Two Blog Posts about using Hypothesis

[Ian Hunt-Isaak]

Hi all,

Deepak Cherian and I both love using Hypothesis to test one of the major open source project we work on (https://icechunk.io/en/stable/overview). We recently wrote some blog posts diving into some fun technical details of how we are using Hypothesis. So I figured I'd share here

Post 1

TLDR:

How we used hypothesis property and stateful testing to move really quickly building a new data format for nD arrays

https://www.earthmover.io/blog/engineering-rigor-in-icechunk-part-1

Post 2

TLDR:

In a single Python process, Hypothesis drives a randomized interleaving of Icechunk 1.x and 2.x calls against the same on-disk repo, and shrinks any failure down to its minimal reproducer

We extend the same trick of running two versions in one process to give us a second cross-version test for garbage collection, where we use Icechunk 1.x itself as the reference model

https://www.earthmover.io/blog/engineering-rigor-in-icechunk-part-3

Happy reading,

Ian