arrayfun.m reworked

[Bill Denney]

Thomas Treichl wrote:
> Hi,
>
> I have finished reworking arrayfun - finally I decided to find the bug 
> in arrayfun.m and not to send my arrayun.cc here. And thanks to David 
> for having a look at the cellfun problems I have had ;)
>
> You have been right John, my implementation arrayfun.cc is still 
> slower then the current implementation because I used the C++ subsref 
> and subsasgn methods for all types of input arguments.
>
> The attached changeset includes the following changes to arrayfun.m:
>  - created new help text
>  - fixed bug in arrayfun function
>  - added a lot of test to check my compatibility needs
>
> I hope it is good enough to be included into the default development 
> branch.
>
> Regards,
>
>   Thomas
>
> PS. I have created a similiar large testset for cellfun, should I 
> prepare a patch for this or is this some kind of overshooting?
Wow, that's a lot of tests.  Out of curiosity, was the bug that it would 
send an empty cell when it shouldn't?

Have a good day,

Bill