Skip to main content

Internal Schema

We have 2 schemas.

The first thing that you need to know when developing for Anvil is that the schema that Anvil uses internally is completely different than the one that the users write (.anv files).

While the schema for written by users is focused on:

  • Being easily understandable for humans
  • Being able to share knowledge to other humans (trough comments)
  • Being easy and lean to write
  • Being divisible in multiple files
  • Organized in the way that you architect your API

The internal schema that Anvil uses is focused on:

  • Being performative to parse by the machine
  • Having everything that you need in one single file
  • Already have all the references resolved
  • Most performative as possible to access all the values that you may need
  • Organized on the way that your program will generate things

To identify the difference between processed / internal schemas and "user-friendly" schemas, we have 2 different types of files:

  • .anv: user friendly schema, the one described here.
  • .anvp: processed schema, made by the machine for the machine

You will never touch a .anvp file, but Anvil will always use this format to communicate to your generator.

When and how does Anvil CLI communicates with the generator?​

Anvil CLI calls your generator when the configuration says so.

Anvil CLI uses stdout, a decision inspired by LSPs. All the information is formatted in JSON string.