bharos commented on code in PR #14515:
URL: https://github.com/apache/gravitino/pull/14515#discussion_r4461600868
##########
common/src/main/java/org/apache/gravitino/auth/ActiveRolesParser.java:
##########
@@ -119,4 +127,98 @@ public static ActiveRoles parse(@Nullable String rawValue)
{
}
return ActiveRoles.of(names);
}
+
+ // Servlet containers such as Jetty expose each header byte as one
ISO-8859-1 char. Re-read those
+ // bytes as UTF-8 so a role name sent as raw UTF-8 matches. A value holding
a char above U+00FF
+ // was already decoded by the container and is kept as is.
+ private static String recoverUtf8(String rawValue) {
+ boolean hasNonAscii = false;
+ for (int i = 0; i < rawValue.length(); i++) {
+ char c = rawValue.charAt(i);
+ if (c > 0xFF) {
+ return rawValue;
+ }
+ hasNonAscii |= c >= 0x80;
+ }
+ return hasNonAscii
+ ? decodeUtf8(rawValue.getBytes(StandardCharsets.ISO_8859_1), rawValue)
+ : rawValue;
+ }
+
+ private static String decodeRoleName(String token, String value) {
+ if (token.indexOf('%') < 0) {
+ return token;
+ }
+ ByteArrayOutputStream bytes = new ByteArrayOutputStream();
+ int i = 0;
+ while (i < token.length()) {
+ if (token.charAt(i) == '%') {
+ bytes.write(decodeEscape(token, i, value));
+ i += 3;
+ continue;
+ }
+ int next = token.indexOf('%', i);
+ int end = next < 0 ? token.length() : next;
+ byte[] literal = token.substring(i,
end).getBytes(StandardCharsets.UTF_8);
+ bytes.write(literal, 0, literal.length);
+ i = end;
+ }
+ return decodeUtf8(bytes.toByteArray(), value);
+ }
+
+ // Decoding can produce names the raw header could not carry: reject blank
names (e.g. "%20") here
+ // so they get a 400, and control characters so a name cannot forge lines in
logs.
+ private static String checkRoleName(String name, String value) {
+ if (name.trim().isEmpty()) {
Review Comment:
Switched the blank check to `StringUtils.isBlank` (matching
`RoleCreateRequest`), so a decoded Unicode-whitespace-only name like
`%E2%80%83` now returns 400; added that case to `testInvalidEncodingThrows`.
_🤖 Addressed by [Claude Code](https://claude.com/claude-code)_
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
