Preserve trailing comments in OrderedImports rule

[itsybitsybootsy]

Problem

OrderedImports runs every source file through a convertToCodeBlockItems pass that walks the categorised Line array and flushes accumulated trivia onto each syntax node via pendingLeadingTrivia. When the input ends with trivia after the last syntax node — typically a free-standing comment that import reordering pushed past the final import — the accumulator is silently dropped at the end of the loop.

import A

// Comment

import B

becomes

import A
import B

The // Comment is silently lost.

Fix

Append any leftover trivia containing comments to the trailing trivia of the last code block item before convertToCodeBlockItems returns. The loop synthesised a trailing run of newlines after every iteration to separate the current line from the next; those are meaningless when no following line exists and are stripped first so the output doesn't gain spurious blank lines.

After the fix:

import A
import B

// Comment

Tests

Added testCommentBetweenImportGroupsIsPreserved to OrderedImportsTests.swift covering the reproducer from the issue.

swift test locally: 913 tests, 1 skipped, 0 failures.

Resolves #1080.

---

This contribution was developed with the help of Claude Code. The fix, regression test, and the full local test suite were reviewed manually before submitting.

[allevato]

Suggested change

	let hasComment = pendingLeadingTrivia.contains { piece in
	switch piece {
	case .lineComment, .blockComment, .docLineComment, .docBlockComment:
	return true
	default:
	return false
	}
	}
	if hasComment, !output.isEmpty {
	if pendingLeadingTrivia.hasAnyComments, !output.isEmpty {

We already have a helper for this in Trivia+Convenience.swift.

[itsybitsybootsy]

Done in a7dc786. Wrapped in Trivia(pieces:) since pendingLeadingTrivia is [TriviaPiece].

[allevato]

Thanks for the bug fix!