mkProbe now prints a usage message showing allowable command line arguments when it is run without arguments.

Extended the EXPRESS Pretty Printer:

• Added command line options -a and -A to the exppp executable to specify that the output should be alphabetized. Without the -a option the order of the TYPEs and entities in the original EXPRESS is preserved when pretty printing an EXPRESS file.

• Fixed a bug in the exppp toolkit where it would crash when handling PROCEDUREs in EXPRESS.

• Extended the NIST Pretty Printer Toolkit to provide string and buffer representations for EXPRESS USE and REFERENCE statements, and for providing access to the EXPRESS SUPERTYPE statements.

• Added additional dictionary constructs for representing EXPRESS: SUPERTYPE statements, WHERE rules within entities, USE FROM clauses, REFERENCE FROM clauses, global rules, PROCEDUREs, and FUNCTIONS. Information about these objects is available but validation based on it is not performed.

• Enhanced fedex_plus C++ code generator to generate the necessary dictionary objects for the above EXPRESS constructs for a particular schema. Also enhanced schema dictionary code generation to add information necessary to generate an EXPRESS schema (i.e. which entities, types, global rules, USE and REFERENCE statements, etc are associated with which EXPRESS schema dictionary object).

• Added capability to generate EXPRESS from the above dictionary constructs (e.g. USE, REFERENCE, global rules, functions, procedures, etc.) and the existing dictionary constructs for EXPRESS schemas, types, entities, and attributes.

• Fixed a bug in the casting functions for the classes handling instances of SELECT TYPEs in EXPRESS. e.g.

  SdaiMeasure_value::operator SCLP23(Real)()
  {
  // was missing stmts such as these...
     if( CurrentUnderlyingType () == config_control_designt_area_measure )
        return  _real;
     if( CurrentUnderlyingType () == config_control_designt_context_dependent_measure )
        return  _real;
  }
  

• Fixed an aggregate element TypeDescriptor (TD) pointer problem in the dictionary. The TD pointer in the aggregate dictionary entry is supposed to point at the TD representing the element type for the aggregate. It pointed at the underlying or base type's TD (e.g. the TD for STRING) instead of for the type which was specified in the EXPRESS. For example where it should have pointed at a TD for a TYPE label - where label is defined as TYPE label = STRING; END_TYPE - it pointed at the TD for STRING. The TD for the aggregate element now points at the TD for label.

• Fixed the casting operator for the class representing EXPRESS enumeration types. It was returning the last element of the enum type as the the default when the enum was unset. This choice predated enums having an unset enum value. It now returns _unset as the default when the enum is unset.

• Cleaned up a bit of leaked memory. The memory leaked was mostly handles to static memory.

• sdaiString not reading a value '' properly.

• sdaiString copy constructor no longer mixes up undefined and unset values.

• Programming constructs in dictionary objects to handle and write WHERE rule and UNIQUE rule comments in EXPRESS. (Comments from originally parsed EXPRESS are not preserved.) Added a data member:

      SCLstring _comment; // Comment to be generated in the EXPRESS.
  

  to classes Where_rule and Uniqueness_rule. Also they have functions:

      SCLstring comment_() const { return _comment; }
      void comment_(const char* c) { _comment = c; }
  

  to manipulate them. When writing EXPRESS for entities and types, the keywords WHERE and UNIQUE are only written if there are actual rules. If there are only comments the keywords are written as a comment.