Skip to content


Subversion checkout URL

You can clone with
Download ZIP
Tools for code analysis, visualizations, or style-preserving source transformation.
OCaml Java PHP Makefile TeX Prolog Other
Latest commit 404134f @aryx aryx Merge pull request #134 from mmcco/master
Remove GNU make version check
Failed to load latest commit information.
.arc * .arc/default-relative-commit:
code_graph cm/cg: fix Gtk warning Critical bla bla Remove xxx
code_map * code_map/Visual.tex.nw: misc
commons * commons/ misc
data [pfff] remove `rm` for nonexistant files
demos CEs on commons/lib-json/ -> external/jsonwheel and also removed
docs * data/cpp_stdlib/macros.h: cleanup, and introduced pfff_macros.h loc…
editors/emacs * editors/emacs/sgrep.el:
external [pfff] add json_out to exported symbols for jsonwheel
globals many more META files for findlib
graph_code Merge branch 'master' of
h_files-format many more META files for findlib
h_program-lang * lang_nw/parsing/lexer_nw.mll: highlight special \t footnote (pad sp…
h_program-visual * commons/ handling .coffee
h_version-control [pfff] export findlib target for h_version-control
h_visualization * h_visualization/META: export, for efuns
lang_bytecode * h_program-lang/ new, extracted from graph_code_pro.og
lang_c * lang_c/parsing/ typo
lang_clang * h_program-lang/ misc
lang_cmt * Makefile: remove CLANG_HACK
lang_cpp CEs on Typedef constructor removal
lang_csharp CEs on highlight_code and entity_code factorization
lang_css * lang_css/parsing/ remove dead type
lang_erlang * lang_erlang/parsing/ cleanup, remove program2
lang_haskell CEs on Typedef constructor removal
lang_html * lang_html/parsing/ cleanup, remove program2
lang_java CEs on highlight_code and entity_code factorization
lang_js * lang_js/analyze/ Fix invalid documentation comments.
lang_lisp * h_program-lang/ factorize lexbuf_to_strpos
lang_ml * lang_ml/analyze/visual/ CE
lang_nw * lang_nw/parsing/lexer_nw.mll: highlight special \t footnote (pad sp…
lang_opa CEs on Typedef constructor removal
lang_php [pfff] Support anonymous and nested shapes
lang_python * lang_python/parsing/ cleanup
lang_rust CEs on Typedef constructor removal
lang_sql/parsing * h_program-lang/ factorize lexbuf_to_strpos
lang_text CEs on commons/lib-json/ -> external/jsonwheel and also removed
lang_web CEs on commons/lib-json/ -> external/jsonwheel and also removed
matcher * matcher/ fix test regressions
mini * mini/readme.txt:
opa * opa/pfff_logger.opa:
scripts clean lang_js/ remove avik's code, remove NotParsedCorrectly, FinalDef
tests Merge pull request #118 from mutantzombie/multiple_patterns
web [pfff] move commons_graph to a separate directory
.arcconfig * .arcconfig: new file
.gitignore * .gitignore:
.travis.yml * .travis.yml: send email to gmail too
Makefile * Makefile: mv in Makefile.config the Wno-invalid-token hack for clang
Makefile.common * Makefile.common: reinstall-findlib [pfff] add travis CI badge to readme
architecture.txt * architecture.txt:
authors.txt * authors.txt: update
bugs.txt * bugs.txt: changed email
changes.txt * changes.txt:
configure Build systems very often (usually?) don't check the make version.
copyright.txt * configure: new option --nocmt
credits.txt * credits.txt: adding john whaley
env.csh initial import into fresh git repo. [pfff] switch to 4.01 * lang_lisp/analyze/ first version, module only deps
find_source.mli factorize more skip_code stuff
gtkrc-2.0 * gtkrc-2.0: example of config
install.txt port to 4.02 (disabled lang_cmt/ for now in configure), fixed
install_linux.txt * install_linux.txt:
install_macos.txt Fix MacOSX install instructions
license.txt initial import into fresh git repo. * new file! moved lpizer here too * -test_dotfile_of_dotdepend, for cruel hessian, * -test_loc support for multiple dir (so can use ski… * todo * misc * handle -o for -db_of_graph_code and generate json by de… * todo Reuse the AST so multiple pattern matching is more efficient. * related work * related work * -luisa
project.el * project.el: update
skip_list.txt * skip_list.txt:
syncweb.txt * syncweb.txt: * yes


Build Status

pfff is a set of tools and APIs to perform static analysis, code visualizations, code navigations, or style-preserving source-to-source transformations such as refactorings on source code. There is good support for C, Java, Javascript and PHP. There is also preliminary support for other languages such as C++, Rust, C#, Html, CSS, Erlang, Lisp, Haskell, Python, OPA and SQL. There is also very good support for OCaml code so that the framework can be used on the code of pfff itself.

For each languages there are mainly 2 libraries, for instance parsing_php.cma and analysis_php.cma, that you can embed in your own application if you need to process PHP code. See the demos/ directory for example of use of the pfff API. See also docs/manual/Parsing_xxx.pdf and docs/manual/Analyzis_xxx.pdf for more documentation on how to use or extend pfff.

pfff is also made of few tools:

  • pfff, which allows to test the different parsers on a single file
  • scheck, a bug finder
  • stags, an Emacs tag generator
  • sgrep, a syntactical grep
  • spatch, a syntactical patch
  • codemap, which is a gtk and cairo based source code visualizer/navigator/searcher leveraging the information computed previously by pfff_db and codegraph.
  • codegraph, a source code indexer and package/module/class dependency visualizer
  • codequery, an interactive tool a la SQL to query information about the structure of a codebase using Prolog as the query engine
  • pfff_db, which does some global analysis on a set of source files and store the data in a marshalled form in a file somewhere (e.g. /tmp/db.json)

For more information, look at the pfff wiki here: as well as the docs/manual/ directory.

Usage for pfff:

$ ./pfff -parse_php demos/foo.php


$ ./pfff -dump_php demos/foo.php

You can also look at ./pfff --help

Usage for pfff_db:

$ ./pfff_db -lang ml -o /tmp/pfff.json ~/pfff

to analyze all the .ml and .mli files under ~/pfff and store metadata information (the database) in /tmp/pfff.json

Usage for codemap:

$ ./codemap ~/pfff

This should launch a gtk-based GUI that allows you to visualize source code and perform some code search.

Usage for codegraph:

$ ./codegraph -lang cmt -build ~/pfff

to generate a graph_code.marshall file in ~/pfff containing all dependency information about the pfff codebase using the typed bytecode .cmt files generated during the compilation of pfff.

$ ./codegraph ~/pfff

This should launch a gtk-based GUI that allows you to visualize source code dependencies.

More information

Look at the pfff wiki here:

Something went wrong with that request. Please try again.