This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [Fortran,Patch] PR32669 Fix actual argument bound checking
- From: "Tobias Burnus" <burnus at ph2 dot uni-koeln dot de>
- To: <burnus at ph2 dot uni-koeln dot de>
- Cc: <fortran at gcc dot gnu dot org>, <gcc-patches at gcc dot gnu dot org>, <burnus at net-b dot de>
- Date: Sun, 8 Jul 2007 12:56:32 +0200 (CEST)
- Subject: Re: [Fortran,Patch] PR32669 Fix actual argument bound checking
- References: <25841.212.202.24.200.1183883360.squirrel@www.webmail.uni-koeln.de>
Tobias Burnus wrote:
> which led to a [...] rejects-valid bug.
Actually not as only a warning is given and not an error.
Any suggestion, which checks should give an error and which only a
warning? Most compilers give an error for most checks (cf.
gfortran.dg/argument_check*), however, for- gfortran.fortran-torture/execute/st_function_1.f90
- gfortran.fortran-torture/execute/st_function.f90
only NAG f95 gives an error, g95 a warning and ifort gives not even a
warning.
(It is always wrong with regards to the Fortran standard [error], but one
can always imagine a program which works despite this, e.g. passing three
elements to a four-elements dummy, but only using three elements of this
array [-> warning?]. Something alike happens for the fortran-torture tests
above.)
Comments?
Tobias