| 1 |
package Whatpm::HTML::Dumper; |
| 2 |
use strict; |
| 3 |
our $VERSION=do{my @r=(q$Revision: 1.4 $=~/\d+/g);sprintf "%d."."%02d" x $#r,@r}; |
| 4 |
|
| 5 |
require Exporter; |
| 6 |
push our @ISA, 'Exporter'; |
| 7 |
|
| 8 |
our @EXPORT = qw(dumptree); |
| 9 |
|
| 10 |
sub dumptree ($) { |
| 11 |
my $node = shift; |
| 12 |
my $r = ''; |
| 13 |
|
| 14 |
my $ns_id = { |
| 15 |
q<http://www.w3.org/1999/xhtml> => 'html', |
| 16 |
q<http://www.w3.org/2000/svg> => 'svg', |
| 17 |
q<http://www.w3.org/1998/Math/MathML> => 'math', |
| 18 |
q<http://www.w3.org/1999/xlink> => 'xlink', |
| 19 |
q<http://www.w3.org/XML/1998/namespace> => 'xml', |
| 20 |
q<http://www.w3.org/2000/xmlns/> => 'xmlns', |
| 21 |
}; |
| 22 |
|
| 23 |
my @node = map { [$_, ''] } @{$node->child_nodes}; |
| 24 |
while (@node) { |
| 25 |
my $child = shift @node; |
| 26 |
my $nt = $child->[0]->node_type; |
| 27 |
if ($nt == $child->[0]->ELEMENT_NODE) { |
| 28 |
my $ns = $child->[0]->namespace_uri; |
| 29 |
unless (defined $ns) { |
| 30 |
$ns = '{} '; |
| 31 |
} elsif ($ns eq q<http://www.w3.org/1999/xhtml>) { |
| 32 |
$ns = ''; |
| 33 |
} elsif ($ns_id->{$ns}) { |
| 34 |
$ns = $ns_id->{$ns} . ' '; |
| 35 |
} else { |
| 36 |
$ns = '{' . $ns . '} '; |
| 37 |
} |
| 38 |
$r .= $child->[1] . '<' . $ns . $child->[0]->manakai_local_name . ">\x0A"; |
| 39 |
|
| 40 |
for my $attr (sort {$a->[0] cmp $b->[0]} map { [do { |
| 41 |
my $ns = $_->namespace_uri; |
| 42 |
unless (defined $ns) { |
| 43 |
$ns = ''; |
| 44 |
} elsif ($ns_id->{$ns}) { |
| 45 |
$ns = $ns_id->{$ns} . ' '; |
| 46 |
} else { |
| 47 |
$ns = '{' . $ns . '} '; |
| 48 |
} |
| 49 |
$ns . $_->manakai_local_name; |
| 50 |
}, $_->value] } |
| 51 |
@{$child->[0]->attributes}) { |
| 52 |
$r .= $child->[1] . ' ' . $attr->[0] . '="'; ## ISSUE: case? |
| 53 |
$r .= $attr->[1] . '"' . "\x0A"; |
| 54 |
} |
| 55 |
|
| 56 |
unshift @node, |
| 57 |
map { [$_, $child->[1] . ' '] } @{$child->[0]->child_nodes}; |
| 58 |
} elsif ($nt == $child->[0]->TEXT_NODE) { |
| 59 |
$r .= $child->[1] . '"' . $child->[0]->data . '"' . "\x0A"; |
| 60 |
} elsif ($nt == $child->[0]->COMMENT_NODE) { |
| 61 |
$r .= $child->[1] . '<!-- ' . $child->[0]->data . " -->\x0A"; |
| 62 |
} elsif ($nt == $child->[0]->DOCUMENT_TYPE_NODE) { |
| 63 |
$r .= $child->[1] . '<!DOCTYPE ' . $child->[0]->name; |
| 64 |
my $pubid = $child->[0]->public_id; |
| 65 |
my $sysid = $child->[0]->system_id; |
| 66 |
if (length $pubid or length $sysid) { |
| 67 |
$r .= ' "' . $pubid . '"'; |
| 68 |
$r .= ' "' . $sysid . '"'; |
| 69 |
} |
| 70 |
$r .= ">\x0A"; |
| 71 |
unshift @node, |
| 72 |
map { [$_, $child->[1] . ' '] } |
| 73 |
sort { $a->node_name cmp $b->node_name } |
| 74 |
values %{$child->[0]->element_types}; |
| 75 |
unshift @node, |
| 76 |
map { [$_, $child->[1] . ' '] } |
| 77 |
sort { $a->node_name cmp $b->node_name } |
| 78 |
values %{$child->[0]->entities}; |
| 79 |
unshift @node, |
| 80 |
map { [$_, $child->[1] . ' '] } |
| 81 |
sort { $a->node_name cmp $b->node_name } |
| 82 |
values %{$child->[0]->notations}; |
| 83 |
unshift @node, |
| 84 |
map { [$_, $child->[1] . ' '] } @{$child->[0]->child_nodes}; |
| 85 |
} elsif ($nt == $child->[0]->PROCESSING_INSTRUCTION_NODE) { |
| 86 |
$r .= $child->[1] . '<?' . $child->[0]->target . ' '; |
| 87 |
$r .= $child->[0]->data . "?>\x0A"; |
| 88 |
} elsif ($nt == $child->[0]->ENTITY_NODE) { |
| 89 |
$r .= $child->[1] . '<!ENTITY ' . $child->[0]->node_name . ' "'; |
| 90 |
$r .= $child->[0]->text_content; |
| 91 |
$r .= '" "'; |
| 92 |
$r .= $child->[0]->public_id if defined $child->[0]->public_id; |
| 93 |
$r .= '" "'; |
| 94 |
$r .= $child->[0]->system_id if defined $child->[0]->system_id; |
| 95 |
$r .= '" '; |
| 96 |
$r .= $child->[0]->notation_name if defined $child->[0]->notation_name; |
| 97 |
$r .= ">\x0A"; |
| 98 |
unshift @node, |
| 99 |
map { [$_, $child->[1] . ' '] } @{$child->[0]->child_nodes}; |
| 100 |
} elsif ($nt == $child->[0]->NOTATION_NODE) { |
| 101 |
$r .= $child->[1] . '<!NOTATION ' . $child->[0]->node_name . ' "'; |
| 102 |
$r .= $child->[0]->public_id if defined $child->[0]->public_id; |
| 103 |
$r .= '" "'; |
| 104 |
$r .= $child->[0]->system_id if defined $child->[0]->system_id; |
| 105 |
$r .= qq[">\x0A]; |
| 106 |
} elsif ($nt == $child->[0]->ELEMENT_TYPE_DEFINITION_NODE) { |
| 107 |
$r .= $child->[1] . '<!ELEMENT ' . $child->[0]->node_name . ' '; |
| 108 |
$r .= $child->[0]->content_model_text; |
| 109 |
$r .= ">\x0A"; |
| 110 |
unshift @node, |
| 111 |
map { [$_, $child->[1] . ' '] } |
| 112 |
sort { $a->node_name cmp $b->node_name } |
| 113 |
values %{$child->[0]->attribute_definitions}; |
| 114 |
} elsif ($nt == $child->[0]->ATTRIBUTE_DEFINITION_NODE) { |
| 115 |
$r .= $child->[1] . $child->[0]->node_name . ' '; |
| 116 |
$r .= [ |
| 117 |
0, 'CDATA', 'ID', 'IDREF', 'IDREFS', 'ENTITY', 'ENTITIES', |
| 118 |
'NMTOKEN', 'NMTOKENS', 'NOTATION', 'ENUMERATION', 11, |
| 119 |
]->[$child->[0]->declared_type] || $child->[0]->declared_type; |
| 120 |
$r .= ' (' . join ('|', @{$child->[0]->allowed_tokens}) . ') '; |
| 121 |
$r .= [ |
| 122 |
0, 'FIXED', 'REQUIRED', 'IMPLIED', 'EXPLICIT', |
| 123 |
]->[$child->[0]->default_type] || $child->[0]->default_type; |
| 124 |
$r .= ' "' . $child->[0]->text_content . '"'; |
| 125 |
$r .= "\x0A"; |
| 126 |
} else { |
| 127 |
$r .= $child->[1] . $child->[0]->node_type . "\x0A"; # error |
| 128 |
} |
| 129 |
} |
| 130 |
|
| 131 |
return $r; |
| 132 |
} # dumptree |
| 133 |
|
| 134 |
## NOTE: Based on <http://wiki.whatwg.org/wiki/Parser_tests>. |
| 135 |
## TDOO: Document |
| 136 |
|
| 137 |
1; |
| 138 |
## $Date: 2008/10/18 08:05:29 $ |
Parent Directory
| 
Revision Log