Pegium's Workflow

Pegium's workflow can be summarized as a short sequence of practical steps:

1. build the repository
2. choose a starting point
3. write the grammar
4. shape the AST and CST
5. resolve cross-references
6. create validations
7. add formatting and LSP services

The exact needs of a real project may go beyond this sequence, but this is the common path we recommend for building a language with Pegium. For advanced topics and one-off customization tasks, continue with the recipes.

Explanation

This workflow breaks down into three parts:

• the initial setup, which you normally do once
• the core language work, which you revisit as the grammar evolves
• advanced customization, which is usually better handled through the recipes

Initial setup

1. Build the Repository

Build Pegium itself, confirm the examples compile, and make sure the test suite is available locally.

2. Choose a Starting Point

Pick the example whose structure is closest to your future language instead of starting from an empty directory.

Core workflow

3. Write the Grammar

Define the surface syntax with PegiumParser, terminals, rules, skippers, and parser expressions.

4. Shape the AST and CST

Use assignments, actions, and CST-backed parsing to build the semantic and source-aware structure of your language.

5. Resolve Cross-References

Export names, compute scopes, and link references so that navigation and workspace features have real semantic targets.

6. Create Validations

Add semantic rules once the AST and references are stable enough to reason about the model.

7. Add Formatting and LSP Services

Finish the user-facing tooling by wiring formatter, hover, completion, rename, and workspace behavior where needed.

Advanced topics

Everything beyond the common workflow is better handled through the recipes and reference pages.