util,readline: NFC-normalize strings before getStringWidth

The assumption here is that decomposed characters render like their
composed character equivalents, and that working with the former
comes with a risk of over-estimating string widths given that
we compute them on a per-code-point basis. The regression test
added here (한글 vs 한글) is an example of that happening.

PR-URL: #33052
Reviewed-By: Gus Caplan <[email protected]>
Reviewed-By: Michaël Zasso <[email protected]>
Reviewed-By: Anto Aravinth <[email protected]>
Reviewed-By: Colin Ihrig <[email protected]>
Reviewed-By: James M Snell <[email protected]>

Author: addaleax

lib/internal/util/inspect.js

@@ -1922,6 +1922,13 @@ function formatWithOptionsInternal(inspectOptions, ...args) {
   return str;
 }
 
+function prepareStringForGetStringWidth(str, removeControlChars) {
+  str = str.normalize('NFC');
+  if (removeControlChars)
+    str = stripVTControlCharacters(str);
+  return str;
+}
+
 if (internalBinding('config').hasIntl) {
   const icu = internalBinding('icu');
   // icu.getStringWidth(string, ambiguousAsFullWidth, expandEmojiSequence)
@@ -1931,8 +1938,8 @@ if (internalBinding('config').hasIntl) {
   // the receiving end supports.
   getStringWidth = function getStringWidth(str, removeControlChars = true) {
     let width = 0;
-    if (removeControlChars)
-      str = stripVTControlCharacters(str);
+
+    str = prepareStringForGetStringWidth(str, removeControlChars);
     for (let i = 0; i < str.length; i++) {
       // Try to avoid calling into C++ by first handling the ASCII portion of
       // the string. If it is fully ASCII, we skip the C++ part.
@@ -1952,9 +1959,7 @@ if (internalBinding('config').hasIntl) {
   getStringWidth = function getStringWidth(str, removeControlChars = true) {
     let width = 0;
 
-    if (removeControlChars)
-      str = stripVTControlCharacters(str);
-
+    str = prepareStringForGetStringWidth(str, removeControlChars);
     for (const char of str) {
       const code = char.codePointAt(0);
       if (isFullWidthCodePoint(code)) {

test/parallel/test-icu-stringwidth.js

@@ -87,3 +87,12 @@ for (let i = 0; i < 256; i++) {
     assert.strictEqual(getStringWidth(char), 1);
   }
 }
+
+{
+  const a = '한글'.normalize('NFD'); // 한글
+  const b = '한글'.normalize('NFC'); // 한글
+  assert.strictEqual(a.length, 6);
+  assert.strictEqual(b.length, 2);
+  assert.strictEqual(getStringWidth(a), 4);
+  assert.strictEqual(getStringWidth(b), 4);
+}