Golang Code Coverage

This is a creation in Article, where the information may have evolved or changed.

When you write automation, you have to count code coverage = =

Original link: https://www.elastic.co/blog/code-coverage-for-your-golang-system-tests

In fact, I saw the original link I did not succeed in statistics, or to see the notes of colleagues to succeed.

Action steps--written in the steps I've actually done

1 Creating the test file for the main function

2 compiling the code, generating the executable file

3 Statistical coverage

A test file for the main function

Package Main//This file is mandatory as otherwise the Filebeat.test binary was not generated correctly.Import ("Testing"  "Flag")varSystemtest *BOOLfunc init () {systemtest= Flag. Bool ("systemtest",false,"Set to True when running system tests")}//Test started when the test binary is started. Only calls Main.Func Testsystem (T *testing. T) {if*systemtest {main ()}}

Attention:

1 Name: Take a look at the main () function of your code the go file name, directly named *_test.go file; For example, the code file name Main_server.go, can be directly named Main_server_test.go

2 contents: Placed in the same directory as the file under test @1

Two compiled code

1 View the main function, which will be the OS. Exit () Change to return (according to the actual situation Return/return 0/return 255 ...)

2 Original Compile command for GO build

Here you use the command go test-c-covermode=count-ldflags "-X main._version_= $VERSION. ${reversion}"-coverpkg./gopath/src/mvdsp/ module/,./gopath/src/mvdsp/extractor/,./gopath/src/mvdsp/mvutil/,./gopath/src/mvdsp/protocol/-O dsp_server.test

①-c indicates that a test binary is generated

②-covermode=count indicates that the generated binary contains coverage count information

③-ldflags is used to write version information to a binary file without the use of an additional versions file

④-coverpkg behind is to statistical coverage of the file source code

⑤-o is followed by the output binary file name

⑥ ... There may be more parameters available, and I don't know.

3 Execute command, generate an executable binary file, copy to the deployment directory

4 Start the service, add parameters after starting the command:-systemtest-test.coverprofile Coverage/coverage.cov

①-systemtest is used to start the main test that I said earlier

②-test.coverprofile used to specify which file the coverage information is written to

Three statistical coverage

1 Performing automation

2 Execute the following command to generate the coverage file Coverage.cov

Pid_server= ' Ps-ef | grep "My_server-systemtest" | Grep-v "grep" | awk ' {print $} ' kill $pid _server

The condition of coverage is that the program needs to exit/end normally, so when the automation is complete, we need to send a message to the program to indicate the end before we can get the coverage file.

For this: Go is required to return, that is, you can kill Server_pid (if the service starts with supervisor, it will pull itself up without affecting subsequent calls)

Do I need to exit C + +? , send P __gcov_flush (), see http://www.cnblogs.com/zhaoxd07/p/5608177.html

Python needs to trigger the Atexit module to register a callback function that can be used with Ctrl + C or Kill-2

3 Generating test reports

Go tool Cover-html=./coverage/coverage.cov-o/data/reports/coverage.html./coverage/gocover-cobertura < Coverage.cov >/data/reports/coverage.xml

You need to install a widget when generating an XML report: Go get Github.com/t-yuki/gocover-cobertura

Use the Gocover-cobertura under the bin in a directory that is easy to execute

4 integrating test reports into Jenkins, you are done.