Be strict about data providers having redundant data

[kubawerlos]

Closes #6211

• 1st commit is the feature itself
• 2nd commit is the changes to align the codebase to follow the new rule

Should this be hidden behind a config option?
Should it make the test risky/with a warning rather than producing error?

[sebastianbergmann]

Should this be hidden behind a config option?

No.

[sebastianbergmann]

Thank you!

[kubawerlos]

@sebastianbergmann you are so fast 👍🏼

Even too fast: #6215

[sebastianbergmann]

I have "downgraded" this from an error to a warning in 7f1969a.

[kubawerlos]

Cool. The warning will be on the test case level, not at the test method level, as the error was.

[sebastianbergmann]

The warning can be seen here: https://github.com/sebastianbergmann/phpunit/blob/main/tests/end-to-end/data-provider/too-many-arguments.phpt#L23

[dereuromark]

This might have a small regression for ...

    /**
     * tests failure messages for assertions
     *
     * @param string $assertion Assertion method
     * @param string $message Expected failure message
     * @param string $url URL to test
     * @param mixed ...$rest
     */
    #[DataProvider('assertionFailureMessagesProvider')]
    public function testAssertionFailureMessages($assertion, $message, $url, ...$rest): void
    {
        $this->expectException(AssertionFailedError::class);
        $this->expectExceptionMessage($message);

        Security::setSalt($this->key);

        $this->get($url);

        call_user_func_array($this->$assertion(...), $rest);
    }

If there is a ...$var maybe the warning should be skipped as it is intentional?

1) Cake\Test\TestCase\TestSuite\IntegrationTestTraitTest::testAssertionFailureMessages
* Data set "assertCookie" has more arguments (5) than the test method accepts (4)

[dereuromark]

Maybe a

$parameters = $method->getParameters();

$hasVariadic = false;

foreach ($parameters as $param) {
    if ($param->isVariadic()) {
        $hasVariadic = true;
        break;
    }
}

And if true we just skip?

[sebastianbergmann]

@dereuromark Please open a new issue if you think you found a bug/regression. Thank you.

[kubawerlos]

It is regression, I've raised the PR with the fix: #6228