Software Engineering Radio - the podcast for professional software developers avatar

Dexa/Software Engineering Radio - the podcast for professional software developers

Learn more
Published Jun 26, 2024

SE Radio 622: Wolf Vollprecht on Python Tooling in Rust

Wolf Vollprecht delves into the intersection of Rust and Python tooling, showcasing the memory safety and performance benefits of Rust, while navigating its compilation challenges. The episode emphasizes innovative solutions for package management, cross-platform binaries, and reproducible builds, particularly spotlighting the PKCE package manager for enhanced infrastructure tooling and compatibility.
Episode Highlights
Software Engineering Radio - the podcast for professional software developers logo

Topics covered

Popular Clips

Episode Highlights

  • Cross-Platform

    Cross-platform binaries are essential for developers aiming to create software that operates seamlessly across different operating systems. explains that tools like Rattler build enable developers to create relocatable binaries that work on Windows, Mac, and Linux, using specific tricks like Patch Elf on Linux and Installnametool on macOS 1. These methods ensure that packages can be installed anywhere on a system, unlike traditional packages that require specific locations.

    Cross platform in that sense that Redlabilt works on all the different platforms that we support. So Windows, Mac and Linux.

    ---

    Additionally, Vollprecht highlights the use of Rust in building tools like PKCE, which leverages existing libraries to manage workflows effectively across platforms 2.

       

    Reproducibility

    Reproducible builds are crucial for ensuring consistency and reliability in software development. emphasizes the importance of reproducibility in the scientific community, where tools like PKCE ensure that results can be replicated even years later by locking down packages to specific versions 3. This is achieved through tight log files and the extensive Conda Forge ecosystem, which retains all package versions.

    You can reproduce this exact program on your computer if you compile something, etcetera.

    ---

    Furthermore, PKCE packages not only Python but also other programming languages, ensuring a comprehensive and isolated build environment 4.

       

    Dependencies

    Managing dependencies effectively is a significant challenge in software development, especially in Python ecosystems. discusses how tools like Mamba have improved the speed of dependency resolution in Conda by integrating faster SAT solvers 5. This enhancement addresses performance issues that arise from the exponential growth of package repositories.

    Conda became really, really slow, unbearable even.

    ---

    Additionally, Vollprecht notes the challenges faced by developers when using package managers like Poetry, which do not bundle Python or compilers, making it difficult to ensure compatibility across different systems 6.

Related Episodes