User Profile

Everything was working fine for 4 years. But since Ventura, when the WKWebView is in focus, the keyDown event are not happening. Not only they don't fire inside of web page input box, but they won't register in the app either. As long as a WKWebView is in focus, the keyboard events won't work. The app does not receive the key events, and the web page will not receive the js events keyDown and friends. This is particularly painful with auto complete input box. Test with an empty project with only a WKWebView with this page https://www.toptal.com/developers/keycode/for/d It does work fine on iPhone and iPad. Message to Apple : You force us to use your Safari engine, yet it is always broken. Don't wait for 8 months to fix it this time, we would be tempted to go full web and skip native app with all the store certification problems.

If we click on a simple html dropdown list on a WKWebview and choose an item in the list, after that we will never be able to click it ever again on that same dropdown (it will never open again), unless we click elsewhere on the webpage (removing the focus that was on the dropdown). That only happens on Mac, on iOS it works fine. This is another Mac Catalyst problem on Mac.

We released our app 2 years ago, it is available on iPhone, iPad and macOS (with catalyst). Since macOS Monterey, the WKWebview is crashing at random, for different customers it will be a different thing that triggers the crash. It could be a specific button or simply one text box inside a web page and it will be that specific thing all the time that will crash the app from now on. For different customers it will be a different trigger. It could be anything. The problem is exactly the same on Intel and M1. Only a restart (or multiple) or a shut down can resolve the problem, but not for everyone, for some people it will be one specific button on one of our 5 000 web pages that will start crashing the app and nothing will ever make it work again. Of course I cleared the cashes programmatically, even deleted the files and uninstalled the app. The crash dump is mostly the same and in my crash log, there is not a single line of code belonging to my code. The crash is really inside the Safari engine or the WKWebview. There are other problems with WKWebview since Monterey : Pasting text Cmd+V in web page paste the text twice and dropdown can only be used once, after that they never open again unless you click on an empty space elsewhere. This is on any web pages, not just ours. Of course none of these problems are happening on iPhone and iPad and they use the exact same code (macCatalyst). Please Apple take the macOS seriously or discontinue the product. Crash Dump Application Specific Information: *** Terminating app due to uncaught exception 'NSInternalInconsistencyException', reason: 'Invalid parameter not satisfying: <SPRoundedWindow: 0x154e8f060>. "frame=!CGRectIsNull(frame)"' terminating with uncaught exception of type NSException abort() called Application Specific Backtrace 0: 0 CoreFoundation 0x00000001c404eb08 __exceptionPreprocess + 240 1 libobjc.A.dylib 0x00000001c3d99e14 objc_exception_throw + 60 2 Foundation 0x00000001c4fc4c80 -[NSMutableDictionary(NSMutableDictionary) initWithContentsOfFile:] + 0 3 AppKit 0x00000001c6b7fbec -[NSWindow _reallySetFrame:] + 900 4 AppKit 0x00000001c6b7f4ac -[NSWindow _oldPlaceWindow:fromServer:] + 228 5 AppKit 0x00000001c6b7e490 -[NSWindow _setFrameCommon:display:fromServer:] + 2220 6 SafariPlatformSupport 0x00000001eaf26f08 __85-[SPSafariPlatformSupport displayOTPAutoFillRelativeToRect:ofView:completionHandler:]_block_invoke + 300

In my WKWebview, if I right click (macCatalyst) or long press (ios) paste, it only paste text once. But if I Cmd-V to paste, it will paste text twice with a space between the 2 copied text. This is my log (canPerform action always return true): -canPerformAction() action : paste: -canPerformAction() action : paste: -canPerformAction() action : paste: -canPerformAction() action : paste: -canPerformAction() action : newWindowForTab: -2021-12-17 09:50:17.101763-0500 myApp[18815:1434268] *** Assertion failure in -[UINSResponderProxy _initWithWrappedResponder:orMenuProxy:forAction:sender:], UINSResponderProxy.m:250 -[UINSResponderProxy _initWithWrappedResponder:orMenuProxy:forAction:sender:]: UINSResponderProxy is improperly wrapping a responder that does not respond to the action. -canPerformAction() action : paste: -canPerformAction() action : paste: -canPerformAction() action : paste:

I create a new URLRequest(url: myUrl) object every time before calling URLSession.shared.dataTask(with: urlRequest). I wait for one request to complete before making another call, so no timing problem possible. If I set it to false, it does not inject cookies. If I set it to true, it does send cookies. So everything is fine? NO! If you sometimes set it to false for some calls and to true for other calls it will always not inject cookies. So if you set it to false, even if you set it to true for other calls it will never ever work. I don't know how they manage to create such a bug, but it require some special kind of evil intentions. Note : I always call the exact same domain and I always have my cookies.

I have a macCatalyst app. I can see the WKWebview pages in the Develop tab of Safari on my Mac when the app is started from Xcode. I can see the WKWebview pages in the Develop tab of Safari on my iPad when started on the device (not started by Xcode) If I start the same app on my Mac but normally this time (not from Xcode), all I get is "No Inspectable Applications". I have set com.apple.security.get-task-allow to 1 like other internet post has suggested, still not working. Please help me, my web colleagues need to be able to debug their webpages in the context of my app running on their mac. My Xcode and macOS is the latest version.

let ciDetector = CIDetector(ofType: CIDetectorTypeRectangle, context: nil, options: [CIDetectorAccuracy: CIDetectorAccuracyHigh]) The code worked for 2 years on macCatalyst. After the Big Sur update it stopped working and now always return nil. On iOS it is always working as expected. If I use CIDetectorTypeText instead, it won't be nil. I suspect Apple broke the CIDetectorTypeRectangle on Big Sur