Testing Graphical User Interfaces¶

In this chapter, we explore how to generate tests for Graphical User Interfaces (GUIs), abstracting from our previous examples on Web testing. Building on general means to extract user interface elements and activate them, our techniques generalize to arbitrary graphical user interfaces, from rich Web applications to mobile apps, and systematically explore user interfaces through forms and navigation elements.

Prerequisites

  • We build on the Web server introduced in the chapter on Web testing.

Automated GUI Interaction¶

In the chapter on Web testing, we have shown how to test Web-based interfaces by directly interacting with a Web server using the HTTP protocol, and processing the retrieved HTML pages to identify user interface elements. While these techniques work well for user interfaces that are based on HTML only, they fail as soon as there are interactive elements that use JavaScript to execute code within the browser, and generate and change the user interface without having to interact with the browser.

In this chapter, we therefore take a different approach to user interface testing. Rather than using HTTP and HTML as the mechanisms for interaction, we leverage a dedicated UI testing framework, which allows us to

  • query the program under test for available user interface elements, and
  • query the UI elements for how they can be interacted with.

Although we will again illustrate our approach using a Web server, the approach easily generalizes to arbitrary user interfaces. In fact, the UI testing framework we use, Selenium, also comes in variants that run for Android apps.

Our Web Server, Again¶

As in the chapter on Web testing, we run a Web server that allows us to order products.

In [5]:
# ignore
if 'CI' in os.environ:
    # Can't run this in our continuous environment,
    # since it can't run a headless Web browser
    sys.exit(0)
In [8]:
db = init_db()

This is the address of our web server:

In [9]:
httpd_process, httpd_url = start_httpd()
print_url(httpd_url)

http://127.0.0.1:8800

Using webbrowser(), we can retrieve the HTML of the home page, and use HTML() to render it.

In [12]:
HTML(webbrowser(httpd_url))

127.0.0.1 - - [16/Jan/2025 11:12:35] "GET / HTTP/1.1" 200 -
Out[12]:
Fuzzingbook Swag Order Form

Yes! Please send me at your earliest convenience


.

Remote Control with Selenium¶

Let us take a look at the GUI above. In contrast to the chapter on Web testing, we do not assume we can access the HTML source of the current page. All we assume is that there is a set of user interface elements we can interact with.

Selenium is a framework for testing Web applications by automating interaction in the browser. Selenium provides an API that allows one to launch a Web browser, query the state of the user interface, and interact with individual user interface elements. The Selenium API is available in a number of languages; we use the Selenium API for Python.

A Selenium web driver is the interface between a program and a browser controlled by the program. The following code starts a Web browser in the background, which we then control through the web driver.

We support both Firefox and Google Chrome.

In [14]:
BROWSER = 'firefox'  # Set to 'chrome' if you prefer Chrome

Setting up Firefox¶

For Firefox, you have to make sure the geckodriver program is in your path.

In [16]:
if BROWSER == 'firefox':
    assert shutil.which('geckodriver') is not None, \
        "Please install the 'geckodriver' executable " \
        "from https://github.com/mozilla/geckodriver/releases"

Setting up Chrome¶

For Chrome, you may have to make sure the chromedriver program is in your path.

In [17]:
if BROWSER == 'chrome':
    assert shutil.which('chromedriver') is not None, \
        "Please install the 'chromedriver' executable " \
        "from https://chromedriver.chromium.org"

Running a Headless Browser¶

The browser is headless, meaning that it does not show on the screen.

In [18]:
HEADLESS = True

Note: If the notebook server runs locally (i.e. on the same machine on which you are seeing this), you can also set HEADLESS to False and see what happens right on the screen as you execute the notebook cells. This is very much recommended for interactive sessions.

Starting the Web driver¶

This code starts the Selenium web driver.

In [19]:
def start_webdriver(browser=BROWSER, headless=HEADLESS, zoom=1.4):
    # Set headless option
    if browser == 'firefox':
        options = webdriver.FirefoxOptions()
        if headless:
            # See https://www.browserstack.com/guide/firefox-headless
            options.add_argument("--headless")
    elif browser == 'chrome':
        options = webdriver.ChromeOptions()
        if headless:
            # See https://www.selenium.dev/blog/2023/headless-is-going-away/
            options.add_argument("--headless=new")
    else:
        assert False, "Select 'firefox' or 'chrome' as browser"

    # Start the browser, and obtain a _web driver_ object such that we can interact with it.
    if browser == 'firefox':
        # For firefox, set a higher resolution for our screenshots
        options.set_preference("layout.css.devPixelsPerPx", repr(zoom))
        gui_driver = webdriver.Firefox(options=options)

        # We set the window size such that it fits our order form exactly;
        # this is useful for not wasting too much space when taking screen shots.
        gui_driver.set_window_size(700, 300)

    elif browser == 'chrome':
        gui_driver = webdriver.Chrome(options=options)
        gui_driver.set_window_size(700, 210 if headless else 340)

    return gui_driver
In [20]:
gui_driver = start_webdriver(browser=BROWSER, headless=HEADLESS)

The geckodriver version (0.34.0) detected in PATH at /Users/zeller/bin/geckodriver might not be compatible with the detected firefox version (135.03); currently, geckodriver 0.35.0 is recommended for firefox 135.*, so it is advised to delete the driver in PATH and retry

We can now interact with the browser programmatically. First, we have it navigate to the URL of our Web server:

In [21]:
gui_driver.get(httpd_url)

We see that the home page is actually accessed, together with a (failing) request to get a page icon:

In [22]:
print_httpd_messages()

127.0.0.1 - - [16/Jan/2025 11:12:38] "GET / HTTP/1.1" 200 -

127.0.0.1 - - [16/Jan/2025 11:12:38] "GET /favicon.ico HTTP/1.1" 404 -

To see what the "headless" browser displays, we can obtain a screenshot. We see that it actually displays the home page.

In [23]:
Image(gui_driver.get_screenshot_as_png())
Out[23]:

Filling out Forms¶

To interact with the Web page through Selenium and the browser, we can query Selenium for individual elements. For instance, we can access the UI element whose name attribute (as defined in HTML) is "name".

In [25]:
name = gui_driver.find_element(By.NAME, "name")

Once we have an element, we can interact with it. Since name is a text field, we can send it a string using the send_keys() method; the string will be translated into appropriate keystrokes.

In [26]:
name.send_keys("Jane Doe")

In the screenshot, we can see that the name field is now filled:

In [27]:
Image(gui_driver.get_screenshot_as_png())
Out[27]:

Similarly, we can fill out the email, city, and ZIP fields:

In [28]:
email = gui_driver.find_element(By.NAME, "email")
email.send_keys("j.doe@example.com")
In [29]:
city = gui_driver.find_element(By.NAME, 'city')
city.send_keys("Seattle")
In [30]:
zip = gui_driver.find_element(By.NAME, 'zip')
zip.send_keys("98104")
In [31]:
Image(gui_driver.get_screenshot_as_png())
Out[31]:

The check box for terms and conditions is not filled out, but clicked instead using the click() method.

In [32]:
terms = gui_driver.find_element(By.NAME, 'terms')
terms.click()
In [33]:
Image(gui_driver.get_screenshot_as_png())
Out[33]:

The form is now fully filled out. By clicking on the submit button, we can place the order:

In [34]:
submit = gui_driver.find_element(By.NAME, 'submit')
submit.click()

We see that the order is being processed, and that the Web browser has switched to the confirmation page.

In [35]:
print_httpd_messages()

127.0.0.1 - - [16/Jan/2025 11:12:39] INSERT INTO orders VALUES ('tshirt', 'Jane Doe', 'j.doe@example.com', 'Seattle', '98104')

127.0.0.1 - - [16/Jan/2025 11:12:39] "GET /order?item=tshirt&name=Jane+Doe&email=j.doe%40example.com&city=Seattle&zip=98104&terms=on&submit=Place+order HTTP/1.1" 200 -
In [36]:
Image(gui_driver.get_screenshot_as_png())
Out[36]:

Navigating¶

Just as we fill out forms, we can also navigate through a website by clicking on links. Let us go back to the home page:

In [37]:
gui_driver.back()
In [38]:
Image(gui_driver.get_screenshot_as_png())
Out[38]:

We can query the web driver for all elements of a particular type. Querying for HTML anchor elements (<a>) for instance, gives us all links on a page.

In [39]:
links = gui_driver.find_elements(By.TAG_NAME, "a")

We can query the attributes of UI elements – for instance, the URL the first anchor on the page links to:

In [40]:
links[0].get_attribute('href')
Out[40]:
'http://127.0.0.1:8800/terms'

What happens if we click on it? Very simple: We switch to the Web page being referenced.

In [41]:
links[0].click()
In [42]:
print_httpd_messages()

127.0.0.1 - - [16/Jan/2025 11:12:39] "GET /terms HTTP/1.1" 200 -
In [43]:
Image(gui_driver.get_screenshot_as_png())
Out[43]:

Okay. Let's get back to our home page again.

In [44]:
gui_driver.back()
In [45]:
print_httpd_messages()
In [46]:
Image(gui_driver.get_screenshot_as_png())
Out[46]:

Writing Test Cases¶

The above calls, interacting with a user interface automatically, are typically used in Selenium tests – that is, code snippets that interact with a website, occasionally checking whether everything works as expected. The following code, for instance, places an order just as above. It then retrieves the title element and checks whether the title contains a "Thank you" message, indicating success.

In [47]:
def test_successful_order(driver, url):
    name = "Walter White"
    email = "white@jpwynne.edu"
    city = "Albuquerque"
    zip_code = "87101"

    driver.get(url)
    driver.find_element(By.NAME, "name").send_keys(name)
    driver.find_element(By.NAME, "email").send_keys(email)
    driver.find_element(By.NAME, 'city').send_keys(city)
    driver.find_element(By.NAME, 'zip').send_keys(zip_code)
    driver.find_element(By.NAME, 'terms').click()
    driver.find_element(By.NAME, 'submit').click()

    title = driver.find_element(By.ID, 'title')
    assert title is not None
    assert title.text.find("Thank you") >= 0

    confirmation = driver.find_element(By.ID, "confirmation")
    assert confirmation is not None

    assert confirmation.text.find(name) >= 0
    assert confirmation.text.find(email) >= 0
    assert confirmation.text.find(city) >= 0
    assert confirmation.text.find(zip_code) >= 0

    return True
In [48]:
test_successful_order(gui_driver, httpd_url)
Out[48]:
True

In a similar vein, we can set up automated test cases for unsuccessful orders, canceling orders, changing orders, and many more. All these test cases would be automatically run after any change to the program code, ensuring the Web application still works.

Of course, writing such tests is quite some effort. Hence, in the remainder of this chapter, we will again explore how to automatically generate them.

Retrieving User Interface Actions¶

To automatically interact with a user interface, we first need to find out which elements there are, and which user interactions (or short actions) they support.

User Interface Elements¶

We start with finding available user elements. Let us get back to the order form.

In [49]:
gui_driver.get(httpd_url)
In [50]:
Image(gui_driver.get_screenshot_as_png())
Out[50]:

Using find_elements(By.TAG_NAME, ) (and other similar find_elements_...() functions), we can retrieve all elements of a particular type, such as HTML input elements.

In [51]:
ui_elements = gui_driver.find_elements(By.TAG_NAME, "input")

For each element, we can retrieve its HTML attributes, using get_attribute(). We can thus retrieve the name and type of each input element (if defined).

In [52]:
for element in ui_elements:
    print("Name: %-10s | Type: %-10s | Text: %s" %
          (element.get_attribute('name'),
           element.get_attribute('type'),
           element.text))

Name: name       | Type: text       | Text: 
Name: email      | Type: email      | Text: 
Name: city       | Type: text       | Text: 
Name: zip        | Type: number     | Text: 
Name: terms      | Type: checkbox   | Text: 
Name: submit     | Type: submit     | Text:
In [53]:
ui_elements = gui_driver.find_elements(By.TAG_NAME, "a")
In [54]:
for element in ui_elements:
    print("Name: %-10s | Type: %-10s | Text: %s" %
          (element.get_attribute('name'),
           element.get_attribute('type'),
           element.text))

Name:            | Type:            | Text: terms and conditions

User Interface Actions¶

Similarly to what we did in the chapter on Web fuzzing, our idea is now to mine a grammar for the user interface – first for an individual user interface page (i.e., a single Web page), later for all pages offered by the application. The idea is that a grammar defines legal sequences of actions – clicks and keystrokes – that can be applied on the application.

We assume the following actions:

  1. fill(<name>, <text>) – fill the UI input element named <name> with the text <text>.
  2. check(<name>, <value>) – set the UI checkbox <name> to the given value <value> (True or False)
  3. submit(<name>) – submit the form by clicking on the UI element <name>.
  4. click(<name>) – click on the UI element <name>, typically for following a link.

This sequence of actions, for instance would fill out the order form:

fill('name', "Walter White")
fill('email', "white@jpwynne.edu")
fill('city', "Albuquerque")
fill('zip', "87101")
check('terms', True)
submit('submit')

Our set of actions is deliberately defined to be small – for real user interfaces, one would also have to define interactions such as swipes, double clicks, long clicks, right button clicks, modifier keys, and more. Selenium supports all of this; but in the interest of simplicity, we focus on the most important set of interactions.

Retrieving Actions¶

As a first step in mining an action grammar, we need to be able to retrieve possible interactions. We introduce a class GUIGrammarMiner, which is set to do precisely that.

In [55]:
class GUIGrammarMiner:
    """Retrieve a grammar of possible GUI interaction sequences"""

    def __init__(self, driver, stay_on_host: bool = True) -> None:
        """Constructor.
        `driver` - a web driver as produced by Selenium.
        `stay_on_host` - if True (default), no not follow links to other hosts.
        """
        self.driver = driver
        self.stay_on_host = stay_on_host
        self.grammar: Grammar = {}

Let us show GUIGrammarMiner in action, using its mine_state_actions() method to retrieve all elements from our current page. We see that we obtain input element actions, button element actions, and link element actions.

In [69]:
gui_grammar_miner = GUIGrammarMiner(gui_driver)
gui_grammar_miner.mine_state_actions()
Out[69]:
frozenset({"check('terms', <boolean>)",
           "click('terms and conditions')",
           "fill('city', '<text>')",
           "fill('email', '<email>')",
           "fill('name', '<text>')",
           "fill('zip', '<number>')",
           "submit('submit')"})

We assume that we can identify a user interface state from the set of interactive elements it contains – that is, the current Web page is identified by the set above. This is in contrast to Web fuzzing, where we assumed the URL to uniquely characterize a page – but with JavaScript, the URL can stay unchanged although the page contents change, and UIs other than the Web may have no concept of unique URLs. Therefore, we say that the way a UI can be interacted with uniquely defines its state.

Models for User Interfaces¶

User Interfaces as Finite State Machines¶

Now that we can retrieve UI elements from a page, let us go and systematically explore a user interface. The idea is to represent the user interface as a finite state machine – that is, a sequence of states that can be reached by interacting with the individual user interface elements.

Let us illustrate such a finite state machine by looking at our Web server. The following diagram shows the states our server can be in:

In [70]:
# ignore
from graphviz import Digraph
In [71]:
# ignore
from GrammarFuzzer import dot_escape
In [72]:
# ignore
dot = Digraph(comment="Finite State Machine")
dot.node(dot_escape('<start>'))
dot.edge(dot_escape('<start>'),
         dot_escape('<Order Form>'))
dot.edge(dot_escape('<Order Form>'),
         dot_escape('<Terms and Conditions>'), "click('Terms and conditions')")
dot.edge(dot_escape('<Order Form>'),
         dot_escape('<Thank You>'), r"fill(...)\lsubmit('submit')")
dot.edge(dot_escape('<Terms and Conditions>'),
         dot_escape('<Order Form>'), "click('order form')")
dot.edge(dot_escape('<Thank You>'),
         dot_escape('<Order Form>'), "click('order form')")
display(dot)
\<start\> <start> \<Order Form\> <Order Form> \<start\>->\<Order Form\> \<Terms and Conditions\> <Terms and Conditions> \<Order Form\>->\<Terms and Conditions\> click('Terms and conditions') \<Thank You\> <Thank You> \<Order Form\>->\<Thank You\> fill(...) submit('submit') \<Terms and Conditions\>->\<Order Form\> click('order form') \<Thank You\>->\<Order Form\> click('order form')

Initially, we are in the <Order Form> state. From here, we can click on Terms and Conditions, and we'll be in the Terms and Conditions state, showing the page with the same title. We can also fill out the form and place the order, having us end in the Thank You state (again showing the page with the same title). From both <Terms and Conditions> and <Thank You>, we can return to the order form by clicking on the order form link.

State Machines as Grammars¶

To systematically explore a user interface, we must retrieve its finite state machine, and eventually cover all states and transitions. In the presence of forms, such an exploration is difficult, as we need a special mechanism to fill out forms and submit the values to get to the next state. There is a trick, though, which allows us to have a single representation for both states and (form) values. We can embed the finite state machine into a grammar, which is then used for both states and form values.

To embed a finite state machine into a grammar, we proceed as follows:

  1. Every state $\langle s \rangle$ in the finite state machine becomes a symbol $\langle s \rangle$ in the grammar.
  2. Every transition in the finite state machine from $\langle s \rangle$ to $\langle t \rangle$ and actions $a_1, a_2, \dots$ becomes an alternative of $\langle s \rangle$ in the form $a_1, a_2, dots$ $\langle t \rangle$ in the grammar.

The above finite state machine thus gets encoded into the grammar

<start> ::= <Order Form>
<Order Form> ::= click('Terms and Conditions') <Terms and Conditions> | 
                 fill(...) submit('submit') <Thank You>
<Terms and Conditions> ::= click('order form') <Order Form>
<Thank You> ::= click('order form') <Order Form>

Expanding this grammar gets us a stream of actions, navigating through the user interface:

fill(...) submit('submit') click('order form') click('Terms and Conditions') click('order form') ...

This stream is actually infinite (as one can interact with the UI forever); to have it end, one can introduce an alternative <end> that simply expands to the empty string, without having any expansion (state) follow.

Retrieving State Grammars¶

Let us extend GUIGrammarMiner such that it retrieves a grammar from the user interface in its current state.

Let us show GUIGrammarMiner() in action. Its method mine_state_grammar() extracts the grammar for the current Web page:

In [81]:
gui_grammar_miner = GUIGrammarMiner(gui_driver)
state_grammar = gui_grammar_miner.mine_state_grammar()
In [82]:
state_grammar
Out[82]:
{'<start>': ['<state>'],
 '<unexplored>': [''],
 '<end>': [''],
 '<text>': ['<string>'],
 '<string>': ['<character>', '<string><character>'],
 '<character>': ['<letter>', '<digit>', '<special>'],
 '<letter>': ['a',
  'b',
  'c',
  'd',
  'e',
  'f',
  'g',
  'h',
  'i',
  'j',
  'k',
  'l',
  'm',
  'n',
  'o',
  'p',
  'q',
  'r',
  's',
  't',
  'u',
  'v',
  'w',
  'x',
  'y',
  'z',
  'A',
  'B',
  'C',
  'D',
  'E',
  'F',
  'G',
  'H',
  'I',
  'J',
  'K',
  'L',
  'M',
  'N',
  'O',
  'P',
  'Q',
  'R',
  'S',
  'T',
  'U',
  'V',
  'W',
  'X',
  'Y',
  'Z'],
 '<number>': ['<digits>'],
 '<digits>': ['<digit>', '<digits><digit>'],
 '<digit>': ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9'],
 '<special>': ['.', ' ', '!'],
 '<email>': ['<letters>@<letters>'],
 '<letters>': ['<letter>', '<letters><letter>'],
 '<boolean>': ['True', 'False'],
 '<state>': ["click('terms and conditions')\n<state-1>",
  "fill('city', '<text>')\nfill('email', '<email>')\nfill('name', '<text>')\ncheck('terms', <boolean>)\nfill('zip', '<number>')\nsubmit('submit')\n<state-2>",
  '<end>'],
 '<state-1>': ['<unexplored>'],
 '<state-2>': ['<unexplored>']}

To better see the structure of the state grammar, we can visualize it as a state machine. We see that it nicely reflects what we can see from our Web server's home page:

In [83]:
fsm_diagram(state_grammar)
start <start> state <state> start->state state-1 <state-1> state->state-1 click('terms and conditions') state-2 <state-2> state->state-2 fill('city', '<text>') fill('email', '<email>') fill('name', '<text>') check('terms', <boolean>) fill('zip', '<number>') submit('submit') end <end> state->end unexplored <unexplored> state-1->unexplored state-2->unexplored

From the start state (<state>), we can go and either click on "terms and conditions", ending in <state-1>, or fill out the form, ending in <state-2>.

In [84]:
state_grammar[GUIGrammarMiner.START_STATE]
Out[84]:
["click('terms and conditions')\n<state-1>",
 "fill('city', '<text>')\nfill('email', '<email>')\nfill('name', '<text>')\ncheck('terms', <boolean>)\nfill('zip', '<number>')\nsubmit('submit')\n<state-2>",
 '<end>']

Both these states are yet unexplored:

In [85]:
state_grammar['<state-1>']
Out[85]:
['<unexplored>']
In [86]:
state_grammar['<state-2>']
Out[86]:
['<unexplored>']
In [87]:
state_grammar['<unexplored>']
Out[87]:
['']

Given the grammar, we can use any of our grammar fuzzers to create valid input sequences:

In [89]:
gui_fuzzer = GrammarFuzzer(state_grammar)
while True:
    action = gui_fuzzer.fuzz()
    if action.find('submit(') > 0:
        break
print(action)

fill('city', '.')
fill('email', 'EB@iYN')
fill('name', '.')
check('terms', True)
fill('zip', '3')
submit('submit')

These actions, however, must also be executed such that we can explore the user interface. This is what we do in the next section.

Executing User Interface Actions¶

To execute actions, we introduce a Runner class, conveniently named GUIRunner. Its run() method executes the actions as given in an action string.

In [91]:
class GUIRunner(Runner):
    """Execute the actions in a given action string"""

    def __init__(self, driver) -> None:
        """Constructor. `driver` is a Selenium Web driver"""
        self.driver = driver

Let us try out GUIRunner and its run() method. We create a runner on our Web server, and let it execute a fill() action:

In [101]:
gui_driver.get(httpd_url)
In [102]:
gui_runner = GUIRunner(gui_driver)
In [103]:
gui_runner.run("fill('name', 'Walter White')")
Out[103]:
("fill('name', 'Walter White')", 'PASS')
In [104]:
Image(gui_driver.get_screenshot_as_png())
Out[104]:

A submit() action submits the order. (Note that our Web server does no effort whatsoever to validate the form.)

In [105]:
gui_runner.run("submit('submit')")
Out[105]:
("submit('submit')", 'PASS')
In [106]:
Image(gui_driver.get_screenshot_as_png())
Out[106]:

Of course, we can also execute action sequences generated from the grammar. This allows us to fill the form again and again, using values matching the type given in the form.

In [107]:
gui_driver.get(httpd_url)
In [108]:
gui_fuzzer = GrammarFuzzer(state_grammar)
In [109]:
while True:
    action = gui_fuzzer.fuzz()
    if action.find('submit(') > 0:
        break
In [110]:
print(action)

fill('city', 'S0.')
fill('email', 'o@i')
fill('name', 'MF')
check('terms', False)
fill('zip', '7')
submit('submit')
In [111]:
gui_runner.run(action)
Out[111]:
("fill('city', 'S0.')\nfill('email', 'o@i')\nfill('name', 'MF')\ncheck('terms', False)\nfill('zip', '7')\nsubmit('submit')\n",
 'PASS')
In [112]:
Image(gui_driver.get_screenshot_as_png())
Out[112]:

Exploring User Interfaces¶

So far, our grammar retrieval and execution of actions is limited to the current user interface state (i.e., the current page shown). To systematically explore a user interface, we must explore all states, notably those ending in <unexplored> – and whenever we reach a new state, again retrieve its grammar such that we may be able to reach other states. Since some states can only be reached by generating inputs, test generation and user interface exploration take place at the same time.

Consequently, we introduce a GUIFuzzer class, which generates inputs for all forms and follows all links, and which updates its grammar (i.e., its user interface model as a finite state machine) every time it encounters a new state.

Let us put GUIFuzzer to use, enabling its logging mechanisms to see what it is doing.

In [135]:
gui_driver.get(httpd_url)
In [136]:
gui_fuzzer = GUIFuzzer(gui_driver, log_gui_exploration=True, disp_gui_exploration=True)

Running it the first time yields a new state:

In [137]:
gui_fuzzer.run(gui_runner)

Action click('terms and conditions') -> <state-1>

In new state <state-1> frozenset({"ignore('Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.')", "click('order form')"})
start <start> state <state> start->state state-1 <state-1> state->state-1 click('terms and conditions') state-2 <state-2> state->state-2 fill('city', '<text>') fill('email', '<email>') fill('name', '<text>') check('terms', <boolean>) fill('zip', '<number>') submit('submit') end <end> state->end state-1->end state-3 <state-3> state-1->state-3 click('order form') unexplored <unexplored> state-2->unexplored state-3->unexplored 
None
Out[137]:
('<state-1>', 'PASS')

The next actions fill out the order form.

In [138]:
gui_fuzzer.run(gui_runner)

Action click('terms and conditions') -> <end>
Out[138]:
('<end>', 'PASS')
In [139]:
gui_fuzzer.run(gui_runner)

Action click('terms and conditions') -> <end>
Out[139]:
('<end>', 'PASS')

At this point, our GUI model is fairly complete already. In order to systematically cover all states, random exploration is not efficient enough, though.

Covering States¶

During exploration as well as during testing, we want to cover all states and transitions between states. How can we achieve this?

It turns out that we already have this. Our GrammarCoverageFuzzer from the chapter on coverage-based grammar testing strives to systematically cover all expansion alternatives in a grammar. In the finite state model, these expansion alternatives translate into transitions between states. Hence, applying the coverage strategy from GrammarCoverageFuzzer to our state grammars would automatically cover one transition after another.

How do we get these features into GUIFuzzer? Using multiple inheritance, we can create a class GUICoverageFuzzer which combines the run() method from GUIFuzzer with the coverage choices from GrammarCoverageFuzzer.

Since the __init__() constructor is defined in both superclasses, we need to define our own constructor that serves both:

In [142]:
inheritance_conflicts(GUIFuzzer, GrammarCoverageFuzzer)
Out[142]:
['__init__']
In [143]:
class GUICoverageFuzzer(GUIFuzzer, GrammarCoverageFuzzer):
    """Systematically explore all states of the current Web page"""

    def __init__(self, *args, **kwargs):
        """Constructor. All args are passed to the `GUIFuzzer` superclass."""
        GUIFuzzer.__init__(self, *args, **kwargs)
        self.reset_coverage()

With GUICoverageFuzzer, we can set up a method explore_all() that keeps on running the fuzzer until there are no unexplored states anymore:

In [144]:
class GUICoverageFuzzer(GUICoverageFuzzer):
    def explore_all(self, runner: GUIRunner, max_actions=100) -> None:
        """Explore all states of the GUI, up to `max_actions` (default 100)."""

        actions = 0
        while (self.miner.UNEXPLORED_STATE in self.grammar and 
               actions < max_actions):
            actions += 1
            if self.log_gui_exploration:
                print("Run #" + repr(actions))
            try:
                self.run(runner)
            except ElementClickInterceptedException:
                pass
            except ElementNotInteractableException:
                pass
            except NoSuchElementException:
                pass

Let us use this to fully explore our Web server:

In [145]:
gui_driver.get(httpd_url)
In [146]:
gui_fuzzer = GUICoverageFuzzer(gui_driver)
In [147]:
gui_fuzzer.explore_all(gui_runner)

Success! We have covered all states:

In [148]:
fsm_diagram(gui_fuzzer.grammar)
start <start> state <state> start->state state-1 <state-1> state->state-1 click('terms and conditions') state-2 <state-2> state->state-2 fill('city', '<text>') fill('email', '<email>') fill('name', '<text>') check('terms', <boolean>) fill('zip', '<number>') submit('submit') end <end> state->end state-1->state click('order form') state-1->end state-2->state click('order form') state-2->end

We can retrieve the expansions covered so far, which of course cover all states.

In [149]:
gui_fuzzer.covered_expansions
Out[149]:
{'<boolean> -> False',
 '<boolean> -> True',
 '<character> -> <digit>',
 '<character> -> <letter>',
 '<character> -> <special>',
 '<digit> -> 0',
 '<digit> -> 1',
 '<digit> -> 2',
 '<digit> -> 3',
 '<digit> -> 4',
 '<digit> -> 5',
 '<digit> -> 6',
 '<digit> -> 7',
 '<digit> -> 8',
 '<digit> -> 9',
 '<digits> -> <digit>',
 '<digits> -> <digits><digit>',
 '<email> -> <letters>@<letters>',
 '<end> -> ',
 '<letter> -> A',
 '<letter> -> B',
 '<letter> -> D',
 '<letter> -> H',
 '<letter> -> J',
 '<letter> -> K',
 '<letter> -> L',
 '<letter> -> M',
 '<letter> -> P',
 '<letter> -> Q',
 '<letter> -> W',
 '<letter> -> Y',
 '<letter> -> b',
 '<letter> -> g',
 '<letter> -> h',
 '<letter> -> l',
 '<letter> -> m',
 '<letter> -> n',
 '<letter> -> q',
 '<letter> -> t',
 '<letter> -> v',
 '<letter> -> y',
 '<letters> -> <letter>',
 '<letters> -> <letters><letter>',
 '<number> -> <digits>',
 '<special> -> .',
 '<start> -> <state>',
 '<state-1> -> <end>',
 '<state-1> -> <unexplored>',
 "<state-1> -> click('order form')\n<state-3>",
 '<state-2> -> <end>',
 '<state-2> -> <unexplored>',
 "<state-2> -> click('order form')\n<state-4>",
 '<state-3> -> <unexplored>',
 '<state-4> -> <unexplored>',
 '<state> -> <end>',
 "<state> -> click('terms and conditions')\n<state-1>",
 "<state> -> fill('city', '<text>')\nfill('email', '<email>')\nfill('name', '<text>')\ncheck('terms', <boolean>)\nfill('zip', '<number>')\nsubmit('submit')\n<state-2>",
 '<string> -> <character>',
 '<string> -> <string><character>',
 '<text> -> <string>',
 '<unexplored> -> '}

Still, we haven't seen all expansions covered. A few digits and letters remain to be used.

In [150]:
gui_fuzzer.missing_expansion_coverage()
Out[150]:
{'<letter> -> C',
 '<letter> -> E',
 '<letter> -> F',
 '<letter> -> G',
 '<letter> -> I',
 '<letter> -> N',
 '<letter> -> O',
 '<letter> -> R',
 '<letter> -> S',
 '<letter> -> T',
 '<letter> -> U',
 '<letter> -> V',
 '<letter> -> X',
 '<letter> -> Z',
 '<letter> -> a',
 '<letter> -> c',
 '<letter> -> d',
 '<letter> -> e',
 '<letter> -> f',
 '<letter> -> i',
 '<letter> -> j',
 '<letter> -> k',
 '<letter> -> o',
 '<letter> -> p',
 '<letter> -> r',
 '<letter> -> s',
 '<letter> -> u',
 '<letter> -> w',
 '<letter> -> x',
 '<letter> -> z',
 '<special> ->  ',
 '<special> -> !',
 "<state-1> -> click('order form')\n<state>",
 "<state-2> -> click('order form')\n<state>"}

Running the fuzzer again and again will eventually cover these expansions too, leading to letter and digit coverage within the order form.

Exploring Large Sites¶

Our GUI fuzzer is robust enough to handle exploration even on nontrivial sites such as fuzzingbook.org. Let us demonstrate this:

In [151]:
gui_driver.get("https://www.fuzzingbook.org/html/Fuzzer.html")
In [152]:
Image(gui_driver.get_screenshot_as_png())
Out[152]:
In [153]:
book_runner = GUIRunner(gui_driver)
In [154]:
book_fuzzer = GUICoverageFuzzer(gui_driver, log_gui_exploration=True)  # , disp_gui_exploration=True)

We explore the first few states of the site, defined in ACTIONS:

In [155]:
ACTIONS = 5
In [156]:
book_fuzzer.explore_all(book_runner, max_actions=ACTIONS)

Run #1
Action click('discussed above') -> <state-7>
In existing state <state>
Replacing expected state <state-7> by <state>
Run #2
Action click('use the code provided in this chapter') -> <state-11>
In new state <state-11> frozenset({"ignore('installation instructions')", "click('the chapter on fuzzers')", "click('')", "click('Fuzzer')", "ignore('official instructions')", "click('Cite')", "ignore('apt.txt file in the binder/ folder')", "ignore('Last change: 2023-11-11 18:18:06+01:00')", "ignore('Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License')", "ignore('Pipenv')", "ignore('')", "ignore('bookutils')", "click('fuzzingbook.Fuzzer')", "click('The Fuzzing Book')", "ignore('requirements.txt file within the project root folder')", "ignore('MIT License')", "ignore('the project page')", "ignore('Imprint')", "ignore('pyenv-win')", "ignore('bookutils.setup')", "click('fuzzingbook.')"})
Run #3
Action click('Intro_Testing') -> <state-10>
In existing state <state>
Replacing expected state <state-10> by <state>
Run #4
Action click('chapter on mining function specifications') -> <state-5>
In new state <state-5> frozenset({"click('ExpectError')", "ignore('Last change: 2024-11-09 17:07:29+01:00')", "ignore('itertools')", "ignore('showast')", "ignore('curated list')", "click('symbolic fuzzing')", "click('Cite')", "ignore('tempfile')", "click('part on semantic fuzzing techniques')", "ignore('Ernst et al, 2001')", "ignore('Use the notebook')", "click('symbolic')", "ignore('MonkeyType')", "ignore('DAIKON dynamic invariant detector')", "click('introduction to testing')", "click('Grammars')", "click('The Fuzzing Book')", "click('GrammarFuzzer')", "click('symbolic interpretation')", "click('domain-specific fuzzing techniques')", "click('Intro_Testing')", "ignore('subprocess')", "click('fuzzingbook.DynamicInvariants')", "click('our chapter with the same name')", "click('concolic fuzzer')", "click('concolic')", "click('chapter on testing')", "ignore('code snippet from StackOverflow')", "ignore('Pacheco et al, 2005')", "ignore('sys')", "ignore('ast')", "ignore('&quot;The state of type hints in Python&quot;')", "click('chapter on coverage')", "ignore('functools')", "ignore('Ammons et al, 2002')", "ignore('Mypy')", "click('the next part')", "ignore('')", "ignore('Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License')", "click('chapter on information flow')", "ignore('typing')", "ignore('bookutils.setup')", "ignore('bookutils')", "ignore('MIT License')", "ignore('PyAnnotate')", "ignore('Imprint')", "click('')", "ignore('inspect')", "click('use the code provided in this chapter')", "click('Coverage')"})
Run #5
Action click('chapter on testing') -> <state-14>
In new state <state-14> frozenset({"click('ExpectError')", "check('11b29b38-9eb5-11ef-9f1d-6298cf1a578f', <boolean>)", "ignore('Beizer et al, 1990')", "click('Timer')", "click('Cite')", "click('Timer module')", "check('1251ba2e-9eb5-11ef-9f1d-6298cf1a578f', <boolean>)", "ignore('random')", "ignore('Use the notebook')", "click('use fuzzing to test programs with random inputs')", "ignore('Shellsort')", "click('The Fuzzing Book')", "click('import it')", "ignore('Newton\xe2\x80\x93Raphson method')", "click('00_Table_of_Contents.ipynb')", "ignore('Python tutorial')", "submit('')", "ignore('math.isclose()')", "click('Web Page')", "ignore('Pezz\xc3\xa8 et al, 2008')", "ignore('&quot;Effective Software Testing: A Developer&#x27;s Guide&quot;')", "click('Background')", "ignore('Maur\xc3\xadcio Aniche, 2022')", "ignore('Last change: 2023-11-11 18:18:06+01:00')", "check('119cfd32-9eb5-11ef-9f1d-6298cf1a578f', <boolean>)", "ignore('Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License')", "ignore('')", "ignore('bookutils')", "ignore('bookutils.setup')", "ignore('MIT License')", "ignore('Myers et al, 2004')", "ignore('Imprint')", "click('')", "click('Guide for Authors')"})

After the first ACTIONS actions already, we can see that the finite state model is quite complex, with dozens of transitions still left to explore. Most of the yet unexplored states will eventually merge with existing states, yielding one state per chapter. Still, following all links on all pages will take quite some time.

In [157]:
# Inspect this graph in the notebook to see it in full glory
fsm_diagram(book_fuzzer.grammar)
start <start> state <state> start->state state->state click('discussed above') state->state click('Intro_Testing') state-1 <state-1> state->state-1 click('A Fuzzing Architecture') state-2 <state-2> state->state-2 click('ExpectError') state-3 <state-3> state->state-3 click('runtime verification') state-4 <state-4> state->state-4 click('Cite') state-5 <state-5> state->state-5 click('chapter on mining function specifications') state-6 <state-6> state->state-6 click('use grammars to specify the input format and thus get many more valid inputs') state-8 <state-8> state->state-8 click('The Fuzzing Book') state-9 <state-9> state->state-9 click('"Introduction to Software Testing"') state-11 <state-11> state->state-11 click('use the code provided in this chapter') state-12 <state-12> state->state-12 click('Introduction to Testing') state-13 <state-13> state->state-13 click('use mutations on existing inputs to get more valid inputs') state-14 <state-14> state->state-14 click('chapter on testing') state-15 <state-15> state->state-15 click('') state-16 <state-16> state->state-16 click('reduce failing inputs for efficient debugging') state-17 <state-17> state->state-17 click('chapter on information flow') state-18 <state-18> state->state-18 click('fuzzingbook.Fuzzer') state-19 <state-19> state->state-19 check('19b2920c-9eb5-11ef-aaff-6298cf1a578f', <boolean>) check('19bb868c-9eb5-11ef-aaff-6298cf1a578f', <boolean>) submit('') end <end> state->end unexplored <unexplored> state-1->unexplored state-2->unexplored state-3->unexplored state-4->unexplored state-5->end state-10 <state-10> state-5->state-10 click('ExpectError') state-26 <state-26> state-5->state-26 click('symbolic fuzzing') state-27 <state-27> state-5->state-27 click('Cite') state-28 <state-28> state-5->state-28 click('part on semantic fuzzing techniques') state-29 <state-29> state-5->state-29 click('symbolic') state-30 <state-30> state-5->state-30 click('introduction to testing') state-31 <state-31> state-5->state-31 click('Grammars') state-32 <state-32> state-5->state-32 click('The Fuzzing Book') state-33 <state-33> state-5->state-33 click('GrammarFuzzer') state-34 <state-34> state-5->state-34 click('symbolic interpretation') state-35 <state-35> state-5->state-35 click('domain-specific fuzzing techniques') state-36 <state-36> state-5->state-36 click('Intro_Testing') state-37 <state-37> state-5->state-37 click('fuzzingbook.DynamicInvariants') state-38 <state-38> state-5->state-38 click('our chapter with the same name') state-39 <state-39> state-5->state-39 click('concolic fuzzer') state-40 <state-40> state-5->state-40 click('concolic') state-41 <state-41> state-5->state-41 click('chapter on testing') state-42 <state-42> state-5->state-42 click('chapter on coverage') state-43 <state-43> state-5->state-43 click('the next part') state-44 <state-44> state-5->state-44 click('chapter on information flow') state-45 <state-45> state-5->state-45 click('') state-46 <state-46> state-5->state-46 click('use the code provided in this chapter') state-47 <state-47> state-5->state-47 click('Coverage') state-6->unexplored state-8->unexplored state-9->unexplored state-11->end state-7 <state-7> state-11->state-7 click('the chapter on fuzzers') state-20 <state-20> state-11->state-20 click('') state-21 <state-21> state-11->state-21 click('Fuzzer') state-22 <state-22> state-11->state-22 click('Cite') state-23 <state-23> state-11->state-23 click('fuzzingbook.Fuzzer') state-24 <state-24> state-11->state-24 click('The Fuzzing Book') state-25 <state-25> state-11->state-25 click('fuzzingbook.') state-12->unexplored state-13->unexplored state-14->end state-48 <state-48> state-14->state-48 click('ExpectError') state-49 <state-49> state-14->state-49 click('Timer') state-50 <state-50> state-14->state-50 click('Cite') state-51 <state-51> state-14->state-51 click('Timer module') state-52 <state-52> state-14->state-52 click('use fuzzing to test programs with random inputs') state-53 <state-53> state-14->state-53 click('The Fuzzing Book') state-54 <state-54> state-14->state-54 click('import it') state-55 <state-55> state-14->state-55 click('00_Table_of_Contents.ipynb') state-56 <state-56> state-14->state-56 click('Web Page') state-57 <state-57> state-14->state-57 click('Background') state-58 <state-58> state-14->state-58 click('') state-59 <state-59> state-14->state-59 click('Guide for Authors') state-60 <state-60> state-14->state-60 check('11b29b38-9eb5-11ef-9f1d-6298cf1a578f', <boolean>) check('1251ba2e-9eb5-11ef-9f1d-6298cf1a578f', <boolean>) check('119cfd32-9eb5-11ef-9f1d-6298cf1a578f', <boolean>) submit('') state-15->unexplored state-16->unexplored state-17->unexplored state-18->unexplored state-19->unexplored state-10->unexplored state-26->unexplored state-27->unexplored state-28->unexplored state-29->unexplored state-30->unexplored state-31->unexplored state-32->unexplored state-33->unexplored state-34->unexplored state-35->unexplored state-36->unexplored state-37->unexplored state-38->unexplored state-39->unexplored state-40->unexplored state-41->unexplored state-42->unexplored state-43->unexplored state-44->unexplored state-45->unexplored state-46->unexplored state-47->unexplored state-7->unexplored state-20->unexplored state-21->unexplored state-22->unexplored state-23->unexplored state-24->unexplored state-25->unexplored state-48->unexplored state-49->unexplored state-50->unexplored state-51->unexplored state-52->unexplored state-53->unexplored state-54->unexplored state-55->unexplored state-56->unexplored state-57->unexplored state-58->unexplored state-59->unexplored state-60->unexplored

We now have all the basic capabilities we need: We can automatically explore large websites; we can explore "deep" functionality by filling out forms; and we can have our coverage-based fuzzer automatically focus on yet unexplored states. Still, there is a lot more one can do; the exercises will give you some ideas.

In [158]:
gui_driver.quit()

Synopsis¶

This chapter demonstrates how to programmatically interact with user interfaces, using Selenium on Web browsers. It provides an experimental GUICoverageFuzzer class that automatically explores a user interface by systematically interacting with all available user interface elements.

The function start_webdriver() starts a headless Web browser in the background and returns a GUI driver as handle for further communication.

In [159]:
gui_driver = start_webdriver()

The geckodriver version (0.34.0) detected in PATH at /Users/zeller/bin/geckodriver might not be compatible with the detected firefox version (135.03); currently, geckodriver 0.35.0 is recommended for firefox 135.*, so it is advised to delete the driver in PATH and retry

We let the browser open the URL of the server we want to investigate (in this case, the vulnerable server from the chapter on Web fuzzing) and obtain a screenshot.

In [160]:
gui_driver.get(httpd_url)
Image(gui_driver.get_screenshot_as_png())
Out[160]:

The GUICoverageFuzzer class explores the user interface and builds a grammar that encodes all states as well as the user interactions required to move from one state to the next. It is paired with a GUIRunner which interacts with the GUI driver.

In [161]:
gui_fuzzer = GUICoverageFuzzer(gui_driver)
In [162]:
gui_runner = GUIRunner(gui_driver)

The explore_all() method extracts all states and all transitions from a Web user interface.

In [163]:
gui_fuzzer.explore_all(gui_runner)

The grammar embeds a finite state automation and is best visualized as such.

In [164]:
fsm_diagram(gui_fuzzer.grammar)
start <start> state <state> start->state state-1 <state-1> state->state-1 click('terms and conditions') state-2 <state-2> state->state-2 fill('city', '<text>') fill('email', '<email>') fill('name', '<text>') check('terms', <boolean>) fill('zip', '<number>') submit('submit') end <end> state->end state-1->state click('order form') state-1->end state-2->state click('order form') state-2->end

The GUI Fuzzer fuzz() method produces sequences of interactions that follow paths through the finite state machine. Since GUICoverageFuzzer is derived from CoverageFuzzer (see the chapter on coverage-based grammar fuzzing), it automatically covers (a) as many transitions between states as well as (b) as many form elements as possible. In our case, the first set of actions explores the transition via the "order form" link; the second set then goes until the "" state.

In [165]:
gui_driver.get(httpd_url)
actions = gui_fuzzer.fuzz()
print(actions)

fill('city', 'U')
fill('email', 'r@z')
fill('name', 'H')
check('terms', True)
fill('zip', '3')
submit('submit')
click('order form')
fill('city', 'q')
fill('email', 'v@p')
fill('name', 's')
check('terms', True)
fill('zip', '4')
submit('submit')

These actions can be fed into the GUI runner, which will execute them on the given GUI driver.

In [166]:
gui_driver.get(httpd_url)
result, outcome = gui_runner.run(actions)
In [167]:
Image(gui_driver.get_screenshot_as_png())
Out[167]:

Further invocations of fuzz() will further cover the model – for instance, exploring the terms and conditions.

Internally, GUIFuzzer and GUICoverageFuzzer use a subclass GUIGrammarMiner which implements the analysis of the GUI and all its states. Subclassing GUIGrammarMiner allows extending the interpretation of GUIs; the GUIFuzzer constructor allows passing a miner via the miner keyword parameter.

A tool like GUICoverageFuzzer will provide "deep" exploration of user interfaces, even filling out forms to explore what is behind them. Keep in mind, though, that GUICoverageFuzzer is experimental: It only supports a subset of HTML form and link features, and does not take JavaScript into account.

In [168]:
# ignore
from ClassDiagram import display_class_hierarchy
from Fuzzer import Fuzzer, Runner
from Grammars import Grammar, Expansion
from GrammarFuzzer import GrammarFuzzer, DerivationTree
In [169]:
# ignore
display_class_hierarchy([GUIFuzzer, GUICoverageFuzzer,
                         GUIRunner, GUIGrammarMiner],
                        public_methods=[
                            Fuzzer.__init__,
                            Fuzzer.fuzz,
                            Fuzzer.run,
                            Fuzzer.runs,
                            Runner.__init__,
                            Runner.run,
                            GUIRunner.__init__,
                            GUIRunner.run,
                            GrammarFuzzer.__init__,
                            GrammarFuzzer.fuzz,
                            GrammarFuzzer.fuzz_tree,
                            GUIFuzzer.__init__,
                            GUIFuzzer.restart,
                            GUIFuzzer.run,
                            GUIGrammarMiner.__init__,
                            GrammarCoverageFuzzer.__init__,
                            GUICoverageFuzzer.__init__,
                            GUICoverageFuzzer.explore_all,
                        ],
                        types={
                            'DerivationTree': DerivationTree,
                            'Expansion': Expansion,
                            'Grammar': Grammar
                        },
                        project='fuzzingbook')
Out[169]:
GUIFuzzer GUIFuzzer __init__() restart() run() fsm_last_state_symbol() fsm_path() set_grammar() update_existing_state() update_new_state() update_state() GrammarFuzzer GrammarFuzzer __init__() fuzz() fuzz_tree() GUIFuzzer->GrammarFuzzer Fuzzer Fuzzer __init__() fuzz() run() runs() GrammarFuzzer->Fuzzer GUICoverageFuzzer GUICoverageFuzzer __init__() explore_all() GUICoverageFuzzer->GUIFuzzer GrammarCoverageFuzzer GrammarCoverageFuzzer GUICoverageFuzzer->GrammarCoverageFuzzer SimpleGrammarCoverageFuzzer SimpleGrammarCoverageFuzzer GrammarCoverageFuzzer->SimpleGrammarCoverageFuzzer TrackingGrammarCoverageFuzzer TrackingGrammarCoverageFuzzer __init__() SimpleGrammarCoverageFuzzer->TrackingGrammarCoverageFuzzer TrackingGrammarCoverageFuzzer->GrammarFuzzer GUIRunner GUIRunner DELAY_AFTER_CHECK DELAY_AFTER_CLICK DELAY_AFTER_FILL DELAY_AFTER_SUBMIT __init__() run() do_check() do_click() do_fill() do_submit() find_element() Runner Runner FAIL PASS UNRESOLVED __init__() run() GUIRunner->Runner GUIGrammarMiner GUIGrammarMiner FINAL_STATE GUI_GRAMMAR START_STATE UNEXPLORED_STATE __init__() follow_link() mine_a_element_actions() mine_button_element_actions() mine_input_element_actions() mine_state_actions() mine_state_grammar() new_state_symbol() Legend Legend •  public_method() •  private_method() •  overloaded_method() Hover over names to see doc

Lessons Learned¶

  • Selenium is a powerful framework for interacting with user interfaces, especially Web-based user interfaces.
  • A finite state model can encode user interface states and transitions.
  • Encoding user interface models into a grammar integrates generating text (for forms) and generating user interactions (for navigating)
  • To systematically explore a user interface, cover all state transitions, which is equivalent to covering all expansion alternatives in the equivalent grammar.

We are done, so we clean up. We shut down our Web server, quit the Web driver (and the associated browser), and finally clean up temporary files left by Selenium.

In [170]:
httpd_process.terminate()
In [171]:
gui_driver.quit()
In [173]:
for temp_file in [ORDERS_DB, "geckodriver.log", "ghostdriver.log"]:
    if os.path.exists(temp_file):
        os.remove(temp_file)

Next Steps¶

From here, you can learn how to

  • fuzz in the large. running a myriad of fuzzers on the same system

Background¶

Automatic testing of graphical user interfaces is a rich field – in research as in practice.

Coverage criteria for GUIs as well as how to achieve them were first discussed in \cite{Memon2001}. Memon also introduced the concept of GUI Ripping \cite{Memon2003} – the process in which the software's GUI is automatically traversed by interacting with all its user interface elements.

The CrawlJax tool \cite{Mesbah2012} uses dynamic state changes in Web user interfaces to identify candidate elements to interact with. As our approach above, it uses the set of interactable user interface elements as a state in a finite-state model.

The Alex framework uses a similar approach to learn automata for web applications. Starting from a set of test inputs, it produces a mixed-mode behavioral model of the application.

Exercises¶

As powerful as our GUI fuzzer is at this point, there are still several possibilities left for further optimization and extension. Here are some ideas to get you started. Enjoy user interface fuzzing!

Exercise 1: Stay in Local State¶

Rather than having each run() start at the very beginning, have the miner start from the current state and explore states reachable from there.

Exercise 2: Going Back¶

Make use of the web driver back() method and go back to an earlier state, from which we could again start exploration. (Note that a "back" functionality may not be available on non-Web user interfaces.)

Exercise 3: Avoiding Bad Form Values¶

Detect that some form values are invalid, such that the miner does not produce them again.

Exercise 4: Saving Form Values¶

Save successful form values, such that the tester does not have to infer them again and again.

Exercise 5: Same Names, Same States¶

When the miner finds a link with a name it has already seen, it is likely to lead to a state already seen, too; therefore, one could give its exploration a lower priority.

Exercise 6: Combinatorial Coverage¶

Extend the grammar miner such that for every boolean value, there is a separate value to be covered.

Exercise 7: Implicit Delays¶

Rather than using explicit (given) delays, use implicit delays and wait for specific elements to appear. these elements could stem from previous explorations of the state.

Exercise 8: Oracles¶

Extend the grammar miner such that it also produces oracles – for instance, checking for the presence of specific UI elements.

Exercise 9: More UI Elements¶

Run the miner on a website of your choice. Find out which other types of user interface elements and actions need to be supported.