Discuss on Groups View on GitHub

Other Languages

Cap’n Proto’s reference implementation is in C++. Implementations in other languages are maintained by respective authors and have not been reviewed by me (@kentonv). Below are the implementations I’m aware of. Some of these projects are more “ready” than others; please consult each project’s documentation for details.

Serialization + RPC
Serialization only
Tools

These are other misc projects related to Cap’n Proto that are not actually implementations in new languages.

Contribute Your Own!

We’d like to support many more languages in the future!

If you’d like to own the implementation of Cap’n Proto in some particular language, let us know!

You should e-mail the list before you start hacking. We don’t bite, and we’ll probably have useful tips that will save you time. :)

Do not implement your own schema parser. The schema language is more complicated than it looks, and the algorithm to determine offsets of fields is subtle. If you reuse the official parser, you won’t risk getting these wrong, and you won’t have to spend time keeping your parser up-to-date. In fact, you can still write your code generator in any language you want, using compiler plugins!

How to Write Compiler Plugins

The Cap’n Proto tool, capnp, does not actually know how to generate code. It only parses schemas, then hands the parse tree off to another binary – known as a “plugin” – which generates the code. Plugins are independent executables (written in any language) which read a description of the schema from standard input and then generate the necessary code. The description is itself a Cap’n Proto message, defined by schema.capnp. Specifically, the plugin receives a CodeGeneratorRequest, using standard serialization (not packed). (Note that installing the C++ runtime causes schema.capnp to be placed in $PREFIX/include/capnp/usr/local/include/capnp by default).

Of course, because the input to a plugin is itself in Cap’n Proto format, if you write your plugin directly in the language you wish to support, you may have a bootstrapping problem: you somehow need to generate code for schema.capnp before you write your code generator. Luckily, because of the simplicity of the Cap’n Proto format, it is generally not too hard to do this by hand. Remember that you can use capnp compile -ocapnp schema.capnp to get a dump of the sizes and offsets of all structs and fields defined in the file.

capnp compile normally looks for plugins in $PATH with the name capnpc-[language], e.g. capnpc-c++ or capnpc-capnp. However, if the language name given on the command line contains a slash character, capnp assumes that it is an exact path to the plugin executable, and does not search $PATH. Examples:

# Searches $PATH for executable "capnpc-mylang".
capnp compile -o mylang addressbook.capnp

# Uses plugin executable "myplugin" from the current directory.
capnp compile -o ./myplugin addressbook.capnp

If the user specifies an output directory, the compiler will run the plugin with that directory as the working directory, so you do not need to worry about this.

For examples of plugins, take a look at capnpc-capnp or capnpc-c++.

Supporting Dynamic Languages

Dynamic languages have no compile step. This makes it difficult to work capnp compile into the workflow for such languages. Additionally, dynamic languages are often scripting languages that do not support pointer arithmetic or any reasonably-performant alternative.

Fortunately, dynamic languages usually have facilities for calling native code. The best way to support Cap’n Proto in a dynamic language, then, is to wrap the C++ library, in particular the C++ dynamic API. This way you get reasonable performance while still avoiding the need to generate any code specific to each schema.

To parse the schema files, use the capnp::SchemaParser class (defined in capnp/schema-parser.h). This way, schemas are loaded at the same time as all the rest of the program’s code – at startup. An advanced implementation might consider caching the compiled schemas in binary format, then loading the cached version using capnp::SchemaLoader, similar to the way e.g. Python caches compiled source files as .pyc bytecode, but that’s up to you.

Testing Your Implementation

The easiest way to test that you’ve implemented the spec correctly is to use the capnp tool to encode test inputs and decode outputs.