1 | require 'rexml/encoding'
|
---|
2 |
|
---|
3 | module REXML
|
---|
4 | # Generates Source-s. USE THIS CLASS.
|
---|
5 | class SourceFactory
|
---|
6 | # Generates a Source object
|
---|
7 | # @param arg Either a String, or an IO
|
---|
8 | # @return a Source, or nil if a bad argument was given
|
---|
9 | def SourceFactory::create_from(arg)
|
---|
10 | if arg.kind_of? String
|
---|
11 | Source.new(arg)
|
---|
12 | elsif arg.respond_to? :read and
|
---|
13 | arg.respond_to? :readline and
|
---|
14 | arg.respond_to? :nil? and
|
---|
15 | arg.respond_to? :eof?
|
---|
16 | IOSource.new(arg)
|
---|
17 | elsif arg.kind_of? Source
|
---|
18 | arg
|
---|
19 | else
|
---|
20 | raise "#{source.class} is not a valid input stream. It must walk \n"+
|
---|
21 | "like either a String, IO, or Source."
|
---|
22 | end
|
---|
23 | end
|
---|
24 | end
|
---|
25 |
|
---|
26 | # A Source can be searched for patterns, and wraps buffers and other
|
---|
27 | # objects and provides consumption of text
|
---|
28 | class Source
|
---|
29 | include Encoding
|
---|
30 | # The current buffer (what we're going to read next)
|
---|
31 | attr_reader :buffer
|
---|
32 | # The line number of the last consumed text
|
---|
33 | attr_reader :line
|
---|
34 | attr_reader :encoding
|
---|
35 |
|
---|
36 | # Constructor
|
---|
37 | # @param arg must be a String, and should be a valid XML document
|
---|
38 | # @param encoding if non-null, sets the encoding of the source to this
|
---|
39 | # value, overriding all encoding detection
|
---|
40 | def initialize(arg, encoding=nil)
|
---|
41 | @orig = @buffer = arg
|
---|
42 | if encoding
|
---|
43 | self.encoding = encoding
|
---|
44 | else
|
---|
45 | self.encoding = check_encoding( @buffer )
|
---|
46 | end
|
---|
47 | @line = 0
|
---|
48 | end
|
---|
49 |
|
---|
50 |
|
---|
51 | # Inherited from Encoding
|
---|
52 | # Overridden to support optimized en/decoding
|
---|
53 | def encoding=(enc)
|
---|
54 | return unless super
|
---|
55 | @line_break = encode( '>' )
|
---|
56 | if enc != UTF_8
|
---|
57 | @buffer = decode(@buffer)
|
---|
58 | @to_utf = true
|
---|
59 | else
|
---|
60 | @to_utf = false
|
---|
61 | end
|
---|
62 | end
|
---|
63 |
|
---|
64 | # Scans the source for a given pattern. Note, that this is not your
|
---|
65 | # usual scan() method. For one thing, the pattern argument has some
|
---|
66 | # requirements; for another, the source can be consumed. You can easily
|
---|
67 | # confuse this method. Originally, the patterns were easier
|
---|
68 | # to construct and this method more robust, because this method
|
---|
69 | # generated search regexes on the fly; however, this was
|
---|
70 | # computationally expensive and slowed down the entire REXML package
|
---|
71 | # considerably, since this is by far the most commonly called method.
|
---|
72 | # @param pattern must be a Regexp, and must be in the form of
|
---|
73 | # /^\s*(#{your pattern, with no groups})(.*)/. The first group
|
---|
74 | # will be returned; the second group is used if the consume flag is
|
---|
75 | # set.
|
---|
76 | # @param consume if true, the pattern returned will be consumed, leaving
|
---|
77 | # everything after it in the Source.
|
---|
78 | # @return the pattern, if found, or nil if the Source is empty or the
|
---|
79 | # pattern is not found.
|
---|
80 | def scan(pattern, cons=false)
|
---|
81 | return nil if @buffer.nil?
|
---|
82 | rv = @buffer.scan(pattern)
|
---|
83 | @buffer = $' if cons and rv.size>0
|
---|
84 | rv
|
---|
85 | end
|
---|
86 |
|
---|
87 | def read
|
---|
88 | end
|
---|
89 |
|
---|
90 | def consume( pattern )
|
---|
91 | @buffer = $' if pattern.match( @buffer )
|
---|
92 | end
|
---|
93 |
|
---|
94 | def match_to( char, pattern )
|
---|
95 | return pattern.match(@buffer)
|
---|
96 | end
|
---|
97 |
|
---|
98 | def match_to_consume( char, pattern )
|
---|
99 | md = pattern.match(@buffer)
|
---|
100 | @buffer = $'
|
---|
101 | return md
|
---|
102 | end
|
---|
103 |
|
---|
104 | def match(pattern, cons=false)
|
---|
105 | md = pattern.match(@buffer)
|
---|
106 | @buffer = $' if cons and md
|
---|
107 | return md
|
---|
108 | end
|
---|
109 |
|
---|
110 | # @return true if the Source is exhausted
|
---|
111 | def empty?
|
---|
112 | @buffer == ""
|
---|
113 | end
|
---|
114 |
|
---|
115 | def position
|
---|
116 | @orig.index( @buffer )
|
---|
117 | end
|
---|
118 |
|
---|
119 | # @return the current line in the source
|
---|
120 | def current_line
|
---|
121 | lines = @orig.split
|
---|
122 | res = lines.grep @buffer[0..30]
|
---|
123 | res = res[-1] if res.kind_of? Array
|
---|
124 | lines.index( res ) if res
|
---|
125 | end
|
---|
126 | end
|
---|
127 |
|
---|
128 | # A Source that wraps an IO. See the Source class for method
|
---|
129 | # documentation
|
---|
130 | class IOSource < Source
|
---|
131 | #attr_reader :block_size
|
---|
132 |
|
---|
133 | # block_size has been deprecated
|
---|
134 | def initialize(arg, block_size=500, encoding=nil)
|
---|
135 | @er_source = @source = arg
|
---|
136 | @to_utf = false
|
---|
137 | # Determining the encoding is a deceptively difficult issue to resolve.
|
---|
138 | # First, we check the first two bytes for UTF-16. Then we
|
---|
139 | # assume that the encoding is at least ASCII enough for the '>', and
|
---|
140 | # we read until we get one of those. This gives us the XML declaration,
|
---|
141 | # if there is one. If there isn't one, the file MUST be UTF-8, as per
|
---|
142 | # the XML spec. If there is one, we can determine the encoding from
|
---|
143 | # it.
|
---|
144 | @buffer = ""
|
---|
145 | str = @source.read( 2 )
|
---|
146 | if encoding
|
---|
147 | self.encoding = encoding
|
---|
148 | elsif /\A(?:\xfe\xff|\xff\xfe)/n =~ str
|
---|
149 | self.encoding = check_encoding( str )
|
---|
150 | else
|
---|
151 | @line_break = '>'
|
---|
152 | end
|
---|
153 | super [email protected]( @line_break )
|
---|
154 | end
|
---|
155 |
|
---|
156 | def scan(pattern, cons=false)
|
---|
157 | rv = super
|
---|
158 | # You'll notice that this next section is very similar to the same
|
---|
159 | # section in match(), but just a liiittle different. This is
|
---|
160 | # because it is a touch faster to do it this way with scan()
|
---|
161 | # than the way match() does it; enough faster to warrent duplicating
|
---|
162 | # some code
|
---|
163 | if rv.size == 0
|
---|
164 | until @buffer =~ pattern or @source.nil?
|
---|
165 | begin
|
---|
166 | # READLINE OPT
|
---|
167 | #str = @source.read(@block_size)
|
---|
168 | str = @source.readline(@line_break)
|
---|
169 | str = decode(str) if @to_utf and str
|
---|
170 | @buffer << str
|
---|
171 | rescue Iconv::IllegalSequence
|
---|
172 | raise
|
---|
173 | rescue
|
---|
174 | @source = nil
|
---|
175 | end
|
---|
176 | end
|
---|
177 | rv = super
|
---|
178 | end
|
---|
179 | rv.taint
|
---|
180 | rv
|
---|
181 | end
|
---|
182 |
|
---|
183 | def read
|
---|
184 | begin
|
---|
185 | str = @source.readline(@line_break)
|
---|
186 | str = decode(str) if @to_utf and str
|
---|
187 | @buffer << str
|
---|
188 | rescue Exception, NameError
|
---|
189 | @source = nil
|
---|
190 | end
|
---|
191 | end
|
---|
192 |
|
---|
193 | def consume( pattern )
|
---|
194 | match( pattern, true )
|
---|
195 | end
|
---|
196 |
|
---|
197 | def match( pattern, cons=false )
|
---|
198 | rv = pattern.match(@buffer)
|
---|
199 | @buffer = $' if cons and rv
|
---|
200 | while !rv and @source
|
---|
201 | begin
|
---|
202 | str = @source.readline(@line_break)
|
---|
203 | str = decode(str) if @to_utf and str
|
---|
204 | @buffer << str
|
---|
205 | rv = pattern.match(@buffer)
|
---|
206 | @buffer = $' if cons and rv
|
---|
207 | rescue
|
---|
208 | @source = nil
|
---|
209 | end
|
---|
210 | end
|
---|
211 | rv.taint
|
---|
212 | rv
|
---|
213 | end
|
---|
214 |
|
---|
215 | def empty?
|
---|
216 | super and ( @source.nil? || @source.eof? )
|
---|
217 | end
|
---|
218 |
|
---|
219 | def position
|
---|
220 | @er_source.stat.pipe? ? 0 : @er_source.pos
|
---|
221 | end
|
---|
222 |
|
---|
223 | # @return the current line in the source
|
---|
224 | def current_line
|
---|
225 | begin
|
---|
226 | pos = @er_source.pos # The byte position in the source
|
---|
227 | lineno = @er_source.lineno # The XML < position in the source
|
---|
228 | @er_source.rewind
|
---|
229 | line = 0 # The \r\n position in the source
|
---|
230 | begin
|
---|
231 | while @er_source.pos < pos
|
---|
232 | @er_source.readline
|
---|
233 | line += 1
|
---|
234 | end
|
---|
235 | rescue
|
---|
236 | end
|
---|
237 | rescue IOError
|
---|
238 | pos = -1
|
---|
239 | line = -1
|
---|
240 | end
|
---|
241 | [pos, lineno, line]
|
---|
242 | end
|
---|
243 | end
|
---|
244 | end
|
---|