Testing screen tracking with UI automation on iOS - ios

So I had this idea to test the implementation of my screen tracking (with Google Analytics) on my app using UI automation.
The original idea was to build a UI script to go through the screens while checking if the tracking events are being sent accordingly. I need this as sometimes I'm not able to compose everything out of view controllers or the events are not forwarded in the expected order. Regardless of that, I should test this aspect of my app as well and I thought that UI automation was the answer.
I have implemented a script to go through the screens using the UI automation instrument and this is working correctly. I even went so far as using tuneup js to make the code more streamlined and easier to follow.
I was expecting to have something like (in general terms, the syntax is only a simplification):
Being on screen X
Tap button A
Expect screen Y and tracking event for the screen Y
However, as far as I was able to check, testing the screen tracking is something that is not possible with the UI automation.
Or am I missing something?
I thought of creating an invisible view that stays on top of all the view hierarchy and changing its name every time a new screen is loaded to allow me to test it with UI automation but the idea sounded a little over the top...
What do you people suggest? Look for another UI automation tool? Do it with unit testing instead?
Thanks in advance for any help

You could use a UIAlertView and inspect those alerts. Instead of sending the analytics events you can pop up the alert so you can check on it in UIAutomation.
Analytics abstraction frameworks like AnalyticsKit provide an easy way to change the analytics provider. And AnalyticsKit even has an example for that (take a look at the AnalyticsKitDebugProvider class). So the changes to your production code are minimal.
You could use a build configuration where you set a build variable to control the initialization of your analytics
id<AnalyticsKitProvider> provider
#ifdef USE_UI_AUTOMATION_ANALYTICS
provider = [[TestAutomationProvider alloc] init];
#else
provider = [[RealProvider alloc] initWithApiKey:API_KEY];
#endif
[AnalyticsKit initializeLoggers:#[provider]];
In UIAutomation you can test for the alert coming up. You can utilize assertions.js out of the tuneup.js package to write a function like this
function checkForAlert()
{
var alert = null;
retry( function() {
log("wait until alert appaers");
alert = UIATarget.localTarget().frontMostApp().alert();
assertNotNull(alert, "No alert found");
assertTrue("The name you can choose for the alert" == alert.name());
}, 5, 1.0);
return alert;
};
This combines waiting for the alert and testing if it finally appear. If the alert not appears, the test will fail.
In your test you use this in the following way:
var analyticAlert = checkForAlert() // if alert appears it will be in the var, otherwise the test fails at this point.
analyticAlert.buttons()["OK"].tap(); // dismiss the alert
To make this work you also need to set an onAlert handler. Otherwise UIAutomation would try to dismiss your alert immediately. This has to be done before your tests code. Alert handling is explained in the UIAutomation docs.
function MyOnAlertHandler(alert)
{
if("The name you choose"==alert.name()) // filter all alerts created by analytics provider
{
return true; // handle alert in your test
}
return false // automaticly dismiss all other
}
UIATarget.onAlert = MyOnAlertHandler; // set the alert handler

Related

Firebase A/B test not counting users when activation event is used on iOS

We're using the current version of the Firebase iOS framework (5.9.0) and we're seeing a strange problem when trying to run A/B test experiments that have an activation event.
Since we want to run experiments on first launch, we have a custom splash screen on app start that we display while the remote config is being fetched. After the fetch completes, we immediately activate the fetched config and then check to see if we received info about experiment participation to reconfigure the next UI appropriately. There are additional checks done before we determine that the current instance, in fact, should be part of the test, thus the activation event. Basically, the code looks like:
<code that shows splash>
…
[[FIRRemoteConfig remoteConfig] fetchWithExpirationDuration:7 completionHandler:^(FIRRemoteConfigFetchStatus status, NSError * _Nullable error) {
[[FIRRemoteConfig remoteConfig] activateFetched];
if (<checks that see if we received info about being selected to participate in the experiment and if local conditions are met for experiment participation>) {
[FIRAnalytics logEventWithName:#"RegistrationEntryExperimentActivation" parameters:nil];
<dismiss splash screen and show next UI screen based on experiment variation received in remote config>
} else {
<dismiss splash screen and show next UI screen>
}
}
With the approach above (which is completely straight-forward IMO) does not work correctly. After spending time with the debugger and Firebase logging enabled I can see in the log that there is a race-condition problem occurring. Basically, the Firebase activateFetched() call does not set up a "conditional user property experiment ID" synchronously inside the activateFetched call but instead sets it up some short time afterward. Because of this, our firing of the activation event immediately after activateFetched does not trigger this conditional user property and subsequent experiment funnel/goal events are not properly marked as part of an experiment (the experiment is not even activated in the first place).
If we change the code to delay the sending of the activation event by some arbitrary delay:
<code that shows splash>
…
[[FIRRemoteConfig remoteConfig] fetchWithExpirationDuration:7 completionHandler:^(FIRRemoteConfigFetchStatus status, NSError * _Nullable error) {
[[FIRRemoteConfig remoteConfig] activateFetched];
if (<checks that see if we received info about being selected to participate in the experiment and if local conditions are met for experiment participation>) {
dispatch_after(dispatch_time(DISPATCH_TIME_NOW, (int64_t)(0.5 * NSEC_PER_SEC)), dispatch_get_main_queue(), ^{
[FIRAnalytics logEventWithName:#"RegistrationEntryExperimentActivation" parameters:nil];
<dismiss splash screen and show next UI screen based on experiment variation received in remote config>
}
} else {
<dismiss splash screen and show next UI screen>
}
}
the conditional user property for the experiment gets correctly setup beforehand and triggered by the event (causing experiment activation and subsequent events being correctly marked as part of the experiment).
Now, this code obviously is quite ugly and prone to possible race-conditions. The delay of 0.5 seconds is conservatively set to hopefully be enough on all iOS devices but ¯_(ツ)_/¯. I've read the available documentation multiple times and tried looking at all available API methods with no success in figuring out what the correct point of starting to send events should be. If the activateFetched method uses an asynchronous process of reconfiguring internal objects, one would expect a callback method that indicates to the caller the point in time when everything is done reconfiguring and ready for further use by the application. Seems the framework engineers didn't anticipate a use-case when someone needs to send the activation event immediatly after remote config profile activation…
Has anyone else experienced this problem? Are we missing something in the API? Is there a smarter way of letting activateFetched finish its thing?
Hope some Firebase engineers can chime-in with their wisdom as well :)
Thanks

Why does this Spectron code block my Electron app after the first click?

This is the first time I'm trying to create automated tests for an Electron app using Spectron. It might be my rusty knowledge of async programming but I don't know why the code below is misbehaving:
it ('should allow me to create an account', function() {
return app.client
.waitUntilWindowLoaded()
.waitForExist('//a[text()="Create Free Account"]')
.click('//a[text()="Create Free Account"]')
.waitForExist('//button[text()="Create Account"]')
.setValue('#Email', "test#test.com")
.setValue('#Password', "Password1!")
.click('//button[text()="Create Account"]')
.waitForExist('//p[contains(text(),"Almost done.")]')
});
The test seems to get as far as the first click(), then it should wait for the App to request a new page, eventually displaying a "Create Account" button. However, for some reason, the app itself seems to block at this point. I know the click is occurring. When I try it manually, the app behaves properly.
Mark
I think you need to wait till the next page loads
Also chain
.pause(3*1000)
after clicking the button
(Or)
.waitUntilWindowLoaded(3*1000)

Xamarin UI test fail to identify System dialog(s) of iOS App

I want to automate our iOS app using xamarin UI test. To access the app, user has to give his/her TouchID. But both Xamarin Test recorder and Repl() failed to identify the Touch ID system dialog (Please see attached screenshot).
This dialog is not developed by our developers and it is a system dialog.
How do I perform the action(s) which associated with this Touch ID?
Thanks
System dialogs is not something you can query with calabash or Xamarin.UITest, which is super annoying.
You would need to use a backdoor to simulate a success or fail state for this. You can read more about these here:
https://developer.xamarin.com/guides/testcloud/uitest/working-with/backdoors/
Basically you define a method in your AppDelegate:
[Export("touchIdBackdoor:")] // notice the colon at the end of the method name
public NSString TouchIdBackdoor(NSString value)
{
if (value == "true")
{
//simulate ok finger press
}
else
{
//simulate failed finger
}
}
Then in your test, when you expect the touch id to appear, you invoke the backdoor:
app.Invoke("touchIdBackdoor:", "true");
To dismiss the touch id and set whatever you need to continue.

Is it possible to dismiss System Alerts using EarlGrey (iOS UI Testing)?

I'm beginning to experiment with EarlGrey a little bit, having done UI Testing using the XCUITest for some months now. I'm running into the classic problem of being unable to dismiss system alerts, which is strange as it looks as though Google implemented a matcher for system alerts called grey_systemAlertViewShown(). I'm trying to detect system alerts using GREYCondition. Here's what I've tried:
- (void)waitForAndDismissSystemAlertForSeconds:(NSInteger)seconds {
GREYCondition *interactableCondition = [GREYCondition conditionWithName:#"isInteractable" block:^BOOL{
// Fails if element is not interactable
NSError *error;
[[EarlGrey selectElementWithMatcher:grey_systemAlertViewShown()] assertWithMatcher:grey_interactable() error:&error];
if (error) {
return NO;
} else {
NSError *allowButtonError;
[[EarlGrey selectElementWithMatcher:grey_accessibilityLabel(#"Allow")] assertWithMatcher:grey_notNil() error:&allowButtonError];
if (!allowButtonError) {
[[EarlGrey selectElementWithMatcher:grey_accessibilityLabel(#"Allow")] performAction:grey_tap()];
}
return YES;
}];
[interactableCondition waitWithTimeout:seconds];
}
I've also tried using addUIInterruptionMonitorWithDescription as described here (but using EarlGrey code to do basically what I am doing above in the interruption monitors): Xcode 7 UI Testing: how to dismiss a series of system alerts in code
Neither approach works. Breakpoints don't fire for the non-error case in my GREYCondition, and the interruption monitor doesn't dismiss my alert either.
Does anybody know if EarlGrey supports dismissing system alerts?
As the docs for grey_systemAlertViewShown indicate, grey_systemAlertViewShown merely checks to see if system alert views are shown. A better usage of the API would be to assert that system alert is not shown (maybe because the test app has mocked out the code that causes system alerts).
Code that taps a button that requests causes system alert to be shown (for ex: requests user's geo location) comes here...
// Assert that in the test app system alert view is not shown because we have mocked out the part of code that requests user location.
[[EarlGrey selectElementWithMatcher:grey_anything()] assertWithMatcher:grey_not(grey_systemAlertViewShown())];
As of this writing system alert views cannot be dismissed by EarlGrey. Alertviews that are launched by the app can be dismissed. The FAQ has a question that indicates that EarlGrey tests will fail if modal dialogs are present.
The best way we found to get around this was to include a launch argument for testing, in which we wouldn't register the app for notifications.
Something like:
if [[[NSProcessInfo processInfo] arguments] containsObject:argument] {
return;
}
before you call
[[UIApplication sharedApplication] registerUserNotificationSettings:settings];
[[UIApplication sharedApplication] registerForRemoteNotifications];
This way, the "Do you want to allow push notifications..." alert won't show.
It is possible to grant all the required permissions using AppleSimulatorUtils util.
This approach eliminates the need to dismiss alerts and saves time.
Install an util by entering next commands in the Terminal app
brew tap wix/brew
brew install applesimutils
And grant permission(s) with
applesimutils --byId <simulator UDID> --bundle <bundle identifier> --setPermissions "notifications=YES"
For more info and examples please refer to
https://github.com/wix/AppleSimulatorUtils
EarlGreyImpl.invoked(fromFile: #file, lineNumber: #line).selectElement(with: grey_text("Click")).perform(grey_tap())
//use above code it might work for your problem

Xcode 7 UI Testing: how to dismiss a series of system alerts in code

I am writing UI test cases using the new Xcode 7 UI Testing feature. At some point of my app, I ask the user for permission of camera access and push notification. So two iOS popups will show up: "MyApp Would Like to Access the Camera" popup and "MyApp Would Like to Send You Notifications" popup. I'd like my test to dismiss both popups.
UI recording generated the following code for me:
[app.alerts[#"cameraAccessTitle"].collectionViews.buttons[#"OK"] tap];
However, [app.alerts[#"cameraAccessTitle"] exists] resolves to false, and the code above generates an error: Assertion Failure: UI Testing Failure - Failure getting refresh snapshot Error Domain=XCTestManagerErrorDomain Code=13 "Error copying attributes -25202".
So what's the best way of dismissing a stack of system alerts in test? The system popups interrupt my app flow and fail my normal UI test cases immediately. In fact, any recommendations regarding how I can bypass the system alerts so I can resume testing the usual flow are appreciated.
This question might be related to this SO post which also doesn't have an answer: Xcode7 | Xcode UI Tests | How to handle location service alert?
Thanks in advance.
Xcode 7.1
Xcode 7.1 has finally fixed the issue with system alerts. There are, however, two small gotchas.
First, you need to set up a "UI Interuption Handler" before presenting the alert. This is our way of telling the framework how to handle an alert when it appears.
Second, after presenting the alert you must interact with the interface. Simply tapping the app works just fine, but is required.
addUIInterruptionMonitorWithDescription("Location Dialog") { (alert) -> Bool in
alert.buttons["Allow"].tap()
return true
}
app.buttons["Request Location"].tap()
app.tap() // need to interact with the app for the handler to fire
The "Location Dialog" is just a string to help the developer identify which handler was accessed, it is not specific to the type of alert.
I believe that returning true from the handler marks it as "complete", which means it won't be called again. For your situation I would try returning false so the second alert will trigger the handler again.
Xcode 7.0
The following will dismiss a single "system alert" in Xcode 7 Beta 6:
let app = XCUIApplication()
app.launch()
// trigger location permission dialog
app.alerts.element.collectionViews.buttons["Allow"].tap()
Beta 6 introduced a slew of fixes for UI Testing and I believe this was one of them.
Also note that I am calling -element directly on -alerts. Calling -element on an XCUIElementQuery forces the framework to choose the "one and only" matching element on the screen. This works great for alerts where you can only have one visible at a time. However, if you try this for a label and have two labels the framework will raise an exception.
Objective - C
-(void) registerHandlerforDescription: (NSString*) description {
[self addUIInterruptionMonitorWithDescription:description handler:^BOOL(XCUIElement * _Nonnull interruptingElement) {
XCUIElement *element = interruptingElement;
XCUIElement *allow = element.buttons[#"Allow"];
XCUIElement *ok = element.buttons[#"OK"];
if ([ok exists]) {
[ok tap];
return YES;
}
if ([allow exists]) {
[allow tap];
return YES;
}
return NO;
}];
}
-(void)setUp {
[super setUp];
self.continueAfterFailure = NO;
self.app = [[XCUIApplication alloc] init];
[self.app launch];
[self registerHandlerforDescription:#"“MyApp” would like to make data available to nearby Bluetooth devices even when you're not using app."];
[self registerHandlerforDescription:#"“MyApp” Would Like to Access Your Photos"];
[self registerHandlerforDescription:#"“MyApp” Would Like to Access the Camera"];
}
Swift
addUIInterruptionMonitorWithDescription("Description") { (alert) -> Bool in
alert.buttons["Allow"].tap()
alert.buttons["OK"].tap()
return true
}
Gosh.
It always taps on "Don't Allow" even though I deliberately say tap on "Allow"
At least
if app.alerts.element.collectionViews.buttons["Allow"].exists {
app.tap()
}
allows me to move on and do other tests.
For the ones who are looking for specific descriptions for specific system dialogs (like i did) there is none :) the string is just for testers tracking purposes. Related apple document link : https://developer.apple.com/documentation/xctest/xctestcase/1496273-adduiinterruptionmonitor
Update : xcode 9.2
The method is sometimes triggered sometimes not. Best workaround for me is when i know there will be a system alert, i add :
sleep(2)
app.tap()
and system alert is gone
God! I hate how XCTest has the worst time dealing with UIView Alerts. I have an app where I get 2 alerts the first one wants me to select "Allow" to enable locations services for App permissions, then on a splash page the user has to press a UIButton called "Turn on location" and finally there is a notification sms alert in a UIViewAlert and the user has to select "OK". The problem we were having was not being able to interact with the system Alerts, but also a race condition where behavior and its appearance on screen was untimely. It seems that if you use the alert.element.buttons["whateverText"].tap the logic of XCTest is to keep pressing until the time of the test runs out. So basically keep pressing anything on the screen until all the system alerts are clear of view.
This is a hack but this is what worked for me.
func testGetPastTheStupidAlerts() {
let app = XCUIApplication()
app.launch()
if app.alerts.element.collectionViews.buttons["Allow"].exists {
app.tap()
}
app.buttons["TURN ON MY LOCATION"].tap()
}
The string "Allow" is completely ignored and the logic to app.tap() is called evreytime an alert is in view and finally the button I wanted to reach ["Turn On Location"] is accessible and the test pass
~Totally confused, thanks Apple.
The only thing I found that reliably fixed this was to set up two separate tests to handle the alerts. In the first test, I call app.tap() and do nothing else. In the second test, I call app.tap() again and then do the real work.
On xcode 9.1, alerts are only being handled if the test device has iOS 11. Doesn't work on older iOS versions e.g 10.3 etc. Reference: https://forums.developer.apple.com/thread/86989
To handle alerts use this:
//Use this before the alerts appear. I am doing it before app.launch()
let allowButtonPredicate = NSPredicate(format: "label == 'Always Allow' || label == 'Allow'")
//1st alert
_ = addUIInterruptionMonitor(withDescription: "Allow to access your location?") { (alert) -> Bool in
let alwaysAllowButton = alert.buttons.matching(allowButtonPredicate).element.firstMatch
if alwaysAllowButton.exists {
alwaysAllowButton.tap()
return true
}
return false
}
//Copy paste if there are more than one alerts to handle in the app
#Joe Masilotti's answer is correct and thanks for that, it helped me a lot :)
I would just like to point out the one thing, and that is the UIInterruptionMonitor catches all system alerts presented in series TOGETHER, so that the action you apply in the completion handler gets applied to every alert ("Don't allow" or "OK"). If you want to handle alert actions differently, you have to check, inside the completion handler, which alert is currently presented e.g. by checking its static text, and then the action will be applied only on that alert.
Here's small code snippet for applying the "Don't allow" action on the second alert, in series of three alerts, and "OK" action on the remaining two:
addUIInterruptionMonitor(withDescription: "Access to sound recording") { (alert) -> Bool in
if alert.staticTexts["MyApp would like to use your microphone for recording your sound."].exists {
alert.buttons["Don’t Allow"].tap()
} else {
alert.buttons["OK"].tap()
}
return true
}
app.tap()
This is an old question but there is now another way to handle these alerts.
The system alert isn't accessibly from the app context of the app you are launched in, however you can access the app context anyway. Look at this simple example:
func testLoginHappyPath() {
let app = XCUIApplication()
app.textFields["Username"].typeText["Billy"]
app.secureTextFields["Password"].typeText["hunter2"]
app.buttons["Log In"].tap()
}
In a vacuum with a simulator already launched and permissions already granted or denied, this will work. But if we put it in a CI pipeline where it gets a brand new simulator, all of the sudden it won't be able to find that Username field because there's a notification alert popping up.
So now there's 3 choices on how to handle that:
Implicitly
There's already a default system alert interrupt handler. So in theory, simply trying to typeText on that first field should check for an interrupting event and handle it in the affirmative.
If everything works as designed, you won't have to write any code but you'll see an interruption logged and handled in the log, and your test will take a couple seconds more.
Explicitly via interruptionmonitor
I won't rewrite the previous work on this, but this is where you explicitly set up an interruptionmonitor to handle the specific alert being popped up - or whatever alerts you expect to happen.
This is useful if the built-in handler doesn't do what you want - or doesn't work at all.
Explicitly via XCUITest framework
In xCode 9.0 and above, you can switch between app contexts fluidly by simply defining multiple XCUIApplication() instances. Then you can locate the field you need via familiar methods. So to do this explicitly would look like the following:
func testLoginHappyPath() {
let app = XCUIApplication()
let springboardApp = XCUIApplication(bundleidentifier: "com.apple.springboard")
if springboardApp.alerts[""FunHappyApp" would like permission to own your soul."].exists {
springboardApp.alerts.buttons["Allow"].tap()
}
app.textFields["Username"].typeText["Billy"]
app.secureTextFields["Password"].typeText["hunter2"]
app.buttons["Log In"].tap()
}
Sounds like the approach to implementing camera access and notifications are threaded as you say, but not physically managed and left to chance when and how they are displayed.
I suspect one is triggered by the other and when it is programatically clicked it wipes out the other one as well (which Apple would probably never allow)
Think of it you're asking for a users permission then making the decision on their behalf? Why? Because you can't get your code to work maybe.
How to fix - trace where these two components are triggering the pop up dialogues - where are they being called?, rewrite to trigger just one, send an NSNotification when one dialogue has been completed to trigger and display the remaining one.
I would seriously discourage the approach of programatically clicking dialogue buttons meant for the user.

Resources