http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51713
Bug #: 51713
Summary: Type mismatch for polymorphic dummy arguments
depending on "use" path
Classification: Unclassified
Product: gcc
Version: 4.6.2
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: fortran
AssignedTo: unassig...@gcc.gnu.org
ReportedBy: b...@stellardeath.org
Created attachment 26204
--> http://gcc.gnu.org/bugzilla/attachment.cgi?id=26204
Test case, only compiles when changed according to the comments within
I get a compile-time type-mismatch error for a subroutine usage depending on
how I import the module containing the derived type declaration.
The subroutine is declared to return a class(mytype_t) pointer, it does so by
associating a pointer to a type(mytype_specific_t), an extension of
type(mytype_t).
The definition of type(mytype_specific_t) is in another module, if I now "use"
this module from within the subroutine itself, I get the following error:
$> gfortran -g3 -fcheck=all -pedantic typetest.f90 -o typetest
typetest.f90:42.18:
call get_mytype(mytype)
1
Error: Actual argument to 'mytype' at (1) must have the same declared type
However, If I put the "use" statement in the module enclosing thus subroutine,
it compiles. I think this is an error, as the definition of
type(mytype_specific_t) should not be necessary outside of the subroutine.
As a source file says more than thousand words, see the attached file on which
I get the error.
Regards,
Lorenz
