[PATCH] D32778: [libcxx] [test] Conditionally workaround C1XX/EDG bugs

Casey Carter via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Tue May 2 18:18:41 PDT 2017


CaseyCarter added inline comments.


================
Comment at: test/support/archetypes.hpp:20
 protected:
+#endif // !TEST_WORKAROUND_C1XX_BROKEN_ZA_CTOR_CHECK
   NullBase() = default;
----------------
STL_MSFT wrote:
> I think that this comment doesn't match the ifndef check, but it's a style question.
I know which style *I* prefer, because I wrote it, and I know which style *you* prefer, because you commented. I'll change this when Marshall or Eric let me know what they prefer here.


================
Comment at: test/support/archetypes.ipp:10
+#ifdef TEST_WORKAROUND_EDG_EXPLICIT_CONSTEXPR
+#define DEFINE_CONSTEXPR
+#else // TEST_WORKAROUND_EDG_EXPLICIT_CONSTEXPR
----------------
STL_MSFT wrote:
> Should you define it to be `inline` instead of nothing?
This macro is typically defined to either `constexpr` or nothing - see the definitions in `archetypes.hpp` above.


https://reviews.llvm.org/D32778





More information about the cfe-commits mailing list