Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-50670

Pipeline job does not fail when serialization of Pipeline steps arguments fails within NonCPS

      Discovered it during testing of JENKINS-50665

      Example (for nonCPS wrapper which may not work correctly as jglick mentioned in the PR):

      nonCPS.exec() { 
          def formatter = new org.jfree.chart.axis.QuarterDateFormat()
          printTimestamp arg1: 'foo', formatter: formatter, persistFormatter: true
      }
      

          [JENKINS-50670] Pipeline job does not fail when serialization of Pipeline steps arguments fails within NonCPS

          Oleg Nenashev added a comment -

          Oleg Nenashev added a comment - Reproduced the issue in https://github.com/jenkinsci/workflow-cps-plugin/pull/218 (autotest)

          Jesse Glick added a comment -

          As designed I think. ArgumentsAction represents a best effort at recording metadata. If something cannot be recorded, so be it—the program continues and metadata is incomplete.

          Jesse Glick added a comment - As designed I think. ArgumentsAction represents a best effort at recording metadata. If something cannot be recorded, so be it—the program continues and metadata is incomplete.

          Sam Van Oort added a comment -

          Yes, as designed. ArgumentAction data is considered 'noncritical.'

          Sam Van Oort added a comment - Yes, as designed. ArgumentAction data is considered 'noncritical.'

          Oleg Nenashev added a comment -

          OK, I will invert the test in the pull request then

          Oleg Nenashev added a comment - OK, I will invert the test in the pull request then

          Sam Van Oort added a comment -

          I created JENKINS-50752 to track the fact that we're not successfully bulletproofing against unserializable Action contents.

          Sam Van Oort added a comment - I created JENKINS-50752 to track the fact that we're not successfully bulletproofing against unserializable Action contents.

          Sam Van Oort added a comment -

          oleg_nenashev the actual argument data is being bulletproofed by JENKINS-50752 implementation now so it CANNOT cause failures elsewhere if it can't be persisted and is instead discarded.

          Sam Van Oort added a comment - oleg_nenashev the actual argument data is being bulletproofed by JENKINS-50752 implementation now so it CANNOT cause failures elsewhere if it can't be persisted and is instead discarded.

            Unassigned Unassigned
            oleg_nenashev Oleg Nenashev
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

              Created:
              Updated:
              Resolved: