From: Albert Chin (postfix-usersmlists.thewrittenword.com)
Date: Tue May 16 2006 - 11:19:09 CDT

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Tue, May 16, 2006 at 11:57:27AM -0400, Victor Duchovni wrote:
> On Tue, May 16, 2006 at 10:39:23AM -0500, Albert Chin wrote:
>
> One more question. This works on your HPUX system with your compiler,
> but is __float80 really a system and machine architecture specific type,
> or is it a compiler dependent type?

It's a native type.

> Does this also work for GCC on the same system?

Dunno. But, GCC recognizes the __float80 type.

> What compiler are you using?

$ cc -V
cc: HP aC++/ANSI C B3910B A.06.07 [Feb 2 2006]

> http://sourceware.org/ml/libc-hacker/2003-03/msg00148.html

From gcc-3.4.3/gcc/config/ia64/ia64.c:
  /* The __float80 type. */
  float80_type = make_node (REAL_TYPE);
  TYPE_PRECISION (float80_type) = 96;
  layout_type (float80_type);
  (*lang_hooks.types.register_builtin_type) (float80_type, "__float80");

> http://www.koders.com/c/fid1EA0B52E10BBC052352C2F3EE7C9B63B0FD92DAA.aspx
>
> The C++ code in the second URL is a useful tool for testing alignment of
> multiple data types with various IA64 compilers.

Cool.

--
albert chin (chinathewrittenword.com)

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]