Re: Extending octave_value::is_bool_type()

[Jordi Gutiérrez Hermoso]

2011/5/4 John W. Eaton <address@hidden>:
> On  4-May-2011, Jordi Gutiérrez Hermoso wrote:
>
> | 2011/5/4 Olaf Till <address@hidden>:
> | > On Tue, May 03, 2011 at 01:12:23PM -0500, Jordi Gutiérrez Hermoso wrote:
> | >> I'm running into this in my bwlabeln implementation. It would be
> | >> convenient if is_bool_type() returned true for matrices of type double
> | >> that only have zeros and ones.  Is this a bad idea? Can I implement
> | >> it?
> | >
> | > But doesn't one need e.g. to discriminate between matrices of numeric
> | > indices, which can be all ones, and logical indices?
> |
> | Ah, I see, yes. I hadn't considered this case. In which case my
> | proposed extension is a bad idea. Never mind, then, I'll do my work
> | another way.
>
> Do the any_element_not_one_or_zero functions defined in the NDArray,
> Matrix, Sparse, floatMatrix, floatNDArray, and intNDArray<T> classes
> help you?

Yes, those fit the bill, thanks. I hadn't seen them.

Maybe this will motivate me to retake that project of providing
Doxygen documentation for liboctave classes.

- Jordi G. H.