[Bug fortran/103472] New: ICE in gfc_conv_ss_startstride, at fortran/trans-array.c:4527

[Bug fortran/103472] New: ICE in gfc_conv_ss_startstride, at fortran/trans-array.c:4527

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103472

            Bug ID: 103472
           Summary: ICE in gfc_conv_ss_startstride, at
                    fortran/trans-array.c:4527
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gscfq@t-online.de
  Target Milestone: ---

Affects versions down to at least r5 :


$ cat z1.f90
program p
   integer, parameter :: a(0) = 0
   print *, (a(1:1))
end


$ gfortran-12-20211128 -c z1.f90
z1.f90:3:15:

    3 |    print *, (a(1:1))
      |               1
Warning: Lower array reference at (1) is out of bounds (1 > 0) in dimension 1
z1.f90:3:15:

    3 |    print *, (a(1:1))
      |               1
Warning: Lower array reference at (1) is out of bounds (1 > 0) in dimension 1
z1.f90:3:20:

    3 |    print *, (a(1:1))
      |                    1
internal compiler error: in gfc_conv_ss_startstride, at
fortran/trans-array.c:4527
0x856d02 gfc_conv_ss_startstride(gfc_loopinfo*)
        ../../gcc/fortran/trans-array.c:4527
0x8b3535 gfc_trans_transfer(gfc_code*)
        ../../gcc/fortran/trans-io.c:2645
0x850627 trans_code
        ../../gcc/fortran/trans.c:2136
0x8b0ece build_dt
        ../../gcc/fortran/trans-io.c:2026
0x850607 trans_code
        ../../gcc/fortran/trans.c:2108
0x8795fe gfc_generate_function_code(gfc_namespace*)
        ../../gcc/fortran/trans-decl.c:7644
0x7fc90e translate_all_program_units
        ../../gcc/fortran/parse.c:6638
0x7fc90e gfc_parse_file()
        ../../gcc/fortran/parse.c:6925
0x84967f gfc_be_parse_file
        ../../gcc/fortran/f95-lang.c:216

[Bug fortran/103472] ICE in gfc_conv_ss_startstride, at fortran/trans-array.c:4527

[mikael at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103472

Mikael Morin <mikael at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mikael at gcc dot gnu.org

--- Comment #1 from Mikael Morin <mikael at gcc dot gnu.org> ---
Created attachment 51982
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51982&action=edit
draft patch

We could workaround it by replacing warnings with errors.
Are there cases where we do want warnings, not errors?

[Bug fortran/103472] ICE in gfc_conv_ss_startstride, at fortran/trans-array.c:4527

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103472

anlauf at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2021-12-12
             Status|UNCONFIRMED                 |NEW
           Priority|P3                          |P4
     Ever confirmed|0                           |1
                 CC|                            |anlauf at gcc dot gnu.org

--- Comment #2 from anlauf at gcc dot gnu.org ---
(In reply to Mikael Morin from comment #1)
> We could workaround it by replacing warnings with errors.
> Are there cases where we do want warnings, not errors?

I would consider the testcase as invalid, so an error appears ok.
But technically the lower bound of a is 1 and upper bound is 0,
so the error message is a bit misleading.

E.g. NAG gives:

Error: pr103472-z1.f90, line 2: First subscript (1) is greater than upper bound
(0) for array A

[Bug fortran/103472] ICE in gfc_conv_ss_startstride, at fortran/trans-array.c:4527

[mikael at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103472

Mikael Morin <mikael at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |14.0
      Known to fail|                            |13.2.0

--- Comment #3 from Mikael Morin <mikael at gcc dot gnu.org> ---
The example from comment #0 is now rejected, which avoids the problem:

comment_0.f90:3:15:

    3 |    print *, (a(1:1))
      |               1
Warning: Lower array reference at (1) is out of bounds (1 > 0) in dimension 1
comment_0.f90:3:15:

    3 |    print *, (a(1:1))
      |               1
Warning: Lower array reference at (1) is out of bounds (1 > 0) in dimension 1
comment_0.f90:3:15:

    3 |    print *, (a(1:1))
      |               1
Error: index in dimension 1 is out of bounds at (1)