Suppress -Warray-bounds false positive in RCLASS_WRITE_SUPER

[k0kubun]

After 4bb1977 switched RCLASS_PRIME_CLASSEXT_WRITABLE_P to FL_TEST_RAW, GCC lost the RB_FL_ABLE-derived hint that klass is not a special const, and started warning that RCLASS_EXT_PRIME(klass) (used by the inlined RCLASS_WRITE_SUPER in class_associate_super) could yield a near-zero pointer. Make the existing VM_ASSERT(klass != 0) precondition visible to the optimizer via RBIMPL_ASSUME so the warning goes away without re-introducing the FL_TEST overhead.

Example warning:

In file included from ../include/ruby/internal/core/rarray.h:32,
                   from ../include/ruby/internal/core.h:23,
                   from ../include/ruby/ruby.h:29,
                   from ../constant.h:13,
                   from ../class.c:20:
  In function ‘rb_obj_write’,
      inlined from ‘RCLASS_WRITE_SUPER’ at ../internal/class.h:484:5,
      inlined from ‘class_associate_super’ at ../class.c:646:9,
      inlined from ‘rb_class_set_super’ at ../class.c:655:12:
  ../include/ruby/internal/gc.h:672:11: warning: array subscript 1 is outside array bounds of ‘rb_classext_t[0]’ {aka ‘struct rb_classext_struct[]’} [-Warray-bounds=]
    672 |     *slot = b;
        |     ~~~~~~^~~
  In function ‘rb_class_set_super’:
  cc1: note: source object is likely at address zero
  Unpacked gems/debug-1.11.1.gem
  In function ‘rb_obj_write’,
      inlined from ‘RCLASS_WRITE_SUPER’ at ../internal/class.h:484:5,
      inlined from ‘class_associate_super.constprop’ at ../class.c:646:9:
  ../include/ruby/internal/gc.h:672:11: warning: array subscript 1 is outside array bounds of ‘rb_classext_t[0]’ {aka ‘struct rb_classext_struct[]’} [-Warray-bounds=]
    672 |     *slot = b;
        |     ~~~~~~^~~
  In function ‘class_associate_super.constprop’:
  cc1: note: source object is likely at address zero
  transdb.h unchanged
  At top level:
  cc1: note: unrecognized command-line option ‘-Wno-self-assign’ may have been intended to silence earlier diagnostics
  cc1: note: unrecognized command-line option ‘-Wno-parentheses-equality’ may have been intended to silence earlier diagnostics
  cc1: note: unrecognized command-line option ‘-Wno-constant-logical-operand’ may have been intended to silence earlier diagnostics