This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: g++ and aliasing bools
- From: Mark Mitchell <mark at codesourcery dot com>
- To: Daniel Berlin <dan at dberlin dot org>
- Cc: Dan Nicolaescu <dann at godzilla dot ICS dot UCI dot EDU>, "gcc at gcc dot gnu dot org" <gcc at gcc dot gnu dot org>
- Date: Thu, 24 Jan 2002 14:29:14 -0800
- Subject: Re: g++ and aliasing bools
- References: <Pine.LNX.4.44.0201241700080.26628-100000@dberlin.org>
> I'm confused as to why you think it's such a tricky thing to do C++
> aliasing that it requires any kind of formal sketch of a proof.
I'm confused as to why you think we shouldn't prove things if we can. :-)
This is a change that can cause very hard to debug problems, both for
users and for us. We should try to be sure we know what we're doing
before we do it.
Things are not very simple in C++ with the new IA64 ABI class layout.
Earlier C++ compilers *literally* translated C++ into C using nested
structs, which would have made things pretty simple. But here we
can have multiple types overlaying the same location.
It's not that the idea won't work -- it might. But it needs to be
proven first.
If it's not a hard proof, what's the big deal? And if it is a hard
proof, that's just an indicating that it's worth doing before we
make the change.
--
Mark Mitchell mark@codesourcery.com
CodeSourcery, LLC http://www.codesourcery.com