On Oct 29, 2012, at 12:59 PM, Chandler Carruth <[email protected]> wrote:
> FWIW, I continue to think we should fix the attribute design first > before spreading its use throughout LLVM and Clang. I'm not going to > review any patches spreading its use until we get the design right, as > I don't have any rational basis on which to evaluate them given a > flawed design. > > It seemed like there was no real disagreement about what the right > long-term design is, and so I'm confused why we aren't just making > that so, and are instead spending reviewer time (and your time!) on a > path that we will eventually reverse. I don't think that anyone is talking about fundamentally changing how the existing per-function attributes are modeled. Bill is interested and working on *generalizing* that to support target specific codegen-level options, but I don't think that invalidates the use case for target-independent ones like optsize or even ssp. What is your specific concern? -Chris _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
