Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!think.com!hsdndev!cmcl2!lanl!jlg
From: jlg@lanl.gov (Jim Giles)
Newsgroups: comp.lang.c
Subject: Re: low level optimization
Message-ID: <21848@lanl.gov>
Date: 18 Apr 91 21:42:59 GMT
References: <21660@lanl.gov> <1991Apr17.190243.24691@watmath.waterloo.edu> <1991Apr18.185442.7546@zoo.toronto.edu>
Organization: Los Alamos Natl Lab, Los Alamos, N.M.
Lines: 19

In article <1991Apr18.185442.7546@zoo.toronto.edu>, henry@zoo.toronto.edu (Henry Spencer) writes:
> [...]
> Note that there is no definition of what "translation" means.  It could
> mean something as trivial as tokenizing the source and storing it for
> later processing.  It does not have to mean classical compilation.

Echoing my own previous arguments is not going to get you anywhere.
I previously have stated that if you move the "backend" processes
of compilation into a post-compilation tool (like the loader), that
you could do interprocedural analysis and still remain standard
comforming.  I also said that this is not what you were recomending,
and that I had seen no production quality compiler/loader environments
fitted to do this.

If you want now to recommend this implementation as opposed to the
non-conforming method of dependent compilation you were previously
recommending, then i have no quarrel with you.

J. Giles