Uploaded image for project: 'Weld'
  1. Weld
  2. WELD-531

Weld swallowing stacktrace in the case of multiple container errors

    Details

    • Type: Bug
    • Status: Closed (View Workflow)
    • Priority: Blocker
    • Resolution: Out of Date
    • Affects Version/s: 1.0.1.Final
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Description

      When there are multiple container errors (for instance, definition errors at startup), Weld cycles through the errors and builds a localized message. However, it drops the cause of the exception. If the message is further up the chain (i.e., in the exception cause), this leads to exceptions such as:

      Exception #0: null

      The user is lost as to what the real problem is. Typically root causes are:

      • NoClassDefFoundError
      • NullPointerException

      The problem is not that the message is null. The problem is that the underlying cause is dropped.

      This came up recently when using Arquillian as reported in this forum post: https://community.jboss.org/message/544395#544395

      When placed in a debugger, it turned out that a NoClassDefFoundError was being thrown. There was no way for the user to trace that to the source.

      I think that what Weld needs to do is follow the cause until it finds a non-null message.

        Gliffy Diagrams

          Attachments

            Issue Links

              Activity

                People

                • Assignee:
                  pmuir Pete Muir
                  Reporter:
                  dan.j.allen Dan Allen
                • Votes:
                  0 Vote for this issue
                  Watchers:
                  0 Start watching this issue

                  Dates

                  • Created:
                    Updated:
                    Resolved: