[patch] Correctly encode retained readonly properties

[Nico Weber]

Thanks for the review!


On Tue, May 7, 2013 at 1:12 PM, jahanian <fjahanian at apple.com> wrote:

> Patch is OK. But I am concerned about producing ABI incompatibility with
> older clangs.
>

That's a valid concern. It _does_ restore ABI compatibility with gcc; some
code in Chromium apparently relied on the original behavior.

Also, similar problem exists
> with the ‘copy’ attribute (and probably others I haven’t tried).
>

'copy' and 'retain' can't be used together as far as I know.


> Let me run this by our runtime guys to see its impact.
>

Thanks!


>
> - Fariborz
>
> On May 7, 2013, at 12:25 PM, Nico Weber <thakis at chromium.org> wrote:
>
> Hi,
>
> clang omits the '&' for retained properties if the property is also
> readonly. The attached patch fixes this – it makes clang match gcc's
> behavior and fixes PR15928.
>
> Ok?
>
> Nico
> <clang-retain.patch>_______________________________________________
> cfe-commits mailing list
> cfe-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130507/3d055a23/attachment.html>