floating point emulation cleanup, 1/3

[Richard Henderson]

On Sat, Mar 02, 2002 at 07:10:27PM -0800, Zack Weinberg wrote:
> - 1750a: I think this is the only target with a floating-point format
> not supported by real.c.  Compile fails, unsurprisingly:

Don't worry about this for now.  real.c should be reorganized
such that it can support multiple fp formats simultaneously
without a tangled mass of ifdefs.  At that point proper support
for 1750a can be added.

As it is, we get incorrect data for this processor, so it's not
so bad.

> The doc patch is not entirely complete; I need to think about where
> the REAL_* macros should be documented (they still exist and are used,
> they're just not for the target to define anymore) and I have other
> changes in trouble.texi right now so I didn't touch that one.

You should be able to poison REAL_ARITHMETIC though, right?

Ok when bootstrap and check passes.



r~