DateTimeScalar incorrectly using NANO_OF_SECOND

[nathanclayton]

Describe the bug
In src/main/java/graphql/scalars/datetime/DateTimeScalar.java, it's incorrectly using three digits of NANO_OF_SECOND in the DateTimeFormatter. If It's a three digit fraction of a second, it should instead be MILLI_OF_SECOND.

Original:

    private static DateTimeFormatter getCustomDateTimeFormatter() {
        return new DateTimeFormatterBuilder()
                .parseCaseInsensitive()
                .append(ISO_LOCAL_DATE)
                .appendLiteral('T')
                .appendValue(HOUR_OF_DAY, 2)
                .appendLiteral(':')
                .appendValue(MINUTE_OF_HOUR, 2)
                .appendLiteral(':')
                .appendValue(SECOND_OF_MINUTE, 2)
                .appendFraction(NANO_OF_SECOND, 3, 3, true)
                .appendOffset("+HH:MM", "Z")
                .toFormatter();
    }

Should be:

    private static DateTimeFormatter getCustomDateTimeFormatter() {
        return new DateTimeFormatterBuilder()
                .parseCaseInsensitive()
                .append(ISO_LOCAL_DATE)
                .appendLiteral('T')
                .appendValue(HOUR_OF_DAY, 2)
                .appendLiteral(':')
                .appendValue(MINUTE_OF_HOUR, 2)
                .appendLiteral(':')
                .appendValue(SECOND_OF_MINUTE, 2)
                .appendFraction(MILLI_OF_SECOND, 3, 3, true)
                .appendOffset("+HH:MM", "Z")
                .toFormatter();
    }

[nathanclayton]

I understand that appendFraction truncates the string without rounding, giving the same value as using MILLI_OF_SECOND. However, this is a consistency thing than more anything else, as it could be confusing to someone who isn't familiar with the DateTimeFormatterBuilder and the truncation rules of the appendFraction method.

[dondonz]

Hello, that's a good point, it's much easier to understand the code to use milliseconds, after all they are meant to be milliseconds there. Otherwise someone might mistakenly think these are nanoseconds.

Would you like to open a PR to improve this?