[m-dev.] Re: pragma fact_table

[Fergus Henderson]

David Matthew OVERTON, you wrote:
> 
> > !BOXED_FLOAT implies sizeof(Float) <= sizeof(Integer),
> > but BOXED_FLOAT does not imply sizeof(Float) > sizeof(Integer).
> 
> Well the comment in runtime/conf.h should be changed then.  This is
> what it says:
> 
> **      BOXED_FLOAT             double precision floats do not fit in a Word,
> **                              and must be boxed.

runtime/conf.h.in:
	Clarify the documentation for BOXED_FLOAT: it does not imply
	that sizeof(Float) <= sizeof(Integer).

diff -u -r1.12 conf.h.in
--- conf.h.in	1997/02/12 07:45:51	1.12
+++ conf.h.in	1997/02/27 05:43:28
@@ -22,8 +22,12 @@
 **
 **	WORD_TYPE		the base type for the definition of Word.
 **	LOW_TAG_BITS		the number of low-order tag bits we can use.
-**	BOXED_FLOAT		double precision floats do not fit in a Word,
-**				and must be boxed.
+**	BOXED_FLOAT		true if double precision floats might not fit
+**				in a Word, and hence must be boxed.
+**				(Note that when bootstrapping from the source
+**				distribution, we initially build things without
+**				BOXED_FLOAT even on machines for which
+**				sizeof(Float) <= sizeof(Integer).)
 **	HAVE_SYSCONF		the machine has the sysconf() syscall.
 **	HAVE_GETPAGESIZE	the machine has the getpagesize() syscall.
 **	HAVE_MEMALIGN		the machine has the memalign() function.

> How can I check that sizeof(Float) > sizeof(Integer) at compile time,
> since sizeof() does not work with the #if preprocessor command?  If
> this means adding a new parameter to the configure script I've got no
> idea how to do this.

It would mean adding a new parameter to the configure script.

But you don't need to check it.  Just write code that works
regardless of the relative sizes.  In fact your code below does that.

> > > union FloatInteger {
> > > 	Float f;
> > > 	Integer i[sizeof(Float)/sizeof(Integer)];
> > > };
> > > 
> > > Integer
> > > hash_float(Float f)
> > > {
> > > 	union FloatInteger fi;
> > > 	size_t i;
> > > 	Integer h = 0;
> > > 
> > > 	fi.f = f;
> > > 
> > > 	for (i = 0; i < sizeof(Float)/sizeof(Integer); i++)
> > > 		h ^= fi.i[i];
> > 
> > This code won't work if sizeof(Float) is not an even multiple of
> > sizeof(Integer).
>
> I can't see why it wouldn't work.  It will compute a hash value, it
> just won't use all the bits of the float.

Oh, my apologies, you're right.  I thought it would use too many bits.

That code is OK.

However, it might be better to use

	if (sizeof(Float) % sizeof(Integer) == 0) {
		... your code above ...
	} else {
		... code using array of chars ...
	}

-- 
Fergus Henderson <fjh at cs.mu.oz.au>   |  "I have always known that the pursuit
WWW: <http://www.cs.mu.oz.au/~fjh>   |  of excellence is a lethal habit"
PGP: finger fjh at 128.250.37.3         |     -- the last words of T. S. Garp.