noalias macro bug [was: Linux and aliasing?]

[mark@codesourcery.com]

>>>>> "Jamie" == Jamie Lokier <egcs@tantalophile.demon.co.uk> writes:

    Jamie> mark@codesourcery.com wrote:
    >> First, it's silly to keep saying this.  We're outside standard
    >> behavior, so "proof" based on standardese is not an issue.
    >> Documented behavior is one issue, and the code in GCC itself is
    >> another issue.

    Jamie> Aren't we talking about conformant code?

No.  There's *no* conformant way to do the kind of thing Linus wants,
except to use memcpy, or other means of copying through a char*.

    Jamie> I still don't get the wording.  Do you mean "permitted to
    Jamie> access *any objects with the same type* as the unionized
    Jamie> fields"

Yes.

    Jamie> I'm worried when you say `noalias' is outside standard
    Jamie> behaviour, since it means we're depending on undocumented
    Jamie> GCC-specific behaviour.  Clearly we don't want to do that.

Then you have to use memcpy.

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com