Remove attempt to support floating-point time_t.

It wasn't tested and probably never worked, no platform used it,
and the latest POSIX no longer allows it.
* Makefile (typecheck): Don't check time_t being 'double'.
* Theory: Document the change.
* difftime.c (difftime):
* localtime.c (differ_by_repeat, tzload, timesub, time2sub):
* private.h (time_t_min, time_t_max):
* zdump.c (absolute_min_time, absolute_max_time, tformat):
Don't try to support floating-point time_t.
* localtime.c, private.h, zdump.c: Don't include float.h.
* localtime.c (truncate_time, double_to_time):
* private.h (TYPE_INTEGRAL):
* zdump.c (checkabsolutes):
Remove; no longer needed.  All uses removed.
* zdump.8 (LIMITATIONS): Remove discussion of floating-point time_t.

difftime.c

@@ -5,7 +5,7 @@
 
 /*LINTLIBRARY*/
 
-#include "private.h"	/* for time_t, TYPE_INTEGRAL, and TYPE_SIGNED */
+#include "private.h"	/* for time_t and TYPE_SIGNED */
 
 double ATTRIBUTE_CONST
 difftime(const time_t time1, const time_t time0)
@@ -16,15 +16,8 @@ difftime(const time_t time1, const time_t time0)
 	*/
 	if (sizeof (double) > sizeof (time_t))
 		return (double) time1 - (double) time0;
-	if (!TYPE_INTEGRAL(time_t)) {
-		/*
-		** time_t is floating.
-		*/
-		return time1 - time0;
-	}
 	if (!TYPE_SIGNED(time_t)) {
 		/*
-		** time_t is integral and unsigned.
 		** The difference of two unsigned values can't overflow
 		** if the minuend is greater than or equal to the subtrahend.
 		*/
@@ -33,7 +26,6 @@ difftime(const time_t time1, const time_t time0)
 		else	return -(double) (time0 - time1);
 	}
 	/*
-	** time_t is integral and signed.
 	** Handle cases where both time1 and time0 have the same sign
 	** (meaning that their difference cannot overflow).
 	*/