Local exploratory test method:
1, user input
What is a user input: a stimulus generated by the environment, leading to the stimulation has been tested in response to the application. Input atoms and atom abstraction input into input, such as clicking a button and the like, string or integer atom is input, the input is abstract interrelated combined into an input.
(1) legally imported illegally enter
The reason / purpose:
1, the forward and reverse test test test basic things to be covered.
2, most developers do not want to write error-handling code, which is easier to enter illegally will lead to error or improper handling.
Mode: The idea of developers to construct an illegal input value. Because developers need to know what is legal input, what is illegal input, the input value for illegal do they think of what kind of error handling, this part is what we need to test. Such as: Enter the outside of the null test requirements specified input, special characters, such as characters long.
(2) input filters
Cause / Objective: an input filter for preventing an illegal function code input value is passed to application software, the input filter is silently exclude illegal input, the application software can be transmitted to the data will be processed as normal data. Testers need to test input filter function is correctly implemented, can bypass the filter input values passed to the application software.
: 1, the test: The input panel Integers only, the input filter input only realize the display integer values, the remaining input will be ignored. Drop-down list, check box is also an input filter, provided only legitimate input.
2, enter a valid value, the filter display correctly. Enter an illegal value, the filter box properly shielded. Ensuring the development of a set input filter range is correct.
3, some way to bypass the filter input. The html code to modify the front end, using a tool or the like to skip verification js.
(3) Check Input
Cause / Objective: to accept an input value, if the input value is legal, then run and then deal with it, otherwise generate an error message and aborts the process. And will alarm, display an error message describing the current situation.
: 1, firmly grasp the error message, read the error message, check that wrong. Error message generally indicates the current input value is identified as the root causes of illegal values and how to modify it to become a legal value.
2. The error message, you can analyze those input values can trigger other error messages, or those input values should not cause an error and the software error.
(4) exception handling code
The reason / purpose: to specify any software errors during operation, exception handling code can handle a wide variety of failure conditions, and not just confined to illegal entry.
Method: 1, error messages for abnormal (usually this information is not there any more specific error contents, pointing to a large range), continues to repeatedly test the same function, continue to use the data or just an exception is thrown slightly change it, will not see continue to cause errors
2, try to run elsewhere, call the function test
3, continuous anomalies, to see whether the program will crash
4, to check whether the log information corresponding to the recording error
(5) General input or unconventional input
The reason / purpose: regular input is entered in the program, the user input will be frequently used. Unconventional enter under special circumstances will occur, or will occur simply by chance. Such as: do not accidentally press Ctrl + c when the user enters a capital C, but actually it should be Shift + c.
: 1, and all the Ctrl, Alt, Esc key combination input
2, special fonts may be used by the end user
3, language testing different countries and special symbols
4, some special characters, characters associated with the operation of the platform software, its value as an input.
(6) Input default or user-supplied input
The reason / purpose: usually have some form selected default value, or the value specified is null. If a field is pre-set default values, and then the field does not contain a default value in the treatment program is often very different.
Method: 1, the blank field without entering a character or a NULL parameter passed to an API. Under normal circumstances they need to continue to run down in the default state.
2, detecting whether the user set the default values where values most likely to use.
3, delete the default value, leave the field blank, submit test.
4, test defaults annex some other value. Such as: a default value, a value larger than the default setting or a small string, to modify the default field in the header, tail few characters, plus a few characters, characters and the like to delete a few operations.
(7) using the output to the input selection guide
The reason / purpose: to define what kind of reaction they want the software generates, and then determine those inputs will lead to a corresponding output, and then enter a value corresponding test. The output roughly divided into legal and illegal export output.
: 1, from the results output to think about the whole issue.
2, the first really want what the program produces output, the user visits the scene, to see how to generate the desired output.
3, first look at the output, and then select a new input, and to ensure that the new output is the result of recalculation (with or different from the original)
4, not initialized and initialized generate input whether different.
5, looking saved up output (page display, file storage, etc.), these values change or alter their function (size, type), to test whether rebuilt on the original value. (Such tests for each function needs to run)
2, state
What is the state: the state of the software is a point in state space, it is a point of the value of all the internal data structure uniquely determined. Software state space is constituted by a cross-product of all internal software variables. Application interacts with its operating environment software and receive input causes a change in state. Software receiving the input, and these input values in the internal memory, the software will change the status.
How to test the software status:
(1), test software is correctly updated its own current state
Is abnormal (2), the current state of the test software to receive certain input
(3) whether the software should not have entered the state.
(4), the tester should be based on their own time and on budget, with consideration of how end-users can take risks, to determine the test as many states.
method:
(1), using the state information to help find enter: Off observed after the state have any effect on the results, and get relevant input status information, you can determine those combinations need to be tested in some way. We need to ensure that important changes in the situation and have been tested.
(2) using the state information to identify key input sequence: When an input state changes cause, then several times using the same series of input leads to a state change, a series of observed changes will not cause an overflow state. Such as: a list of stored data is too large, such as shopping cart, add a shopping cart items are not too much of. By observing the extent of accumulation of these states, the same content is repeated to see whether side effects. (Such as continuous deleted, continuous click?)
3, code path
Code path refers to a series of code statements, in fact, in the statement software starts running, terminates in a specific statement.
Testers need to have clear procedures in which the branch, especially without touching code or use code coverage tools. Some branches have not been tested to the very existence of a defect is possible. In addition to branch, loop, etc. can also lead to the total number of code path is large, that is, there will be possibilities.
4, user data:
When handling a huge amount of data or software file, testers need environment simulation test corresponding environment. How simulation data corresponding to the data and ensure real and effective, testers need to address the problem.
the way:
(1) Try using real data
(2) When you remove or add data test, pay attention to restore data
(3) Note that the information processing sensitive data.
5, operating environment:
Operating environment itself is an input software, when running the software environment used are changes that may be effective.
Method: operating system, browser, configuration files, network status, other applications and other software-related