I am trying to create a UI test in Android Studio which will navigate through the various screens of my application and take screenshots when I tell it to.
I am new to Android Studio and Android programming in general; I have a decent understanding of XML and Java, but I don't know much about build files and I am not very good at using Android Studio, it seems.
I started this endeavor a couple weeks ago, and the first solution I tried was to use uiautomator. However, the documentation on that page (and seemingly just about everywhere else) is geared towards development with Eclipse, which I would like to avoid using for this project if possible.
The next thing I tried was Espresso. After I overcame some issues with implementing Espresso into my project, I was able to write tests with Espresso which would navigate through the screens of my application. However, unlike uiautomator, Espresso does not have built-in functionality to take screenshots at this time.
I first attempted to solve this problem of being unable to take screenshots with Espresso by writing custom code; as I'm still unfamiliar with Android, I wasn't really sure how to go about that, so I searched for help on the Internet (How to programmatically take a screenshot in Android?). However, I was unable to get the solutions I found to function from inside the test file.
Somebody recommended the usage of this tool: https://github.com/rtyley/android-screenshot-lib but I could not figure out how to import that into my project.
I eventually came back to uiautomator; I was still having a lot of trouble importing it into my project, and some people said that Robotium would help with that. I got Robotium to work, but I still could not import uiautomator.
It has been probably one month since I started using Android Studio, and in that time, I've had nothing but trouble simply getting the software to function properly. For the sake of brevity, I've omitted all the problems I have managed to solve on my own, but, to put it bluntly, I'm at the end of my patience.
TL;DR
If somebody could either:
-explain in the simplest possible way how to import uiautomator into an Android Studio project (I have read a lot of documentation about how to import external libraries into a project, but they all tell me to add a 'libs' folder to my project, but do not specify which type of folder to use [Java Resource Folder? Assets Folder? Module? etc.], and/or they tell me to go into Project Structure, select my app, go to dependencies, and choose "Import as Module," which does not work...)
OR
-explain how best to take a screenshot from inside of an Espresso test, including any instructions on how to import any required libraries
OR
-explain in detail some other way to create a UI test that can take screenshots...
...I would really appreciate it. I've spent days trying to figure out how to do this, and I am so frustrated. Many people have asked similar questions, but the answers are either too vague or the problems aren't close enough to my own.
Thanks!
Alright, after much trouble, I've found a very simple solution. It took me a very long time to work out, but if anyone else needs to do something similar, I'll put my conclusion here.
First of all, the testing framework that is easiest to use with Android Studio, it seems, is Espresso. Setting up Espresso is fairly simple; most of the instructions can be found here: https://code.google.com/p/android-test-kit/wiki/EspressoSetupInstructions Make sure you read it carefully -- it tells you basically everything you need to know, but I missed some important details and that caused me a lot of trouble.
If you browse around that Espresso site, it tells you just about everything you need to know about how to write Espresso tests. It was a little frustrating for me because, if I wrote a test and the test failed, my device would then have connection issues with my laptop and I would have to disconnect and reconnect the USB cord I was using. I think this had something to do with the fact that I was using a Nexus 7 with a Windows 8 laptop, which has given me some problems in other areas, so you may not encounter this issue yourself.
Now, unlike uiautomator, the documentation of which claims to have support for taking screenshots, Espresso does not have built-in support for taking screenshots. That means you'll have to figure out a different way to take screenshots. My solution was to create a new class (called HelperClass, in my case) inside my androidTest package and add this method to it.
public static void takeScreenshot(String name, Activity activity)
{
//slightly modified version of solution from http://www.ssaurel.com/blog/how-to-programmatically-take-a-screenshot-in-android/
//I added "/Pictures/" to my path because that's the folder where I wanted to store my screenshots -- you might not have that folder on your device, so you might want to replace "/Pictures/" with just "/" until you decide where you want to store the screenshots
String path = Environment.getExternalStorageDirectory().getAbsolutePath() + "/Pictures/" + name;
View v = activity.getWindow().getDecorView().getRootView();
v.setDrawingCacheEnabled(true);
Bitmap bitmap = Bitmap.createBitmap(v.getDrawingCache());
v.setDrawingCacheEnabled(false);
OutputStream out = null;
File imageFile = new File(path);
//the following line will help you find where the screen will be stored on your device
Log.v("Screenshot", "The image file path is " + imageFile.getPath());
try {
out = new FileOutputStream(imageFile);
// choose JPEG format
bitmap.compress(Bitmap.CompressFormat.JPEG, 90, out);
out.flush();
} catch (FileNotFoundException e) {
// manage exception
} catch (IOException e) {
// manage exception
} finally {
try {
if (out != null) {
out.close();
}
} catch (Exception exc) {
}
}
}
In order for this function to work, you will also have to add the following line to your manifest file.
<uses-permission android:name="android.permission.WRITE_EXTERNAL_STORAGE"/>
Without that, the function above will throw a FileNotFoundException every time you run it.
Finally, to call the takeScreenshot function from inside your Espresso test code, use this line (assuming you called your class HelperClass... if not, use the name of your class instead.
HelperClass.takeScreenshot("Whatever you want to call the file", getActivity());
Finding where your screenshots are stored can be a little difficult if you don't know where to look. I added a line of code to the takeScreenshot function that would print the filepath to LogCat, but I was using the file explorer on my computer to look for the screenshots on my Nexus (which was, of course, connected to the computer), and I couldn't find that path. However, I got a file explorer application on my tablet which made it very easy to find where the files were located in relation to everything else.
My solution may not be the simplest and it certainly isn't the best -- you'll fill your device up with screenshots before long if you aren't careful to delete the ones you don't need anymore, and I haven't got any idea how one would go about saving the screenshots directly to, say, a computer connected to the tablet via USB. That would certainly be helpful. However, if you really need a simple UI test that takes screenshots, and you're frustrated to no end like I was, this solution should probably help. I certainly found it useful.
I hope this helps somebody else -- it definitely solved my problems, at least for now.
Of course if you don't have all the restrictions that I did when I had to write a UI test that took screenshots, the other posts in this thread probably work much better.
You should give AndroidViewClient/culebra a try. Using culebra GUI, you can automatically generate a test case that interacts with your app and takes screenshot exactly when you indicate so.
I'm trying to automate uploading an image file to a server to which I don't have a better form of access (FTP, etc.). I'm using Node.js with Selenium and Chromedriver, and everything is going well until I need to simulate special keys. I've researched a bit, and have found nothing so far. Mac specific answers are ok for this.
I tried this from another answer somewhere, but it's not working for me:
driver.findElement(webdriver.By.name(uploadName))
.click()
.sendKeys(Keys.COMMAND + Keys.SHIFT + 'g')
.sendKeys(imgPath)
.sendKeys(Keys.RETURN);
It fails out because Keys is undefined, but I haven't run across any other ways to get at the commands.
After further research, it seems that this is the wrong way to do this. Selenium and Chromedriver don't really support OS level interaction, but there is an easier way.
driver.findElement(webdriver.By.name(uploadName)).sendKeys(imgPath);
This will target uploadName and then give it the file path passed into sendKeys. This bypasses the messy OS file dialogs.
I have created a powershell module (.psm1) file that includes a few other powershell scripts. We use it for sharepoint.
So basically, here's what happens:
I have a deploy script that retrieves the module location from the registry
It loads the module using the Import-Module cmdlet (using -force switch)
This module in turn loads the Sharepoint 2010 snap in and a few other scripts that I created
It runs runs a deployment script that references functions from the included scripts
It also runs a command line application and sends the output directly to the screen
The script will usually work the first time. However, after a few number of tries the commandline tool will stop working and sending output to the screen altogether. And if I try to run a commandline tool (not a cmd-let) after running my script, it don't worky anymore: no output, nothing is done. Its just the same as hitting enter on a blank prompt. anything powershell specific or running GUI applications will work fine but running any console application will not produce any concievable results. the only solution to this, is to just close my powershell and open it again. it will work for usually once and I will have to close it again. our users certainly wont be happy about that..
The most 'notable' things on the script:
scriptblocks are used extensively (for logging), a script block is sent to a handler that executes it using invokecommand and logs the step
its manipulating sharepoint objects
all objects are properly disposed of
no static variables are created nor changed
There are a few global variables shared across all scripts
What I have tried:
I striped my code to a bare minimum: loading an xml file, and restaring a few windows services but I'm still getting this intermittently. I have no idea which part of the code could cause this. I would love to post the code, but our company policy forbids me to. so my aplogies..
Update as per the comment below:
here's roughly how I use codeblocks. I have this function below that is used everytime I want to make the user aware of a task that I'm executing and what it outcome is.
function DoTask($someString, $scriptBlock, $param)
{
try
{
OutputTaskDescription $someString
InvokeCommand $scriptBlock -ArgumentList $param
OutputResultOK
}
catch
{
OutputResultError $_.tostring()
}
}
it could then be used like this:
$stringVar = "something"
$SpSite = new-spsite
deploySomething 'Deploying something' -param $spsite -ScriptBlock {
dosomethingToObject $stringvar
dosomethingToObject $spSite.Name
}
it would then output something like:
Deploying Something ------------- OK
Deploying Something ------------- ERROR
Also notice that I pass the $spsite in the argument list and I just use the string directly. I still don't understand how this works but it seems like I can access all primitive typed variables even without passing them as arguments but I have to pass more complex objects are params, else they dont have any value.
Update:
after much searching and days of pain. I have found others with the same pain. My code exhibits the same exact symptoms as described here:
http://connect.microsoft.com/PowerShell/feedback/details/496326/stability-problem-any-application-run-fails-with-lastexitcode-1073741502
I guess there is no solution yet to this problem.
After a little while I've noticed that if I've ran some very memory intensive functions, I too have gotten that behavior where everything you try to execute just goes to the prompt again. I'd recommend setting Set-PsDebug -Trace 2 to see what those functions are actually doing. I fixed my issue by doing this and figuring out how to make my functions more efficient.
I'm developing a COM server to be used from Excel VBA. When I update the server (edit code, unregister, re-register) Excel seems to carry on using the original version of the COM server, not the updated version. The only way I have found to get it to use the updated version is to close and re-open Excel, which gets a bit irritating. Is there a way to force Excel to use the newly registered version (maybe some kind of "clear cache" option)?
More details:
The server is being developed in Python using win32com.
In VBA I'm doing something like:
set obj=CreateObject("Foo.Bar")
obj.baz()
Where Foo.Bar is the COM server I have registered in the registry.
If I unregister the server then run the VBA code, I get a "can't create object" error from VBA, so it must realise that something is going on. But once I reregister it picks up the old version.
Any hints appreciated!
Thanks,
Andy
I've found a solution to my problem - the general idea is to set things up so that the main COM server class dynamically loads the rest of the COM server code when it is called. So in Python I've created a COM server class that looks something like:
import main_code
class COMInterface:
_public_methods_ = [ 'method1' ]
_reg_progid_ = "My.Test"
_reg_clsid_ = "{D6AA2A12-A5CE-4B6C-8603-7952B711728B}"
def methods(self, input1,input2,input3):
# force python to reload the code that does the actual work
reload(main_code)
return main_code.Runner().go(input1,input2,input3)
The main_code module contains the code that does the actual work and is reloaded each time the COM method is called. This works as long as the inputs don't change. There will presumably be a runtime penalty for this, so might want to remove the reload for the final version, but it works for development purposes.
Just a suggestion, have you tried to Unload the object? Maybe create a button in your excel spredsheat that force to unload the object.
I hope it helps
I am using the Scriptom extension to Groovy 1.7.0 to automate some processing using Excel 2007 under Windows XP.
This always seems to leave an Excel process running despite my calling quit on the excel activeX object. (There is a passing reference to this phenomenon in the Scriptom example documentation too.)
Code looks like:
import org.codehaus.groovy.scriptom.ActiveXObject;
def xls = new ActiveXObject("Excel.Application")
xls.Visible = true
// do xls stuff
xls.Quit()
The visible excel window does disappear but an EXCEL process is left in the task manager (and more processes pile up with each run of the script).
There are no error message or exceptions.
Can anyone explain why the Excel process is left behind and is there any way to prevent it from happening?
This works:
xls.Quit()
Scriptom.releaseApartment()
The javadocs state:
In some cases, the JVM can close
before everything is cleaned up, which
can leave automation servers
(especially Excel) hanging. Call this
before your script exits to get
correct behavior from automation
servers.
Looks like you are missing
xls.release();
like it is done here.