Re: [PATCH] rvalue reference implementation for C++0x

[Russell Yanofsky <russ at yanofsky dot org>]

On Thu, 2007-03-15 at 13:09 -0700, Mark Mitchell wrote:
> Doug Gregor wrote:
> > Just so we know what we're dealing with... a quick survery shows about
> > 157 uses of REFERENCE_TYPE in the C++ front end, plus another 55 in
> > the common bits. That's compared to the 22 uses of TYPE_REF_IS_RVALUE
> > introduced by Russell's patch.
> 
> I agree.
> 
> There's no way to avoid analyzing all current users of REFERENCE_TYPE to
> decide if they have to change for rvalue references.  But, if most of
> the cases are going to be the same for rvalue and lvalue references,
> then it's better to leave them the using the same code.

If I implement the approach favored by Andrew Pinski and Gabriel Dos
Reis of creating a reference type node in the C++ front end, which gets
converted into a REFERENCE_TYPE node during gimplification, I'll have to
change all of the current uses of REFERENCE_TYPE in the front end.
You're agreeing with Doug that this approach isn't ideal, I guess, but
would it be acceptable? I haven't done any work yet, I just want to
figure out which direction to go in.

-- 
-  Russell Yanofsky (PGP ID: 0x5FAA0216)
-  http://russ.yanofsky.org/
--

Attachment: signature.asc
Description: This is a digitally signed message part