Is there any way to automate SFSafariViewController? I like the Xcode 7 UI test feature, but it seems it does not support SFSafariViewController automation. Some of the UI flows I am testing require a web browser so the app uses SFSafariViewController to make it safer vs a web view.
If it's similar to launching extensions (currently broken with direct interactions), try tapping the screen at the point where the element you're looking for is:
Example of tapping an action sheet which launches an extension:
func tapElementInActionSheetByPosition(element: XCUIElement!) {
let tableSize = app.tables.elementBoundByIndex(0).frame.size
let elementFrame = element.frame
// get the frame of the cancel button, because it has a real origin point
let CancelY = app.buttons["Cancel"].frame.origin.y
// 8 is the standard apple margin between views
let yCoordinate = CancelY - 8.0 - tableSize.height + elementFrame.midY
// tap the button at its screen position since tapping a button in the extension picker directly is currently broken
app.coordinateWithNormalizedOffset(CGVectorMake(elementFrame.midX / tableSize.width, yCoordinate / app.frame.size.height)).tap()
}
Note: You must tap at the XCUIApplication query layer. Tapping the element by position doesn't work.
For now Xcode 9.3 has support for that, but it doesn't work properly because of annoying Xcode bug.
In test you can print app.webViews.buttons.debugDescription or app.webViews.textFields.debugDescription, and it prints correct information, but after tap or typeText you have crash.
To workaround you can parse debugDescription, extract coordinates and tap by coordinate. For text field you can insert text via "Paste" menu.
private func coordinate(forWebViewElement element: XCUIElement) -> XCUICoordinate? {
// parse description to find its frame
let descr = element.firstMatch.debugDescription
guard let rangeOpen = descr.range(of: "{{", options: [.backwards]),
let rangeClose = descr.range(of: "}}", options: [.backwards]) else {
return nil
}
let frameStr = String(descr[rangeOpen.lowerBound..<rangeClose.upperBound])
let rect = CGRectFromString(frameStr)
// get the center of rect
let center = CGVector(dx: rect.midX, dy: rect.midY)
let coordinate = XCUIApplication().coordinate(withNormalizedOffset: .zero).withOffset(center)
return coordinate
}
func tap(onWebViewElement element: XCUIElement) {
// xcode has bug, so we cannot directly access webViews XCUIElements
// as workaround we can check debugDesciption, find frame and tap by coordinate
let coord = coordinate(forWebViewElement: element)
coord?.tap()
}
Article about that: https://medium.com/#pilot34/work-with-sfsafariviewcontroller-or-wkwebview-in-xcode-ui-tests-8b14fd281a1f
Full code is here: https://gist.github.com/pilot34/09d692f74d4052670f3bae77dd745889
Related
On my iOS app I use a Google Maps view with a GMSLayer, the transparency renders correctly on a real device but the background is white on Simulator, I can't make the screenshots for the App Store...
And this is the code I use :
// Implement GMSTileURLConstructor
// Returns a Tile based on the x,y,zoom coordinates, and the requested floor
let urls: GMSTileURLConstructor = {(x, y, zoom) in
let url = "http://wxs.ign.fr/[KEY]/geoportail/wmts?LAYER=TRANSPORTS.DRONES.RESTRICTIONS&FORMAT=image/png&SERVICE=WMTS&VERSION=1.0.0&REQUEST=GetTile&STYLE=normal&TILEMATRIXSET=PM&TILEMATRIX=" + String(zoom) + "&TILEROW="+String(y)+"&TILECOL="+String(x)
return URL(string: url)
}
// Create the GMSTileLayer
let layer = GMSURLTileLayer(urlConstructor: urls)
layer.zIndex = 0
layer.map = mapView
class TestTileLayer: GMSSyncTileLayer {
override func tileFor(x: UInt, y: UInt, zoom: UInt) -> UIImage {
let url = URL(string : "http://wxs.ign.fr/[KEY]/geoportail/wmts?LAYER=TRANSPORTS.DRONES.RESTRICTIONS&FORMAT=image/png&SERVICE=WMTS&VERSION=1.0.0&REQUEST=GetTile&STYLE=normal&TILEMATRIXSET=PM&TILEMATRIX=" + String(zoom) + "&TILEROW="+String(y)+"&TILECOL="+String(x))
return UIImage(data: try! Data(contentsOf: url!))!
}
}
I'm sure the image given by the URL uses transparency and I tried to manually change the white pixel to transparent but it doesn't change anything...
This is a simulator problem only it has been mentioned here.
However as #a.munzer suggested some work around you can take a screenshot from a real device and just edit it, shouldn't be that much of a problem.
I started out with the template project which you get when you choose ARKit project. As you run the app you can see the ship and view it from any angle.
However, once I allow camera control and tap on the screen or zoom into the ship through panning the ship gets stuck to camera. Now wherever I go with the camera the ship is stuck to the screen.
I went through the Apple Guide and seems like the don't really consider this as unexpected behavior as there is nothing about this behavior.
How to keep the position of the ship fixed after I zoom it or touch the screen?
Well, looks like allowsCameraControl is not the answer at all. It's good for SceneKit but not for ARKit(maybe it's good for something in AR but I'm not aware of it yet).
In order to zoom into the view a UIPinchGestureRecognizer is required.
// 1. Find the touch location
// 2. Perform a hit test
// 3. From the results take the first result
// 4. Take the node from that first result and change the scale
#objc private func handlePan(recognizer: UIPinchGestureRecognizer) {
if recognizer.state == .changed {
// 1.
let location = recognizer.location(in: sceneView)
// 2.
let hitTestResults = sceneView.hitTest(location, options: nil)
// 3.
if let hitTest = hitTestResults.first {
let shipNode = hitTest.node
let newScaleX = Float(recognizer.scale) * shipNode.scale.x
let newScaleY = Float(recognizer.scale) * shipNode.scale.y
let newScaleZ = Float(recognizer.scale) * shipNode.scale.z
// 4.
shipNode.scale = SCNVector3(newScaleX, newScaleY, newScaleZ)
recognizer.scale = 1
}
}
Regarding #2. I got confused a little with another hitTest method called hitTest(_:types:)
Note from documentation
This method searches for AR anchors and real-world objects detected by
the AR session, not SceneKit content displayed in the view. To search
for SceneKit objects, use the view's hitTest(_:options:) method
instead.
So that method cannot be used if you want to scale a node which is a SceneKit content
I use the following code to move the cursor position to 5 characters from the beginning of a UITextField:
txtView.selectedRange = NSMakeRange(5, 0);
Now, if I have a cursor at an arbitrary position as shown in the image below, how can I move the cursor up, down, left and right?
Left and right should be more or less easy. I guess the tricky part is top and bottom. I would try the following.
You can use caretRect method to find the current "frame" of the cursor:
if let cursorPosition = answerTextView.selectedTextRange?.start {
let caretPositionRect = answerTextView.caretRect(for: cursorPosition)
}
Then to go up or down, I would use that frame to calculate estimated position in UITextView coordinates using characterRange(at:) (or maybe closestPosition(to:)), e.g. for up:
let yMiddle = caretPositionRect.origin.y + (caretPositionRect.height / 2)
let lineHeight = answerTextView.font?.lineHeight ?? caretPositionRect.height // default to caretPositionRect.height
// x does not change
let estimatedUpPoint = CGPoint(x: caretPositionRect.origin.x, y: yMiddle - lineHeight)
if let newSelection = answerTextView.characterRange(at: estimatedUpPoint) {
// we have a new Range, lets just make sure it is not a selection
newSelection.end = newSelection.start
// and set it
answerTextView.selectedTextRange = newSelection
} else {
// I guess this happens if we go outside of the textView
}
I haven't really done it before, so take this just as a general direction that should work for you.
Documentation to the methods used is here.
I am creating a simple app with ARKit in which I add some text to the scene to the tapped position:
#objc func tapped(sender: UITapGestureRecognizer){
let sceneView = sender.view as! ARSCNView
let tapLocation = sender.location(in: sceneView)
let hitTest = sceneView.hitTest(tapLocation, types: .featurePoint)
if !hitTest.isEmpty{
self.addTag(tag: "A", hitTestResult: hitTest.first!)
}
else{
print("no match")
}
}
func addTag(tag: String, hitTestResult: ARHitTestResult){
let tag = SCNText(string:tag, extrusionDepth: 0.1)
tag.font = UIFont(name: "Optima", size: 1)
tag.firstMaterial?.diffuse.contents = UIColor.red
let tagNode = SCNNode(geometry: tag)
let transform = hitTestResult.worldTransform
let thirdColumn = transform.columns.3
tagNode.position = SCNVector3(thirdColumn.x,thirdColumn.y - tagNode.boundingBox.max.y / 2,thirdColumn.z)
print("\(thirdColumn.x) \(thirdColumn.y) \(thirdColumn.z)")
self.sceneView.scene.rootNode.addChildNode(tagNode)
}
It works, but I have problem with the orientation of the text. When I add it with the camera's original position, the text orientation is ok, I can see the text frontwise (Sample 1). But when I turn camera to the left / right, and add the text by tapping, I can see the added text from the side (Sample 2).
Sample 1:
Sample 2:
I know there should be some simple trick to solve it, but as a beginner in this topic I could not find it so far.
You want the text to always face the camera? SCNBillboardConstraint is your friend:
tagNode.constraints = [SCNBillboardConstraint()]
Am I correct in saying that you want the text to face the camera when you tap (wherever you happen to be facing), but then remain stationary?
There are a number of ways of adjusting the orientation of any node. For this case I would suggest simply setting the eulerAngles of the text node to be equal to those of the camera, at the point in which you instantiate the text.
In your addTag() function you add:
let eulerAngles = self.sceneView.session.currentFrame?.camera.eulerAngles
tagNode.eulerAngles = SCNVector3(eulerAngles.x, eulerAngles.y, eulerAngles.z + .pi / 2)
The additional .pi / 2 is there to ensure the text is in the correct orientation, as the default with ARKit is for a landscape orientation and therefore the text comes out funny. This applies a rotation around the local z axis.
It's also plausible (and some may argue it's better) to use .localRotate() of the node, or to access its transform property, however I like the approach of manipulating both the position and eulerAngles directly.
Hope this helps.
EDIT: replaced Float(1.57) with .pi / 2.
I have an App with Today Widget. So I would like to perform some UI testing on it.
I found a way to open Today/Notifications panel. It seems easy:
let statusBar = XCUIApplication().statusBars.elementBoundByIndex(0)
statusBar.swipeDown()
But then I can't find a way to do something useful. It is possible to record UI interactions in Today/Notifications panel, but such code can't reproduce my actions.
First you need to open Today View, you can use this way:
let app = XCUIApplication()
// Open Notification Center
let bottomPoint = app.coordinate(withNormalizedOffset: CGVector(dx: 0, dy: 2))
app.coordinate(withNormalizedOffset: CGVector(dx: 0, dy: 0)).press(forDuration: 0.1, thenDragTo: bottomPoint)
// Open Today View
let springboard = XCUIApplication(bundleIdentifier: "com.apple.springboard")
springboard.scrollViews.firstMatch.swipeRight()
Then, to access everything you need, just use springboard, for example:
let editButton = springboard.buttons["Edit"]
There's a similar problem testing extensions. I've found that what you must do is tap the element at where it is on the screen rather than the element itself in order to drive the interaction. I haven't tested this with your scenario, but I haven't found anything un-tappable via this method yet.
Here is a Swift example of tapping the "X" button on the Springboard for an app icon, which similarly cannot be tapped via typical interaction:
let iconFrame = icon.frame // App icon on the springboard
let springboardFrame = springboard.frame // The springboard (homescreen)
icon.pressForDuration(1.3) // tap and hold
// Tap the little "X" button at approximately where it is. The X is not exposed directly
springboard.coordinateWithNormalizedOffset(CGVectorMake((iconFrame.minX + 3) / springboardFrame.maxX, (iconFrame.minY + 3) / springboardFrame.maxY)).tap()
By getting the frame of the superview and the subview, you can calculate where on the screen the element should be. Note that coordinateWithNormalizedOffset takes a vector in the range [0,1], not a frame or pixel offset. Tapping the element itself at a coordinate doesn't work, either, so you must tap at the superview / XCUIApplication() layer.
More generalized example:
let myElementFrame = myElement.frame
let appFrame = XCUIApplication().frame
let middleOfElementVector = CGVectorMake(iconFrame.midX / appFrame.maxX, iconFrame.midY / appFrame.maxY)
// Tap element from the app-level at the given coordinate
XCUIApplication().coordinateWithNormalizedOffset(middleOfElementVector).tap()
If you need to access the Springboard layer and go outside your application, you can do so with:
let springboard = XCUIApplication(privateWithPath: nil, bundleID: "com.apple.springboard")
springboard.resolve()
But you'll need to expose some private XCUITest methods with Objective-C:
#interface XCUIApplication (Private) {
- (id)initPrivateWithPath:(id)arg1 bundleID:(id)arg2;
}
#interface XCUIElement (Private) {
- (void) resolve;
}