Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!ut-sally!utah-cs!utah-gr!thomas
From: thomas@utah-gr.UUCP (Spencer W. Thomas)
Newsgroups: net.lang.c
Subject: Lint (was Re: enum function bug?)
Message-ID: <1815@utah-gr.UUCP>
Date: Wed, 1-Oct-86 13:53:34 EDT
Article-I.D.: utah-gr.1815
Posted: Wed Oct  1 13:53:34 1986
Date-Received: Fri, 3-Oct-86 00:51:39 EDT
References: <299@sdchema.sdchem.UUCP> <86900054@haddock> <572@opus.nbires.UUCP> <226@BMS-AT.UUCP> <3635@umcp-cs.UUCP>
Reply-To: thomas@utah-gr.UUCP (Spencer W. Thomas)
Organization: University of Utah CS Dept
Lines: 45


I once looked at the innards of our (4.2) lint, and concluded that it
does type matching on structure data types by comparing
	1. Length of the data item
	2. A hash of the data item's type name.

Thus, we have:

================ x.c ================
struct s
{
    char i[4];
};

struct s
func()
{
    struct a;

    a.i[0] = 'a';
    return a;
}

================ y.c ================
struct s
{
    long a;
};

main()
{
    struct s b, func();

    b = func();
    return 0;
}


Running 'lint x.c y.c' gives (!)

x.c:
y.c:

-- 
=Spencer   ({ihnp4,decvax}!utah-cs!thomas, thomas@utah-cs.ARPA)