bug-gnulib
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Problem with stdbool.h redefining true/false even in C++


From: Michael Goffioul
Subject: Re: Problem with stdbool.h redefining true/false even in C++
Date: Wed, 29 Aug 2012 16:44:20 +0100

On Wed, Aug 29, 2012 at 3:59 PM, Paul Eggert <address@hidden> wrote:
I'm no C++ expert but this one looks like there's a straightforward
fix: just use C++'s bool if available.  That's what GCC's stdbool.h does,
anyway.  With other compilers this may be an issue of C and C++
disagree with how _Bool is implemented but I suppose we can burn
those bridges when we come to them.

Here's the patch I installed into gnulib.
Please give it a try, and further patches are welcome.

>From 067dea8fdc99b30a826c0c8aff8060fe4f8095bf Mon Sep 17 00:00:00 2001
From: Paul Eggert <address@hidden>
Date: Wed, 29 Aug 2012 07:52:32 -0700
Subject: [PATCH] stdbool: be more compatible with mixed C/C++ compiles

* lib/stdbool.in.h (_Bool, true, false) [__cplusplus]:
Define to bool, true, false, respectively, as GCC's builtin
stdbool.h does.  Problem reported by Michael Goffioul in
<http://lists.gnu.org/archive/html/bug-gnulib/2012-08/msg00143.html>.
---
 ChangeLog        |  8 ++++++++
 lib/stdbool.in.h | 48 +++++++++++++++++++++++++++++-------------------
 2 files changed, 37 insertions(+), 19 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 63f651f..f484d28 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,11 @@
+2012-08-29  Paul Eggert  <address@hidden>
+
+       stdbool: be more compatible with mixed C/C++ compiles
+       * lib/stdbool.in.h (_Bool, true, false) [__cplusplus]:
+       Define to bool, true, false, respectively, as GCC's builtin
+       stdbool.h does.  Problem reported by Michael Goffioul in
+       <http://lists.gnu.org/archive/html/bug-gnulib/2012-08/msg00143.html>.
+
 2012-08-28  Jim Meyering  <address@hidden>

        revert last change: it was not needed
diff --git a/lib/stdbool.in.h b/lib/stdbool.in.h
index e58f211..924c8ff 100644
--- a/lib/stdbool.in.h
+++ b/lib/stdbool.in.h
@@ -66,24 +66,18 @@
 # undef true
 #endif

-/* For the sake of symbolic names in gdb, we define true and false as
-   enum constants, not only as macros.
-   It is tempting to write
-      typedef enum { false = 0, true = 1 } _Bool;
-   so that gdb prints values of type 'bool' symbolically. But if we do
-   this, values of type '_Bool' may promote to 'int' or 'unsigned int'
-   (see ISO C 99 6.7.2.2.(4)); however, '_Bool' must promote to 'int'
-   (see ISO C 99 6.3.1.1.(2)).  So we add a negative value to the
-   enum; this ensures that '_Bool' promotes to 'int'.  */
-#if defined __cplusplus || (defined __BEOS__ && !defined __HAIKU__)
+#ifdef __cplusplus
+# define _Bool bool
+#else
+# if defined __BEOS__ && !defined __HAIKU__
   /* A compiler known to have 'bool'.  */
   /* If the compiler already has both 'bool' and '_Bool', we can assume they
      are the same types.  */
-# if address@hidden@
+#  if address@hidden@
 typedef bool _Bool;
-# endif
-#else
-# if !defined __GNUC__
+#  endif
+# else
+#  if !defined __GNUC__
    /* If @HAVE__BOOL@:
         Some HP-UX cc and AIX IBM C compiler versions have compiler bugs when
         the built-in _Bool type is used.  See
@@ -103,19 +97,35 @@ typedef bool _Bool;
           "Invalid enumerator. (badenum)" with HP-UX cc on Tru64.
         The only benefit of the enum, debuggability, is not important
         with these compilers.  So use 'signed char' and no enum.  */
-#  define _Bool signed char
-# else
+#   define _Bool signed char
+#  else
    /* With this compiler, trust the _Bool type if the compiler has it.  */
-#  if address@hidden@
+#   if address@hidden@
+   /* For the sake of symbolic names in gdb, define true and false as
+      enum constants, not only as macros.
+      It is tempting to write
+         typedef enum { false = 0, true = 1 } _Bool;
+      so that gdb prints values of type 'bool' symbolically.  But then
+      values of type '_Bool' might promote to 'int' or 'unsigned int'
+      (see ISO C 99 6.7.2.2.(4)); however, '_Bool' must promote to 'int'
+      (see ISO C 99 6.3.1.1.(2)).  So add a negative value to the
+      enum; this ensures that '_Bool' promotes to 'int'.  */
 typedef enum { _Bool_must_promote_to_int = -1, false = 0, true = 1 } _Bool;
+#   endif
 #  endif
 # endif
 #endif
 #define bool _Bool

 /* The other macros must be usable in preprocessor directives.  */
-#define false 0
-#define true 1
+#ifdef __cplusplus
+# define false false
+# define true true
+#else
+# define false 0
+# define true 1
+#endif
+
 #define __bool_true_false_are_defined 1

 #endif /* _GL_STDBOOL_H */

Thanks, that should indeed redefining true and false (I already tried that solution with octave and it solved the problem). But why also include the '#define bool _Bool' line in the patch and only do it in C? AFAIK "bool" is a standard C++ type, there's no need to redefine it.

Michael.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]