
On Wed, Oct 3, 2012 at 7:41 PM, Matt Calabrese <rivorus@gmail.com> wrote:
Also, I'm going to warn you now because I went down this same exact path with Generic -- you're going to have to change your approach for representing givens -- datamembers and even function parameters will not work. The problem is, you need to account for checking of "void", but with this implementation, you'll get a hard error instead of a compile-time bool value. You need to get everything into a context where SFINAE can take place. My approach with the old-style generic was to do all of the checking in a template specialization pattern list (as opposed to in a function template definition), and I used template reference parameters for the givens.
I'm not sure how much help it'll be, but you can look at the .hei files in Boost.Generic's std_concept headers to see how I do the checks exactly. All the tests for all of my concepts explicitly check results when using void for each of the concept arguments, along with other "problem" types (such as function types, since a function type that appears in a function parameter list is treated as a function pointer type). The approach I use should work for your N3351 concepts implementation as well, but it might not be worth trying to figure out the preprocessor output. Anyway, my recommendation is just make all of your givens template arguments (not function arguments) and do all of your SFINAE stuff as a part of a specialization of that template. -- -Matt Calabrese