On 15 June 2013 14:58, Duncan Jones <dun...@wortharead.com> wrote:
> On 14 June 2013 07:55, Duncan Jones <dun...@wortharead.com> wrote:
>> On 13 June 2013 21:26, Phil Steitz wrote:
>>> On 6/13/13 1:03 AM, Duncan Jones wrote:
> <snip>
>>
>>> I am not sure where this little method would fit in [lang]; but it might
>>> make sense to add there in the simple form above (with RANDOM
>>> static, I assume).
>>
>> Indeed, I had assumed any addition to [lang] would just use a standard
>> source of randomness (i.e. a static Random or SecureRandom instance).
>> However, I'm struggling to see an existing class that could
>> accommodate the functionality and creating a new "RandomUtils" class
>> feels a bit heavyweight, unless others can think of additional methods
>> that would be useful.
>
> I gave it some further thought and realised there were some other
> methods that would prove useful in such a "RandomUtils" class. I've
> created https://issues.apache.org/jira/browse/LANG-900 and attached an
> implementation. I welcome comments.
In particular, I welcome comments on whether or not to constrain the
start and end range values to be non-negative.
In the current implementation (attached to the bug), I enforced
non-negative values, e.g. for `nextInt(X, Y)`, both X and Y must be
non-negative. This simplified the implementation because it meant that
"Y - X" couldn't be larger than the maximum value of an int. However,
this could be changed for nextInt with (untested) code such as:
public static int nextInt(int startInclusive, int endExclusive) {
Validate.isTrue(endExclusive >= startInclusive,
"Start value must be smaller or equal to end value.");
if (startInclusive == endExclusive) {
return startInclusive;
}
long range = endExclusive - startInclusive;
return (int) startInclusive + nextLong(0, range);
}
This may work for ints, but it gets more difficult for longs. Any
thoughts on this?
Duncan
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org
