Compiler error when char byte in UTF32 (`\U...`) or hex (`\x...`) notation

[Booksbaum]

A byte can be represented as integer number (97uy and its Hex/Oct/Bin formats) -- but also as char ('a'B).

Besides normal char representation ('a'), Char supports several other formats: 3-digit decimal ('\097'), 2-digit hexadecimal ('\x61'), 4-digit hexadecimal ('\u0061', UTF16), and 8-digit hexadecimal ('\U00000061', UTF32).

It would make sense for byte to support the same representations. And it does support some -- but not 2-digit and 8-digit hexadecimal:

[
    'a'B
    '\097'B
    '\x61'B
//  ^^^^^^ This value is not a function and cannot be applied.
    '\u0061'B
    '\U00000061'B
//  ^^^^^^^^^^^^ This value is not a function and cannot be applied.
]

(fable repl, sharplab)

Expected behavior

Byte supports same notations as normal char.

Actual behavior

Byte fails for 2-digit (\x) and 8-digit (\U) hex formats -- but works with other formats (including 4-digit hex (\u)).

Known workarounds

Use another format. It's quite trivial to change \x & \U to \u (4-digits) as byte doesn't support large values.

Note: \x & \U are both supported in byte string array:

> "a\097\x61\u0061\U00000061"B;;
val it: byte array = [|97uy; 97uy; 97uy; 97uy; 97uy|]

Unless not handling \x & \U is desired behaviour (which I don't think because: a) why shouldn't it? and b) it's supported in byte string array), I would attempt to fix it (I think should just be slight adjustment to lexer by copying from char, adding range check and then some tests)

Related information

• Operating system: Windows 11 x64
• dotnet --version: 7.0.307, 8.0.100-preview.7.23376.3