[cfe-commits] r154844 - in /cfe/trunk: include/clang/AST/ include/clang/Basic/ include/clang/Parse/ include/clang/Sema/ lib/Parse/ lib/Sema/ test/CXX/class/class.mem/ test/CXX/expr/expr.prim/expr.prim.general/ test/CXX/temp/temp.decls/temp.variad

[Richard Smith]

Fixed in r155293.

On Fri, Apr 20, 2012 at 1:25 PM, Richard Smith <richard at metafoo.co.uk>wrote:

> This change causes a regression in C++11 mode:
>
> struct A {
>   virtual ~A();
>
>   struct Inner {
>     ~Inner() throw();
>   };
>   Inner inner;
> };
>
> We don't parse A::Inner's exception specification until after we call
> ActOnFields for A and try to compute A::~A's exception specification.
> Consequently, since A::~A calls Inner::~Inner, A::~A's exception
> specification is computed as EST_Delayed, and never fixed. That causes a
> rejects-valid for code like this:
>
> struct B {
>   virtual ~B() {}
> };
>
> struct C : B {
>   virtual ~C() {}
>   A a;
> };
>
> Here, we assert when checking that C's exception specification is no
> looser than B's, because we can't compute C's exception spec.
>
> On Mon, Apr 16, 2012 at 5:24 PM, Douglas Gregor <dgregor at apple.com> wrote:
>
>>
>> On Apr 16, 2012, at 12:44 PM, Richard Smith <richard at metafoo.co.uk>
>> wrote:
>>
>> On Mon, Apr 16, 2012 at 11:27 AM, Douglas Gregor <dgregor at apple.com>wrote:
>>
>>> Author: dgregor
>>> Date: Mon Apr 16 13:27:27 2012
>>> New Revision: 154844
>>>
>>> URL: http://llvm.org/viewvc/llvm-project?rev=154844&view=rev
>>> Log:
>>> Implement the last part of C++ [class.mem]p2, delaying the parsing of
>>> exception specifications on member functions until after the closing
>>> '}' for the containing class. This allows, for example, a member
>>> function to throw an instance of its own class. Fixes PR12564 and a
>>> fairly embarassing oversight in our C++98/03 support.
>>>
>>
>> This feature was removed from the language in DR1308 (which was voted
>> into the WP at Kona).
>>
>>
>> Richard and I have discussed this, and we think that removing this
>> feature from the language is in error. We'll continue to implement the
>> pre-DR1308 wording and bring the issue up for discussion again at the new
>> committee meeting.
>>
>> - Doug
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20120421/955dc896/attachment.html>