Skip to content

behappycc/b2g-monkey

 
 

Repository files navigation

#Ver. 0.1.0

Build Status

b2g-monkey is a python-based crawler for Firefox OS (codename: Boot to Gecko, or b2g) apps. It automatically explores the app under test by clicking candidate clickables with corresponding text input on encountered screens. You can use it to:

  • Explore the interface, clickables and input of the app under test, and see the state transition graph.
  • Find and fill all forms with random or pre-saved input (For now, it only deals with the input fields in forms.)
  • Detect invariants when crawling the app.

Getting Started

(Tested on Windows 10, Firefox 42.0 (w/ Simulator 2.2), b2g-43.0a1.en-US.win32 and Python 2.7.9)

Clone the project. Than install the prerequisite (Admin privilege may be needed):

pip install -e path/to/b2g-monkey/gaiatest-0.33 
pip install -r path/to/b2g-monkey/requirement.txt

There are two ways to activate b2g simulator:

  1. Install Firefox and WebIDE. In WebIDE, install b2g Simulator 2.2, and install your app into the simulator. Keep the App Name and App ID.

  2. Download b2g desktop client (b2g-43.0a1.en-US.win32). Unzip it and run b2g.exe to activate the simulator. Skip the first time settings, and then go to:

  3. Settings -> Display, to turn off Screen Timeout

  4. Settings -> Screen Lock, to disable the screen lock

Install You apps in the b2g desktop client. Keep the App Name and App ID.

In controller.py, replace the values in

config = B2gConfiguration('APP_NAME', 'APP_ID')

with yours. (By default the Contact App would be crawled.)

To start crawling, run

python controller.py

Find report.html, state.html, logs, captured doms and screenshots in

path/to/b2g-monkey/trace/YYYYMMDDHHMMSS/

Customization

Clickable tags

By default four tags are treated as candidate clickables when parsing the dom (refer to dom_analyzer.py):

Tag('a')
Tag('button')
Tag('input', {'type': 'submit'})
Tag('input', {'type': 'button'})

You can get, add or remove tags from this list of DomAnalyzer. For example, to add tag:

DomAnalyzer.add_clickable_tags(Tag('button', {'type': 'reset'}))

Then all buttons like <button type="reset">Click Me</button> will be treated as candidate clickables.

Invariant

Invariants are rules to check every time a new state (screen) is discovered. If any invariant is found violated in a state, the crawler will stop digging and display execution sequences to the invariant in report. By default there is only one invariant: FileNotFoundInvariant (refer to invariant.py). You can add StringInvariant or TagInvariant into configuration like this:

from invariant import TagInvariant, StringInvariant
config.add_invariant(
    TagInvariant('a',
			[{'name': 'class', 'value': 'sister'},
             {'name': 'id', 'value': 'link2'},
             {'name': 'string', 'value': 'Bobby'}])
)
config.add_invariant(
    StringInvariant("display this page because the file cannot be found.")
)

Normalizer

Normalizers are tag removers or string filters applied to DOM document when comparing DOMs. By default there are three normalizers in DomAnalyzer (refer to dom_analyzer.py):

_normalizers = [
        TagNormalizer(['head']),
        AttributeNormalizer('class'),
        TagWithAttributeNormalizer('section', 'class', 'hide')
    ]

You have to manually add or delete normalizers in the list. There are four normalizers (refer to normalizer.py and NormalizerTestCase in test.py)

AttributeNormalizer(attr_list): All attributes not in attr_list will be deleted.

TagContentNormalizer(tag_list): Content in tag will be deleted for tag in tag_list.

TagNormalizer(tag_list): Content in tag and tag itself will be deleted for tag in tag_list.

TagWithAttributeNormalizer: Remove tag with: 1. matched name, attr containing value; 2. matched name, tag content containing value

Running with real device

(Only tested on InFocus New Tab F1)

  1. Make sure you have adb tool in your environment, and can see your device through adb devices
  2. Put the App under test in footer section
executor = B2gExecutor(config.get_app_name(), config.get_app_id(), device=True)

Screenshots

Demo

Report

State

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 80.5%
  • JavaScript 9.3%
  • HTML 6.0%
  • Shell 2.6%
  • CSS 1.6%