Hi!
It might be poorly understanding something, but I am under the impression that
some warning from '-Wall' only appear in conjunction with '-O2'.
Here is a simplified case which seems to exhibit this behaviour: do not look
for much meaning, in this very case the warning turns out to be a bit
unecessary, but no way the compiler could figure that out. Anyway, this is just
one example, and I believe that to be systematic.
It might be nice at some point if someone could fix that, but it's only
compilation diagnostics after all...
Thanks!
Philippe
PS: some sources which trigger this behaviour:
!
SUBROUTINE Check ( Number )
!
IMPLICIT NONE
INTEGER :: Value
INTEGER, INTENT(IN) :: Number
!
SELECT CASE ( Number )
CASE ( 1 )
Value = 11
CASE ( 2 )
Value = 12
CASE ( 3 )
Value = 13
CASE DEFAULT
CALL Some_Routine_Performing_Orderly_Stop
END SELECT
IF ( Value .LT. 0 ) THEN
CALL Some_Routine_Issueing_Some_Warnings
CALL Some_Routine_Performing_Orderly_Stop
ENDIF
!
RETURN
END SUBROUTINE Check
!
PPS: what I get compiling that:
gfortran -O2 -g -Wall -c Sources.f90
Sources.f90: In function 'check':
Sources.f90:18: warning: 'value' is used uninitialized in this function
whereas 'gfortran -g -Wall -c Sources.f90' gives no warning of any kind
PPPS: my build is not up to date but still relatively recent:
Using built-in specs.
Target: x86_64-unknown-linux-gnu
Configured with: /USER/philippe/Irix/Gcc_Sources/configure
--prefix=/WORK/philippe/Tools/Gcc --enable-languages=c,fortran
--disable-maintainer-mode --disable-shared
--with-mpfr=/WORK/philippe/Tools/Mpfr --with-gmp=/WORK/philippe/Tools/Gmp
Thread model: posix
gcc version 4.2.0 20060626 (experimental)
--
Summary: One warning seems to appear only with '-O2' even though
'-Wall' is specified
Product: gcc
Version: 4.2.0
Status: UNCONFIRMED
Severity: minor
Priority: P3
Component: fortran
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: P dot Schaffnit at access dot rwth-aachen dot de
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28200
