This is a plugin for GCC, which links against libpython, and (I hope) allows you to invoke arbitrary Python scripts from inside the compiler. The aim is to allow you to write GCC plugins in Python.

The plugin is Free Software, licensed under the GPLv3 (or later).

It's still at the "experimental proof-of-concept stage"; expect crashes and tracebacks (I'm new to insides of GCC, and I may have misunderstood things).

It's already possible to use this to add additional compiler errors/warnings, e.g. domain-specific checks, or static analysis. One of my goals for this is to "teach" GCC about the common mistakes people make when writing extensions for CPython, but it could be used e.g. to teach GCC about GTK's reference-counting semantics, or about locking in the Linux kernel, or about signal-safety in APIs.

Other ideas include visualizations of code structure. Given a gcc.CFG instance, gccutils.render_to_dot(cfg) and gccutils.invoke_dot(cfg) will use graphviz and eog to plot a handy visualization of a control flow graph, showing the source code interleaved with GCC's GIMPLE internal representation.

• GCC: 4.6 or later (it uses APIs that weren't exposed to plugins in 4.5)
• Python: tested with 2.7 and 3.2; it may work with earlier versions

• "six": The libcpychecker code uses the "six" Python compatibility library to smooth over Python 2 vs Python 3 differences, both at build-time and run-time:

  http://pypi.python.org/pypi/six/

I use:

make

to build the plugin and run the tests

You can also use:

make demo

to demonstrate the new compiler errors.

All of my coding so far has been on Fedora 15 x86_64, using:

gcc-plugin-devel-4.6.0-0.15.fc15.x86_64

and I don't know to what extent it will be compatible with other versions and architectures.

The code also makes some assumptions about the Python version you have installed (grep for "PyRuntime" in the .py files). I've been using:

python-devel-2.7.1-5.fc15.x86_64
python-debug-2.7.1-5.fc15.x86_64
python3-debug-3.2-0.9.rc1.fc15.x86_64
python3-devel-3.2-0.9.rc1.fc15.x86_64

but you may have to hack up the PyRuntime() invocations in the code to get it to build on other machines. Ultimately I want the plugin to be buildable against multiple python versions, so there could be a python27.so, python27-debug.so, python-32mu.so, python-32-dmu.so, etc (c.f. PEP-3149)

There isn't an installer yet. In theory you should be able to add these arguments to the gcc invocation:

gcc -fplugin=python.so -fplugin-arg-python-script=PATH_TO_SCRIPT.py OTHER_ARGS

and have it run your script as the plugin starts up.

The plugin automatically adds the absolute path to its own directory to the end of its sys.path, so that it can find support modules, such as gccutils.py and libcpychecker.

The exact API is still in flux; you can currently connect to events by registering callbacks e.g. to be called for each function in the source at different passes.

It exposes GCC's various types as Python objects, within a "gcc" module. You can see the API by running:

import gcc
help(gcc)

from within a script.

This is currently three projects in one:

gcc-python-*: the plugin for GCC. The entrypoint (init_plugin) is in gcc-python.c

libcpychecker and cpychecker.py: a Python library (and a driver script), written for the plugin, in which I'm building new compiler warnings to help people find bugs in CPython extension code.

cpybuilder: a handy module for programatically generating C source code for CPython extensions. I use this both to generate parts of the GCC plugin, and also in the selftests for the cpychecker script. (I initially attempted to use Cython for the former, but wrapping the "tree" type hierarchy required more programatic control)

Coding style: Python and GCC each have their own coding style guide for C. I've chosen to follow Python's (PEP-7), as I prefer it (although my code is admittedly a mess in places).

You'll find API documentation within the "docs" directory, written in the reStructuredText format (as is this file, in fact). If you have Sphinx installed (http://sphinx.pocoo.org/), you can regenerate these docs using:

make html

within the docs directory. Sphinx is the python-sphinx package on a Fedora/RHEL box.

More detailed documentation can be seen within docs/getting-involved.rst

Enjoy! David Malcolm <dmalcolm@redhat.com>