Xref: utzoo comp.unix.questions:21861 comp.lang.perl:1236
Path: utzoo!attcan!uunet!munnari.oz.au!metro!pta!mcc!chris
From: chris@mcc.UUCP (Chris Robertson)
Newsgroups: comp.unix.questions,comp.lang.perl
Subject: Re: using variable in a regular expression
Message-ID: <106@mcc.UUCP>
Date: 1 May 90 11:44:53 GMT
References: <4122@uudell.dell.com> <1990Apr20.231018.512@iwarp.intel.com> <40438@apple.Apple.COM> <6358@star.cs.vu.nl> <6400@star.cs.vu.nl>
Reply-To: chris@mcc.UUCP (Chris Robertson)
Organization: MCC Software Group, Sydney, Australia
Lines: 17

In article <6400@star.cs.vu.nl> maart@cs.vu.nl (Maarten Litmaath) writes:
>In article <6358@star.cs.vu.nl> I wrote:
[Re using a shell variable in a sed expression]
>)...  What would happen if $1 were to contain a slash?
>)I guess sed wouldn't be so pleased with that.

In order to avoid this, make your sed delimiters something uncommon,
like ^A, or my own favourite, ^V.  Or even some fairly ordinary character
not likely to occur in a variable, like ";" -- e.g.:

	sed -e ";$original;$new;" $file > $newfile

Most sed's I've encountered will accept this.

-- 
"Down in the dumps?  I TOLD you you'd     |    Chris Robertson
 need two sets..."                        |  chris@mcc.pyrsyd.oz