FULL PRODUCT VERSION :
java version 1.7.0_02


ADDITIONAL OS VERSION INFORMATION :
Independent of OS. Error in source code.

A DESCRIPTION OF THE PROBLEM :
In the sourcecode for java.util.GregorianCalendar there is a comment starting at line 411 concerning minimum and maximum values for fields retrieved with constants ERA, YEAR, MONTH... The minimum for DAY_OF_WEEK_IN_MONTH is -1 according to line 424. But in the declaration of MIN_VALUES[] which is following the comment, there is the line 445: 1, // DAY_OF_WEEK_IN_MONTH. Is the comment with the minimum value of -1, or the respective element in MIN_VALUES[] correct?
I'll paste the whole questionable comment with the following decleration here (from GregorianCalendar.java):
 /*
     * <pre>
     * Greatest Least
     * Field name Minimum Minimum Maximum Maximum
     * ---------- ------- ------- ------- -------
     * ERA 0 0 1 1
     * YEAR 1 1 292269054 292278994
     * MONTH 0 0 11 11
     * WEEK_OF_YEAR 1 1 52* 53
     * WEEK_OF_MONTH 0 0 4* 6
     * DAY_OF_MONTH 1 1 28* 31
     * DAY_OF_YEAR 1 1 365* 366
     * DAY_OF_WEEK 1 1 7 7
     * DAY_OF_WEEK_IN_MONTH -1 -1 4* 6
     * AM_PM 0 0 1 1
     * HOUR 0 0 11 11
     * HOUR_OF_DAY 0 0 23 23
     * MINUTE 0 0 59 59
     * SECOND 0 0 59 59
     * MILLISECOND 0 0 999 999
     * ZONE_OFFSET -13:00 -13:00 14:00 14:00
     * DST_OFFSET 0:00 0:00 0:20 2:00
     * </pre>
     * *: depends on the Gregorian change date
     */
    static final int MIN_VALUES[] = {
        BCE, // ERA
        1, // YEAR
        JANUARY, // MONTH
        1, // WEEK_OF_YEAR
        0, // WEEK_OF_MONTH
        1, // DAY_OF_MONTH
        1, // DAY_OF_YEAR
        SUNDAY, // DAY_OF_WEEK
        1, // DAY_OF_WEEK_IN_MONTH
        AM, // AM_PM
        0, // HOUR
        0, // HOUR_OF_DAY
        0, // MINUTE
        0, // SECOND
        0, // MILLISECOND
        -13*ONE_HOUR, // ZONE_OFFSET (UNIX compatibility)
        0 // DST_OFFSET
    };


REPRODUCIBILITY :
This bug can be reproduced always.