Lavisys/mailsweep
2
0
Fork 0
Code Issues 2 Pull requests Projects Releases 7 Packages Wiki Activity Actions
Page: ADR 0009 retire python baseline and complete rust cutover
Pages
0001 deepen classification safety auditor 0002 deepen rules registry 0003 prune legacy script shims 0004 decouple rfc codecs from mailbox gateway 0005 collapse workflow into mailbox pipeline 0006 decouple gateway orchestration into pipeline 0007 incremental scan cache and progress feedback 0008 port mailsweep to synchronous rust 0009 retire python baseline and complete rust cutover 0010 automated multi platform forgejo release workflow ADR 0001 deepen classification safety auditor ADR 0002 deepen rules registry ADR 0003 prune legacy script shims ADR 0004 decouple rfc codecs from mailbox gateway ADR 0005 collapse workflow into mailbox pipeline ADR 0006 decouple gateway orchestration into pipeline ADR 0007 incremental scan cache and progress feedback ADR 0008 port mailsweep to synchronous rust ADR 0009 retire python baseline and complete rust cutover ADR 0010 automated multi platform forgejo release workflow ADRs CLI Reference Domain Context FAQ Home Internals Python API Rules Schema Rust API
1 ADR 0009 retire python baseline and complete rust cutover
MailSweep Doc Bot edited this page 2026-08-30 16:44:17 +00:00
Table of Contents

9. Retire Python Baseline and Complete Rust Cutover

Date: 2026-08-30

Status

Accepted

Context

Following the implementation of ADR-0008, MailSweep was ported to a high-performance synchronous Rust crate architecture.

During the migration phase, the legacy Python implementation (src/mailsweep/, test_mailsweep.py, scripts/forgejo.py, and pyproject.toml) was retained alongside the Rust codebase as an oracle and golden parity verification suite.

With 100% parity verified across all 66 Rust integration and unit tests (covering 13-language categorization heuristics, RFC MIME codecs, IMAP gateway, transactional attachment inspection, backup archives with SHA-256 verification, safety auditor, and Forgejo issue tracker tooling), the Python baseline has completed its migration lifecycle.

Retaining Python sources introduces unnecessary maintenance overhead, dual-stack dependency noise (uv.lock, pyproject.toml, __pycache__), and documentation ambiguity.

Decision

We have retired all Python code and standardized the repository entirely on native Rust:

  1. Source Code Cleanup:

    • Removed src/mailsweep/ directory and all Python domain modules.
    • Removed legacy test runner test_mailsweep.py.
    • Removed scripts/forgejo.py in favor of the compiled Rust binary src/bin/forgejo.rs.

  2. Packaging and Dependency Cleanup:

    • Removed Python packaging configurations (pyproject.toml, uv.lock).
    • Cleaned up Python caches and virtual environment artifacts.

  3. Documentation & Tooling Standardization:

    • Updated documentation, CLI references, and examples to use cargo run -- or the compiled mailsweep binary.
    • Replaced docs/python-api.md with docs/rust-api.md documenting the native Rust crate interface.
    • Updated Wiki synchronization workflow (.forgejo/workflows/wiki-sync.yaml) and documentation index.

Consequences

  • Delivers a pure, single-language Rust repository with zero external runtime or Python dependencies.
  • Eliminates dual-maintenance overhead and simplifies the build, test, and CI lifecycle to standard cargo workflows.
  • Retains full backward-compatible JSON schema interoperability for scan summaries, cleanup plans, and rulesets.
  • Preserves all 13-language classification heuristics, safety tier hierarchies, and backup guarantees.

📬 MailSweep Wiki