Re: Infix "*" in ABSTRACT_STRING

[Paolo Redaelli]

Il 30/12/23 13:19, Eric Bezault ha scritto:

> On
>       30/12/2023 10:36, Patrick Cloarec wrote:
>       
> > It is unnatural to me to define a binary
> >         operator as the feature of the class of one object. infix is the
> >         less intuitive construction offered by Eiffel, in my humble
> >         opinion.
> >         
> >       
>       
>       They are available in one way or another in many
>       
>       other languages. For example, in Python:
>       
>       
>       >>> "Ho "*3+"! Merry Christmas"
>       
>       'Ho Ho Ho ! Merry Christmas'
>       
>       
>

I forgot to express my pleasant surprise on discovering you registered here.

It is not at all obvious nor expected, but very, very wellcome.

Having said that I think that this simple desire raise a legit issue about the expressiveness of Eiffel. I am also convinced that Raphael is right, such a "beautiful" feature shall not be in STRING nor in INTEGER.

Name it repeated infix "*" alias "×"


it seems to me that it actually belongs to


STRING is a NATIVELY_STORED_STRING which is an ABSTRACT_STRING that inherits: COMPARABLE, HASHABLE, SEARCHABLE, STORABLE, TRAVERSABLE

The more general the class where we put "repeated" the better.

But this also raise another question: why STRING is not a COLLECTION[CHARACTER]? 

Because def __mul__ in Python works also on lists and tuples... at a first glance it seems that this is a feature belonging to TRAVERSABLE or to INDEXABLE.

Let's say what can we see in FreeELKS (https://www.eiffel.org/files/doc/static/23.09/libraries/base/index.html BTW, it is quite convoluted to come here from eiffel.org...): neither STRING_32 or STRING_8 offer a feature like repeated (but it offers nice operators such as three_way_comparison alias "⋚" 🙂); they are INDEXABLE.

COLLECTION, a CONTAINER that has fill (other: CONTAINER [G]) that Fill the container with as many other items of other as possible.

So, as far as I can say both STRING and COLLECTION, both Liberty of FreeELKS are INDEXABLE. 

Does "repeated" query really belongs here?

Lastly, I want to reassure Raphael: no, I do not want to put this into curtiss! 

(but I would put alias "[]"  alongside item where meaningful)