Skills of writing test cases efficiently

This topic will not discuss whether it is necessary to write detailed test cases , On the premise that it is determined to deliver detailed test cases , Share how to write test cases more efficiently .

Align test case requirements

First 、 Specify the test case document objectives and requirements to be completed , Templates 、 Range 、 Particle size, etc .

• Use case document users ： Testers
• Scope of use case document ： Cover all requirements of the product
• Use case template content ： Number 、 modular 、 Sub module 、 Test function point 、 Preset conditions 、 data 、 step 、 Expected results 、 priority 、 Use case types 、 Linked requirements 、( Prepared by 、 Update time 、 Executor 、 state 、 execution time 、 Execution results )
• Test case granularity ： Positive and negative use cases of all functions
• Person in charge of test case acceptance ： See you for a long time ( Align target )

Quickly understand the product

Get familiar with the product business background and technical architecture as soon as possible , So as to outline the overall framework of test cases . Any product , Can eventually be mapped to 【 Horizontal scaling 】+【 Vertical stratification 】 Complete the use case coverage under the combination mode of .

Horizontal business expansion

It refers to the auxiliary level of the product , What are the total business scenarios , What capabilities are provided , That is, the functional panorama of the top layer of the product .

Vertical architecture layering

It refers to the analysis from the technical architecture level of the product , The current products can be divided into several layers , So that the use case verification is to verify and cover the use case from different levels .

Take the big data cloud platform of a cloud as an example , The core of big data cloud platform is cluster . The big data cloud platform cluster is composed of one or more virtual machine instances Hadoop、Flink、ZooKeeper colony . With Hadoop For example , There are usually some running on each virtual machine instance daemon process ( for example ,NameNode、DataNode、ResouceManager and NodeManager), Various big data service components can also be installed on the cluster ( for example ：HBase、Hive、Presto、Spark etc. ).

Panoramic circuit diagram of horizontal core business functions of big data cloud platform ( With Hadoop Clusters, for example ), Its core processes are ：Hadoop Create cluster -> Cluster management -> Big data component management -> Virtual host management -> ... ->Hadoop Cluster release ; The functional panorama is shown in the figure 1 Shown ：

Panorama of big data cloud platform functions

The vertical core architecture of big data cloud platform is simplified into the following four layers , Pictured 2：

• At the top ： Portal console interface of big data cloud platform 【UI】
• Second floor ： The back-end platform of cloud big data API【OpenApi】
• The second floor ： The server of big data cloud platform 【 Big data service components 】
• At the bottom ： Infrastructure of big data cloud platform 【 Cloud server 】

Big data cloud platform architecture diagram

Make a quick plan

Use case coverage

Starting from the panorama of product business functions , around PRD(Product Requirement Document)、 Combined with the vertical architecture level , The use case has no dead corner and fully covers the product ( On scope ).

(1) Widen horizontally 【 Width 】, The main life cycle of its products ranges from large modules to small modules 、 The main function to the secondary function is gradually expanded , Borrow the fish bone diagram to comb ( Here's the picture 3) or Xmind Brain map to sort out . First sort out the inside , Then sort out the service or product scenarios of external docking ( Such as ： Message center 、 Cost center 、 Alarm Center 、 Document center 、 Data development and so on ).

(2) After the lateral expansion divergence is completed , Start longitudinal excavation 【 depth 】, such as , The core architecture of big data cloud platform is divided into four layers , Each layer needs to be taken apart ：

• At the top ：UI Layer end-to-end use case walk through ( As described above ), From the top floor UI Operational testing is in addition to inspection UI result 、 Also ensure that the actual results on the underlying cluster server are consistent with the interface display
• Second floor ： The second layer is the portal backend Api, Call directly OpenApi The relevant test cases cover
• The second floor ： Direct operation, use or strong intervention Hadoop Cluster service component 、 Test the quality of the whole big data cloud platform ; Because there are many service components on the big data platform ( There are more than thirty ), In addition to the use of a single service , More importantly, multiple common service components should be combined for verification
• At the bottom ： Direct operation, use or strong intervention in the server layer ( increase 、 Delete 、 stop 、 restart 、 expand 、 shrink 、 l 、 The Internet 、 disk 、 Software configuration, etc ), Test the quality of the whole big data cloud platform

So far , The whole big data cloud platform Hadoop All test cases of the cluster have been sorted out .

Use case design method

Starting from the type of test , Functional and non functional test cases cover . There is no need to deliver non functional use cases this time , So don't expand ; Functional use case design method ：

• Division of equivalence class ( Positive equivalence class 、 Negative equivalence class )
• Boundary value analysis ( Within the boundary 、 Outside the boundary )
• Decision table analysis
• Cause and effect diagram
• Wrong guess

Use case writing principles

• Resolution principle ： The full text establishes a unified boundary . such as ： Take the module as the boundary 、 When there is correlation and interaction between different modules 、 Preset conditions as boundaries , The contents of the preset conditions are put in the upstream module for verification .
• Priority principle ：【 establish 】【 see 】【 Use ( Start stop, etc )】【 modify 】【 Delete 】 In order 【 Main scene 】 first 、【 Secondary scenario 】 secondly 【 Example 】 first 、【 Counter example 】 secondly
• Basic principles ： The use case is not repeated 、 There is no omission , Principle of unicity 、 That is, a use case covers only one scenario and clear steps 、 There is no ambiguity in the clear expected results Repeated execution results in the same

Write tips quickly

Set uniform standards

Take a cloud big data cloud platform product as an example , A lot of functions are required in a unified way , To this end, design a set of standardized use cases ：

• such as ： Create a new page , Form entry , Unified requirements ( If required 、 Length limit 、 Character requirements ), Design a set of standardized use cases , For reuse by other pages .
• such as ： Authority test cases for each module , Design unified standard use cases ;
• such as ： be-all OpenApi test , All for return codes 200、400、401、403、405、500 The scene test ;
• such as ： Big data platform services 30 Multiple , Each service is different , But the operation is similar ： add to 、 start-up 、 stop it 、 Modify the configuration 、 Deploy , Design unified standard use cases for this purpose ( Do you have a visual sense of code refactoring at the moment , Define a standard method 、 For you to call repeatedly ).

Extract common components

Take a cloud big data cloud platform product as an example , Which includes 10 More than one list page , For each list, there is a paging component 、 Screening 、 Search for 、 Sort , The use cases of these common components are drawn as 【 Common component use cases 】, Design a set of standardized use cases , Relevant pages can be reused .

Be careful ： In the unified standard use case , Variable items use {ABC} To replace , such as ： When filtering cluster status in the cluster view list , Unify... In standard use cases {ABC} Replace with { State of the cluster } that will do .

Batch writing and automatic generation

In the case writing process , Found a lot of things except { A name or field } Different , Everything else is the same , At this time, you can write in batch ( Such as ： With the help of Sublime Or directly transfer variables and generate them with code ), This can also greatly improve the writing efficiency .

Writing OpenApi Related test cases , Directly define a set of OpenApi Standard use cases , With QA The designed standard use case is the template , Then write code to generate use cases , By reading the OpenApi Of Json file , Quickly generate 71 individual Api Test cases for , near 1000 Detailed test cases , Efficient .

Replace with full text

When writing use cases ,QA Personnel must use unified language or format , First, it is convenient for readers 、 Second, it is convenient to find and replace , That is, through full-text search and replacement, you can Quick maintenance use case .

There was a demand change ： From the original first level menu A001 Lower secondary menu B002, Change to level one C001 Next D002; Because in the use case of the whole product , Enter the secondary menu from the primary menu , All use ：A001->B002 This format , This demand change , Direct full-text search and replacement with one click .

As mentioned earlier, several sets of unified standard use cases have been designed , When new pages are reused , Replace variable contents directly , Generate the current use case . Or what the requirements change is just the content of the unified standard use case , Find and replace with full text 、 Use case maintenance in one minute .

All in all , We must sum up a set of our own methods to deal with such a huge amount of writing workload , Otherwise, it cannot be completed in a short time . And the second trick of writing use cases efficiently , It is inseparable from reusability 、 Look for commonalities 、 Refine unified standards , Borrow some means or tools to automatically generate .

Send you a word at the end ： Draw the boundaries of the field 、 High reuse 、 Low coupling .