krakenjs / belter Goto Github PK

The versions for some dependencies are currently outdated.

User agent for Yandex Chromium Browser is being recognized as a Safari browser.

Mozilla/5.0 (iPad; CPU OS 14_8 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/14.0 YaBrowser/21.11.1.882.11 Mobile/15E148 Safari/604.1

Need to add an isYandex and check that it doesn't exist in isSafari check.

This project isn't written in TypeScript, but many TypeScript users would like to use published typings. For this issue, create TypeScript typings that cover the exported functions, and contribute it back to DefinitelyTyped!

This issue is great for someone who wants to learn more about TypeScript!

Does any line of code require mocking in order to be tested? If so, then Chai, Sinon, and Sinon-chai should be installed as dependencies, or else karma-sinon-chai won't be providing any use.

I see here that this is the commit in which chai, sinon, and sinon-chai were replaced with grumbler-scripts.

You can see here that grumbler-scripts does not have chai, sinon, nor sinon-chai as dependencies. Instead, it has karma-sinon-chai

It seems to me that this repo is being functionally tested rather than unit tested, so I'm not sure sinon and chai are needed at all. Is that why they were removed in the first place? Some one please confirm!

• Create a Changelog based on the keep a Changelog format
• Update the Changelog with at least the last five commits.
• Keep the Changelog updated when a new version gets released.
• Use the Unreleased section to track unreleased changes.

The onResize method in dom.js currently uses offsetWidth and offsetHeight to get the elements dimensions.

These two attributes always return integer values--disregarding the fractional part

• https://developer.mozilla.org/en-US/docs/Web/API/HTMLElement/offsetWidth
• https://bugs.chromium.org/p/chromium/issues/detail?id=360889

As some elements can have subpixel values, perhaps its worth it considering not to round the width and height to achieve a better precision.

I believe this is also causing the zoid library to sometimes generate slightly wrong iframe sizes, since the method seems to be used here

• this can cause scrollbars to appear in the iframes if the dimensions generated are smaller than the element dimensions

In case you are interested, I opened a draft PR with a suggested approach:

• https://github.com/krakenjs/belter/pull/66/files

The function checks if the document body is present. If the document is not ready, it waits for the window to be ready and checks again if the document body is present. If the body is present, it returns the body. If not throws an error. This is the code:

 return waitForDocumentReady().then(() => {
            if (document.body) {
                return document.body;
            }

            throw new Error('Document ready but document.body not present');
        });

But, waitForDocumentReady calls isDocumentReady which in turn checks:

if (document.body && document.readyState === 'complete) return true`

That means the document cannot be ready without the body being ready. This causes an error while testing, when I mocked document.body = null. Also leads to throw being unreachable.

When I installed the repo I ran npm i and npm build then got a bunch of errors.

After some searching online then I found this:
https://www.npmjs.com/package/karma-chrome-launcher#headless-chromium-with-puppeteer

Usage:
npm i -D puppeteer karma-chrome-launcher

After I ran this command my tests started working. Is this something that should be updated in the README.md?

Are there any other ways the testing structure could improve?

I notice that each folder in the tests folder has an index.js file in it. Is there a clean way to make it such that we do not have to use an index.js file inside every folder of the tests folder?

Every single time a person adds a new folder inside tests, there are two index.js file that has to be updated, and then another to be created. The created index.js would then have to be updated every time someone makes a new testing file, and also must be created in the new folder that the person added to tests.

It would be less confusing if we did not have to use so many index.js, while still maintaining an organized structure. My preliminary thoughts are to make a single script that replaces the need to make a lot of index.js files.

There are a LOT of useful utilities in belter, however there aren't any unit tests in this repo! 😬 If anyone's interested in picking this up, can you write unit tests for at least two files in src/? 😄

'npm run lint' fails due to Extra Semicolon issue as below