[lmi-commits] [lmi] master 0db24270 3/6: Use gcc-predefined endianness macros consistently

[Greg Chicares]

branch: master
commit 0db24270e611b9a10d59a2e337880f3b2e5b6ccc
Author: Gregory W. Chicares <gchicares@sbcglobal.net>
Commit: Gregory W. Chicares <gchicares@sbcglobal.net>

    Use gcc-predefined endianness macros consistently
---
 md5.cpp               |  4 ++--
 round_glibc.c         | 25 ++++++++-----------------
 test_coding_rules.cpp |  3 ---
 3 files changed, 10 insertions(+), 22 deletions(-)

diff --git a/md5.cpp b/md5.cpp
index 7807acbc..6324ca1f 100644
--- a/md5.cpp
+++ b/md5.cpp
@@ -74,9 +74,9 @@
 #include "md5.hpp"
 
 #if defined _LIBC
-# if __BYTE_ORDER__ == __BIG_ENDIAN
+# if __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
 #  define WORDS_BIGENDIAN 1
-# endif // __BYTE_ORDER__ == __BIG_ENDIAN
+# endif // __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
 #endif // defined _LIBC
 
 /* intel x86 is little-endian */
diff --git a/round_glibc.c b/round_glibc.c
index e14bf5c2..37d17729 100644
--- a/round_glibc.c
+++ b/round_glibc.c
@@ -49,15 +49,6 @@
 //   
https://lists.freedesktop.org/archives/release-wranglers/2004-August/000925.html
 typedef uint32_t u_int32_t;
 
-// This file is intended only for use with MinGW-w64, which is always
-// little endian. It seemed best to preserve alternative definitions
-// in case they're ever wanted for some unanticipated purpose, and
-// to provide these aliases mapping the macros used in the unmodified
-// sources to gcc's standard predefinitions:
-#define __FLOAT_WORD_ORDER __FLOAT_WORD_ORDER__
-#define LITTLE_ENDIAN      __ORDER_LITTLE_ENDIAN__
-#define BIG_ENDIAN         __ORDER_BIG_ENDIAN__
-
 // 
https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/ieee754/dbl-64/s_round.c
 /* Round double to integer away from zero. */
 // 
https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/ieee754/flt-32/s_roundf.c
@@ -97,7 +88,7 @@ typedef uint32_t u_int32_t;
 /* A union which permits us to convert between a double and two 32 bit
    ints.  */
 
-#if __FLOAT_WORD_ORDER == BIG_ENDIAN
+#if __FLOAT_WORD_ORDER__ == __ORDER_BIG_ENDIAN__
 typedef union
 {
   double value;
@@ -108,9 +99,9 @@ typedef union
   } parts;
   uint64_t word;
 } ieee_double_shape_type;
-#endif // __FLOAT_WORD_ORDER == BIG_ENDIAN
+#endif // __FLOAT_WORD_ORDER__ == __ORDER_BIG_ENDIAN__
 
-#if __FLOAT_WORD_ORDER == LITTLE_ENDIAN
+#if __FLOAT_WORD_ORDER__ == __ORDER_LITTLE_ENDIAN__
 typedef union
 {
   double value;
@@ -121,7 +112,7 @@ typedef union
   } parts;
   uint64_t word;
 } ieee_double_shape_type;
-#endif // __FLOAT_WORD_ORDER == LITTLE_ENDIAN
+#endif // __FLOAT_WORD_ORDER__ == __ORDER_LITTLE_ENDIAN__
 
 /* Get two 32 bit ints from a double.  */
 
@@ -178,7 +169,7 @@ do {                                                        
    \
 /* A union which permits us to convert between a long double and
    three 32 bit ints.  */
 
-#if __FLOAT_WORD_ORDER == BIG_ENDIAN
+#if __FLOAT_WORD_ORDER__ == __ORDER_BIG_ENDIAN__
 typedef union
 {
   long double value;
@@ -190,9 +181,9 @@ typedef union
     u_int32_t lsw;
   } parts;
 } ieee_long_double_shape_type;
-#endif // __FLOAT_WORD_ORDER == BIG_ENDIAN
+#endif // __FLOAT_WORD_ORDER__ == __ORDER_BIG_ENDIAN__
 
-#if __FLOAT_WORD_ORDER == LITTLE_ENDIAN
+#if __FLOAT_WORD_ORDER__ == __ORDER_LITTLE_ENDIAN__
 typedef union
 {
   long double value;
@@ -204,7 +195,7 @@ typedef union
     unsigned int empty:16;
   } parts;
 } ieee_long_double_shape_type;
-#endif // __FLOAT_WORD_ORDER == LITTLE_ENDIAN
+#endif // __FLOAT_WORD_ORDER__ == __ORDER_LITTLE_ENDIAN__
 
 /* Get three 32 bit ints from a double.  */
 
diff --git a/test_coding_rules.cpp b/test_coding_rules.cpp
index fc17b618..622b69fd 100644
--- a/test_coding_rules.cpp
+++ b/test_coding_rules.cpp
@@ -998,9 +998,6 @@ bool check_reserved_name_exception(std::string const& s)
     // Compiler specific: glibc.
         ,"_GLIBCXX_DEBUG"
         ,"_LIBC"
-        ,"__BIG_ENDIAN"
-        ,"__FLOAT_WORD_ORDER"
-        ,"__LITTLE_ENDIAN"
     // Compiler specific: EDG; hence, como, and also libcomo.
         ,"__asm"
         ,"__COMO__"