Union schemas as a basis for a C++ extractor

An extractor is a program which processes source code and outputs facts about the code in a software exchange format (SEF). An SEF can be further specified by a schema, analogous to a schema for a database. This paper explains how two such schemas can be combined into a union schema as the basis for creating an extractor. We describe CPPX (C Plus Plus eXtractor), which transforms a schema designed as the internals of a compiler (GCC) to a schema designed for software exchange (Datrix). CPPX performs this transformation as a pipelined sequence of sub-transformations. At each stage in the pipeline, the intermediate data conforms to the union of the two schemas