Re: [csmith-project/csmith] 717b1c: Fixed a bug related to --no-math64 and --no-longlo...

[John Regehr <regehr@cs.utah.edu>]

If Csmith needs to know the sizes (and alignments) of integer types at 
program generation time, then it must read these parameters from a file 
like CIL does (see below) instead of just taking this from the host 
platform.

The host platform is totally irrelevant.  Remember that a lot of users 
(like the guy who mailed this morning, and like me) are testing embedded 
compilers.

Of course, the best solution is for Csmith to not care about this stuff. 
 My guess is that this is a very easy goal to achieve.

John




	 sizeof_short          = 2;
	 sizeof_int            = 4;
	 sizeof_bool           = 1;
	 sizeof_long           = 4;
	 sizeof_longlong       = 8;
	 sizeof_ptr            = 4;
	 sizeof_enum           = 4;
	 sizeof_float          = 4;
	 sizeof_double         = 8;
	 sizeof_longdouble     = 12;
	 sizeof_void           = 1;
	 sizeof_fun            = 1;
	 size_t                = "unsigned int";
	 wchar_t               = "int";
	 alignof_short         = 2;
	 alignof_int           = 4;
	 alignof_bool          = 1;
	 alignof_long          = 4;
	 alignof_longlong      = 4;
	 alignof_ptr           = 4;
	 alignof_enum          = 4;
	 alignof_float         = 4;
	 alignof_double        = 4;
	 alignof_longdouble    = 4;
	 alignof_str           = 1;
	 alignof_fun           = 1;
	 alignof_aligned       = 16;
	 char_is_unsigned      = false;
	 const_string_literals = true;
	 underscore_name       = false;
	 __builtin_va_list     = true;
	 __thread_is_keyword   = true;
	 little_endian         = true;