Skip to content

Expected exception semantics #1798

Closed
Closed
@mindplay-dk

Description

@mindplay-dk

The semantics of expected exception testing could be better.

When you call setExpectedException() is just overwrites an internal test-condition, which means you can't do more than one assertion within a single test.

What's worse though, if you have two statements after a call to setExpectedException(), and you're expecting the first statement to trigger an exception, but it doesn't, and instead the second statement, which shouldn't, triggers the same type of exception - your test comes up green, but may actually be covering up two bugs.

What's missing, is some way to indicate precisely which statement is expected to generate the exception - for example:

/**
 * Check for an expected exception, which must be thrown.
 *
 * @param mixed    $exceptionName
 * @param string   $exceptionMessage
 * @param int      $exceptionCode
 * @param callable $function
 *
 * @void
 */
public function assertException($exceptionName, $exceptionMessage, exceptionCode, $function)
{
    try {
        $function();
    } catch (Exception $e) {
        // ... check type/message/code and record assertion as passed/failed
        return;
    }
    // ... mark assertion as failed (not thrown)
}

Using a closure, you can designate precisely which statement is expected to throw - and, you can test for any number of expected exceptions in the same test body.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions