TESTS

About testing in the hledger project, as of 201809.

Kinds of tests

"Here, then, is a list of properties of tests. Not all tests need to exhibit all properties. However, no property should be given up without receiving a property of greater value in return.
  • Isolated — tests should return the same results regardless of the order in which they are run.
  • Composable — if tests are isolated, then I can run 1 or 10 or 100 or 1,000,000 and get the same results.
  • Fast — tests should run quickly.
  • Inspiring — passing the tests should inspire confidence
  • Writable — tests should be cheap to write relative to the cost of the code being tested.
  • Readable — tests should be comprehensible for reader, invoking the motivation for writing this particular test.
  • Behavioral — tests should be sensitive to changes in the behavior of the code under test. If the behavior changes, the test result should change.
  • Structure-insensitive — tests should not change their result if the structure of the code changes.
  • Automated — tests should run without human intervention.
  • Specific — if a test fails, the cause of the failure should be obvious.
  • Deterministic — if nothing changes, the test result shouldn’t change.
  • Predictive — if the tests all pass, then the code under test should be suitable for production." --Kent Beck

  1. Unit tests

    Unit tests exercise small chunks of functionality. In hledger, that means a function. So, many of our functions have one or more unit tests. These are mostly in hledger-lib, with a few in hledger.

    Our unit tests use the tasty test runner, tasty-hunit HUnit-style tests, and some helpers from Hledger.Utils.Test, such as:

    • tests and test aliases for testGroup and testCase
    • assert* helpers for constructing various kinds of assertions

    We would like our unit tests to be:

    • easy to read (clear, concise)
    • easy to write (low boilerplate, low cognitive load)
    • easy to maintain (easy to edit, easy to refactor, robust)
    • easy to associate with the code under test (easy to view/jump between code & test, easy to estimate coverage)
    • and scalable (usable for all devs, easy to run and select, suitable for small/large modules/packages).

    Here's the current pattern (let us know if you see a better way):

    module Foo (
  ...
  tests_Foo -- export this module's and submodules' tests
)
where
import Hledger  -- provides Hledger.Utils.Test helpers
import Bar      -- submodules, providing tests_Bar etc.
import Baz

functionA = ...
functionB = ...
functionC = ...
functionD = ...

tests_Foo = tests "Foo" [ -- define tests at the end of each module

   -- a group of several named tests for functionA
   tests "functionA" [
     test "a basic test"           $ assertBool "" SOMEBOOL
    ,test "a pretty equality test" $ SOMEEXPR @?= EXPECTEDVALUE
    ,test "a pretty parsing test"  $ assertParseEq PARSER INPUT EXPECTEDRESULT
    ,test "a multiple assertions test" $ do
      A @?= B
      doSomeIO
      C @?= D
    ]

   -- a single test containing multiple unnamed assertions for functionB
  ,test "functionB" $ do
     assertBool "" BOOL
     EXPR @?= VALUE

  ,tests_Foo            -- aggregate submodule tests
  ,tests_Bar
  ]

    Here are some real-world examples.

    The unit tests are shipped as part of the hledger executable, and can always be run via the test command (hledger test).

  2. Doc tests

    Like unit tests, but defined inside functions' haddock documentation, in the style of a GHCI transcript. These test functionality, provide usage examples in the API docs, and test those examples, all at once. They are a bit more finicky and slower than unit tests. See doctest for more.

    doctests do not work on Mac with GHC 8.4+, out of the box. See ghc#15105 for current status and a workaround.

  3. Functional tests

    Functional tests test the overall functioning of the program. For hledger, that means running hledger with various inputs and options and checking for the expected output. This exercises functionality in the hledger and hledger-lib packages. We do this with shelltestrunner. Tests are defined in files named *.test under hledger/test/, grouped by component (command or topic name). For more about these, see the README there.

  4. Code tests

    We have some tests aimed at testing eg code quality, generally runnable via just. Eg just haddocktest, just hlinttest.

  5. Package test suites

    Haskell tools like stack and cabal recognise test suites defined in a package's cabal file (or package.yaml file). These can be run via stack test, cabal test etc., and they are required to build and pass by services like Stackage. Here are the currently hledger package test suites:

    package test suite what it runs hledger-lib doctests doctests hledger-lib easytests unit tests hledger test builtin test command (hledger's + hledger-lib's unit tests) hledger-ui
    hledger-web

Coverage

This means how thoroughly the code is tested - both in breadth (are all parts of the code tested at least a little ?) and in depth (are all possible code paths, states, situations tested ?).

Our current test coverage can be summarised like so:

package unit doc functional hledger-lib X X X hledger X X hledger-ui
hledger-web

There are ways to generate detailed coverage reports for haskell unit tests, at least. It would be useful to set this up for hledger.

How to run tests

Run unit tests:

$ just unittest

Run doctests:

$ just doctest

Run functional tests (and unit tests, now):

$ stack install shelltestrunner
$ just functest

Run the package tests (unit tests, maybe doctests, but not functional tests) of all or selected packages.

$ stack test [PKG]

Run "default tests: package plus functional tests":

$ just test

Test generation of haddock docs:

$ just haddocktest

Run built-in hledger/hledger-lib unit tests via hledger command:

$ hledger test  # test installed hledger
$ stack build hledger && stack exec -- hledger test  # test just-built hledger
$ hledger test --help
test [TESTPATTERN] [SEED]
  Run the unit tests built in to hledger-lib and hledger,
  printing results on stdout and exiting with success or failure.
  Tests are run in two batches: easytest-based and hunit-based tests.
  If any test fails or gives an error, the exit code will be non-zero.
  If a pattern argument (case sensitive) is provided, only easytests
  in that scope and only hunit tests whose name contains it are run.
  If a numeric second argument is provided, it will set the randomness
  seed for easytests.

Rebuild and rerun hledger/hledger-lib doc tests via ghcid:

$ just ghcid-doctest

See all test-related just rules:

$ just h test