PPI::Node - Abstract PPI Node class, an Element that can contain other Elements
PPI::Node isa PPI::Element
# Create a typical node (a Document in this case)
my $Node = PPI::Document->new;
# Add an element to the node( in this case, a token )
my $Token = PPI::Token::Word->new('my');
$Node->add_element( $Token );
# Get the elements for the Node
my @elements = $Node->children;
# Find all the barewords within a Node
my $barewords = $Node->find( 'PPI::Token::Word' );
# Find by more complex criteria
my $my_tokens = $Node->find( sub { $_[1]->content eq 'my' } );
# Remove all the whitespace
$Node->prune( 'PPI::Token::Whitespace' );
# Remove by more complex criteria
$Node->prune( sub { $_[1]->content eq 'my' } );
The PPI::Node class provides an abstract base class for the Element
classes that are able to contain other elements PPI::Document,
PPI::Statement, and PPI::Structure.
As well as those listed below, all of the methods that apply to
PPI::Element objects also apply to PPI::Node objects.
The scope method returns true if the node represents a lexical scope
boundary, or false if it does not.
The add_element method adds a PPI::Element object to the end of a
PPI::Node. Because Elements maintain links to their parent, an
Element can only be added to a single Node.
Returns true if the PPI::Element was added. Returns undef if the
Element was already within another Node, or the method is not passed
a PPI::Element object.
The elements method accesses all child elements structurally within
the PPI::Node object. Note that in the base of the PPI::Structure
classes, this DOES include the brace tokens at either end of the
structure.
Returns a list of zero or more PPI::Element objects.
Alternatively, if called in the scalar context, the elements method
returns a count of the number of elements.
The first_element method accesses the first element structurally within
the PPI::Node object. As for the elements method, this does include
the brace tokens for PPI::Structure objects.
Returns a PPI::Element object, or undef if for some reason the
PPI::Node object does not contain any elements.
The last_element method accesses the last element structurally within
the PPI::Node object. As for the elements method, this does include
the brace tokens for PPI::Structure objects.
Returns a PPI::Element object, or undef if for some reason the
PPI::Node object does not contain any elements.
The children method accesses all child elements lexically within the
PPI::Node object. Note that in the case of the PPI::Structure
classes, this does NOT include the brace tokens at either end of the
structure.
Returns a list of zero of more PPI::Element objects.
Alternatively, if called in the scalar context, the children method
returns a count of the number of lexical children.
The schildren method is really just a convenience, the significant-only
variation of the normal children method.
In list context, returns a list of significant children. In scalar context, returns the number of significant children.
The child method accesses a child PPI::Element object by its
position within the Node.
Returns a PPI::Element object, or undef if there is no child
element at that node.
The lexical structure of the Perl language ignores 'insignificant' items, such as whitespace and comments, while PPI treats these items as valid tokens so that it can reassemble the file at any time. Because of this, in many situations there is a need to find an Element within a Node by index, only counting lexically significant Elements.
The schild method returns a child Element by index, ignoring
insignificant Elements. The index of a child Element is specified in the
same way as for a normal array, with the first Element at index 0, and
negative indexes used to identify a "from the end" position.
The contains method is used to determine if another PPI::Element
object is logically "within" a PPI::Node. For the special case of the
brace tokens at either side of a PPI::Structure object, they are
generally considered "within" a PPI::Structure object, even if they are
not actually in the elements for the PPI::Structure.
Returns true if the PPI::Element is within us, false if not, or undef
on error.
The find method is used to search within a code tree for
PPI::Element objects that meet a particular condition.
To specify the condition, the method can be provided with either a simple
class name (full or shortened), or a CODE/function reference.
# Find all single quotes in a Document (which is a Node)
$Document->find('PPI::Quote::Single');
# The same thing with a shortened class name
$Document->find('Quote::Single');
# Anything more elaborate, we so with the sub
$Document->find( sub {
# At the top level of the file...
$_[1]->parent == $_[0]
and (
# ...find all comments and POD
$_[1]->isa('PPI::Token::Pod')
or
$_[1]->isa('PPI::Token::Comment')
)
} );
The function will be passed two arguments, the top-level PPI::Node
you are searching in and the current PPI::Element that the condition
is testing.
The anonymous function should return one of three values. Returning true
indicates a condition match, defined-false (0 or '') indicates
no-match, and undef indicates no-match and no-descend.
In the last case, the tree walker will skip over anything below the
undef-returning element and move on to the next element at the same
level.
To halt the entire search and return undef immediately, a condition
function should throw an exception (i.e. die).
Note that this same wanted logic is used for all methods documented to
have a \&wanted parameter, as this one does.
The find method returns a reference to an array of PPI::Element
objects that match the condition, false (but defined) if no Elements match
the condition, or undef if you provide a bad condition, or an error
occurs during the search process.
In the case of a bad condition, a warning will be emitted as well.
If the normal find method is like a grep, then find_first is
equivalent to the Scalar::Util first function.
Given an element class or a wanted function, it will search depth-first through a tree until it finds something that matches the condition, returning the first Element that it encounters.
See the find method for details on the format of the search condition.
Returns the first PPI::Element object that matches the condition, false
if nothing matches the condition, or undef if given an invalid condition,
or an error occurs.
The find_any method is a short-circuiting true/false method that behaves
like the normal find method, but returns true as soon as it finds any
Elements that match the search condition.
See the find method for details on the format of the search condition.
Returns true if any Elements that match the condition can be found, false if
not, or undef if given an invalid condition, or an error occurs.
If passed a PPI::Element object that is a direct child of the Node,
the remove_element method will remove the Element intact, along
with any of its children. As such, this method acts essentially as a
'cut' function.
The prune method is used to strip PPI::Element objects out of a code
tree. The argument is the same as for the find method, either a class
name, or an anonymous subroutine which returns true/false. Any Element
that matches the class|wanted will be deleted from the code tree, along
with any of its children.
The prune method returns the number of Element objects that matched
and were removed, non-recursively. This might also be zero, so avoid a
simple true/false test on the return false of the prune method. It
returns undef on error, which you probably should test for.
- Move as much as possible to PPI::XS
See the support section in the main module.
Adam Kennedy <adamk@cpan.org>
Copyright 2001 - 2008 Adam Kennedy.
This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.
The full text of the license can be found in the LICENSE file included with this module.