[Ubuntu-phone] [Design] Installed fonts in Ubuntu Touch
Hi all, What fonts will come installed by default in Ubuntu Touch? Obviously the Ubuntu font will be there, but what else? The Nimbus family? The URW++ fonts? How about Bitstream Charter? I'm asking because I'm working on a Epub reader for the App Showdown. I want to give the user the option of specifying the font. Rather than overwhelming the user with all the fonts on the system, I thought I should pick out 5-8 good ones for them to choose from. For simplicity, I thought I'd stick with those installed by default. Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] [SDK][Theming] Styling a compound widget
Hi all, This email is going to run on a bit, so let me put my questions here at the top and motivate them afterwards: 1) Is it possible to adjust the style of a single compound widget? (That is, a QML element with several sub-elements. I don't know the terminology here.) 2) If not (and I haven't found one), should there be? I'm working on an epub reader for the app showdown [1]. One of the options I provide is the font in which the epub will be displayed. I want to do this with an OptionSelector, and I'd like to display the name of each font in that font. It looks like this is impossible with the SDK as it currently stands. (Please correct me if I'm wrong.) With simple QML widgets, like a Rectangle, you can adjust the properties of the widget to change its styling. But the widget I need to adjust, an OptionSelectorDelegate, is a compound widget -- the text is in a Label inside of the OptionSelectorDelegate, and there's no way for me to access this. These Labels don't have ids, and even if they did those ids wouldn't be accessible outside of that .qml file, for some reason. I came up with a simple fix to this: Give the Labels ids, and then give the OptionSelectorDelegate properties referencing the label. Then when I make an OptionSelectorDelegate, I can adjust the font of the label through the textLabel.fontFamily property. I've written up a patch and submitted it to Launchpad [2]. I asked about this on IRC, and mzanetti opined that this was not the right approach. Internals shouldn't be exposed like this. And while I disagree, this is not an unreasonable position. Instead, mzanetti suggested that the OptionSelectorDelegate should become themable, and then I could provide my own theme to do the styling. It may be that it should be themable, but this wouldn't solve my problem. I don't want to redo the theming of the widget; I just want to tweak the font. I could do this, theoretically, by creating a new widget based on the OptionSelectorDelegateStyle widget and tweaking the font there. But this runs into three problems: 1) Not all the *Style widgets are exposed in the Theme directories. (I recently ran into this problem when trying to theme a Slider; see [3].) 2) I would have to inherit from a specific theme. If that changes in the future, due to system changes or user choice, my styling would be completely wrong. 3) I would still need access to the internal Label of the style widget, which is the very thing that got us going down this path! Hence my questions: Is there a way to set the font in an OptionSelectorDelegate? There doesn't seem to be. Should there be? I think so. How should this work? This I don't know enough about to comment. Thanks, Robert [1] https://github.com/rschroll/beru [2] https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1220533 [3] https://github.com/rschroll/beru/commit/927b2fd377d7cc86b0633a2707b03c3a79a330db -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [SDK][Theming] Styling a compound widget
Hi Zsombor,
Thanks for the detailed reply! There's still some things I'm confused
about, but this may reflect more my lack of understanding of QML. So
don't be afraid to tell me that these are stupid questions.
On Mon, Sep 9, 2013 at 8:41 AM, Zsombor Egri
wrote:
2) If not (and I haven't found one), should there be?
Perhaps. It all depends on the use cases and UX. But for sure there
would be use cases where having such a feature in theme engine would
be beneficial.
If you're going to ask for a use case for each feature, you're dooming
yourself to a long game of Whack-a-Mole, where you have to keep adding
features each time someone comes up for a reasonable use for them. IMO,
you should either declare none of them will be allowed (a mistake, I
think) or allow access to everything and trust us devs to use that power
responsibly.
(And by 'you', I mean the Ubuntu SDK in general, not Zsombor specifically!)
2) I would have to inherit from a specific theme. If that changes
in the future, due to system changes or user choice, my styling
would be completely wrong.
Themes provided by the SDK will always be in sync with the
components. So if you need to override some properties from the
style, you will be most likely on the safe side.
Does this mean that users won't be able to choose their own themes? If
so, what does this mean for our vision-impaired users that need bold
white text on a black background to be able to use their devices?
CustomStyle.qml in Ubuntu.Components.Themes.Ambiance
Item {
property string firstLabelFamily: "Ubuntu"
property string secondLabelFamily: "Courrier New"
}
CustomComponent.qml
StyledItem {
id: custom
Label {
font.family: __styleInstance.firstLabelFamily
text: "First line"
}
Label {
font.family: __styleInstance.secondLabelFamily
text: "Second line"
}
style: Theme.createStyleComponent("CustomStyle.qml", custom)
}
In this way the CustomStyle doesn't do anything else but styles the two labels
inside the CustomComponent. Same thing is needed in OptionSelectorDelegate.
Maybe I'm just being dense, but I don't see how this helps me out. I
don't want to implement a CustomOptionSelectorDelegate that uses the
same style as the default one; I want to take the existing one and tweak
one aspect of its style. It seems to me that either (1) there has to be
a property (whether in the Delegate itself or in its associated Style)
that sets the font for me to tweak (but then you're back to Whack-a-Mole
trying to expose the appropriate things), or (2) there needs to be
access to the sub-widgets (whether in the Delegate itself or in its
associated Style) for me to adjust directly.
Robert
--
Mailing list: https://launchpad.net/~ubuntu-phone
Post to : ubuntu-phone@lists.launchpad.net
Unsubscribe : https://launchpad.net/~ubuntu-phone
More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [SDK][Theming] Styling a compound widget
On Wed, Sep 11, 2013 at 8:43 AM, Michał Sawicz wrote: I feel like there has been a misunderstanding here... IIUC, you want an OptionSelector with: Ubuntu DejaVu Liberation ... Each of those in the font it describes, right? IMO that's the exact place where you create a custom OptionSelectorDelegate that will change the font for its own Label. I agree that exposing the font on the standard OptionSelectorDelegate is not the way to go - if you need to customize it - just build your own with the tweaks you need. That's what I'm doing now. But I want this CustomOptionSelectorDelegate to look just like the standard OptionSelectorDelegate, except for font. Right now, I've copied all the code from OptionSelectorDelegate to my own copy, which gets tweaked. But this is stupid: Every time the SDK folk update OptionSelectorDelegate, I'm going to have to copy the new code into my Custom version. And if the OptionSelectorDelegate is different in 13.10 and 14.04, I'm going to need to do version sniffing! You need to also remember that "exposing 2 I'm new to QML, but isn't this why import statements have version numbers? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] Whither the Content Hub?
Hi all, Beru, my epub reader app, is currently hamstrung by apparmor restrictions. Instead of being allowed to read any books the user puts in her home directory, it can only read epub files if they're stored in ~/.local/share/com.ubuntu.developer.rschroll.beru/. I was told that the Content Hub would provide a way around this silliness, but I haven't heard any updates on it. So, is the Content Hub a thing? Is it in 13.10? Is it backported to previous releases? Is it turned on by default? How do I use it? The only documentation I've found on it is the source, where there's a short description that's more polemic than tutorial and several brief, uncommented examples. Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] How do I know an app is safe to install?
On Tue, Oct 15, 2013 at 9:16 AM, Michael Zanetti wrote: App Developers don't want to publish their code and the vast majority of users doesn't seem to care about anything security at all anyways. It's a sad situation for people like us who actually DO care about security. However, I haven't given up hope that at some point someone will set up some App Repository for Ubuntu Touch which requires developers to upload a source package, the binary will be built on the trusted server and the exact same source archive published along with the binary. Personally, I'd much prefer to publish through this system. Not only would it provide more assurance for my users that my source actually is my source, it would free me from having to play with pbuilder and chroots to build click packages. I doubt that I'm typical in this regard, but one datum is better than none, right? Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] Accessing the user's files in a click app
Hi all, Since my previous questions about the content hub didn't get me anywhere, I must have been asking stupid questions. Allow me to back up to a more basic topic. This is the workflow I wish to enable: 1) The user copies (via rsync or Ubuntu One folder or whatever) files to their home directory or subdirectory thereof of their device. 2) The user uses my app to display the contents of those files. 3) Optionally, my app can download more files of a similar type to the same location, so that the user can rsync (or whatever) these files back to their computer. Is this possible with click apps in from the app store? If so, how? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] problems submitting application
It turns out that deleting the app doesn't actually delete it -- it just hides it from you. It still exists, and this is what's blocking you from creating a new app with the same name. Somebody with admin power can go in and undelete the existing app, letting you submit a new version. You may want to ask in IRC for some help (#ubuntu-app-devel, I think). Robert Why yes, I did have this problem before. How did you guess? On Fri, Oct 25, 2013 at 11:36 AM, brian larochelle wrote: If this isn't the proper place, I would appreciate it if you could direct me to the correct place. When submitting my app the review had problems with my *.desktop file. So after correcting them I figured I could resubmit the app of the same version. I clicked the resubmit button like a dummie and caused the same click package to go into review, which would have the same problem. I decided to just delete the app from MyApps and start fresh. That prolly wasn't a good idea. Now when trying to publish my app, on the General submit screen, in Package name field i'm getting the following: Package name feed-the-aggregator You already have a Click Package with this package name. The package name for this application. Letters, numbers or hyphens allowed only. Perhaps there are some db entries left over from my previous attempts that can be cleared? I appreciate any assistance. Thanks -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Accessing the user's files in a click app
Hi all, Thanks for the many replies, all. It's a bit overwhelming, and I fear I may have missed something. I'm going to try to summarize; please correct my mistakes: 1) There's no way to do what I want to do right now. 2) The content hub may provide the capabilities I want in the future, but this may not look like file access. 3) There may be additional ways to access the file system in the far future. As several of you have correctly surmised, this is for my epub reader Beru [1], which is currently limited to reading epubs stored in ~/.local/share/com.ubuntu.developer.rschroll.beru/, which isn't exactly convenient for the user to access. But I asked this in general, because I don't want an epub-specific solution. I want something that will work for anyone who's trying to build a viewer for Ubuntu Touch. Let me try to clarify a few points: Jamie Strandboge wrote: your app would only get a copy of the file in the app-specific directory but no way to sync the file back out to its original location (in the future, we should/may also support passing a file descriptor instead of a full copy-- but this is an implementation detail that won't help your use case). [...] In short, your use case falls into the category of 'backup software' and backup software is not supported by the appstore at this time. You're crediting me with too much ambition. I don't want to modify or keep a copy of the epub files, I just want to read them. The write access I want is to be able to download new epub files, but I'm less concerned about that (at the moment). Marc Deslauriers wrote: The first message in this thread sounds like he wants to access arbitrary directories in the user's home folder. His original post however seems to indicate he wants his ebook reader to be able to read and write ebooks into arbitrary locations in the user's home directory. I'd prefer arbitrary locations, to allow the user to choose where to save their files and to allow for run-time localization. But if we restrict it to a give sub-directory at compile time, that wouldn't be the worst thing in the world. Perhaps I'm thinking too much about files because that's what I'm used to. What I really want to do is access content that the user already has on the device and add additional content for the user to take off the device. I had thought that's what the Content Hub was about, but... Our application security model has each application owning its own data, and sharing it with other applications. In other words, the ebook reader should be storing its files in its own directory, and offering them to other applications via the content hub. How does this work for content the user already has? The main point of Beru is to let the user read the epubs the user already owns. Can the user only copy the content over after they've chosen which reader to use? That seems backwards. Furthermore, what if I become evil (well, more evil than usual) and stop offering Beru's books to other apps? This locks users into using Beru (or forces them to wade through hidden directories to recover their content). This approach seems completely counter-intuitive to me. Is the rationale behind it documented somewhere? Michael Zanetti wrote: On Friday 25 October 2013 08:13:20 Marc Deslauriers wrote: 5 Oh, nice. What is the problem then why Robert's use cases don't work? One problem is that there's not "epub_files" policy. But even if there was, it doesn't solve the general case for fpub files, gpub files, etc. Michael Zanetti wrote: Wouldn't it make sense to have such a policy group for the "Documents" folder? that would probably be just what Robert needs, no? Given that eBooks (which I think is Robert'main use case) are quite big (in terms of distribution) nowadays I could even imagine adding a ~/Books folder with appropriate capabilities could make sense. What do you think? When I first submitted Beru, I asked for read and write access to ~/Books. This was denied. But again, this is a special-case solution to what seems to me to be a general problem. Thanks for the answers, Robert [1] rschroll.github.io/beru/ -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Accessing the user's files in a click app
Thank you, thank you, thank you! This is exactly the explanation I've been searching for, and I've felt stupid not being able to find it. Basically, the answer is to wait -- we're not seeing the full picture yet. A few more questions while we're waiting: 1) When we get to the point where apps get "pure" content, will there be some way for them to identify that content to request it again later? Beru has a list of books ordered by last reading date. Right now, the books are identified by their filename. Will there be a similar way when we just get a file handle, for example, that we can remember the content ID and ask for it again later? 2) Will all of this work if you're not running Unity? If I want Beru to work for other DEs, will I need two code paths, one to handle file system access, the other to handle content hub access? 3) While we're waiting for all of this to happen, does it make sense to give apps access to a "Content" directory in the user's home directory? There would be no guarantee of safety or persistence of data stored here, but I'm not counting on that for epub files in Beru. Potentially, this could be turned into a FUSE filesystem backed by the Content Hub once that's ready. Thanks again, Robert On Fri, Oct 25, 2013 at 3:01 PM, Jamie Strandboge wrote: On 10/25/2013 12:03 PM, Robert Schroll wrote: ... I'd prefer arbitrary locations, to allow the user to choose where to save their files and to allow for run-time localization. But if we restrict it to a give sub-directory at compile time, that wouldn't be the worst thing in the world. Perhaps I'm thinking too much about files because that's what I'm used to. What I really want to do is access content that the user already has on the device and add additional content for the user to take off the device. I had thought that's what the Content Hub was about, but... I think you are thinking too much about files. The design of Unity, scopes, content-hub, etc is more about content and less about file names and locations. I see several things here that will help you and others when they are implemented: * system file picker/dialog content provider for the content-hub. This is an unconfined process. Beru talks to content-hub, content-hub calls file picker, user chooses what to read, content-hub gives it to Beru. This might be a copy of the file or in the future a file handle. AIUI, the goal is to abstract those kinds of details away so app deverlopers don't have to worry about them * epub content provider for the content-hub. This is the equivalent of the gallery content provider for ebub files. Ie, it can be fancier/prettier/more useful then a traditional file dialog. Otherwise works just like above (fyi, you can see the content-hub in action by changing the background on the phone) * epub unity scope. This allows users to search for/within epub files and launch them directly in an app that can handle them * sharing to other apps implemented-- ie, you can share files from the file manager, browser, etc to Beru so it can read them The main problem right now is everything isn't written yet. Specifically, I think there is only one content provider: the gallery-app. More need to be written but once more are written multiple content providers need to be handled gracefully (ie, should the epub, fpub, gpub, or file picker be used?). Sharing to other apps isn't implemented nor is there an epub scope (both would also need a way for Beru to easily hook into them). Our application security model has each application owning its own data, and sharing it with other applications. In other words, the ebook reader should be storing its files in its own directory, and offering them to other applications via the content hub. How does this work for content the user already has? The main point of Beru is to let the user read the epubs the user already owns. Can the user only copy the content over after they've chosen which reader to use? That seems backwards. Furthermore, what if I become evil (well, more evil than usual) and stop offering Beru's books to other apps? This locks users into using Beru (or forces them to wade through hidden directories to recover their content). This approach seems completely counter-intuitive to me. Is the rationale behind it documented somewhere? That is indeed counter-intuitive and is not the intent of the design, however it would work at this moment in time. In the future the content-hub and scopes solve this-- the user puts content wherever he/she wants and then can search for it (scope), open it from within Beru (content-hub) or launch Beru from another app (sharing) and Beru doesn't know or care about file locations of content. I'm n
[Ubuntu-phone] NemoMobile FolderListModel
Hi all, In my epub reader app Beru, I use nemomobile.folderlistmodel to get a list of epub files on the device. This used to work on the devices, but a few weeks ago, users started reporting errors like: module "org.nemomobile.folderlistmodel" is not installed I notice that the music app, from which I originally stole this code, got rid of it some time ago. But the file manager still depends on this. So, is this still in the Ubuntu Touch images, or has it been removed? If it's there, what do I need to do to get access to it? If it's gone, how does the file manager manage to use it? And for the future, is there some way to note in the click package what (deb) packages it depends on, so users can get something better than runtime errors when those packages get removed from the images? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] NemoMobile FolderListModel
On Tue, Nov 19, 2013 at 12:50 AM, Sergio Schvezov wrote: On Mon, Nov 18, 2013 at 11:15 PM, Robert Schroll wrote: In my epub reader app Beru, I use nemomobile.folderlistmodel to get a list of epub files on the device. 2 If it's gone, how does the file manager manage to use it? 6 Thanks. Was there an easy way to tell this? I couldn't figure it out from the sources. On a related subject, I notice that Qt.labs.folderlistmodel is now listed in the API docs. Is this the same thing? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] [Click] Valid architecture values
Hi all, What are the valid values for the architecture key in a manifest file in a click package? I know it's "armhf" for arm devices. What about for x86 and x86_64 devices? (Yes, I know there's no need for anything other than arm right now. I have a little script that assembles click packages, and I figured I might as well get it right now.) Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Click] Valid architecture values
Thanks! On Mon, Nov 25, 2013 at 9:23 PM, Robert Park wrote: I'd like to see that... anything that makes packaging even easier than clicks can only be good ;-) Here's the script, updated with your input: https://github.com/rschroll/beru/blob/restricted/makeclick.sh It is very simple, but if anyone wants to use it as a jumping-off point for their own project, please go ahead. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] [SDK] Toolkit in PPA broken?
Hi all, I recently discovered that my app, built on 12.04 with the SDK PPA, is broken on devices. This is because of a version mismatch. The PPA for 13.10 and 14.04 is up to 0.1.46+14.04.20131129bzr873, but the version I have on my 12.04 box is 0.1.46+13.10.20130925.1bzr780. Looking at the build records [1], it seems that this package hasn't been building for 12.04, 12.10, and 13.04 for the last two months. Is this a known problem? Is it being worked on? When can we expect a fix? The specific problem for me is that the OptionSelector API changed somewhere between these two versions. Are these API changes announced somewhere, or do I just have to wait for runtime failures to find out? Thanks, Robert [1] https://launchpad.net/~ubuntu-sdk-team/+archive/ppa/+builds?build_text=ubuntu-ui-toolkit&build_state=depwait -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [SDK] Toolkit in PPA broken?
On Mon, Dec 2, 2013 at 1:06 PM, Tim Peeters wrote: API changes are listed in the CHANGES file, and announced on this mailing list. There have been changes to the OptionSelector API in r796 but they were reverted to avoid problems in r825, so neither r780 or r873 should be affected. If you tell me the specific problems/error messages that you have with the OptionSelector then I can try to help you fix it. I'm using a custom OptionSelectorDelegate [1] based on the stock version from early September (judging from my commit date). This works as expected on my 12.04 system stuck on r780; it just doesn't open on an up-to-date 13.10 system (on a VM), a recent emulator instance, or a device (don't know the version, since it isn't mine). Doing a diff between the versions of OptionSelector.qml, I suspect the problem is that listContainer.isExpanded has changed to listContainer.currentlyExpanded. The OptionSelector has also gained currentlyExpanded and itemHeight properties and an expansionCompleted signal, but I'd be surprised if these were causing problems. Presumably the fix will be as simple as rebasing my delegate on the updated stock OptionSelectorDelegate. I don't think the user base is large enough that I have to worry about supporting both APIs. This will be easier for me to do if my 12.04 system gets the update, but it wouldn't be impossible if not. At the risk of sounding insufferable, I'd like to point out that I was worried about these sorts of incompatibilities from when I was first pointed in this direction [2]. I don't know if it's worth reviving that discussion or not. I think I need to pay a bit more attention to the CHANGES file, since I missed both the changes in r796 and their reversion in r825. Thanks, Robert [1] https://github.com/rschroll/beru/blob/restricted/ui/components/StylableOptionSelectorDelegate.qml [2] https://lists.launchpad.net/ubuntu-phone/msg04044.html -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [SDK] Toolkit in PPA broken?
On Tue, Dec 3, 2013 at 11:34 AM, Tim Peeters wrote: So the two main issues here are: 1. UITK packages for 12.04 are not up-to-date (and perhaps not compatible with newer versions) I just managed to find an existing bug about this: https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1237045. It was reported almost two months ago, which fits with the timing I've noticed. 2. In order to change the looks of an option selector delegate, you need to copy the full OptionSelectorDelegate code Submitted here: https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1257529. Thanks for the help! Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [SDK] Toolkit in PPA broken?
On Tue, Dec 3, 2013 at 8:50 PM, Robert Schroll wrote: On Tue, Dec 3, 2013 at 11:34 AM, Tim Peeters wrote: So the two main issues here are: 1. UITK packages for 12.04 are not up-to-date (and perhaps not compatible with newer versions) I just managed to find an existing bug about this: https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1237045. It was reported almost two months ago, which fits with the timing I've noticed. This bug has been marked as fixed (Thumbnailer) and invalid (Ubuntu UI Toolkit). But the ubuntu-sdk-team PPA hasn't gotten the updated thumbnailer, so neither package has built yet. Is there someone I should ping about this? 2. In order to change the looks of an option selector delegate, you need to copy the full OptionSelectorDelegate code Submitted here: https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1257529. Should I CC someone else on this bug? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] U1dB Documentation and Advice
On Sat, Jan 4, 2014 at 11:45 AM, Nekhelesh Ramananthan wrote: This email is concerning persistent storage options for an Ubuntu Touch application developer. I hope this doesn't turn out to be a big email, but in any case here is a short summary of this email. What is the future of U1dB-Qt? Currently the documentation and tutorials for using U1dB is quite limited and lacking in several area. Would it be possible to engage with the original U1dB-Qt developers to improve this? [...] I did find some links to U1dB-Qt resources [1],[2],[3] which help a little but is still missing a big chunk of features. If it should help, I wrote up a post on using U1DB procedurally back when I was trying to learn it: http://rschroll.github.io/beru/2013/09/04/using-u1db-procedurally.html. Anyone writing up better documentation is welcome to use what I wrote, or perhaps just the correct parts thereof. On Sat, Jan 4, 2014 at 3:34 PM, David Marceau wrote: From the sounds of it you don't want to be limited to using *U1dB *and *Qt LocalStorage*. No problem. *You can pick your favourite C++ based persistent solution from the standard Ubuntu repositories but targetted for Ubuntu Touch. I don't mean to get too distracted on this tangent, but this is a negative, not a positive, for the developer trying to learn the Ubuntu SDK. (See the "Paradox of Choice".) One of the things Ubuntu does very well is make opinionated choices on our behalf (even when they're wrong!), and that should be the case here as well. While it's nice that the developer who wants to use mongodb can, that doesn't help the developer who just wants persistent storage. For them, there should be one, and only one, obvious way to do it. U1DB seems positioned to be that way, but its documentation is wanting. This is certainly a problem. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] Click to deb conversion?
Hi all, From the little that I understand of it, click packages are just simplified deb packages. So, is it possible to convert a click package into a deb? If so, how? I ask because I'm working to make my Ubuntu Touch app usable on the desktop as well as the phone. While you can install click packages on the desktop, most people are probably more familiar with debs. Since I'm already set up to build click packages, I thought I'd see if it's possible to convert those into debs before trying to learn how to build debs directly. Thanks in advance, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Click to deb conversion?
On Sun, Jan 5, 2014 at 12:50 PM, Dimitri John Ledkov wrote: Whilst the implementation detail is correct that both are tar archives inside of an ar. Policy wise it's not possible to simply convert a .click into .deb and expect it to be properly functional - it will not install correctly, executables will not be available on the path, .desktop file will not get indexed, one will not be able to find that app from the dash, etc... Good to know. I was hoping you could just add a debian directory to the click package and have everything work, but that's not the case. On Sun, Jan 5, 2014 at 6:23 AM, Fabio Colella wrote: This -> http://forum.xda-developers.com/showthread.php?t=2265480 can help you! ;) Perfect! Thanks a bunch for that link. I haven't had a chance to try that yet, but it looks well explained. Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Click to deb conversion?
On Mon, Jan 6, 2014 at 1:02 AM, Robert Schroll wrote: On Sun, Jan 5, 2014 at 6:23 AM, Fabio Colella wrote: This -> http://forum.xda-developers.com/showthread.php?t=2265480 can help you! ;) Perfect! Thanks a bunch for that link. I haven't had a chance to try that yet, but it looks well explained. I've managed to get everything set up to build debs with debuild. But before I try uploading the source to a PPA, I though I should try building with pbuilder to make sure I have the Build-Depends: correct. Unfortunately, this fails due to unmet dependencies. As best I can tell from the output, this is due to the various qt dependencies, though I'm not sure exactly what the problem is. Here's the output: http://paste.ubuntu.com/6716591/, and here's the control file: http://paste.ubuntu.com/6716688/. Does this make sense to any of you? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Using a zip file?
In case anyone out there actually does have to handle zip archives: The QuaZIP project (http://quazip.sourceforge.net/) provides a Qt interface to zip and unzip archives. This is what I'm currently using in my epub reader. (Epubs are just glorified zip archives.) There are also several pure-Javascript zip libraries out there. However, they are targeted for the browser JS environment, not the more limited QML JS environment. I did get JSZip (http://stuk.github.io/jszip/) working with QML before (details: http://rschroll.github.io/beru/2013/08/16/javascript-libraries.html), but it was slow and occasionally buggy. Robert On Wed, Feb 5, 2014 at 1:07 PM, Sam Bull wrote: On mer, 2014-02-05 at 17:45 +0100, Michael Zanetti wrote: But actually this begs the question: Why do you want to package up your files into a .zip? click packages compress their data automatically. So while the files would be unpackaged in the source repository, the actual resulting click package should be smaller without you compressing them yourself. Alright, I'll have a go at using it uncompressed, and see what size the click package ends up as, if that's the case, it should be fine without handling compressed files at all. -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [SDK] Expandable items
If you have a compatible license (and I have no idea what that means in terms of QML components, but GPLv3 is obviously OK), you can just include those new components in your project. You'd probably need to adjust the imports a bit and rename the data attribute back to children, but that's not too much work. I'm doing a similar thing in my app, Beru. I need access to the internals of the OptionSelector, and I've learned the hard way that I can't count on them not changing in the SDK. So I've copied a version into my app to use in place of the one provided by the SDK. Inelegant, perhaps, but it works. Robert On Mon, Feb 10, 2014 at 11:17 AM, Sam Bull wrote: Does nobody have any idea? There's nothing more we can develop on our app until this is merged. On mer, 2014-02-05 at 15:56 +, Sam Bull wrote: Are expandable items going to be available soon? It looks like it's been ready to merge for weeks: https://code.launchpad.net/~mzanetti/ubuntu-ui-toolkit/expanding-listitem/+merge/195602 Sorry to pester, but this has been holding up completion of our app for a few months now. -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] Fwd: Re: [SDK] Expandable items
Missed the list with my first reply. On Wed, Feb 12, 2014 at 4:50 AM, Christian Dywan wrote: Be sure to file a bug https://bugs.launchpad.net/ubuntu-ui-toolkit/+filebug with your use case(s), what API is missing. So the OptionSelector can be improved for everyone using the SDK. The specific problem, with patch: https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1220533 The general problem: https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1257529 Better yet, post about it with the bug URL here afterwards, so more people are aware of it. https://lists.launchpad.net/ubuntu-phone/msg04021.html https://lists.launchpad.net/ubuntu-phone/msg05460.html I didn't mean to suggest that Sam not report the problem, just that he could continue development while waiting for that bug to be fixed. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] New SDK and UI Toolkit releases
On Fri, Feb 14, 2014 at 3:42 PM, Zoltán Balogh wrote: 2. The UITK is back on Precise We had a long pending problem with releasing the 12.04 backport of the UI Toolkit due to a cluster of difficultes. We had to backport several dependencies and tackle with the fact hat the UI Toolkit is a fast moving on the edge project where we use the latest bits from the 14.04 development. Given that, keeping the 12.04 edition of the UI Toolkit is at least challenging. Several of us have found that we are unable to complete the upgrade to the current packages on 12.04. The qtdeclarative5-ubuntu-ui-toolkit-plugin package is stuck at 0.1.46+13.10.20130925.1bzr780precise0, with 0.1.46+14.04.20131216-0ubuntu1-0precise1 listed as available. However, installing this would require the removal of cordova-ubuntu-2.8, cordova-ubuntu-2.8-examples, qtcreator-plugin-ubuntu, qtdeclarative5-cordova-2.8-plugin, qtdeclarative5-ubuntu-ui-extras-browser-plugin, ubuntu-html5-container, ubuntu-sdk, and ubuntu-ui-toolkit-examples. I'm assuming that I don't want to do that. (If I do, please say so.) This leaves the SDK in a somewhat broken state. It seems that any app that uses a Toolbar produces the warning, Connections: Cannot assign to non-existent property "onSelectedIndexChanged" None of the headers shows up, either. I've also gotten several seg faults while tracking down other issues, but I don't know where the blame lies for that. Let me know if there's more info I can give you, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] New SDK and UI Toolkit releases
On Tue, Feb 18, 2014 at 7:04 AM, Zoltán Balogh wrote: I pushed an update with the right version number and tweaked dependencies. Please try now and see what it does. It should not remove any packages :) and if it offers that please do not say yes. It upgrades just fine, but when I try to run anything, I get the same errors reported by Julien: file:///home/rschroll/touch/beru/ui/main.qml:15 Type MainView unavailable file:///usr/lib/x86_64-linux-gnu/qt5/qml/Ubuntu/Components/MainView.qml:18 module "Ubuntu.Unity.Action" version 1.1 is not installed Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] New SDK and UI Toolkit releases
On Tue, Feb 18, 2014 at 2:54 PM, Zoltán Balogh wrote: I have just pushed the 0.1.46+14.04.20140212-0ubuntu1-0precise3 to the SDK PPA This release reverts an MR[1] exclusively for 12.04 desktop installation. That fixes things for me! Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] [HTML5] Local storage options for HTML5 apps
Hi all, I'm trying to figure out what options I have for storing data in an HTML app. Based on testing in 13.10, it appears that Web Storage and Web SQL are supported, while IndexedDB are FileSystem are not. Is this correct? Is this situation expected to persist, or will Web SQL be phased out and IndexedDB and/or FileSystem phased in? Will U1DB bindings appear? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [HTML5] Local storage options for HTML5 apps
Thanks for the info. I'll use Web Storage for now, and look into U1DB when it comes out. Robert On Wed, Mar 12, 2014 at 9:48 PM, Alexandre Abreu wrote: Hi Robert, On Wed, Mar 12, 2014 at 4:16 PM, Robert Schroll wrote: Hi all, I'm trying to figure out what options I have for storing data in an HTML app. Based on testing in 13.10, it appears that Web Storage and Web SQL are supported, while IndexedDB are FileSystem are not. Is this correct? correct, at the moment, Is this situation expected to persist, or will Web SQL be phased out and IndexedDB and/or FileSystem phased in? Yes, no ETA yet, but we expect this to happen in the short/midterm, Will U1DB bindings appear? yes, definitely yes, it is part of the bindings that we plan to support soon, -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] Triggering on-screen keyboard with HTML5 SDK
Hi all, I'm working on a crossword client using the HTML5 SDK. Obviously, being able to enter letters is important, so the on-screen keyboard should be visible whenever the user is working on a puzzle. But I can't figure out how to trigger it to show up. Is there some Javascript to call or something in Cordova? On Google+, Adnane Belmadiaf suggested making a text input and setting the focus on that. In testing with the browser in an emulator, this didn't work; the input would have focus, but the OSK would not appear. For extra credit, - Do I have to do anything to prevent the OSK when there's a physical keyboard? - Can I turn on caps lock for the OSK? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Triggering on-screen keyboard with HTML5 SDK
On Mon, Mar 24, 2014 at 1:13 PM, Alexandre Abreu wrote: While a programmatic solution might be handy in the meantime, can you have a text field that can be the receiver for the user's input & with which the user could trigger the keyboard to popup? you could track down the onchange events & pick up the currently entered "word"? That might work, but it'd be mildly annoying both to program and to use. I also don't know how to detect whether the user has a physical keyboard, so we'd be stuck with this for all devices. If I'm going to go through the work of building a work around, I wonder if it's better to write my own QML wrapper, instead of relying on ubuntu-html5-app-launcher, and create my own javascript hooks to show and hide the keyboard. BTW, there is a bug about this now: https://bugs.launchpad.net/ubuntu-html5-theme/+bug/1296462 Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Triggering on-screen keyboard with HTML5 SDK
On Tue, Mar 25, 2014 at 10:04 AM, Alexandre Abreu wrote: Hi Robert, The gist of the issue is OSK not popping up when js sets the focus in a text field. I'll agree that fixing this would be enough for me to get by with. But... I dont think that having a way to programatically popup the OSK w/o a given "target" (text input field, etc.) is sound as a use case for a given API. It opens the gate for quite a few edge cases and adds quite a bit of complexity to have a reliable interface. It does not even work in QML afaik (you need a text field and can only hide() the inputMethod). I admit I haven't though about this much, but I'm not sure I see the problem. Both HTML and QML have well established rules for how keyboard focus behaves when the focus isn't in a text entry, so that's not a problem. I guess there's a worry about when a keyboard should hide after an app requests it be opened. But a "you break it, you bought it" philosophy seems appropriate: once an app explicitly shows or hides they keyboard, the automatic bit turns off and the app is responsible for handling the keyboard state in the future. This isn't such an esoteric feature. Off the top of my head, I'm aware of another app (Word Chain) that needs alphabetic input without a text entry. It solves the problem by implementing its own OSK. The alternative to having the SDK provide this is to have every app implement it's own solution, hack, or work-around. This means that each will behave slightly differently, which can't be desired. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Triggering on-screen keyboard with HTML5 SDK
On Tue, Mar 25, 2014 at 5:10 PM, Robert Schroll wrote: On Tue, Mar 25, 2014 at 10:04 AM, Alexandre Abreu wrote: Hi Robert, The gist of the issue is OSK not popping up when js sets the focus in a text field. I'll agree that fixing this would be enough for me to get by with. I've played around with this a bit more, and it seems that the OSK fails to appear only if you try to activate it before the user has interacted with the WebView. While this is a problem in general, I escape as I only want the OSK for an inner page of the application. So I've added a dummy text input to my HTML to act as my focus trigger. It doesn't work with display: none, but I can make it a pixel big and toss it somewhere out of the way. (If I make it zero pixels in size, the web process crashes.) But -- pressing OSK buttons while this has focus doesn't seem to generate any key events, either for the document in general or for the input specifically! Is this a known problem? Is there a work around? Or is there a different set of events to which I should be listening? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Triggering on-screen keyboard with HTML5 SDK
Here's a test case that demonstrates both of these issues: http://paste.ubuntu.com/7158427/ Clicking on the "Click here" text focuses the text input. This works in the browser, the HTML5 launcher, and the device (or, the emulator, at least). In this last case, the OSK appears as desired. But the fact that there's a click is important. In a previous test case, I made the focus happen after 5 seconds. If you had clicked anywhere, you'd get the OSK; if not, you'd get focus without the keyboard. Once focus is on the input, the keydown, keyup, and keypress events will trigger text to be added to the three bottom lines. This works in the browser and in the HTML5 launcher, but it does not work on the device (er, emulator). Robert On Wed, Mar 26, 2014 at 2:22 PM, Alexandre Abreu wrote: Hi Robert, On Wed, Mar 26, 2014 at 1:56 PM, Robert Schroll wrote: On Tue, Mar 25, 2014 at 5:10 PM, Robert Schroll wrote: On Tue, Mar 25, 2014 at 10:04 AM, Alexandre Abreu wrote: Hi Robert, The gist of the issue is OSK not popping up when js sets the focus in a text field. I'll agree that fixing this would be enough for me to get by with. I've played around with this a bit more, and it seems that the OSK fails to appear only if you try to activate it before the user has interacted with the WebView. While this is a problem in general, I escape as I only want the OSK for an inner page of the application. afaik (and I have to further validate it), it fails to appear when js has set the focus to a given (text) element and the user clicks on the text element, no OSK pops up unless the user manually forces the focus elsewhere and back to the text field, So I've added a dummy text input to my HTML to act as my focus trigger. It doesn't work with display: none, but I can make it a pixel big and toss it somewhere out of the way. (If I make it zero pixels in size, the web process crashes.) But -- pressing OSK buttons while this has focus doesn't seem to generate any key events, either for the document in general or for the input specifically! Is this a known problem? Is there a work around? Or is there a different set of events to which I should be listening? I'll give it a try, but it should work; are you sure the events are not caught by some element? -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Triggering on-screen keyboard with HTML5 SDK
On Wed, Mar 26, 2014 at 3:48 PM, Alexandre Abreu
wrote:
On Wed, Mar 26, 2014 at 3:04 PM, Robert Schroll
wrote:
Once focus is on the input, the keydown, keyup, and keypress events
will trigger text to be added to the three bottom lines. This works
in the browser and in the HTML5 launcher, but it does not work on
the device (er, emulator).
they are not supported yes w/ the OSK, but you can still make it work
by trapping the 'input' event:
document.querySelector("input").addEventListener('input', function(e)
{ console.log(e.target.value); });
Thanks. That may be the work-around I need to make my original
work-around work.
Is there a bug about this I should be watching?
Thanks,
Robert
--
Mailing list: https://launchpad.net/~ubuntu-phone
Post to : ubuntu-phone@lists.launchpad.net
Unsubscribe : https://launchpad.net/~ubuntu-phone
More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Triggering on-screen keyboard with HTML5 SDK
On Mon, Mar 31, 2014 at 9:42 AM, Alexandre Abreu wrote: On Wed, Mar 26, 2014 at 3:59 PM, Robert Schroll wrote: Is there a bug about this I should be watching? not at this point, Should I submit a bug report? If so, against which project? Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Deploy QMake (*.pro) C++/QML project?
On Mon, Mar 31, 2014 at 8:14 PM, José Expósito wrote: Does someone know if it is possible to generate a click package just by preparing an output directory like that? I have checked the click command help but it is not very extensive... Does someone know where could I get more information? As I understand it (not well at all), the 'click build' command just takes the contents of a directory you specify and wraps everything up in the correct packaging. As long as you have a manifest.json file in that directory, everything should be fine. When I wrote Beru, I ended up writing my own script to generate the click packages. It copies all the files that are needed into a temporary directory and then runs 'click build' on that. (It also tar-and-feathers the directory for my future reference, but that's extraneous. And it does everything twice, since I build two different versions each time, because reasons.) Here's the code, if you're interested: https://github.com/rschroll/beru/blob/master/makeclick.sh I make no claim that this is an intelligent way to do things, but it's worked so far. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Triggering on-screen keyboard with HTML5 SDK
I've got a kludge in place right now that seems to properly trigger the OSK when needed without messing things up on the desktop. The ugly details are here: https://github.com/rschroll/crosswords/commit/e714295baff5232b9bc6bcd7fcb621f4e09a8ca5 But the RuntimeApi looks like it could be very useful. Thanks for the heads-up, Robert On Thu, Apr 3, 2014 at 10:33 AM, Alexandre Abreu wrote: I thouhgt you might be interested, On Tue, Mar 25, 2014 at 9:43 AM, Robert Schroll wrote: That might work, but it'd be mildly annoying both to program and to use. I also don't know how to detect whether the user has a physical keyboard, ... The runtime-api javascript binding for HTML5 apps has now been released (along w/ the new content hub one) and offers something for this http://developer.ubuntu.com/api/html5/sdk-14.04/RuntimeApi.RuntimeApi/ geInputMethodName -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Landing team 13.08.15
Thanks for the explanation! Just to make sure I've got it: devel isn't actually dangerous right now -- it's just not going to get any love until devel-proposed manages to stabilize. rc-proposed is available, but under the "If it breaks, you get both halves" warranty. I'm speaking from a position of complete ignorance here, but: Is there that much of a chance that an image that works on mako will fail on manta or flo? I would assume that most of the driver issues have been worked out, and most of the updates for these machines would be "pure" software. If so, what about promoting images for manta and flo once they've passed QA on mako? Obviously you can't control for unknown unknowns, but it seems that this should be significantly safer than a -proposed channel. Obviously, we'd make it clear that these images aren't guaranteed in the same way those for mako are, and maybe this should only happen for rc, not stable, but I think it would be a nice intermediate between the working-but-old devel and new-but-maybe-broken rc-proposed. Or maybe I'm completely wrong. Thanks, Robert On Fri, Aug 14, 2015 at 4:26 AM, Łukasz 'sil2100' Zemczak wrote: Hello, W dniu 14.08.2015 o 03:53, Robert Schroll pisze: On Thu, Aug 13, 2015 at 2:14 PM, Łukasz 'sil2100' Zemczak wrote: Please DO NOT use the devel/devel-proposed channels for your phones if not needed. Perhaps then we should stop recommending devel for those wishing to "track the latest development" on developer.ubunutu.com [1]? Although devel hasn't been updated in four months, so maybe that's supposed to be self-evident. Although what is written there is partially true, I will try to update it with more details so it's evident how the real situation looks like. This means that all the feature development actually happens on the vivid-based channels, e.g. stable, rc and rc-proposed. Neither stable or rc have images for manta or flo. Should we be on rc-proposed, or would it be better to stay on the (stale) devel channel? The concept was that flo and manta are supported for promotion into the devel channels, while the rc set only focuses on the devices we have time to support with full force. We didn't have time or resources to do a devel promotion for a long time and now - and seeing the current state of wily touch, I would suppose we won't have one in the nearest time. The main issue is of course time. With really limited resources we do not have the man-power for taking care of so many devices. I think we might need to re-visit which channels should be recommended and supported for the flo and manta devices anyway. What would I propose? I wouldn't want to give wrong hints here as I have no idea what's the state of flo on rc-proposed, but that channel at least gets developer focus. This developer focus is mainly for krillin arale and mako, but still better than nothing. This might sound a bit strong but we simply don't want users to get the wrong impression by using the wrong channels. When I asked four months ago about flo, I was specifically told to use devel. Was that wrong at the time, or did things change? Was there an announcement of the change? If this is such a concern, the documentation [1] really ought to be updated. The situation with flo and manta is a bit more complicated. The warning I sent out in my landing e-mail was more about the many stories of users switching their mako, arale and krillin devices to devel-proposed to get 'the latest of the latest', wheres this in this channel it's not entirely true. All in all - the flo and manta situations need to be resolved. I will bring it up with the product team and see what we can do about the current situation. Robert [1] https://developer.ubuntu.com/en/start/ubuntu-for-devices/image-channels/, section "Channel selection guide" Cheers, -- Łukasz 'sil2100' Zemczak Foundations Team lukasz.zemc...@canonical.com www.canonical.com -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Landing team 13.08.15
On Fri, Aug 14, 2015 at 12:27 PM, Łukasz 'sil2100' Zemczak wrote: Well... that's the *theory*, but in practice it's sadly a bit more problematic ;) As we say in physics: in theory there's no difference between theory and practice; in practice, there is. That being said, we discussed all this slightly in our landing team meeting and also partially with the product team today. There is no definite decision yet, but what we'll think of doing is actually including at least flo to the list of community devices we'll try promoting on every main update. We won't be doing all the regression tests on it, but do something similar to what we do for mako - make sure that it passes sanity tests. So I suppose this way flo users would be able to use the rc/stable channels for their convenience. Do images in the rc channel have equivalents in the rc-proposed channel? If so, someone (and I'll volunteer to be someone) could run the images that would have been promoted on flo and see how they would have worked. If most of them mostly work, we could be someone confident about this plan. If there are significant problems, then it might not be such a great idea. Obviously, past performance is no guarantee of future results, but it might help with the decision. Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Landing team 25.08.15
On Tue, Aug 25, 2015 at 6:13 PM, Łukasz 'sil2100' Zemczak wrote: Once we get a green light regarding the testing we will proceed with the promotion plan as mentioned yesterday: mako, flo and arale released tomorrow, krillin and vegetahd next week. You almost slipped this past me, but I notice that flo is now included in the rc and stable channels. Is this part of a test, or is the intention for flo to continue to see rc and stable updates? Either way, I'm glad to see it! Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Document privacy
On Mon, Sep 21, 2015 at 11:38 AM, Jamie Strandboge wrote: I can say that the upcoming storage framework specification deals with (vastly) improving sharing as well. Is this specification available somewhere for public comment? I've been frustrated by the content hub many times, so I'd like to help ensure its successor will solve these problems. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] updates through emulator
On Thu, Oct 22, 2015 at 1:55 PM, Nicholas Skaggs wrote: AFAIK, yes, you cannot update the emulator. That said, I never tried toggling the update through the command line, or attempting hackery to force it through. According to the wiki (https://wiki.ubuntu.com/Touch/Emulator): To update to a latest image; it is safe to create over the previous instance (upgrade option comming soon). As I recall, it's been coming soon for at least the last year and a half, so don't get too excited. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] QML ContentStore example?
On Fri, Oct 23, 2015 at 8:38 AM, Jani Monoses wrote: I have searched in vain for an example of a minimal use of ContentStore - such that imported files do not by default end up under ~/.cache where they will be deleted from. This may or may not be your problem but: There is a bug wherein a file imported into a ContentStore with ContentType.All is sent to ~/.cache/ instead of of ~/.local/share/. https://bugs.launchpad.net/ubuntu/+source/content-hub/+bug/1429691 If that's not the problem, let me know. I have an example using the ContentStore somewhere around here Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] QML ContentStore example?
On Fri, Oct 23, 2015 at 1:06 PM, Michael Zanetti wrote: If that's not the problem, let me know. I have an example using the ContentStore somewhere around here I'd be interested in that. Can you share please? I've had a git repository sitting around with some content hub examples for a while. I keep meaning to finish them off, document them, and share them, but I never got around to it. So I'm going to skip the first two steps and go straight to sharing. An important detail that the official docs elide over is that there are two types of transfers. I've dubbed them "push" transfers, initiated by the app sending the content, and "pull" transfers, initiated by the receiving app. To try to produce minimal examples, I've written separate examples for each case. At the moment, I only have apps that import content. The pull-import branch [1] requests content from either the default picture source or from a picture source of your choosing. The push-import branch [2] can receive pictures exported from any source. An important limitation of the ContentStore system is that it only works for pull transfers [3]. Therefore this system is only used in the pull-import branch. I added it in this commit [4]. As you'll see, it's pretty easy to use. You just need to create a ContentStore object and pass it as an argument to the peer.request() calls. On Fri, Oct 23, 2015 at 1:40 PM, Jani Monoses wrote: I had even looked at beru sources on github but saw you do not actually use the ContentStore but build upon item.move(). I do this because of the failure of the ContentStore system to work with push transfers. I have to use item.move() in this case. And once I've written that code, it works just fine for the pull transfer case, so I see no reason to switch that to using the ContentStore. And that's my general recommendation. If you have any plans to support push imports, don't bother with ContentStore. Robert [1] https://github.com/rschroll/chtest/tree/pull-import [2] https://github.com/rschroll/chtest/tree/push-import [3] https://bugs.launchpad.net/ubuntu/+source/content-hub/+bug/1430075 [4] https://github.com/rschroll/chtest/commit/d8796be6f25ed0e4d904b0fed38c1dcc487f5265 -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] JavaScript bindings for Scopes
On Tue, Nov 3, 2015 at 7:47 AM, Marcus Tomlinson wrote: We’d like to announce the release of the JavaScript bindings for Scopes. Great news! I've long wondered why I had to muck about in C++ to push JSON data from one place to another. No more! I'll be trying this out soon. Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Spying features in Ubuntu
On Thu, Nov 5, 2015 at 6:55 PM, Peter Bittner wrote: I've accidentally come across a Richard Stallman interview on Ubuntu. He's mainly concerned about the spying feature of Ubuntu's dash when you do a search on your device. RMS is displaying his usual flair for the dramatic. In Unity 7, searches in the Dash go to all scopes by default. If you have remote scopes installed (and there are some installed by default), your searches will go out on the network. In earlier versions, this was not well advertised and difficult to disable. Since then, I believe there is more notification of this during install and a single button to turn off all remote searches by default. It's not set to "Off" by default, and that's not good enough for RMS. Will the spying feature then also be on all phones one day? In Unity 8 [1], the phone and future-desktop environment, scopes work differently. Searches go to a single scope at a time. Some scopes will return local results, some will return remote results, and some will aggregate results from other scopes (potentially both local and remote). I don't know what the default scope will be, but I would anticipate an aggregate scope sourcing some remote content. Since the user can easily control which scope is being used, this set up strikes me as much more reasonable than the Unity 7 model. Whether it's good enough for RMS is another matter. I hope these details are right, Robert [1] Yes, a single project has produced two incompatible versions of a package named "Unity". Let us take a minute to appreciate the irony. -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Syncthing
On Thu, Dec 17, 2015 at 11:47 AM, Neil McPhail wrote: One of the requirements for the Open Store submission is that the app be open source, with a link to the source used to build the app. Looking at the build documentation, this looks easier than I thought. They provide a docker build image which can build all architectures from git automatically, so it should be fairly easy to get an ARM build. The point of this restriction is to give Michael Zanetti a chance to do a security review of the code. I don't know for sure, but he may be willing to accept a binary provided by a reputable open-source project like Syncthing. I'm sure he'd insist on being able to verify that binary; I'd suggest making the downloading of it part of your build process. But do get in touch with Michael before doing a lot of work in any direction on this. He's a nice guy, and I'm sure he'd be happy to figure out a solution with you. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Challenge] Ebook reader
On Mon, Feb 1, 2016 at 10:33 AM, Neil McPhail wrote: The app is an ebook reader, and should be capable of displaying books in a variety of formats (epub, txt, mobi etc). It should also have a text-to-speech engine, and the displayed text should follow the text-to-speech engine in real time. These things are relatively trivial to implement. However, some things are not so easy to achieve on Ubuntu: As someone who actually has built an ebook reader for the Ubuntu phone [1], I figured I might have some insight here. At the moment, Beru supports Epub, PDF, and CBZ files, though not particularly well in the latter two cases. TXT file support would be pretty easy, actually, but I didn't know that was actually a thing. (Is that actually a thing?) As for text-to-speech, my idea of trivial is somewhat different than yours. 1) It should be possible to add books from internet sites such as Project gutenberg Early versions of Beru actually had a built-in web browser to allow this. Advances in the content hub (and churn and bugs in the web browser and downloader components) convinced me to give this up. However, to get any support for ebooks in the content hub, I had to submit a bug against the content hub itself, which obviously doesn't scale. And once that was fixed, I still have to wait for all other apps to add support for this new content hub. I use the present tense purposefully there. So, last I checked, the web browser would let you open Epub and PDF files, but not CBZ files, with Beru. The file manager would only let you open PDF files with Beru. Unless you triggered it via the "Import from Content Hub" option within Beru, in which case it'll pass any type of file ever. Clear? The root cause of this insanity is that the content hub doesn't actually say anything about what the content types are. Each application is responsible for determining file type, and therefore must be updated every time anyone anywhere adds a new file type. This problem has been solved on the desktop since before Ubuntu was a gleam in the BDFL's eye, so I'm not sure why we're having it here. The obvious (to me) solution is using MIME types to define content types, but that bug [2] just had it's 20-month anniversary. I don't think the documentation on Content Hub is clear or comprehensive enough, and I see that view echoed on the IRC channels. I agree, and wrote a rather extensive bug report about some of the problems I saw [3]. It's been "In progress" for 10 months now. 2) It should be possible to add books via USB cable from computer This was the first feature past basic display that I wanted, and it's still not possible. Every time I mention this, "Yeah, that'd be useful, but you can't do it yet. Maybe when the content hub ...?" A relevant bug [4]. This has irritated me to the extent that there is a version of Beru in the Open Store that actually makes this work [5]. It lets the user specify which directory Beru should watch for new books. Once you've connected your device over USB, you can just drag your books over and things just work. Of course, this is a heinous security vulnerability. There's an important bit for user freedom that I think you've missed: 2.5) It should be possible to open the books you've added with any ebook reader. This isn't possible right now, since the books Beru can read are kept in its own silo, where another reader can't get at them. If Beru offered an export option, things wouldn't be so bad, but I'm an evil monopolist trying to exploit first-mover and lock-in advantages. Mwa-ha-ha-ha-ha! 3) It should be possible to add books via Bluetooth from another device The ebook reader shouldn't know anything about Bluetooth, IMO. Bluetooth transfer would be nice, but it should look just like USB transfer -- hey there are some new books in this directory I've been told to watch! Cool! I'll open them if you'd like me to. 4) It should be possible to add books over WiFi from computer Whoa now! Don't get greedy. 5) It should be possible to store or add books from an SD card Hahahahahahahaha... 9) It should be possible to sync the library, reading positions etc to a central source I purposefully architected Beru so that the reading positions and other settings were stored in their own U1DB that could be synchronized over Ubuntu One. And then Ubuntu One shut down. Frankly, though, the solution to me would seem to be to keep configuration in a central place, so you can use Dropbox, etc, to synchronize your settings without the app needing to be aware of this. The XDG directory structure makes this easy. The Ubuntu confinement model makes this nearly impossible. 10) It should be possible to cut and paste from books in the app Ya know, I don't know if this will work in Beru. I suspect not, because any touch triggers a link or a page change. Part of my reasoning behind working on an ebo
Re: [Ubuntu-phone] [Challenge] Ebook reader
On Mon, Feb 1, 2016 at 5:07 PM, Rodney Dawes wrote: Ubuntu One is not shut down. What was shut down was the file/music sync/streaming/storage/store services. U1DB is still up AFAIK, and Ubuntu One is still the SSO account infrastructure for Ubuntu. I was under the impression, perhaps mis-formed, that Ubuntu One (the database) would have allowed database synchronizing via Ubuntu One (the file sync service), which you signed into with your Ubuntu One (the authentication service) login. If that's not the case, the confusion is your* own damn fault for giving them all the same name. Also, anyone can run a U1DB server and you could sync the db in your app to any U1DB server that supports sync. Are there any public U1DB servers out there? If not, this is about as useful to most people as being able to compile your own kernel. Robert * referring to Canonical as a whole here, not any individual. -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Challenge] Ebook reader
On Mon, Feb 1, 2016 at 6:04 PM, Neil McPhail wrote: It seems that the most useful core apps have additional apparmor blessings which are not available to ordinary devs using the default store, This is perhaps the most frustrating part of the whole experience. Beru is not fundamentally different than the document viewer or the gallery. These apps faced the same difficulties in sharing data that I did with Beru. But instead of figuring out how to solve these problems in a general, extensible way, we just special-cased them and moved on. These apps could have provided valuable dog-fooding experiences and useful reference code. Instead, they're red herrings to confuse developers new to the platform. And yes, I know that these restrictions are only in the default store and we could yadda yadda freedom GPL copyleft. But I think we all know that this isn't *really* a solution. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Challenge] Ebook reader
On Mon, Feb 1, 2016 at 10:06 PM, Rodney Dawes wrote: It was always clearly stated that the U1DB server was still up, and that it was not done via file synchronization. I'm afraid I completely missed this point up til now. I looked back at the announcement of the shutdown of Ubuntu One (not to be confused with Ubuntu One!) [1], and the only only reference to U1DB I found was in the second to last paragraph, which says, The shutdown will not affect ... the backend U1DB database service. Even after your explanation, I'm still not quite sure how to parse that. I'm pretty sure the Qt U1DB implementation that exists on the phone was also written mostly after the Ubuntu One file sync services were taken offline, or at least announced that the service would be taken offline. Would be a bit odd to continue promoting an API for a service that was dead, don't you think? :) All the documentation I've seen for U1DB has focused on local use. I assumed that the network syncing was being put on the back burner. That seemed more sensible than giving the same name to three different components that had nothing to do with each other. There is u1db.ubuntu.com of course. $ nslookup u1db.ubuntu.com Server: 8.8.8.8 Address:8.8.8.8#53 ** server can't find u1db.ubuntu.com: NXDOMAIN I'm assuming you've mistyped the domain, but Googling "U1DB server" isn't giving me anything useful. How are people supposed to find out about this? I thought there was a blog post somewhere about how to sync a U1DB with Ubuntu One using the account on the phone, in QML, but I can't seem to find it now. Maybe Christian (kalikiana) can help clarify that when he comes back on-line. I'd love to see this, if anyone can find it. Thanks, Robert [1] http://blog.canonical.com/2014/04/02/shutting-down-ubuntu-one-file-services/ -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Challenge] Ebook reader
On Tue, Feb 2, 2016 at 4:55 AM, Christian Dywan wrote: The sync branch didn't make it past proof of concept - see https://code.launchpad.net/~uonedb-qt/u1db-qt/syncWithU1 - it basically works but the code needed cleaning up when my time then was increasingly going into UI Toolkit work, and the aforementioned design problems are still unsolved. I'd be very happy to see this work being picked up if there's enough interest, especially with concrete use cases to solve, as looking back the original API was not written in the best possible way, especially the Synchronizer. Thanks, Christian! As I said, my intension had been to use this to synchronize reading position across devices. But when documentation didn't emerge and the file sharing service was shut down, I figured it was dead. I don't have time to look at this in any detail right now, but it'll join my long list of things that it'd be nice to do sometime. If and when I get to it, I'll ping you if I have any questions. On Tue, Feb 2, 2016 at 2:43 AM, Selene Scriven wrote: Plain text is, and probably always will be, a thing. I have quite a few books in plain text format. They are usually .txt.gz or .txt.bz2 though. Epubs are just zipped HTML files, so supporting gzipped text files wouldn't be such a big thing. Another thing to add to my list of things to do in my copious free time. As for syncing, have you considered doing it via Calibre support? No, I actually like my users. :) I kid, but Calibre represents the opposite of my philosophy in many ways. If there's something I can do to make life better for Calibre users, I'll try to help. But Beru will never require the use of Calibre. On Tue, Feb 2, 2016 at 7:04 AM, Sam Bull wrote: How would OwnCloud file sync know how to sync settings and such from certain apps? That would require a moderately complex and hacky solution, which I believe would also require making the phone writeable (not recommended). Well, I'm sure each app could have a place to let you specify your OwnCloud credentials. Or it may be listed as a service provider. If you wanted to do it yourself, you'd need to break application confinement. As the Open Store shows, you can do that without making the phone writable. On Tue, Feb 2, 2016 at 8:21 AM, Neil McPhail wrote: (1),(2),(3),(4) are really the same problem. Passing data and sharing information/content between apps and directories is not mature enough yet. What is the preferred way forward here? Improve the capabilities and documentation of Content Hub or relax apparmor restrictions to certain "shared" directories? I would add (5) to this list as well. Once there's a mechanism for reading / writing outside of silos, I'd image the SD card could be subject to this as well. Maybe I'm missing some wrinkle here? (6),(7),(8) any input from the people looking after lifecycle management? Again, will existing frameworks be improved, new frameworks added or lifecycle management restrictions diluted? Is there an ETA for one or all of these? There've been discussion of lifecycle management in the context of GPS tracking apps. Check the archives if you haven't seen them; I forget what the conclusion there was. (9) I'm not quite following the thread of what is still switched on and what is switched off In summary, there is a synchronization service that could be used for this, but it's un-advertised, un-documented, and perhaps unused. Thanks all for your help, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Difficulty Sending Messages to this List (was Re: (no subject)
I have had a number of messages from yahoo accounts end up in my (gmail) spam foulder. Generally, Gmail says, Why is this message in Spam? It's from an address in the yahoo.com domain but has failed yahoo.com's required tests for authentication. There are (at least) two obvious things that could be causing this. 1) The sender is sending their email from their own server, not Yahoo's, causing this check to fail. This doesn't seem to be the case in the one example I checked out. 2) Google is being distracted by the fact the the email is coming from a Canonical server, not from a Yahoo server. This is true, but it doesn't explain why no other hosts have this difficulty. (Perhaps Yahoo is misconfigured?) In any event, the fact that someone somewhere has misconfigured their email is hardly cause for us to abandon email altogether. Robert On Fri, Mar 18, 2016 at 11:17 AM, Rodney Dawes wrote: No, that is not the problem. Sending a message with a different e-mail address will have your message land in the moderator queue, and you will get an e-mail saying so. If it goes through and ends up in the spam folders of people subscribed to the list, then the only relation to e-mail address used, may be that certain e-mail providers often have more spam sent through them, and thus, your e-mail address may be identified as spam. Beyond that, sometimes the e-mail message simply reads as if it could be spam, as not everyone is a native English speaker, and sometimes certain combinations of words and various typos, will simply be caught by spam filters. I have definitely seen mails from others on this list end up in the spam folder in gmail. But my local spamassassin config is a bit smarter, and it doesn't tend to send them to spam, when I go pull them out of gmail's spam folder. On Fri, 2016-03-18 at 18:25 +0100, Peter Bittner wrote: Have them check whether they use the same, identical email address for posting messages that they have subscribed with. I'm 99% sure this is the problem. It's like faxes. Why the hack do people lift the receiver and then hang up when they hear the fax signal? Probably because the system is not intelligent enough to not let them lift the receiver when a fax is coming in. (That would be the actual solution. And the phone shouldn't even ring.) I think when you are a software developer or a manager you should be able to understand the underlying problem. But of course, subscription to mailing list is an old-fashioned thing that just works. Just that it was not designed to times when everyone has 3 or more different email addresses. Welcome to 2016! Difficult to fix. At least difficult to quick-fix. Peter 2016-03-18 17:42 GMT+01:00 Randall Ross : > > I'm wondering how widespread this issue is... I've had two friends > this > week tell me that although they can receive ubuntu-phone messages, > they > cannot post. Is this a case of over-aggressive spam filters, or > perhaps a > moderation queue issue? > > Can the list admin investigate? > > Cheers, > Randall. > > On 03/18/2016 09:18 AM, Krzysztof Tataradziński wrote: > > Hello, > I'm only forwarding ;) > > " > -- Forwarded message -- > From: Marcin Hofmann > Date: 2016-03-17 18:01 GMT+01:00 > Subject: Mailing list bug > To: ubuntu-phone@lists.launchpad.net > > > I do not know if You noticed that lots of messages from this > mailing list > land in a spam folder of our e-mail boxes? The way we communicate > does NOT > work! How are we expected to create a better world when the > communication > chain is broken? A few days ago I sent a message with a problem > that nobody > answered. As I found out later someone found my message by mistake > in spam. > Wouldn't it be a better to make a forum with a clear topic list so > no of our > messages are gone and so that we can easily find some old > information, sent > months ago that we now badly need? If I were I manager in a company > like > Canonical I'd ban using mailing list. The spam filters nowadays are > to > strong and block important information. > > Cheers > > GTriderXC > " > > Best regards, > Krzysztof Tataradziński > https://launchpad.net/~ktatar156 -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] Building in an arm chroot with Qt Creator
Hi all, Previously, I had built binaries for armhf in my own chroot. But it's really out of date, so I thought I'd give the new system built into Qt Creator a try. Following the instructions [1], I got the chroot got built correctly, as far as I can tell, but when go to add a kit based on it to my project, the relevant kit is greyed out. Hovering over it, I get a warning message that "No Qt version set in kit". Sure enough, when I select it in the "Manage Kits..." dialog, no Qt version is selected. The only two available, however, are "Qt 5.0.2 in PATH (System)" and "Qt 5.0.2 (qt5)". Neither will work, though, since both are for x86 (/usr/lib/x86_64-linux-gn/qt5/bin/qmake). I've tried to add another Qt version with the path to the armhf version of qmake in the chroot, but the dialog won't accept that, since that path doesn't exist in the current filesystem. What am I doing wrong? If it matters, I'm on Ubuntu 14.04, upgraded from 12.04, using the SDK team PPA. Thanks, Robert P.S. Is Qt Creator not supposed to have a menu any more? It makes step one of [1] rather hard when you don't have a "Tools" menu! [1] http://developer.ubuntu.com/apps/sdk/tutorials/building-cross-architecture-click-applications/ -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Building in an arm chroot with Qt Creator
On Sun, Jul 6, 2014 at 4:39 AM, David Planella wrote: Do you have the Sysroot and Compiler fields set correctly? I seem to. The one difference I note is that I have the debugger set to "System GDB"; setting it to "Ubuntu SDK Debugger" gives an error that '/usr/bin/gdb-multiarch' is not found. I don't think that's the problem, but I can try to figure out how to install that if you think it'd help. If you or anyone else wants to help troubleshoot, the project is Beru, available at https://github.com/rschroll/beru, and I've attached my .user file, in case that's important. FWIW, when I began Beru, lots of the current framework wasn't in place yet, so I've probably made several choices that are now non-standard. But I don't know why any of them would be causing this. Thanks, Robert beru.pro.user Description: XML document -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Building in an arm chroot with Qt Creator
On Sun, Jul 6, 2014 at 2:13 PM, Zoltán Balogh wrote: One thing before that could help us to investigate is to try out the QtC 3.1.1 and the latest SDK tools: https://launchpad.net/~ubuntu-sdk-team/+archive/tools-development I'll give that a try this evening and let you know how it goes. Once you have upgraded (can be ppa-purge'ed and so downgraded easily later) please create new emulator and check out how the kits are. Do you mean an new chroot? FWIW, I've logged into the chroot and tried to do the compilation by hand. This is what happens: (click-ubuntu-sdk-14.10-armhf)root@copper:/home/rschroll/touch/beru-arm# qmake qmake: could not find a Qt installation of '' (click-ubuntu-sdk-14.10-armhf)root@copper:/home/rschroll/touch/beru-arm# qtchooser -l 5 qt5-arm-linux-gnueabihf qt5-x86_64-linux-gnu qt5 Why does an arm chroot have a x86 version of qt? Oh well, let's specify the arm one: (click-ubuntu-sdk-14.10-armhf)root@copper:/home/rschroll/touch/beru-arm# qmake -qt=qt5-arm-linux-gnueabihf qmake: could not exec '/usr/lib/arm-linux-gnueabihf/qt5/bin/qmake': No such file or directory (click-ubuntu-sdk-14.10-armhf)root@copper:/home/rschroll/touch/beru-arm# ls -l /usr/lib/arm-linux-gnueabihf/qt5/bin/qmake -rwxr-xr-x 1 root root 140 Jul 1 12:37 /usr/lib/arm-linux-gnueabihf/qt5/bin/qmake ?! Thanks for the help, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Building in an arm chroot with Qt Creator
On Sun, Jul 6, 2014 at 2:30 PM, Robert Schroll wrote: (click-ubuntu-sdk-14.10-armhf)root@copper:/home/rschroll/touch/beru-arm# qmake -qt=qt5-arm-linux-gnueabihf qmake: could not exec '/usr/lib/arm-linux-gnueabihf/qt5/bin/qmake': No such file or directory (click-ubuntu-sdk-14.10-armhf)root@copper:/home/rschroll/touch/beru-arm# ls -l /usr/lib/arm-linux-gnueabihf/qt5/bin/qmake -rwxr-xr-x 1 root root 140 Jul 1 12:37 /usr/lib/arm-linux-gnueabihf/qt5/bin/qmake ?! I think I understand this at least: (click-ubuntu-sdk-14.10-armhf)root@copper:/home/rschroll# file /bin/ls /bin/ls: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.24, BuildID[sha1]=64d095bc6589dd4bfbf1c6d62ae985385965461b, stripped It looks to me that this armhf chroot is actually a x86_64 chroot. The version of qmake I was trying to run is an arm executable, so this is why it was complaining about not finding the file. Anyway, I've added the PPA, updated, made a new chroot, and all the problems seem to persist. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Building in an arm chroot with Qt Creator
On Mon, Jul 7, 2014 at 3:18 AM, Zoltán Balogh wrote: Your project is a qmake project. Sadly qmake projects are not supported by the SDK. In order to build and package your project for armhf target you need to convert the project to cmake. Perhaps this should be the error message, rather than something about being unable to find Qt? Is there a guide or conversion script for going from qmake to cmake? Beru includes several third-party components using qmake, and I don't relish the idea of converting them all the cmake by trial-and-error. Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Building in an arm chroot with Qt Creator
On Mon, Jul 7, 2014 at 2:08 PM, Nicholas Skaggs
wrote:
The docs contain the fact you need to be cmake as part of the
building section, but it could have easily been missed. Does this
need to be more explicit?
It also appears after the place where I got stuck ("For an existing
project:"), so I never read that section carefully. ("Cut the red
wire. But first...")
In any event, people don't read the directions. And when they do, they
assume the rules don't apply to them. So adding a more descriptive
error message to Qt Creator ("You cannot use this kit because you are
not using CMake.") is the best solution, in my opinion.
Robert
--
Mailing list: https://launchpad.net/~ubuntu-phone
Post to : ubuntu-phone@lists.launchpad.net
Unsubscribe : https://launchpad.net/~ubuntu-phone
More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] Application crash on utopic but not trusty
Hi all, I've got a rather puzzling bug, and I'm out of ideas to attack it. Maybe someone here can offer some help. The full details are here https://github.com/rschroll/beru/issues/52; I'll try to summarize the important details below. Beru is a QML application that contains several compiled modules. The most recent version, 0.9.6, was compiled back in January on a saucy chroot. The reporter found it to be working on trusty r303. Upon upgrading to utopic r113, he found Beru would reliably crash the first time it tried to read information about an Epub file (using one of these compiled modules). There are no logs in /var/crash. I've recompiled Beru in a brand-new utopic chroot. That version fails in an identical manner. I have been unable to get a utopic emulator past the setup pages, so I am unable to reproduce the issue. Any ideas on what's going wrong or how to troubleshoot it? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Application crash on utopic but not trusty
On Tue, Jul 8, 2014 at 4:09 AM, Albert Astals Cid wrote: Run it with gdb/valgrind? How does one do that? Every time I try to launch an application from an ssh session, I get QUbuntu: Could not create application instance Aborted (core dumped) Obviously, there's some special sauce that I'm missing. Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Application crash on utopic but not trusty
On Tue, Jul 8, 2014 at 11:34 AM, Michał Sawicz wrote: On 08.07.2014 17:25, Robert Schroll wrote: How does one do that? Every time I try to launch an application from an ssh session, I get QUbuntu: Could not create application instance Aborted (core dumped) Obviously, there's some special sauce that I'm missing. A temporary workaround for that (until we get proper support for gdb runs) is passing: --desktop_file_hint=/path/to/your/file.desktop It's not working for me, I'm afraid. For example: phablet@ubuntu-phablet:/usr/share/click/preinstalled/com.ubuntu.dropping-letters/0.1.2.2.52$ qmlscene -qt5 ./dropping-letters.qml --desktop_file_hint=/usr/share/click/preinstalled/com.ubuntu.dropping-letters/0.1.2.2.52/dropping-letters.desktop QUbuntu: Could not create application instance Aborted (core dumped) Do I need to do some equivalent of "export DISPLAY=:0.0"? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Application crash on utopic but not trusty
On Tue, Jul 8, 2014 at 12:19 PM, Ted Gould wrote: The supported way is to use QtCreator, which I think that Benjamin got working (cc'd to check). It sets up a custom click that can be used for debugging and connects back to GDB on the host. Then you're running the app under confinement in the same manner as it normally runs, just with a gdb connection. Thanks, but I don't think this is going to work for me: 1) I need the user to run the debugging, since I'm unable to reproduce the crash. I can ask them to run something over a ssh connection, but asking them to download Qt Creator, download my source, set up a chroot, build the source, create the click, install it, and run it seems like a bit much! 2) As we established yesterday, my build process, based on qmake, won't work with the default systems that expect cmake. Any other suggestions? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Application crash on utopic but not trusty
On Tue, Jul 8, 2014 at 1:17 PM, Rodney Dawes wrote: Have you tried to recreate the problem inside a utopic lxc, on the native architecture of your system? No. I was considering making a i386 binary and testing it on the i386 emulator, but these problems getting gdb to run worry me on that approach. I've never used lxc before, but it sounds like a better way to go. Any getting-started guides to check out? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Application crash on utopic but not trusty
On Tue, Jul 8, 2014 at 1:47 PM, Michał Sawicz wrote: Right... you fell victim of this bug: https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1325984 Just use, say /usr/share/applications/webbrowser-app.desktop in the desktop_file_hint arg and it should work... It will show up as the browser, but hey... it's a workaround... Ha! It works. Thanks for all the help, everyone. I'll be back shortly, when I get a stack trace I can't understand :) Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Application crash on utopic but not trusty
Hi all, For the record, the problem was that I was assuming a pointer would be initialized as NULL. Evidently something changed between trusty and utopic, and that was no longer the case. Explicitly setting it NULL on initialization fixed the problem. Thanks for the help, everyone, Robert On Tue, Jul 8, 2014 at 2:00 PM, Robert Schroll wrote: On Tue, Jul 8, 2014 at 1:47 PM, Michał Sawicz wrote: Right... you fell victim of this bug: https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1325984 Just use, say /usr/share/applications/webbrowser-app.desktop in the desktop_file_hint arg and it should work... It will show up as the browser, but hey... it's a workaround... Ha! It works. Thanks for all the help, everyone. I'll be back shortly, when I get a stack trace I can't understand :) Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] QtWebKit WebViews appear black on i386 emulator
Hi all, I'm running r113 on an i386 emulator. It seems to be working pretty well, but all QtWebKit WebViews appear solid black. This means that HTML5 apps are broken, as are the parts of QML apps that rely on these WebViews. Oxide WebViews, however, are displayed just fine. I assume this is a problem with the emulator and/or my system. Perhaps it's an OpenGL issue? Has anyone else seen something similar? Any suggestions of solutions to try? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Are the menus in Ubuntu-SDK gone?
On Sat, Jul 12, 2014 at 9:31 AM, Nekhelesh Ramananthan wrote: The menu should be present. Are you missing the Build menu item or the entire Qtcreator menus? If the latter, you might want to check if you have the package appmenu-qt5 installed on your system as it enables the global menus for qt5 applications. If this doesn't work, please specify which desktop environment you are using. I'm having the same problem. Gnome Shell 3.10 on Ubuntu 14.04, using the Ubuntu SDK team PPA. appmenu-qt5 0.3.0+14.04.20140415-0ubuntu1 is installed. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Are the menus in Ubuntu-SDK gone?
On Sat, Jul 12, 2014 at 10:28 AM, Nekhelesh Ramananthan wrote: I ran Qtcreator on KDE and noticed that the menus were missing. I figured this could be due to KDE not supporting global menus. On removing appmenu-qt5 and restarting Qtcreator I got the menus back. Perhaps you could try the same? That did it for me. Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] Persisting cookies in Oxide
Hi all, My app Beru [1] is a QML app, but does a lot of its work inside a QtWebKit WebView. One of those things is to persist settings in a cookie stored in the WebView [2]. With QtWebKit, this worked fine. I'm trying to port it to the Oxide WebView, but this is no longer working. Cookies persist for a given execution, but they are apparently cleared when Beru is restarted. Is there a setting I need to adjust to allow the cookies to persist? If not, is this something that will be added? Thanks, Robert [1] https://github.com/rschroll/beru [2] https://github.com/rschroll/beru/blob/master/html/styling.js#L63 -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Persisting cookies in Oxide
On Tue, Jul 15, 2014 at 8:26 PM, Alexandre Abreu wrote: There might be a few reasons for that. Are you directly using the Oxide.WebView or do you use it through the Ubuntu.Web.UbuntuWebView component? I'm using the Oxide.WebView directly, since I need to set some custom user scripts. If you are directly using the Oxide.WebView you might not have set the dataPath for the WebContext of the Oxide WebView. Unless things have changed, an empty dataPath means that the cookies/data resources are not persisted, That seems to have done it. I already have a C++ module that returns QStandardPaths::DataLocation (amongst other things), so I could replicate the behavior of the UbuntuWebView easily enough. Do you know if it's possible to make Oxide read the cookies stored by QtWebKit? If so, does anyone know where QtWebKit stored its cookies? Also although less likely since the above might put your in the right direction, you might want to play with the sessionCookieMode of the WebContext associated with your Oxide webview and set it to WebContext.SessionCookieModePersistent. It seems to work, even though sessionCookieMode is set to SessionCookieModeEphemeral. Is this expected? Should I change it to Persistent to be safe in the future? I don't know if this is relevant to my issue, but I get these messages on the console, regardless of whether I'm setting the WebContext properties: Cannot set WebContext.sessionCookieMode once the context is in use Cannot set WebContext.dataPath once the context is in use I have no idea what code is causing this, but I'm pretty sure it's not anything I'm doing. Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Persisting cookies in Oxide
On Wed, Jul 16, 2014 at 7:53 AM, Olivier Tilloy wrote: There’s not code readily available in oxide to import QtWebKit cookies, but I guess you could write some if you really need to. QtWebKit stores its cookies under $dataLocation/.QtWebKit/cookies.db, in an sqlite database. Thanks. I hadn't thought to look for a hidden directory inside the hidden path. It seems to work, even though sessionCookieMode is set to SessionCookieModeEphemeral. Is this expected? Should I change it to Persistent to be safe in the future? See https://code.google.com/p/chromium/codesearch#chromium/src/content/public/browser/cookie_store_factory.h&q=SessionCookieMode&sq=package:chromium&l=42 for an explanation of what the various values for SessionCookieMode mean. Or I could just realize that "SessionCookieMode" probably has something to do with session cookies, which these cookies aren't. This is easier to do now that I'm awake. I don't know if this is relevant to my issue, but I get these messages on the console, regardless of whether I'm setting the WebContext properties: Cannot set WebContext.sessionCookieMode once the context is in use Cannot set WebContext.dataPath once the context is in use I have no idea what code is causing this, but I'm pretty sure it's not anything I'm doing. That looks similar to https://bugs.launchpad.net/webbrowser-app/+bug/1324855, the fact that you’re not using Ubuntu.Web seems to suggest the issue is in oxide itself, would you mind commenting on the bug report to explain how beru is also affected? It turns out that this is caused by another Page, which has an Ubuntu.Components.Extra.Browser(0.2).UbuntuWebView in it, which, if I'm not mistaken, is the same as Ubuntu.Web.WebView. Thanks for the help, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Persisting cookies in Oxide
On Wed, Jul 16, 2014 at 11:22 AM, Olivier Tilloy wrote: That’s correct. The latter is preferred over the former (and is properly documented at http://developer.ubuntu.com/api/qml/sdk-14.10/Ubuntu.Web/). For whatever reason, Ubuntu.Web isn't installed on my 14.04 system (with SDK team PPA), while both Oxide and the new UbuntuWebView are. I don't know if this is a known issue or not. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Community participation
On Thu, Jul 17, 2014 at 8:11 AM, Nathan Haines wrote: On 07/17/2014 05:02 AM, Oliver Propst wrote: Sure, but I think if they really want expand and create the a similar atmosphere around it as the desktop operating system they should try to work closer with the the community. They can't work more closely with the community, really. Consider this bug: https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1276808. Yes, it's publicly available and open for comments from everyone, but the description is simply a link to a private canonical.com document. That's not truly open for community participation. Now, I'm sure this is just a mistake -- whoever posted that to the bug probably did so with the intention that everyone could view it. And since we all live logged into Google, it never occurs to people who have access that other people don't. However, it does imply that Canonical is making development decisions in private Google documents. So don't tell me "they can't work more closely with the community"! That said, I'm not convinced that this is actually a problem. One of the things I like best about Ubuntu is that it is developed with strong direction and vision, and doesn't merely follow the majority view of the community. It may well be that that spec had to be developed in private, since it was part of this directed vision. That's fine! So my complaint isn't that Ubuntu isn't completely community-driven, it's that we keep saying it is. Let's not. Then when we get emails like Oliver's, treat it as a request to alter the balance. Robert (I hope this email makes sense. I rewrote it several times, changed my conclusion in the process, and almost deleted it. I'm hoping that it's better to get these opinions out there, rather than leaving them hidden, but I reserve the right to decide tomorrow that I was all mistaken today.) -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] dlopen: cannot load any more object with static TLS
Hi all, I've been trying, and largely failing, to get Beru [1] to run on emulators and devices. I think I've squashed a bug related to apparmor permissions [2], but now when running on a i386 emulator I'm getting the error: file:///usr/lib/i386-linux-gnu/qt5/qml/Ubuntu/Components/Extras/Browser/UbuntuWebView01.qml:20 plugin cannot be loaded for module "QtWebKit": Cannot load library /usr/lib/i386-linux-gnu/qt5/qml/QtWebKit/libqmlwebkitplugin.so: (dlopen: cannot load any more object with static TLS) I'm trying to figure out what that means. (Can it not load more objects with static TLS, or can it no longer load objects with static TLS? And from what language was this message translated?) I'm assuming TLS stands for "Thread-Local Storage", not "Transport Layer Security". If so, a Stack Overflow answer [3] suggests that this is a problem when cross-compiling for i386 from x86_64. Does this mean that this won't be a problem on an arm device? If you feel like testing, please note that this bug is not deterministic. It only occurs sporadically (perhaps related to the memory addresses allocated), so please run the program several times before deciding that it doesn't happen. There's an issue [4] tracking this problem, if you want more details. Respond here or there as you prefer. Thanks, Robert [1] https://github.com/rschroll/beru [2] https://bugs.launchpad.net/qtcreator-plugin-ubuntu/+bug/1358004 [3] http://stackoverflow.com/questions/14892101/cannot-load-any-more-object-with-static-tls [4] https://github.com/rschroll/beru/issues/56 -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] dlopen: cannot load any more object with static TLS
We've been unable to reproduce this problem on either an arm emulator or device, so it looks like it's specific to the i386 architecture. I'd still love to hear people's thoughts about what's going wrong and how to work around it. Thanks, Robert On Tue, Aug 19, 2014 at 11:24 AM, Robert Schroll wrote: Hi all, I've been trying, and largely failing, to get Beru [1] to run on emulators and devices. I think I've squashed a bug related to apparmor permissions [2], but now when running on a i386 emulator I'm getting the error: file:///usr/lib/i386-linux-gnu/qt5/qml/Ubuntu/Components/Extras/Browser/UbuntuWebView01.qml:20 plugin cannot be loaded for module "QtWebKit": Cannot load library /usr/lib/i386-linux-gnu/qt5/qml/QtWebKit/libqmlwebkitplugin.so: (dlopen: cannot load any more object with static TLS) I'm trying to figure out what that means. (Can it not load more objects with static TLS, or can it no longer load objects with static TLS? And from what language was this message translated?) I'm assuming TLS stands for "Thread-Local Storage", not "Transport Layer Security". If so, a Stack Overflow answer [3] suggests that this is a problem when cross-compiling for i386 from x86_64. Does this mean that this won't be a problem on an arm device? If you feel like testing, please note that this bug is not deterministic. It only occurs sporadically (perhaps related to the memory addresses allocated), so please run the program several times before deciding that it doesn't happen. There's an issue [4] tracking this problem, if you want more details. Respond here or there as you prefer. Thanks, Robert [1] https://github.com/rschroll/beru [2] https://bugs.launchpad.net/qtcreator-plugin-ubuntu/+bug/1358004 [3] http://stackoverflow.com/questions/14892101/cannot-load-any-more-object-with-static-tls [4] https://github.com/rschroll/beru/issues/56 -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] ownCloud app.
On Wed, Aug 27, 2014 at 6:46 PM, Daniel Holm wrote: 3. Added a WebView (Webkit since I can't get the Oxide stuff to work...) Make sure you add the "webview" policy group to your apparmor file and increase the version to 1.1. Otherwise, you'll get obscure errors from icu_util.cc. I've filed a bug about the lack of discoverability here: https://bugs.launchpad.net/qtcreator-plugin-ubuntu/+bug/1358004 Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Errors after installing ubuntu-sdk
On Thu, Sep 4, 2014 at 3:18 PM, Andrey Kozhevnikov wrote: Thank you, i just updated to 14.10 to match qt 5.3 version. I want to clarify: there are no way to compile and run qmake .pro project on emulator right now? I'm not strict to QtC, can build using cli, if possible. You can, but you have to do everything by hand: Make your own chroot (or VM), install dependencies, copy in your source, build it, create a click package, copy it to the emulator, install it, and launch it. I used to do this, but it was a pain. Thanks to Randy Olive, I'm set up with CMake now, and a single click builds, packages, installs, and runs my project on the emulator. (Well, about 50% of the time -- I just keep hitting the button until it all works.) It's worth the pain of the transition, especially if you can get someone else to do it for you :). If it helps, here are the commits where we made the switch from qmake to cmake: https://github.com/rschroll/beru/commit/4b82800b71ed6b1cd295e75d3e7372d6b2c5327e https://github.com/rschroll/beru/commit/1c0c113746d67f5aa1bcc3b0fc8d0f3c3823e66e Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] Testing Translations
Hi all,
I recently added translations to my QML app, and I'm trying to test
them now. But I'm having trouble getting even a basic case to work.
Can someone say what I'm doing wrong?
I've attached a sample program that demonstrates my confusion. It
reports the selected language, sets a bogus domain, and translates the
string "Done" twice, once with tr() and once with dtr() and the "apt"
domain. Then it sets i18n.domain = "apt" and does both translations
again. Here's what I get:
$ LANGUAGE=es qmlscene i18n.qml
Language: en_US.UTF-8
Dummy domain set
tr: Done
dtr: Listo
Correct domain set
tr: Done
dtr: Done
Note that the language is not detected, but the first call of dtr()
gets the translation correct. (The first call of tr() should, and
does, fail.) However, when the domain is set, neither translation
works.
I can also do:
$ LANG=es qmlscene i18n.qml
(process:26027): Gtk-WARNING **: Locale not supported by C library.
Using the fallback 'C' locale.
Language:
Dummy domain set
tr: Done
dtr: Done
Correct domain set
tr: Done
dtr: Done
The same thing happens when setting LC_ALL or LC_MESSAGES. FWIW, a
simple Python script [1] works with no problems.
Can someone point out what I'm doing wrong? I'm using 14.04 with the
SDK team PPA, if that matters.
Robert
[1] Namely,
import gettext
gettext.textdomain("apt")
print gettext.gettext("Done"), gettext.dgettext("apt", "Done")
import QtQuick 2.0
import Ubuntu.Components 0.1
Item {
Component.onCompleted: {
console.log("Language:", i18n.language);
i18n.domain = "blah";
console.log("Dummy domain set");
console.log(" tr:", i18n.tr("Done"))
console.log(" dtr:", i18n.dtr("apt", "Done"));
i18n.domain = "apt";
console.log("Correct domain set");
console.log(" tr:", i18n.tr("Done"))
console.log(" dtr:", i18n.dtr("apt", "Done"));
}
}
--
Mailing list: https://launchpad.net/~ubuntu-phone
Post to : ubuntu-phone@lists.launchpad.net
Unsubscribe : https://launchpad.net/~ubuntu-phone
More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Testing Translations
On Tue, Sep 9, 2014 at 7:35 PM, John Lenton wrote: $ LC_ALL=es_AR.UTF-8 LANGUAGE=es qmlscene i18n.qml qml: Language: es_AR.UTF-8 qml: Dummy domain set qml: tr: Done qml: dtr: Listo qml: Correct domain set qml: tr: Listo qml: dtr: Listo Unfortunately, that doesn't work for me. I get the GTK warning and no translations working, just as when I set LANG=es. Perhaps that's a sign something is broken on my system? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] Fwd: Re: Testing Translations
Missed the list. -- Forwarded message -- From: Robert Schroll Subject: Re: [Ubuntu-phone] Testing Translations Date: Tue, 09 Sep 2014 20:03:34 -0400 To: John Lenton On Tue, Sep 9, 2014 at 7:53 PM, John Lenton wrote: Do you have the appropriate locales installed? I have the relevant es.mo file installed. Is that what you're asking about? I haven't installed system-wide support for Spanish, but that doesn't keep me from launching specific programs with LANGUAGE set. Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Fwd: Re: Testing Translations
On Wed, Sep 10, 2014 at 2:11 AM, David Planella wrote: 1. I think you might not have Spanish translations installed on your system. That means that probably the es_ES.UTF-8 locale is not generated. Try starting the language support app from the Dash and installing Spanish from there. That did it. Thanks! But I'm going to complain anyway: Every other program I've tested works fine if LANGUAGE is set but the language pack isn't installed, so long as the appropriate .mo file is in place. Why should qmlscene/i18n work any differently? And why does dtr() work, but only if the default domain is set to something else?!? 2. In Ubuntu translations are shipped in language packs. But only for the default apps, surely. Anything in a PPA will need to ship it's own translations, and that's what I'm trying to test. Since I have the necessary .mo file from the PPA, why should I need the language pack? (I used apt in my example precisely because it would already be installed for everyone.) 3. The language specification for LANGUAGE and LANG is different. For LANGUAGE, it is just the 2-letter or 3-letter language code; e.g. es (or a list of them separated by semicolon); for LANG it's the full locale, e.g. es_ES.UTF-8. I see. It seems that Python's gettext is more permissive. It allowed LANG=es, but nothing else I've tested does. Thanks for setting me straight. Is this worth submitting a bug about? If so, against which project? Thanks again, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] [Design] Popover pointing to header buttons
Hi all, My app Beru has a main screen that features a list that can be sorted in several different ways. The sorting UI consists of a Sort toolitem that opens a popover with three different sorts to choose from. When the toolitems were in a toolbar, this worked fine. When they got moved to the header, the popover no longer points to the correct spot. When I reported this on Launchpad [1], I was told, Popovers pointing to buttons in the new header is not supported. Design recommends to use a pagestack in those cases. For specific needs of an app, please talk to designers for a solution Is this true? Tabs are currently handled with a popover pointing to a button in the header, so it seems odd not to allow this mechanism for other buttons. I don't understand how I'm supposed to solve this with a pagestack. Am I really supposed to create a new page to hold a single option? What solution does the design team suggest? For what it's worth, I'm currently faking the pointer by putting a rectangle behind where I know the header item will be and pointing to that. It looks and works just fine [2]. Thanks, Robert [1] https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1337699 [2] http://rschroll.github.io/beru/assets/ss-library-author.png -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Design] Popover pointing to header buttons
On Thu, Oct 9, 2014 at 7:48 AM, Giorgio Venturi wrote: Yeah, I totally agree with Tim, that's the recommended way of dealing with sorting & filtering options That does look nice. Thanks! Unfortunately, it's not available on 14.04. As that's one of my targets, I'll have to stick with the work-around for now. Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Design] Popover pointing to header buttons
On Fri, Oct 10, 2014 at 6:40 PM, Nekhelesh Ramananthan wrote: I am curious, are you targeting Trusty desktop or trusty phone channel? The desktop. I started Beru because I wanted a simple epub reader for the desktop, so that's actually my focus. Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] ANNOUNCEMENT: Vivid and ubuntu-rtm landing procedures
On Fri, Oct 24, 2014 at 12:23 PM, Rodney Dawes wrote: But I don't see any way for supporting multiple different framework versions in an app, on a system that is designed to be a single rolling release. I've asked this before [1], but I've yet to receive a satisfactory response: What about the desktop? The killer feature of this new SDK, we've been told, is that it'll work on the phone and the desktop equally well. So, 1) Is the desktop moving to a rolling release? As an LTS-to-LTS kinda guy, I certainly hope not. 2) Will the SDK be kept up to date via backports or PPAs on the desktop? So far, it hasn't happened. 3) Will the SDK be completely backwards compatible? Obviously not -- the point of versioning it is to add features. 4) Will there be some way for apps to react to the framework they find themselves running under? So far I've heard "no". Without any of these, we app authors have to (a) target a framework old enough to be on the oldest desktop we're interested (likely a LTS, and do you really want people using Ubuntu.Components 0.1 until 2016?) and (b) hope that the old frameworks are completely forward-compatible on newer images (they aren't). Already, I find that the only way to get my app to run both on a current device and the 14.04 desktop is to do Loader hacks to detect which framework is supported [2]. It'd be much better if there were a straight-forward way to do this. Robert [1] https://plus.google.com/108747901910183509998/posts/GZLFqo3GJVS [2] https://github.com/rschroll/beru/commit/6ffc8f0a8f659afdce7b04ac3bb49ff58ae65563 -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] Scopes questions
Hi all, I started playing around with scopes and have run into a few questions. If there's a better place to ask about this, please let me know. Otherwise: 1) Is it possible to delay a query until the user hits "Enter"? Or failing that, until a second or so has passed from the last key stroke? I'm experimenting with a service that rate-limits queries, and the partial queries won't lead to useful results anyway. I'd prefer an explicit indication from the user that they're done, but I'm not sure if that's possible. I'm guessing that the delay is possible, but I don't know if that should go in the client or the query or somewhere else. 2) Is it possible to display images without any rescaling or cropping? I can tell the service the maximum size for images, but the actual size and aspect ratio will depend on the content. Both in the query and the preview, I'd like to insert those images as-is, but the defaults rescale and, in the query view, crop them. In a related issue, I've set ShapeImages=false in the .ini file, but that doesn't seem to have any effect. That could be related to... 3) How do you get an updated .ini file to be used? I'm testing on an emulator (r243), and it seems to be caching the old .ini files. The only way I can get it to recognize an updated one is to restart the emulator. Surely there's a better way? 4) Is there some guide to making asynchronous queries? The tutorials I'm working off of seem to be synchronous, although the docs suggest the different parts may be running in different threads. But I'd like to fire off a query, get and display some results, and then fire off some more queries based on those results and display them when those replies get back. I've never worked with asynchronous code in C++, so a pointer to a C++ tutorial may be helpful. This may be relevant for (1). Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Scopes questions
On Mon, Nov 3, 2014 at 3:49 PM, Rodney Dawes wrote: There should be a short delay (~100ms) already, after the last key was pressed, before the query is sent. Maybe this needs to be closer to 300ms or so, as typing on the phone can be pretty slow, unlike typing with a real keyboard. If you are finding the current delay is not long enough, please file a bug that the queries are happening too fast. I'm testing on an emulator, where typing is *really* slow, so I'm not in any position to judge this empirically. But 100ms is roughly 100 wpm, which is a rather fast clip! Is this something that could be made a setting in the .ini file? If you're typing a mathematical expression, as in my case, you're probably going to be moving significantly slower than if you're typing a buddy's handle. However, when a new query is sent to the scope, the old query has cancel() called on it. You may need to add additional code to close the connection to the server, or similar, when you receive the cancel call. Presumably a scope could set up a timeout and only send out the HTTP request if it doesn't get a cancel() in that time. Any gotchas with that approach? 2) Is it possible to display images without any rescaling or cropping? I don't think so. The card size and aspect ratio you specify in the scope, tells the dash how to draw the images. To fit properly, the dash will scale and may crop, the images. This is a problem. Some responses will be mathematical formulae. They can be formatted as MathML or LaTeX, but unless I'm mistaken, the scope can only display plain text. The server can also respond with these rendered as images, but these look ugly if a short formula is scaled to stretch all the way across the screen. 3) How do you get an updated .ini file to be used? It looks like this was an issue with the older emulator. I've just updated to rtm v5, and this is no longer a problem. It is basically up to the developer of the scope to decide how to do that. You will need to block in the Query::run, hold the reference to the Reply pointer, and manage asynchronous calls in the background, via additional threads or event loops. Am I correct in understanding that the results show up in the scope each time reply->push is called? If so, I could probably get nearly the same result by making a series of synchronous requests with pushes in between. Presumably we'd also need to be checking if cancel had been called anywhere in there. Thanks for the detailed answers! Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] Scopes questions
Thanks, all, for the helpful explanations. I think I'm finally starting to grok what's going on with scopes. I need some time to digest everything you've told me, but I'll be back with more questions soon. One comment: There seems to be several reasons to prefer asynchronous behavior, but the sample projects shipped with QtCreator and the tutorials on developer.ubuntu.com all use synchronous requests. A simple example using asynchronous requests would be very helpful. Obviously such an example will be a bit more complicated. But we're already dealing with a system in which various things are happening in different threads. Async requests ought to fit into this existing mindset pretty easily. On Mon, Nov 3, 2014 at 4:50 PM, Rodney Dawes wrote: I do think 100ms is probably to fast, personally. 300ms would be much better, and on par with average users. I'm happy to file a bug about this, if you can tell me which project to file it against. Thanks again, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] [Scopes] Responding to preview actions
Hi all, I'm trying to understand how to have a scope respond to the user's activation of a preview action, by running a new query for example. This only documentation I've seem on it is in the API docs, and that's not entirely clear. I don't understand the difference between ScopeBase::perform_action and ScopeBase::activate, for instance. I've put up a test case I made on Launchpad [1]. You can enter a number and it presents some multiples of that number. Each preview page has a button that I want to drive another query showing the multiples of that number. But clicking on that button does nothing. I've given my scope both perform_action and activate methods, but neither is being called. What am I doing wrong? Thanks, Robert [1] https://code.launchpad.net/~rschroll/+junk/multiples -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Scopes] Responding to preview actions
On Fri, Nov 7, 2014 at 1:53 AM, Michi Henning wrote: You set the "uri" attribute with builder.add_tuple(). That tells the shell that you want that uri to be activated when the button is pressed. Don't set the "uri" attribute, and your activate() method is called as expected. Aha! I had forgotten the uri attribute for results before, which kept them from working. So I figured I might as well give these actions uri attributes to avoid that problem. Little did I know I've updated my code, and it's now running the new query when you press the button on the preview page. The odd thing is that it doesn't switch back to the query view--it stays on the preview view when you click the button. When you click the back arrow, you're returned to the query view with the new query having been run. Is this a bug in my code, a bug in the scope display, or desired behavior? Thanks for all the help! Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Scopes] Responding to preview actions
On Fri, Nov 7, 2014 at 1:08 PM, Pawel Stolowski wrote: I think you're experiencing https://bugs.launchpad.net/ubuntu/+source/unity-scopes-shell/+bug/1335761 That'd seem to be the problem. Good to know it's not my fault for once. One more question on this subject: Is it possible to intercept the signal to display a result, and do something else there instead. For instance, could I get the click on a number and immediately display that number's multiples, instead of going to the preview screen. Thanks for the help, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Scopes] Responding to preview actions
Finally got a chance to try this, and it works like a charm. Thanks! Robert On Mon, Nov 10, 2014 at 8:27 AM, Pawel Stolowski wrote: Ah, correction, I totally forgot about canned queries; you can use them to do what you want (if I got your idea right): just use a canned query (see CannedQuery::to_uri()) as an uri of your result and that should fire off your new query when you activate the result. Pawel On 10.11.2014 10:06, Pawel Stolowski wrote: Hi Robert, n 07.11.2014 20:37, Robert Schroll wrote: On Fri, Nov 7, 2014 at 1:08 PM, Pawel Stolowski wrote: I think you're experiencing https://bugs.launchpad.net/ubuntu/+source/unity-scopes-shell/+bug/1335761 That'd seem to be the problem. Good to know it's not my fault for once. One more question on this subject: Is it possible to intercept the signal to display a result, and do something else there instead. For instance, could I get the click on a number and immediately display that number's multiples, instead of going to the preview screen. No, that's not possible. While technically it wouldn't be difficult to support, we deliberately decided to always force previews to have consistent UI (only exception is for installed Apps, which we treat special in the shell). Cheers, Pawel Thanks for the help, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] [Scope] Failure to authenticate with online accounts on recent emulator images
Hi all, As some of you may know, I'm working on a Gmail scope [1] that uses the online accounts Google account for authentication. This works (sporadically) in RTM image 5 on the emulator, but I have been unable to get it to work with images 6 or 7. Specifically, the scope creates an OnlineAccountClient and goes through the results of get_service_statuses() to find one that's authenticated. If none are found, it uses register_account_login_item() to give the user a button to press to get the account set up. On images 6 and 7, the button is always shown, and pressing it always leads to a Google login page. This occurs whether or not there's a Google account already created, and whether or not the scope is allowed to use this. If the login page is filled out, another Google account is created in online accounts (with the scope's access enabled), but the scope will never use it. I've tried restarting the scope, restarting signond, restarting unity8, and restarting the emulator, but the same behavior occurs time and time again. Does this behavior indicated a problem in my scope? If so, how do I go about debugging it? If not, is it a known problem with online accounts? Thanks, Robert [1] https://github.com/rschroll/gmail-scope -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Scope] Failure to authenticate with online accounts on recent emulator images
On Tue, Nov 18, 2014 at 2:12 AM, Alberto Mardegan wrote: I suspect that there might be something wrong with the parameters you pass to the OnlineAccountsClient constructor: indeed, the service-type parameter there should not be "email", but the same value you have in the element of your .service file. That was indeed the problem! Thanks for finding it. Note that the documentation for the OnlineAccountClient API [1] describes service-type as 'The type of service (E.g. "sharing")', while the developer guide [2] uses the RDN for the service type in the XML file. I don't know which is correct, but these two should be harmonized. Though, I doubt that this can be the reason for the bug you are experiencing. And it's especially weird that this works on image 5 and not in the later ones; maybe something changes in unity-scopes-api? The image 5 emulator was the one on which I spent hours of cargo-cult programming getting this to work in the first place. (This was before the developer guide was released.) So it's possible that I had accidentally made these two values the same at one point. Would they be sticky after login succeeded once? I'm willing to leave this as an unsolved mystery, but if you think it's interesting, let me know and I can dig deeper. Thanks again, Robert [1] http://developer.ubuntu.com/api/scopes/sdk-14.10/unity.scopes.OnlineAccountClient/#af1139a1b81d9d95c0c84ea52d51dc07c [2] http://developer.ubuntu.com/apps/platform/guides/online-accounts-developer-guide/#oa_packaging -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Scope] Failure to authenticate with online accounts on recent emulator images
On Tue, Nov 18, 2014 at 2:59 AM, Robert Schroll wrote: Note that the documentation for the OnlineAccountClient API [1] describes service-type as 'The type of service (E.g. "sharing")', while the developer guide [2] uses the RDN for the service type in the XML file. I don't know which is correct, but these two should be harmonized. [1] http://developer.ubuntu.com/api/scopes/sdk-14.10/unity.scopes.OnlineAccountClient/#af1139a1b81d9d95c0c84ea52d51dc07c [2] http://developer.ubuntu.com/apps/platform/guides/online-accounts-developer-guide/#oa_packaging Against which of these should I file a bug? Thanks, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
[Ubuntu-phone] [Scope] More authentication questions
Hi all, I'm completely confused about how I should handle authentication over the life of my scope. I'll try to explain my specific problems below, but if you think it's best to ignore those and give me an overview of the recommended approach, that'd be great too. My scope [1] uses Google oauth2 for authentication, and can't do anything meaningful without it. I followed the example from the API documentation and authenticate in the Query::run() function. I store the key for later use, as I hit the Google servers from other places from the code. This works fine for a while, but if the scope is open for long enough, eventually Google will reply with a 401 Invalid Credentials response. I assume there's only a limited validity to the key I get. I could detect this response and then re-run the authentication, but 1) It'd be better to know that the key was expired before hand. Is there any way to know this? 2) The logical place to put this check would be inside my get() method, my post(), method and my batch() method. But this is already getting redundant, and I'm not sure what to do when authentication fails. Right now, the check is in the Query::run(), so I can run oa_client.register_account_login_item(). Is it better to handle the oa_client close to the network or close to the UI? I've also heard that scope needs to handle being killed at arbitrary times, which presumably means it may start up in an arbitrary state. (I think this is suppressed when running from Qt Creator, right?) This complicates the issue further, since now I have to guard each API access against authentication, not just the "first" in the standard workflow. (This may make worries about expiring keys moot, though.) How to do you handle authentication failures that could occur anywhere? Are there best practices documented somewhere, or am I on my own? Thanks, Robert [1] https://github.com/rschroll/gmail-scope -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
Re: [Ubuntu-phone] [Scope] More authentication questions
On Wed, Nov 19, 2014 at 1:54 AM, Marcus Tomlinson wrote: Its really up to you where you want to query the oa_client. Once the OA backend has requested and given you an access token, subsequent calls to retrieve statuses from oa_client will have negligible overhead (In fact, even re-instantiating a new OnlineAccountClient object on every query will have negligible overhead as its actually the OA backend that caches the token). Therefore, the easiest, stateless way to do this is to simply construct a new OnlineAccountClient object everywhere you need the token (as shown in the API example). This way it doesn’t matter what state your scope is in / was in, and you’ll always guarantee an up-to-date token. The OnlineAccountClient class was designed to construct and destruct efficiently for this very purpose. I did not appreciate this. I had thought each construction of an OnlineAccountClient would hit the oauth server again, so I was trying to do my own caching. But it looks like I needn't do that. I'll make the change an instantiate a new client each time I need it. I assume this will also protect me from expiring tokens? Not sure I understand what you mean “start up in an arbitrary state”. The scope will simply start again from fresh (i.e. a fresh process is started). I had been assuming that the scope would always run a query first, so my authentication check is only on the query. But if the dash is displaying the results of a query and the scope is killed and restarted, the first action it takes might be a preview. This is just another reason to not try to cache the credentials but to create a new client for each request. So yeah, wherever in your code you could expect an authentication failure, you’ll need to handle it there and then. I’m not sure I understand what you mean by anywhere though. Can you give me an example of a point where you could get an authenticating failure that would leave you in an unrecoverable state? One example: On my previews, I have several actions that trigger a call to the server. What should I do when that call fails? I'm not in a position to add any UI elements here, so I don't see how to prompt the user to log in again. However, if I construct a new OnlineAccountsClient for each of these calls, I shouldn't have to worry about tokens timing out (right?), so the only failure mode would be the online accounts backend being unable to authenticate. But the backend has obviously already authenticated at least once, since I was able to get to a preview page. They only ways I can get into this state are if the remote server goes down (in which case I can't do anything about it) or if the user removes their online account in the middle of using my scope (in which case they deserve brokenness.) If you do want to pass a single oa_client object around (instead of reconstructing a new one every time you need it), what you should be able to do is simply call oa_clinet->refresh_service_statuses() when your token is no longer valid (401 returned). This should refresh the held token so that the next time you call get_service_statuses() you’ll receive the new one. ... Does that sound more like what you need? That's what I'd need with my current approach. But you've already convinced me that my current approach is wrong. :) Thanks for the detailed answers, Robert -- Mailing list: https://launchpad.net/~ubuntu-phone Post to : ubuntu-phone@lists.launchpad.net Unsubscribe : https://launchpad.net/~ubuntu-phone More help : https://help.launchpad.net/ListHelp
