That did it! I... never would've guessed that, from the error.
Post not yet marked as solved
Hey, greay, did you ever find a solution for your problem? I'm having a similar problem with an app I'm working on where I have a one-and-done binding that functions as expected the first time, but then doesn't fire again upon changes. Any guidance would be greatly appreciated. :) No, I ended up moving completely away from SwiftUI because I need to support older versions of macOS.
Post not yet marked as solved
I should add/clarify that both of these apps are already for sale on the store. We are NOT trying to change anything under "App Information" – like the localizable name. We want that to stay as-is. All we're trying to do now is add localized copy for the stuff in the update: What's new, promotional text, description, etc.
Post not yet marked as solved
Hmm, ok, that suggests to me that the problem isn't in this construction itself, but something further up.As I hinted, but should've made more explicit, the [Check] is a property on another struct. The larger structure has some pretty deeply-nested subviews, too much to put in a forum post, so I was hoping this small snippet would be enough to highlight what was going wrong. It wasn't – but since this works, I'm thinking I might be able to slowly grow out from this & poinpoint the problem.
Post not yet marked as solved
It's in an existing application, so the change actually breaks some functionality.In the app on the store, certain on-screen buttons are disabled when the keyboard appears. They're re-enabled when you press submit, dismissing the keyboard. Since the notifications are no longer sent, those buttons never get re-enabled.Basically, the text entry is a modal operation, and the existing code relies on those notifications to exit that mode. Might not be the best design, but I didn't write the original code.
Post not yet marked as solved
But it's not always visible. I would still expect it to send the will/did show notifications when it appears. And send the will/did hide notifications when it's dismissed. If this is in fact by design, it should be documented.Having to check whether the starting/ending frame's bounds is zero seems incredibly backwards to me.
