Perl replacement for sed with command file?

[derekpc]

I am trying to create a perl script that will read arguments from a pipe dilimted file with each line of the file containing a search term and a replace term delimited by the pipe. I want it to work somewhat as sed will with the -f switch and a command file.
What I have works well if I put only literal strings in my searchterms file, but if I try it with perl regular expressions I am completely unsuccessful. I am not sure just how to process the terms so that perl will handle them as regexes. Here is what I am trying to do:

			open TERMS, "$searchterms" or die "Can't open the replacement terms file because: $!\n";
			while (<TERMS>) {
				($srch,$repl) = split(/\|/);
				chomp($srch);
				chomp($repl);
				$file =~ s!$srch!$repl!ig;
			}
			close TERMS;

Here is a sample searchterms file

[URL unfurl="true"]http://www.thisdomain.com|http://www.that.domain.com[/URL]
(<a href ? = ?&quot;?[URL unfurl="true"]http://www\.)thisdomain(\.[/URL][a-d]{3,})|$1thatdomain$2

When run with terms as in the first line os the searchterms I get the subsitution expected. But when run with the regex my substitution is exactly like

$1thatdomain$2

.

I have been hunting the newsgroups and a fairly extensive perl library but am obviously missing the cogent point. Can anyone point me in the right direction?

Thanks
Derek

 

[MikeLacey]

Hi Derek,

Don't know if this will work, but it's something to try:

$file =~ s!"$srch"!"$repl"!ig;
Mike

"Experience is the comb that Nature gives us, after we are bald."

Is that a haiku?
I never could get the hang
of writing those things.

 

[derekpc]

Mike,
I should have mentioned that I did try that, but no luck. Also tried incorporating qr{}, but no luck with that either. Obviously, I am learning as I go.

Thanks for the input.
Derek

 

[derekpc]

I have found a method that seems to work can test thoroughly until I get back to work but I have tried this with a limited set of data and get the substitution I expect:

while (<TERMS>) {
      my $srch = $_;
	chomp($srch);
	eval (&quot;\$file =~ $srch&quot;) 
}
close TERMS;

and put the full subsitution in my searchterms file as follows:

s!(<a href ?= ?&quot;?)([URL unfurl="true"]http://www.)this(.domain/docs)!$1$2that$3!ig[/URL]

As I say, it seems to do what I want, but I am still interested in knowing how to properly handle the code with the split as I attempted first, if anyone would like to comment.
Derek

 

[mountainbiker]

I came across this in perldocs FAQ:

Passing Regexes

To pass regexes around, you'll need to be using a release of Perl sufficiently recent as to support the qr// construct, pass around strings and use an exception-trapping eval, or else be very, very clever.

Here's an example of how to pass in a string to be regex compared using qr//:

sub compare($$) {
my ($val1, $regex) = @_;
my $retval = $val1 =~ /$regex/;
return $retval;
}
$match = compare("old McDonald", qr/d.*D/i);

Notice how qr// allows flags at the end. That pattern was compiled at compile time, although it was executed later. The nifty qr// notation wasn't introduced until the 5.005 release. Before that, you had to approach this problem much less intuitively. For example, here it is again if you don't have qr//:

sub compare($$) {
my ($val1, $regex) = @_;
my $retval = eval { $val1 =~ /$regex/ };
die if $@;
return $retval;
}

$match = compare("old McDonald", q/($?i)d.*D/);

Make sure you never say something like this:

return eval "\$val =~ /$regex/"; # WRONG

or someone can sneak shell escapes into the regex due to the double interpolation of the eval and the double-quoted string. For example:

$pattern_of_evil = 'danger ${ system("rm -rf * &&quot } danger';

eval "\$string =~ /$pattern_of_evil/";

Those preferring to be very, very clever might see the O'Reilly book, Mastering Regular Expressions, by Jeffrey Friedl. Page 273's Build_MatchMany_Function() is particularly interesting. A complete citation of this book is given in perlfaq2.

 

[Coderifous]

There would be more to it to make it 100% fool-proof, but try a combination of mikelacey's example, and the 'e' modifier:

$file =~ s!$srch!"$repl"!ieg;

The e causes the right side of the sub-regex to be evaluated as code. The code in question would simply return the string between the quotes, with all the wonderful interpolation you could ever dream of.

If including $variable's in the $srch string, you may need to do an eval on that as well:

$srch = eval { "$srch" };
$file =~ s/$srch/"$repl"!ieg;

Make sense?

--jim

 

[derekpc]

Mountainbiker and Jim,

Thanks very much for the suggestions. I am aware of the security issue in evaling the full regex, but, as it is not running as a CGI and it is a reasonably closed system (I can track who uses it and punish perpetrators of foul deeds) I decided to take the risk.

I will definately take a look at "Mastering Regular Expressions" pg 273 and try the e modifier, about the only one I didn't think to try before.

Thanks again.

Derek

 

[Coderifous]

Beware that in a regex, if someone wants to be a real beast, could do this (don't run this by the way):

$srch = '(?{qx+rm -r /+}).';
$var =~ s/$srch/blahblahbla/;

Regex's - so powerful.

--jim

 

[mrn]

Have you seen a2p (Awk 2 Perl) & s2p (Sed 2 Perl) do a search on google, to get more info. --
| Mike Nixon
| Unix Admin
|

http://www.instantflashshop.com

----------------------------