A binary code parser converts the machine code representation of a program, library, or code snippet to abstractions such as the instructions, basic blocks, and functions that the binary code represents. The ParseAPI is a multi-platform library for creating such abstractions from binary code sources. The current incarnation uses the Dyninst SymtabAPI as the default binary code source; all platforms and architectures handled by the SymtabAPI are supported. The ParseAPI is designed to be easily extensible to other binary code sources. Support for parsing binary code in memory dumps or other formats requires only implementation of a small interface as described in this document.
This API provides the user with a control flow-oriented view of a binary code source. Each code object such as a program binary or library is represented as a top-level collection containing the functions, basic blocks, and edges that represent the control flow graph. A simple query interface is provided for retrieving lower level objects like functions and basic blocks through address or other attribute lookups.
ParseAPI is distributed with Dyninst, and can be downloaded here
