Bug in java.time.Duration

Question

I need to parse Durations from strings. Java 8 provide a method for that taking the ISO-8601 standard as a basis:

Duration.parse("p10d"); // parses as ten days
Duration.parse("pt1h"); // parses as one hour

As the standard states that "it is permitted to omit the 'T' character by mutual agreement" some of the Javadoc examples of Durations.parse() leave out the T . According to them, the following expression should parse as "-6 hours and +3 minutes":

"P-6H3M"

But I found that all expressions ommitting the T throw a DateTimeParseException . Is that a bug in the parse() method or am I missing something?

Answer 1

In the JavaDoc of parse() :

The ASCII letter "T" must occur before the first occurrence, if any, of an hour , minute or second section.

That means you have to include T whenever you use H , M , or S .

---

The examples are wrong though:

"P-6H3M"    -- parses as "-6 hours and +3 minutes"
"-P6H3M"    -- parses as "-6 hours and -3 minutes"
"-P-6H+3M"  -- parses as "+6 hours and -3 minutes"

Answer 2

The regular expression used by Duration.parse is:

private static final Pattern PATTERN =
        Pattern.compile("([-+]?)P(?:([-+]?[0-9]+)D)?" +
                "(T(?:([-+]?[0-9]+)H)?(?:([-+]?[0-9]+)M)?(?:([-+]?[0-9]+)(?:[.,]([0-9]{0,9}))?S)?)?",
                Pattern.CASE_INSENSITIVE);

The input P-6H3M is not matched by this regular expression. If it is changed to

"(T?(?:([-+]?[ ...

in the fourth line (note the ? after the T ), the examples match (test it on http://regexpal.com/ ).

So it looks like you have found an inconsistency between code and JavaDoc.