Avoid some substring allocations during Uri host normalization

[MihaZupan]

Avoids 1 or 2 temporary string allocations - the argument to Normalize and (sometimes) its result.
Only applies to Uris with non-ASCII values.

Method	Toolchain	Mean	Error	Ratio	Allocated	Alloc Ratio
NewUri	main	397.0 ns	1.59 ns	1.00	440 B	1.00
NewUri	pr	377.6 ns	1.60 ns	0.95	368 B	0.84

[MemoryDiagnoser(false)]
public class UriBench
{
    [Benchmark]
    public Uri NewUri() => new("http://some.host.with.ümlauts/");
}

[dotnet-policy-service]

Tagging subscribers to this area: @dotnet/ncl
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull Request Overview

This PR reduces temporary string allocations during URI host normalization by introducing a span-based normalization helper and optimizing Bidi control character stripping.

• Introduces NormalizeAndConcat to normalize a span and concatenate it without intermediate allocations.
• Refactors StripBidiControlCharacters to return a boolean and optional output, avoiding new strings when no characters are removed.
• Updates Uri.CheckAuthorityHelper calls to use the new helpers.

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
src/libraries/System.Private.Uri/src/System/UriHelper.cs	Added NormalizeAndConcat and refactored StripBidiControlCharacters overloads for allocation improvements.
src/libraries/System.Private.Uri/src/System/Uri.cs	Updated host normalization logic to use NormalizeAndConcat and new StripBidiControlCharacters overloads.

Comments suppressed due to low confidence (3)

src/libraries/System.Private.Uri/src/System/UriHelper.cs:23

• Add unit tests for NormalizeAndConcat to verify behavior with null start values, empty spans, and spans containing non-ASCII characters to ensure correct concatenation and normalization.

        public static string NormalizeAndConcat(string? start, ReadOnlySpan<char> toNormalize)

src/libraries/System.Private.Uri/src/System/UriHelper.cs:610

• Add unit tests for the new StripBidiControlCharacters overloads to cover both the fast path (no removal) and removal path, including the backingString fallback behavior.

        public static string StripBidiControlCharacters(ReadOnlySpan<char> strToClean, string? backingString = null)

src/libraries/System.Private.Uri/src/System/UriHelper.cs:23

• [nitpick] Consider adding XML documentation to NormalizeAndConcat to explain its purpose, expected parameters (including null handling), and behavior for consumers and future maintainers.

        public static string NormalizeAndConcat(string? start, ReadOnlySpan<char> toNormalize)

[rzikm]

LGTM