Bundled plugin loading mechanism leads to larger than necessary memory use

XMLWordPrintable

      The mechanism we use to read bundled plugins involves the atlassian-bundled-plugins.zip being read via the class loader and extracted into individual plugins, which are then loaded again via the PluginClassLoader.

      However, on Tomcat at least, the WebappClassLoader will retain the atlassian-bundled-plugins.zip for the duration of the server running, which means for Confluence 3.1-beta1 you have 45Mb of memory being needlessly used.

      I've made this minor because 45Mb is a pretty small percentage of Confluence's entire memory requirement, but it's still quite a load of memory.

            Assignee:
            Daniel (Inactive)
            Reporter:
            Paul Curren
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: