Bogus standard conformance warning with operator(*) overloading

[Harald]

Hello,

ifort v12 gives a bogus warning for the following case:

% ifort -stand -c ifort-bogus-warning.f90
ifort-bogus-warning.f90(34): warning #6191: Fortran 2003 requires an arithmeticdata type in this context. [SIZE]
real :: z(size (x,1)*size (x,2)) ! <- ifort warns here
--------------------------^
ifort-bogus-warning.f90(34): warning #6191: Fortran 2003 requires an arithmeticdata type in this context. [SIZE]
real :: z(size (x,1)*size (x,2)) ! <- ifort warns here
-------------------------------------^

Here's the source:

ifort-bogus-warning.f90

When I comment out the interface operator(*), the warning disappears.
I think that the code is fully standard conforming.

Harald

[Harald]

I forgot to mention: commenting out the "interface size" part
makes the warning also disappear.

Harald

[Steven_L_Intel1]

Thanks - we'll take a look at this.

[Steven_L_Intel1]

Escalated as issue DPD200170370.

[Steven_L_Intel1]

Fixed in a future release.