Review

The Rule constructor in org.joda.time.tz.ZoneInfoCompiler.java does not
check the number of elements in the StringTokenizer obtained from parsing
the timezone file. There is an assumption that the input TimeZone file
will always be valid, leading to runtime exceptions with no good error
message when the file is invalid. This pull request adds a potential fix
and a test for this issue.

• missing/condition/value_or_state

org/joda/time/tz/ZoneInfoCompiler.java

Rule(StringTokenizer)

class ZoneInfoCompiler.Rule {
        Rule(StringTokenizer st) {
            iName = st.nextToken().intern();
            iFromYear = parseYear(st.nextToken(), 0);
            iToYear = parseYear(st.nextToken(), iFromYear);
            if (iToYear < iFromYear) {
                throw new IllegalArgumentException();
            }
            iType = parseOptional(st.nextToken());
            iDateTimeOfYear = new DateTimeOfYear(st);
            iSaveMillis = parseTime(st.nextToken());
            iLetterS = parseOptional(st.nextToken());
        }

}

class TokenizerCount {
  String pattern(StringTokenizer st) {
    if (st.countTokens() < 1) {
      throw new IllegalArgumentException("too few tokens");
    }
    return st.nextToken();
  }
}