Why does Perl allow mutual "use" relationships between modules?

Question

Let's say there are two modules that mutually use each other:

package a;
use b;
sub p {}

1;

package b;
use a;
1;

I think that it is systematically wrong to write code like the above, because the two modules will endlessly copy each other's code to themselves, but I can successfully run the following code, which makes me very surprised. Could any of you explain all of this to me?

#! /usr/bin/perl
use a;
a->p();

Answer 1

As far as I remember "use" perl directive performs checking whether module is already loaded. It's done by calling require() function. So, there is no endless copy.

Answer 2

There are (at least) three different ways of loading something: use, require and do.

use is basically a pimped require and perldoc states for require: require demands that a library file be included if it hasn't already been included. So no problem there.

do is a different story. It executes the file and is more or less like eval or C's #include. Mutual inclusion via do should be fatal.

Answer 3

because the two modules will endlessly copy each other's code to themselves

No, they won't, as you demonstrated with the code that surprised you by working. Perl keeps a record in %INC of which modules have been loaded with use or require and will not attempt to reload them if they get used or required again.