On Sep 15, 2010, at 1:54 PM, Sebastian Redl wrote:
>
> On Sep 15, 2010, at 1:41 PM, John McCall wrote:
>
>> On Sep 15, 2010, at 10:06 AM, Sebastian Redl wrote:
>>> On Sep 15, 2010, at 3:14 AM, John McCall wrote:
>>>> Implement generalized copy elision
>>>
>>> Does this mean we can finally finish our rvalue ref implementation?
>>
>> Can you explain how this was blocking that?
>
> I think the main blocker for rvalue refs is that we're not correctly
> preferring moving over copying when the copy constructor could be elided.
That's entirely a Sema issue, while John's change mainly affects IR generation.
Rvalue references should be relatively easy to finish, now that we have the
rest of initialization sorted out, and Sebastian's already handled the
classification bits.
- Doug
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
