Conversation
| @@ -1 +1 @@ | |||
| 2.4.2 | |||
| 2.4.4 | |||
There was a problem hiding this comment.
Same remark as for the other PR regarding system version
| */ | ||
| func process(folder: Path, withPrefix prefix: String = "") -> [Catalog.Entry] { | ||
| return (try? folder.children().sorted(by: <).flatMap { | ||
| return (try? folder.children().sorted(by: <).compactMap { |
There was a problem hiding this comment.
Not sure I get this change
folder.children().sorted(by:<) is an array
the closure calls process which returns an array too
So in this use case we should be using the monadic flatMap (the one that wasn't renamed), not the non-monadic [T].compactMap(f: (T) -> T?) -> [T] one? So why would the compiler want us to change it there, did we get types mixed up and didn't do what we thought we were doing before, using the non monadic one all along by mistake?
There was a problem hiding this comment.
The process(item: $0, withPrefix: prefix) that's called in the compactMap closure has as type:
func process(item: Path, withPrefix prefix: String) -> Catalog.Entry?There was a problem hiding this comment.
Oh right it's not the same process method as process(folder: withPrefix: ) -> [Catalog.Entry], right. 👍
| guard let descRefs = (descs as? [CTFontDescriptor]) else { return [] } | ||
|
|
||
| return descRefs.flatMap { desc -> Font? in | ||
| return descRefs.compactMap { desc -> Font? in |
There was a problem hiding this comment.
(… As compared to here where the change is justified as we operate on an array but the closure returns an optional, so here it's indeed intended to be the non-monadic compactMap and this change is expected)
|
I'm assuming that with the ruby change this PR should be ok, and go ahead and merge it. |
2 participants
Note that to fix some warnings, we have to use
compactMapin the SwiftGenKit code, which might be considered a breaking change (Swift 4.1). I don't know if this warrants a major version bump for SwiftGenKit (and SwiftGen).