Legend:
Library
Module
Module type
Parameter
Class
Class type
Library
Module
Module type
Parameter
Class
Class type
Base test functions.
Add a function to be called before each test start.
Used to reset counters such as the ones which are used to choose default process names.
Add a function to be called by Test.run
before it does anything.
val fail : ?__LOC__:string -> ('a, Format.formatter, unit, 'b) format4 -> 'a
Log an error and stop the test right here.
If the optional location __LOC__
is provided, it is prepended to the error message. You would typically use it simply as Test.fail ~__LOC__ "..."
to prepend location of the failure.
val register :
__FILE__:string ->
title:string ->
tags:string list ->
(unit -> unit Lwt.t) ->
unit
Register a test.
The __FILE__
argument, which should be equal to __FILE__
(i.e. just write Test.run ~__FILE__
), is used to let the user select which files to run from the command-line.
One should be able to infer, from title
, what the test will do. It is typically a short sentence like "addition is commutative"
or "server runs until client tells it to stop"
.
The list of tags
must be composed of short strings which are easy to type on the command line (lowercase letters, digits and underscores). Run the test executable with --list
to get the list of tags which are already used by existing tests. Try to reuse them if possible.
The last argument is a function f
which implements the test. If f
needs to spawn external processes, it should use the Process
module to do so, so that those processes are automatically killed at the end of the test. Similarly, if this function needs to create temporary files, it should declare them with Temp
, and if it needs to have promises that run in the background, it should use Background.register
(and not Lwt.async
in particular).
If f
raises an exception, act as if fail
was called (without the error location unfortunately).
You can call register
several times in the same executable if you want it to run several tests. Each of those tests should be standalone, as the user is able to specify the list of tests to run on the command-line.
The test is not actually run until you call run
.
Run registered tests that should be run.
Call this once you have registered all tests. This will check command-line options and run the tests that have been selected, or display the list of tests.
Get the current worker id.
In single-process mode (with -j 1
), this always returns None
.
In multi-process mode, this returns either:
None
(if called in the scheduler process, i.e. outside of a test);Some id
where 0 <= id < Cli.options.job_count
and where id
uniquely identifies the current worker process that is running the current test.