commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gilles (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MATH-1019) "shuffle" method broken (in class "o.a.c.m.random.RandomDataGenerator")
Date Fri, 09 Aug 2013 13:10:47 GMT

    [ https://issues.apache.org/jira/browse/MATH-1019?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13734748#comment-13734748
] 

Gilles commented on MATH-1019:
------------------------------

With the replacement, the unit test "testNextPermutation()" passes but it uses a fully shuffled
array where there is no difference in behaviour between the old (in "RandomDataGenerator")
and the fixed (in "MathArrays") code.

                
> "shuffle" method broken (in class "o.a.c.m.random.RandomDataGenerator")
> -----------------------------------------------------------------------
>
>                 Key: MATH-1019
>                 URL: https://issues.apache.org/jira/browse/MATH-1019
>             Project: Commons Math
>          Issue Type: Bug
>    Affects Versions: 3.2
>            Reporter: Gilles
>             Fix For: 3.3
>
>
> The method does not abide by its contract: elements before the "end" index are included
in the shuffle.
> {code}
> /**
>  * Uses a 2-cycle permutation shuffle to randomly re-order the last elements 
>  * of list.
>  *
>  * @param list list to be shuffled
>  * @param end element past which shuffling begins
>  */
> private void shuffle(int[] list, int end) {
>     int target = 0;
>     for (int i = list.length - 1; i >= end; i--) {
>         if (i == 0) { // XXX "0" should be "end"
>             target = 0; // XXX "0" should be "end"
>         } else {
>             // NumberIsTooLargeException cannot occur
>             target = nextInt(0, i); // XXX "0" should be "end"
>         }
>         int temp = list[target];
>         list[target] = list[i];
>         list[i] = temp;
>     }
> }
> {code}
> I'm going to introduce the above corrections in the new implementation to be located
in "MathArrays" (cf. issue MATH-1010).

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message