Client still comes with error prone hashValue() Method

[GoogleCodeExporter]

current hashValue() code is:

[code]
    @Override
    public int hashCode() {
        if (StringUtils.isNumeric(id)) {
            return Integer.valueOf(id);
        } else {
            return id.hashCode();
        }
    }
[/code]

StringUtils.isNumeric does not guarantee that Integer.valueOf() works, since in 
case of high numeric values a NumberFormatException will be thrown nevertheless.

Apart of that i doubt that testing on numeric plus converting into integer will 
be less complex than the default String hashValue().

Suggestion is to simply always use the String default hashValue() if there is 
no benchmark showing that it is grossly inefficient.

Original issue reported on code.google.com by S.No...@gmail.com on 2 Dec 2013 at 7:01

[GoogleCodeExporter]

correction, the method is hashCode() as shown in the code snippet.

A short test confirmed that the default hashCode() in String outperforms the 
current override by orders of magnitude. (test run over random Strings/Numeric 
Strings with 1000 calls to the hashCode calculation, the default method even 
outperformed a djb hash algorithm i used as reference value)

Original comment by S.No...@gmail.com on 2 Dec 2013 at 7:43

[GoogleCodeExporter]

This issue was closed by revision r4738.

Original comment by mondain on 5 Dec 2013 at 12:15

• Changed state: Fixed

[GoogleCodeExporter]

Original comment by mondain on 5 Dec 2013 at 12:16