«

Eg

class Foo
{
static const float f = 0.0f; // error
static const int i = 0; // ok
};

What's the reasoning behind this seemingly arbitrary limitation?

Martin

Martin Rennix wrote:
> class Foo
> {
> static const float f = 0.0f; // error
> static const int i = 0; // ok
> };
>
> What's the reasoning behind this seemingly arbitrary limitation?

I would guess portability, but don't quote me on that.

Ask in comp.std.c++ -- they hold/discuss the reasoning behind any
decisions that made it to the Standard.

V
--
Please remove capital 'A's when replying by e-mail
I do not respond to top-posted replies, please don't ask

Martin Rennix wrote:
> Eg
>
> class Foo
> {
> static const float f = 0.0f; // error
> static const int i = 0; // ok
> };
>
> What's the reasoning behind this seemingly arbitrary limitation?
>
> Martin
>
The compiler is only required to know how to do integer math at compile
time.

On Apr 18, 8:41 am, Ron Natalie <r...@spamcop.net> wrote:
> MartinRennixwrote:
> > Eg
>
> > class Foo
> > {
> > static const float f = 0.0f; // error
> > static const int i = 0; // ok
> > };
>
> > What's the reasoning behind this seemingly arbitrary limitation?
>
> > Martin
>
> The compiler is only required to know how to do integer math at compile
> time.

Is that because in the "olden days" you couldn't guarantee that a FPU
would be available, or the equivalent software floating point library?

On Apr 18, 5:20 am, Martin Rennix <martin.ren...@gmail.com> wrote:
> On Apr 18, 8:41 am, Ron Natalie <r...@spamcop.net> wrote:

> > MartinRennixwrote:
> > > Eg

> > > class Foo
> > > {
> > > static const float f = 0.0f; // error
> > > static const int i = 0; // ok
> > > };

> > > What's the reasoning behind this seemingly arbitrary limitation?

> > The compiler is only required to know how to do integer math at compile
> > time.

Also, the fact that there is no need for it. It's really a
hack, which was added for integers is because certain uses
require an integral constant expression.

> Is that because in the "olden days" you couldn't guarantee that a FPU
> would be available, or the equivalent software floating point library?

It's because even today, cross-compilers exist, and floating
point results depend on the actual floating point being used.
Something like 2. * 3.14 will have different values on different
machines. And the standard doesn't impose emulation of the
target floating point hardware.

--
James Kanze (GABI Software) email:
Conseils en informatique orientée objet/
Beratung in objektorientierter Datenverarbeitung
9 place Sémard, 78210 St.-Cyr-l'École, France, +33 (0)1 30 23 00 34

Martin Rennix wrote:

>
> Is that because in the "olden days" you couldn't guarantee that a FPU
> would be available, or the equivalent software floating point library?
>
I think it was primarily there for crosscompilers. Integer math is
pretty easy to do universally, floating point is a bit more involved.
Not everybody uses the IEEE fp math.