hurl/integration
2024-07-06 23:02:51 +02:00
..
hurl Unify Retry and Repeat types with Count. 2024-07-06 23:02:51 +02:00
hurlfmt Add repeat option per request 2024-06-19 22:22:00 +02:00
README.md Update integration README. 2024-04-09 13:58:01 +02:00
test_pattern.py Replace ~~~ in pattern files by standard regex 2024-06-09 11:53:16 +02:00
test_script.py Apply delay only once per entry, no matter how many retry. 2024-06-18 17:38:13 +02:00

Hurl Integration Tests Suite

Introduction

In the Hurl project, there are three type of tests:

  • Rust unit tests: run at the root of the project with cargo test --lib
  • Rust integration tests: run at the root of the project with cargo test. You will also run unit test with this command. To run Rust integration tests, you need to launch a local test server (see below).

These tests are "classic" Rust tests and should not surprise a Rust developer.

Along with tests, we have an extensive integration test suite. These tests launch scripts to run hurl and hurlfmt and test various options and Hurl files. To run these tests you have to set up a local server (see below) All these tests will be performed automatically in Hurl CI/CD, on various OS, for every pull request.

Set up Test Local Server

Python 3.9+

The local test server is a Flask application, so you will need Python 3.9+ on your machine. You can create a Python virtual environment and install required dependencies:

$ python3 -m venv .venv
$ source .venv/bin/activate
$ pip install --requirement bin/requirements-frozen.txt

Proxy

Some integration tests need a proxy. You can use mitmproxy or squid.

Start local server

You can use the scripts bin/test/test_prerequisites.sh / bin/test/test_prerequisites.ps1 depending on your OS to start the local test server and proxy. Once launch, there is:

Now, everything is ready to run the integration tests!

Integration Tests

Organisation

Integration tests to test hurl binary are grouped in integration/hurl directory:

  • hurl/tests_ok: every test there must be successful (exit code 0). The Hurl files in this folder are formatted with hurlfmt.
  • hurl/tests_ok_not_linted: every test here must be successful but are not necessary formatted with hurlfmt. This way we can ensure that there is no regression even if a Hurl file doesn't follow a stricter format.
  • hurl/tests_failed: every test must fail (exit code different from 0). Tests are syntactically correct, so the error raised by the test is a runtime error.
  • hurl/tests_error_parser: every test is not a syntactically correct Hurl file. We test here the parsing error message.
  • [hurl/ssl]: tests SSL features (server and client certificates, etc...)
  • [hurl/unix_socket]: tests Unix Socket

Integration tests to test hurlfmt binary are grouped in integration/hurlfmt directory:

Files Description

An integration test consists of:

  • two runnable scripts: one for Linux, macOS (foo.sh) and one for Windows (foo.ps1). These are the integration tests that we want to execute.
  • a Hurl file (foo.hurl)
  • a Flask endpoint (foo.py). This is the server side used by the Hurl file. You can add as many assert as you want to test that our Hurl client conforms to what is expected. Generally, each integration test has its own Flask endpoint, even if there is some duplication between tests.
  • an expected stdout file (foo.out). This file is the expected value for stdout. This file is not dependent from the OS, as we want a Hurl file to have the same stdout on any OS. If the stdout have some variant data (like timestamp for instance), one can use a patterned expected file, with ~~~ for wildcard matching (foo.out.pattern)
  • an expected stderr file (foo.err). This file is the expected value for stderr. This file is not dependent from the OS, as we want a Hurl file to have the same stderr on any OS. Like stdout, stderr expected file can be patterned (foo.err.pattern)
  • an expected exit code (foo.exit). This file is the expected value of the script. If absent, the default exit code is 0. is a JSON view of the Hurl source file and can serve to convert from/to Hurl format.
  • an expected list of curl commands. This list is the curl command equivalent to each request in the Hurl file (which is logged in --verbose/--very-verbose mode). Each curl command is run against the server.

To run all integration tests:

$ cd integration/hurl
$ python3 integration.py
$ cd integration/hurlfmt
$ python3 integration.py

To run a particular integration test without any check:

$ cd integration/hurl
$ tests_ok/hello.sh

To run a particular integration test with all check (stdout, stderr, HTML/JSON export etc...):

$ cd integration/hurl
$ python3 ../test_script.py tests_ok/hello.sh

Sample

include.sh:

#!/bin/bash
set -Eeuo pipefail
hurl tests_ok/include.hurl --include --verbose

include.ps1:

Set-StrictMode -Version latest
$ErrorActionPreference = 'Stop'
hurl tests_ok/include.hurl --include --verbose

include.hurl:

GET http://localhost:8000/include

HTTP 200
`Hello`

include.py:

from app import app
from flask import Response


@app.route("/include")
def include():
    return Response("Hello")

include.out.pattern:

HTTP/1.1 200
Server: Werkzeug/~~~ Python/~~~
Date: ~~~
Content-Type: text/html; charset=utf-8
Content-Length: 5
Server: Flask Server
Connection: close

Hello

include.html:

<pre><code class="language-hurl"><span class="hurl-entry"><span class="request"><span class="line"><span class="method">GET</span> <span class="url">http://localhost:8000/include</span></span>
</span><span class="response"><span class="line"></span>
<span class="line"><span class="version">HTTP</span> <span class="number">200</span></span>
<span class="line"><span class="string">`Hello`</span></span>
</span></span><span class="line"></span>
</code></pre>

include.json:

{"entries":[{"request":{"method":"GET","url":"http://localhost:8000/include"},"response":{"status":200,"body":{"type":"text","value":"Hello"}}}]}

include.curl:

curl 'http://localhost:8000/include'