This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFC] Fold floating-point comparisons


On Wed, 26 May 2004, Paolo Bonzini wrote:
> The same as the RTL operator LTGT, basically a negated UNEQ as you
> guessed.  This 14th (!) operator is only present in RTL codes, but
> not in tree codes.

Ah.  Another example of missing internals documentation.  My approach
was to document the semantics first, and then once documented propose
a patch to fix PR rtl-optimization/15649.

> > Floating point equality and inequality operators, i.e. EQ_EXPR,
> > NE_EXPR and UNEQ_EXPR don't trap.  So I'd recommend that the same
> > apply for your new LTGT_EXPR.
>
> I guessed the same, but the question is actually what LTGT does...

This is the direct benefit to a programmer of writing internal
documentation.  The person that writes the documentation gets to
define the semantics.  In this case, we get to choose that LTGT
shouldn't trap :>


> And I of course will write them.  I had no time to write regtests of
> any kind, document, and even run preexisting regtests -- I just posted
> a preliminary patch to ensure the basic idea is fine.

Yep.  The basic idea itself looks good.

Roger
--


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]