Edit's x, y and s match whole runes, as addresses do

mvzr matches bytes, and only the addr path widened a match to runes, so
,x/./a/|/ on é split it into invalid UTF-8 and ,s/./X/g gave a combining
sequence four X's. Edit's matches now go through the same rune alignment
(and an empty match steps a rune, not a byte). runeStart no longer walks a
stray continuation byte back onto the ASCII before it: an invalid byte is
a place of its own.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

1 files changed, 15 insertions, 7 deletions

diff --git a/src/modal.zig b/src/modal.zig
index dd7d2b57..11eda7b9 100644
--- a/src/modal.zig
+++ b/src/modal.zig
@@ -1327,18 +1327,26 @@ pub fn lineEndOffset(text: []const u8, line: usize) usize {
 /// acme's do, not grapheme clusters: a lone combining mark or a `\r` is a
 /// place of its own, whatever a terminal draws in one cell.
 pub fn runeStart(text: []const u8, off: usize) usize {
-    var o = @min(off, text.len);
+    const o = @min(off, text.len);
+    if (o == text.len or text[o] & 0xC0 != 0x80) return o;
+    var s = o;
     var steps: u8 = 0;
-    while (o > 0 and o < text.len and text[o] & 0xC0 == 0x80 and steps < 3) : (steps += 1) o -= 1;
-    return o;
+    while (s > 0 and text[s] & 0xC0 == 0x80 and steps < 3) : (steps += 1) s -= 1;
+    // Only a lead byte whose sequence reaches `o`, its bytes all
+    // continuations, makes `o` the inside of a rune; a stray continuation
+    // byte (invalid UTF-8) is a place of its own.
+    const n = std.unicode.utf8ByteSequenceLength(text[s]) catch return o;
+    if (s + n <= o or s + n > text.len) return o;
+    for (text[s + 1 .. s + n]) |c| if (c & 0xC0 != 0x80) return o;
+    return s;
 }
 
 /// `off`, or the end of the rune it falls inside.
 pub fn runeEnd(text: []const u8, off: usize) usize {
-    var o = @min(off, text.len);
-    var steps: u8 = 0;
-    while (o < text.len and text[o] & 0xC0 == 0x80 and steps < 3) : (steps += 1) o += 1;
-    return o;
+    const o = @min(off, text.len);
+    const s = runeStart(text, o);
+    if (s == o) return o;
+    return s + (std.unicode.utf8ByteSequenceLength(text[s]) catch 1);
 }
 
 /// The rune boundary after the one at `off`.