You've ruled out CPAN modules, so Expect can't help you. But it would solve your problem for you if you'd let it.

Why not CPAN? There are many articles around here that will tell you how to use CPAN when you think you can't so unless your boss is prohibiting you from using external libraries, then it is probably possible to use CPAN.

Having said that, I agree that in some situations it is better to have a larger but self contained script, than a small one with a forest of dependencies.

Back to your question. Your simplest answer is to read from the STDIN filehandle, and look for a 'Y', though that will not be as reliable or portable as using Expect.

Most Unix systems have a little program called "yes"¹, which simply repeatedly outputs a string (by default "y\n") in order to answer questions like the one you mentioned. With that, the task essentially becomes a piped open:

#!/usr/bin/perl -l
# foo.pl

print "Hello world. Should I continue? (y/n)";
my $answer = <STDIN>;

print "Ok, you said $answer";
[download]

#!/usr/bin/perl

open my $foo, "yes | ./foo.pl |" or die $!;

while (<$foo>) {
    print ": $_";
}
[download]

$ ./876596.pl
: Hello world. Should I continue? (y/n)
: Ok, you said y
:
[download]

___

¹ if you don't have it, you can trivially write it yourself...

Update:

Of course, if you want to actually read the questions (and only respond with 'y' to "Hello world. Should I continue?", instead of every question), you'd have to set up bidirectional communication, e.g. using IPC::Open3.

No, it is not just two lines of code

I know asking people to write code for you is frowned upon, but in this case wouldn't it be faster than typing a response in English sentences? :)

See perlipc

Here's a link that you might find helpful. Chapter 1 covers "Hello World" and program structure which, combined with the other responses to your post, should get you started on your own.

Assuming you are on some UNIX like system. You don't really need any CPAN modules. You can do everything "the hard way". Create 2 pipes, do fork, reopen STDOUT and STDIN of the child into appropriate pipe ends, then execute your program. Now the parent will see both STDOUT and STDIN of your program. Of course, you have to make sure, that the program does not buffer its responses, so that perl script can see them. And make sure the perl script also avoids buffering.

Read 'perldoc perlipc' to find details and easier alternatives.