PH03966: TESTCONDUCTOR: WRONG RHAPSODYâ ™S CODE GENERATION FOR TEST COMPONENTS WITH FLOWPORTS

IBM Rational Rhapsody 8.4

APAR status

• Closed as program error.

Error description

• When creating a test architecture for a class which uses ports,
  it doesn't compile.
  The problem is that in TC_at_x_of_class_0 a setter setX() is
  generated:
  
  TC_at_x_of_class_0_setX(TC_at_x_of_class_0* const me, const
  mystruc* _x) where the data-argument 'const mystruc* _x'
  conforms to C_CG.Type.In of mystruc, but should be 'mystruc _x'
  to conform to the way it is called by class_0 (class_0 declares
  a function pointer to the setter on receiver flowport. Here
  'void TC_at_x_of_class_0_setX(TC_at_x_of_class_0* const me,
  mystruc _x)' would be appropriate):
  
  struct class_0_t {
  RiCReactive ric_reactive;
  /*#[ ignore */
  void (* const x_cb)(void* const,mystruc);
  void* const x_trg;
  /*#]*/
  int a; /*## attribute a */
  int b; /*## attribute b */
  mystruc x; /*## attribute x */
  /*#[ ignore */
  RhpInteger rootState_subState;
  RhpInteger rootState_active;
  /*#]*/
  };
  
  Besides the compile error at the end of the generated setter,
  for 'me->x = _x;', which should read 'me->x = *_x;' instead if
  the signature of the setter was correct, invocation of
  TC_at_x_of_class_0_setX() with a mystruc Argument instead of an
  mystruc* argument leads to interpretation of a structure value
  as pointed (->crash).
  
  BTC-ES could fix the setter according to the actual cg for
  setters, but this would mean to rely on an inconsistency in cg:
  attribute setter cg does not take C_CG.Type.In into account (set
  to 'const $type*') but generates a setter with plain value as
  argument (as if C_CG.Type.In was set to '$type'). A real fix
  would require a common solution of cg and TestConductor
  treatment of setters.
  
  This occurs for basically every class which uses ports (almost
  100% of our classes).
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * Rhapsody                                                     *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * TestConductor: Wrong Rhapsodyâ ™s code generation for test   *
  * components with flowports                                    *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  

Problem conclusion

• The generation of test code has been corrected, it now
  corresponds to Rhapsody CG behavior.
  Note: It might be necessary to apply "Clean TestPackage" on the
  test architecture to make sure test code is generated correctly.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  TLOGIC RHAPSODY

• Fixed component ID

  5724V74RP

Applicable component levels

• R830 PSY

     UP