perl -MO=Xref YIKES.pl
[download]

Your Perl code is messed-up, by the way. You've got a double-quote midway through that first line, and near the end of the third line. And those percents aren't going to work either. {grin}

-- Randal L. Schwartz, Perl hacker
Be sure to read my standard disclaimer if this is a reply.

Thanks for the suggestion! I especially like the idea that I can also find out where they were defined and used. That's just like getting a cherry on top of the icing on the cake!!

Just so folks won't think I'm a complete idiot: ;-D

I agree whole-heartedly that the Perl pseudo code won't work (I noted that it wouldn't work in the original post). It's similar to what was given to me to own (minus some weird characters I inserted in my sample pseudo code to show how the variables might appear in the rest of the 8000 lines) (see my original post Sub-initiate needs help getting started for the history of that 8000 line Perl file if you're interested).

     perl -MO=Xref YIKES.pl

Works like a charm! Thanks!

The easy (and wrong solution) is to use a regex on that.

use strict;
use warnings;
my %scalars;

# This is the long and wordy version of the line that follows this whi
+le loop.
while ( my $line = <> ) {
    my @scalars_on_this_line = ( $line =~ /(\$\w+)/g );
    foreach my $scalar ( @scalars_on_this_line ) {
        $scalars{ $scalar } = 1;
    }
}

# This is the faster (because of the hash slice )and smaller (undef va
+lues instead of lots of '1' values) method. I'd write it this way.
@scalars{ /(\$\w+)/g } = () while <>;

$\ = $, = "\n";
print sort keys %scalars;
[download]

The harder method would be to use something like B::Xref on your data. Try doing perl -MB::Xref somescript.pl and seeing how that works for you.

<INPUT TYPE='HIDDEN' NAME='g_emplid' VALUE='<TMPL_VAR NAME="g_emplid">
+'>
<INPUT TYPE='HIDDEN' NAME='g_oprid' VALUE='<TMPL_VAR NAME="g_oprid">'>
<INPUT TYPE='HIDDEN' NAME='g_LogonName' VALUE='<TMPL_VAR NAME="g_Logon
+3Name">'>
<INPUT TYPE='HIDDEN' NAME='g_projects' VALUE='<TMPL_VAR NAME="g_projec
+ts2">'>
<INPUT TYPE='HIDDEN' NAME='g_projects_valid' VALUE='<TMPL_VAR NAME="g_
+projects_valid">'>
<INPUT TYPE='HIDDEN' NAME='g_projects_valid' VALUE='<TMPL_VAR NAME="g_
+projects_valid">'>
[download]

use HTML::Template;
my $h = new HTML::Template(filename=>'output.tmpl');
$h->param
(
    g_emplid          => $g_emplid,
    g_oprid           => $g_oprid,
    g_LogonName       => $g_LogonName,
    g_projects        => $g_projects,
    g_projects_valid  => $g_projects_valid,
    g_projects_valid  => $g_projects_valid,
);

print $h->output;
[download]

Excellent suggestion! Also, thank you for giving me an example of how to use it. I'm so new at this, and trying to dig through a program handed off to me. The original program has a lot of print statements in it in order to generate the HTML pages, and I'd like to slowly weed those out in future versions. (See Sub-initiate needs help getting started for history on this program I'm tackling).

Take the following examples, for example:

'This might $look like a scalar, but because of the single quotes, it'
+s not to be interpolated'

/\w+$/ # Looks like the scalar $/, but isn't.

"You owe me \$$money dollars!" # The first $ is escaped.

${$hello[10]}  # Which scalar do you want?

$hello[10] # Does an array element count as a scalar?

$hello{Fred} # Does a hash element count as a scalar?

@array = split /\$/, "Hello$world$here$I$come!";

$text =~ s/(.)(.+)/$2$1/; # $1 and $2 are scalars,
                            do you want them?
[download]

Ok, enough. There are infinately more examples of cruel and usual situations where your regexp searching for scalars is going to have to be grotesque, for it to do what you want. And even then it probably won't always work.

merlyn suggested using the B::Xref module. I can't think of better advice. My point to this post was to try to convey why that is good advice.

And though it's a moot point, since you're going to use B::Xref (right?) I did want to comment on your question of capturing only unique instances of a scalar, because the discussion applies to so many other situations...

Pushing matches onto a stack (into an array) will do nothing for guaranteeing uniqueness. But using a hash will. Hash keys are always unique. Therefore, hashes provide a perfect way of checking to see if a given key already exists, and a foolproof method of ensuring that duplicates can't possibly be made to exist. Any time you're considering uniqueness to be an essential attribute of something you're storing, think of using a hash.

Update: Thanks Not_a_Number for pointing out my misspelling of merlyn. It's been corrected.

Dave

"If I had my life to do over again, I'd be a plumber." -- Albert Einstein

What a difference a 'y' makes.

merlin suggested...

Er, merlin or merlyn ? :-)

dave

use strict;
use warnings;

require('YIKES.pl');

use Devel::Symdump;
use Data::Dumper;

my @array = Devel::Symdump->scalars('main');
print Dumper \@array;
[download]

Maybe this is obvious, but using require on input files can be dangerous because it will run the code in that file.