Paul Eggert wrote:
> The output (when compiling Bison) contained oodles of "An enumerated
> type is mixed with another type" diagnostics, because the compiler
> didn't like mixing int values like 0 and 1 (which "false" and "true"
> evaluated to) with bool values. There were so many diagnostics, and
> they were formatted weirdly enough, that it was hard to find the real
> diagnostics that were more important.
> ...
> Here is a public message that talks about a similar problem with the
> SGI compilers:
>
> http://lists.gnu.org/archive/html/bug-textutils/2002-02/msg00020.html
Thanks, I'm applying the patch below. It tacifies IRIX cc.
> The obvious solution is to stop using enum for bool values.
I don't know what you will propose, but
- "enum" is fundamentally the right concept for this; "enum" in C was
introduced for exactly this kind of symbolic values,
- Regarding the stdbool and stdint modules, I prefer an approach that
doesn't break unrelated platforms. We have already invested some effort
in these modules. I'm not interested in retesting it on platforms
from Linux libc5 to HP-UX 10 to BeOS.
Bruno
2006-01-25 Bruno Haible <[EMAIL PROTECTED]>
* stdbool_.h (_Bool) [IRIX cc]: Define as 'signed char', to avoid
warnings.
Reported by Paul Eggert.
*** stdbool_.h 24 Jan 2006 22:00:13 -0000 1.8
--- stdbool_.h 25 Jan 2006 13:52:11 -0000
***************
*** 90,97 ****
enum { false = 0, true = 1 };
# endif
# else
! # if defined __SUNPRO_C && (__SUNPRO_C < 0x550 || __STDC__ == 1)
! /* Avoid stupid "warning: _Bool is a keyword in ISO C99". */
# define _Bool signed char
enum { false = 0, true = 1 };
# else
--- 90,100 ----
enum { false = 0, true = 1 };
# endif
# else
! # if (defined __SUNPRO_C && (__SUNPRO_C < 0x550 || __STDC__ == 1)) ||
(defined __sgi && !defined __GNUC__)
! /* With SunPRO C, avoid stupid
! "warning: _Bool is a keyword in ISO C99".
! With IRIX cc, avoid stupid
! "warning(1185): enumerated type mixed with another type". */
# define _Bool signed char
enum { false = 0, true = 1 };
# else
_______________________________________________
bug-gnulib mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-gnulib
