gmpxx implicit conversion between types

Marc Glisse marc.glisse at inria.fr
Wed Mar 9 09:27:01 CET 2011


On Wed, 9 Mar 2011, Joerg Arndt wrote:

> * Torbjorn Granlund <tg at gmplib.org> [Mar 08. 2011 20:27]:
>> When dealing with "heavy" types like these, is it in the C++ spirit to
>> allow implicit conversions?  Will we do any good by disallowing them?
>
> Implicit: as few as possible.
>
> A good guidance are computer algebra systems.  E.g.:
> Divide an integer by an integer ==> rational

It's too late for that one, mpz_class has euclidean / and % operations, 
and it's useful to have a type with this behavior.

> Any operation of type with float ==> float.

Any operation with an mpf yields an mpf. On the other hand, mpz+double 
gives a mpz (just describing the current situation here).

> Another thing (if not already done):
> make _all_ constructors "explicit".

Constructors from a string or a mp*_srcptr are already explicit.
Constructors from int (and the likes) are implicit, and I can't really see 
anything wrong with that.

-- 
Marc Glisse


More information about the gmp-discuss mailing list