trlc 2.0.4

Treat Requirements Like Code

Treat Requirements Like Code (TRLC)

TRLC is a domain-specific language developed at BMW for writing (and linking) requirements with meta-data.

The repository contains:

• The language reference manual for TRLC. Previous versions are also available.

• A pure Python reference implementation of TRLC.

The implementation is not very fast, but designed to be pedantically correct in following the language definition. The tools also include a powerful static analysis tool to find issues with types and user-defined check rules.

The Python implementation can be used for several purposes:

• It can be used to validate other TRLC implementations.

• It can be used to validate a body of requirements (e.g. a CI check that all requirements are well formed)

• The API can be used to write other tools based on TRLC (for example a tool to render the requirements in HTML, a tool to diff requirements or perform an impact analysis, or a tool to perform software traceability, etc.)

Documentation

For normal users

• Tutorial (read this as a first introduction)
• Bazel integration (how to use TRLC inside a Bazel build)
• User manual: TRLC linter (the user manual for the TRLC static analysis and linter)
• Release Notes (read this to find out whats new)
• License

For advanced users

• Language Reference Manual (for language lawyers)
• Python API Documentation (API reference for end-users)
• AST Hierarchy (overview over all classes in the AST)

For TRLC developers

• Set up development environment
• Lexer/Parser Hierarchy (overview over all classes releated to the lexing and parsing of TRLC)
• Tool Architecture Overview
• TRLC Static Checker Slides (Kinda incomplete, designed to go along with a code walkthrough)
• Requirements Coverage Report
• Code Coverage Report

Tools Available

• TRLC_RST Convert TRLC Requirements for Sphinx Build

Dependencies

Run-time

• 3.8 <= Python3 <= 3.14
• PyVCG
• PyPI CVC5 (required when using the --verify option)

Optional dependency (not installed automatically):

• Binary CVC5 (An alternative to PyPI CVC5, make sure to rename the binary to cvc5 and put it on your PATH).