[PATCH] Support for __has_cpp_attribute (N4200)

[Aaron Ballman]

On Thu, Nov 13, 2014 at 10:29 PM, Richard Smith <richard at metafoo.co.uk> wrote:
> LGTM
>
> There's a good chance we'll bump the version of
> __has_cpp_attribute(deprecated) in the next revision of SD-6, and we'll
> probably want to say "you get the new version only in C++1z mode" when we do
> so. No need to add support for this just yet, but it's something we'll need
> to think about.

This is true -- I'll have to think about ways to differentiate between
the versions we put in Attr.td somehow.

Committed in r221991.

Thanks!

~Aaron
>
> On Thu, Nov 13, 2014 at 2:47 PM, Aaron Ballman <aaron at aaronballman.com>
> wrote:
>>
>> This patch should complete our support for the SD-6 standing document,
>> based off N4200
>> (http://www.open-std.org/JTC1/SC22/WG21/docs/papers/2014/n4200.htm).
>> It implements support for __has_cpp_attribute, based off our existing
>> support for __has_attribute.
>>
>> Note, C++11-style attributes will now be required (by tablegen) to
>> provide a non-default value for the new Version field, associated with
>> the attribute spelling. If we decide we want __has_attribute to report
>> version-specific information for all attributes, we can generalize to
>> cover that later.
>>
>> ~Aaron
>
>