[Patch, fortran] PR38324 - Wrong lbound given to allocatable components

[Dominique Dhumieres]

Paul,

Your patch works as advertised on my tests. Regtesting passes but for:

FAIL: gfortran.dg/alloc_comp_basics_1.f90  -O0  scan-tree-dump-times original "builtin_free" 26
FAIL: gfortran.dg/alloc_comp_basics_1.f90  -O1  scan-tree-dump-times original "builtin_free" 26
FAIL: gfortran.dg/alloc_comp_basics_1.f90  -O2  scan-tree-dump-times original "builtin_free" 26
FAIL: gfortran.dg/alloc_comp_basics_1.f90  -O3 -fomit-frame-pointer  scan-tree-dump-times original "builtin_free" 26
FAIL: gfortran.dg/alloc_comp_basics_1.f90  -O3 -fomit-frame-pointer -funroll-loops  scan-tree-dump-times original "builtin_free" 26
FAIL: gfortran.dg/alloc_comp_basics_1.f90  -O3 -fomit-frame-pointer -funroll-all-loops -finline-functions  scan-tree-dump-times original "builtin_free" 26
FAIL: gfortran.dg/alloc_comp_basics_1.f90  -O3 -g  scan-tree-dump-times original "builtin_free" 26
FAIL: gfortran.dg/alloc_comp_basics_1.f90  -Os  scan-tree-dump-times original "builtin_free" 26

Since I don't see why the patch should change the memory allocation/deallocation
the patch for gfortran.dg/alloc_comp_basics_1.f90: Count of builtin is 26
seems a leftover from something else. (Also I don't see why this test as a 
dg-options "-O2).

Thanks for the patch,

Dominique