[clang] Make sure C99/C11 features in <float.h> are provided in C++11

[ldionne]

Summary

Previously, those #defines were only provided in C or when GNU extensions were
enabled. We need those #defines in C++11 and above, too.

[ldionne]

I'll ship this. @eli.friedman I think this is your playground -- if you want any changes to happen post-review, please LMK and I will gladly cooperate.

[efriedma]

Formally, I don't think C11 is a normative reference for C++11 or C++14, only C++17 (see [intro.refs] in the standard).

[jfb]

In D58149#1397382, @efriedma wrote:

Formally, I don't think C11 is a normative reference for C++11 or C++14, only C++17 (see [intro.refs] in the standard).

Right, this was changed in wg21.link/p0063r3
That being said, we *can* offer these C11 features as extensions IIUC.

[ldionne]

In D58149#1397390, @jfb wrote:

In D58149#1397382, @efriedma wrote:

Formally, I don't think C11 is a normative reference for C++11 or C++14, only C++17 (see [intro.refs] in the standard).

You're right. I don't mind submitting a patch that enables it for C++17 and above only if that's what you want, however...

Right, this was changed in wg21.link/p0063r3
That being said, we *can* offer these C11 features as extensions IIUC.

... I think it's also fine to have it in C++11.

Your choice, @eli.friedman

[rsmith]

In D58149#1397499, @ldionne wrote:

In D58149#1397390, @jfb wrote:

In D58149#1397382, @efriedma wrote:

Formally, I don't think C11 is a normative reference for C++11 or C++14, only C++17 (see [intro.refs] in the standard).

You're right. I don't mind submitting a patch that enables it for C++17 and above only if that's what you want, however...

Right, this was changed in wg21.link/p0063r3
That being said, we *can* offer these C11 features as extensions IIUC.

... I think it's also fine to have it in C++11.

In C++11, this is technically valid:

#define FLT_TRUE_MIN 1
#include <cfloat>
#if FLT_TRUE_MIN != 1
#error float.h redefined my macro
#endif

and likewise

#include <cfloat>
constexpr float FLT_TRUE_MIN = 0.0001f;

... but both will fail with this patch, so providing these macros is technically non-conforming. I don't know to what extent we should care about such uses, though. We generally leak additional non-conforming names from the system <*.h> headers into the user's namespace if the libc headers put them there. This would, however, probably be the first time that we put them there ourselves.

[jfb]

In D58149#1397532, @rsmith wrote:

In D58149#1397499, @ldionne wrote:

In D58149#1397390, @jfb wrote:

In D58149#1397382, @efriedma wrote:

Formally, I don't think C11 is a normative reference for C++11 or C++14, only C++17 (see [intro.refs] in the standard).

You're right. I don't mind submitting a patch that enables it for C++17 and above only if that's what you want, however...

Right, this was changed in wg21.link/p0063r3
That being said, we *can* offer these C11 features as extensions IIUC.

... I think it's also fine to have it in C++11.

In C++11, this is technically valid:

#define FLT_TRUE_MIN 1
#include <cfloat>
#if FLT_TRUE_MIN != 1
#error float.h redefined my macro
#endif

and likewise

#include <cfloat>
constexpr float FLT_TRUE_MIN = 0.0001f;

... but both will fail with this patch, so providing these macros is technically non-conforming. I don't know to what extent we should care about such uses, though. We generally leak additional non-conforming names from the system <*.h> headers into the user's namespace if the libc headers put them there. This would, however, probably be the first time that we put them there ourselves.

🎻 ← tiny unicode violin for anyone this would break ☺️

[ldionne]

I agree with the comments. I think we should strive to be strictly conforming. I make that argument all the time for libc++, so I'll be consistent :-).

https://reviews.llvm.org/D58289

lib/Headers/float.h

 
 /* Undefine anything that we'll be redefining below. */
 #  undef FLT_EVAL_METHOD
 #  undef FLT_ROUNDS
 #  undef FLT_RADIX
 #  undef FLT_MANT_DIG
 #  undef DBL_MANT_DIG
 #  undef LDBL_MANT_DIG
-#  if __STDC_VERSION__ >= 199901L || !defined(__STRICT_ANSI__)
+#  if __STDC_VERSION__ >= 199901L || !defined(__STRICT_ANSI__) || __cplusplus >= 201103L
 #    undef DECIMAL_DIG
 #  endif
 #  undef FLT_DIG
 #  undef DBL_DIG
 #  undef LDBL_DIG
 #  undef FLT_MIN_EXP
 #  undef DBL_MIN_EXP
 #  undef LDBL_MIN_EXP

 #  undef DBL_MAX
 #  undef LDBL_MAX
 #  undef FLT_EPSILON
 #  undef DBL_EPSILON
 #  undef LDBL_EPSILON
 #  undef FLT_MIN
 #  undef DBL_MIN
 #  undef LDBL_MIN
-#  if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__)
+#  if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__) || __cplusplus >= 201103L
 #    undef FLT_TRUE_MIN
 #    undef DBL_TRUE_MIN
 #    undef LDBL_TRUE_MIN
 #    undef FLT_DECIMAL_DIG
 #    undef DBL_DECIMAL_DIG
 #    undef LDBL_DECIMAL_DIG
 #    undef FLT_HAS_SUBNORM
 #    undef DBL_HAS_SUBNORM
 #    undef LDBL_HAS_SUBNORM
 #  endif
 #endif
 
 /* Characteristics of floating point types, C99 5.2.4.2.2 */
 
 #define FLT_EVAL_METHOD __FLT_EVAL_METHOD__
 #define FLT_ROUNDS (__builtin_flt_rounds())
 #define FLT_RADIX __FLT_RADIX__
 
 #define FLT_MANT_DIG __FLT_MANT_DIG__
 #define DBL_MANT_DIG __DBL_MANT_DIG__
 #define LDBL_MANT_DIG __LDBL_MANT_DIG__
 
-#if __STDC_VERSION__ >= 199901L || !defined(__STRICT_ANSI__)
+#if __STDC_VERSION__ >= 199901L || !defined(__STRICT_ANSI__) || __cplusplus >= 201103L
 #  define DECIMAL_DIG __DECIMAL_DIG__
 #endif
 
 #define FLT_DIG __FLT_DIG__
 #define DBL_DIG __DBL_DIG__
 #define LDBL_DIG __LDBL_DIG__
 
 #define FLT_MIN_EXP __FLT_MIN_EXP__

 #define FLT_EPSILON __FLT_EPSILON__
 #define DBL_EPSILON __DBL_EPSILON__
 #define LDBL_EPSILON __LDBL_EPSILON__
 
 #define FLT_MIN __FLT_MIN__
 #define DBL_MIN __DBL_MIN__
 #define LDBL_MIN __LDBL_MIN__
 
-#if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__)
+#if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__) || __cplusplus >= 201103L
 #  define FLT_TRUE_MIN __FLT_DENORM_MIN__
 #  define DBL_TRUE_MIN __DBL_DENORM_MIN__
 #  define LDBL_TRUE_MIN __LDBL_DENORM_MIN__
 #  define FLT_DECIMAL_DIG __FLT_DECIMAL_DIG__
 #  define DBL_DECIMAL_DIG __DBL_DECIMAL_DIG__
 #  define LDBL_DECIMAL_DIG __LDBL_DECIMAL_DIG__
 #  define FLT_HAS_SUBNORM __FLT_HAS_DENORM__
 #  define DBL_HAS_SUBNORM __DBL_HAS_DENORM__

test/Headers/float.c

 // RUN: %clang_cc1 -fsyntax-only -verify -std=c89 -ffreestanding %s
 // RUN: %clang_cc1 -fsyntax-only -verify -std=c99 -ffreestanding %s
 // RUN: %clang_cc1 -fsyntax-only -verify -std=c11 -ffreestanding %s
+// RUN: %clang_cc1 -fsyntax-only -verify -xc++ -std=c++11 -ffreestanding %s
+// RUN: %clang_cc1 -fsyntax-only -verify -xc++ -std=c++14 -ffreestanding %s
+// RUN: %clang_cc1 -fsyntax-only -verify -xc++ -std=c++17 -ffreestanding %s
 // expected-no-diagnostics
 
 /* Basic floating point conformance checks against:
     - N1570 draft of C11 Std.
     - N1256 draft of C99 Std.
     - http://port70.net/~nsz/c/c89/c89-draft.html draft of C89/C90 Std.
 */
 /*
     C11,    5.2.4.2.2p11,   pp. 30
     C99,    5.2.4.2.2p9,    pp. 25
-    C89,    2.2.4.2 
+    C89,    2.2.4.2
 */
 #include <float.h>
 
 #ifndef FLT_RADIX
     #error "Mandatory macro FLT_RADIX is missing."
 #elif   FLT_RADIX < 2
     #error "Mandatory macro FLT_RADIX is invalid."
 #endif

 #elif   LDBL_MANT_DIG < 2
     #error "Mandatory macro LDBL_MANT_DIG is invalid."
 #endif
 #if ((FLT_MANT_DIG > DBL_MANT_DIG) || (DBL_MANT_DIG > LDBL_MANT_DIG))
     #error "Mandatory macros {FLT,DBL,LDBL}_MANT_DIG are invalid."
 #endif
 
 
-#if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__)
+#if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__) || __cplusplus >= 201103L
     #ifndef FLT_DECIMAL_DIG
         #error "Mandatory macro FLT_DECIMAL_DIG is missing."
     #elif   FLT_DECIMAL_DIG < 6
         #error "Mandatory macro FLT_DECIMAL_DIG is invalid."
     #endif
     #ifndef DBL_DECIMAL_DIG
         #error "Mandatory macro DBL_DECIMAL_DIG is missing."
     #elif   DBL_DECIMAL_DIG < 10
#ifdef DBL_HAS_SUBNORM
         #error "Macro DBL_HAS_SUBNORM should not be defined."
     #endif
     #ifdef LDBL_HAS_SUBNORM
         #error "Macro LDBL_HAS_SUBNORM should not be defined."
     #endif
 #endif
 
 
-#if __STDC_VERSION__ >= 199901L || !defined(__STRICT_ANSI__)
+#if __STDC_VERSION__ >= 199901L || !defined(__STRICT_ANSI__) || __cplusplus >= 201103L
     #ifndef DECIMAL_DIG
         #error "Mandatory macro DECIMAL_DIG is missing."
     #elif   DECIMAL_DIG < 10
         #error "Mandatory macro DECIMAL_DIG is invalid."
     #endif
 #else
     #ifdef DECIMAL_DIG
         #error "Macro DECIMAL_DIG should not be defined."

 
 /* Internal consistency checks */
 _Static_assert(FLT_RADIX == __FLT_RADIX__, "");
 
 _Static_assert(FLT_MANT_DIG == __FLT_MANT_DIG__, "");
 _Static_assert(DBL_MANT_DIG == __DBL_MANT_DIG__, "");
 _Static_assert(LDBL_MANT_DIG == __LDBL_MANT_DIG__, "");
 
-#if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__)
+#if __STDC_VERSION__ >= 201112L || !defined(__STRICT_ANSI__) || __cplusplus >= 201103L
 _Static_assert(FLT_DECIMAL_DIG == __FLT_DECIMAL_DIG__, "");
 _Static_assert(DBL_DECIMAL_DIG == __DBL_DECIMAL_DIG__, "");
 _Static_assert(LDBL_DECIMAL_DIG == __LDBL_DECIMAL_DIG__, "");
 #endif
 
-#if __STDC_VERSION__ >= 199901L || !defined(__STRICT_ANSI__)
+#if __STDC_VERSION__ >= 199901L || !defined(__STRICT_ANSI__) || __cplusplus >= 201103L
 _Static_assert(DECIMAL_DIG == __DECIMAL_DIG__, "");
 #endif
 
 _Static_assert(FLT_DIG == __FLT_DIG__, "");
 _Static_assert(DBL_DIG == __DBL_DIG__, "");
 _Static_assert(LDBL_DIG == __LDBL_DIG__, "");
 
 _Static_assert(FLT_MIN_EXP == __FLT_MIN_EXP__, "");