kruskal_wallis_test() does not handle ties correctly

[John W. Eaton]

On 12-Feb-2008, Timo Lindfors wrote:

| the kruskal_wallis() issue I reported in 2005 at
| 
| http://velveeta.che.wisc.edu/octave/lists/archive//bug-octave.2005/msg01081.html
| 
| appears to have gone largerly forgotten, I still get wrong results
| with mercurial head. Could you please reconsider the suggested fixes
| represented in the thread? If the archive is to be the last message on
| the issue is
| 
| http://velveeta.che.wisc.edu/octave/lists/archive//bug-octave.2005/msg01310.html

I applied these changes.  Sorry for dropping them, it was not
intentional.

jwe