Modernize XmlUTF8TextWriter with u8/inline out's.

[jlennox]

I couldn't find any tests for this area of code. I am not certain there are any. I have performed no testing. This was done because while browsing I noticed this code would benefit from some modernizing.

Is there a better method to copy a ReadOnlySpan<byte> to a byte[] than value.CopyTo(new Span<byte>(buffer, offset, buffer.Length));?

I followed the if (value.Length < bufferLength) semantics but as currently used, that condition will always be true.

[stephentoub]

Is there a better method to copy a ReadOnlySpan to a byte[] than value.CopyTo(new Span(buffer, offset, buffer.Length));?

If the goal is to copy to the beginning of the buffer, you'd just do value.CopyTo(buffer) and rely on the implicit conversion from an array to a span. If the goal is to copy to some non-0 offset, then you can do buffer.AsSpan(offset). You'd almost never see new Span<byte>(buffer, offset, buffer.Length), as that would fail for any offset other than 0 (since the length is the full length of the buffer).

[stephentoub]

Suggested change

	value.CopyTo(new Span<byte>(buffer, offset, buffer.Length));
	value.CopyTo(buffer.AsSpan(offset));

If the code in this PR is passing tests, it seems like we're missing some tests, as a non-0 offset value here will result in exceptions.

[stephentoub]

Do we still need this WriteUTF8Chars method now that there's the WriteUTF8Bytes method below? Any remaining call sites could be changed from WriteUTF8Chars(value, offset, count) to WriteUTF8Chars(value.AsSpan(offset, count)).

[stephentoub]

Thanks for the PR. It seems like this might be duplicating changes already being made in https://github.com/dotnet/runtime/pull/71478/files#diff-b500e28b981633d459d67a89d678a1f71e05f9497d1ac49859ff86b53b3ab836R225?

[HongGit]

@StephenMolloy and @mconnew can you please take a look?

[StephenMolloy]

Addressed a couple points. Looks good.

[StephenMolloy]

Test failures are unrelated. (See #83901 and #83986 which are currently preventing clean CI runs.)