Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!mailrus!tut.cis.ohio-state.edu!bloom-beacon!athena.mit.edu!peter
From: peter@athena.mit.edu (Peter J Desnoyers)
Newsgroups: comp.lang.c
Subject: more on cdecl (no flames)
Message-ID: <4834@bloom-beacon.MIT.EDU>
Date: 24 Apr 88 06:16:13 GMT
References: <11216@brl-adm.ARPA> <2111@chinet.UUCP> <4403@garfield.UUCP> <20126@think.UUCP> <601@vsi.UUCP>
Sender: daemon@bloom-beacon.MIT.EDU
Reply-To: peter@athena.mit.edu (Peter J Desnoyers)
Distribution: na
Organization: Massachusetts Institute of Technology
Lines: 17

I just saw something interesting in the man pages for some Harris
computer I use (I'm embarrassed I don't know the model number) 
that is relevant to the (civilized part of the) cdecl discussion 
that was going on earlier:

If log, exp, etc. are declared as "fortran float log();" instead of
"double log();" then the compiler will generate the inline 
instructions flog, fexp,... (or whatever the names were)

It seems, like cdecl, to be a reasonable way to solve a non-portable
problem. I think the reason why cdecl angered so many people is
because it is a solution (not a bad one) to a botch brought on by
failure to standardize calling conventions, aggrevated by the design
of the processor.

				Peter Desnoyers
				peter@athena.mit.edu