I ran into a problem compiling some code today and one symptom of the problem I had showed up because config.h was included multiple times. If there was a guard against multiple inclusion...something like: #ifndef INCLUDED_CONFIG_H #define INCLUDED_CONFIG_H <what's currently in config.h> #endif I wouldn't have had the problem, or at least the problem would have been masked. In my particular case, someone was #defining int64 to __int64 (I'm running with MS native tools). Before that, config.h was fine, but afterwards it choked since it's got a line like this: /* Define on systems that have an int64_t type. */ #define HAVE_INT64_T 0 /* Define to 1 if the system has the type `__int64'. */ #define HAVE___INT64 1 /* Define to the type of a signed integer type of width exactly 64 bits if such a type exists and the standard includes do not define it. */ /* #undef int64_t */ /* Try to define a 64 bit integer type */ #if !HAVE_INT64_T #ifdef HAVE___INT64 typedef __int64 int64_t; #else #error "no 64 bit type available" #endif #endif The actual offending line is this one: typedef __int64 int64_t; which after the #define became typedef __int64 __int64; I can't decide whether I'm glad I know about this lingering problem in my code or whether I just want it to work. I'm leaning towards the latter. Anyway, I could use an education about why there isn't a guard against multiple inclusion built in to autoheader, as well as a way to add one myself. AH_TOP puts the first part where I want, but I've got multiple AH_BOTTOM calls and I can't see a way to force the final #endif to really go last in the file. automake 1.10 (slightly patched so .deps is created properly) autoconf 2.61 Thanks for your help. -DB _______________________________________________ Autoconf mailing list Autoconf@xxxxxxx http://lists.gnu.org/mailman/listinfo/autoconf
