[Bug middle-end/78524] [7 regression] failure of ACATS c41104a at -O0

["rguenth at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78524

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
The MIN_EXPR it builds is correct now though.  Looks like a bug in Adas
max_size to me when invoked on

 <min_expr 0x7ffff5ea77d0
    type <integer_type 0x7ffff5e80dc8 p__char___XDLU_87__90
        type <integer_type 0x7ffff61e7498 character sizes-gimplified public
visited string-flag QI
            size <integer_cst 0x7ffff61e2ca8 constant visited 8>
            unit size <integer_cst 0x7ffff61e2cc0 constant visited 1>
            align 8 symtab 0 alias set 4 canonical type 0x7ffff61e7498
precision 8 min <integer_cst 0x7ffff61e2cf0 -128> max <integer_cst
0x7ffff61e2d20 127> context <translation_unit_decl 0x7ffff61ed168 D.15> RM min
<integer_cst 0x7ffff61e2cd8 0> RM max <integer_cst 0x7ffff61e2c78 255> debug
type <integer_type 0x7ffff61e73f0 character>
            pointer_to_this <pointer_type 0x7ffff6202000 access_character>>
        public unsigned string-flag QI size <integer_cst 0x7ffff61e2ca8 8> unit
size <integer_cst 0x7ffff61e2cc0 1>
        align 8 symtab 0 alias set 4 canonical type 0x7ffff5e80dc8 precision 8
min <integer_cst 0x7ffff5e967c8 0> max <integer_cst 0x7ffff6238300 255> context
<function_decl 0x7ffff5ea0e00 p> RM size <integer_cst 0x7ffff5e96828 7> RM min
<integer_cst 0x7ffff5e967f8 87> RM max <integer_cst 0x7ffff5e96810 90>
        chain <type_decl 0x7ffff6207ab0 p__char___XDLU_87__90>>

    arg 0 <component_ref 0x7ffff5ea6270 type <integer_type 0x7ffff5e80dc8
p__char___XDLU_87__90>
        visited
        arg 0 <placeholder_expr 0x7ffff5e969f0 type <record_type 0x7ffff5ea4bd0
p__rec>
            visited>
        arg 1 <field_decl 0x7ffff5ea5688 d type <integer_type 0x7ffff5e80dc8
p__char___XDLU_87__90>
            visited nonaddressable decl_4 VOID file p.adb line 7 col 13
            align 8 offset_align 1 context <record_type 0x7ffff5ea4bd0 p__rec>
initial <integer_cst 0x7ffff61e2cc0 1>>>
    arg 1 <integer_cst 0x7ffff5e96ab0 type <integer_type 0x7ffff5e80dc8
p__char___XDLU_87__90> constant 88>>

we end up with mismatched type lhs/rhs.  The placeholder is transformed to

(gdb) p debug_tree (lhs)
 <integer_cst 0x7ffff5e967f8 type <integer_type 0x7ffff61e7498 character>
constant 87>

a signed integer constant.