Using Sigasi CLI in Jenkins CI

Recently, Sigasi launched its command line tool, Sigasi CLI . With Sigasi CLI, you can run the same powerful code validation as in Sigasi Visual HDL (SVH) from the command line. This means that code validation is now possible without a GUI.

In this article, we demonstrate how you can run Sigasi CLI in a Continuous Integration (CI)  environment, namely Jenkins . In a CI setup, code validation with CLI becomes part of regression testing and will help you maintain the quality of your HDL code. In the near future, we’re planning more articles on CLI and continuous integration, so be sure to check back.

Code validation in Sigasi has a couple of different aspects. On one hand, syntax checking verifies that your code is syntactically correct. Syntax errors are always errors and are flagged as such. On the other hand, we have code linting, where Sigasi checks for code that is syntactically correct but is likely incorrect nonetheless. Examples are a VHDL process with an incomplete sensitivity list, or a case statement which doesn’t cover all cases. A third, optional aspect of code validation is code style checking, which checks, among other things, whether names in your design comply with your team’s naming conventions. For example, some teams want names of constants as all caps or signal names with an s_ prefix. Linting and code style problems may be flagged as errors, warnings, or info, or they may be ignored depending on your settings.

For the purposes of this article, we’ve set up a Gitlab project , taken from the Sigasi Studio 4.x VHDL built-in tutorial. As the bare tutorial project is incomplete, it contains a number of problems which Sigasi CLI will flag. Feel free to make your own copy of the project and explore further. Sigasi customers, as always, are welcome to contact support  with further questions.

The remainder of this article is organized as follows: First, we briefly discuss the setup of Jenkins itself. Then we show two different ways of setting up the Jenkins project, either as a simple freestyle project, or as a more advanced pipeline project. We also demonstrate how to present the problems found by Sigasi CLI using Jenkins’ warnings-ng  plugin.

Jenkins setup

For our demonstration of how to use Sigasi CLI in Jenkins, we have used an almost-out-of-the-box Jenkins setup. We’re using Rocky Linux 8  as an OS. You can use this procedure , starting from step 2, to install Jenkins on it. During the initial Jenkins setup, we installed the recommended plugins, and we’ve kept the default Jenkins executors on the main Jenkins node. In addition to that, we’ve installed the warnings-ng  plugin to present CLI’s results in Jenkins.

Jenkins freestyle project

The simple approach to running Sigasi CLI in Jenkins uses a freestyle project. It only takes a few simple steps:

  1. On the Jenkins dashboard, click New Item, enter a name for the new project, select Freestyle project, and click OK.
  2. On the next screen (Configuration), go to the Source Code Management section, select Git and enter https://gitlab.com/sigasi/public/sigasi-cli-ci.git for Repository URL. Set the Branch Specifier to */main .
  3. Continue to Build Steps, add an Execute shell build step and enter these two lines as the command. Replace the license and the install path with the correct ones for your setup.
export SIGASI_LM_LICENSE_FILE=27000@my_license_server
/sigasi_cli_install_path/sigasi-cli verify . --fail-on-error
  1. Click Save

Now your project is ready to run. Click Build Now to have Jenkins run Sigasi CLI on your project. The build will fail because the design contains some errors. Click on the build (under Build History) and then Console Output to inspect the results.

In this case, we have configured the job to fail if Sigasi CLI reports errors in the HDL code by using --fail-on-error. Depending on your project’s needs, you can change that using a different command line argument to Sigasi CLI:

  • --fail-on-error makes the Jenkins job fail if errors are found in the HDL code
  • --fail-on-warning makes the Jenkins job fail if errors or warnings are found
  • --fail-on-error makes the Jenkins job fail if errors, warnings, or infos are reported
  • without any of the above command line options, the job will fail only if Sigasi CLI was unable to verify the HDL code, e.g. if no license was available.

Reporting Sigasi CLI’s result with warnings-ng

Checking Sigasi CLI’s results in the job’s console log is not particularly convenient. The good news is that Jenkins has some tricks to make our life easier. The Warnings-NG  plugin helps us present CLI’s result nicely. Once the plugin is installed in Jenkins, our Jenkins project needs some small changes to use it:

  1. Add command line options so CLI produces an XML file for warnings-ng. The shell command now looks as follows:
export SIGASI_LM_LICENSE_FILE=27000@my_license_server
/sigsi_cli_install_path/sigasi-cli verify . --fail-on-error --warnings-ng -o sigasi-warnings.xml
  1. Configure warnings-ng to pick up Sigasi CLI’s results. Add a Post-build Action: Record compiler warnings and static analysis results, and configure as follows:
  • Tool: Native Analysis Model Format
  • Fileset ‘includes’: **/sigasi-warnings.xml
  • Custom ID: sigasi
  • Custom Name: Sigasi CLI
  • Open the advanced options and check Enable recording for failed builds

Now run the Jenkins job again. When the job finishes, the build result view in Jenkins contains a warnings-ng section with the number of problems found.

Warnings-NG in job overview

Click on the number of warnings to see more details, e.g. an overview of problems per file…

Warnings-NG file details

… or details of each problem Sigasi CLI found in the code.

Warnings-NG error in code

Jenkins pipeline project

In this example, we demonstrate how to run Sigasi CLI as a pipeline step of a Jenkins declarative pipeline. A file named Jenkinsfile in the GitLab project repository defines the pipeline steps. Before running the pipeline, you’ll need to set the Sigasi CLI PATH and license configuration in Jenkinsfile. As you can’t do that directly in our GitLab project, you’ll need to make your own fork if you want to try this.

// Set the absolute path of Sigasi CLI
String sigasiCLI = '/path/to/sigasi-cli'

pipeline {
    agent 'any'

    environment {
        // Configure Sigasi CLI license.
        SIGASI_LM_LICENSE_FILE = '27000@my_license_server'
    }

In this case, we only define one stage, in which we run Sigasi CLI code verification. This stage could become a part of your CI pipeline, along with regression tests, automatic synthesis, and so on. As in the previous example, we’ll have Sigasi CLI write its results in an XML file, which the warnings-ng plugin will use to present them. Depending on how you want to define a successful or failed Jenkins job, you can add --fail-on-... flags to Sigasi CLI as discussed earlier.

    stages {
        stage('Sigasi CLI Verify') {
            steps {
                sh """
                    ${sigasiCLI} verify --warnings-ng -o sigasi-verify.xml .
                """
                // Add `--fail-on-error` , `--fail-on-warning` or `--fail-on-info`
                // on the CLI command line above as required
            }
        }
    }
    post {
        // Report Sigasi CLI errors and warnings to Jenkins' Warnings-NG plugin
        always {
            archiveArtifacts artifacts: 'sigasi-verify.xml'
            recordIssues(
                enabledForFailure: true,
                aggregatingResults: true,
                tool: issues(pattern: 'sigasi-verify.xml', analysisModelId: 'sigasi')
            )
        }
    }

Error/warning reporting with Warnings-NG is defined as a post-processing step in the pipeline. As with the Freestyle job example, the warnings-ng plugin picks up the XML from Sigasi CLI and presents the result in a user-friendly way.

Once the Jenkinsfile is ready and pushed to the repository, it’s time to define the Jenkins pipeline project.

  1. On the Jenkins dashboard, click New Item, enter a name for the new project, select Pipeline project and click OK.
  2. On the next screen (Configuration), go to the Pipeline section. Set Definition to Pipeline script from SCM, then select Git as SCM. Enter your repository’s URL for Repository URL. Set the Branch Specifier and Credentials as appropriate. Leave the Script Path as its default value Jenkinsfile.
  3. Click Save

Now you can run the project, sit back and relax while the job runs, and then check the results.

Conclusion

In this article, we demonstrated two different ways to run Sigasi CLI verification in Jenkins, either in a freestyle project or in a pipeline project. In both cases, Jenkins’ warnings-ng plugin was used to present CLI’s findings in a user-friendly way.

See also

comments powered by Disqus