Re: Two or three initial observations on the fp prining patch

[Jerry Quinn <jlquinn at optonline dot net>]

Gabriel Dos Reis writes:
 > Jerry Quinn <jlquinn@optonline.net> writes:
 > 
 > | Paolo Carlini writes:
 > |  > The next step is dealing with the various configury bits:
 > |  > 
 > |  > * IEEE, IBM, or VAX
 > |  > For this, one of the most difficult, we should probably start from
 > |  > the macros present in v2 code and generalize it to the architectures
 > |  > currently supported.
 > | 
 > | Out of curiosity, how many supported platforms are not ieee?
 > | 
 > | Also, we need to determine if long double is an 80 bit or 128 bit
 > | representation.  In the code, I added a macro IEEE_QUAD.
 > 
 > See GCC/gcc/real.[hc].  I would oppose to anything that would mean
 > duplication of the work done there.  If we need anything that is
 > available there, we should hard at ways to communicate with the
 > compiler that does not mean code duplication. 

Absolutely.  I started with the code from libio for convenience sake,
and pretended the configury to get the job was in place so I could
develop.  I only did that under the assumption that there was a
right way to do it.

I see that struct real_format captures the variability between
different float formats, and REAL_MODE_FORMAT is a macro that selects
the format struct to use.  However, I don't see where REAL_MODE_FORMAT
is defined, and more importantly, how we export this to the library.

Also, how can we export FLOAT_BIG_WORDS_ENDIAN to the library?

Or are you thinking that we want to pull real.c into the library in
some manner?

Any help is appreciated :-)

Thanks,
Jerry Quinn