Possibility to use scm variable in multibranch pipeline

[testuser7]

It will be nice if we can use scm variable in multibranch pipeline jenkinsfile.

[tomasbjerre]

Do you have an example of where it does not work?

[testuser7]

For now I need to manually specify projectKey and repoSlug:

                if (env.CHANGE_ID) {
                    step([$class: 'ViolationsToBitbucketServerRecorder',
                          config: [useUsernamePasswordCredentials        : true,
                                   projectKey                            : 'prokey',
                                   repoSlug                              : 'repository-name',
                                   pullRequestId                         : env.CHANGE_ID,
                                   commentOnlyChangedContent             : false,
                                   commentOnlyChangedContentContext      : 0,
                                   createCommentWithAllSingleFileComments: false,
                                   createSingleFileComments              : true,
                                   violationConfigs                      : [[pattern : '.*/checkstyle-result\\.xml$',
                                                                             reporter: 'CHECKSTYLE']]]])
                }

Simplified configuration example:

                if (env.CHANGE_ID) {
                    step([$class: 'ViolationsToBitbucketServerRecorder',
                          config: [repositoryConfig                      : scm,
                                   pullRequestId                         : env.CHANGE_ID,
                                   commentOnlyChangedContent             : false,
                                   commentOnlyChangedContentContext      : 0,
                                   createCommentWithAllSingleFileComments: false,
                                   createSingleFileComments              : true,
                                   violationConfigs                      : [[pattern : '.*/checkstyle-result\\.xml$',
                                                                             reporter: 'CHECKSTYLE']]]])
                }

[tomasbjerre]

Are you sure scm has these values? And if it does, cant you do something like this?

projectKey                            : env.repositoryConfig.projectKey,

[testuser7]

Yes, I can simply use: checkout scm to checkout repository.
Method scm.getKey() return full repository url.

There is no any environment variables containing project key or repository name.

[tomasbjerre]

Perhaps if you use https://github.com/tomasbjerre/pull-request-notifier-for-bitbucket and let it provide these variables for you? Then I think you can do:

projectKey                            : env.PULL_REQUEST_TO_REPO_PROJECT_KEY,

If you parameterize the job with a String parameter like PULL_REQUEST_TO_REPO_PROJECT_KEY

[testuser7]

Right now I'm using Bitbucket Branch Source Plugin with Post Webhooks for Bitbucket. Direct usage of scm variable will be most reasonable.

[tomasbjerre]

Perhaps you can get the project key from getRepoOwner and repository from getRepository. But I'm just guessing here =)

[testuser7]

I'm not sure if I can access this methods from pipeline groovy script.

[jansohn]

I'd also appreciate this feature.

[jetersen]

Definitely possible to get the ID and even more: https://github.com/jenkinsci/office-365-connector-plugin/blob/53f1823060c4a7aba72294a08c0efba6751664d5/src/main/java/jenkins/plugins/office365connector/ActionableBuilder.java#L59-L89

Here you get the full URL so ya to the PR even
String urlString = oma.getObjectUrl();

[amandel]

For me the following pipeline snipped could collect all needed data:

@NonCPS
void publishBuildResults() {
  try {
    def gitUrl = scmVars.GIT_URL_1
    // https://myserver.example.com/bitbucket/scm/SETOOLS/setools-sbs-bitbucket.git
    def urlMatcher = gitUrl =~ '^(https?://[^/]*/bitbucket)/scm/([^/]*)/([^/]*).git$'
    if (urlMatcher) {
      def bitbucketServer = urlMatcher[0][1]
      def projectKey = urlMatcher[0][2]
      def repoSlug = urlMatcher[0][3]
      def changeId = pipeline.env.CHANGE_ID;
      def credentialIdTemp = pipeline.scm.getUserRemoteConfigs()[0].credentialsId;
      pipeline.ViolationsToBitbucketServer([
          bitbucketServerUrl                    : bitbucketServer,
          commentOnlyChangedContent             : true,
          commentOnlyChangedContentContext      : 5,
          createCommentWithAllSingleFileComments: false,
          createSingleFileComments              : true,
          maxNumberOfViolations                 : 99999,
          keepOldComments                       : true,
          projectKey                            : projectKey,
          pullRequestId                         : changeId,
          repoSlug                              : repoSlug,
          credentialsId                         : credentialIdTemp,
          commentTemplate                       : """
**{{violation.severity}}**: {{violation.message}}
*Reporter: {{violation.reporter}}{{#violation.rule}} //  Rule: {{violation.rule}}{{/violation.rule}}*
""",
          violationConfigs            : [
              // Many more formats available, check https://github.com/tomasbjerre/violations-lib
              [parser: 'SONAR', pattern: '.*/target/sonar/sonar-report.json', reporter: 'SONAR']
          ]
      ])
    } else {
      pipeline.echo("Does not look like a bitbucket server - think about migrating?")
    }
  } catch (error) {
    // do not leak errors but report them.
    pipeline.echo("Error during result publish: " + error)
  }
}

[robross0606]

From what I can tell, this is really more the fault of Bitbucket Branch Source Plugin because it doesn't expose the Bitbucket JSON payload data (project, slug, etc.) from the webhook trigger into the pipeline.