Do not write BOM to results

[martincostello]

Do not write UTF-8 BOM to results files.

Fixes #3070.

All the tests choose their own encoding, rather than using the default constructor the real exporters use, e.g.:

BenchmarkDotNet/tests/BenchmarkDotNet.Tests/Helpers/ExporterTestHelper.cs

Lines 13 to 26 in 2252a3b

	using var stream = new MemoryStream();
	using (var writer = new CancelableStreamWriter(stream, Encoding.UTF8, 1024, leaveOpen: true))
	{
	await exporter.ExportAsync(summary, writer, cancellationToken).ConfigureAwait(false);
	}
	stream.Position = 0;
	using var reader = new StreamReader(stream, Encoding.UTF8);
	string? line;
	while ((line = await reader.ReadLineAsync().ConfigureAwait(false)) != null)
	{
	logger.WriteLine(line);
	}
	}

vs:

BenchmarkDotNet/src/BenchmarkDotNet/Exporters/ExporterBase.cs

Line 38 in 2252a3b

using var writer = new CancelableStreamWriter(fileStream);

I'm happy to add/update some tests for this issue if you have opinions on how you'd like to verify this exactly.

[Copilot]

Pull request overview

This PR fixes an unintended behavior change where result files (e.g., JSON exports) are written with a UTF-8 BOM, which breaks some downstream consumers (notably JavaScript JSON parsing in CI workflows).

Changes:

• Change CancelableStreamWriter’s default encoding from Encoding.UTF8 (BOM-emitting) to a UTF-8 encoding configured to not emit a BOM.
• Introduce a cached UTF8NoBOM encoding instance used by the default CancelableStreamWriter(Stream) constructor.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.