Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!ncar!ico!haddock!karl
From: karl@haddock.ima.isc.com (Karl Heuer)
Newsgroups: comp.lang.c
Subject: Re: strstr sources: a summary of responses
Keywords: ANSI lib, strstr
Message-ID: <17546@haddock.ima.isc.com>
Date: 28 Aug 90 02:43:54 GMT
References: <1158@umvlsi.ecs.umass.edu>
Reply-To: karl@haddock.ima.isc.com (Karl Heuer)
Organization: Interactive Systems, Cambridge, MA 02138-5302
Lines: 15

In article <1158@umvlsi.ecs.umass.edu> srivasta@umvlsi.ecs.umass.edu writes:
>[From the comments introducing John Lacey's Boyer-Moore code:]
>   Some rough timings found this implementation about twice as fast as 
>   the following straight string search.  ...
>   while ( text[i] && pattern[j] ) {
>       i++;
>       j = ( text[i] == pattern[j] ) ? j + 1 : 0;
>     }

I hope it's also more correct.  As written, this won't even handle
strstr("a", "a"); and even if you put the increment in a more likely place,
there's no provision for doing less than a full reset of the search--so
strstr("ababac", "abac") loses.

Karl W. Z. Heuer (karl@kelp.ima.isc.com or ima!kelp!karl), The Walking Lint