I’ll leave it to others to explain whether this is the intended behavior, but
the further I get from properties being idempotent, the more grief I’ve had.
Especially when (as I assume is your real use case) I tried to assign something
to itself to force a didSet {} during viewDidLoad().
If I may read your mind further: Yes, your prepareForSegues will fill view
controllers’ state before the view is instantiated, and therefore before its UI
can be made to conform. Give in: Have the setter store the value in a private
temporary, attempt to update its UI element through a conditional chain from
its outlet (a no-op if the outlet isn’t filled yet), and have viewDidLoad()
call a batch vars-to-UI func to bring everything up to the current state.
Of course there are side-effects; that’s what didSet {} is for. But
assignment-to-self is more apt to be a typo for some other assignment than not;
and it’s easy to work around by factoring the side effects into their own func
— and then you can replace the self-assignment with an explicit call to the
service you wanted to accomplish. Giving that service a name isolates it from
additional tasks you might want to do at didSet {} time, and makes it clear at
the point-of-call exactly what you want to do.
I can imagine an optimizer or runtime that assumes an assignment that doesn’t
change the value can be skipped, but I’m sure it can’t be done without breaking
some source. (Source which deserves breaking, but one picks one’s battles…)
— F
> On 21 Feb 2017, at 7:57 PM, Saagar Jha via swift-users <swift-users@swift.org
> <mailto:swift-users@swift.org>> wrote:
>
> It appears that trying to assign a variable to itself is an error:
>
> var foo = 5
> foo = foo // error: assigning a variable to itself
>
> However, I was thinking about this and was wondering why this is an error and
> not a warning. If, for example, the assignment had a side effect (e.g. a
> didSet block) I’d think it would be warning, possibly ignorable with a set of
> parentheses. Thoughts?
>
_______________________________________________
swift-users mailing list
swift-users@swift.org
https://lists.swift.org/mailman/listinfo/swift-users
