1
|
require "rexml/element"
|
2
|
require "rexml/xmldecl"
|
3
|
require "rexml/source"
|
4
|
require "rexml/comment"
|
5
|
require "rexml/doctype"
|
6
|
require "rexml/instruction"
|
7
|
require "rexml/rexml"
|
8
|
require "rexml/parseexception"
|
9
|
require "rexml/output"
|
10
|
require "rexml/parsers/baseparser"
|
11
|
require "rexml/parsers/streamparser"
|
12
|
require "rexml/parsers/treeparser"
|
13
|
|
14
|
module REXML
|
15
|
# Represents a full XML document, including PIs, a doctype, etc. A
|
16
|
# Document has a single child that can be accessed by root().
|
17
|
# Note that if you want to have an XML declaration written for a document
|
18
|
# you create, you must add one; REXML documents do not write a default
|
19
|
# declaration for you. See |DECLARATION| and |write|.
|
20
|
class Document < Element
|
21
|
# A convenient default XML declaration. If you want an XML declaration,
|
22
|
# the easiest way to add one is mydoc << Document::DECLARATION
|
23
|
# +DEPRECATED+
|
24
|
# Use: mydoc << XMLDecl.default
|
25
|
DECLARATION = XMLDecl.default
|
26
|
|
27
|
# Constructor
|
28
|
# @param source if supplied, must be a Document, String, or IO.
|
29
|
# Documents have their context and Element attributes cloned.
|
30
|
# Strings are expected to be valid XML documents. IOs are expected
|
31
|
# to be sources of valid XML documents.
|
32
|
# @param context if supplied, contains the context of the document;
|
33
|
# this should be a Hash.
|
34
|
def initialize( source = nil, context = {} )
|
35
|
super()
|
36
|
@context = context
|
37
|
return if source.nil?
|
38
|
if source.kind_of? Document
|
39
|
@context = source.context
|
40
|
super source
|
41
|
else
|
42
|
build( source )
|
43
|
end
|
44
|
end
|
45
|
|
46
|
def node_type
|
47
|
:document
|
48
|
end
|
49
|
|
50
|
# Should be obvious
|
51
|
def clone
|
52
|
Document.new self
|
53
|
end
|
54
|
|
55
|
# According to the XML spec, a root node has no expanded name
|
56
|
def expanded_name
|
57
|
''
|
58
|
#d = doc_type
|
59
|
#d ? d.name : "UNDEFINED"
|
60
|
end
|
61
|
|
62
|
alias :name :expanded_name
|
63
|
|
64
|
# We override this, because XMLDecls and DocTypes must go at the start
|
65
|
# of the document
|
66
|
def add( child )
|
67
|
if child.kind_of? XMLDecl
|
68
|
@children.unshift child
|
69
|
elsif child.kind_of? DocType
|
70
|
# Find first Element or DocType node and insert the decl right
|
71
|
# before it. If there is no such node, just insert the child at the
|
72
|
# end. If there is a child and it is an DocType, then replace it.
|
73
|
insert_before_index = 0
|
74
|
@children.find { |x|
|
75
|
insert_before_index += 1
|
76
|
x.kind_of?(Element) || x.kind_of?(DocType)
|
77
|
}
|
78
|
if @children[ insert_before_index ] # Not null = not end of list
|
79
|
if @children[ insert_before_index ].kind_of DocType
|
80
|
@children[ insert_before_index ] = child
|
81
|
else
|
82
|
@children[ index_before_index-1, 0 ] = child
|
83
|
end
|
84
|
else # Insert at end of list
|
85
|
@children[insert_before_index] = child
|
86
|
end
|
87
|
child.parent = self
|
88
|
else
|
89
|
rv = super
|
90
|
raise "attempted adding second root element to document" if @elements.size > 1
|
91
|
rv
|
92
|
end
|
93
|
end
|
94
|
alias :<< :add
|
95
|
|
96
|
def add_element(arg=nil, arg2=nil)
|
97
|
rv = super
|
98
|
raise "attempted adding second root element to document" if @elements.size > 1
|
99
|
rv
|
100
|
end
|
101
|
|
102
|
# @return the root Element of the document, or nil if this document
|
103
|
# has no children.
|
104
|
def root
|
105
|
elements[1]
|
106
|
#self
|
107
|
#@children.find { |item| item.kind_of? Element }
|
108
|
end
|
109
|
|
110
|
# @return the DocType child of the document, if one exists,
|
111
|
# and nil otherwise.
|
112
|
def doctype
|
113
|
@children.find { |item| item.kind_of? DocType }
|
114
|
end
|
115
|
|
116
|
# @return the XMLDecl of this document; if no XMLDecl has been
|
117
|
# set, the default declaration is returned.
|
118
|
def xml_decl
|
119
|
rv = @children[0]
|
120
|
return rv if rv.kind_of? XMLDecl
|
121
|
rv = @children.unshift(XMLDecl.default)[0]
|
122
|
end
|
123
|
|
124
|
# @return the XMLDecl version of this document as a String.
|
125
|
# If no XMLDecl has been set, returns the default version.
|
126
|
def version
|
127
|
xml_decl().version
|
128
|
end
|
129
|
|
130
|
# @return the XMLDecl encoding of this document as a String.
|
131
|
# If no XMLDecl has been set, returns the default encoding.
|
132
|
def encoding
|
133
|
xml_decl().encoding
|
134
|
end
|
135
|
|
136
|
# @return the XMLDecl standalone value of this document as a String.
|
137
|
# If no XMLDecl has been set, returns the default setting.
|
138
|
def stand_alone?
|
139
|
xml_decl().stand_alone?
|
140
|
end
|
141
|
|
142
|
# Write the XML tree out, optionally with indent. This writes out the
|
143
|
# entire XML document, including XML declarations, doctype declarations,
|
144
|
# and processing instructions (if any are given).
|
145
|
#
|
146
|
# A controversial point is whether Document should always write the XML
|
147
|
# declaration (<?xml version='1.0'?>) whether or not one is given by the
|
148
|
# user (or source document). REXML does not write one if one was not
|
149
|
# specified, because it adds unneccessary bandwidth to applications such
|
150
|
# as XML-RPC.
|
151
|
#
|
152
|
# See also the classes in the rexml/formatters package for the proper way
|
153
|
# to change the default formatting of XML output
|
154
|
#
|
155
|
# _Examples_
|
156
|
# Document.new("<a><b/></a>").serialize
|
157
|
#
|
158
|
# output_string = ""
|
159
|
# tr = Transitive.new( output_string )
|
160
|
# Document.new("<a><b/></a>").serialize( tr )
|
161
|
#
|
162
|
# output::
|
163
|
# output an object which supports '<< string'; this is where the
|
164
|
# document will be written.
|
165
|
# indent::
|
166
|
# An integer. If -1, no indenting will be used; otherwise, the
|
167
|
# indentation will be twice this number of spaces, and children will be
|
168
|
# indented an additional amount. For a value of 3, every item will be
|
169
|
# indented 3 more levels, or 6 more spaces (2 * 3). Defaults to -1
|
170
|
# trans::
|
171
|
# If transitive is true and indent is >= 0, then the output will be
|
172
|
# pretty-printed in such a way that the added whitespace does not affect
|
173
|
# the absolute *value* of the document -- that is, it leaves the value
|
174
|
# and number of Text nodes in the document unchanged.
|
175
|
# ie_hack::
|
176
|
# Internet Explorer is the worst piece of crap to have ever been
|
177
|
# written, with the possible exception of Windows itself. Since IE is
|
178
|
# unable to parse proper XML, we have to provide a hack to generate XML
|
179
|
# that IE's limited abilities can handle. This hack inserts a space
|
180
|
# before the /> on empty tags. Defaults to false
|
181
|
def write( output=$stdout, indent=-1, trans=false, ie_hack=false )
|
182
|
if xml_decl.encoding != "UTF-8" && !output.kind_of?(Output)
|
183
|
output = Output.new( output, xml_decl.encoding )
|
184
|
end
|
185
|
formatter = if indent > -1
|
186
|
#if transitive
|
187
|
if trans
|
188
|
REXML::Formatters::Transitive.new( indent, ie_hack )
|
189
|
else
|
190
|
REXML::Formatters::Pretty.new( indent, ie_hack )
|
191
|
end
|
192
|
else
|
193
|
REXML::Formatters::Default.new( ie_hack )
|
194
|
end
|
195
|
formatter.write( self, output )
|
196
|
end
|
197
|
|
198
|
|
199
|
def Document::parse_stream( source, listener )
|
200
|
Parsers::StreamParser.new( source, listener ).parse
|
201
|
end
|
202
|
|
203
|
private
|
204
|
def build( source )
|
205
|
Parsers::TreeParser.new( source, self ).parse
|
206
|
end
|
207
|
end
|
208
|
end
|