docs(todo): retag TODO markers by root cause

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

1 files changed, 4 insertions, 4 deletions

diff --git a/crates/shirabe-symfony-string/src/unicode_string.rs b/crates/shirabe-symfony-string/src/unicode_string.rs
index 9048f77b..21117437 100644
--- a/crates/shirabe-symfony-string/src/unicode_string.rs
+++ b/crates/shirabe-symfony-string/src/unicode_string.rs
@@ -6,7 +6,7 @@ pub struct UnicodeString {
 }
 
 impl UnicodeString {
-    // TODO(phase-c): the real constructor runs `normalizer_normalize` (Unicode NFC normalization),
+    // TODO(unicode): the real constructor runs `normalizer_normalize` (Unicode NFC normalization),
     // which has no Rust std equivalent and would need a dedicated normalization implementation.
     // Normalization is skipped here, which is only correct for already-NFC input such as ASCII.
     pub fn new(string: &str) -> Self {
@@ -15,7 +15,7 @@ impl UnicodeString {
         }
     }
 
-    // TODO(phase-c): ASCII-only provisional implementation. The faithful `width()` uses `wcswidth`
+    // TODO(unicode): ASCII-only provisional implementation. The faithful `width()` uses `wcswidth`
     // with the Unicode width tables to treat wide characters as width 2 and skip zero-width /
     // combining characters; here every character counts as width 1, correct only for ASCII. The
     // ANSI/control-character stripping driven by `ignore_ansi_decoration` is likewise not handled.
@@ -34,7 +34,7 @@ impl UnicodeString {
         width
     }
 
-    // TODO(phase-c): the faithful `length()` uses `grapheme_strlen` (extended grapheme clusters),
+    // TODO(unicode): the faithful `length()` uses `grapheme_strlen` (extended grapheme clusters),
     // which needs Unicode segmentation tables with no Rust std equivalent and no permitted crate.
     // Approximated with the code-point count, exact only when no combining/multi-code-point
     // clusters are present (e.g. ASCII).
@@ -42,7 +42,7 @@ impl UnicodeString {
         shirabe_php_shim::mb_strlen(&self.string, "UTF-8")
     }
 
-    // TODO(phase-c): the faithful `slice()` uses `grapheme_substr` (grapheme-cluster offsets), which
+    // TODO(unicode): the faithful `slice()` uses `grapheme_substr` (grapheme-cluster offsets), which
     // needs Unicode segmentation tables with no std equivalent and no permitted crate. Approximated
     // with code-point offsets via `mb_substr`, exact only without combining/multi-code-point clusters.
     pub fn slice(&self, start: i64, length: Option<i64>) -> Self {