Handle zero-width marks in string width (#19538)

Co-authored-by: fisker Cheung <lionkay@gmail.com>

src/utilities/get-string-width.js

@@ -8,6 +8,8 @@ import {
 import { isNarrowEmojiCharacter } from "narrow-emojis";
 
 const notAsciiRegex = /[^\x20-\x7F]/;
+// Exclude [`Spacing Mark`](https://www.compart.com/en/unicode/category/Mc) because spacing marks contribute horizontal width.
+const zeroWidthMarkRegex = /[\p{Nonspacing_Mark}\p{Enclosing_Mark}]/u;
 
 // Similar to https://github.com/sindresorhus/string-width
 // We don't strip ansi, always treat ambiguous width characters as having narrow width.
@@ -42,13 +44,8 @@ function getStringWidth(text) {
       continue;
     }
 
-    // Ignore combining characters
-    if (codePoint >= 0x300 && codePoint <= 0x36f) {
-      continue;
-    }
-
-    // Ignore Variation Selectors
-    if (codePoint >= 0xfe00 && codePoint <= 0xfe0f) {
+    // Ignore zero-width marks, including combining marks and variation selectors
+    if (zeroWidthMarkRegex.test(character)) {
       continue;
     }