Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!zaphod.mps.ohio-state.edu!uwm.edu!ux1.cso.uiuc.edu!ux1.cso.uiuc.edu!m.cs.uiuc.edu!marick
From: marick@m.cs.uiuc.edu
Newsgroups: comp.software-eng
Subject: Re: Project experience with C++
Message-ID: <39400118@m.cs.uiuc.edu>
Date: 13 Aug 90 14:06:00 GMT
References: <26@<294>
Lines: 19
Nf-ID: #R:<294:26:m.cs.uiuc.edu:39400118:000:765
Nf-From: m.cs.uiuc.edu!marick    Aug 13 09:06:00 1990


For what it's worth:

I've been analysing bugs from the bug newsgroups (for a completely
different reason).  Of the 79 I've got, 9 could be caught by strict
type checking.  6 could have been caught by lint, 1 might have been
caught by lint, and 2 would have required subtyping of ints (and that
such subtypes actually be used).

(In the case of the one that might have been caught by lint, I didn't
save enough code to check whether it actually would be caught.  It was
a case where a new field was added to a struct but the old [partial]
initializor wasn't changed.  It wouldn't be caught if the shifted
initial values happened to be type-compatible with new fields they
matched.)

Brian Marick
Motorola @ University of Illinois
marick@cs.uiuc.edu, uiucdcs!marick