issue with large array size inputs for JavaBigDecimalParser.parseBigDecimal

[pjfanning]

This relates to FasterXML/jackson-databind#4694 (comment)

A Jackson user hit an issue with the test case below. The JavaBigDecimalParser.parseBigDecimal call works for the String input but with the char array input, the wrong result is generated. The wrong value is -1.1000E-648.

    @Test
    void issueDatabind4694() {
        final String str = "-11000.0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000";
        final BigDecimal expected = new BigDecimal(str);
        assertEquals(expected, JavaBigDecimalParser.parseBigDecimal(str));
        final char[] arr = str.toCharArray();
        assertEquals(expected, JavaBigDecimalParser.parseBigDecimal(arr, 0, arr.length));
    }

[cowtowncoder]

Without digging deeper, exception suggests like there was an off-by-one bug somewhere (one or more extra/missing digits).

[wrandelshofer]

Yes, this is a off-by-one bug. I have not properly kept up the code of the char-array implementations in sync with the byte-array and CharSequence implementations.

I am going to fix this asap.

[wrandelshofer]

Fixed in #81