foos | FOOS 's Overcoming Operating System

I recently started writing my first SwiftUI app and ran into an design issue, which can be demonstrated by the simple code below. I know what the issue is, but I wonder why it occurred and what's the recommended way to address it.

In the demo, there are two views. The first is a list view, the second is a detail view. The detail view contains a button to delete the item being displayed. Clicking on the button crashes the app due to the forced unwrapping in the detail view. I didn't expect the crash, because from my understanding when the data model changes, SwiftUI should regenerate the entire view hierarchy. That is, it calls ContentView.body, which calls FooListView.body, which goes through items in the data model and creates NavigationLink for each item.

Since the data model has been changed, there is only one item left. So I don't think SwiftUI would create FooDetailView (or call its body) for the item deleted. If so, how come the FooDetailView code crashed? I tried to debug the code but didn't find much useful information. I believe the FooDetailView that crashed the app is the one that contained the deleted item. That I don't understand. Since SwiftUI regenerates view hierarchy, how could that old view left uncleaned up?

Can anyone explain a bit how you understand it? And how do you address the issue? I currently think out two ways. The first is to pass all the params needed by detail view to avoid accessing data model. But I don't think this approach scales. The second is to not use forced wrapping. That should work fine, but I doubt if this is the recommended way to do it.

BTW, another similar setup to generate the crash is to use three views: list view -> detail view -> delete view. When user clicks on button in delete view, the detail view will crash.

Thanks.

Update:

1. SwiftUI may recall an existing view's body which contains stale data.

@jrturton I was aware that changes to @EnvironmentObject would cause view's body get called. What I didn't realize was that SwiftUI migtht recalled an existing view's body which contain stale data. I never read any discussion about this on the net. Do you know why SwiftUI do that?

I had always thought that when state changes, SwiftUI would regenerate the entire view hierarchy from top down by calling Content.body. If it was so, FooDetailView would always has the up-to-date data when it gets called and there wouldn't be the issue. I had the understanding because SwiftUI is advertised as a state driven architecture, and app developers are supposed to declare the UI based on the current state. By "current" I mean the new state, not the previous state. That's the reason why I thought it should be fine to use forced unwrapping.

1. I doubt if passing all params to detail view is a general solution.

First, this doesn't scale well. For example, suppose Foo is associated with another struct Bar (that is, Foo has a property containing Bar's id) and we want to display Bar's name in detail view, then we will need to add bar name to the params. For a complicated Item, its detail view may contain a lot of things which are determined at runtime, it will be hard to prepare everything ahead by the caller.

More importantly, once we pass these params to detail view, they are effectively outside data model and can easily go stale. It would be issue if user performs delete action using these stale data.

1. Passing binding doesn't solve the crash issue on its own.

@lorem-ipsum, thanks for your suggestion on using binding. I wasn't aware that ForEach can take binding. What's more, I have also beening think if it's good practice to pass binding, instead of regular params, in SwiftUI (I don't know that anwser yet).

That said, passing binding doesn't solve the crash issue on its own, because when the data model changes, SwiftUI still recalls the existing detail view's body which contain stale data.

1. The solution?

I think the root cause is that, although SwiftUI is advertised as a state driven architecture, a view's body may get called with stale data. So the data model's api should deal with invalid params. Hornestly speaking, this isn't a design decison which I prefer to. I usually think the caller should only pass valid params to data model API. Otherwise it's an architecture issue that should be resolved on the caller side in the first place. Unfortunately it seems that's the case with SwiftUI.

(Note: Thank all for pointing out that deletion code should be in data model. I knew that. I didn't do it because I spent long time investigating the issue in my app and was exhausted when I prepared the example code.)

I'm looking to do replacements in unknown third-party inputs in strings that sometimes have quotes among them.

I want to replace a wholeword whereever it occurs unless it's in double or single-quotes, and unless the quote is escaped.

Example: Replacing FOO by BAR

Input:

It is possible to neglect some elements during Collectors.groupingBy or Collectors.partitioningBy?

Of course I know I can place a .filter() anywhere in a stream. But my problem is that I have to run a rather complex and expensive evaluation, to decide in which "group" my objects should be divided into.

Moreover, there are always many objects that I would want to neglect during collecting.

Example: imagine a List, that I want to split into 2 lists. That's easy, but how can I additionally neglect all objects that don't fit into my evaluated condition?

I have a bug in my code which I don't quite understand. According to the documentation, std::vector.back() returns a reference to the last element in the container, so here's what I did: (live here)

I have a vector x of type Eigen::VectorXi with more than 2^31-1 entries, which I would like to return to R. I can do that by copying x entry-wisely to a new vector of type Rcpp::IntegerVector, but that seems to be quite slow.

I am wondering:

1. whether there is a more efficient workaround;
2. why in the following reproducible example Rcpp::wrap(x) doesn't work.

test.cpp

I have the following code that deletes the signal during one of the callbacks from the signal:

I am trying to create a select query with a simple where-clause using Haskell's beam. From https://haskell-beam.github.io/beam/user-guide/queries/select/#where-clause, I believed that this would work:

In a pipeline like

is it somehow possible to use the new searable in combination with a @FetchRequest?

I have a code like this: