From: Bruno Luong on
I have this piece of code that crashes my Matlab 2010a 64 bits, Vista.

a=zeros(1,2e4);
a(1000:1000:end)=NaN;
a(logical(0./a)) = 0;

With 2007a, it throw an error of the third statement "NaN's cannot be converted to logicals."

Can someone else confirm so I can report to TMW as a bug?

Admittedly this piece of code is strange, but the crashes indicates something is not right in the built-in basic library.

Bruno
From: Oliver Woodford on
"Bruno Luong" wrote:
> I have this piece of code that crashes my Matlab 2010a 64 bits, Vista.
>
> a=zeros(1,2e4);
> a(1000:1000:end)=NaN;
> a(logical(0./a)) = 0;
>
> With 2007a, it throw an error of the third statement "NaN's cannot be converted to logicals."
>
> Can someone else confirm so I can report to TMW as a bug?
>
> Admittedly this piece of code is strange, but the crashes indicates something is not right in the built-in basic library.
>
> Bruno

This also crashes my MATLAB 2010a on 32 bit Windows XP.
Oliver
From: Jan Simon on
Dear Bruno!

> I have this piece of code that crashes my Matlab 2010a 64 bits, Vista.
>
> a=zeros(1,2e4);
> a(1000:1000:end)=NaN;
> a(logical(0./a)) = 0;
>
> With 2007a, it throw an error of the third statement "NaN's cannot be converted to logicals."
>
> Admittedly this piece of code is strange, but the crashes indicates something is not right in the built-in basic library.
>
> Bruno

Perhaps the level of strangeness can be reduced?
a = zeros(1, 5);
a(logical([0, 1, NaN, 0 / 0, 0 / NaN])) = 0
or:
b = 0;
b(logical(0 / b)) = 0
I do not have 2010a to test this.

In 2009a the expected error message "NaN's cannot be converted to logicals." appears, as you surely know already, Bruno.

Thanks Bruno for finding this bug! Jan
From: Bruno Luong on
"Jan Simon" <matlab.THIS_YEAR(a)nMINUSsimon.de> wrote in message <hpfij0$91v$1(a)fred.mathworks.com>...

>
> Perhaps the level of strangeness can be reduced?
> a = zeros(1, 5);
> a(logical([0, 1, NaN, 0 / 0, 0 / NaN])) = 0
> or:
> b = 0;
> b(logical(0 / b)) = 0

Hi Jan and Olivier,

Both of the above still work. The crash appears with larger size like in the original code.

I have reported the bug.

Thanks,

Bruno
From: Matt J on
"Oliver Woodford" <o.j.woodford.98(a)cantab.net> wrote in message <hpfh9q$i41$1(a)fred.mathworks.com>...

> This also crashes my MATLAB 2010a on 32 bit Windows XP.
======

R2009b as well.