<addr_expr <realpart_expr <x>>> ?

[Diego Novillo]

On Sat, 2004-01-10 at 20:18, Richard Henderson wrote:
> So, like, where did execute/20030910-1.c and the requirement
> that (&__real__ x) be well-formed come from?
> 
Blech.  What if we expanded COMPLEX_TYPE into vanilla COMPONENT_REFs? 
So that __real__ x, is emitted as COMPONENT_REF <x, real>.

That would mean that we need to expand all the complex operations either
at gimplification time, or at parse time.  But we would at least be able
to treat complex and structures similarly.


Diego.