Narrow StringOutput.OutputType to non-optional

[broken-circle]

StringOutput.output(from:) never returns an optional string, so the associated OutputType does not need to be optional. This PR narrows its type to String. This is a breaking API change, but one I think is worth landing before 1.0.

The current behavior is that the user always gets a non-optional String. Invalid byte sequences are replaced with the Unicode replacement character (U+FFFD), which is the behavior of String(decoding:as:) that output(from:) already uses via String(decodingBytes:as:). This also makes StringOutput consistent with DataOutput, whose OutputType is non-optional Data.

The optional type was introduced in iCharlesHu/Subprocess/main@7f9cea8 and brought over to this repository in the initial commit. I couldn't find a reason the type was optional. The type permits nil, but no code path produces it; if returning nil on invalid input is behavior we want, let me know and I'll look into it as a follow-up PR. String(validating:as:) would be the appropriate API, but it's gated on macOS 15+ (Subprocess supports macOS 13+), so the implementation will require availability-gated fallbacks.

[broken-circle]

@iCharlesHu, somewhat related question: the 1.0 beta release includes the deprecated .standardOutput/.standardError aliases on FileDescriptorOutput (the where Self == FileDescriptorOutput extension in Output.swift), each marked // TODO: remove for 1.0. Now that the beta has shipped with them present, is removal before 1.0 still the plan, or do they need to stay through 1.0 for source compatibility? I can open a follow-up PR to remove them if you want.

[jakepetroules]

Personally I think String(decoding:as:) + non-Optional is the right default, if people really want to validate that the text doesn't contain invalid bytes they could always use bytes and do the parsing themselves.

[broken-circle]

Exactly what I was thinking!