Code Coverage for dotnet Csharp projects in command-line mode
Integrate RKTracer with C# projects with dotnet build and generate code coverage
The application should be built and testable
You should be able to build and test the application using the dotnet build command in the command line mode. To generate code coverage for the dotnet based application, you need to prefix rktracer to dotnet build command.
dotnet build prime.csproj
rktracer dotnet build prime.csproj
RKTracer tool makes internal adjustments to dotnet build to instrument all the source files compiled. By default, rktracer instruments for multiple-condition code coverage, and you see the following code coverage information in the HTML report.
- Function Coverage
- Line Coverage
- Statement Coverage
- Multi-Condition Code Coverage
Example dotnet build:
Step 1: Enable RKTracer Tool
Before you enable the RKTracer tool for instrumentation, you need to clean the project, as the RKTracer tool will instrument the project at the source code level.
dotnet build prime.csproj
rktracer dotnet build prime.csproj
- RKTracer instruments C# source file with “c:\rktracer/lib/librkliz-x86_64-w64-mingw32.exe ” and writes out the instrumented file to the “prime.cs” file in a preprocessed directory. Also, it stores the metadata, structure of instrumentation, and copy of source file content in JSON file. The tool will refer rktracer.config for internal configuration at the time of instrumentation.
- RKTracer automatically adds runtime library to dotnet build.
Step 2: Test instrumented application.
You can do unit testing/functional testing. In this case, it’s functional testing.
Post-testing the RKTracer tool will save coverage data to the rk-coverage.txt file. In case if your testing instrumented application in a server machine you need to copy the rk-coverage.txt file from the server to the application working folder to build the machine. You can skip and go to the fourth step to generate reports if you see the rk-coverage.txt file.
Coverage data file not found after testing.
RKTracer tool uses the following approaches to save coverage data.
- Run unit/functional testing, and at the end RKTracer tool will automatically save coverage data on the host/embedded target/server machine, that is, write coverage data to the rk-coverage.txt file.
- Send a signal to the application running in an infinite loop to trigger and save coverage data.
- The application under test has crashed in the middle of testing on the host or target device. RKTracer saves coverage data to rk-coverage.txt and generate reports to check the point of a crash in reports.
- Embedded devices without a file system or not connected to host while testing. In this case, rktracer RAW mode comes in to picture. You need to save the coverage data from the RKTracer variable.
Suppose you cannot find the rk-coverage.txt file in the project working folder or target server machine. An application under test has not exited within the proper shutdown hook, i.e., Runtime. get runtime().addShutdownHook(new Thread()), or your application might be running in an infinite loop or live application
Save coverage data when a particular function is executed at runtime testing.
Identify a function/method name in your application (ensure this function/method is not excluded from RKTracer instrumentation or file with this function). Then copy the function name and edit the rktracer.config in the RKTracer installation folder.
For C# programming language: Add a new line at the end of rktracer.config file with “c#-function-writeout” and function name. Here the function name is case sensitive, and you should also add empty braces, i.e., () at the end of a function name. In case if you have a function with an input parameter, it should still be empty braces. You can add multiple function names with one space to the list, as shown below.
- c#-function-writeout main()
- c#-function-writeout main() OnCreate() fun()
After adding function name to rktracer.config file. You need to rebuild the application to re-instrument using rktracer. Coverage data will be saved to the rk-coverage.txt file when this function executes at runtime testing.
Send a signal to save coverage data.
We can send a signal to the application running in an infinite loop to save coverage data to rk-coverage.txt.
- Windows: send a kill signal using keys in the keyboard, i.e., Control+Pause/break. Contact support if you would like to add additional signal command to save coverage data at runtime testing.
- Linux/Unix: Send a SIGWINCH signal: killall -SIGWINCH a.out or kill -SIGWINCH Process ID number.
Step 3: Generate Code Coverage HTML reports.
Once you have the rk-coverage.txt file, you need to run the command “rkresults” to generate HTML reports. Ensure that you run the “rkresults” command in the project root folder or where we have the “rktracer” folder.
rkresults command will search coverage data file “rk-coverage.txt” and map with JSON files (generated during instrumentation) in the rktracer folder and generate html reports. You can manually open the HTML report using index.html.
Generate code coverage for selected functions/files/folder
Generate coverage reports for the selected folders
Suppose you need code coverage for source files from three different folders, i.e., core, keys, power, and ignore coverage for folders firmware_loader and lockdown. Edit rktracer.config in the RKTracer installation folder and go to the end of the file add the following information.
instrument */power/* */core/* */keys/*
never */firmware_loader/* */lockdown/*
ignore *.cs = Ignores all C# source files from instrumentation
Instrument = Instrument source files from given folders
never = ignore selected folder
Generate coverage reports only for selected functions from three different files.
fun_X() in source-file-X
fun_Y() in source-file-Y
fun_Z() in source-file-Z
Suppose you need code coverage for selected functions from three different files. Edit rktracer.config in the RKTracer installation folder and go to the end of the file and set the following variables as shown below.
instrument *source-file-X.cs *source-file-Y.cs *source-file-Z.cs
function-instrument fun_X() fun_Y() fun_Z()
ignore *.cs = Ignore all C# programming source files
instrument *file-X.cs *file-Y.cs *file-Z.cs = Instrument only these three source files.
function-ignore * = Then ignore all functions in the above three files.
function-instrument fun_X() fun_Y() fun_Z() = Don’t ignore these three functions from these three files from instrumentation.