Skip to main content

A subset of Python that can be compiled to C++ meta-functions using the py2tmp compiler

Project description

Build Status

TMPPy is a subset of Python that can be compiled to C++ meta-functions using the py2tmp compiler. This project is aimed at C++ library developers whose libraries include a non-trivial amount of C++ metaprogramming.

Compared to writing C++ metaprogramming code directly, using TMPPy allows that code to be expressed in a more concise and readable way, provides static type checking (avoiding some classes of instantiation-time errors) and produces optimized C++ meta-functions, reducing the compile time for the C++ compilation.


As an example, let’s write a metafunction (aka type trait class) add_pointer_multiple such that:

  • add_pointer_multiple<T, 0>::type is T
  • add_pointer_multiple<T, 1>::type is T*
  • add_pointer_multiple<T, 2>::type is T**
  • (and so on)

This can be written as a template, as follows:

template <typename T, int64_t n>
struct add_pointer_multiple {
    using type = typename add_pointer_multiple<T, n - 1>::type*;

template <typename T>
struct add_pointer_multiple<T, 0> {
    using type = T;

However this syntax is quite verbose and not very readable. For more complex metafunctions this becomes a significant issue, leading to more bugs and more effort when debugging or maintaining the code.

Some C++ metaprogramming libraries (notably Boost’s MPL library) can be used to reduce the verbosity, however that comes at the price of slower compile times.

Using TMPPy, the above can be written as:

def add_pointer_multiple(t: Type, n: int) -> Type:
    if n == 0:
        return t
        return Type('T*', T=add_pointer_multiple(t, n-1))

And this TMPPy code can then be compiled to C++ code equivalent to the metafunction above (without the overhead of e.g. MPL).

For more information on TMPPy, see the wiki.


TMPPy is released under the Apache 2.0 license. See the LICENSE file for details.

This is not an official Google product.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Filename, size & hash SHA256 hash help File type Python version Upload date
TMPPy-0.1.3-py3-none-any.whl (62.0 kB) Copy SHA256 hash SHA256 Wheel py3
TMPPy-0.1.3.tar.gz (52.6 kB) Copy SHA256 hash SHA256 Source None

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page