There is a functional requirement for MIDP7 TCK to simplify TCK configuration.

To satisfy this requirement, besides obvious things like merging some questions
in the interview, we split TCK confifuration onto two independent parts:
device setup and test run environment setup. Our scenario can be described
as follows:

============================================================
"New scenario is pretty much similar to JDTS 2.0's

Currently, we have single Configuration where "licensee" should answer
all questions, including:
  1) Device features description. For example, how many colors are supported by the Toolkit,
       or if serial port is available on the device.

   2) Test run setup, not related to network settings. For example, whether to export tests or not,
        or Agent/Server/JT verbosity level.

   3) Network setup and the values which may and/or must vary across different JT machines. For example,
       HTTP(s) server host/port, DTF setup, optional baud rates for serial link (since PC's UART might support
       some or not).


So, as you have understood, the suggestion is to split TCK setup into two parts:
template and configuration. The following claims apply:

    - It will be possible to create work directory without template. In this case,
      all questions 1), 2) and 3) will be shown in the configuration

    - It will be possible to create a template. In the template interview wizard UI, only 1) questions will be shown.

    - If configuration is created on top of template, only 2) and 3) questions will be shown.


Next:
  - 2) questions will have default values corresponding to the typical test run.

  - 3) questions will be unanswered in the Configuration, requiring the answers.

  - We'll provide template file answers for the RI, with only 1) answers included. RI/TCK QA will
    only need to load template and answer the questions corresponding to the RI.

 - Licensee might have created single template for their implementation and place it into the shared
   folder (similarly to JDTS 2.0). Test operators just load the template and configure their
   network/hardware/test run setup.

=================================================================

Briefly, I have implemented desired behavior, but implementation includes numerous hacks,
and spotted several problems:

 - In the TCK interview constructor, it is impossible to realize if it is a Template or Configuration
   while in our TCKs, the path is being built in the constructor body, i.e., I have to
   keep following code in MidTCKInterview:
   ======================= MidTckInterview.java ================================================
      private void evaluateIfTemplate() {
        StackTraceElement[] strace = Thread.currentThread().getStackTrace();
        for (StackTraceElement stackTraceElement : strace) {
            if (stackTraceElement.getClassName().equals("com.sun.javatest.exec.BasicSessionControl") &&
                    stackTraceElement.getMethodName().equals("newConfig")) { // "New Configuration"
                setTemplate(false);
            }
        }
    }
   =============================================================================================

 - Templates behavior is not clear at all
 ========================================
   In my CtxMgr the following takes place:
    featureToggles[FeatureManager.SHOW_TEMPLATE_UPDATE] = true;
    featureToggles[FeatureManager.WD_WITHOUT_TEMPLATE] = false;
    featureToggles[FeatureManager.TEMPLATE_CREATION] = true;

    My expectation is is that the sequence must be applicable during template/workdir/configuration creation:
    - create and save a template
    - create workdir (provide a pointer at saved template file)
    - create a configuration for the workdir (and base it on the template)

    Nevertheless, "New template..." becomes active only AFTER work directory is created
    and another template ws loaded (due to WD_WITHOUT_TEMPLATE=false) already.
    I think that this is a bug, please object.