[libcxx-commits] [PATCH] D70631: Microsoft's floating-point to_chars powered by Ryu and Ryu Printf

[Jorg Brown via Phabricator via libcxx-commits]

jorgbrown added inline comments.


================
Comment at: libcxx/include/xcharconv_ryu.h:193
+  // The casts here help MSVC to avoid calls to the __allmul library function.
+  const uint32_t __aLo = static_cast<uint32_t>(__a);
+  const uint32_t __aHi = static_cast<uint32_t>(__a >> 32);
----------------
zoecarver wrote:
> Is it well-defined behavior to cast from a larger integer to a smaller one? 
Yes.  See http://eel.is/c++draft/conv#integral which says:

> A prvalue of an integer type can be converted to a prvalue of another integer type. A prvalue of an unscoped enumeration type can be converted to a prvalue of an integer type.
>
> If the destination type is bool, see [conv.bool]. If the source type is bool, the value false is converted to zero and the value true is converted to one.
>
> Otherwise, the result is the unique value of the destination type that is congruent to the source integer modulo N, where N is the width of the destination type.


================
Comment at: libcxx/include/xcharconv_ryu.h:224
+  _LIBCPP_ASSERT(__dist < 64, "");
+#ifdef _WIN64
+  _LIBCPP_ASSERT(__dist > 0, "");
----------------
zoecarver wrote:
> Is there a way this could be generalized to more 64 bit machines? 
This is inside of the third-part of a large 3-way #if, where the first of the 3 forks (line 146) handles _M_X64 (MSVC support for Intel x64), and the second of the 3 forks (line 165) handles the availability of __int128 in the compiler.  I'm not certain of the existence of 64-bit machines that wouldn't be handled by the those other two forks.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D70631/new/

https://reviews.llvm.org/D70631