This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug fortran/83344] Use of uninitialized memory with ASSOCIATE and strings
- From: "jb at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Thu, 14 Dec 2017 21:14:06 +0000
- Subject: [Bug fortran/83344] Use of uninitialized memory with ASSOCIATE and strings
- Auto-submitted: auto-generated
- References: <bug-83344-4@http.gcc.gnu.org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83344
--- Comment #9 from Janne Blomqvist <jb at gcc dot gnu.org> ---
With the patch in #2 (which in this case is equivalent to your patch in #8) I
get on my charlen->size_t branch:
❯ gfortran -O0 -Wall -c a22.f90
a22.f90:20:0:
associate(w4 => trim(s))
Warning: ‘.w4’ is used uninitialized in this function [-Wuninitialized]
a22.f90:20:0:
associate(w4 => trim(s))
Error: size of variable ‘w4’ is too large
The warning is entirely correct, if one looks at the 003t.original dump one can
see that .w4 is never assigned to:
foo ()
{
character(kind=1) s[1:4];
{
integer(kind=8) .w4;
character(kind=1) w4[1:.w4];
{
character(kind=1) * pstr.0;
integer(kind=8) len.1;
sizetype D.3727;
sizetype D.3728;
_gfortran_string_trim (&len.1, (void * *) &pstr.0, 4, &s);
D.3727 = (sizetype) len.1;
D.3728 = (sizetype) .w4;
if (D.3728 != 0)
{
__builtin_memmove ((void *) &w4, (void *) pstr.0, (unsigned long)
MIN_EXPR <NON_LVALUE_EXPR <D.3728>, NON_LVALUE_EXPR <D.3727>>);
if (NON_LVALUE_EXPR <D.3727> < NON_LVALUE_EXPR <D.3728>)
{
__builtin_memset ((void *) &w4 + NON_LVALUE_EXPR <D.3727>, 32,
(unsigned long) (NON_LVALUE_EXPR <D.3728> - NON_LVALUE_EXPR <D.3727>));
}
}
if (len.1 > 0)
{
__builtin_free ((void *) pstr.0);
}
}
L.1:;
}
}
What the "w4 is too large" error message means I'm not sure. It might be that
the uninitialized kind=8 .w4 value is large enough to trigger some kind of
"this must be non-sensical" error in the middle-end.