Implement relops for arguments always resulting in false.

[awulkiew]

The purpose is to simplify writing generic code with relational operations. Consider e.g.:

bg::equals(linestring, polygon);
bg::touches(point, linestring);
bg::crosses(point, linestring);
bg::overlaps(point, linestring);

before: compilation error,
this PR: compiles and always returns false.

[vissarion]

That's useful, thanks!

[mloskot]

Nitpicking: & ) -> &) :-)

[awulkiew]

@mloskot Well idk ;) do we have gudelines for that? AFAIU we put space after & and before , so I assumed that the missing name is represented by space. Or am I mistaken and all of the spaces should be removed?

[barendgehrels]

AFAIK we don't have a written guideline for that.
I find this readable, but I'm also fine with &)

[mloskot]

Having nothing constructive to say, I went nit-picking and bike-shedding. Certainly, ignoreable

[barendgehrels]

Certainly nice to hear from you Mateusz!

[mloskot]

Barend, I'm happy to be revolving around still

[awulkiew]

Ok, I saw this in Geometry code but it is possible that it was me who put it in the first place. I don't have an opinion regarding it, but if it's ignorable I'll leave it as it is. ;)

[barendgehrels]

So this is only for geometry-type-combinations which can, by definition, never return true?

[awulkiew]

@barendgehrels Yes, other cases are already handled. There is no combination which always results in true.

[barendgehrels]

@barendgehrels Yes, other cases are already handled. There is no combination which always results in true.

Great, thanks

[barendgehrels]

AFAIK we don't have a written guideline for that.
I find this readable, but I'm also fine with &)

[awulkiew]

Thanks for reviews.