Meta-monomorphizing Specializations

This repository contains the artifact for the paper "Meta-monomorphizing Specializations". The artifact consists of three main components: a prototype implementation, an instrumentation tool for empirical analysis, and a benchmark evaluation suite.

---

Table of Contents

• Overview
• Repository Structure
• Requirements
• Getting Started
  • Prototype Implementation (spec-trait-impl)
  • Instrumentation Tool (spec-trait-inst)
  • Evaluation Benchmarks (spec-trait-evaluation)
• Reproducing the Empirical Evaluation
• Pre-computed Results
• Contact

---

Overview

This artifact provides:

1. spec-trait-impl/ — A prototype implementation of trait specialization in Rust, featuring procedural macros and compile-time trait ordering.

2. spec-trait-inst/ — An instrumentation tool built as a cargo/rustc wrapper that statically analyzes Rust projects to identify trait specialization opportunities.

3. spec-trait-evaluation/ — A benchmark suite that compares spec!/#[when] compile-time specialization against alternative dispatch strategies (naive, enum, runtime TypeId), measuring runtime performance, build time, and binary size. See the evaluation README for details.

---

Repository Structure

spec-trait/
├── spec-trait-impl/              # Prototype implementation
│   └── crates/
│       ├── spec-trait-bin/       # Binary crate (entry point)
│       ├── spec-trait-macro/     # Procedural macros for specialization
│       ├── spec-trait-order/     # Compile-time trait ordering
│       └── spec-trait-utils/     # Shared utilities
│
├── spec-trait-inst/              # Instrumentation tool
│   ├── src/                      # Source code
│   └── tests/
│
└── spec-trait-evaluation/        # Benchmark evaluation suite
    ├── benches-crate/            # Criterion benchmarks
    ├── notebooks/                # Analysis notebooks
    ├── results/                  # Pre-computed CSV results
    └── scripts/                  # Metric collection scripts
        ├── scripts/              # Analysis scripts and results
        │   ├── analysis.sh       # Main analysis script
        │   ├── repos.txt         # List of repositories analyzed
        │   ├── summary-90.csv    # Pre-computed results (90th percentile)
        │   └── summary-99.csv    # Pre-computed results (99th percentile)
        ├── snippets/             # Test snippets
        └── workspaces/           # Test workspaces

---

Requirements

• Rust nightly toolchain (nightly-2026-01-01)
• Git (for cloning repositories during analysis)
• Bash (for running the analysis script)

---

Getting Started

Prototype Implementation (spec-trait-impl)

The prototype implementation showcases our approach to trait specialization through procedural macros.

Build

cd spec-trait-impl
cargo build

Run Tests

cd spec-trait-impl
cargo test

---

Instrumentation Tool (spec-trait-inst)

The instrumentation tool analyzes Rust projects to identify existing and potential trait specialization patterns.

Setup

First, install the required nightly toolchain with all necessary components:

rustup toolchain install nightly-2026-01-01
rustup component add --toolchain nightly-2026-01-01 \
    rust-src rustc-dev llvm-tools-preview miri rust-analyzer clippy

Build

cd spec-trait-inst
cargo build

Run Tests

Execute the test suite on all example workspaces:

cd spec-trait-inst
cargo test --no-fail-fast -- --test-threads=1 --nocapture

Running on a Single Project

To analyze a specific Rust project:

cd /path/to/target/project

# On Linux:
LD_LIBRARY_PATH=$(rustc --print sysroot)/lib \
    /path/to/spec-trait-inst/target/debug/cargo-spec-trait-inst

# On macOS:
DYLD_LIBRARY_PATH=$(rustc --print sysroot)/lib \
    /path/to/spec-trait-inst/target/debug/cargo-spec-trait-inst

Note: Additional logging can be enabled by setting RUST_LOG=debug or RUST_LOG=trace. Use RUST_LOG_STYLE=always to force colored output.

---

Reproducing the Empirical Evaluation

The empirical evaluation from the paper can be fully reproduced using the provided analysis script. This script clones repositories from a curated list, runs the instrumentation tool on each, and aggregates the results.

Running the Analysis

cd spec-trait-inst/tests
./scripts/analysis.sh

Script Options

Usage: analysis.sh [-f <file>] [-w <workdir>] [-l <logs>] [-k] [-c] [-h]
  -f  File with newline-separated repo URLs (default: ./scripts/repos.txt)
  -w  Working directory for cloned repos (default: ./workspaces)
  -l  Logs directory (default: ./logs)
  -c  Continue mode: append to existing summary file
  -k  Keep cloned repositories after analysis
  -h  Show help message

Example: Analyze a Custom Set of Repositories

cd spec-trait-inst/tests
./scripts/analysis.sh -f /path/to/custom-repos.txt -k

---

Pre-computed Results

For convenience, we provide pre-computed analysis results in the spec-trait-inst/tests/scripts/ directory:

File	Description
summary-90.csv	Analysis results with 90th percentile threshold
summary-99.csv	Analysis results with 99th percentile threshold
repos.txt	List of analyzed repositories (top Rust crates from crates.io and popular GitHub projects)

These results were generated by running analysis.sh on the repositories listed in repos.txt. The CSV files contain detailed metrics including:

• Repository metadata (crate name, downloads, stars)
• Function and trait counts
• Specialization opportunity metrics (existing and potential)
• Analysis performance data (time, memory usage)