Suspicious mismatch between comments and code

[Shnatsel]

The comment indicates len-(STD_MIN_MATCH-1), note the -1, while the actual code uses +1:

zlib-rs/zlib-rs/src/deflate/longest_match.rs

Lines 302 to 309 in 0c161ed

	/* Try hash head at len-(STD_MIN_MATCH-1) position to see if we could get
	* a better cur_match at the end of string. Using (STD_MIN_MATCH-1) lets
	* us include one more byte into hash - the byte which will be checked
	* in main loop now, and which allows to grow match by 1.
	*/
	let [scan0, scan1, scan2, ..] = scan[len - (STD_MIN_MATCH + 1)..] else {
	panic!("index out of bounds");
	};

Changing the code to -1 to match by the comment doesn't break any tests. So this codepath might be untested?

FWIW upstream zlib-ng also has this mismatch:

https://github.com/zlib-ng/zlib-ng/blob/c2712b8a345191f6ed79558c089777df94590087/match_tpl.h#L204-L209

[folkertdev]

This is fun, the implementation that zlib-ng is basing that on is actually different, it uses scan_end = scan + len - MIN_MATCH + 1; so without parens.

https://github.com/gildor2/fast_zlib/blob/6141d21116692cd95c0c5dc17106758c372b9fc4/Sources/match.h#L216