ICSE 2026 Artifact Submission

Decades of GNU Patch and Git Cherry-Pick: Can We Do Better?

This artifact contains the reproduction package for our paper Decades of GNU Patch and Git Cherry-Pick: Can We Do Better? which has been accepted to the 48th International Conference on Software Engineering (ICSE 2026).
The artifact is licensed under a dual MIT - Apache 2.0 license.

Purpose

The artifact should allow other researchers to reproduce our experiments and confirm the results which reported in our paper.
We provide instructions that allow reproduction of the evaluation presented in Sections 4 through 6 of our paper.
The reproduction is executed in a Docker container.

Provenance

The Preprint of our paper can be found online.
After official publication, it becomes available under https://doi.org/10.1145/3744916.3764537.

The artifact itself is available on Zenodo and Github.

Content

The reproduction package consists of three main parts:

1. mpatch: The implementation of our novel match-based patcher, written in Rust.
2. Mined cherries: Our dataset of cherry picks mined from 5,000 GitHub repositories.
3. Empirical evaluation: Our empirical evaluation of various language-agnostic patchers.

Obtaining the artifact

The easiest way obtain the artifact is to clone the artifact from GitHub:

git clone https://github.com/VariantSync/patching-with-matching-eval.git

Note

For more information, see the README.md file in the root folder of the reproduction package.

ICSE 2026 Camera Ready

Decades of GNU Patch and Git Cherry-Pick: Can We Do Better?

This is the initial release of our reproduction package for our paper Decades of GNU Patch and Git Cherry-Pick: Can We Do Better? which has been accepted to the 48th International Conference on Software Engineering (ICSE 2026)

Content

The reproduction package consists of three main parts:

1. mpatch: The implementation of our novel match-based patcher, written in Rust.
2. Mined cherries: Our dataset of cherry picks mined from 5,000 GitHub repositories.
3. Empirical evaluation: Our empirical evaluation of different language-agnostic patchers.

Note

For more information, see the README.md file in the root folder of the reproduction package.