> On 7 Jul 2016, at 21:07, G B via swift-evolution <[email protected]>
> wrote:
>
> It has always seemed odd to me that `case`s use a colon as a delimiter rather
> than curly braces like everything else. Is there a reason for this other
> than the legacy of C-like languages?
>
> If I wanted to write a series of branching `if` \ `else` statements I would
> do it like so:
>
> if x==0 { print(0) }
> else if x==1 { print (1) }
> else if x==2 { print(2) }
> else { print("other”) }
>
> I believe all flow control is wrapped in curly braces, except for `case`s
> inside a `switch`:
>
> switch x {
> case 0: print(0)
> case 1: print(1)
> case 2: print(2)
> default: print("other")
> }
>
>
> I feel like this would be more consistent with the rest of the syntax:
>
> switch x {
> case 0 { print(0) }
> case 1 { print(1) }
> case 2 { print(2) }
> default { print("other”) }
> }
>
> The colon syntax evokes a label, but the modern, complex `case` statements in
> Swift don’t act much like labels.
While I could maybe see this being an option, we already kind of have it thanks
to do{} blocks:
switch x {
case 0: do {
print(0)
}
case 1: do {
print(1)
}
case 2: do {
print(2)
}
default: do {
print("other")
}
}
Not quite as pretty, but perfectly functional right now, but will many people
use it if given the chance? If you need scoping then you probably have
conditionals anyway, or you can just do{} blocks for those cases only, I don't
really see the need to require them on all of them.
Personally I don't see the need; the case labels themselves are highlighted
pretty clearly thanks to being keywords, and act as both the opening for their
block, and the close of any previous block, meanwhile the whole switch has
enclosing braces already. There's also the issue of scope; you can fall through
from a case, so technically speaking the only scope may be the switch itself,
as you could fall through every single case and thus visit everything on the
way down.
So yeah, I think the construct kind of already makes sense as it is, it's just
a bit of weird one conceptually._______________________________________________
swift-evolution mailing list
[email protected]
https://lists.swift.org/mailman/listinfo/swift-evolution
