[Bug c++/70403] A null pointer check removed with -O2 even with -fno-delete-null-pointer-checks

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70403

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|---                         |INVALID

--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> ---
The reduced testcase doesn't have the pattern you describe in the description.

With the preprocessed source I can confirm your observation (the NULL pointer
check is gone) with GCC 4.9.2.  With GCC 5 the whole function
NewLogicalSwitch::CpSelect is optimized to

  <bb 2>:
  _profile63.mpRecord = 0B;
  _profile63.mThisBufferId = 0;
  msEnabled.9_15 = msEnabled;
  if (msEnabled.9_15 != 0)
    goto <bb 3>;
  else
    goto <bb 4>;

  <bb 3>:
  ProfilePoint::Enter (&_profile63, &__PRETTY_FUNCTION__, 0);

  <bb 4>:
  tbrGet ();

  <bb 5>:
  *rFailureReason_5(D) = 0;
  __builtin_unreachable ();

<L125>:
  msEnabled.10_16 = msEnabled;
  if (msEnabled.10_16 != 0)
    goto <bb 7>;
  else
    goto <bb 8>;

  <bb 7>:
  ProfilePoint::Exit (&_profile63);

  <bb 8>:
  _14 = __builtin_eh_pointer (2);
  __builtin_unwind_resume (_14);

which means there is something wrong with your code.  Very likely it is

    McCpId* pCpId = static_cast<McCpId*>(&newCpId);

and it's further use.  You can't use an object of McId via a pointer to
McCpId but you call

    pCpId->toString (cpIdBuf, sizeof (cpIdBuf));

on it.