Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-6976

In incremental build mode - if a breaking commit is done to a module while another module is broken, the build eventually succeeds

    • Icon: Bug Bug
    • Resolution: Duplicate
    • Icon: Major Major
    • maven-plugin
    • None
    • Hudson version 1.353

      Two unrelated modules A and B. According to the parent pom, B is built after A (but does not depend on it).

      1. John makes a commit to module A that breaks it.
      2. Hudson forks a build that compiles module A and fails.
      3. Paul makes a commit to module B that breaks it also.
      4. Hudson forks a build that intends to compile module B, but first compiles module A (it is still failed) and fails.
      5. John fixes module A.
      6. Hudson forks a build that compiles A only (B is not among its dependencies).

      ==> Build is marked successful while there is a "hidden time-bomb" in module B.

          [JENKINS-6976] In incremental build mode - if a breaking commit is done to a module while another module is broken, the build eventually succeeds

          kutzi added a comment -

          I think the behaviour was fixed, so that unstable/failing modules are alway rebuild.
          Can you retest with a current version of Hudson?

          kutzi added a comment - I think the behaviour was fixed, so that unstable/failing modules are alway rebuild. Can you retest with a current version of Hudson?

          Have just verified with version 1.386. Still the same problem. The problem is that Hudson doesn't know that the module B is failing because it has never built it after the breaking change described at the item #3.

          victor_bronstein added a comment - Have just verified with version 1.386. Still the same problem. The problem is that Hudson doesn't know that the module B is failing because it has never built it after the breaking change described at the item #3.

          kutzi added a comment -

          I think this is the same problem as JENKINS-5764

          kutzi added a comment - I think this is the same problem as JENKINS-5764

            Unassigned Unassigned
            victor_bronstein victor_bronstein
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: