streaming-crypto

A Rust cryptographic library with optional FFI and Python bindings.

Features

• Rust API (default): Pure Rust functions.
• FFI API: C ABI functions for integration with C/C++ and other languages.
• PyO3 API: Python bindings, installable via pip.

Usage

Rust API

cargo add streaming-crypto

FFI API

cargo build --features ffi-api

Python API

pip install streaming-crypto

Documentation

• API Docs on docs.rs
• Crate on crates.io
• PyPI package

✅ Badge Breakdown

• CI badge → shows GitHub Actions build/test status.
• Crates.io badge → shows latest published version on crates.io.
• Docs.rs badge → links to auto‑generated Rust documentation.
• PyPI badge → shows latest published version on PyPI.
• License badge → signals open‑source license clearly.

1. Repository setup

We need to create the GitHub repository in advance on GitHub.com.

• Push our initial codebase there.
• In the repo settings, add our secrets (CARGO_REGISTRY_TOKEN, PYPI_API_TOKEN) under Settings → Secrets and variables → Actions.
• This ensures our workflow has the environment it needs when we eventually tag a release.

2. Controlling publish triggers

Our workflow is configured to run on:

on:
  push:
    tags:
      - 'v*.*.*'

That means only pushes with a tag like v0.1.0 or v1.2.3 will trigger publish.

• If we just push commits to main (without tags), the publish job won’t run.
• Our CI workflow (build/test) runs on push and pull_request to main, but publish is tag‑only.

3. Recommended flow

• Step 1: Create the repo on GitHub.
• Step 2: Push our initial codebase (no tags yet). This will run CI (build/test) but not publish.
• Step 3: Configure secrets (CARGO_REGISTRY_TOKEN, PYPI_API_TOKEN).
• Step 4: Once CI is green and we’re ready for a release, create a version tag locally:

git tag v0.1.0
git push origin v0.1.0

• Step 5: That tag push will trigger the publish workflow, which will publish to crates.io and PyPI.

Set environment python for test

1. Before: PyO3 was trying to link against Python 3.13 because either a cached build or environment variable made it think our Python was 3.13.

2. Action: We forced PyO3 to use Python 3.12 or current version of pyenv explicitly:

   export PYTHON_SYS_EXECUTABLE="$(pyenv which python3)"
   export PYO3_PYTHON="$(pyenv which python3)"
   export PYO3_NO_PYTHON_LINK=1
   cargo clean
   

3. Result: Cargo rebuilds PyO3 and all dependent crates from scratch. Now cargo run --bin check_python correctly detects Python 3.12.12 and does not crash on missing libpython3.13.dylib.

4. cargo test will work now