C14.13 r1: query and line fold case the same way; final sigma folds to sigma

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

tasks/reviews/C14.13-r1.md

@@ -0,0 +1,3 @@
+FAIL
+
+- apps/ely-mark/src/folder.rs:62 — Source text is lowercased character by character, but the query is lowercased as a whole string. These differ for context-sensitive Unicode casing: `ΟΣ` becomes `οσ` in `column_of`, whereas the identical query becomes `ος` (final sigma). A document containing `ΟΣ` therefore produces no result when searching for `ΟΣ`, in either current-file or folder search; the previous whole-string implementation matches. Native Windows and WSL probes both reproduce `hits=[]` (logs: `.work/C14.13-r1/windows-sigma.log`, `.work/C14.13-r1/linux-sigma.log`). Use consistent Unicode normalization for query and source while preserving the mapping to original character columns, and add a regression covering this exact-text match and its casing variants.