SF-0033: Implement String.Encoding.ianaName and String.Encoding(ianaName:).

[YOCKOW]

Implementation of the proposal: #1243 (SF-0033)

[YOCKOW]

@swift-ci Please test

[YOCKOW]

@itingliu

I've made this PR ready for review.
I'm happy if you take a look at this. Thank you.

cc: Foundation Workgroup
@adam-fowler @iCharlesHu @Lukasa @designatednerd @jmschonfeld @lorentey @stephentyrone @itingliu @tomerd @parkera

[YOCKOW]

@itingliu
Thank you for reviewing. I'm going to modify implementation reflecting your feedback.

[YOCKOW]

@swift-ci Please test

[parkera]

@swift-ci test

[parkera]

Looking good to me after the recent simplifications. Will also want @itingliu approval.

[YOCKOW]

Thank you @parkera.
I'm going to wait for @itingliu.

Just in case, what's changed:

• Rewrote util script in Swift
• Simplified the logic to parse names
  • Use UInt8(UTF8.UnitCode)
  • Remove redundant tokenizers.

[itingliu]

I also like this simplified version. Thank you! just had a few minor comments

[itingliu]

Would it work if you write

let possibilities: [IANACharset] = {...

and below just do

         func __determineEncoding() -> String.Encoding? {
            for ianaCharset in possibilities {
                if ianaCharset.matches(charsetName) {
                    return encoding
                }
            }
            return nil
        }

Am I missing anything?

[YOCKOW]

Won't we lose encoding for return encoding in that implementation?

[itingliu]

And to simply this even further, it looks like this logic inside func __determineEncoding() doesn't even needed to be extracted in a function since it's only used once?

[YOCKOW]

The reason why nested function is used is just because of readability since self = nil; return is invalid in Swift.

Sorry, we can write self = encoding; return...
I'm fixing it.

[YOCKOW]

I've fixed String.Encoding(ianaName:).
I hope that this is in line with your intention.