Test plan for Utf8StringLiterals feature.

[AlekseyTs]

Speclet - https://github.com/dotnet/csharplang/blob/main/proposals/csharp-11.0/utf8-string-literals.md
Feature branch - https://github.com/dotnet/roslyn/tree/features/Utf8StringLiterals
Championed issue - dotnet/csharplang#184

Compiler

• LangVer
  • for u8 suffix
• compiler test plan
• parsing
  • u8 and U8 suffix allowed
  • no trivia between string constant and u8 suffix
  • single and multi-line verbatim strings
  • single and multi-line raw string literals
  • interpolated strings with suffix disallowed
  • disallowed in # directives
• natural type of "abc"u8 is ReadOnlySpan<byte>: NaturalType_01
• no target type of constant string with u8 suffix to byte[] or Span<byte>: ImplicitConversions_02
• explicit conversion from constant string with u8 suffix to Span<byte>, or byte[] is disallowed: ExplicitConversions_02
• implicit/explicit conversion from constant string to ReadOnlySpan<byte>, Span<byte>, or byte[] is disallowed: ImplicitConversions_01
• implicit/explicit conversion from non-constant string to ReadOnlySpan<byte>, Span<byte>, or byte[] is disallowed: NoConversionFromNonConstant_01, _02
• implicit/explicit conversion from null to ReadOnlySpan<byte>, Span<byte>, or byte[] is allowed (unchanged): NoBehaviorChangeForConversionFromNullLiteral_01
• implicit/explicit conversion from (object)null to ReadOnlySpan<byte>, Span<byte>, or byte[] is disallowed: ImplicitConversions_04, ExplicitConversions_04
• implicit/explicit conversion from (string)null to ReadOnlySpan<byte>, Span<byte>, or byte[] is disallowed: ImplicitConversions_03, ExplicitConversions_03
• error reported for invalid UTF8 in u8 string: InvalidContent_02, _03
• pattern matching ReadOnlySpan<byte> against string constants: PatternMatching_01, _02, _03
• Expression tree use or implicit conversion and u8 literal disallowed
• UTF8 byte[] stored in <PrivateImplementationDetails> if Length > 2
• pattern matching ReadOnlySpan<byte> against string constants: PatternMatching_01, _02, _03
• Expression tree use or implicit conversion and u8 literal disallowed
• UTF8 byte[] stored in <PrivateImplementationDetails> if Length > 2
• concatenation: UserDefinedConcatenation_01, etc.
• const ReadOnlySpan<byte> x = "abc"u8; is disallowed: NotConstant_01
• void M(Span<byte> x = "abc"u8) is disallowed: DefaultParameterValues_01, _02
• "abc"u8 and implicit conversion disallowed in attribute value: AttributeArgument_01, etc.
• nullability of "abc"u8: NullableAnalysis_03
• Missing and invalid type and members: ReadOnlySpan<T>, Span<T>, etc.
• Document breaking changes (none)
• IOperation: IOperationTests_IUtf8StringOperation
• GetTypeInfo()
• GetConversion()
• EE
• EnC with <PrivateImplementationDetails>
• SyntaxNormalizer: string, single and multi-line raw string
• null terminator: NullTerminate_01, _02

Public API

• API review New public APIs for UTF8 String Literals #61847

IDE

• SyntaxClassifier: string, single and multi-line raw string
• QuickInfo: byte[] natural type

Debugger

• Can/should we show the string (rather than just bytes) when hovering over such variables? (There's discussion on csharplang on this and decompilation experience)

Other

• LDM: readonlySpanOfByte is "abc" or readonlySpanOfByte is "abc"u8? (also reported here and here) [Proposal]: UTF-8 Literal Pattern Matching Support csharplang#6161
• LDM: confirm whether the conversion part of the feature is desirable (answer: no)
• LDM: confirm whether the suffix part of the feature is desirable (answer: yes)
• LDM: allow concatenation? (issue) Implement concatenation operator for "u8" literals. #62044
• Update utf8-string-literals.md to reflect LDM decisions
  • require u8 suffix: notes
  • natural type ReadOnlySpan<byte>: notes
  • no target type to byte[]: notes
  • null terminated: notes
  • concatenation: notes1, notes2

[ericwj]

Re

• interpolated strings with suffix disallowed

Moved.

[AlekseyTs]

@ericwj This issue is not for general purpose discussions on topics related to the feature. If you have an issue to report/discuss, please open a separate dedicated issue for that.

[CyrusNajmabadi]

@ericwj not sure what you mean:

The curlies are not interpolation holes here.

--

Just saw Aleksey's comment. If you want to discuss this further, please open a separate issue/discussion thread. Thanks!

[cston]

Closing as completed.