Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!ucla-cs!rutgers!cmcl2!lanl!cochiti.lanl.gov!jlg
From: jlg@cochiti.lanl.gov (Jim Giles)
Newsgroups: comp.lang.c
Subject: Re: low level optimization
Message-ID: <21815@lanl.gov>
Date: 18 Apr 91 17:08:34 GMT
References: <21527@lanl.gov> <15870@smoke.brl.mil> <21660@lanl.gov> <15881@smoke.brl.mil>
Sender: news@lanl.gov
Organization: Los Alamos National Laboratory
Lines: 24

In article <15881@smoke.brl.mil>, gwyn@smoke.brl.mil (Doug Gwyn) writes:
|> [...]                                         As I and others
|> have recently told you, sometimes this must be assumed and
|> other times it need not be, depending on context.

You can tell me that until you are blue in the face.  I have never seen
any of these hypothetical contexts in which anonymous pointer arguments
passed in as parameters can be safely assumed not to be aliased.
Please give an example.  Please state why you consider the example 
be representative of a common case in C.  Since I have never seen
such, I suspect that if an example exists it is _very_ obscure.

|> [... aliasing of pointers to mixed type ...]
|> 
|> We know you're a Fortran bigot, but that's ridiculous.

What's rediculous about it?  Even the C standard says it's illegal.
Why can't I complain about it?  It is a _very_ common cause of errors.

Finally, what has this particular complaint have to do with Fortran?
I would have made the same comment if I had been comparing C to _any_
well designed language (Pascal, Ada, Scheme, ...).

J. Giles