Class: Arachni::Page::DOM

Inherits:
Object show all
Defined in:
lib/arachni/page/dom.rb,
lib/arachni/page/dom/transition.rb

Overview

Static DOM snapshot as computed by a real browser.

Author:

Defined Under Namespace

Classes: Error, Transition

Constant Summary collapse

IGNORE_FROM_HASH =

Ignore these elements when calculating a #hash.

Set.new([ 'text', 'p' ])

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(options) ⇒ DOM

Returns a new instance of DOM.

Parameters:

Options Hash (options):



57
58
59
60
61
62
63
64
65
66
 
# File 'lib/arachni/page/dom.rb', line 57

def initialize( options )
    @page                 = options[:page]
    self.url              = options[:url]                   || @page.url
    self.digest           = options[:digest]
    @transitions          = options[:transitions]           || []
    @data_flow_sinks      = options[:data_flow_sinks]       || []
    @execution_flow_sinks = options[:execution_flow_sinks]  || []
    @skip_states          = options[:skip_states]           ||
        Support::LookUp::HashSet.new( hasher: :persistent_hash )
end

Instance Attribute Details

#data_flow_sinks ⇒ Array

Returns Browser::Javascript::TaintTracer#data_flow_sinks data.

Returns:



36
37
38
 
# File 'lib/arachni/page/dom.rb', line 36

def data_flow_sinks
  @data_flow_sinks
end

#digest ⇒ String

Returns String digest of the DOM tree.

Returns:

  • (String) —

    String digest of the DOM tree.



44
45
46
 
# File 'lib/arachni/page/dom.rb', line 44

def digest
  @digest
end

#execution_flow_sinks ⇒ Array

Returns Browser::Javascript::TaintTracer#execution_flow_sinks data.

Returns:



40
41
42
 
# File 'lib/arachni/page/dom.rb', line 40

def execution_flow_sinks
  @execution_flow_sinks
end

#page ⇒ Page (readonly)

Returns Page to which this DOM state is attached.

Returns:

  • (Page) —

    Page to which this DOM state is attached.



52
53
54
 
# File 'lib/arachni/page/dom.rb', line 52

def page
  @page
end

#skip_states ⇒ Support::LookUp::HashSet

Returns:



27
28
29
 
# File 'lib/arachni/page/dom.rb', line 27

def skip_states
  @skip_states
end

#transitions ⇒ Array<Transition>

Returns Transitions representing the steps required to convert a Arachni::Page::DOM snapshot to a live Browser page.

Returns:



32
33
34
 
# File 'lib/arachni/page/dom.rb', line 32

def transitions
  @transitions
end

#url ⇒ String

Returns URL of the page as seen by the user-agent, fragments and all.

Returns:

  • (String) —

    URL of the page as seen by the user-agent, fragments and all.



48
49
50
 
# File 'lib/arachni/page/dom.rb', line 48

def url
  @url
end

Class Method Details

.from_rpc_data(data) ⇒ DOM

Parameters:

Returns:



238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
 
# File 'lib/arachni/page/dom.rb', line 238

def self.from_rpc_data( data )
    instance = allocate
    data.each do |name, value|

        value = case name
                    when 'transitions'
                        value.map { |t| Transition.from_rpc_data t }

                    when 'data_flow_sinks'
                        value.map do |entry|
                            Browser::Javascript::TaintTracer::Sink::DataFlow.from_rpc_data( entry )
                        end.to_a

                    when 'execution_flow_sinks'
                        value.map do |entry|
                            Browser::Javascript::TaintTracer::Sink::ExecutionFlow.from_rpc_data( entry )
                        end.to_a

                    when 'skip_states'
                        skip_states = Support::LookUp::HashSet.new(
                            hasher: :persistent_hash
                        )
                        skip_states.collection.merge( value || [] )
                        skip_states

                    else
                        value
                end

        instance.instance_variable_set( "@#{name}", value )
    end
    instance
end

Instance Method Details

#==(other) ⇒ Object 



277
278
279
 
# File 'lib/arachni/page/dom.rb', line 277

def ==( other )
    hash == other.hash
end

#===(other) ⇒ Bool

Note:

Removes the URL strings of both DOMs from each other’s document before comparing.

Returns ‘true` if the compared DOM trees are effectively the same, `false` otherwise.

Parameters:

Returns:

  • (Bool) —

    ‘true` if the compared DOM trees are effectively the same, `false` otherwise.



287
288
289
 
# File 'lib/arachni/page/dom.rb', line 287

def ===( other )
    digest_without_urls( other ) == other.digest_without_urls( self )
end

#depth ⇒ Integer

Returns Depth of the current DOM – sum of #transitions Arachni::Page::DOM::Transition#depths that had to be triggered to reach the current state.

Returns:



93
94
95
 
# File 'lib/arachni/page/dom.rb', line 93

def depth
    @transitions.map { |t| t.depth }.inject(&:+).to_i
end

#hash ⇒ Object 



272
273
274
275
 
# File 'lib/arachni/page/dom.rb', line 272

def hash
    # TODO: Maybe raise error if #digest is not set?
    digest.persistent_hash
end

#playable_transitions ⇒ Object 



97
98
99
 
# File 'lib/arachni/page/dom.rb', line 97

def playable_transitions
    transitions.select { |t| t.playable? }
end

#print_transitions(printer, indent = '') ⇒ Object 



101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
 
# File 'lib/arachni/page/dom.rb', line 101

def print_transitions( printer, indent = '' )
    longest_event_size = 0
    page.dom.transitions.each do |t|
        longest_event_size = [t.event.to_s.size, longest_event_size].max
    end

    page.dom.transitions.map do |t|
        padding = longest_event_size - t.event.to_s.size + 1
        time    = sprintf( '%.4f', t.time.to_f )

        if t.event == :request
            printer.call "#{indent * 2}* [#{time}s] #{t.event}#{' ' * padding} => #{t.element}"
        else
            url = nil
            if t.options[:url]
                url = "(#{t.options[:url]})"
            end

            printer.call "#{indent}-- [#{time}s] #{t.event}#{' ' * padding} => #{t.element} #{url}"

            if t.options[:cookies] && t.options[:cookies].any?
                printer.call "#{indent * 2}-- Cookies:"

                t.options[:cookies].each do |name, value|
                    printer.call  "#{indent * 3}* #{name}\t=> #{value}\n"
                end
            end

            if t.options[:inputs] && t.options[:inputs].any?
                t.options[:inputs].each do |name, value|
                    printer.call  "#{indent * 2}* #{name}\t=> #{value}\n"
                end
            end
        end
    end
end

#push_transition(transition) ⇒ Object

Parameters:



86
87
88
 
# File 'lib/arachni/page/dom.rb', line 86

def push_transition( transition )
    @transitions << transition
end

#restore(browser, take_snapshot = true) ⇒ Browser?

Loads the page and restores it to its captured state.

Parameters:

  • browser (Browser) —

    Browser to use to restore the DOM.

Returns:

  • (Browser, nil) —

    Live page in the ‘browser` if successful, `nil` otherwise.



145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
 
# File 'lib/arachni/page/dom.rb', line 145

def restore( browser, take_snapshot = true )
    # Preload the associated HTTP response since we've already got it.
    browser.preload( page )

    # First, try to load the page via its DOM#url in case it can restore
    # itself via its URL fragments and whatnot.
    browser.goto url, take_snapshot: take_snapshot

    playables = playable_transitions

    # If we've got no playable transitions then we're done.
    return browser if playables.empty?

    browser_page = browser.to_page

    # We were probably led to an out-of-scope page via a JS redirect,
    # bail out.
    return if browser_page.code == 0

    # Check to see if just loading the DOM URL was enough.
    #
    # Of course, this check will fail some of the time because even if the
    # page can restore itself via its URL (using fragment data most probably),
    # the document may still be different from when our snapshot was captured.
    #
    # However, this check doesn't cost us much so it's worth a shot.
    if browser_page.dom === self
        browser.print_debug "Loaded snapshot by URL: #{url}"
        return browser
    end

    browser.print_debug "Could not load snapshot by URL (#{url}), " <<
        'will load by replaying transitions.'

    # The URL restore failed, so, navigate to the pure version of the URL and
    # replay its transitions.
    browser.preload( page )

    playables.each do |transition|
        next if transition.play( browser )

        browser.print_debug "Could not replay transition for: #{url}"
        playables.each do |t|
            browser.print_debug "-#{t == transition ? '>' : '-'} #{transition}"
        end

        return
    end

    browser
end

#to_h ⇒ Hash

Returns:



198
199
200
201
202
203
204
205
206
207
 
# File 'lib/arachni/page/dom.rb', line 198

def to_h
    {
        url:                  url,
        transitions:          transitions.map(&:to_hash),
        digest:               digest,
        skip_states:          skip_states,
        data_flow_sinks:      data_flow_sinks.map(&:to_hash),
        execution_flow_sinks: execution_flow_sinks.map(&:to_hash)
    }
end

#to_hash ⇒ Object 



208
209
210
 
# File 'lib/arachni/page/dom.rb', line 208

def to_hash
    to_h
end

#to_rpc_data ⇒ Hash

Returns Data representing this instance that are suitable the RPC transmission.

Returns:

  • (Hash) —

    Data representing this instance that are suitable the RPC transmission.



224
225
226
227
228
229
230
231
232
233
 
# File 'lib/arachni/page/dom.rb', line 224

def to_rpc_data
    {
        'url'                  => url,
        'transitions'          => transitions.map(&:to_rpc_data),
        'digest'               => digest,
        'skip_states'          => skip_states ? skip_states.collection.to_a : [],
        'data_flow_sinks'      => data_flow_sinks.map(&:to_rpc_data),
        'execution_flow_sinks' => execution_flow_sinks.map(&:to_rpc_data)
    }
end

#to_s ⇒ Object Also known as: inspect 



212
213
214
215
216
217
218
219
 
# File 'lib/arachni/page/dom.rb', line 212

def to_s
    s = "#<#{self.class}:#{object_id} "
    s << "@url=#{@url.inspect} "
    s << "@transitions=#{transitions.size} "
    s << "@data_flow_sinks=#{@data_flow_sinks.size} "
    s << "@execution_flow_sinks=#{@execution_flow_sinks.size}"
    s << '>'
end