzen_> ooo :D
[19:29] <balloons> https://developer.ubuntu.com/en/apps/qml/tutorials/qml-unit-testing/ and https://developer.ubuntu.com/en/apps/qml/tutorials/qml-integration-testing/
[19:31] <ahayzen_> balloons, thats gonna help sooooooo much :)
[19:39] <balloons> ahayzen_, I hope so. The autopilot tests too in the example are much updated
[19:39] <ahayzen_> :)
[19:40] <balloons> they were old and didn't incorporate all the new goodies.
[19:45] <zyga> balloons: woot, that's pretty nice!
[19:46] <balloons> I hope so.. now to go on mission and get folks to write some unit tests!
[19:47] <zyga> balloons: we're strugging with checkbox unit tests
[19:47] <zyga> balloons: autopilot crashes on mesa/llvm when we run our tests on vms
[19:48] <zyga> balloons: we're pretty stuck with that
[19:48] <balloons> zyga, struggling because?
[19:48] <balloons> ohh.. well, nothing with qmltestrunner then, just autopilot?
[19:48] <zyga> balloons: everything works on device and on non-virtualized desktops and on virtualized machines
[19:48] <zyga> balloons: stuff crashes on canonistack
[19:48] <zyga> balloons: with xvfb
[19:48] <zyga> balloons: qt crashes there on some low level mesa/llvm bug
[19:48]  * zyga looks for the link
[19:48] <zyga> (in essence, we cannot run our tests in CI)
[19:49] <zyga> I was wondering how core apps get CId
[19:51] <zyga> balloons: http://paste.ubuntu.com/10585366/
[19:51] <zyga> balloons: LLVM ERROR: Cannot select: intrinsic %llvm.x86.sse41.pblendvb
[19:51] <zyga> balloons: have you seen anything like that?
[19:53] <balloons> zyga, this is on canonistack where exactly? is this CI run?
[19:53] <balloons> the core apps all run in CI supported environments
[19:54] <balloons> what that amounts to though is a desktop vm with Xfvb
[19:54] <zyga> balloons: we have a juju-spawned tarmac instance
[19:54] <balloons> why not run on device directly?
[19:54] <zyga> balloons: we don't have that and this is not device-centric
[19:54] <balloons> I know CI has had some trouble with virtualizing some of these apps
[19:55] <zyga> balloons: our tree has many components that get tested uniformly on multiple releases
[19:55] <zyga> balloons: and we're not using anything the CI team came up with as back when we started CI story was poor and what we wanted didn't work
[19:55] <zyga> balloons: so we've built our own and sticked to it
[19:55] <zyga> balloons: tarmac spawns lxc/vm for each supported release
[19:56] <zyga> balloons: inside we provision dependencies and run our tests
[19:56] <zyga> balloons: (for lots of thingds)
[19:56] <zyga> balloons: adding autopilot is a new thing and it fails for us now :/
[19:57] <balloons> ahh.. well, such is life rolling your own I guess, heh ;-(
[19:57] <zyga> balloons: well, hence the question, how does CI run tests for core apps?
[19:57] <balloons> can you run inside a different environment? does it matter?
[19:57] <balloons> zyga, Xvfb
[19:57] <zyga> balloons: that's what we're using
[19:57] <zyga> balloons: and it works everywhere except for canonistack somehow
[19:57] <zyga> balloons: we suspect it might be the raw hardware (old machines?)
[19:58] <zyga> balloons: as it all works on 1st gen core i7 and younger
[19:58] <zyga> balloons: we can run it on any laptop in the back like we used to
[19:59] <balloons> zyga, well what if you run it on bare metal, even provisioned?
[19:59] <balloons> does it change things?
[19:59] <zyga> balloons: it works everywhere except for canonistack so far
[20:00] <zyga> balloons: we cannot provision bare metal on canonistac, can we>
[20:04] <balloons> zyga, no; no magic powers there afaik
[20:05] <zyga> balloons: we'll keep digging
[20:05] <zyga> balloons: there are reports of bug like that all over the place but it seems to be fixed
[20:05] <balloons> zyga, I guess I was saying 2 things. 1, talk to CI as they have had similar issues and they might have a solution. 2) find somewhere else to run the tests for now if you can
[20:06] <balloons> afaik, for now CI is sticking with bare metal, but I recall someone calling they had it working  . . .
[20:06] <zyga> balloons: yeah, I plan on working closer with the CI team starting from this month
[20:06] <zyga> balloons: to unify what we do
[20:10] <dobey> zyga: you are trying to run autopilot tests without a dbus session?
[20:10] <zyga> dobey: oh, curious question, I don't know, let me look
[20:10] <zyga> https://code.launchpad.net/~kissiel/checkbox/autopilot-testing/+merge/252731
[20:10] <zyga> this is the merge request that triggered this
[20:11] <dobey> zyga: well, the python trace is clearly complaining about an issue trying to get the dbus address :)
[20:11] <zyga> I suspect so
[20:11] <zyga> +xvfb-run --server-args='-screen 0 1024x768x24 -ac +render -noreset' make autopilot
[20:11] <zyga> that's what we run
[20:11] <zyga> (+ is from the diff)
[20:11] <balloons> zyga, you need a window manager too
[20:11] <zyga> ohh :)
[20:11] <zyga> so what's the minimal set of this to do?
[20:11] <balloons> well, probably not for checkbox, but it will simulate better using one
[20:11] <dobey> running autopilot tests in CI like this is a pain in general
[20:12] <balloons> we use a modified autopilot3-sandbox script, and use ratpoison
[20:12] <zyga> balloons: where can I get that?
[20:12] <zyga> and what is ratpoison?
[20:12] <zyga> we could use dbus-launch to get a session bus
[20:13] <zyga> but to me that still looks like it will crash on llvm :/
[20:13] <balloons> zyga, also have you used or looked at adt?
[20:13] <zyga> oh, yes, we use that but not for CI
[20:13] <dobey> zyga: this is to test checkbox built as a click?
[20:13] <zyga> this is before packaging
[20:13] <zyga> no, this gets a merged but not commited branch
[20:13] <balloons> zyga, http://bazaar.launchpad.net/~autopilot/autopilot/trunk/view/head:/bin/autopilot3-sandbox-run
[20:13] <zyga> runs all tests on all components (python, C++, jobs, etc)
[20:14] <zyga> one of that is autopilot
[20:14] <dobey> zyga: you're running your own tarmac setup for checkbox, and not having your branches landed by the CI team's stuff?
[20:14] <zyga> dobey: yes
[20:14] <balloons> zyga, ratposion is a simple autofocusing window manager. it plays real nice with autopilot and touch apps
[20:14] <zyga> dobey: this has been running for 2+ years now
[20:14] <zyga> balloons: ah, I get it
[20:14] <zyga> dobey: we'd love to discuss how to simplify what we need to run
[20:14] <zyga> dobey: but bear in mind we're not just running click
[20:15] <dobey> zyga: sure. i wish i could run my own taramc for the stuff i have to work on
[20:15] <zyga> dobey: we run tests on 12.04, 14.04, 14.10 and 15.04 for mobile, desktop and server components
[20:15] <zyga> dobey: well, what stops you?
[20:15] <dobey> zyga: sure, but CI has jobs which build a click, install it on a device, and run the autopilot tests on an actual running device
[20:15] <zyga> dobey: we have our stuff charmed
[20:15] <zyga> dobey: with documented configs and scripts
[20:15] <zyga> dobey: ah, yeah
[20:16] <zyga> dobey: we'd love that iff that was a service we can tie into
[20:16] <dobey> zyga: management and the whole "fitting into the standard UE-wide processes" bit is what stops me really
[20:16] <zyga> dobey: not a all-or-nothing :/
[20:16] <zyga> dobey: yeah, I understand
[20:16] <zyga> dobey: I'd love travis-like service for all ubuntu engineering
[20:16] <zyga> dobey: where you just stick a trivial file in your git tree
[20:16] <zyga> dobey: and magic happens elsewhere
[20:16] <dobey> zyga: i'm not sure what you'd specifically need that wouldn't be doable with the UE CI setup
[20:16] <zyga> dobey: supporting all the ad-hoc CI costs a lot
[20:17] <dobey> ugh yaml
[20:17] <zyga> dobey: well, whatever
[20:17] <zyga> dobey: yaml is irrelevant
[20:17] <zyga> dobey: it's the amount of automation on the other hand
[20:17] <zyga> other *side*
[20:17] <dobey> yeah, i'm just a little annoyed that we're actually using it for things :)
[20:17] <zyga> we have nothing like that in ubuntu
[20:17] <zyga> using what? travis?
[20:17] <dobey> yaml
[20:18] <zyga> well, could have been worse
[20:18] <dobey> how is the UE CI nothing like travis?
[20:18] <zyga> at least it's not xml :P
[20:18] <dobey> i mean, setup isn't trivial, but it should be able to do everything that's needed
[20:18] <zyga> yaml is moredately easier to write and has support for comments, parses are all bad and had a history of buffer overflows and such
[20:19] <zyga> dobey: yeah, but there's nothing that provides an integrated package like travis+github
[20:19] <dobey> i don't understand that
[20:19] <zyga> dobey: where anyone can start a new project, write 3 lines and have total CI support
[20:19] <zyga> dobey: do you know what travis is?
[20:19] <dobey> oh well
[20:19] <dobey> yes i know what travis is
[20:19] <zyga> dobey: sorry, I'm not making much sense, it's late and I'm tired
[20:20] <zyga> dobey: I just wish there was something like that for folks working on ubuntu
[20:20] <dobey> but complaining that it's not a 3 line file you dump in the repo that gets you set up in the current CI train, is not a useful complaint as to why we shouldn't use UE CI
[20:20] <zyga> dobey: or at least for canonical folks at first, when it is in beta
[20:20] <zyga> dobey: I don't even know how to begin using UE CI
[20:20] <dobey> zyga: it's more than 3 lines and not directly in your tree, but we have something like that
[20:20] <zyga> dobey: I wish I could
[20:20] <zyga> dobey: I tried to find out many times
[20:20] <zyga> dobey: it's not discoverable
[20:21] <zyga> dobey: and I don't even know what the CI train does in practice, if that covers what we're using so far
[20:21] <dobey> ok, "not discoverable" is a useful complaint that you should make to the CI team (or just your manager, so they can complain to the CI team)
[20:21] <dobey> zyga: #ubuntu-ci-eng and "cihelp: i need help understanding how to get a project set up in CI train" :)
[20:22] <dobey> zyga: but maybe better for you to do on monday at this point :)
[20:23] <zyga> dobey: cihelp is a bot or something?
[20:23] <dobey> no, it's just a term that all the ci folks have highlighted in their clients
[20:23] <dobey> to get the attention of the ci vanguard
[22:19] <mhall119> popey: docviewer appears to have lot URL dispatcher/Arguements support
[22:19] <popey> hmm?
[22:19] <popey> que?
[22:19] <mhall119> lost
[22:19] <mhall119> not lot
[22:20] <popey> file a bug pls
[22:24] <mhall119> maybe it's content-hub it's missing
[22:26] <mhall119> popey: https://bugs.launchpad.net/ubuntu-docviewer-app/+bug/1432066
[22:26] <popey> thanks
[22:26] <popey> ah no
[22:26] <popey> image support is removed
[22:26] <popey> document viewer is a document viewer
[22:27] <popey> not an image viewer.
[22:31] <mhall119> it used to be an image viewer
[22:31] <mhall119> sounds like we need a new core app project :)
[22:32] <popey> we have gallery
[22:32] <popey> thats what it's for
[22:33] <mhall119> I just want to *view* an image, not add it to my collection
[22:34] <mhall119> you don't add image attachments to Shotwell on the desktop
[23:03] <popey> mhall119: some do
[23:03] <popey> mhall119: an image is not a document though
[23:04] <popey> docviewer is too heavyweight for mere images.
[23:04] <popey> yes, make an image viewer :)
[23:48] <nik90> balloons: ping