Welcome! Thank you for considering the README file.
Contents
- What is bparser?
- Why using bparser?
- Getting started
- Documentation
- Credits
- Contact
- Enclosed documents
It is a library for parsing ANSII C99 code. It currently supports C99 almost completely (with some minor lookup problems in rare cases) + struct attributes. C11 is not yet completely supported.
For generating the parser, it uses bison and flex. The corresponding rules were taken from Jutta Degener, quut.com and have been filled with code.
- bparser has no library dependencies (bison and flex are used for code generating only)
- bparser is a very fast parser (due to being an LALR parser)
- bparser is minimal: a very small library with no more features than you need
- bparser is easy to use and extend. It uses visitor- and similar concepts.
C-Code is basically parsed like this:
- The flexer reads tokens (using GNU flex, see files lexer.l and lexer_common.h), and the parser (see parser.y) builts an incomplete AST (abstract syntax tree) from them. These files can be hard to read.
- The AST is yet incomplete. visitors are run on its leafs in order to
- assign parent pointers for all leafs
- pass the token geometry (source file, line and column) to all leaves
- complete type info for the AST nodes (type_completor.h). E.g. for the occurence of a variable, you might want to know where it has been defined and what type it is.
All this is being done with the routine from ast.h.
For then using the AST, you might want to implement your own visitor. A few useful visitors already exist in src/utils.h. For more details about the AST nodes themselves, see node.h. This file also explains how C grammar works.
Type make doc for doxygen. TODO: more stuff here...
I thank Jutta Degener et al. for providing the useful ANSI C Yacc grammar.
Feel free to give feedback. The author's e-mail address is shown if you execute this in a shell:
printf "\x6a\x6f\x68\x61\x6e\x6e\x65\x73\x40\x6c\x6f\
\x72\x65\x6e\x7a\x2d\x68\x6f\x2e\x6d\x65\x0a"