https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81062
Bug ID: 81062
Summary: Allocate on assignment with reshape(eoshift(...))
Product: gcc
Version: unknown
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: fortran
Assignee: unassigned at gcc dot gnu.org
Reporter: tkoenig at gcc dot gnu.org
Target Milestone: ---
program main
real, dimension(:), allocatable :: r
real, dimension(3,3) :: a
call random_number(a)
r = reshape(eoshift(a,1),shape(r))
print *,size(r)
print *,r
end program main
$ gfortran resh.f90
$ $ ./a.out
0
$ valgrind ./a.out
==17442== Memcheck, a memory error detector
==17442== Copyright (C) 2002-2015, and GNU GPL'd, by Julian Seward et al.
==17442== Using Valgrind-3.12.0 and LibVEX; rerun with -h for copyright info
==17442== Command: ./a.out
==17442==
==17442== Conditional jump or move depends on uninitialised value(s)
==17442== at 0x4F96C30: _gfortran_reshape_r4 (reshape_r4.c:93)
==17442== by 0x400BA2: MAIN__ (in /home/ig25/Krempel/Cshift/Cs-test/a.out)
==17442== by 0x400D4A: main (in /home/ig25/Krempel/Cshift/Cs-test/a.out)
==17442==
0
The uninitialized variable is shape_data:
90 for (n = 0; n < rdim; n++)
91 {
92 shape_data[n] = shape->base_addr[n * GFC_DESCRIPTOR_STRIDE(shape,0)];
93 if (shape_data[n] <= 0)
94 {
Maybe we are missing something along the lines of the fix for PR 58883 here.
