Sharing StorageItems from a byte[] array - windows-8

I want to implement the Share source contract in my WinRT C# Metro app (Windows Release Preview). My app is storing arbitrary files. Not in the filesystem, but instead I get the data over a WCF service as byte[]. Now I want to share such "files" in my app.
The only possibility I've seen with a standard data format is using the SetStorageItems() method on the DataPackage. Thus I'm facing the challenge to convert the data from my byte array to a StorageFile, which can be shared. I found the StorageFile.CreateStreamedFileAsync() method and wanted to use it in this way:
// filename: string
// fileContent: byte[]
// ... setting DataPackage title and description ...
DataRequestDeferral deferral = args.Request.GetDeferral();
var file = await Windows.Storage.StorageFile.CreateStreamedFileAsync(filename,
async stream => await stream.WriteAsync(fileContent.AsBuffer()), null);
args.Request.Data.SetStorageItems(new List<IStorageItem> { file });
deferral.Complete();
It compiles fine, but it doesn't work as expected. I've tried the sharing with the standard Mail app. The Mail share view opens and I can create a new mail. The file is shown without thumbnail (as expected), but the e-mail can't be sent. It's showing the sending progress for several minutes and then an error occurs: "Couldn't share {filename} with Mail.". The share charm shows "Something went wrong" and "[...] Mail can't share right now. Try again later.".
It works perfectly when I load the StorageFile from the file system: the mail opens and is sent within seconds, no problems here. So either I'm using CreateStreamedFileAsync() wrong or there's a bug in this method, what do you think?

In the callback passed into CreateStreamedFileAsync, you need to actually dispose of the object - that signals to the OS that you are done.
Wrote a complete example here

The Mail app is not a target for sharing files. From http://blogs.msdn.com/b/b8/archive/2012/06/14/building-the-mail-app.aspx: "Mail supports sharing text, links, and pictures."
Remember that there are 2 parts of the Share contract: Share sources and Share targets. As you know, there are many different data formats that can be shared between them, like text, pictures, URIs, and files. The full list of the different data formats that are supported is at http://msdn.microsoft.com/en-us/library/windows/apps/hh771179.aspx.
I recommend that you use the Share Target Sample app to test that your file is being shared properly - share to this and it will display everything that is being shared from your app as a source (and it does accept files for sharing). You can download it from http://code.msdn.microsoft.com/windowsapps/Sharing-Content-Target-App-e2689782. You can also use the Share Source Sample app as an example and leverage code from this app; you can download it from http://code.msdn.microsoft.com/windowsapps/Sharing-Content-Source-App-d9bffd84.
Hope that helps!

Ok, perhaps the preview version of the Mail app doesn't handle the sharing target contract correctly. Using the SDK sample app "Sharing Content Target App" from http://code.msdn.microsoft.com/windowsapps/Sharing-Content-Target-App-e2689782, sharing a StorageItem created in memory with the StorageFile.CreateStreamedFileAsync() method posted above works fine.
Thus, that's the way you should go when you want to share in-memory byte[] arrays. For testing, make sure that the share target app doesn't run in Visual Studio when you want to share data from another app with it. Then the sharing sidebar mysteriously will disappear automatically...

Related

Firefox Web Extension API - Get Downloads Folder

Is it possible using the web extensions API to get the default download folder for the current profile? I need to send it via native messaging to an external app.
I feel like https://developer.mozilla.org/en-US/Add-ons/WebExtensions/API/downloads should have it but it does't seem to.
Your best bet is probably making a dummy download and sending the "filename" property that you get back as a variable to your external app.
You would use the browser.downloads.onChanged event to get a reference to the filename value:
browser.downloads.onChanged.addListener(listener);
function listener(changed){
if(changed.filename != null){
// Do something
// Remove downloads.onChanged listener
browser.downloads.onChanged.removeListener(listener);
}
}
browser.downloads.download({url: dummyUrl});
https://developer.mozilla.org/en-US/Add-ons/WebExtensions/API/downloads/onChanged

Capture and save a photo in XAML/C# for a Windows store application

I'm trying to take and save a photo using a windows surface device.
I'm using the code below to take a photo and this work but I'd like to automatically create a directory on the device's local drive and save this photo there without any dialog prompts.
So the code I use to capture to photo is as follows:
CameraCaptureUI camera = new CameraCaptureUI();
StorageFile file = await camera.CaptureFileAsync(CameraCaptureUIMode.Photo);
if (file!=null)
{
using (IRandomAccessStream ras=await file.OpenAsync(FileAccessMode.Read))
{
BitmapImage source = new BitmapImage();
source.SetSource(ras);
imageBuildingPhoto.Source = source; // this is just an image control.
}
}
So after this I'd like to automatically save the photo to a new directory. e.g.
My Pictures\NewDirectory\Photo1.jpg
Anybody got any idea how I can do this?
This is a windows store application written using C#4.5 and XAML.
Thanks in advance
Use the CopyAsync method on the StorageFile object you get back (file). You can specify a directory and file name. If you need to create your own directory structure, you will need to enable access to the appropriate library in the Package Manifest then create it in code. You will then use the StorageFolder class and its CreateFolderAsync method to create folders.
http://aka.ms/30Days has some great resources for learning about scenarios like this. Might be worth checking out.
Your code will need to look to see if that folder exists and create it if it does not. Your app will need to declare the capability to access the user's Photos library in the app manifest, too.
To take a picture, your code is correct. I have a walkthrough in case you want to verify it against some other code: http://blog.jerrynixon.com/2012/10/walkthrough-capturing-photos-in-your.html
To interact with the file system, this can be tricky, but I have a longer write up on that if you want to reference it: http://blog.jerrynixon.com/2012/06/windows-8-how-to-read-files-in-winrt.html
The answer to your question is, yes you can. I have done it in my own apps. Now, it's just a matter of you implementing it in yours. You will find it to be pretty easy.

iOS custom email attachment handling in iOS6

My app allows for a text file to be emailed to other users with or without images and audio. When there's no images or audio, then the app sends the text file "as is" with a custom extension (e.g. text.xxx). When there's audio and images, the app zips the text file along with the images and audio into a file named something like text.xxx.zip.
Prior to iOS 6 this worked fine. Pre iOS 6, the app was given the URL of the text.xxx.zip file. Now, with iOS 6, it appears that the file is already unzipped by Mail, and only the text.xxx is sent to the openURL handler.
Does anyone out there have any experience with this? Suggestions for a better approach? I'm thinking I'll need to come up with a unique extension for the zipped case...
So after pounding my head on this for the last 24 hours or so, this is what "solves" the problem:
1) Change any existing attachment filenames from test.xxx.zip to test.zip. It looks iOS6 Mail is assuming that anything of the form filename.xxx.zip, is really just a compressed version of filename.xxx. (Maybe that's a standard somewhere?) Also noted that if I changed the filename to test.yyy.zip it then said it couldn't open the attachment. (presumably since no one registered for the ".yyy" extension).
2) Rewrite code to not use .zip extension in the future to prevent similar issues.
I also discovered that for multiple document types (e.g., .xxx, .zzz) you must specify a different mime type for each in the UTI declaration - otherwise Mail appends the first UTI extension it finds to the object and then calls openURL. So, in other words, if you're set up to handle a flat file (.xxx) and a zip file (.zzz), but you use the same mime type (e.g. application/myappname) and "xxx" is defined first, when openURL is called for your file "test.zzz", it will actually pass it to openURL as "test.xxx".

Unraveling the confusion about Embedded Resources

EDIT: Read answer number 1 from Tim Schmelter and then use this question for examples of how to embed resources and access them at runtime.
The subject of embedded resources comes up a lot, especially with people asking how to access the embedded files at runtime. Things get more confusing because Visual Studio gives you 2 different ways of embedding a resource, and different ways of accessing those resources at runtime. The problem is that depending on which method you used to embed the resource, the method you’re trying to use to access the file at runtime might not work. This post is an attempt to clear up all the confusion that I see out there, but I also have a question that nobody can seem to answer factually: Why is the size of my compiled program TWICE the size of the embedded resource (sometimes)? For example if I embed a 20MB file into my project, why does my program compile to 40MB? I haves asked this question in the past and nobody was able to reproduce my results. I found that the reason they were not able to reproduce was because they were embedding the file in a different way. See here:
Method 1:
Double-click on My Project to open the property pages and go to the Resources Tab. Now click Add Resource > Add Existing File. Browse to the file you want to embed. For this example I’m using an executable. You will now see your file on the Resources Tab:
You will also see that a folder named Resources was created under your project and the embedded file has been placed in this folder:
EDIT: THIS NEXT STEP WAS THE PROBLEM. TURNS OUT THAT WHEN YOU ADD A FILE VIA THE RESOURCES TAB YOU SHOULD NOT SET THE BUILD ACTION TO EMBEDDED RESOURCE. Counter intuitive to say the least!
Now with the file selected, look down at the properties window for the file and change the build action to Embedded Resource: (this step should ONLY be performed when you add a file via method 2).
Now compile your program. You will see that the size of your compiled program is at least double the size of your embedded resource. This does not happen with method 2. See here:
Method 2:
Right-click on your project name and choose Add > Existing Item. Browse to your file, and this time you will notice that while it was indeed placed under your project, there was no Resources folder created:
Now once again select the file and change the Build Action to Embedded Resource and compile. This time the size of the compiled program will be as you expected - about the size of the embedded file and not double the size as with method 1.
Which method you use to embed your file will determine which method you can use to access the file at runtime. For method 1 this is very simple, all you have to do is:
My.Computer.FileSystem.WriteAllBytes(Path, My.Resources.ResourceName, Append)
Where Path is the location and name for the file you want to save on the harddrive, ResourceName is the name of the embedded resource that you see in the project window (minus any extension), and Append is whether or not you want to create a new file or overwrite an existing file. So for example, using test.exe from the above images, I could save that file to the C drive like this:
My.Computer.FileSystem.WriteAllBytes(“C:\test.exe”, My.Resources.test, False)
Couldn’t be easier.
Method 2 however doesn’t appear to give you access to My.Resources so it gets a little more complicated. You have to create a Stream to hold the resource, put the stream into a byte array, then write the bytes out to the file system. The simplest way I have found to do this is like this:
Using s As Stream = Assembly.GetExecutingAssembly().GetManifestResourceStream(Project.ResourceName)
Dim bytes(s.Length) As Byte
s.Read(bytes, 0, bytes.Length)
File.WriteAllBytes(OutputFile, bytes)
End Using
With this method ResourceName must contain the file extension AND project name so using our example from above we can just do:
Using s As Stream = Assembly.GetExecutingAssembly().GetManifestResourceStream(WindowsApplication1.test.exe)
Dim bytes(s.Length) As Byte
s.Read(bytes, 0, bytes.Length)
File.WriteAllBytes(“C:\test.exe”, bytes)
End Using
Text-based files are a little different:
Dim output As String
Using sr As StreamReader = New StreamReader(Assembly.GetExecutingAssembly().GetManifestResourceStream(WindowsApplication1.test.txt))
output = sr.ReadToEnd()
End Using
Using sw As StreamWriter = New StreamWriter(“C:\test.txt”)
sw.Write(output)
End Using
Having struggled with this in the past I hope this will help someone. And if you think you can explain factually why method 1 of embedding a resource bloats my compiled program to double its size, I would really appreciate it.
I assume that Method 1 is adding the files twice.
http://www.vbdotnetforums.com/vb-net-general-discussion/42670-visual-basic-net-2008-get-resource-file-io-stream.html#post121923
At least that is the conclusion of the thread above.
Quote:
You went to the Resources page of the project properties and added the files there, right? You then went into the Solution Explorer and change the Build Action of the files to Embedded Resource, right? That's why you were doubling the file size: you were adding each file twice.
There are two different ways to add resources: on the Resources page of the project properties and in the Solution Explorer. You do NOT do both. If you want to use GetManifestResourcestream then you do NOT use the Resources page. You add the files to the project in the Solution Explorer manually, then you set the Build Action to Embedded Resource.
In future, do one or the other, not both.
Add a file to the Resources page of the project properties and then access it via My.Resources. This will automatically add the file to the project in the Solution Explorer but the Build Action will be None and it should be left that way.
Add the file to the project in the Solution Explorer by using Add New Item or Add Existing Item. Set the Build Action of the file to Embedded Resource and then access the resource using GetManifestResourceStream.
Just an update for anyone who wants to use this code. The code actually writes one additional byte to the file due to zero-based declaration of the byte array.
To get an exact copy of the original file change the code to:
Using s As Stream = Assembly.GetExecutingAssembly().GetManifestResourceStream(WindowsApplication1.test.exe)
Dim bytes(s.Length-1) As Byte
s.Read(bytes, 0, bytes.Length)
File.WriteAllBytes(“C:\test.exe”, bytes)
End Using

Does anyone know of a Cocoa/Obj-C library that can be used to gather application usage data

I would like to be able to gather info like how often certain windows are opened, what types of user data are accessed, how often menu items are clicked, etc. Does anyone know of a 3rd party (open source or commercial) Cocoa/Obj-C library or plugin that would allow me to gather this info?
I have used pinch media in the past, and they merged with Flurry. Library was simple to use and was setup in around 40 minutes.
I don't know any library for that but at least to get informed about when the user switches the front application you can install an event handler like this:
EventTypeSpec eventType;
eventType.eventClass = kEventClassApplication;
eventType.eventKind = kEventAppFrontSwitched;
EventHandlerUPP handlerUPP = NewEventHandlerUPP(FrontAppSwitchedDetector_callback);
OSStatus status=InstallApplicationEventHandler(handlerUPP,1,&eventType,self,&_eventHandlerRef);
... and when receiving an callback you may get the current front application process:
pascal OSStatus FrontAppSwitchedDetector_callback(EventHandlerCallRef nextHandler,EventRef theEvent,void* userData)
{
ProcessSerialNumber newSerial;
GetFrontProcess(&newSerial);
//to something with that ....
return (CallNextEventHandler(nextHandler, theEvent));
}