Simple test harness for describing HTTP tests
The tests use a HOCON (Human Optimised Config Object Notation) file to create a suite of tests.
For example:
tests { // tests wrapper
httpbin-get { // creates a test named httpbin-get
url="http:https://httpbin.org/get" // gives the test a URL
pathTests= [
{ path="url", testExpression="http:https://httpbin.org/get" } // defines some jsonpath tests
]
}
httpbin-xml {
url="http:https://httpbin.org/xml"
acceptHeader=application/xml // request XML instead of JSON
pathTests = [
{ path="/slideshow/slide/title", testExpression="\\w" } // xpath tests are also supported
]
}
httpbin-post {
url="http:https://httpbin.org/post"
method=POST // Send a POST, PUT, PATCH, DELETE, etc instead of GET
contentType=application/json // Content type of the HTTP request body
entity=example.json // File that contains the data for the HTTP request body
pathTests= [
{ path="json.q", testExpression="1" }
]
}
httpbin-robots {
url="http:https://httpbin.org/robots.txt"
acceptHeader=text/html
script=example.groovy // evaluate a groovy script that returns a boolean
}
}
Path tests can be a JSON path or XPath expression, automatically detected based on the content type of the return service.
Path testing occurs as follows:
- Path expression is evaluated against the JSON or XML document and returns a list of nodes
- XML documents extract the text content of the nodes, JSON documents are assumed to only return single value nodes
- If the list is empty then the test fails (unless the test is marked as optional=true
- The testExpression is parsed as a regular expression
- If any entry in the list does not contain the regular expression, the test fails.
Using spock, you can create a specification like this:
class FunctionalTestSpec extends Specification {
@Shared HttpTestRunner runner
def setupSpec() {
runner = HttpTestRunner.Builder.withClasspathResolver('test.conf').groovyScriptEvaluator().build()
}
def cleanupSpec() {
}
@Unroll
def "test #test.name"() {
setup:
when:
def response = runner.run(test)
then:
response.statusCode == 200
for (def pt : test.pathTests)
assert response.evaluatePath(pt)
if (test.script)
assert runner.evaluateScriptTest(test.script, response)
where:
test << runner.tests
}
}This will read test.conf from the classpath and unroll each entry in the tests wrapper into it's own individual test.
The cli subproject contains a version of the Spock test above which can be run via the command line. Execute it like this:
java -jar functional-test-harness-jar-with-dependencies.jar example.conf
Where example.conf is a path to a configuration file. It will produce output like this:
$ java -jar target/functional-test-harness-jar-with-dependencies.jar example.conf
Running httpbin-post (1/4)
HTTP Request was successful... ✓
Evaluating path json.q against 1... ✓
Running httpbin-xml (2/4)
HTTP Request was successful... ✓
Evaluating path /slideshow/slide/title against \w... ✓
Running httpbin-robots (3/4)
HTTP Request was successful... ✓
Evaluating script: example.groovy... ✓
Running httpbin-get (4/4)
HTTP Request was successful... ✓
Evaluating path url against http:https://httpbin.org/get... ✓
$ echo $?
0
The return code of the process indicates the number of failed tests.