package XML::Generator::DOM; =head1 NAME XML::Generator::DOM XML::Generator subclass for producing DOM trees instead of strings. =head1 SYNOPSIS use XML::Generator::DOM; my $dg = XML::Generator::DOM->new(); my $doc = $dg->xml($dg->xmlcmnt("Test document."), $dg->foo({'baz' => 'bam'}, 42)); print $doc->toString; yields: 42 =head1 DESCRIPTION XML::Generator::DOM subclasses XML::Generator in order to produce DOM trees instead of strings (see L and L). This module is still experimental and its semantics might change. Essentially, tag methods return XML::DOM::DocumentFragment objects, constructed either from a DOM document passed into the constructor or a default document that XML::Generator::DOM will automatically construct. Calling the xml() method will return this automatically constructed document and cause a fresh one to be constructed for future tag method calls. If you passed in your own document, you may not call the xml() method. Below, we just note the remaining differences in semantics between XML::Generator methods and XML::Generator::DOM methods. =cut use strict; use Carp; use base 'XML::Generator'; use XML::DOM; use vars qw( $AUTOLOAD $VERSION ); $VERSION = '0.1'; =head1 CONSTRUCTOR These configuration options are accepted but have no effect on the semantics of the returned object: escape, pretty, conformance and empty. =head1 TAG METHODS Subsequently, tag method semantics are somewhat different for this module compared to XML::Generator. The primary difference is that tag method return XML::DOM::DocumentFragment objects. Namespace and attribute processing remains the same, but remaining arguments to tag methods must either be text or other XML::DOM::DocumentFragment objects. No escape processing, syntax checking, or output control is done; this is all left up to XML::DOM. =cut sub new { my $class = shift; my $dom; for (my $i = 0; $i < $#_; $i+=2) { if ($_[$i] eq 'dom_document') { $dom = $_[$i+1]; unless (UNIVERSAL::isa($dom, 'XML::DOM::Document')) { croak "argument to 'dom' option not an XML::DOM::Document object"; } splice @_, $i, 2; last; } } if (ref $class) { $AUTOLOAD = 'new'; return $class->AUTOLOAD(@_); } my $this = $class->SUPER::new(@_); $this->{'dom'} = $dom || XML::Generator::DOM::util::new_dom_root(); return $this; } =head1 SPECIAL TAGS All special tags are available by default with XML::Generator::DOM; you don't need to use 'conformance' => 'strict'. =head2 xmlpi(@args) Arguments will simply be concatenated and passed as the data to the XML::DOM::ProcessingInstruction object that is returned. =cut sub xmlpi { my $this = shift; my $root = $this->{dom}; my $tgt = shift; return $root->createProcessingInstruction($tgt, join '', @_); } =head2 xmlcmnt Escaping of '--' is done by XML::DOM::Comment, which replaces both hyphens with '-'. An XML::DOM::Comment object is returned. =cut sub xmlcmnt { my $this = shift; my $root = $this->{dom}; my $xml = join '', @_; return $root->createComment($xml); } my $config = 'XML::Generator::util::config'; =head2 xmldecl Returns an XML::DOM::XMLDecl object. Respects 'version', 'encoding' and 'dtd' settings in the object. =cut sub xmldecl { my $this = shift; my $root = $this->{dom}; my $version = $this->$config('version') || '1.0'; my $encoding = $this->$config('encoding') || undef; my $standalone = $this->xmldtd($this->$config('dtd')) ? "no" : "yes"; return $root->createXMLDecl($version, $encoding, $standalone) } =head2 xmldecl Returns an XML::DOM::DocumentType object. =cut sub xmldtd { my($this, $dtd) = @_; my $root = $this->{dom}; $dtd ||= $this->$config('dtd'); return unless $dtd && ref($dtd) eq "ARRAY"; return $root->createDocumentType(@{ $dtd }); } =head2 xmlcdata Returns an XML::DOM::CDATASection object. =cut sub xmlcdata { my $this = shift; my $data = join '', @_; my $root = $this->{dom}; return $root->createCDATASection($data); } =head2 xml As described above, xml() can only be used when dom_document was not set in the object. The automatically created document will have its XML Declaration set and the arguments to xml() will be appended to it. Then a new DOM document is automatically generated and the old one is returned. This is the only way to get a DOM document from this module. =cut sub xml { my $this = shift; my $root = $this->{dom}; if ($root != $XML::Generator::DOM::util::root) { croak "xml() method not allowed when dom_document option specified"; } $this->{dom} = XML::Generator::DOM::util::new_dom_root(); $root->setXMLDecl($this->xmldecl()); $root->appendChild($_) for @_; return $root; } sub AUTOLOAD { my $this = shift; (my $tag = $AUTOLOAD) =~ s/.*:://;; my $root = $this->{'dom'}; my($namespace, $attr, @args) = $this->XML::Generator::util::parse_args(@_); my $xml = $root->createDocumentFragment(); my $node = $xml->appendChild($root->createElement("$namespace$tag")); if ($attr) { while (my($k, $v) = each %$attr) { unless ($k =~ /^[^:]+:/) { $k = "$namespace$k"; } $node->setAttribute($k, $v); } } for (@args) { if (UNIVERSAL::isa($_, 'XML::DOM::Node')) { $node->appendChild($_); } else { $node->appendChild($root->createTextNode($_)); } } return $xml; } package XML::Generator::DOM::util; use XML::DOM; use vars qw($root $parser); $parser = XML::DOM::Parser->new; sub new_dom_root { $root = $parser->parse('<_/>'); $root->removeChild($root->getFirstChild); return $root; } 1;