[Bug c++/11517] [3.3/3.4 Regression]g++ fails to properly convert pointer expressions in conditional expressions.

["gdr at integrable-solutions dot net" <gcc-bugzilla@gcc.gnu.org>]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11517



------- Additional Comments From gdr at integrable-solutions dot net  2003-07-14 15:42 -------
Subject: Re:  [3.3/3.4 Regression]g++ fails to properly convert pointer expressions in conditional expressions.

"bangerth at dealii dot org" <gcc-bugzilla@gcc.gnu.org> writes:

| I am not at all sure that the code is legal, and even less that it is
| a regression. The standard says that the two operands shall have the same
| type, or that one should be convertible to the other one. This is not
| the case here, where both would have to be converted to pointer-to-base.
| 
| FWIW, icc doesn't accept the code either. Gaby, would you mind
| explaining on which part of 5.16 you base your opinion?

It is quite simple: I was presented with different behaviour of
GCC-2.95.x and GCC-3.3.1 on something which intuitively appeared clear
to me.  I was under the impression that "composite pointer type" can
also brings to common base class.  It appears not.  Sorry for the
confusion. 

-- Gaby