srfi-64.html

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html lang="en">
  <head>
    <title>SRFI 64: A Scheme API for test suites</title>
    <meta charset="utf-8" />
    <meta content="width=device-width, initial-scale=1" name="viewport" />
    <link href="https://srfi.schemers.org/srfi.css" rel="stylesheet" type="text/css" />
    <link href="/favicon.png" rel="icon" sizes="192x192" type="image/png" />
<style type="text/css">
  /* So var inside pre gets same font as var in paragraphs. */
  var { font-family: monospace; }
</style>
  </head>

<body>
<h1><a href="https://srfi.schemers.org/"><img alt="SRFI surfboard logo" class="srfi-logo" src="https://srfi.schemers.org/srfi-logo.svg" /></a>64: A Scheme API for test suites</h1>

<p>by Per Bothner<br />
<code><a href="mailto:per@bothner.com">&lt;per@bothner.com&gt;</a></code></p>

<h2>Status</h2>

<p>This SRFI is currently in <em>final</em> status.  Here is <a href="https://srfi.schemers.org/srfi-process.html">an explanation</a> of each status that a SRFI can hold.  To provide input on this SRFI, please send email to <code><a href="mailto:srfi+minus+64%20%20+at+srfi+dotschemers+dot+org">srfi-64  @<span class="antispam">nospam</span>srfi.schemers.org</a></code>.  To subscribe to the list, follow <a href="https://srfi.schemers.org/srfi-list-subscribe.html">these instructions</a>.  You can access previous messages via the mailing list <a href="https://srfi-email.schemers.org/srfi-64">archive</a>.</p>
<ul>
  <li>Received: 2005-01-07</li>
  <li>Draft: 2005-01-28--2005-03-28</li>
  <li>Revised: <a href="https://srfi.schemers.org/srfi-64/srfi-64-1.3.html">2005-10-18</a>  
  </li><li>Revised: <a href="https://srfi.schemers.org/srfi-64/srfi-64-1.5.html">2006-02-24</a>  
  </li><li>Final: <a href="https://srfi.schemers.org/srfi-64/srfi-64-1.6.html">2006-06-18</a></li>
  <li>Revised to fix errata: 2016-08-11</li>
  <li>Revised to fix links and to refer to RackUnit, the new name of SchemeUnit: 2023-11-29</li>
</ul>


<h2>Abstract</h2>
<p>
This defines an API for writing <dfn>test suites</dfn>, to make it easy
to portably test Scheme APIs, libraries, applications, and implementations.
A test suite is a collection of <dfn>test cases</dfn> that execute
in the context of a <dfn>test-runner</dfn>.  This specification
also supports writing new test-runners, to allow customization
of reporting and processing the result of running test suites.</p>

<h2>Rationale</h2>

<p>The Scheme community needs a standard for writing test suites.
Every SRFI or other library should come with a test suite.
Such a test suite must be portable, without requiring any
non-standard features, such as modules.  The test suite implementation
or "runner" need not be portable, but it is desirable that it be
possible to write a portable basic implementation.</p>
<p>
There are other testing frameworks written in Scheme, including
<a href="https://docs.racket-lang.org/rackunit/">RackUnit</a>.
However RackUnit is not portable.
It is also a bit on the verbose side.
It would be useful to have a bridge between this framework and RackUnit
so RackUnit tests could run under this framework and vice versa.
There exists also at least one Scheme wrapper providing a Scheme interface
to the <q>standard</q> <a href="https://www.junit.org/">JUnit</a> API for Java.
It would be useful to have a bridge so that tests written using this
framework can run under a JUnit runner.
Neither of these features are part of this specification.</p>
<p>
This API makes use of implicit dynamic state, including an
implicit <q>test runner</q>.  This makes the API convenient
and terse to use, but it may be a little less elegant and <q>compositional</q>
than using explicit test objects, such as JUnit-style frameworks.
It is not claimed to follow either object-oriented or functional design
principles, but I hope it is useful and convenient to use and extend.</p>
<p>
This proposal allows converting a Scheme source file to a
test suite by just adding a few macros.  You don't have to
write the entire file in a new form, thus you don't have to
re-indent it.</p>
<p>
All names defined by the API start with the prefix <code>test-</code>.
All function-like forms are defined as syntax.  They may be implemented
as functions or macros or built-ins.  The reason for specifying them as
syntax is to allow specific tests to be skipped without evaluating sub-expressions, or for implementations
to add features such as printing line numbers or catching exceptions.</p>

<h2>Specification</h2>

<p>While this is a moderately complex specification,
you should be able to write simple test suites after just reading the
first few sections below.  More advanced functionality, such
as writing a custom test-runner, is at the end of the specification.</p>

<h3>Writing basic test suites</h3>
<p>Let's start with a simple example.
This is a complete self-contained test-suite.</p>

<pre>
;; Initialize and give a name to a simple testsuite.
(test-begin "vec-test")
(define v (make-vector 5 99))
;; Require that an expression evaluate to true.
(test-assert (vector? v))
;; Test that an expression is eqv? to some other expression.
(test-eqv 99 (vector-ref v 2))
(vector-set! v 2 7)
(test-eqv 7 (vector-ref v 2))
;; Finish the testsuite, and report results.
(test-end "vec-test")
</pre>
<p>
This testsuite could be saved in its own source file.
Nothing else is needed:
We do not require any top-level forms, so it is easy
to wrap an existing program or test to this form, without adding indentation.
It is also easy to add new tests, without having to name individual
tests (though that is optional).</p>
<p>
Test cases are executed in the context of a <dfn>test runner</dfn>,
which is an object that accumulates and reports test results.
This specification defines how to create and use custom test runners,
but implementations should also provide a default test runner.
It is suggested (but not required) that loading the above
file in a top-level environment will cause the
tests to be executed using an implementation-specified default test runner,
and <code>test-end</code> will cause a summary to be displayed
in an implementation-specified manner.</p>

<h4>Simple test-cases</h4>
<p>
Primitive test cases test that a given condition is true.
They may have a name.
The core test case form is <code>test-assert</code>:</p>
<pre>
(<b>test-assert</b> [<var>test-name</var>] <var>expression</var>)
</pre>
<p>
This evaluates the <var>expression</var>.
The test passes if the result
is true; if the result is false, a test failure is reported.
The test also fails if an exception is raised, assuming the implementation
has a way to catch exceptions.
How the failure is reported depends on the test runner environment.
The <var>test-name</var> is a string that names the test case.
(Though the <var>test-name</var> is a string literal in the examples,
it is an expression.  It is evaluated only once.) 
It is used when reporting errors, and also when skipping tests,
as described below.
It is an error to invoke <code>test-assert</code>
if there is no current test runner.</p>
<p>
The following forms may be more convenient than
using <code>test-assert</code> directly:</p>
<pre>
(<b>test-eqv</b> [<var>test-name</var>] <var>expected</var> <var>test-expr</var>)
</pre>
<p>
This is equivalent to:</p>
<pre>
(test-assert [<var>test-name</var>] (eqv? <var>expected</var> <var>test-expr</var>))
</pre>
<p>
Similarly <code>test-equal</code> and <code>test-eq</code>
are shorthand for <code>test-assert</code> combined with
<code>equal?</code> or <code>eq?</code>, respectively:</p>
<pre>
(<b>test-equal</b> [<var>test-name</var>] <var>expected</var> <var>test-expr</var>)
(<b>test-eq</b> [<var>test-name</var>] <var>expected</var> <var>test-expr</var>)</pre>
<p>
Here is a simple example:</p>
<pre>
(define (mean x y) (/ (+ x y) 2.0))
(test-eqv 4 (mean 3 5))
</pre>
<p>For testing approximate equality of inexact reals
we can use <code>test-approximate</code>:</p>
<pre>
(<b>test-approximate</b> [<var>test-name</var>] <var>expected</var> <var>test-expr</var> <var>error</var>)
</pre>
<p>
This is equivalent to (except that each argument is only evaluated once):</p>
<pre>
(test-assert [test-name]
  (and (&gt;= test-expr (- expected error))
       (&lt;= test-expr (+ expected error))))
</pre>

<h4>Tests for catching errors</h4>
<p>
We need a way to specify that evaluation <em>should</em> fail.
This verifies that errors are detected when required.</p>
<pre>
(<b>test-error</b> [[<var>test-name</var>] <var>error-type</var>] <var>test-expr</var>)
</pre>
<p>
Evaluating <var>test-expr</var> is expected to signal an error.
The kind of error is indicated by  <var>error-type</var>.</p>
<p>
If the <var>error-type</var> is left out, or it is
<code>#t</code>, it means "some kind of unspecified error should be signaled".
For example:</p>
<pre>
(test-error #t (vector-ref '#(1 2) 9))
</pre>
<p>
This specification leaves it implementation-defined (or for a future
specification) what form <code><var>test-error</var></code> may take,
though all implementations must allow <code>#t</code>.
Some implementations may support
<a href="https://srfi.schemers.org/srfi-35/srfi-35.html">SRFI-35's conditions</a>,
but these are only standardized for 
<a href="https://srfi.schemers.org/srfi-36/srfi-36.html">SRFI-36's I/O conditions</a>, which are seldom useful in test suites.
An implementation may also allow implementation-specific
<q>exception types</q>.
For example Java-based implementations may allow
the names of Java exception classes:</p>
<pre>
;; Kawa-specific example
(test-error &lt;java.lang.IndexOutOfBoundsException&gt; (vector-ref '#(1 2) 9))
</pre>
<p>
An implementation that cannot catch exceptions should skip
<code>test-error</code> forms.</p>

<h4>Testing syntax</h4>
<p>
Testing syntax is tricky, especially if we want to
check that invalid syntax is causing an error.
The following utility function can help:</p>
<pre>
(<b>test-read-eval-string</b> <var>string</var>)
</pre>
<p>
This function parses <var>string</var> (using <code>read</code>)
and evaluates the result.
The result of evaluation is returned from <code>test-read-eval-string</code>.
An error is signalled if there are unread characters after the
<code>read</code> is done.
For example:<br />
<code>(test-read-eval-string "(+ 3 4)")</code> <i>evaluates to</i> <code>7</code>.<br />
<code>(test-read-eval-string "(+ 3 4")</code> <i>signals an error</i>.<br />
<code>(test-read-eval-string "(+ 3 4) ")</code> <i>signals an error</i>,
because there is extra <q>junk</q> (<i>i.e.</i> a space) after the
list is read.
</p>
<p>
The <code>test-read-eval-string</code> used in tests:</p>
<pre>
(test-equal 7 (test-read-eval-string "(+ 3 4)"))
(test-error (test-read-eval-string "(+ 3"))
(test-equal #\newline (test-read-eval-string "#\\newline"))
(test-error (test-read-eval-string "#\\newlin"))

;; Skip the next 2 tests unless srfi-62 is available.
(test-skip (cond-expand (srfi-62 0) (else 2)))
(test-equal 5 (test-read-eval-string "(+ 1 #;(* 2 3) 4)"))
(test-equal '(x z) (test-read-string "(list 'x #;'y 'z)"))
</pre>

<h4>Test groups and paths</h4>
<p>A <dfn>test group</dfn> is a named sequence of forms containing testcases,
expressions, and definitions.
Entering a group sets the <dfn>test group name</dfn>; leaving a
group restores the previous group name.
These are dynamic (run-time) operations, and a group has no
other effect or identity.
Test groups are informal groupings: they are neither
Scheme values, nor are they syntactic forms.</p>
<!--(More formal <q>test suite</q> values are introduced below.)-->
<p>
A test group may contain nested inner test groups.
The <dfn>test group path</dfn> is a list of the currently-active
(entered) test group names, oldest (outermost) first.</p>
<pre>
(<b>test-begin</b> <var>suite-name</var> [<var>count</var>])
</pre>
<p>A <code>test-begin</code> enters a new test group.
The <var>suite-name</var> becomes the current test group name,
and is added to the end of the test group path.
Portable test suites should use a string literal for <var>suite-name</var>;
the effect of expressions or other kinds of literals is unspecified.</p>
<p>
<b>Rationale:</b> In some ways using symbols would be preferable.
However, we want human-readable names, and standard Scheme does not
provide a way to include spaces or mixed-case text in
literal symbols.</p>
<p>
The optional <var>count</var> must match the number of
test-cases executed by this group.
(Nested test groups count as a single test case for this count.)
This extra test may be useful to catch cases where a test doesn't
get executed because of some unexpected error.</p>
<p>
Additionally, if there is no currently executing test runner,
one is installed in an implementation-defined manner.</p>
<pre>
(<b>test-end</b> [<var>suite-name</var>])
</pre>
<p>
A <code>test-end</code> leaves the current test group.
An error is reported if the <var>suite-name</var> does not
match the current test group name.
<!--If it does match an earlier
name in the test group path, intervening groups are left.
--></p>
<p>
Additionally, if the matching <code>test-begin</code>
installed a new test-runner, then the <code>test-end</code>
will uninstall it, after reporting the accumulated test
results in an implementation-defined manner.</p>

<pre>
(<b>test-group</b> <var>suite-name</var> <var>decl-or-expr</var> ...)
</pre>
<p>
Equivalent to:</p>
<pre>
(if (not (test-to-skip% <var>suite-name</var>))
  (dynamic-wind
    (lambda () (test-begin <var>suite-name</var>))
    (lambda () <var>decl-or-expr</var> ...)
    (lambda () (test-end <var>suite-name</var>))))
</pre>
<p>This is usually equivalent to executing the <var>decl-or-expr</var>s
within the named test group.  However, the entire group is skipped
if it matched an active <code>test-skip</code> (see later).
Also, the <code>test-end</code> is executed in case of an exception.</p>

<h4>Handling set-up and cleanup</h4>
<pre>
(<b>test-group-with-cleanup</b> <var>suite-name</var>
  <var>decl-or-expr</var> ...
  <var>cleanup-form</var>)
</pre>
<p>
Execute each of the <code><var>decl-or-expr</var></code> forms in order
(as in a <code><var>&lt;body&gt;</var></code>),
and then execute the <code><var>cleanup-form</var></code>.
The latter should be executed even if
one of a <code><var>decl-or-expr</var></code> forms raises an exception
(assuming the implementation has a way to catch exceptions).</p>
<p>For example:</p>
<pre>
(let ((f (open-output-file "log")))
  (test-group-with-cleanup "test-file"
    (do-a-bunch-of-tests f)
    (close-output-port f)))
</pre>
<p><b>Erratum note:</b> <a href="https://github.com/scheme-requests-for-implementation/srfi-64/blob/4470ffdec71b1cf61633b664958a3ce5e6997710/srfi-64.html">Earlier versions</a> had a non-working example.
</p>
<!--
<h3>Test suites</h3>
<p>
<i>(Not sure how useful this is, given <code>test-group</code>).</i>
<p>A <dfn>test suite</dfn> is a test group that can (and must) be
executed explicitly.
<pre>
(test-suite <var>suite-name</var> <var>decl-or-expr</var> ...)
</pre>
<p>
The <df>test suite path</dfn> is the list of names of currently
running testsuites, from outermost (oldest) to innermost (newest).
<p>A <code>test-suite</code> form is equivalent to:
<pre>
(test-suite-register <var>suite-name</var>
  (test-group <var>suite-name</var> <var>decl-or-expr</var> ...))
</pre>
You can run previously registered test suite:
<pre>
(test-suite-run <var>suite-name</var>)
</pre>
-->

<h3>Conditonal test-suites and other advanced features</h3>
<p>
The following describes features for controlling which tests to execute,
or specifying that some tests are <em>expected</em> to fail.</p>

<h4>Test specifiers</h4>
<p>Sometimes we want to only run certain tests, or we know that
certain tests are expected to fail.
A <dfn>test specifier</dfn> is a one-argument function that takes a test-runner
and returns a boolean.  The specifier may be run before a test is performed,
and the result may control whether the test is executed.
For convenience, a specifier may also be a non-procedure value,
which is coerced to a specifier procedure, as described below for
<code><var>count</var></code> and <code><var>name</var></code>.</p>
<p>
A simple example is:</p>
<pre>
(if <var>some-condition</var>
  (test-skip 2)) ;; skip next 2 tests
</pre>
<p>
<code>(<b>test-match-name</b> <var>name</var>)</code><br />
The resulting specifier matches if the current test name (as
returned by <code>test-runner-test-name</code>) is <code>equal?</code> to
<var>name</var>.</p>
<p>
<code>(<b>test-match-nth</b> <var>n</var> [<var>count</var>])</code><br />
This evaluates to a <em>stateful</em> predicate: A counter keeps track of
how many times it has been called.
The predicate matches the <var>n</var>'th time it is called
(where <code>1</code> is the first time), and
the next <code>(- <var>count</var> 1)</code> times,
where <var>count</var> defaults to <code>1</code>.</p>
<p>
<code>(<b>test-match-any</b> <var>specifier</var> ...)</code><br />
The resulting specifier matches if any <var>specifier</var>
matches.
Each <code><var>specifier</var></code> is applied, in order,
so side-effects from a later <code><var>specifier</var></code> happen
even if an earlier <code><var>specifier</var></code> is true.</p>
<p>
<code>(<b>test-match-all</b> <var>specifier</var> ...)</code><br />
The resulting specifier matches if each <var>specifier</var>
matches.
Each <code><var>specifier</var></code> is applied, in order,
so side-effects from a later <code><var>specifier</var></code> happen
even if an earlier <code><var>specifier</var></code> is false.</p>
<p>
<code><var>count</var> </code><i>(i.e. an integer)</i><br />
Convenience short-hand for: <code>(test-match-nth 1 <var>count</var>)</code>.</p>
<p>
<code><var>name</var> </code><i>(i.e. a string)</i><br />
Convenience short-hand for <code>(test-match-name <var>name</var>)</code>.</p>

<h4>Skipping selected tests</h4>
<p>In some cases you may want to skip a test.</p>
<pre>
(<b>test-skip</b> <var>specifier</var>)
</pre>
<p>Evaluating <code>test-skip</code> adds the
resulting <var>specifier</var>
to the set of currently active skip-specifiers.
Before each test (or <code>test-group</code>)
the set of active skip-specifiers are applied to the active test-runner.
If any specifier matches, then the test is skipped.</p>
<p>
For convenience, if the <var>specifier</var> is a string that
is syntactic sugar for <code>(test-match-name <var>specifier</var>)</code>.
For example:</p>
<pre>
(test-skip "test-b")
(test-assert "test-a")   ;; executed
(test-assert "test-b")   ;; skipped
</pre>
<p>
Any skip specifiers introduced by a <code>test-skip</code>
are removed by a following non-nested <code>test-end</code>.</p>
<pre>
(test-begin "group1")
(test-skip "test-a")
(test-assert "test-a")   ;; skipped
(test-end "group1")      ;; Undoes the prior test-skip
(test-assert "test-a")   ;; executed
</pre>

<h4>Expected failures</h4>
<p>
Sometimes you know a test case will fail, but you don't have time
to or can't fix it.   Maybe a certain feature only works on certain platforms.
However, you want the test-case to be there
to remind you to fix it.  You want to note that 
such tests are expected to fail.</p>
<pre>
(<b>test-expect-fail</b> <var>specifier</var>)
</pre>
<p>
Matching tests (where matching is defined as in <code>test-skip</code>)
are expected to fail.  This only affects test reporting,
not test execution.  For example:</p>
<pre>
(test-expect-fail 2)
(test-eqv ...) ;; expected to fail
(test-eqv ...) ;; expected to fail
(test-eqv ...) ;; expected to pass
</pre>

<h3>Test-runner</h3>
<p>
A <dfn>test-runner</dfn> is an object that runs a test-suite,
and manages the state.  The test group path, and the sets skip and
expected-fail specifiers are part of the test-runner.
A test-runner will also typically accumulate statistics about executed tests.
</p>
<p>
<code>(<b>test-runner?</b> <var>value</var>)</code><br />
True iff <code><var>value</var></code> is a test-runner object.</p>
<p>
<code>(<b>test-runner-current</b>)</code><br />
<code>(<b>test-runner-current</b> <var>runner</var>)</code><br />
Get or set the current test-runner.
If an implementation supports parameter objects
(as in <a href="https://srfi.schemers.org/srfi-39/srfi-39.html">SRFI-39</a>),
then <code>test-runner-current</code> can be a parameter object.
Alternatively, <code>test-runner-current</code> may be implemented
as a macro or function 
that uses a fluid or thread-local variable, or a plain global variable.</p>
<p>
<code>(<b>test-runner-get</b>)</code><br />
Same as <code>(test-runner-current)</code>, but throws an exception
if there is no current test-runner.</p>
<p>
<code>(<b>test-runner-simple</b>)</code><br />
Creates a new simple test-runner, that prints errors and a summary
on the standard output port.</p>
<p>
<code>(<b>test-runner-null</b>)</code><br />
Creates a new test-runner, that does nothing with the test results.
This is mainly meant for extending when writing a custom runner.</p>
<p>
Implementations <em>may</em> provide other test-runners, perhaps
a <code>(test-runner-gui)</code>.</p>
<p><code>(<b>test-runner-create</b>)</code><br />
Create a new test-runner.  Equivalent to
<code>((test-runner-factory))</code>.</p>
<p>
<code>(<b>test-runner-factory</b>)</code><br />
<code>(<b>test-runner-factory</b> <var>factory</var>)</code><br />
Get or set the current test-runner factory.
A factory is a zero-argument function that creates a new test-runner.
The default value is <code>test-runner-simple</code>,
but implementations may provide a way to override the default.
As with <code>test-runner-current</code>, this may be a parameter object,
or use a per-thread, fluid, or global variable.</p>

<h4>Running specific tests with a specified runner</h4>
<p>
<code>(<b>test-apply</b> [<var>runner</var>] <var>specifier</var> ... <var>procedure</var>)</code><br />
Calls <var>procedure</var> with no arguments using the specified
<var>runner</var> as the current test-runner.
If <var>runner</var> is omitted,
then <code>(test-runner-current)</code> is used.
(If there is no current runner, one is created as in <code>test-begin</code>.)
If one or more <var>specifier</var>s are listed then only tests matching
the <var>specifier</var>s are executed.  A <var>specifier</var> has the same form
as one used for <code>test-skip</code>.  A test is executed
if it matches any of the <var>specifier</var>s in the
<code>test-apply</code> <em>and</em> does not match any
active <code>test-skip</code> specifiers.</p>
<p>
<code>(<b>test-with-runner</b> <var>runner</var> <var>decl-or-expr</var> ...)</code><br />
Executes each <var>decl-or-expr</var> in order in a context
where the current test-runner is <var>runner</var>.</p>

<h3>Test results</h3>
<p>Running a test sets various status properties in the current test-runner.
This can be examined by a custom test-runner,
or (more rarely) in a test-suite.</p>

<h4>Result kind</h4>
<p>Running a test may yield one of the following
status symbols:</p>
<dl>
<dt><code>'pass</code></dt><dd>The test passed, as expected.</dd>
<dt><code>'fail</code></dt><dd>The test failed (and was not expected to).</dd>
<dt><code>'xfail</code></dt><dd>The test failed and was expected to.</dd>
<dt><code>'xpass</code></dt><dd>The test passed, but was expected to fail.</dd>
<dt><code>'skip</code></dt><dd>The test was skipped.</dd>
</dl>
<p>
<code>(<b>test-result-kind</b> [<var>runner</var>])</code><br />
Returns one of the above result codes from the most recent tests.
Returns <code>#f</code> if no tests have been run yet.
If we've started on a new test, but don't have a result yet,
then the result kind is <code>'xfail</code> if the test is expected to fail,
<code>'skip</code> if the test is supposed to be skipped,
or <code>#f</code> otherwise.</p>
<p>
<code>(<b>test-passed?</b> [<var>runner</var>])</code><br />
True if the value of <code>(test-result-kind [<var>runner</var>])</code>
is one of <code>'pass</code> or <code>'xpass</code>.
This is a convenient shorthand that might be useful
in a test suite to only run certain tests if the previous test passed.</p>

<h4>Test result properties</h4>
<p>
A test runner also maintains a set of more detailed <q>result properties</q>
associated with the current or most recent test.  (I.e. the properties of the
most recent test are available as long as a new test hasn't started.)
Each property has a name (a symbol) and a value (any value).
Some properties are standard or set by the implementation;
implementations can add more.</p>
<p>
<code>(<b>test-result-ref</b> <var>runner</var> '<var>pname</var> [<var>default</var>])</code><br />
Returns the property value associated with the <var>pname</var> property name.
If there is no value associated with <code>'<var>pname</var></code>
return <var>default</var>,
or <code>#f</code> if <var>default</var> isn't specified.</p>
<p>
<code>(<b>test-result-set!</b> <var>runner</var> '<var>pname</var> <var>value</var>)</code><br />
Sets the property value associated with the <var>pname</var>
property name to <var>value</var>.
Usually implementation code should call this function, but it may be
useful for a custom test-runner to add extra properties.</p>
<p>
<code>(<b>test-result-remove</b> <var>runner</var> '<var>pname</var>)</code><br />
Remove the property with the name <code>'<var>pname</var></code>.</p>
<p>
<code>(<b>test-result-clear</b> <var>runner</var>)</code><br />
Remove all result properties.
The implementation automatically calls <code>test-result-clear</code>
at the start of a <code>test-assert</code> and similar procedures.</p>
<p>
<code>(<b>test-result-alist</b> <var>runner</var>)</code><br />
Returns an association list of the current result properties.
It is unspecified if the result shares state with the test-runner.
The result should not be modified; on the other hand, the result
may be implicitly modified by future <code>test-result-set!</code> or
<code>test-result-remove</code> calls.
However, a <code>test-result-clear</code> does not modify the returned
alist.  Thus you can <q>archive</q> result objects from previous runs.</p>

<h4>Standard result properties</h4>
<p>
The set of available result properties is implementation-specific.
However, it is suggested that the following might be provided:</p>
<dl>
<dt><code>'result-kind</code></dt>
<dd>The result kind, as defined previously.
This is the only mandatory result property.<br />
<code>(test-result-kind <var>runner</var>)</code> is equivalent to:<br />
<code>(test-result-ref  <var>runner</var> 'result-kind)</code>
</dd>
<dt><code>'source-file</code></dt>
<dt><code>'source-line</code></dt>
<dd>If known, the location of test statements (such as <code>test-assert</code>)
in test suite source code.</dd>
<dt><code>'source-form</code></dt>
<dd>The source form, if meaningful and known.</dd>
<dt><code>'expected-value</code></dt>
<dd>The expected non-error result, if meaningful and known.</dd>
<dt><code>'expected-error</code></dt>
<dd>The <code><var>error-type</var></code>
specified in a <code>test-error</code>, if it meaningful and known.</dd>
<dt><code>'actual-value</code></dt>
<dd>The actual non-error result value, if meaningful and known.</dd>
<dt><code>'actual-error</code></dt>
<dd>The error value, if an error was signalled and it is known.
The actual error value is implementation-defined.</dd>
</dl>

<h3>Writing a new test-runner</h3>
<p>This section specifies how to write a test-runner.
It can be ignored if you just want to write test-cases.</p>


<h4>Call-back functions</h4>
<p>
These call-back functions are <q>methods</q> (in the object-oriented sense)
of a test-runner. A method <code>test-runner-on-<var>event</var></code>
is called by the implementation when <var>event</var> happens.</p>
<p>
To define (set) the callback function for <var>event</var> use the following expression.
(This is normally done when initializing a test-runner.)
<br />
<code>(test-runner-on-<var>event</var>! <var>runner</var> <var>event-function</var>)</code></p>
<p>
An <var>event-function</var> takes a test-runner argument, and possibly other arguments, depending on the <var>event</var>.</p>
<p>
To extract (get) the callback function for <var>event</var> do this:<br />
<code>(test-runner-on-<var>event</var> <var>runner</var>)</code></p>
<p>
To extract call the callback function for <var>event</var> use the following expression.
(This is normally done by the implementation core.)<br />
<code>((test-runner-on-<var>event</var> <var>runner</var>) <var>runner</var> <var>other-args</var> ...)</code></p>
<p>
The following call-back hooks are available.</p>
<p>
<code>(<b>test-runner-on-test-begin</b> <var>runner</var>)</code><br />
<code>(<b>test-runner-on-test-begin!</b> <var>runner</var> <var>on-test-begin-function</var>)</code><br />
<code>(<var>on-test-begin-function</var> <var>runner</var>)</code><br />
The <var>on-test-begin-function</var> is called at the start of an
individual testcase, before the test expression (and expected value) are
evaluated.
</p>
<p>
<code>(<b>test-runner-on-test-end</b> <var>runner</var>)</code><br />
<code>(<b>test-runner-on-test-end!</b> <var>runner</var> <var>on-test-end-function</var>)</code><br />
<code>(<var>on-test-end-function</var> <var>runner</var>)</code><br />
The <var>on-test-end-function</var> is called at the end of an
individual testcase, when the result of the test is available.</p>
<p>
<code>(<b>test-runner-on-group-begin</b> <var>runner</var>)</code><br />
<code>(<b>test-runner-on-group-begin!</b> <var>runner</var> <var>on-group-begin-function</var>)</code><br />
<code>(<var>on-group-begin-function</var> <var>runner</var> <var>suite-name</var> <var>count</var>)</code><br />
The <var>on-group-begin-function</var> is called by a <code>test-begin</code>,
including at the start of a <code>test-group</code>.
The <var>suite-name</var> is a Scheme string,
and <var>count</var> is an integer or <code>#f</code>.</p>
<p>
<code>(<b>test-runner-on-group-end</b> <var>runner</var>)</code><br />
<code>(<b>test-runner-on-group-end!</b> <var>runner</var> <var>on-group-end-function</var>)</code><br />
<code>(<var>on-group-end-function</var> <var>runner</var>)</code><br />
The <var>on-group-end-function</var> is called by a <code>test-end</code>,
including at the end of a <code>test-group</code>.</p>
<p>
<code>(<b>test-runner-on-bad-count</b> <var>runner</var>)</code><br />
<code>(<b>test-runner-on-bad-count!</b> <var>runner</var> <var>on-bad-count-function</var>)</code><br />
<code>(<var>on-bad-count-function</var> <var>runner</var> <var>actual-count</var> <var>expected-count</var>)</code><br />
Called from <code>test-end</code> (before the <var>on-group-end-function</var>
is called) if an <var>expected-count</var> was specified by the matching
<code>test-begin</code> and the <var>expected-count</var> does not match
the <var>actual-count</var> of tests actually executed or skipped.</p>
<p>
<code>(<b>test-runner-on-bad-end-name</b> <var>runner</var>)</code><br />
<code>(<b>test-runner-on-bad-end-name!</b> <var>runner</var> <var>on-bad-end-name-function</var>)</code><br />
<code>(<var>on-bad-end-name-function</var> <var>runner</var> <var>begin-name</var> <var>end-name</var>)</code><br />
Called from <code>test-end</code> (before the <var>on-group-end-function</var>
is called) if a <var>suite-name</var> was specified, and it did not that the
name in the matching <code>test-begin</code>.</p>
<p>
<code>(<b>test-runner-on-final</b> <var>runner</var>)</code><br />
<code>(<b>test-runner-on-final!</b> <var>runner</var> <var>on-final-function</var>)</code><br />
<code>(<var>on-final-function</var> <var>runner</var>)</code><br />
The <var>on-final-function</var> takes one parameter (a test-runner)
and typically displays a summary (count) of the tests.
The <var>on-final-function</var> is called after called the
<var>on-group-end-function</var> correspondiong to the outermost
<code>test-end</code>.
The default value is <code>test-on-final-simple</code> which writes
to the standard output port the number of tests of the various kinds.
</p>
<p>
The default test-runner returned by <code>test-runner-simple</code>
uses the following call-back functions:<br />
<code>(<b>test-on-test-begin-simple</b> <var>runner</var>)</code><br />
<code>(<b>test-on-test-end-simple</b> <var>runner</var>)</code><br />
<code>(<b>test-on-group-begin-simple</b> <var>runner</var> <var>suite-name</var> <var>count</var>)</code><br />
<code>(<b>test-on-group-end-simple</b> <var>runner</var>)</code><br />
<code>(<b>test-on-bad-count-simple</b> <var>runner</var> <var>actual-count</var> <var>expected-count</var>)</code><br />
<code>(<b>test-on-bad-end-name-simple</b> <var>runner</var> <var>begin-name</var> <var>end-name</var>)</code><br />
You can call those if you want to write your own test-runner.</p>

<h4>Test-runner components</h4>
<p>
The following functions are for accessing the other components of a test-runner.
They would normally only be used to write a new test-runner or
a match-predicate.</p>
<p>
<code>(<b>test-runner-pass-count</b> <var>runner</var>)</code><br />
Returns the number of tests that passed, and were expected to pass.</p>
<p>
<code>(<b>test-runner-fail-count</b> <var>runner</var>)</code><br />
Returns the number of tests that failed, but were expected to pass.</p>
<p>
<code>(<b>test-runner-xpass-count</b> <var>runner</var>)</code><br />
Returns the number of tests that passed, but were expected to fail.</p>
<p>
<code>(<b>test-runner-xfail-count</b> <var>runner</var>)</code><br />
Returns the number of tests that failed, and were expected to pass.</p>
<p>
<code>(<b>test-runner-skip-count</b> <var>runner</var>)</code><br />
Returns the number of tests or test groups that were skipped.</p>
<p>
<code>(<b>test-runner-test-name</b> <var>runner</var>)</code><br />
Returns the name of the current test or test group, as a string.
During execution of <code>test-begin</code> this is the name of the
test group; during the execution of an actual test, this is the name
of the test-case.
If no name was specified, the name is the empty string.</p>
<p>
<code>(<b>test-runner-group-path</b> <var>runner</var>)</code><br />
A list of names of groups we're nested in, with the outermost group first.</p>
<p>
<code>(<b>test-runner-group-stack</b> <var>runner</var>)</code><br />
A list of names of groups we're nested in, with the outermost group last.
(This is more efficient than <code>test-runner-group-path</code>,
since it doesn't require any copying.)</p>
<p>
<code>(<b>test-runner-aux-value</b> <var>runner</var>)</code><br />
<code>(<b>test-runner-aux-value!</b> <var>runner</var> <var>on-test</var>)</code><br />
Get or set the <code>aux-value</code> field of a test-runner.
This field is not used by this API or the <code>test-runner-simple</code>
test-runner, but may be used by custom test-runners to store extra state.</p>
<p>
<code>(<b>test-runner-reset</b> <var>runner</var>)</code><br />
Resets the state of the <var>runner</var> to its initial state.
</p>

<h4>Example</h4>
<p>This is an example of a simple custom test-runner.
Loading this program before running a test-suite will install
it as the default test runner.</p>
<pre>
(define (my-simple-runner filename)
  (let ((runner (test-runner-null))
	(port (open-output-file filename))
        (num-passed 0)
        (num-failed 0))
    (test-runner-on-test-end! runner
      (lambda (runner)
        (case (test-result-kind runner)
          ((pass xpass) (set! num-passed (+ num-passed 1)))
          ((fail xfail) (set! num-failed (+ num-failed 1)))
          (else #t))))
    (test-runner-on-final! runner
       (lambda (runner)
          (format port "Passing tests: ~d.~%Failing tests: ~d.~%"
                  num-passed num-failed)
	  (close-output-port port)))
    runner))

(test-runner-factory
 (lambda () (my-simple-runner "/tmp/my-test.log")))
</pre>

<h2>Implementation</h2>
<p>
The test implementation uses <code>cond-expand</code>
(<a href="https://srfi.schemers.org/srfi-0/srfi-0.html">SRFI-0</a>)
to select different code depending on certain SRFI names (<code>srfi-9</code>,
<code>srfi-34</code>, <code>srfi-35</code>, <code>srfi-39</code>),
or implementations (<code>kawa</code>).
It should otherwise be portable to any R5RS implementation.</p>

<p><a href="testing.scm">testing.scm</a></p>

<h3>Examples</h3>
<p>Here is <a href="srfi-25-test.scm"><code>srfi-25-test.scm</code></a>,
based converted from Jussi Piitulainen's
<a href="https://srfi.schemers.org/srfi-25/test.scm"><code>test.scm</code></a>
for <a href="https://srfi.schemers.org/srfi-25/srfi-25.html">SRFI-25</a>.</p>

<h3>Test suite</h3>
<p>
Of course we need a test suite for the testing framework itself.
This suite <code><a href="srfi-64-test.scm">srfi-64-test.scm</a></code>
was contributed by Donovan Kolbly
<a href="mailto:donovan@rscheme.org"><code>&lt;donovan@rscheme.org&gt;</code></a>.</p>

<h2>Copyright</h2>
<p>
Copyright (C) Per Bothner (2005, 2006)</p>
<p>
Permission is hereby granted, free of charge, to any person obtaining
a copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:</p>
<p>
The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software.</p>
<p>
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.</p>
<hr />
<address>Author: <a href="mailto:per@bothner.com">Per Bothner</a></address>
<address>Editor: <a href="mailto:srfi%20minus%20editors%20at%20srfi%20dot%20schemers%20dot%20org">Francisco Solsona</a></address>
<p>
<!-- Created: Sat Sep  4 12:26:07 PDT 2004 -->
<!-- hhmts start -->
Last modified: Thu Mar 31 19:49:52 MST 2011
<!-- hhmts end -->
</p>
  </body>
</html>