Class: Arachni::BrowserCluster::Worker

Inherits:
Arachni::Browser show all
Defined in:
lib/arachni/browser_cluster/worker.rb

Overview

Overrides some Arachni::Browser methods to make multiple browsers play well with each other when they’re part of a Arachni::BrowserCluster.

Author:

Constant Summary collapse

RESPAWN_WHEN_WINDOW_COUNT_REACHES =

Returns We can’t just close all windows because PhantomJS for some reason freezes after we do it a lot of times and we can’t just leave open windows accumulate, so we’ve got to take more drastic measures and kill it when the amount of open windows reaches or exceeds this setting.

Returns:

  • (Integer) —

    We can’t just close all windows because PhantomJS for some reason freezes after we do it a lot of times and we can’t just leave open windows accumulate, so we’ve got to take more drastic measures and kill it when the amount of open windows reaches or exceeds this setting.

5

Constants inherited from Arachni::Browser

Arachni::Browser::ELEMENT_APPEARANCE_TIMEOUT, Arachni::Browser::HTML_IDENTIFIERS, Arachni::Browser::PHANTOMJS_SPAWN_TIMEOUT, Arachni::Browser::WATIR_COM_TIMEOUT

Instance Attribute Summary collapse

Attributes inherited from Arachni::Browser

#javascript, #page_snapshots_with_sinks, #pid, #preloads, #transitions, #watir

Class Method Summary collapse

Instance Method Summary collapse

Methods inherited from Arachni::Browser

#cache, #capture?, #capture_snapshot, #captured_pages, #clear_buffers, #cookies, #each_element_with_events, executable, #explore_and_flush, #fire_event, #flush_page_snapshots_with_sinks, #flush_pages, #goto, has_executable?, #load, #load_delay, #on_fire_event, #on_new_page, #on_new_page_with_sink, #on_response, #page_snapshots, #preload, #response, #selenium, #skip_path?, #snapshot_id, #source, #source_with_line_numbers, #start_capture, #stop_capture, #to_page, #trigger_event, #url, #wait_for_timers

Methods included from Support::Mixins::Observable

included

Methods included from Utilities

#available_port, #caller_name, #caller_path, #cookie_decode, #cookie_encode, #cookies_from_document, #cookies_from_file, #cookies_from_response, #exception_jail, #exclude_path?, #follow_protocol?, #form_decode, #form_encode, #forms_from_document, #forms_from_response, #full_and_absolute_url?, #generate_token, #get_path, #hms_to_seconds, #html_decode, #html_encode, #include_path?, #links_from_document, #links_from_response, #normalize_url, #page_from_response, #page_from_url, #parse_set_cookie, #path_in_domain?, #path_too_deep?, #port_available?, #rand_port, #random_seed, #redundant_path?, #regexp_array_match, #remove_constants, #request_parse_body, #seconds_to_hms, #skip_page?, #skip_path?, #skip_resource?, #skip_response?, #to_absolute, #uri_decode, #uri_encode, #uri_parse, #uri_parse_query, #uri_parser, #uri_rewrite

Methods included from UI::Output

#debug?, #debug_off, #debug_on, #disable_only_positives, #included, #mute, #muted?, #only_positives, #only_positives?, #print_bad, #print_debug, #print_debug_backtrace, #print_debug_level_1, #print_debug_level_2, #print_debug_level_3, #print_error, #print_error_backtrace, #print_exception, #print_info, #print_line, #print_ok, #print_status, #print_verbose, #reroute_to_file, #reroute_to_file?, reset_output_options, #unmute, #verbose?, #verbose_on

Constructor Details

#initialize(options = {}) ⇒ Worker

Returns a new instance of Worker.



47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
 
# File 'lib/arachni/browser_cluster/worker.rb', line 47

def initialize( options = {} )
    javascript_token  = options.delete( :javascript_token )
    @master           = options.delete( :master )

    @max_time_to_live = options.delete( :max_time_to_live ) ||
        Options.browser_cluster.worker_time_to_live
    @time_to_live     = @max_time_to_live

    @job_timeout      = options.delete( :job_timeout ) ||
        Options.browser_cluster.job_timeout

    # Don't store pages if there's a master, we'll be sending them to him
    # as soon as they're logged.
    super options.merge( store_pages: false )

    @javascript.token = javascript_token

    @done_signal = Queue.new

    start_capture

    return if !@master
    start
end

Instance Attribute Details

#job ⇒ Job (readonly)

Returns Currently assigned job.

Returns:

  • (Job) —

    Currently assigned job.



35
36
37
 
# File 'lib/arachni/browser_cluster/worker.rb', line 35

def job
  @job
end

#job_timeout ⇒ Integer

Returns:

  • (Integer) 


38
39
40
 
# File 'lib/arachni/browser_cluster/worker.rb', line 38

def job_timeout
  @job_timeout
end

#master ⇒ BrowserCluster (readonly)

Returns:



31
32
33
 
# File 'lib/arachni/browser_cluster/worker.rb', line 31

def master
  @master
end

#max_time_to_live ⇒ Integer

Returns:

  • (Integer) 


41
42
43
 
# File 'lib/arachni/browser_cluster/worker.rb', line 41

def max_time_to_live
  @max_time_to_live
end

#time_to_live ⇒ Integer (readonly)

Returns Remaining time-to-live measured in jobs.

Returns:

  • (Integer) —

    Remaining time-to-live measured in jobs.



45
46
47
 
# File 'lib/arachni/browser_cluster/worker.rb', line 45

def time_to_live
  @time_to_live
end

Class Method Details

.name ⇒ Object 



200
201
202
 
# File 'lib/arachni/browser_cluster/worker.rb', line 200

def self.name
    "BrowserCluster Worker##{object_id}"
end

Instance Method Details

#distribute_event(page, element, event) ⇒ Object

Direct the distribution to the master and let it take it from there.

See Also:



155
156
157
158
159
160
161
162
163
164
165
166
167
 
# File 'lib/arachni/browser_cluster/worker.rb', line 155

def distribute_event( page, element, event )
    master.queue @job.forward_as(
        @job.class::EventTrigger,
        resource: page,
        element:  element,
        event:    event
    )
    true
# Job may have been marked as done or the cluster may have been shut down.
rescue BrowserCluster::Job::Error::AlreadyDone,
    BrowserCluster::Error::AlreadyShutdown
    false
end

#inspect ⇒ Object 



191
192
193
194
195
196
197
198
 
# File 'lib/arachni/browser_cluster/worker.rb', line 191

def inspect
    s = "#<#{self.class} "
    s << "pid=#{@pid} "
    s << "job=#{@job.inspect} "
    s << "last-url=#{@last_url.inspect} "
    s << "transitions=#{@transitions.size}"
    s << '>'
end

#run_job(job) ⇒ Array<Page>

Returns Pages which resulted from firing events, clicking JavaScript links and capturing AJAX requests.

Parameters:

Returns:

  • (Array<Page>) —

    Pages which resulted from firing events, clicking JavaScript links and capturing AJAX requests.

See Also:



79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
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
 
# File 'lib/arachni/browser_cluster/worker.rb', line 79

def run_job( job )
    @job = job
    print_debug "Started: #{@job}"

    # PhantomJS may have crashed (it happens sometimes) so make sure that
    # we've got a live one before running the job.
    # If we can't respawn, then bail out.
    return if browser_respawn_if_necessary.nil?

    begin
        with_timeout @job_timeout do
            exception_jail false do
                begin
                    @job.configure_and_run( self )
                rescue Selenium::WebDriver::Error::WebDriverError
                    browser_respawn
                end
            end
        end
    rescue TimeoutError => e
        print_debug "Job timed-out after #{@job_timeout} seconds: #{@job}"

        # Could have left us with a broken browser.
        browser_respawn
    end

    begin
        watir.cookies.clear
    # Working window was closed by JS (probably), start from scratch.
    rescue Selenium::WebDriver::Error::NoSuchWindowError
        browser_respawn
    end

    decrease_time_to_live
    browser_respawn_if_necessary

    print_debug "Finished: #{@job}"

    true
rescue Selenium::WebDriver::Error::WebDriverError
    browser_respawn
    nil
ensure
    @javascript.taint = nil

    clear_buffers

    # The jobs may have configured callbacks to capture pages etc.,
    # remove them.
    clear_observers

    @job = nil
end

#shutdown(wait = true) ⇒ Object

Note:

If there is a running job it will wait for it to finish.

Shuts down the worker.



172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
 
# File 'lib/arachni/browser_cluster/worker.rb', line 172

def shutdown( wait = true )
    return if @shutdown
    @shutdown = true

    # Keep checking to see if any of the 'done' criteria are true.
    kill_check = Thread.new do
        sleep 0.1 while browser_alive? && wait && @job
        @done_signal << nil
    end

    # If we've got a job running wait for it to finish before closing the
    # browser otherwise we'll get Selenium errors and zombie processes.
    @done_signal.pop
    kill_check.join
    @consumer.kill if @consumer

    super()
end

#trigger_events ⇒ Object

We change the default scheduling to distribute elements and events to all available browsers ASAP, instead of building a list and then consuming it, since we’re don’t have to worry about messing up our page’s state in this setup.

See Also:



139
140
141
142
143
144
145
146
147
148
149
 
# File 'lib/arachni/browser_cluster/worker.rb', line 139

def trigger_events
    root_page = to_page

    each_element_with_events do |element, events|
        events.each do |name, _|
            distribute_event( root_page, element, name.to_sym )
        end
    end

    true
end