header2whatever 0.1

Generate files from C/C++ headers using jinja2 templates

header2whatever (h2w)

Generate arbitrary files from C/C++ header files using CppHeaderParser to read the input files and Jinja2 templates to generate the outputs.

This grew out of a desire to generate pybind11 wrapping code from C++ header files. pybind11gen was created, but then I wanted to generate more things…

There are still rough edges, and the documentation is mostly nonexistent, but pull requests with fixes/improvements are very welcome!

Install

pip install header2whatever

Usage

First, you need to create a jinja2 template that represents whatever you want to generate from the header file. For example, maybe you want to describe the functions in yaml:

---
{% for header in headers %}
{% for fn in header.functions %}
{{ fn.name }}:
  returns: {{ fn.returns }}
  params:
  {% for param in fn.parameters %}
  - { name: {{ param.name }}, type: "{{ param.type }}" }
  {% endfor %}

{% endfor %}
{% endfor %}

And let’s say you have the following header file foo.h:

void some_fn(int i);
int returns_int(int p1, char* p2);

You can execute the following:

h2w foo.h -o foo.yml

And you’ll get the following output:

---
returns_int:
  returns: int
  params:
  - { name: p1, type: "int" }
  - { name: p2, type: "char *" }

some_fn:
  returns: void
  params:
  - { name: i, type: "int" }

As you can see, while this is a silly example, this approach is very flexible and fairly powerful.

Currently, the data structure passed to the template isn’t documented – but it’s a filtered version of whatever CppHeaderParser outputs when it parses a header.

See the examples folder for more examples.

Batch mode

If you need to process multiple files, or just want to record the parameters for autogenerating a file without writing a shell script, batch mode is useful. You pass two parameters: a yaml file with the configuration, and an output directory to write the files to.

Using data from external sources

Sometimes you want to mix in data that CppHeaderParser can’t give you. If you pass the --yaml option, it will load the yaml into a dictionary and make it available to the template as the ‘data’ variable.

You can also pass key=value parameters via the --param option, and the specified keys will be available to the template.

Custom processing

When you need to do more complex logic that a jinja2 template just isn’t appropriate for, you can specify a python file to load custom hooks from.

See [the default hooks](header2whatever/default_hooks.py) for documentation.

License

Apache 2

Author

Dustin Spicuzza ([e-mail])