package qcheck

  1. Overview
  2. Docs
Module type
Class type
include module type of struct include QCheck_base_runner end

Runners for Tests

Once you built some tests using QCheck2.Test.make, you need to run the tests. This module contains several runners, which are designed to run every test and report the result.

By default, you can use run_tests in a test program as follows:

let testsuite = [
  Test.make ...;
  Test.make ...;

let () =
  let errcode = QCheck_base_runner.run_tests ~verbose:true testsuite in
  exit errcode

which will run the tests, and exit the program. The error code will be 0 if all tests pass, 1 otherwise.

run_tests_main can be used as a shortcut for that, also featuring command-line parsing (using Arg) to activate verbose mode and others.


val random_state : unit -> Random.State.t

Access the current random state

val verbose : unit -> bool

Is the default mode verbose or quiet?

val long_tests : unit -> bool

Is the default mode to run long tests or nor?

val set_seed : int -> unit

Change the random_state by creating a new one, initialized with the given seed.

val set_verbose : bool -> unit

Change the value of verbose ()

val set_long_tests : bool -> unit

Change the value of long_tests ()

val get_time_between_msg : unit -> float

Get the minimum time to wait between printing messages.

  • since 0.9
val set_time_between_msg : float -> unit

Set the minimum tiem between messages.

  • since 0.9

Event handlers

type counter = private QCheck_base_runner.counter = {
  1. start : float;
  2. expected : int;
  3. mutable gen : int;
  4. mutable passed : int;
  5. mutable failed : int;
  6. mutable errored : int;

The type of counter used to keep tracks of the events received for a given test cell.

type handler = QCheck_base_runner.handler = {
  1. handler : 'a. 'a QCheck2.Test.handler;

A type to represent polymorphic-enough handlers for test cells.

type handler_gen = colors:bool -> debug_shrink:out_channel option -> debug_shrink_list:string list -> size:int -> out:out_channel -> verbose:bool -> counter -> handler

An alias type to a generator of handlers for test cells.

val default_handler : handler_gen

The default handler used.

val debug_shrinking_choices : colors:bool -> out:out_channel -> name:string -> 'a QCheck2.Test.cell -> step:int -> 'a -> unit

The function used by the default handler to debug shrinking choices. This can be useful to outside users trying to reproduce some of the base-runner behavior.

  • since 0.19

Run a Suite of Tests and Get Results

val run_tests : ?handler:handler_gen -> ?colors:bool -> ?verbose:bool -> ?long:bool -> ?debug_shrink:out_channel option -> ?debug_shrink_list:string list -> ?out:out_channel -> ?rand:Random.State.t -> QCheck2.Test.t list -> int

Run a suite of tests, and print its results. This is an heritage from the "qcheck" library.

  • returns

    an error code, 0 if all tests passed, 1 otherwise.

  • parameter colors

    if true (default), colorful output

  • parameter verbose

    if true, prints more information about test cases (default: false)

  • parameter long

    if true, runs the long versions of the tests (default: false)

  • parameter debug_shrink

    debug_shrink:(Some ch) writes a log of successful shrink attempts to channel ch, for example ~debug_shrink:(Some (open_out "mylog.txt")). Use together with a non-empty list in ~debug_shrink_list.

  • parameter debug_shrink_list

    the test names to log successful shrink attempts for, for example ~debug_shrink_list:["list_rev_is_involutive"]. Requires ~debug_shrink to be Some ch.

  • parameter out

    print output to the provided channel (default: stdout)

  • parameter rand

    start the test runner in the provided RNG state

val run_tests_main : ?argv:string array -> QCheck2.Test.t list -> 'a

Can be used as the main function of a test file. Exits with a non-0 code if the tests fail. It refers to run_tests for actually running tests after CLI options have been parsed.

The available options are:

  • "--verbose" (or "-v") for activating verbose tests
  • "--seed <n>" (or "-s <n>") for repeating a previous run by setting the random seed
  • "--long" for running the long versions of the tests

Below is an example of the output of the run_tests and run_tests_main function:

random seed: 438308050
generated  error;  fail; pass / total -     time -- test name
[✓] (1000)    0 ;    0 ; 1000 / 1000 --     0.5s -- list_rev_is_involutive
[✗] (   1)    0 ;    1 ;    0 /   10 --     0.0s -- should_fail_sort_id
[✗] (   1)    1 ;    0 ;    0 /   10 --     0.0s -- should_error_raise_exn
[✓] (1000)    0 ;    0 ; 1000 / 1000 --     0.0s -- collect_results

--- Failure --------------------------------------------------------------------

Test should_fail_sort_id failed (11 shrink steps):

[1; 0]

=== Error ======================================================================

Test should_error_raise_exn errored on (62 shrink steps):


exception QCheck_runner_test.Error
Raised at file "example/", line 20, characters 20-25
Called from file "src/", line 839, characters 13-33

+++ Collect ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Collect results for test collect_results:

4: 207 cases
3: 190 cases
2: 219 cases
1: 196 cases
0: 188 cases

failure (1 tests failed, 1 tests errored, ran 4 tests)
include module type of struct include QCheck_ounit end

Conversion of tests to OUnit Tests

  • since 0.9
val to_ounit_test : ?verbose:bool -> ?long:bool -> ?rand:Random.State.t -> QCheck2.Test.t -> OUnit.test

to_ounit_test ~rand t wraps t into a OUnit test

  • parameter verbose

    used to print information on stdout (default: verbose())

  • parameter rand

    the random generator to use (default: random_state ())

val to_ounit_test_cell : ?verbose:bool -> ?long:bool -> ?rand:Random.State.t -> _ QCheck2.Test.cell -> OUnit.test

Same as to_ounit_test but with a polymorphic test cell

val (>:::) : string -> QCheck2.Test.t list -> OUnit.test

Same as OUnit.(>:::) but with a list of QCheck2 tests

val to_ounit2_test : ?rand:Random.State.t -> QCheck2.Test.t -> OUnit2.test

to_ounit2_test ?rand t wraps t into a OUnit2 test

  • parameter rand

    the random generator to use (default: a static seed for reproducibility), can be overridden with "-seed" on the command-line

val to_ounit2_test_list : ?rand:Random.State.t -> QCheck2.Test.t list -> OUnit2.test list

to_ounit2_test_list ?rand t like to_ounit2_test but for a list of tests

OUnit runners

QCheck provides some custom runners for OUnit tests.

Note that OUnit.run_test_tt or OUnit.run_test_tt_main can be used as well, in particular when QCheck tests are mixed with normal unit tests.

For OUnit2 you can use OUnit2.run_test_tt_main.

val run : ?argv:string array -> OUnit.test -> int

run test runs the test, and returns an error code that is 0 if all tests passed, 1 otherwise. This is the default runner used by the comment-to-test generator.

  • parameter argv

    the command line arguments to parse parameters from (default Sys.argv)

  • raises Arg.Bad

    in case argv contains unknown arguments

  • raises Arg.Help

    in case argv contains "--help"

    This test runner displays execution in a compact way, making it good for suites that have lots of tests.

    Output example:

    random seed: 101121210
    random seed: 101121210
    Error: tests>error_raise_exn
    test `error_raise_exn` raised exception `QCheck_ounit_test.Error`
    on `0 (after 62 shrink steps)`
    Raised at file "example/", line 19, characters 20-25
    Called from file "src/", line 846, characters 13-33
    Failure: tests>fail_sort_id
    Ran: 4 tests in: 0.74 seconds.

TAP-compatible test runner, in case we want to use a test harness. It prints one line per test.


Innovation. Community. Security.