October 11, 2004
The following code (Linux, 0.102) demonstrates that an array of real's is not sorted correctly:

> import std.stdio;
>                                                                                 int main() {
>   real[1000] array;
>   for(int i=999; i>=0; i--)
>     array[999-i] = (cast(real)i)/8;
>                                                                                   real[] sorted_copy = array.dup;
>   sorted_copy.sort;
>                                                                                   for(int i=0; i<1000; i++)
>     writef("%f\n", sorted_copy[i]);
>                                                                                   return 0;
> }

The first several lines of ouput are

> 0.375000
> 0.000000
> 0.500000
> 0.625000
> 0.250000
> 0.125000
> 2.000000

October 13, 2004
Russ Lewis schrieb:
> The following code (Linux, 0.102) demonstrates that an array of real's
> is not sorted correctly:
>
> > import std.stdio;
> >
> > int main() {
> >   real[1000] array;
> >   for(int i=999; i>=0; i--)
> >     array[999-i] = (cast(real)i)/8;
> >
> >   real[] sorted_copy = array.dup;
> >   sorted_copy.sort;
> >
> >   for(int i=0; i<1000; i++)
> >     writef("%f\n", sorted_copy[i]);
> >
> >   return 0;
> > }
>
> The first several lines of ouput are
>
> > 0.375000
> > 0.000000
> > 0.500000
> > 0.625000
> > 0.250000
> > 0.125000
> > 2.000000

added to destress as svn://svn.kuehne.cn/dstress/run/sort_03.d

in addition the following striped down code segfaults:
# real[10] array;
# real[] copy = array.dup;
# copy.sort;

testcase svn://svn.kuehne.cn/dstress/run/dup_04.d

Thomas