On Thu, Jan 19, 2017 at 7:49 AM, Aaron Ballman <aaron.ball...@gmail.com> wrote: > On Thu, Jan 19, 2017 at 4:37 AM, Malcolm Parsons > <malcolm.pars...@gmail.com> wrote: >> On 19 January 2017 at 03:47, Aaron Ballman <aaron.ball...@gmail.com> wrote: >>> It is not used in an unevaluated context -- that is a bug. >> >> It is an evaluated expression, but is it odr-used? >> >> C++14 [basic.def.odr] p3: >> >> A variable x whose name appears as a potentially-evaluated expression >> ex is odr-used by ex unless applying the lvalue-to-rvalue conversion >> (4.1) to x yields a constant expression (5.20) that does not invoke >> any nontrivial functions and, if x is an object, ex is an element of >> the set of potential results of an expression e, where either the >> lvalue-to-rvalue conversion (4.1) is applied to e, or e is a >> discarded-value expression (Clause 5). ... >> >> 5.20 [expr.const] p3: >> >> An integral constant expression is an expression of integral or >> unscoped enumeration type, implicitly converted to a prvalue, where >> the converted expression is a core constant expression. [ Note: Such >> expressions may be used as array bounds (8.3.4, 5.3.4), as bit-field >> lengths (9.6), as enumerator initializers if the underlying type is >> not fixed (7.2), and as alignments (7.6.2). — end note ] >> >> I read that as kDelta is not odr-used. >> >> GCC and ICC don't require kDelta to be captured either. > > You are correct, it is not an odr use. MSVC is wrong to require the capture.
I wasn't thinking about that kind of odr-unuse when reviewing your patch, so I am starting to think that perhaps it's not worth distinguishing unevaluated contexts or not in the diagnostic. :-( If we could do it, then great (we seem to be able to do it for regular variable use: http://coliru.stacked-crooked.com/a/4bde9b5daf48956a), but if not, then I think we should just go back to the original wording that says it's not required to be captured (in all cases, not distinguishing odr-use) and put in a FIXME with the test cases that could have an improved diagnostic (including the test case talked about here, which we should add). What do you think? ~Aaron > > ~Aaron > >> >> -- >> Malcolm Parsons _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits