https://gcc.gnu.org/bugzilla/show_bug.cgi?id=38220
Joost VandeVondele <Joost.VandeVondele at mat dot ethz.ch> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
CC| |burnus at gcc dot gnu.org,
| |Joost.VandeVondele at mat dot
ethz
| |.ch
Resolution|FIXED |---
--- Comment #6 from Joost VandeVondele <Joost.VandeVondele at mat dot ethz.ch>
---
I believe this introduced a regression, C_LOC is not pure.
Just quoting from here https://software.intel.com/en-us/forums/topic/269061
The "standard intrinsic procedures" are those listed in section 13.7 of F2003.
Procedures in intrinsic modules are not included. I happen to be writing this
from a meeting of the Fortran standards committee (J3) and posed this question
- the response from one member was "because we forgot to issue the Ivory Soap
certificates". None of the procedures in ISO_C_BINDING are pure.
The example why this restriction is a good idea (inspired by PR64247) is the
fact that if it where, any pure procedure could return different results even
if the variables have the same value (as opposed to address):
!
! are functions containing C_LOC pure ?
!
MODULE M1
CONTAINS
PURE INTEGER*8 FUNCTION F(a) RESULT(I)
USE ISO_C_BINDING, ONLY: C_LOC
INTEGER*8, INTENT(IN), TARGET :: a
I=TRANSFER(C_LOC(a),I)
END FUNCTION
END MODULE M1
PROGRAM TEST
USE M1
INTEGER*8 :: i,ri
INTEGER*8 :: j,rj
i=4 ; j=4
ri=F(i)
rj=F(j)
IF (ri.NE.rj) CALL ABORT()
END PROGRAM
ifort rejects this code, gfortran and cft compile but trigger the runtime
abort, pgf90 compiles and runs without abort.
CCing Tobias as he approved the patch...
