Why am I unable to load a Perl library when using the `do` function?

Question

I'm new to Perl, and I'm updating an old Perl website. Every .pl file seems to have this line at the top:

do "func.inc";

So I figured I could use this file to tag on a subroutine for global use.

func.inc

#!/usr/bin/perl
sub foobar
{
    return "Hello world";
}

index.pl

#!/usr/bin/perl
do "func.inc";
print "Content-type: text/html\n\n";
print foobar();

However, I get this error:

Undefined subroutine &main::foobar called at /path/to/index.pl line 4.

Both files are in the same directory, and there's tones of subs in func.inc already which are used throughout the website. However, the script works in the Linux production environment, but does not work for my Windows 7 dev environment (I'm using ActivePerl).

Update:

It looks like the file is not being included; the sub works if the file is included using an absolute path...

do "C:/path/to/func.inc";

... so it looks like relative paths don't work for my local dev environment, but they work in the production environment through. But this is no good for me, because the absolute path on my dev machine will not work for the live server.

How do I get do to work using a relative path on my Windows 7 dev machine?

Update 2:

I was using the Perl -T switch. Unfortunately this removes "." from @INC, and so stops us from using relative paths for do. I removed this switch and the old code is working now. I'm aware that this is not good practice, but unfortunately I'm working with old code, so it seems that I have no choice.

Answer 1

I would first check if the file was actually loaded, the documentation for do mentions that it updates %INC if the file was found. There is also more information in the documentation.

Answer 2

You use the subroutine the same way that you'd use any other subroutine. It doesn't matter that you loaded it with do. However, you shouldn't use do for that. Check out the "Packages" chapter in Intermediate Perl for a detailed explanation of loading subroutines from other files. In short, use require instead.

See the documentation for do. You need to have func.inc (which you can also just call func.pl since pl is "perl library") in one of the directories where Perl will look for libraries. That might be different than the directory that has index.pl. Put func.inc in @INC somewhere, or add its directory to @INC. do also doesn't die if it can't load the file, so it doesn't tell you that it failed. That's why you shouldn't use do to load libraries. :)

Answer 3

make sure you have func.inc in the correct path.

do "func.inc" 

means you are saying func.inc is in the same path as your perl script. check the correct path and then do this

do "/path/func.inc"

Answer 4

Making sure the path is correct, use:

#!/usr/bin/perl 
require("func.inc");
print "Content-type: text/html\n\n"; 
print foobar(); 

Answer 5

The perlfunc documentation for do reads

• do EXPR
  Uses the value of EXPR as a filename and executes the contents of the file as a Perl script.

  do 'stat.pl';
  

  is just like

  eval `cat stat.pl`;
  

  except that it's more efficient and concise, keeps track of the current filename for error messages, searches the @INC directories, and updates %INC if the file is found.

So to see all this in action, say C:\Cygwin\tmp\mylib\func.inc looks like

sub hello {
  print "Hello, world!\n";
}

1;

and we make use of it in the following program:

#!/usr/bin/perl

use warnings;
use strict;

# your code may have unshift @INC, ...
use lib "C:/Cygwin/tmp/mylib";

my $func = "func.inc";

do $func;

# Now we can just call it. Note that with strict subs enabled,
# we have to use parentheses. We could also predeclare with
# use subs qw/ hello /; 
hello();

# do places func.inc's location in %INC
if ($INC{$func}) {
  print "$0: $func found at $INC{$func}\n";
}
else {
  die "$0: $func missing from %INC!";
}

Its output is

Hello, world!
./prog: func.inc found at C:/Cygwin/tmp/mylib/func.inc

As you've observed, do ain't always no crystal stair, which the do documentation explains:

If do cannot read the file, it returns undef and sets $! to the error. If do can read the file but cannot compile it, it returns undef and sets an error message in $@. If the file is successfully compiled, do returns the value of the last expression evaluated.

To check all these cases, we can no longer use simply do "func.inc" but

unless (defined do $func) {
  my $error = $! || $@;
  die "$0: do $func: $error";
}

Explanations for each case are below.

do cannot read the file

If we rename func.inc to nope.inc and rerun the program, we get

./prog: do func.inc: No such file or directory at ./prog line 12.

do can read the file but cannot compile it

Rename nope.inc back to func.inc and delete the closing curly brace in hello to make it look like

sub hello {
  print "Hello, world!\n";

1;

Running the program now, we get

./prog: do func.inc: Missing right curly or square bracket at C:/Cygwin/tmp/mylib/func.inc line 4, at end of line
syntax error at C:/Cygwin/tmp/mylib/func.inc line 4, at EOF

do can read the file and compile it, but it does not return a true value.

Delete the 1; at the end of func.inc to make it

sub hello {
  print "Hello, world!\n";
}

Now the output is

./prog: do func.inc:  at ./prog line 13.

So without a return value, success resembles failure. We could complicate the code that checks the result of do, but the better choice is to always return a true value at the end of Perl libraries and modules.

Note that the program runs correctly even with taint checking (-T) enabled. Try it and see! Be sure to read Taint mode and @INC in perlsec.