Date: Mon, 19 Oct 1998 08:58:51 -0500
From: Eric Rudd <rudd AT cyberoptics DOT com>
Subject: Re: libc math function upgrade work
To: djgpp-workers AT delorie DOT com
Message-id: <362B459B.4A66C6A3@cyberoptics.com>
Organization: CyberOptics
MIME-version: 1.0
X-Mailer: Mozilla 4.05 [en] (Win95; U)
Content-type: text/plain; charset=us-ascii
Content-transfer-encoding: 7bit
References: <Pine DOT SUN DOT 3 DOT 91 DOT 981014202437 DOT 11190A-100000 AT is>
Reply-To: djgpp-workers AT delorie DOT com

I successfully re-built libc.a last weekend with my new math routines.
I noticed that the routine

   POW2.S

is currently in the "ansi" subdirectory under libc, but pow2() is not
specified by ANSI.  I will move it to the "compat" subdirectory, but,
looking ahead, I noticed that a draft of C9x has specified "exp2()"
rather than "pow2()", so both names should perhaps point to the same
routine. I also noticed that there was an indirect mechanism for
invoking the pow2() subroutine, through stub0029.s.  Does anyone know
why there is this stub mechanism, rather than a direct call?  Should I
make another stub so the routine can be called by the C9x name exp2()?

I think a cleaner solution would be to simply include two global labels
in the pow2.s file, but there may be some system issues that I don't
understand that necessitated the stub system.

-Eric Rudd