Hook::WrapSub - wrap subs with pre- and post-call hooks


  use Hook::WrapSub qw( wrap_subs unwrap_subs );

  wrap_subs \&before, 'some_func', 'another_func', \&after;

  unwrap_subs 'some_func';


This module lets you wrap a function, providing one or both of functions that are called just before and just after, whenever the wrapped function is called.

There are a number of other modules that provide the same functionality as this module, some of them better. Have a look at the list in SEE ALSO, below, before you decide which to use.


This function enables intercepting a call to any named function; handlers may be added both before and after the call to the intercepted function.

For example:

  wrap_subs \&before, 'some_func', \&after;

In this case, whenever the sub named 'some_func' is called, the &before sub is called first, and the &after sub is called afterwards. These are both optional. If you only want to intercept the call beforehand:

  wrap_subs \&before, 'some_func';

You may pass more than one sub name:

  wrap_subs \&before, 'foo', 'bar', 'baz', \&after;

and each one will have the same hooks applied.

The sub names may be qualified. Any unqualified names are assumed to reside in the package of the caller.

The &before sub and the &after sub are both passed the argument list which is destined for the wrapped sub. This can be inspected, and even altered, in the &before sub:

  sub before {  
    ref($_[1]) && $_[1] =~ /\bARRAY\b/
      or croak "2nd arg must be an array-ref!";
    @_ or @_ = qw( default values );
    # if no args passed, insert some default values

The &after sub is also passed this list. Modifications to it will (obviously) not be seen by the wrapped sub, but the caller will see the changes, if it happens to be looking.

Here's an example that causes a certain method call to be redirected to a specific object. (Note, we use splice to change $_[0], because assigning directly to $_[0] would cause the change to be visible to the caller, due to the magical aliasing nature of @_.)

  my $handler_object = new MyClass;

    sub { splice @_, 0, 1, $handler_object },

  my $other_object = new MyClass;

  # even though the method is invoked on
  # $other_object, it will actually be executed
  # with a 0'th argument = $handler_obj,
  # as arranged by the pre-call hook sub.

Package Variables

There are some Hook::WrapSub package variables defined, which the &before and &after subs may inspect.


This is the fully qualified name of the wrapped sub.


This is a list which strongly resembles the result of a call to the built-in function caller; it is provided because calling caller will in fact produce confusing results; if your sub is inclined to call caller, have it look at this variable instead.


This contains the result of the call to the wrapped sub. It is empty in the &before sub. In the &after sub, it will be empty if the sub was called in a void context, it will contain one value if the sub was called in a scalar context; otherwise, it may have any number of elements. Note that the &after function is not prevented from modifying the contents of this array; any such modifications will be seen by the caller!

This simple example shows how Hook::WrapSub can be used to log certain subroutine calls:

  sub before {
    print STDERR <<"    EOF";
      About to call $Hook::WrapSub::name( @_ );

  sub after {
    print STDERR <<"    EOF";
      Called $Hook::WrapSub::name( @_ );
      Result=( @Hook::WrapSub::result )
      or @Hook::WrapSub::result = qw( default return );
    # if the sub failed to return something...

Much more elaborate uses are possible. Here's one one way it could be used with database operations:

  my $dbh; # initialized elsewhere.

    sub {


    sub {
      # examine result of sub call:
      if ( $Hook::WrapSub::result[0] ) {
        # success
      else {
        # failure


This removes the most recent wrapping of the named subs.

NOTE: Any given sub may be wrapped an unlimited number of times. A "stack" of the wrappings is maintained internally. wrap_subs "pushes" a wrapping, and unwrap_subs "pops".


Hook::LexWrap provides a similar capability to Hook::WrapSub, but has the benefit that the caller() function works correctly within the wrapped subroutine.

Sub::Prepend lets you provide a sub that will be called before a named sub. The caller() function works correctly in the wrapped sub.

Sub::Mage provides a number of related functions. You can provide pre- and post-call hooks, you can temporarily override a function and then restore it later, and more.

Class::Hook lets you add pre- and post-call hooks around any methods called by your code. It doesn't support functions.

Hook::Scope lets you register callbacks that will be invoked when execution leaves the scope they were registered in.

Hook::PrePostCall provides an OO interface for wrapping a function with pre- and post-call hook functions. Last updated in 1997, and marked as alpha.

Hook::Heckle provides an OO interface for wrapping pre- and post-call hooks around functions or methods in a package. Not updated sinc 2003, and has a 20% failed rate on CPAN Testers.

Moose::Manual::MethodModifiers describes Moose's mechanism for hooking a superclass's method. The before and after subs are called immediately before or after the specified methods are called. The around sub wraps the superclass method, and can even decide not to invoke the superclass method.

Class::Method::Modifiers provides a Moose-style mechanism for a subclass to have before, after, or around method modifiers.

Class::Wrap provides the wrap() function, which takes a coderef and a package name. The coderef is invoked every time a method in the package is called.

Sub::Versive lets you stack pre- and post-call hooks. Last updated in 2001.



This module was written by John Porter <>

It is now being maintained by Neil Bowers.


This is free software. This software may be modified and/or distributed under the same terms as Perl itself.