[Bug tree-optimization/128591] fold_const_call folds strstr(s, "") to s during early CCP without calling check_nul_terminated_array, silently suppressing -Wstringop-overread on non-NUL-terminated arrays

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=128591

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |diagnostic, false-negative
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2026-11-10
     Ever confirmed|0                           |1

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed.  But the issue is really that late diagnostics are fragile - GCC is
an optimizing compiler first and foremost, not a static analysis engine.

It sounds like we might want a -fdefer-optimizations-inhibiting-diagnostics
given that adding -W flags may not affect code generation.

Another option is to keep the original as

# DEBUG D.124 = __builtin_strstr (a, "");

or so and diagnose those.

Consider

const char *probe_empty_needle (const char *a)
{
  return __builtin_strstr(a, "");
}

where the exact 'a' might never be exposed or only exposed with LTO after
inlining.