• Icon: Bug Bug
    • Resolution: Unresolved
    • Icon: Minor Minor
    • core
    • Linx Fedora 20 x64

      • Suppose build dependencies like:
        • project dist depends on project A and project B
        • both project A and project B have 'Trigger only if build is stable'
          in the 'Build other projects' section
      • project A and project B are both starting a build at the same time.
      • a build for project dist is scheduled
      • project A build succeeds
      • the build for project dist blocks because project B is still building
      • project B build fails
      • project dist builds (even though project B failed!) and fails

      The build for project dist should have been cancelled

          [JENKINS-25712] Build triggering logic is wrong

          Daniel Beck added a comment -

          Not a bug, because that's simply not how those triggers work. They are acting independently.

          Look into the various workflow related plugins for Jenkins to see if any of them can do what you want here (e.g. MultiJob, Build Pipeline, Build Flow, the early stage Workflow plugin, etc.)

          Daniel Beck added a comment - Not a bug, because that's simply not how those triggers work. They are acting independently. Look into the various workflow related plugins for Jenkins to see if any of them can do what you want here (e.g. MultiJob, Build Pipeline, Build Flow, the early stage Workflow plugin, etc.)

          Ferry Huberts added a comment -

          They must not be independent.
          'dist' depends on both A and B.
          Note the 'and' which technically translates to '&&'.
          'independent' technically translates to '||'

          Therefore the implementation is not a correct reflection of the dependency chain.
          Hence (to me) it really is a bug

          Ferry Huberts added a comment - They must not be independent. 'dist' depends on both A and B. Note the 'and' which technically translates to '&&'. 'independent' technically translates to '||' Therefore the implementation is not a correct reflection of the dependency chain. Hence (to me) it really is a bug

          Daniel Beck added a comment -

          Suit yourself. That's not how those triggers work and will likely never be fixed, as there are almost guaranteed to be numerous users who are expecting the current behavior.

          Daniel Beck added a comment - Suit yourself. That's not how those triggers work and will likely never be fixed, as there are almost guaranteed to be numerous users who are expecting the current behavior.

            Unassigned Unassigned
            fhuberts Ferry Huberts
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated: