[Bug c/82765] [6/7/8 Regression] ICE at -Os on valid code on x86_64-linux-gnu: in tree_to_shwi, at tree.c:6611

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82765

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code
           Priority|P3                          |P2
          Component|tree-optimization           |c
            Version|unknown                     |8.0
   Target Milestone|---                         |6.5

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
I believe this is a frontend bug.  We're parsing the literal into a __int128
type and using that to index the array.

IMHO the FE should truncate array indices to a type equal or smaller than
sizetype 
given the domain type of the array is

 <integer_type 0x7ffff68b4dc8
    type <integer_type 0x7ffff68a9000 sizetype public unsigned DI
        size <integer_cst 0x7ffff6891cd8 constant 64>
        unit-size <integer_cst 0x7ffff6891cf0 constant 8>
        align:64 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type
0x7ffff68a9000 precision:64 min <integer_cst 0x7ffff6891d08 0> max <integer_cst
0x7ffff68934c0 18446744073709551615>>
    DI size <integer_cst 0x7ffff6891cd8 64> unit-size <integer_cst
0x7ffff6891cf0 8>
    align:64 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type
0x7ffff68b4dc8 precision:64 min <integer_cst 0x7ffff6891d08 0> max <integer_cst
0x7ffff6891d08 0>>