In this repository you will find a Dockerfile that creates a CodeChecker client that will analyze a branch from LSTS' DUNE repository. After analysis, it attempts to publish a report to a CodeChecker server. Below, I will break down how to use this repository.
The first piece of the puzzle come as a CodeChecker Server. This server will contain a database that will store each analysis report you send it. To access them you simply need to open the correct port of the host server on an html page (http://server_host:port). Thankfully a Dockerfile with a working server already exists on Docker Hub. Simply pull the image:
docker pull codechecker/codechecker-web
And run it:
docker run -d \
-p 8001:8001 \
-v /home/$USER/codechecker_workspace:/workspace \
codechecker/codechecker-web:latest
The -p argument will expose your local port 8001 to the container's local machine's 8001 port and with -v you will attach a local diretory to a directory inside the container. By sharing volumes you will store the report database outside the container. This way your reports won't be lost if the container, for whatever reason, shuts down.
If you're inside LSTS's lab network, there should be a CodeChecker instance running on the Jenkins Server.
For the next piece, the container created by this docker file will install all needed dependencies to build DUNE's source code. Similar to the previous Dockerfile this will also contain a complete CodeChecker installation. Except, in this case we won't be hosting a server, instead, we will be running the actual analyzer and publishing the report to the server.
After the environment is set up, it builds DUNE while generating a .json file that contains all compilation commands that happen during said build.
CodeChecker log --build "make -j8" --output /opt/codechecker/compile_commands.json
The compile_commands.json file is what CodeChecker uses to then analyze the source code.
CodeChecker analyze compile_commands.json --config client_config.json -o output
The output of the above command will be a folder that includes reports on all the files that were analyzed during the build process.
At the end of this Readme there is an also an explanation on how to use the client_config file.
Finally we will upload this report to the server through the following command.
CodeChecker store output --name=run_name --url=http://host_server_ip:port_server/product_name
These arguments are important so pay attention! The way reports are divided and related to each other are through products which include different runs. Let's say you are analyzing different branches from DUNE. Ideally you would divide each branch into different produts. In the case you are analyzing the same branch but, with new changes made, you would upload the reports to the same product but with a different name.
To choose the product name and server location you should edit the --url and to differentiate between runs you should change the --name parameter.
If you made it to the end you should be able to open an html page that looks something like this:
The DUNE project is extensive and includes specific source code coming directly from various' sensor manufacturers that we don't necessarily want to analize. Because of this, it is both usefull to ignore some of the code checkers included and some source files.
To achieve this, you can edit 2 files.
To ignore the checkers you should edit the settings/client_config.json and specify the checker's name in the like so:
"--disable", "cppcheck-uninitMemberVar",
To ignore files, you should specify the path in the settings/skip_text.json. The example will skip all files that include the name "vendor" in its path.
-*/vendor/*