Well, you could do something like this:

sub test_sub {
    my $error_msg = sub1() || sub2() || sub3();
    return $error_msg;
}
[download]

This works because (quoting from perlop) "Binary "||" performs a short-circuit logical OR operation. That is, if the left operand is true, the right operand is not even evaluated", so $error_msg will contain the first non-zero return value or zero if all subs are successful, and no more subroutines are called after one has returned a non-zero value.

Of course, if your program does other things between each subroutine call then this won't help.

Yes, I forgot to specify that here are many other instructions between them. However, thank you for your comment.
Roberto

You might also find Best Practices for Exception Handling of interest.

# starting point
sub start {
   eval {
       call1();
       call2();
   };
   return 0 if $@;
   return 1;
}

sub call1 {
   # code here
   die "Error in call1" if $something_nasty;
   call3();
   # more code
}

sub call2 {
   # code here
   die "Error in call2" if $something_nasty;
   # more code here..
}

sub call3 {
   # code here
   die "Error in call3" if $something_nasty;
   # more code here..
}
[download]

In other words use die and eval BLOCK to trow and catch exceptions, and you won't have to check every subroutine call, just catch the error at the place you want to process it.

See the die and eval entries in perlfunc for more info.

-- 
Joost       downtime n. The period during which a system
            is error-free and immune from user input.
[download]

$error_msg = sub1();
return $error_msg if ($error_msg ne ());

you can do this:

return $error_param if not sub1();

Godzirra!
Destroying Neo Tokyo since 1954

It's possible to do better than that. Remember Perl's notions of truth can be applied to more than just numeric values. Just return "" if there's no error (which evaluates to false), and return whatever error string is appropriate otherwise, which will evaluate to true.