[SR-9749] The encodedOffset of String.Index has been changed to either UTF-8 or UTF-16.

[norio-nomura]

Previous ID	SR-9749
Radar	rdar://problem/47605131
Original Reporter	@norio-nomura
Type	Bug
Status	Resolved
Resolution	Done

Environment

swift-DEVELOPMENT-SNAPSHOT-2019-01-23-a
swift-5.0-DEVELOPMENT-SNAPSHOT-2019-01-23-a

Additional Detail from JIRA

Votes	0
Component/s	Standard Library
Labels	Bug
Assignee	@milseman
Priority	Medium

md5: 66ece592c380ddd277127ea29c649351

Issue Description:

The problem is:

1. Before merging [String] Use a UTF-8 representation for native strings #20315, The encodedOffset of String.Index was specified as UTF-16 offset, and it actually worked as expected.

   /// The offset into a string's UTF-16 encoding for this index.
   @inlinable // FIXME(sil-serialize-all)
   public var encodedOffset : Int {

   https://github.com/apple/swift/pull/20315/files#diff-8ec9c169d4bc3c678838b4067481979eL130

   /// Creates a new index at the specified UTF-16 offset.
   ///
   /// - Parameter offset: An offset in UTF-16 code units.
   @inlinable // FIXME(sil-serialize-all)
   public init(encodedOffset offset: Int) {

   https://github.com/apple/swift/pull/20315/files#diff-8ec9c169d4bc3c678838b4067481979eL106

2. A lot of codes were written that expect that behavior (including swift-corelibs-foundation)
   e.g. codes using encodedOffset in Swift on GitHub can be found: https://github.com/search?l=Swift&q=encodedOffset&type=Code

3. However, the change is not written in changelog etc., and it is understood only from the description of UTF-16 deleted from each documentation comment. In fact, swift-corelibs-foundation did not recognize that change was necessary, which caused SR-9454.

I think that some countermeasures are necessary:

• [MUST] Behavior changes should be documented (eg within changelog).

• [MAY] Some code change
  Mark init(encodedOffset:) and encodedOffset as deprecated
  or
  Introduce new property exposing encode information(UTF-8 or UTF-16) to String

/cc: @milseman

[norio-nomura]

I confirmed that no change was described in document concerning this behavior change. 🙁
https://developer.apple.com/documentation/swift/string/index/2922827-init?changes=latest_minor
https://developer.apple.com/documentation/swift/string/index/2922802-encodedoffset?changes=latest_minor
https://developer.apple.com/documentation/xcode_release_notes/xcode_10_2_beta_release_notes/swift_5_release_notes_for_xcode_10_2_beta

[milseman]

#22108

Will need release-noting as well.

[norio-nomura]

I am very happy that it seems to be adopting the best solution (I did not write on the issue) that I imagined! 😃

[norio-nomura]

Another example to fix an issue caused by wrong usage of encodedOffset.
apple/swift-corelibs-foundation#1894

[milseman]

https://github.com/apple/swift-evolution/blob/master/proposals/0241-string-index-explicit-encoding-offset.md