Cache

Karva keeps a small on-disk cache of previous test runs. Today it powers two things: per-test duration history (used to schedule the slowest tests first under parallelism) and the list of tests that failed in the last run.

The cache lives under the platform cache directory, namespaced per project root.

Re-running just the failures

--last-failed (or --lf) restricts the run to whichever tests failed in the previous invocation:

1

karva test --last-failed

A typical fix-it-up loop:

1
2
3

karva test                # see the failures
karva test --last-failed  # iterate on just those
karva test                # confirm the full suite passes again

Combine with --watch to keep iterating until they all pass:

1

karva test --watch --last-failed

If the last run had no failures, --last-failed runs nothing.

Disabling cache reads

--no-cache disables reading the cache for the current run. Tests are scheduled without duration hints and --last-failed becomes a no-op. Cache files are still written so subsequent runs without --no-cache have fresh data.

1

karva test --no-cache

Managing the cache

Two karva cache subcommands manage cache contents directly:

1
2

karva cache prune  # keep only the most recent run
karva cache clean  # remove the cache directory entirely

prune is the safer of the two — it reclaims space without losing the data the next --last-failed would use. Reach for clean if the cache gets corrupted, or after upgrading karva across a cache-format change.