Chapter 3  Managing Data and Memory in Mixed-Language Applications

ECLⁱPS^e is a software engine for constraint propagation and search tasks. As such, it represents its data in a form that is different from how it would be represented in a traditional C/C++ program. In particular, the ECLⁱPS^e data representation supports automatic memory management and garbage collection, modifications that can be undone in a search context, referential transparency and dynamic typing.

In a mixed-language application, there are two basic ways of communicating information between the components coded in the different languages:

Conversion:

When data is needed for processing in another language, it can be converted to the corresponding representation. This technique is appropriate for simple data types (integers, strings), but can have a lot of overhead for complex structures.

Sharing:

The bulk of the data is left in its original representation, referred to by a handle, and interface functions (or methods) provide access to its components when required.

Both techniques are supported by the ECLⁱPS^e/C and ECLⁱPS^e/C++ interface.

• Constructing ECLⁱPS^e data
  • ECLⁱPS^e atoms and functors
  • Building ECLⁱPS^e terms
  • Building atomic ECLⁱPS^e terms
  • Building ECLⁱPS^e lists
  • Building ECLⁱPS^e structures
• Converting ECLⁱPS^e data to C data
  • Converting simple ECLⁱPS^e terms to C data types
  • Decomposing ECLⁱPS^e terms
• Referring to ECLⁱPS^e terms
• Passing generic C or C++ data to ECLⁱPS^e
  • Wrapping and unwrapping external data in an ECLⁱPS^e term
  • The method table