Port to Rust #704

Implementing liblouis in C has many benefits:

• runs basically on any platform from desktop to embed system
• enables bindings to many languages

on the other hand it has a number of downsides:

• real cross platform is tricky. The autotools build system isn't really native to windows for example
• the standard libs are poor, so we have home grown implementations of hash maps, lists, a hand written parser, etc.
• the manual memory management leads to quite a few memory leaks and probably buffer overruns

This is why the idea of a port of liblouis to a modern language has been around for some time. But in order to keep the good parts we need to be able to provide a C ABI.

That is where Rust comes in:

• it is a modern language
  • with proper data structures, collections, better error handling, etc
• cross platform build seems easy
• it can export and consume a C ABI
• it can be linked to C code

Now rewriting a piece of software such as liblouis seems like a daunting task. That is where the idea of a gradual port comes into play. Rewrite it piece by piece.

Some people have apparently done this successfully for other legacy C code bases:

• Porting Zopfli from Rust to C
• Federico's blog on a gradual port of librsvg
  • slides from a talk
• Working Effectively with Legacy Code by Michael Feathers
• https://c2rust.com/