peg/leg — recursive-descent parser generators for C

peg and leg are tools for generating recursive-descent parsers: programs that perform pattern matching on text. They processes a Parsing Expression Grammar (PEG) [Ford 2004] to produce a program that recognises legal sentences of that grammar. peg processes PEGs written using the original syntax described by Ford; leg processes PEGs written using slightly different syntax and conventions that are intended to make it an attractive replacement for parsers built with lex and yacc. Unlike lex and yacc, peg and leg support unlimited backtracking, provide ordered choice as a means for disambiguation, and can combine scanning (lexical analysis) and parsing (syntactic analysis) into a single activity.

The current stable version is 0.1.19. Links are below.

The development version is 0.1.20 and is updated periodically. If you want to help find bugs please try it out. A tarball is linked in the version history.

Download the source code: peg-0.1.19.tar.gz
Browse the source code: peg-0.1.19
Read the manual page: peg.1.html
Enjoy the Emacs mode: leg-mode.el
peg is distributed under the MIT license. It will not infect your project with a contagious disease license if you decide to modify it for your own use. The parsers that peg and leg create are unencumbered and you are free to use and/or distribute them any way you like.

If you find bugs or have suggestions, please send them to piumarta via gmail.

References

Version history

2023-11-19 peg-0.1.20 (development, subject to change)
Experimental support for variables and semantic values in ‘inline actions’ executed during recognition (separate from the parser actions that are run after the input is fully recognised).
Don't print #line directives for headers if -P was given and pathnames on WIN32 have \ changed to / to keep #line directives happy (thanks to Ben Hines for the problem report).

2023-03-06 peg-0.1.19 (stable)
Experimental support for re-entrant parsing through yyparsefrom_r().

2016-07-22 peg-0.1.18
Add -P option to disable #line directives. Emit #line directives for all actions.
2016-07-14 peg-0.1.17
Emit #line directives for header and trailer in the generated source file.
2016-06-25 peg-0.1.16
Add @{...} actions that are performed during matching.
2013-12-17 peg-0.1.15
Calls to YY_FREE fixed (thanks to Andrew Dunham).
2013-12-01 peg-0.1.14
Documentation typos fixed (thanks to Giulio Paci).
2013-08-16 peg-0.1.13
Predicate actions can refer to yytext (thanks to Grégory Pakosz). Hexadecimal character escapes are supported by leg (thanks to Hugo Etchegoyen).
2013-07-20 peg-0.1.12
Use BSD-licensed getopt() in Windows build. Verbose mode handles Variable nodes.
2013-06-03 peg-0.1.11
Add error actions via "~" operator. Support declaration of local variables at the top level of semantic actions. Dynamically grow data structures to remove artificial limits on rule recursion (thanks to Alex Klinkhamer). Many small changes to better support C++. Add build files for Win32 and MacOS (thanks to Fyodor Sheremetyev). Update manual page to describe new features.
2012-04-29 peg-0.1.9
Move global state into a structure to facilitate reentrant and thread-safe parsers (thanks to Dmitry Lipovoi).
2012-03-29 peg-0.1.8
Allow nested, matched braces within actions.
2011-11-25 peg-0.1.7
Fix matching of 8-bit chars to allow utf-8 sequences in matching expressions (thanks to Grégory Pakosz).
2011-11-24 peg-0.1.6
Allow octal escapes in character classes.
2011-11-24 peg-0.1.5
Remove dwarf sym dirs when cleaning. Fix size calculation when resizing text buffers. Backslash can be escaped.
2009-08-26 peg-0.1.4
Fix match of a single single quote character. Rename getline -> nextline to avoid C namespace conflict.
2007-09-13 peg-0.1.3
Allow matched braces inside leg actions. Handle empty rules. Handle empty grammars.
2007-08-31 peg-0.1.2
Grow buffers while (not if) they are too small. Remove dependencies on grammar files. Add more basic examples.
2007-05-15 peg-0.1.1
First public release.