Troubleshooting CodeQL code scanning in your CI system

If you're having problems with the CodeQL runner, you can troubleshoot by using these tips.

Code scanning is available in public repositories, and in private repositories owned by organizations with an Advanced Security license. For more information, see "GitHub's products."

In this article

Did this doc help you?

Help us make these docs great!

All GitHub docs are open source. See something that's wrong or unclear? Submit a pull request.

Make a contribution

Or, learn how to contribute.

Note: The CodeQL runner is currently in beta and subject to change.

The init command takes too long

Before the CodeQL runner can build and analyze code, it needs access to the CodeQL bundle, which contains the CodeQL CLI and the CodeQL libraries.

When you use the CodeQL runner for the first time on your machine, the init command downloads the CodeQL bundle to your machine. This download can take a few minutes. The CodeQL bundle is cached between runs, so if you use the CodeQL runner again on the same machine, it won't download the CodeQL bundle again.

To avoid this automatic download, you can manually download the CodeQL bundle to your machine and specify the path using the --codeql-path flag of the init command.

No code found during the build

If the analyze command for the CodeQL runner fails with an error No source code was seen during the build, this indicates that CodeQL was unable to monitor your code. Several reasons can explain such a failure.

  1. Automatic language detection identified a supported language, but there is no analyzable code of that language in the repository. A typical example is when our language detection service finds a file associated with a particular programming language like a .h, or .gyp file, but no corresponding executable code is present in the repository. To solve the problem, you can manually define the languages you want to analyze by using the --languages flag of the init command. For more information, see "Configuring code scanning in your CI system."

  2. You're analyzing a compiled language without using the autobuild command and you run the build steps yourself after the init step. For the build to work, you must set up the environment such that the CodeQL runner can monitor the code. The init command generates instructions for how to export the required environment variables, so you can copy and run the script after you've run the init command.

    • On macOS and Linux:
      $ . codeql-runner/codeql-env.sh
    • On Windows, using the Command shell (cmd) or a batch file (.bat):
      > call codeql-runner\codeql-env.bat
    • On Windows, using PowerShell:
      > cat codeql-runner\codeql-env.sh | Invoke-Expression

    The environment variables are also stored in the file codeql-runner/codeql-env.json. This file contains a single JSON object which maps environment variable keys to values. If you can't run the script generated by the init command, then you can use the data in JSON format instead.

    Note: If you used the --temp-dir flag of the init command to specify a custom directory for temporary files, the path to the codeql-env files might be different.

  3. The code is built in a container or on a separate machine. If you use a containerized build or if you outsource the build to another machine, make sure to run the CodeQL runner in the container or on the machine where your build task takes place. For more information, see "Running CodeQL code scanning in a container."

Did this doc help you?

Help us make these docs great!

All GitHub docs are open source. See something that's wrong or unclear? Submit a pull request.

Make a contribution

Or, learn how to contribute.