Partitioning test runs in CI

cargo-nextest supports both hash-based and count-based partitioning. Here’s an example of count-based partitioning, running tests in bucket 1 of 2.

Tests not in the current bucket are marked skipped.

Hash-based partitioning is similar, except buckets are specified in the format , where is the current bucket and is the number of buckets. The hash is completely deterministic, and is based on a combination of the binary and test names. For sufficiently large test runs, hash-based partitioning produces roughly the same number of tests per bucket.