Array.from can be incorrectly defined by 3rd party js libs

[eoghanmurray]

Preflight Checklist

• I have searched the issue tracker for a bug report that matches the one I want to file, without success.

What package is this bug report for?

rrweb

Version

latest version with #1272 applied

Expected Behavior

inline styles recorded correctly

Actual Behavior

inline styles recorded as follows:

<style type="text/css">
[object CSSStyleRule][object CSSStyleRule][object CSSStyleRule][object CSSStyleRule][object CSSStyleRule][object CSSStyleRule][object CSSStyleRule][object CSSStyleRule][object CSSStyleRule ...
</style>

Steps to Reproduce

#1272 makes full use of the second map argument of the built in Array.from

I discovered today that MooTools < 1.6 implements their own version of Array.from, which lacks the second argument, so if an old version of mootools is on a website it messes things up — or also an old version of Prototype (e.g. 1.7.1)

I did a similar 'guard' for Date.now being overwritten in #1196
Is it a good idea to do the same in this case?

There might be small performance loss from an extra function, but main question would be whether it is our responsibility to patch over this sort of thing?

Here is a test to undo changes from 3rd party libraries:

	if (Array.from([1], (x)=>x*2)[0] !== 2) {
		// https://stackoverflow.com/a/8580576/6691
		const dummy_iframe = document.createElement("iframe");
		document.documentElement.appendChild(dummy_iframe);
		Array.from = dummy_iframe.contentWindow.Array.from;  // a clean working version
		//Array.prototype.from = dummy_iframe.contentWindow.Array.prototype.from;  // this is given in SO answer but doesn't work for some reason
		document.documentElement.removeChild(dummy_iframe);
	}

Testcase Gist URL

No response

Additional Information

No response

[billyvg]

Just ran into this with a few customers, an alternative solution would be to not use the 2nd argument and explicitly call map()

[eoghanmurray]

Actually I completely forgot I had this issue open, but suggested the same fix again which got merged with #1464 so this should no longer be an issue if you are using the latest code.
Open up a new issue or PR if 1464 doesn't fix things for you.