Hi, I discovered a problem with gfortran 4.3.1 when compiling code with pure function with pointer argument. Inside the pure function, the argument is passed further to an other function. This seems to work, as long this other function is an intrinsic function, but seems to fail for user defined pure functions. The effect is demonstrated by the code below. Best regards, Bálint -------------------------------------------------------- module TestPure implicit none type T1 character(10) :: str end type T1 contains pure function getT1Len(self) result(t1len) type(T1), pointer :: self integer :: t1len t1len = getStrLen(self%str) !t1len = len_trim(self%str) ! <- this would work end function getT1Len pure function getStrLen(str) result(length) character(*), intent(in) :: str integer :: length length = len_trim(str) end function getStrLen end module TestPure program Test use TestPure implicit none type(T1), pointer :: pT1 allocate(pT1) pT1%str = "test" write (*,*) getT1Len(pT1) deallocate(pT1) end program Test ------------------------------------------------------
CONFIRM. Works with other compilers. The constraint checked for is: "C1272 In a pure subprogram any designator with a base object that is in common or accessed by host or use association, is a dummy argument of a pure function, is a dummy argument with INTENT (IN) of a pure subroutine, or an object that is storage associated with any such variable, shall not be used in the following contexts: [...] (5) As an actual argument associated with a dummy argument with INTENT (OUT) or INTENT (INOUT) or with the POINTER attribute." (From F2003.) At a glance, it seems as if this applies, but as the non-normative note indicates, C1272 does not apply in this case: "NOTE 12.45 Pure subroutines are included to allow subroutine calls from pure procedures in a safe way [...] The constraints for pure subroutines are based on the same principles as for pure functions, except that side effects to INTENT (OUT), INTENT (INOUT), and pointer dummy arguments are permitted." However, I failed to find this in the normative part; I will try a bit more to figure out why C1272 does not apply.
(In reply to comment #1) This is due to an error in interface.c. The following fixes the fault, bootstraps and regtests OK. I will apply as obvious just as soon as I can. Paul Index: gcc/fortran/interface.c =================================================================== *** gcc/fortran/interface.c (revision 136870) --- gcc/fortran/interface.c (working copy) *************** check_intents (gfc_formal_arglist *f, gf *** 2379,2385 **** return FAILURE; } ! if (a->expr->symtree->n.sym->attr.pointer) { gfc_error ("Procedure argument at %L is local to a PURE " "procedure and has the POINTER attribute", --- 2379,2385 ---- return FAILURE; } ! if (f->sym->attr.pointer) { gfc_error ("Procedure argument at %L is local to a PURE " "procedure and has the POINTER attribute",
Subject: Bug 36526 Author: pault Date: Wed Jun 25 23:04:33 2008 New Revision: 137125 URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=137125 Log: 2008-06-25 Paul Thomas <pault@gcc.gnu.org> PR fortran/36526 * interface.c (check_intents): Correct error where the actual arg was checked for a pointer argument, rather than the formal. 2008-06-25 Paul Thomas <pault@gcc.gnu.org> PR fortran/36526 * gfortran.dg/proc_formal_proc_2.f90: New test. Added: trunk/gcc/testsuite/gfortran.dg/pure_formal_proc_2.f90 Modified: trunk/gcc/fortran/ChangeLog trunk/gcc/testsuite/ChangeLog
this bug seems fixed in 4.4.0, should it be closed?
(In reply to comment #4) > this bug seems fixed in 4.4.0, should it be closed? > Joost, I forgot the PR and missed your prompt - I'll apply it to 4.3 and close it. Thanks Paul
Subject: Bug 36526 Author: pault Date: Thu Nov 27 22:20:27 2008 New Revision: 142248 URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=142248 Log: 2008-11-27 Paul Thomas <pault@gcc.gnu.org> PR fortran/36526 * interface.c (check_intents): Correct error where the actual arg was checked for a pointer argument, rather than the formal. 2008-11-27 Paul Thomas <pault@gcc.gnu.org> PR fortran/36526 * gfortran.dg/pure_formal_proc_2.f90: New test. Added: branches/gcc-4_3-branch/gcc/testsuite/gfortran.dg/pure_formal_proc_2.f90 Modified: branches/gcc-4_3-branch/gcc/fortran/ChangeLog branches/gcc-4_3-branch/gcc/fortran/interface.c branches/gcc-4_3-branch/gcc/testsuite/ChangeLog
Fixed on trunk and 4.3. Thanks for the report. Paul