Do not treat an escaped backslash at the end of a line as a hard break

[bobzhang]

Split out of #152. Found by differential fuzzing against markdown-it-py (commonmark preset).

Bug: a backslash at the end of a line is a hard line break, but not when it is itself escaped by another backslash.

Repro: a\\ followed by a newline and b rendered a\<br />b. Expected a\ followed by a soft break.

Fix: Token::newline treats the final \ as a hard break only if the run of backslashes ending the line has odd length. The run is counted with StringView::rev_iter().take_while(...).count().

Spec examples: none (the spec suite passes either way).

Tests: src/cmark_html/escaped_backslash_line_end_test.mbt.

Depends on: nothing, independent of the other PRs.

[bobzhang]

Automated review: Codex CLI (codex exec, model reasoning effort high, read-only sandbox), 1 round, on git diff against the base branch.

Verdict: signed off, no changes requested.

I sign off on the code review. No concrete correctness, regression, or code-quality problems found.

The odd/even backslash check follows CommonMark 0.31.2’s escape rules and hard-break rules. The library iterator approach is appropriate, and the tests cover the essential two-, three-, and four-backslash cases. Inspection found no regressions involving trailing spaces, CRLF, block endings, or code spans.

Codex could not run moon test in its read-only sandbox. The tests were run separately on native, js and wasm-gc and pass.