Re: PSPP-BUG: bug in binomial_execute()?

[Ben Pfaff]

Thanks.

I don't see a change from "break" to "continue" in the loop.  Was
that part actually correct?  (I guess that the "break" is really
in do_binomial(), not binomial_execute().)

John Darrington <address@hidden> writes:

> I pushed a fix for this.  Thanks for reporting it.
>
> J'
>
> On Sat, Jan 17, 2009 at 09:53:18PM -0800, Ben Pfaff wrote:
>      Ben Pfaff <address@hidden> writes:
>      
>      > The binomial_execute() function allocates two arrays of frequency
>      > structures.  Then, if the caller passed in category values, it
>      > initializes just the first element of each array with those
>      > category values.  I think that this is incorrect: it should
>      > initialize every element of these arrays with the category
>      > values.
>      >
>      > Am I right?
>      
>      Also, should the "break;" in that function be a "continue;"?  I
>      think that, as is, it skips processing all the remaining
>      variables when one missing value is found.

-- 
Ben Pfaff 
http://benpfaff.org