Skip to main content

A small python based build file generator targeting ninja

Project description

Latest PyPI version Build Status


pyrate is a small python based build file generator targeting ninja(s).

It allows to describe the build process of small projects in a very simple way. This description is then turned into ninja build files, that enable a very quick turnaround of project builds.


The following presents the necessary steps to quickly test the waters with this tool. These commands will install pyrate, generate the ninja build file, build and execute a small executable:

pip install pyrate-build
echo -e '#include <cstdio>\nint main() { printf("Ahoy World!"); return 0; }' > test.cpp
echo -e "executable('test', 'test.cpp')" >


pyrate is very easy to deploy - there are no particular installation steps to use it once the single script is available somewhere. It can even board the project directory of your project and simply get called from there. The only dependency of the software is having a working python installation. pyrate should work out of the box with all python versions between 2.4 and 3.4.

It is also possible to get the latest version from the Python Package Index with:

pip install pyrate-build


pyrate can be directly invoked with the name of the build configure script and the optional parameter ‘–output’ to specify the name of the generated ninja build file.

pyrate --output

When the script is started, it first changes the current directory to the directory containing the build configuration script, so all path names are relative to it.

If pyrate is placed in a directory listed in the PATH environment variable (as automatically done by pip install pyrate-build), the build configure script can be made executable to invoke pyrate automatically by starting the build config script with:

#!/usr/bin/env pyrate

Build File Configuration Syntax

The build configuration for pyrate is written in python - so the full power of python can be used to construct and describe the build process. Several classes, functions and variables are available to ease and customize the configuration of the build process.

Specifying build input

The build input files that is given to one of the functions described further below can be specified either as a space separated string or as a list of strings / items:

  • "<file1> <file2> ..."
  • ["<file1>", "<file2>", ...]

Besides specifying these lists it by hand, there are many ways to construct such a list. Common methods include calling the python function os.listdir or using the helper function match provided by pyrate:

  • match(selector, dir_name = '.')

This functions allows to select files from a directory using a string consisting of black / white listing path name patterns. The selector '*.cpp -test*.cpp test3.cpp *.h' for example selects all files ending with ‘.h’ and ‘.cpp’, with the exception of those ‘.cpp’ files that start with ‘test’ and are not called ‘test3.cpp’.

The list of build inputs can also contain external packages, build targets or any other kind of BuildSource (explained later).

Defining build targets

There are four global helper functions to define object files, executables and libraries based on a list of build inputs (which can be files, other targets or externals)

  • executable(name, input_list, compiler_opts = None, linker_opts = None)
  • shared_library(name, input_list, compiler_opts = None, linker_opts = None)
  • static_library(name, input_list, compiler_opts = None, linker_opts = None)
  • object_file(name, input_list, compiler_opts = None)

Each function returns a build target object, that can be used as input / dependency of another function. If multiple executables / libraries or object files with the same name but different inputs / options are defined, pyrate will ensure that the output will have a unique name (by appending a hash based suffix as needed).

The input list of these functions may contain:

  • strings (file names that are processed according to the rules specified by the packages in the compiler dictionary),
  • build targets (as returned by these functions themselves) or
  • external dependencies (retrieved using find_external or explicitly defined).

These functions exist as global functions and as member functions of a so-called build context, that describes how these functions are processed. The global functions are just executing within the default build context.

By default, all build targets that are defined by the above functions (or direct API calls) are built. In order to select these default targets, the global variable default_targets can be set to a list of targets:

  • default_targets = [<target>,...] (None == all targets are built)

External dependencies

The build environment / dependencies of external packages can be expressed using the following functions / variables:

  • find_external(name, ...)

The function find_external searches for some external dependency (built-in or self-defined) with the given name and returns either None or a representation of the dependency. The function takes additional positional and keyword arguments that depend on the external package. A common argument for this function is a version selector, that is supplied through a global variable:

  • version

The comparisons with this variable (eg. version >= 4.1) will create new version instance that is used by the external package finder. This allows for example to write find_external('clang', version >= 3.5) to discover a clang installation with version 3.5 or later. Currently only a small number of built in external packages are available (listed under Externals), but it is easy to add new packages that are recognized.

Configuration of the build environment

The build context allows for example to define implicit dependencies that are automatically included in all generated object files, executables or libraries. An instance of such a build context is created with:

  • Context(...) - the most important parameters are:
    • implicit_input, implicit_object_input, implicit_static_library_input, implicit_shared_library_input and implicit_executable_input

The default context used by these global function can be set using the variable:

  • default_context = Context(...)

Finally, the used default compilers can be configured via the global variable

  • compiler

This is a dictionary that contains links to external packages that provide the basic rules and parameters that are used to build the source. This dictionary can be modified, but should not be overwritten.


Currently the following built in externals are supported (listed with possible find_external arguments):

  • gcc
    • version - specifies required version (eg. version >= 5.2)
    • std - C/C++ language standard version (eg. gnu++14)
    • compiler_flags - flags that are used during the compilation stage
    • static_flags, shared_flags, exe_flags - flags that are used during the linking stage
  • clang
    • version - specifies required version (eg. version > 3.5)
    • std - C/C++ language standard version (eg. c++14)
    • compiler_flags - flags that are used during the compilation stage
    • static_flags, shared_flags, exe_flags - flags that are used during the linking stage
    • version - specifies required version (eg. version >= 2.6)
  • python
    • version - specifies required version (eg. version >= 2.6)
  • swig - The swig package also provides a member function to describe the generation of automated interface code
    • version - specifies required version (eg. version > '3.0.2')
    • wrapper(target_language, library_name, interface_filename, libs = [<targets>...])
  • pthread - posix thread library


The basic pyrate build configuration file for a simple C++ project with a single source file producing a single executable looks like this:

executable('test', ['test.cpp'])

A more complicated example is presented in the following code fragment. It demonstrates how to

  • change the default compiler to clang,
  • define a native static and dynamic library from a set of files selected by wildcards,
  • generate several executables accessing to the shared library and
  • generate a wrapper library to access the C++ library from python (if swig is available).
clang = find_external('clang', version >= 3.7, std = 'c++11')
if clang:
    compiler['C++'] = clang

lib_files = match("*.cpp -test* -mylib.cpp")
static_library('libFoo', lib_files, compiler_opts = '-O3')
lib_reference = shared_library('libFoo', lib_files)

python = find_external('python', version > 2)
swig = find_external('swig')
if swig and python:
    swig.wrapper('python', 'mylib', 'mylib.i', libs = [lib_reference])

for fn in match("test*.cpp"):
    executable(fn.replace('.cpp', '.exe'), [fn, lib_reference])

Many more examples with an increasing level of complexity are available in the github repository.

Project details

Download files

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

Files for pyrate-build, version 0.1.7
Filename, size File type Python version Upload date Hashes
Filename, size pyrate_build-0.1.7-py2.py3-none-any.whl (16.8 kB) File type Wheel Python version 2.7 Upload date Hashes View
Filename, size pyrate-build-0.1.7.tar.gz (15.3 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page