max / makenotwork

tagtree: soften optimistic suggest() latency docstring Name the tier-2 segment-fallback cost honestly (full O(tags*segments) scan, tens of microseconds at low thousands) instead of claiming every suggest call finishes in single-digit microseconds. Deepaudit 07-20 NOTE. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Author: Max Johnson <me@maxj.phd> · 2026-07-24 19:26 UTC

Commit: b9c8d7ba3dc6ec1191d5875dd80534fcb667ecb2

Parent: fe095ea

1 file changed, +4 insertions, -2 deletions

M shared/tagtree/src/lib.rs +4 -2

			@@ -702,8 +702,10 @@ pub fn batch_rename(operations: &[(&str, &str)], index: &mut TagIndex) -> usize
702	702		///
703	703		/// Backed by a sorted `Vec<String>`: binary search for path-prefix matching
704	704		/// (O(log n) to find the start, then linear scan for results), with a segment-prefix
705		-	/// fallback for mid-tag matches. At typical corpus sizes (hundreds to low thousands
706		-	/// of tags), the entire `suggest` call completes in single-digit microseconds.
	705	+	/// fallback for mid-tag matches. A path-prefix keystroke at typical corpus sizes
	706	+	/// (hundreds to low thousands of tags) completes in single-digit microseconds; a
	707	+	/// keystroke that falls through to the segment fallback pays a full O(tags * segments)
	708	+	/// scan, into the tens of microseconds at low thousands.
707	709		///
708	710		/// ```
709	711		/// use tagtree::TagIndex;