Patch -- Floating point conversions

[Dave Brolley]

Hi,

The following patch adds floating point conversions between
single and double precision to the cgen-fpu code.

Ok to commit?

Dave
2000-03-23  Dave Brolley  <brolley@redhat.com>

	* cgen-fpu.h: Rename extsfdf to fextsfdf. Rename truncdfsf to
	ftruncdfsf.
	* cgen-accfp.c (fextsfdf): New function.
	(ftruncdfsf): New function.
	(cgen_init_accurate_fpu): Initialize fextsfdf and ftruncdfsf.