lclint-interest message 108
From matteo@dotto.usr.dsi.unimi.it Thu Sep 26 11:55:59 1996
From: Matteo Vaccari
Subject: Re: Please subscribe
To: lclint-interest@larch.lcs.mit.edu
Date: Thu, 26 Sep 1996 17:46:31 +0200 (MET DST)
X-Mailer: ELM [version 2.4 PL24]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1829
> Please post a brief message introducing yourself and describing your
> professional interests. Feel free to also include any ideas you
> have for lclint development or comments on your experiences using
> lclint.
I am a PhD student at the University of Milano. My main field of
interest is in derivation of programs, usually by means of
paper-and-pencil. I also did some work with automated theorem proving
systems, mostly HOL. I like HOL, but I prefer to concentrate on
finding nice notations and formalisms to make derivations simpler and
crisper.
While I think that learning program derivation techniques greatly
enhances one's programming skills, it is clear that these techniques
are not sufficient. I'm very much in favor of powerful static
analysis tools, rather than powerful debuggers.
My interest in lclint is simply for personal use in my programming
work. Speaking of which, since I do most of my programming on the
Macintosh, I hope to be able to use lclint to check Mac C code. There
are a few C extensions that are commonplace on the Mac; for instance
this is a common idiom in header files:
pascal void Debugger(void) = 0xA9FF;
and also this:
short ApFontID:0x984;
(the latter means that variable ApFontID should be found at the given
address).
My plan is to copy a set of Mac header files to a Linux machine, and
edit them to take away all this clutter. Then I could check my Mac
sources on Linux.
As far as I understand, in order to fully benefit from lclint I should
also add annotations to function prototypes; it is a daunting task
since there are several hundreds of them. If someone else interested
we could share some work (I do a few headers, you do some others).
Regards,
Matteo Vaccari
--
Matteo Vaccari http://eolo.usr.dsi.unimi.it/~matteo
David
Evans
University of Virginia, Computer Science
evans@cs.virginia.edu