String character offset <-> byte offset

[tera]

@SDGGiesbrecht

It is never valid to use a String.Index on any string instance besides the one you got it from. Calling == across indices from two different string instances violates that. In this case, the string you created from source (Swift) has its memory in UTF‐8 under the hood, but the string you got from UIKit (Objective C) has its memory in UTF‐16 under the hood. Thus when you call == , under the hood you are comparing apples to oranges, and you will not get a meaningful answer. (The String.Index type does not store the string it came from—otherwise the index would use more memory than the string itself—so the two indices have no way to work out what each would mean in the other’s format even if they wanted to.)

I have a helper structure that helps working with string indices:

struct StringIndex {
    let string: String
    let index: String.Index
}

What's the best way to assert that the two arbitrary indices belong to "the same" string? I presume "a.string == b.string" is not the right check as two different strings might compare equal, e.g.:

U+00C5 (LATIN CAPITAL LETTER A WITH RING ABOVE)  ==
U+0041 + U+030A (LATIN CAPITAL LETTER A + COMBINING RING ABOVE)

Would this check on utf8 representations be correct for the assert?

a.utf8CString == b.utf8CString

Note, in my code all indices are created with "string.startIndex", "string.index(after:)", etc. and I am not mixing and matching indices created via different views... (it would also be cool if I can assert that fact as well, if that's possible / necessary).