10 KiB
| name | description | metadata | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| basex-test | Use this skill when working with BaseX XQuery unit tests |
|
Unit Functions
This BaseX module contains annotations and functions for performing XQUnit tests. Each test is an XQuery function with the %unit:test annotation.
Introduction
The more complex a software application grows, the more error-prone it gets. This is why testing frameworks exist, which provide a standardized, automated way of testing software. The xUnit frameworks (such as SUnit or JUnit) allow testing of atomic units of a program, such as single functions and algorithms.
This module borrows heavily from the existing frameworks: it provides various annotations for testing XQuery functions. Unit functions are provided to assert the validity of arbitrary conditions expressed in XQuery and to raise errors whenever a condition is not satisfied. Some additional functions exist to run all unit tests of the current module or a set of specified library modules.
Usage
Tests are started via the [TEST](Commands#test) command. It compiles all XQuery modules in a given file
or directory and runs all functions that are annotated with %unit:test. A test report is generated and returned, which resembles the format returned by other xUnit testing frameworks, such as the Maven Surefire Plugin (see below).
Conventions
All annotations, functions and errors in this module are assigned to the http://basex.org/modules/unit namespace, which is statically bound to the unit prefix.
Annotations
unit:test
| Syntax | unit:test() unit:test("expected", CODE) |
| Summary | With this annotation, a function can be marked as unit test. It will be evaluated if a test report is created for the module in which this function is located. error can be supplied as additional string argument. It is followed by CODE, which must be a valid EQName string. If the function expression does not raise that error, the test will fail. |
| Examples | xquery<br>declare %unit:test function local:void() { () };<br>The following test will be successful, as it does nothing (and, hence, nothing wrong). - - - xquery<br>declare %unit:test('expected', "err:XPTY0004") function local:add() {<br> 123 + 'strings and integers cannot be added'<br>};<br>The following test will be successful, as the function body will raise err:XPTY0004. |
unit:before
| Syntax | unit:before() unit:before(FUNCTION) |
| Summary | A function decorated with this annotation will be evaluated before each unit test as a separate transaction. FUNCTION can be supplied as additional argument. It must be a valid EQName string. If specified, the function will only be evaluated before a function with the given name is tested. This extension is e. g. helpful if the results of updates need to be tested. |
| Examples | xquery<br>declare %updating %unit:before("local:check") function local:before-check() {<br> db:create('test-db')<br>};<br>declare %updating %unit:test function local:check() {<br> unit:assert(db:exists('test-db'))<br>};<br>The first function will be evaluated before the actual test. |
unit:after
| Syntax | unit:after() unit:after(FUNCTION) |
| Summary | A function decorated with this annotation will be evaluated after each unit test as a separate transaction. FUNCTION can be supplied as additional argument. It must be a valid EQName string. If specified, the function will only be evaluated after a function with the given name is tested. |
unit:before-module
| Syntax | unit:before-module() |
| Summary | If a function is decorated with this annotation, it will be evaluated before all unit tests in the current module as a separate transaction. |
unit:after-module
| Syntax | unit:after-module() |
| Summary | If a function is decorated with this annotation, it will be evaluated after all unit tests in the current module as a separate transaction. |
unit:ignore
| Syntax | unit:ignore() unit:ignore(MESSAGE) |
| Summary | If a function is decorated with this annotation, it will temporarily be ignored by the test suite runner. |
Functions
unit:assert
Signature
unit:assert( $test as item()*, $info as item() := () ) as empty-sequence()
SummaryAsserts that the effective boolean value of the specified $test is true and returns an empty sequence. Otherwise, raises an error. The effective boolean value of an expression can be explicitly computed by using the fn:boolean function. The default failure message can be overridden with the $info argument.Errors
fail |
An assertion failed, or an error was raised. |
unit:assert-equals
Signature
unit:assert-equals( $returned as item()*, $expected as item()*, $info as item() := () ) as empty-sequence()
SummaryAsserts that the specified arguments are equal according to the rules of the [fn:deep-equal](Standard_Functions#fn:deep-equal) function. Otherwise, raises an error. The default failure message can be overridden with the $info argument.Errors
fail |
An assertion failed, or an error was raised. |
unit:fail
Signature
unit:fail( $info as item() := () ) as empty-sequence()
SummaryRaises a unit error. The default failure message can be overridden with the $info argument.Errors
fail |
An assertion failed, or an error was raised. |
Example
The following XQUnit module tests.xqm contains all available unit annotations:
Query
module namespace test = 'http://basex.org/modules/xqunit-tests';
(:~ Initializing function, which is called once before all tests. :)
declare %unit:before-module function test:before-all-tests() {
()
};
(:~ Initializing function, which is called once after all tests. :)
declare %unit:after-module function test:after-all-tests() {
()
};
(:~ Initializing function, which is called before each test. :)
declare %unit:before function test:before() {
()
};
(:~ Initializing function, which is called after each test. :)
declare %unit:after function test:after() {
()
};
(:~ Function demonstrating a successful test. :)
declare %unit:test function test:assert-success() {
unit:assert(<a/>)
};
(:~ Function demonstrating a failure using unit:assert. :)
declare %unit:test function test:assert-failure() {
unit:assert((), 'Empty sequence.')
};
(:~ Function demonstrating a failure using unit:assert-equals. :)
declare %unit:test function test:assert-equals-failure() {
unit:assert-equals(4 + 5, 6)
};
(:~ Function demonstrating an unexpected success. :)
declare %unit:test("expected", "err:FORG0001") function test:unexpected-success() {
()
};
(:~ Function demonstrating an expected failure. :)
declare %unit:test("expected", "err:FORG0001") function test:expected-failure() {
1 + <a/>
};
(:~ Function demonstrating the creation of a failure. :)
declare %unit:test function test:failure() {
unit:fail("Failure!")
};
(:~ Function demonstrating an error. :)
declare %unit:test function test:error() {
1 + <a/>
};
(:~ Skipping a test. :)
declare %unit:test %unit:ignore("Skipped!") function test:skipped() {
()
};
By running TEST tests.xqm, the following report will be generated (timings may differ):
Result
<testsuites time="PT0.256S">
<testsuite name="file:///C:/Users/user/Desktop/test.xqm"
time="PT0.212S" tests="8" failures="4" errors="1" skipped="1">
<testcase name="assert-success" time="PT0.016S"/>
<testcase name="assert-failure" time="PT0.005S">
<failure line="30" column="15">
<info>Empty sequence.</info>
</failure>
</testcase>
<testcase name="assert-equals-failure" time="PT0.006S">
<failure line="35" column="22">
<returned item="1" type="xs:integer">9</returned>
<expected item="1" type="xs:integer">6</expected>
<info>Item 1: 6 expected, 9 returned.</info>
</failure>
</testcase>
<testcase name="unexpected-success" time="PT0.006S">
<failure>
<expected>FORG0001</expected>
</failure>
</testcase>
<testcase name="expected-failure" time="PT0.004S"/>
<testcase name="failure" time="PT0.004S">
<failure line="50" column="13">
<info>Failure!</info>
</failure>
</testcase>
<testcase name="error" time="PT0.004S">
<error line="55" column="6" type="FORG0001">
<info>Cannot cast to xs:double: "".</info>
</error>
</testcase>
<testcase name="skipped" skipped="Skipped!" time="PT0S"/>
</testsuite>
</testsuites>
Errors
| Code | Description |
|---|---|
fail |
An assertion failed, or an error was raised. |
no-args |
A test function must have no arguments. |
private |
A test function must not be private. |
Changelog
Version 9.0
- Updated: error codes updated; errors now use the module namespace
Version 8.0.2
- Updated: (expected) errors are compared by QNames instead of local names (including namespaces).
Version 8.0
- Added:
[unit:fail](#unit:fail), 0-argument signature. - Updated: the info argument of functions can now be an arbitrary item.
- Updated: infos are now represented in an
infochild element. - Updated:
[unit:before](#unit:before)and[unit:after](#unit:after)can be extended by a filter argument. - Deleted:
UNIT0006(ignore results returned by functions).
Version 7.9
- Added: TEST command
- Removed:
[unit:test](#unit:test),unit:test-uris
Version 7.8
- Added:
[unit:assert-equals](#unit:assert-equals) - Updated: enhanced test report output
Version 7.7
- Added: New module added.