[J3] random pure quibble, was RE: Comments on paper 21-196 (F18/025)

[Malcolm Cohen]

> I find it unsettling that ALLOCATE

>statements can appear in pure

>code, but that DEALLOCATE

>statements cannot.

 

I would find that slightly worrying too, but I don’t see any normative text that supports that assertion.

 

For example,

 

module junk

contains

  pure subroutine s(a,b)

    real,intent(Inout) :: a,b

    real,pointer :: c

    Allocate(c,Source=b)

    a = c

    Deallocate(c)

  end subroutine

end module

 

is accepted by three compilers I tried.

 

Is there something I missed?

 

Cheers,

-- 

..............Malcolm Cohen, NAG Oxford/Tokyo.

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.j3-fortran.org/pipermail/j3/attachments/20211027/3cb540e3/attachment.htm>