fixed mixing symbols and letters in unicode detector on pre N
This commit is contained in:
parent
14f9a51b8b
commit
edfe4a0ccd
|
@ -117,6 +117,10 @@ public class IrregularUnicodeDetector {
|
|||
final int length = word.length();
|
||||
for (int offset = 0; offset < length; ) {
|
||||
final int codePoint = word.codePointAt(offset);
|
||||
offset += Character.charCount(codePoint);
|
||||
if (!Character.isLetter(codePoint)) {
|
||||
continue;
|
||||
}
|
||||
Character.UnicodeBlock block = normalize(Character.UnicodeBlock.of(codePoint));
|
||||
List<String> codePoints;
|
||||
if (map.containsKey(block)) {
|
||||
|
@ -126,7 +130,6 @@ public class IrregularUnicodeDetector {
|
|||
map.put(block, codePoints);
|
||||
}
|
||||
codePoints.add(String.copyValueOf(Character.toChars(codePoint)));
|
||||
offset += Character.charCount(codePoint);
|
||||
}
|
||||
return map;
|
||||
}
|
||||
|
|
Loading…
Reference in a new issue