These are a couple of software test tools. At the moment there are just a bunch of scripts to try out various ideas I'm having. Don't expect fancy install scripts or even a python egg.
- Requirements:
- Windows (Sorry!)
- pyHook: http://sourceforge.net/projects/pyhook/postdownload?source=dlp
- docopt: pip install docopt
- pywin32: http://sourceforge.net/projects/pywin32/files/
Similar to http://developer.android.com/tools/help/monkey.html for Android, this script simulates pseudo-random user input on a Windows Application. For example mouse movement, clicks, keyboard input, ...
Test software by sending a huge amount of user input to it. This is a very low-cost test. A configuration can be set up in a couple of minutes. If run long enough it will eventually execute a combination of user inputs that will cause a bug in the program under test (That's the theory anyway).
The kind of user inputs performed can be configured in a configuration file. Each line contains a user input. During runtime the program rnadomly chooses from those lines and performs the corresponding user input.
The basic format of a line is: action:parameter
Example:
mouse_left_up_or_down
type:'Test'
key:'enter'
The following command shows all possible user inputs with their parameters:
python monkey_test.py --show-functions --detail
Some user inputs aks for keys as parameters. A list of all valid keys can be shown with:
python monkey_test.py --show-keys
To start the test, run:
python monkey_test.py --config FILE
A semi-transparent window will appear. Drag and resize the window so that it covers the area to test (user inputs will only be executed within this window) and hit the red start button. By default 1000 user inputs will be simulated.
To increase that value use:
python monkey_test.py --config FILE --actions=2000
Alternatively it is possible to specify how long (in minutes) the test should run:
python monkey_test.py --config FILE --duration=5
The test will usually run until the end. This poses a problem if, for example, the program under test crashes during the test, it will continue simulating user inputs. Whatever was under underneath the program under test will receive those user inputs (possibly the windows desktop). To avoid this problem it is possible to monitor a log file and stop the user inputs if certain strings appear in the log file:
python monkey_test.py --config FILE --check-log C:\test.log ERROR
Regardless, I always recommend to run monke_test.py in a disposable VM. There is just too much that can go wrong if random inputs are send to a PC. Don't blame me if you accidentally format your disk!
Running monkey test again will result in exactly the same user inputs so you can run it again to reproduce issues. the only problem is that it often runs to fast to be able to observe anything. It is possible to slow it down after a certain amount of user inputs. If you simulate 1000 user inputs and you notice a problem after about half of it, it makes sense to slow down after 450 user inputs to be able to see what is happening:
python monkey_test.py --config FILE --slow-after 450
More options are described in the help:
python monkey_test.py --help
capture.py: Captures user input like mouse movement and keys. When pressing a special key (F12) it takes a screenshot.
replay.py: Replays previously captured user input. It also takes screenshots at the same stages as during the capture.
There are two use cases:
1. A tester can run this tool in the background while testing a software. When a bug is found it can easily be easily reproduced.
2. It can act as a quick & dirty regression test for developers. Before making any changes to the code, the developer can click through all functionality in the software and take screenshots regularly. After modifying the code, the developer can replay the previously recorded actions. This will result in two sets of screenshots. One from before the modification and one after. These screenshots can then be compared and checked for (unwanted) differences. A tool like http://www.imagemagick.org/script/compare.php can help with this.
To capture inputs:
python capture.py TEST_FOLDER
While capturing there are two special keys:
- F12: Takes a screenshot and saves it into a folder 'expected' under the specified test folder
- F11: Stops the recording
To replay inputs:
python replay.py TEST_FOLDER
Screenshots will be taken and put in a folder with the current date/time as its name.
The input will always be recorded relative to the current active window. This ensures that the replay will still work even if the window has been moved in the meantime.