the observation is that profiling tools (like callgrind/valgrind) will assign a
fraction of the instructions to line 2 (PUBLIC S1) in a profile, similarly a
debugger will jump to line 2 if one steps through the execution of this
program. This is not particularly intuitive (since this line just specifies
an attribute and is not executed), and doesn't happen if e.g. ifort is used for
compilation.
MODULE M1
PUBLIC S1
CONTAINS
SUBROUTINE S1(A,n)
INTEGER :: n
INTEGER :: A(n,n,n,n)
A(1,1,1,1)=0
END SUBROUTINE S1
END MODULE M1
PROGRAM TEST
USE M1
INTEGER :: A(2,2,2,2),i
DO i=1,10000000
CALL S1(A,2)
ENDDO
END PROGRAM TEST
--
Summary: debug info points to unexpected line
Product: gcc
Version: 4.5.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: debug
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: jv244 at cam dot ac dot uk
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40823
