MyFaces Core
  1. MyFaces Core
  2. MYFACES-730

1.1.1RC3 jar interferes with Tomcat undeploy

    Details

    • Type: Bug Bug
    • Status: Open
    • Priority: Major Major
    • Resolution: Unresolved
    • Affects Version/s: 1.1.1
    • Fix Version/s: None
    • Component/s: General
    • Labels:
      None
    • Environment:
      XP SP 2, Tomcat 5.0.28, JDK 1.4.2_08

      Description

      Did a test upgrade of a MyFaces 1.1.0 based app to 1.1.1 RC3. The app is deployed in Tomcat 5.0.28.

      After the application is deployed and I use it for a little bit, I try to undeploy it. The undeploy fails because the myfaces-all.jar file is still open after the application is stopped. The only workaround is to stop Tomcat and restart it.

      The 1.1.0 myfaces-all.jar file did not exhibit this behavior.

      My experience is that this occurs if there is an open reference to a resource in the jar file.

        Activity

        Hide
        Simon Kitching added a comment -

        This report will be a whole lot of pain to deal with .

        Firstly, I presume it only happens on Windows. Such file-locking problems don't happen on non-windows systems.
        Secondly, I expect it would mean looking through every patch applied on the 1_1_1 branch since 1.1.0 to see if any are likely candidates for causing this.

        This issue does seem similar to one reported here recently:
        http://issues.apache.org/bugzilla/show_bug.cgi?id=37034

        Note however that this is not a digester bug as such; the patch was just accepted into digester as a workaround.
        And note also that while myfaces uses digester this isn't likely to be the cause of the reported problem, as
        digester has used the code "new InputSource(String)" [which apparently triggers the problem] for many releases.

        Regards,

        Simon

        Show
        Simon Kitching added a comment - This report will be a whole lot of pain to deal with . Firstly, I presume it only happens on Windows. Such file-locking problems don't happen on non-windows systems. Secondly, I expect it would mean looking through every patch applied on the 1_1_1 branch since 1.1.0 to see if any are likely candidates for causing this. This issue does seem similar to one reported here recently: http://issues.apache.org/bugzilla/show_bug.cgi?id=37034 Note however that this is not a digester bug as such; the patch was just accepted into digester as a workaround. And note also that while myfaces uses digester this isn't likely to be the cause of the reported problem, as digester has used the code "new InputSource(String)" [which apparently triggers the problem] for many releases. Regards, Simon
        Hide
        Steve Peterson added a comment -

        Is code review really the only way to get to this? I'd be willing to help but don't know where to start.

        I just tried to run it down by expanding the jar into the WEB-INF/classes but, of course, the webapp cleanly deploys in that situation.

        Note that the problem occurs during initialization; the jar is open at the point that the webapp is initialized, but before the first request is processed.

        http://tomcat.apache.org/faq/windows.html#lock has some notes on what might cause the problem.

        Show
        Steve Peterson added a comment - Is code review really the only way to get to this? I'd be willing to help but don't know where to start. I just tried to run it down by expanding the jar into the WEB-INF/classes but, of course, the webapp cleanly deploys in that situation. Note that the problem occurs during initialization; the jar is open at the point that the webapp is initialized, but before the first request is processed. http://tomcat.apache.org/faq/windows.html#lock has some notes on what might cause the problem.
        Hide
        Matthias Weßendorf added a comment -

        what's with newer versions?
        (nothing tested here from my side)

        Show
        Matthias Weßendorf added a comment - what's with newer versions? (nothing tested here from my side)
        Hide
        Steve Peterson added a comment -

        We're still on 1.1.1; I'll try it out on 1.1.3 when I get an opportunity.

        Show
        Steve Peterson added a comment - We're still on 1.1.1; I'll try it out on 1.1.3 when I get an opportunity.
        Hide
        Simon Kitching added a comment -

        Any feedback on this? If not, this should be closed as it is a hard-to-reproduce bug (only on original poster's machine) and is now 1.5 years old with no additional info.

        Show
        Simon Kitching added a comment - Any feedback on this? If not, this should be closed as it is a hard-to-reproduce bug (only on original poster's machine) and is now 1.5 years old with no additional info.

          People

          • Assignee:
            Unassigned
            Reporter:
            Steve Peterson
          • Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

            Dates

            • Created:
              Updated:

              Development