Re: [PATCH] Perform constant folding of math builtins

[Gabriel Dos Reis <gdr at integrable-solutions dot net>]

"Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> writes:

|  > From: Gabriel Dos Reis <gdr@integrable-solutions.net>
|  > 
|  > "Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> writes:
|  > 
|  > | Assuming I remembered everything correctly, these are mathematical
|  > | identities.  But I'm not sure which if any are valid/invalid in the
|  > | presence of NaNs, etc.
|  > 
|  > Well, I'm not sure we want to restart the discussion about what
|  > -funsafe-transformations should mean, but if we were to consider all
|  > those mentioned transformations (which are valid under appropriate
|  > conditions) I think we ought to clearly state what we want
|  > -funsae-transformation to do.
|  > -- Gaby
| 
| The descriptions of the various fast math flags in invoke.texi
| (including -funsafe-math-optimizations if that's what you meant) are
| all pretty explicit and clear IMHO.

I don't think I agree with you that the documentation is explicit and
clear.  For example, it doesn't appear to me as explicit and clear 
that the description of "-ffast-mast" does consider the divergences in
IEEE-754 and ISO rules, notably between IEEE-754 and LIA-1, or
divergences between IEEE-754 specifications and ISO C specifications.
Ditto for "-funsafe-math-optimizations".


|  If an FP expert would comment on that, I'd be grateful.

I can't find why the above strikes me as excluding me from the set of
"FP expert".  Oh, well.

-- Gaby