Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!cs.utexas.edu!uunet!crdgw1!sungod!davidsen
From: davidsen@sungod.crd.ge.com (William Davidsen)
Newsgroups: comp.unix.xenix
Subject: Re: lint vs prototypes
Message-ID: <1406@crdgw1.crd.ge.com>
Date: 30 Jul 89 22:12:02 GMT
References: <4998@ucdavis.ucdavis.edu>
Sender: news@crdgw1.crd.ge.com
Reply-To: davidsen@crdos1.UUCP (bill davidsen)
Organization: General Electric Corp. R&D, Schenectady, NY
Lines: 22


  I use prototypes all the time. I even wrote a little package to
generate the prototype files which I include as .h files in my programs.
lint is a neat tool, but it has three shortcomings:
	1) it doesn't find all the problems
	2) it finds non-problems
	3) it doesn't fix any problems (where protypes cause
	   coersion)

  Therefore lint by itself is not a complete solution. I usually run
with prototypes and pass lint of a program when it's about to go into
another production release. I like the MSC/Xenix compilers with the
warnings turned up, although you may have to filter out some which are
not meaningful in most cases.

  Prototypes can be left in a program, while lint is a tool rather than
a part of a well designed program. I use both, but I find prototypes
catch more errors, since I don't have to run an extra step to see the
errors.
	bill davidsen		(davidsen@crdos1.crd.GE.COM)
  {uunet | philabs}!crdgw1!crdos1!davidsen
"Stupidity, like virtue, is its own reward" -me