Xref: utzoo comp.unix.questions:8203 comp.unix.wizards:9882 Path: utzoo!attcan!uunet!lll-winken!lll-tis!ames!mailrus!emv From: emv@mailrus.cc.umich.edu (Edward Vielmetti) Newsgroups: comp.unix.questions,comp.unix.wizards Subject: Re: cgrep (context-grep) Keywords: grep Message-ID: <571@mailrus.cc.umich.edu> Date: 14 Jul 88 02:58:02 GMT References: <2985@cvl.umd.edu> Sender: usenet@mailrus.cc.umich.edu Reply-To: emv@mailrus.cc.umich.edu (Edward Vielmetti) Followup-To: comp.unix.questions Distribution: na Organization: University of Michigan Computing Center, Ann Arbor Lines: 24 In article <2985@cvl.umd.edu> arensb@lemuria.UUCP (Andrew Arensburger) writes: > > I'm thinking of writing a utility called cgrep, which works just >like grep, but instead of writing out just the line on which the pattern >was found, would write out a couple of lines surrounding the line in >question. when I needed to do this recently, this is what I cooked up on the spur of the moment. It works, though it's by no means optimal. cgrep: #!/bin/sh grep -v $1 $2 | diff -c - $2 all depends on whether your diff can handle standard input or now. mine can (Apollo SR 10 beta). You might note that there was a long discussion of 'gre', a new project from at&t research to do grep right, and as I recall the prevailing view from the "tools, pipes, filters" crowd was that context doesn't belong in grep. The "features, features, features" crowd thought otherwise. I'd be interested in a cgrep if you wrote one. --Ed