# crap4rust Roadmap
This document tracks the planned evolution of `cargo-crap4rust` beyond the currently shipped `0.3.x` release.
For what is available today, see [IMPLEMENTED-FEATURES.md](IMPLEMENTED-FEATURES.md). For released versions, see [CHANGELOG.md](CHANGELOG.md).
## Product Direction
crap4rust aims to be a practical Rust-native CRAP analysis tool that identifies functions that are both complex and under-tested.
The long-term direction is:
- strong Rust-aware complexity analysis
- low-friction Cargo integration
- useful local and CI reporting
- stable machine-readable outputs
- a reusable library surface for embedding and automation
## Guiding Principles
- Prefer cognitive complexity as the default Rust-facing metric
- Keep the first-class workflow centered on Cargo projects and workspaces
- Avoid noise from generated code, macro expansion, and compiler artifacts
- Keep reports actionable, not just descriptive
- Add ecosystem integrations only after the core ranking remains trustworthy on real Rust codebases
## Current Baseline
The shipped `0.3.x` line already provides:
- `cargo crap4rust` as a published Cargo subcommand
- per-function CRAP scoring
- internal cognitive-complexity scoring
- automatic `cargo llvm-cov` JSON generation when needed
- explicit coverage-file input
- multi-package workspace aggregation
- console reporting and threshold-based exit codes
That baseline is intentionally small. The roadmap below is about what comes next, not what is already complete.
## Planned Phases
### Phase 2: Richer Reporting
Goal: make local output more useful without changing the core analysis model.
Planned scope:
- JSON output for automation and downstream tooling
- HTML and Markdown reports for people-facing consumption
- output-file support instead of stdout-only reporting
- sorting, top-N filtering, and optional clean-function display
- clearer project summaries and report metadata
Exit criteria:
- one machine-readable format is stable enough for CI consumption
- one human-readable file format is useful enough to archive or share
### Phase 3: CI and Baselines
Goal: make crap4rust usable as a gating tool in automated pipelines.
Planned scope:
- baseline files for score comparison over time
- regression detection
- SARIF output for code-scanning workflows
- documented GitHub Actions examples
- stable non-zero exits for regression-specific failures
Exit criteria:
- a team can use crap4rust in CI to fail on CRAP regressions, not just absolute thresholds
### Phase 4: Configuration and Policy
Goal: let teams adopt the tool without passing every preference through the command line.
Planned scope:
- `crap4rust.toml` support
- per-workspace and per-crate defaults
- suppression rules with explicit reasons
- stricter policy controls around ignored functions
- optional risk modifiers such as `unsafe` weighting
Exit criteria:
- common project policy can live in versioned configuration rather than shell scripts
### Phase 5: Deeper Analysis Fidelity
Goal: improve the trustworthiness and flexibility of the analysis engine.
Planned scope:
- alternative complexity sources or engine integration
- alternative coverage-source support beyond `cargo llvm-cov` JSON
- better filtering for macros, tests, and generated code
- richer handling of closures and other Rust-specific units when justified
- versioned report schema and formula metadata where needed
Exit criteria:
- complexity and coverage inputs are flexible without making default usage harder
### Phase 6: Public API and Ecosystem Integrations
Goal: make crap4rust useful as a component, not only as a command.
Planned scope:
- a stable public library API
- formatter and provider extension points where they prove necessary
- editor and ecosystem integrations such as VS Code or code-scanning adapters
- optional compatibility outputs for external dashboards or legacy systems
Exit criteria:
- third-party tools can embed crap4rust without shelling out to the CLI for every workflow
## Deferred Ideas
These ideas are intentionally not prioritized until the core roadmap is further along:
- broad plugin architecture before the core data model stabilizes
- niche output formats before JSON and one strong human-readable file format land
- advanced policy systems before basic configuration exists
- ecosystem integrations that outrun the stability of the core analysis engine
## Success Measure
The roadmap is succeeding if each phase improves one of these outcomes:
- more trustworthy ranking of risky functions
- easier adoption in local development and CI
- clearer automation and reporting surfaces
- broader reuse without destabilizing the default CLI workflow
## Revision Policy
This roadmap is directional, not contractual. Phases may be reordered or narrowed if real-world use shows that a smaller, sharper scope is the better engineering decision.
