Test general review

answering question


Concept: verify whether the software meets user requirements

Difference between testing and development

Commissioning and testing

  1. objective
    Test: check whether the software realizes the functions that should be realized;
    Debugging: the programmer checks whether the software realizes the function that the code should realize.

  2. role
    Developer: Programmer
    Testing: testers and developers

  3. stage
    Development: debug only in the development phase
    Testing: throughout the life cycle of software development


Test: the technical requirements are wide, but the depth is lower. Require to use various tools and understand some languages
Development: specific technical requirements and high depth

Why software testing

  1. interest
  2. skill
  3. Compressive capacity
  4. responsibility
    Why not develop and test
    Learned development knowledge = = > learned software testing = = > interested = = > learned development to test better



Concept: permissions and skills required to meet user needs or formal documents

Testers are involved in testing during the requirements phase

User requirements: what do users want to do, eat braised meat
Software requirements: what do you need to do to make braised meat? Buy meat, seasoning, etc


Concept: if and only if the specification exists and is reasonable, the function and requirement specification of the software do not meet the specification, which indicates that it is a software error;
When the software specification does not exist, the user's demand is reasonable, and the software function is inconsistent with the demand specification, it indicates that the software is wrong.

In other words, no matter whether the specification exists or not, as long as the user's requirements are reasonable, but the software function does not meet the requirements specification, it is a bug

test case

Concept: a set of sets initiated to the test system; Including: test environment, test steps, test data, expected results, title, test function module, priority, importance, etc.

Development model

Waterfall model

Spiral model

Iterative and incremental model

Agile model

Roles: Product Manager (PO), project manager (SM), R & D team (ST)
Product release meeting = = > iteration planning meeting = = > daily station meeting during development = = > demonstration meeting = = > review meeting

Software test model

When did the tester get involved in the test?

  • W model, intervene in demand analysis;
    Disadvantages: serial execution, does not support agile development;
    V model, intervene after coding.
    Disadvantages: unable to intervene in the requirements phase


Software testing process (software testing life cycle)

Requirement analysis = = > test plan = = > test design / development = = > test execution = = > Test Report

How to describe a BUG

Test version (system version) + test environment (browser version, etc.) + test steps + test data + expected results + actual results + attachments (error screenshots and error logs)

BUG level


When a new function of the system goes online and it is found that the user crashes when using a function and cannot be repaired, how can the tester deal with it?

  1. Fallback Version (system stable version)
  2. Pay close attention to the implementation and repair of the problem

BUG life cycle

After the developer fixes a BUG, the BUG is still found during the test. What are the possible reasons

  1. Development inspection, whether to submit new code after repairing the code;
  2. The tester did not re extract the code of the test environment;
  3. It is true that the developer did not modify it successfully.

How to deal with conflicts between testers and developers due to bugs?

  1. Check whether the BUG description is clear and correct;
  2. Convince developers from the perspective of users;
  3. The ranking of bugs is reasonable and justified. Don't exaggerate;
  4. Testers constantly improve their level and can find bugs and their causes, as well as solutions;
  5. Hold a tripartite meeting (product manager, developer and tester) to discuss BUG solutions.

Use case

Design test case basis
Design test cases according to requirements
Analyze requirements, refine requirements, extract function points / test points from requirements, and then design test cases with specific methods of design and testing;

Black box test design test case method
Equivalence class, boundary value, cause and effect diagram, scenario design method, wrong guess method, orthogonal method

Test case design method

Equivalence class

Concept: divide the input (output is considered only in special cases) into several equivalent classes, and select a test case from each equivalent class for testing. If the test case passes, it means that the equivalent class represented by the test case passes the test.

Usage: there are too many test cases to be exhaustive

Valid equivalence class: enter meaningful input for this function (enter legal user name)

Invalid equivalence class: meaningless input for this function (illegal user name input)

Boundary value

Concept: design test cases according to the boundary of input

Cause and effect diagram

Usage: there are multiple conditions for input, and different combinations of inputs correspond to different inputs
Logical relation: and or non identity
Use steps: analyze possible inputs and outputs = = > find out the preference between inputs and outputs = = > draw the cause and effect diagram = = > draw the cause and effect diagram into the judgment table = = > write test cases according to the judgment table

  1. Analyze possible inputs and outputs
    Input: order submission, amount greater than 300, with red envelope
    Output: with or without discount

  2. Find out the preference between input and output

  3. Draw a cause and effect diagram

  4. Draw a decision table: there are 3 inputs and 2 outputs, and the corresponding number of columns 2 ^ 3 = 8

Scene method

Concept: combine isolated function points in the system according to certain logic to form a scenario / business, and test the scenario

First, determine each function point in the scene, and determine the possible input and different output of the function point

error guessing

Concept: the first mock exam is based on the intuition, experience and knowledge of the tester, and then the design of test cases is designed.

Orthogonal method

Concept: study a test case design method of multi factors (different inputs) and multi levels (different inputs and different values). According to the orthogonality, select the optimal input combination for the test, analyze the test results, and use the results to measure the overall test results

How to test TCP and UDP protocols?
The functions and characteristics of TCP and UDP are analyzed respectively

  • TCP
    Function: transmit data
    Features: reliable (whether ACK is received or not), three handshakes (test according to different signals sent each time of the three handshakes), data flow, four waves, etc

  • UDP
    Function: transmit data
    Features: datagram, unreliable, fast.


Divided by development stage

unit testing

Classification: Junit unit test framework (white box), unittest unit test framework (black box)
Basis: detailed design documents

integration testing

Concept: integration test, also known as joint test (joint commissioning) and assembly test, assembles program modules with appropriate integration strategies to test the correctness of system interfaces and integrated functions. The main purpose of integration is to check whether the interface between software units is correct.
Test method: white box test

Smoke testing

The object of smoke test is each newly compiled software version that needs formal test. The purpose is to confirm that the basic function of the software is normal and can carry out subsequent formal test, that is, system test. The executor of the smoke test is the version compiler.

System test

The software system is regarded as the test of a system. It includes testing the function, performance and the software and hardware environment in which the software runs. Most of the time is in the system test execution stage, including regression test and smoke test.
Test method: black box test

regression testing

Concept: check whether the newly introduced code has an impact on the old functions

Acceptance test

Acceptance testing is the last test operation before deploying software. It is the last stage of technical testing, also known as delivery testing. The purpose of acceptance test is to ensure that the software is ready. According to the project contract, assignment and acceptance basis documents agreed by both parties, the software system meets the original requirements

By implementation organization

α test

Non testing, non developer testing

β test

Real users test at any time and place (internal test)

Divided by manual

Manual test

Manual testing is that people enter use cases one by one and then observe the results. Corresponding to machine testing, it is a relatively primitive but necessary one.

Advantages: automation cannot replace exploratory testing and testing of divergent thinking results.
Disadvantages: slow execution efficiency, large quantity and easy to make mistakes.

automated testing

Write automated scripts to test the designed test cases

Automated testing cannot replace manual testing

Run code

Static test

Do not run the code, analyze the code for testing

dynamic test

Run the code, check the similarities and differences between the expected results and the actual results, and analyze the differences, operation efficiency, correctness, etc

Divided by whether to view code

Black box test (unittest)

Regard the software as a black box, do not care about the logical structure implementation of the internal code of the software, only care about the input and output

White box test (Junit)

Test whether the internal logic and structure of the program have realized the corresponding functions

Test methods: statement coverage, logic coverage (decision coverage / condition coverage / decision combination / condition combination / decision and condition combination), path coverage (if / else / try / catch), loop coverage (for / while)

Divided by test object

Business test

Concept: it is the function defined by the tester to run the modules of the system in series, simulate the actual workflow of the real user, and meet the user's needs

Interface test

  1. Layout layout
  2. Font size, thickness, etc
  3. Is the picture clear
  4. Page controls (buttons, scroll bars, checkboxes, etc.)

Fault Tolerance Testing

Concept: when the system has an exception due to external reasons, it can handle it by itself, do not show the exception directly to the user, and give the user a friendly prompt

Usability test

Ease of use is the concentrated embodiment of the adaptability, functionality and effectiveness of interaction. Ease of use belongs to the category of ergonomics. Ergonomics is a discipline that designs everyday things as easy-to-use and practical


Compatibility mainly refers to whether the software can operate well, whether it will have an impact, whether the software and hardware can work efficiently, and whether it will affect the system crash.

Document test

In the actual testing, the most common is the testing of user files, such as manuals, etc. some companies will also test the requirements documents to ensure the quality of the requirements documents.

performance testing

Check whether the system meets the performance specified in the requirements specification.

Safety test

Security testing is a relatively independent field, which needs more professional knowledge. For example, web security testing needs to be familiar with various network protocols
TCP\HTTP, firewall, CDN, familiar with vulnerabilities of various operating systems, familiar with routers, etc. in terms of software, familiar with various attack methods, such as SQL injection, Xss, etc.

Memory leak test

The user allocated memory space and did not recycle it, resulting in more and more memory occupied.

Automated testing (selenium)

webdriver basic usage

How to locate elements

# Locate with id
 Input box id = kw  Send Zhang Jiawen to the input box
driver.find_element_by_id("kw").send_keys("Zhang Jiawen")
# Baidu click

# Locate with name
driver.find_element_by_name("wd").send_keys("Zhang Jiawen")

# class name positioning
driver.find_element_by_class_name("s_ipt").send_keys("Zhang Jiawen")
driver.find_element_by_class_name("bg s_btn").click()

# Link textpositioning
driver.find_element_by_link_text("Post Bar").click()

# partial positioning

# tag name location

# xpath positioning
driver.find_element_by_xpath("//*[@ id ='kw '] ". Send_keys (" Zhang Jiawen ")

# css selector
driver.find_element_by_css_selector("#kw").send_keys(" Zhang Jiawen ")

wait for



similarities and differences
Same: the same as time.wait(10), if neither of them finds the next required element after waiting for 10s, an exception will be thrown
Different: time.sleep() is a fixed wait, while driver. Implicitly_wait() is to wait until it is found, rather than wait

Get browser title and URL

# Get title
# Get URL

Browser window size

# Set browser window
# Browser maximization

Forward and backward

# forward

# back off


# Package required
from selenium.webdriver.common.keys import Keys

# usage


# Package required
from selenium.webdriver.common.action_chains import ActionChains

# Right click
name = driver.find_element_by_id("su")

# double-click

Multiple selection (CheckBox)

for input in inputs:
    if input.get_attribute('type') == 'checkbox':

Drop down box

# Positioning through css selector
driver.find_element_by_css_selector("#ShippingMethod > option:nth-child(5)").click()

# Array subscript



Upload file


Hierarchical framework

From the default frame to f2, you need to go to f1 first;
From f2 to the default frame, you also need to go to f1 first;

unittest framework

test fixture

from selenium import webdriver
import unittest
import time
import os
# Function of unittest.TestCase: the class Unit inherits unittest.TestCase
class Unit(unittest.TestCase):

    def setUp(self):
        print("run setUp")
        self.driver = webdriver.Chrome()
        self.url = "https://www.baidu.com/"

    def tearDown(self):
        print("run tearDown")
if __name__=="__main__":

test suite

from Test import unitDemo
from Test import unitDemo1
import unittest

def createsuite():
    # addTest loads one test case of a test script into the test suite at a time
    suite = unittest.TestSuite()


def createsuite1():

    # Load all test cases in a test script into the suite
    suite = unittest.TestSuite()

    return suite

def createsuite2():
    # TestLoader
    # Load all test cases in a test script into the suite
    suite1 = unittest.TestLoader().loadTestsFromTestCase(unitDemo.Unit)
    suite2 = unittest.TestLoader().loadTestsFromTestCase(unitDemo1.Unit)
    suite = unittest.TestSuite([suite1, suite2])

    return suite

def createsutie3():
    # Add all test scripts named in some form in a folder to the test suite
    discover = unittest.defaultTestLoader.discover("../Test", pattern="unitDemo*.py",top_level_dir=None)

    return discover

if __name__=="__main__":
    suite = createsuite1()
    runner = unittest.TextTestRunner(verbosity=2)

Ignore test case execution

# Note: ignore test case execution
def test_abaidu(self):
	driver = self.driver
	url = "https://www.baidu.com/"
	title = driver.title
	    self.assertEqual("pighead_use Baidu Search", title, msg="Unequal")


Function: judge whether the expected results are consistent with the actual results

# Assert
    def test_abaidu(self):
        driver = self.driver
        url = "https://www.baidu.com/"
        title = driver.title
            self.assertEqual("pighead_use Baidu Search", title, msg="Unequal")


    # Operation sequence 0-9 A-Z a-z
    # screenshot
    def test_Abaidu(self):
        driver = self.driver
        url = self.url
        driver.find_element_by_id("kw").send_keys("It's you")
        title = driver.title

            self.assertEqual("123 It's you_use Baidu Search",title,msg="Unequal")
            self.saveScreenShot(driver,"It's you.png")


Function: output all results to HTML report

import os
import sys
import time
import unittest
import HTMLTestRunner
def createsuite():
    discovers = unittest.defaultTestLoader.discover("../Test", pattern="unitDemo*.py", top_level_dir=None)
    return discovers

if __name__=="__main__":
    #1. Create a folder
    # Current path
    curpath = sys.path[0]
    # sys.path is an array of path collections
    # When the resultreport folder under the current path does not exist, create one
    if not os.path.exists(curpath+'/resultreport'):

    # 2. Solve the problem of duplicate naming
    now = time.strftime("%Y-%m-%d-%H %M %S", time.localtime(time.time()))
    # The file name is the path plus the name of the file
    # Prepare HTML report output file
    filename = curpath + '/resultreport/'+ now + 'resultreport.html'
    # Open the HTML file and wb write it
    with open(filename, 'wb') as fp:
        # The parameters in parentheses are the parameters in the HTML report
        runner = HTMLTestRunner.HTMLTestRunner(stream=fp, title=u"Test report",
                                               description=u"Use case execution", verbosity=2)
        suite = createsuite()

Data driven

Installation kit: pip install ddt

from selenium import webdriver
import unittest
import time
from ddt import ddt, unpack, file_data, data
import sys, csv

def getTxt(file_name):

    # ([cat, cat Baidu search], [dog, dog Baidu search], [pig, pig Baidu search])
    rows = []
    path = sys.path[0]
    with open(path + '/data/' + file_name, 'rt',encoding='utf-8') as f:
        readers = csv.reader(f, delimiter=',', quotechar='|')
        next(readers, None)
        for row in readers:
            # Cat, cat_ Baidu search
            for i in row:
        return rows

def getCsv():
    # Read csv file
    values = []
    with open('data/test_help_json.csv', encoding='utf-8') as f:
        f_scv = csv.reader(f)
        for i in f_scv:
        return values

class Baidu(unittest.TestCase):
    def setUp(self):
        self.driver = webdriver.Chrome()
        self.base_url = "http://www.baidu.com/"

    def tearDown(self):

    # json data
    # Get one at a time
    def test_baidu1(self, value):
        driver = self.driver


    # Use data to put data directly
    def test_baidu2(self, value):
        driver = self.driver
        self.assertEqual(value,driver.title,msg="Non conformity")

    @data(['pig',u'pig_Baidu Encyclopedia'],['pig',u'pig_Baidu Encyclopedia'])
    # unpack is required to use two parameters
    def test_baidu3(self, value, expected_value):
        driver = self.driver
        driver.get(self.base_url + "/")

        self.assertEqual(expected_value, driver.title, msg="Not as expected")

	# Read txt text
    def test_baidu4(self, value, expected_value):
        driver = self.driver
        driver.get(self.base_url + "/")

        self.assertEqual(expected_value, driver.title, msg="Not as expected")
	# Read json file
    def test_baidu5(self,value,expected_value):
        driver = self.driver
if __name__=="__main__":

Tags: unit testing Testing

Posted on Thu, 09 Sep 2021 14:00:29 -0400 by rg_22uk