Testcafe: Selector waits eternally for elements to appear when the date is mocked by using client scripts.

[Thormod]

What is your Test Scenario?

For my Integration tests, I need to mock the browser date and to accomplish that I'm using clientScripts injecting the mockdate module:

const clientScriptsObject = [
  { module: 'mockdate' },
  { content: "MockDate.set('2020-04-19')" },
];

For this issue, I created a small project that goes to Wikipedia and tries to search for something but using an incorrect selector:

const mockDate = `
    Date.now = function () {
        return 1566408247;
    };
`;

fixture('My first fixture').page('https://www.wikipedia.org/');

test('users can search', async (t) => {
  await t
    .click(Selector('#searchInput'))
    .click(Selector('#this-selector-does-not-exist'))
    .typeText(Selector('#searchInput'), 'vikings')
    .click(Selector('button[type=submit]'));
}).clientScripts({ content: mockDate });
 ```
### What is the Current behavior?
<!-- Describe the behavior you see and consider invalid. -->
 When the tests execute ` .click(Selector('#this-selector-does-not-exist'))` the `Selector` get stuck waiting for the element to appear, ignoring the default (and setted) selector timeouts:

![image](https://user-images.githubusercontent.com/7811197/90932366-1e51e480-e3c4-11ea-82e0-bf891797d13f.png)


### What is the Expected behavior?
<!-- Describe what you expected to happen. -->
To raise an error specifying that the selector does not exist.

![image](https://user-images.githubusercontent.com/7811197/90932418-3a558600-e3c4-11ea-8fbe-bd3932b5f447.png)


### What is your web application and your TestCafe test code?
<!-- Share a public accessible link to your application or provide a simple app which we can run. -->
You can access the code here:
https://github.com/Thormod/testcafe-issue

<details>
<summary>Your website URL (or attach your complete example):</summary>

<!-- Provide your website URL or attach a sample. Note: if your website requires any additional access procedures like authentication, please ask the website owner to send us a written confirmation at [[email protected]](mailto:[email protected]) in a free text form. It will allow the DevExpress staff to remotely access the website and its internal resources for research, testing, and debugging purposes. -->
</details>

<details>
<summary>Your complete test code (or attach your test files):</summary>

<!-- Paste your test code here: --> 
```js

Your complete configuration file (if any):

Your complete test report:

Screenshots:

Steps to Reproduce:

Just run npm run test on the example project, or mock the Date.now() function by using clientScript

Your Environment details:

• testcafe version: ^1.9.1
• node.js version: v10.19.0
• command-line arguments: testcafe chrome
• browser name and version: Chrome latest
• platform and version: MacOs Catalina 10.15.6

[jhenaoz]

i have a similar issue when i mock date object in test cafe, I guess the testcafe timeouts use the same object we are mocking?

[leonleo997]

I'm having the same issue when I try to avoid external dependencies with mocks and I'm dealing with dates. I'm interested

[salorrego]

I'm also having the same issue with the mock dates. I'm also interested, this also happens on Linux Mint 19.3 Tricia

[AlexandroGil]

im having the exact same issue

[DavidMosquera]

I'm also having the same issue

[Serandrus]

I'm also having the same issue! And I see I'm not the only one

[miherlosev]

Hi @Thormod

Thank you for the detailed information. I've reproduced the problem.

[Thormod]

@miherlosev Do you know when the new version will be released? PD: Thank you for solving this bug!

[AndreyBelym]

A release candidate version [email protected] is already published to NPM. You can test it in your environment if you do not mind installing preliminary versions. Once we test it and create a changelog, we will publish a stable release.

[Thormod]

@AndreyBelym @miherlosev There are some TypeErrors in the new version:

Uncaught TypeError: Cannot read property 'destUrl' of null
    at Function.t._sameOriginCheck (worker-hammerhead.js:1)
    at r.fetch (worker-hammerhead.js:1)
    at u (b54a5bd5-05e5-4a32-97e0-f1d459ace1fb:2)
    at b54a5bd5-05e5-4a32-97e0-f1d459ace1fb:2
    at xt (b54a5bd5-05e5-4a32-97e0-f1d459ace1fb:2)
    at Object.bt [as getArrayBuffer] (b54a5bd5-05e5-4a32-97e0-f1d459ace1fb:2)
    at l.s [as loadVectorData] (b54a5bd5-05e5-4a32-97e0-f1d459ace1fb:2)
    at l.loadTile (b54a5bd5-05e5-4a32-97e0-f1d459ace1fb:2)
    at ke.loadTile (b54a5bd5-05e5-4a32-97e0-f1d459ace1fb:2)
    at __call$ (b54a5bd5-05e5-4a32-97e0-f1d459ace1fb:1)

It seems to be related with the fetch requests:

[AndreyBelym]

I cannot reproduce the issue with Wikipedia.org. But I created my own example, which demonstrates a very similar issue: https://github.com/DevExpress/testcafe-hammerhead/issues/2428. The version with the fix is already published to NPM as [email protected]. As my example and the fix is based on assumptions only, there is a risk that it doesn't fix your specific issue. Could you please try to install [email protected] and provide a sample test page if you still can reproduce the issue in your environment?

[Thormod]

@AndreyBelym It worked! Thank u for solving this so fast!