hyhac

A HyperDex Haskell Client

What is HyperDex? HyperDex.org describes it so:

HyperDex is a next generation key-value store with a wide array of features. HyperDex may be a good fit for you if your application needs to:

• Store rich datatypes such as strings, integers, floats, lists, maps, and sets.
• Search and retrieve objects efficiently by secondary attributes.
• Ensure that your data is up to date at all times with strong consistency guarantees.
• Protect your data against a configurable number of simultaneous failures.
• Perform atomic transactions over multiple objects, in a fast, scalable NoSQL data store.

HyperDex's key features -- namely its rich API, strong consistency, and fault tolerance -- provide strong guarantees to applications that are not matched by other NoSQL systems.

Authors

Aaron Friel Niklas Hambüchen

Building

This repository now builds with modern ghcup + cabal tooling. After installing GHCup and sourcing ~/.ghcup/env, use the checked-in wrapper so Cabal sees the HyperDex build tree and the local libgmp.so shim it may need on hosts without the libgmp-dev linker symlink:

scripts/cabal.sh build -f tests lib:hyhac test:tests

By default the wrapper expects the sibling HyperDex checkout at ../HyperDex. If your HyperDex build is somewhere else, set HYPERDEX_ROOT=/path/to/HyperDex before invoking the script.

Testing

The test suite depends on a local HyperDex coordinator and daemon. Run the deterministic wrapper below instead of starting daemons by hand:

scripts/test-with-hyperdex.sh

That wrapper:

• starts a private one-node HyperDex cluster in a temporary directory;
• exports the in-tree HyperDex runtime variables needed by an uninstalled build;
• waits for the daemon to show up as AVAILABLE before it runs Haskell tests;
• runs cabal test -f tests test:tests with --test-seed=1 for repeatable property-test input;
• shuts the cluster down automatically and removes temporary state on success.

You can override the fixed property-test seed with HYPERDEX_TEST_SEED=<n>.

For a fast deterministic smoke test that does not require a running HyperDex daemon, you can run just the pure CBString property group:

scripts/cabal.sh test -f tests test:tests \
  --test-show-details=direct \
  --test-option=--plain \
  --test-option=--test-seed=1 \
  --test-option=--select-tests='hyhac-tests/CBString API Tests*'

Automation

The isolated hyhac worktree also carries repo-local helper scripts under .agent/:

make -C .agent smoke
make -C .agent check

make -C .agent smoke runs the fast CBString subset. make -C .agent check builds hyhac, starts a temporary HyperDex cluster, and runs the full suite. If the sibling HyperDex checkout at ../HyperDex is not already built, build it first with:

sudo ../HyperDex/.agent/setup.sh

That setup path builds HyperDex in ../HyperDex/target, so the effective runtime root for hyhac is ../HyperDex/target. The .agent/check.sh helper detects that automatically, and you can also set it explicitly:

HYPERDEX_ROOT=../HyperDex/target make -C .agent check

To exercise the GitHub Actions workflow locally with nektos/act, run:

make -C .agent act-ci

If the HyperDex branch you need is not yet aaronfriel/HyperDex@master, set the workflow variables HYPERDEX_REPOSITORY and HYPERDEX_REF in GitHub, or pass them to act locally:

make -C .agent act-ci -- --var HYPERDEX_REPOSITORY=aaronfriel/HyperDex --var HYPERDEX_REF=master