Typeshare does not perform the bulk of the business logic and code generation during the compile-time stage as a procedural macro would normally do, but instead analyzes the code via a separately-invoked command line tool. This analytical process is not ideal as it currently stands, using many workarounds and homemade shortcuts that may not always hold true in more complex scenarios. Furthermore, this haphazard approach to parsing has lead to an inconsistent standard format for configuration within the macro - for example, using name-value statements for defining language-specific auto-derives, even though that will make it difficult to add other language-specific options in the future.

This project seeks to build a more robust parser for Typeshare's attribute system, and creating a better foundation that new functionality can be built on top of.