Sometimes, when you testing edge cases of your code that cause error, the code/frameworks that you use may call console.error
. A common case that I encounter is testing error boundary in React.
The console.error
is trying to be helpful, however when the error is expected in the tests, the long call stack is cluttering the log, which may cause us miss out important warning/message.
To fix that, the common approach is to mock the console.error
call.
beforeAll(() => {
jest.spyOn(global.console, 'error').mockImplementation(() => {});
});
afterAll(() => {
global.console.error.mockRestore();
});
However, this approach has a downside: if there is any error logged during the test, you can’t see the error at all.
To make sure that will not happen, the closest thing you can do is to assert how many the console method has been called.
beforeAll(() => {
jest.spyOn(global.console, 'error').mockImplementation(() => {});
});
afterAll(() => {
global.console.error.mockRestore();
});
test('my test case title', () => {
// the tests
expect(console.error).toHaveBeenCallTimes(2);
});
Then if there is any error logged that is not expected, at least the tests will catch it.