`CharacterClass.anyUnicodeScalar` crashes

Question

`CharacterClass.anyUnicodeScalar` crashes

hamishknight opened this issue 2 years ago · comments

For the following:

import RegexBuilder

let r = Regex {
  CharacterClass.anyUnicodeScalar
}
_ = try! r.wholeMatch(in: "x")

Crashes with:

Fatal error: Should have been handled by tree conversion

However it's not clear this API should still exist, as it was removed from https://github.com/apple/swift-evolution/blob/main/proposals/0363-unicode-for-string-processing.md.

Hamish Knight · Answer 1 · Mon Jul 04 2022 23:09:01 GMT+0800 (China Standard Time)

cc @natecook1000, should this API be removed?

Michael Ilseman · Answer 2 · Tue Jul 05 2022 00:33:39 GMT+0800 (China Standard Time)

anyUnicodeScalar would be equivalent to anyCharacter. The interesting cases is when you want to go a step higher in semantics and match anyCharacter from scalar semantics.

Hamish Knight · Answer 3 · Tue Jul 05 2022 18:22:34 GMT+0800 (China Standard Time)

@milseman Just to be clear, it was explicitly removed from the Unicode proposal, and is currently listed as an alternative considered. I also noticed that it includes \O. I have put up a PR to bring the implementation in-line with the proposal: #540.