Fixed BMPString encoding

[Craz1k0ek]

• Fixed Incorrect ASN1BMPString encoding ExpressibleByStringLiteral #110 by using UTF-16 to encode the string literal
• Added a unit test to test the added functionality

[Lukasa]

Suggested change

	[UInt8(codeUnit >> 8), UInt8(codeUnit & 0xFF)]
	[UInt8(truncatingIfNeeded: codeUnit >> 8), UInt8(truncatingIfNeeded: codeUnit)]

Separately here, I think we need to handle the case of UTF-16 surrogate pairs. We should probably detect these, and fatalError out. Seeing either a high or low surrogate is an error.

[Craz1k0ek]

I've updated the code. Not sure if the error message is clear enough and whether it's better to have a separate error for high and low surrogates

[Lukasa]

Nice, thanks!

[Lukasa]

Seems like the format checker needs to be appeased.

[Craz1k0ek]

Seems like the format checker needs to be appeased.

I checked out what he was complaining about yesterday, but I couldn't figure it out. Is it the triple property tuple for testing? If you let me know, I'll update the code.

[Lukasa]

It wants this diff applied:

diff --git a/Sources/SwiftASN1/Basic ASN1 Types/ASN1Strings.swift b/Sources/SwiftASN1/Basic ASN1 Types/ASN1Strings.swift
index bd976a6..d20b058 100644
--- a/Sources/SwiftASN1/Basic ASN1 Types/ASN1Strings.swift	
+++ b/Sources/SwiftASN1/Basic ASN1 Types/ASN1Strings.swift	
@@ -284,15 +284,19 @@ public struct ASN1BMPString: DERImplicitlyTaggable, BERImplicitlyTaggable, Hasha
 
     @inlinable
     public init(stringLiteral value: StringLiteralType) {
-        guard value.utf16.allSatisfy ({ codeUnit in
-            !(0xD800...0xDFFF).contains(codeUnit)
-        }) else {
+        guard
+            value.utf16.allSatisfy({ codeUnit in
+                !(0xD800...0xDFFF).contains(codeUnit)
+            })
+        else {
             fatalError("BMPString cannot contain characters outside the Basic Multilingual Plane: '\(value)'")
         }
 
-        self.bytes = ArraySlice(value.utf16.flatMap { codeUnit in
-            [UInt8(truncatingIfNeeded: codeUnit >> 8), UInt8(truncatingIfNeeded: codeUnit)]
-        })
+        self.bytes = ArraySlice(
+            value.utf16.flatMap { codeUnit in
+                [UInt8(truncatingIfNeeded: codeUnit >> 8), UInt8(truncatingIfNeeded: codeUnit)]
+            }
+        )
     }
 
     @inlinable
diff --git a/Tests/SwiftASN1Tests/ASN1StringTests.swift b/Tests/SwiftASN1Tests/ASN1StringTests.swift
index d338806..a43bd40 100644
--- a/Tests/SwiftASN1Tests/ASN1StringTests.swift
+++ b/Tests/SwiftASN1Tests/ASN1StringTests.swift
@@ -121,7 +121,7 @@ final class ASN1StringTests: XCTestCase {
                 "中文",
                 [78, 45, 101, 135],
                 [30, 4, 78, 45, 101, 135]
-            )
+            ),
         ]
         

Basically, two indentation tweaks, and a trailing comma.

[Lukasa]

Looks like this bit got missed:

diff --git a/Sources/SwiftASN1/Basic ASN1 Types/ASN1Strings.swift b/Sources/SwiftASN1/Basic ASN1 Types/ASN1Strings.swift
index b5073a6..d20b058 100644
--- a/Sources/SwiftASN1/Basic ASN1 Types/ASN1Strings.swift	
+++ b/Sources/SwiftASN1/Basic ASN1 Types/ASN1Strings.swift	
@@ -285,7 +285,7 @@ public struct ASN1BMPString: DERImplicitlyTaggable, BERImplicitlyTaggable, Hasha
     @inlinable
     public init(stringLiteral value: StringLiteralType) {
         guard
-            value.utf16.allSatisfy ({ codeUnit in
+            value.utf16.allSatisfy({ codeUnit in
                 !(0xD800...0xDFFF).contains(codeUnit)
             })
         else {

[dnadoba]

FWIW you can now test these fatalErrors with exit tests: https://github.com/swiftlang/swift-evolution/blob/main/proposals/testing/0008-exit-tests.md