Add presence check for property cause on Error.prototype

[legendecas]

Property cause should not present on Error.prototype.

[rwaldron]

• Some of this duplicates existing tests.
• Please put these in separate files, in the appropriate directories (where you will find that some of this is already covered)

[jugglinmike]

Thanks for the patch! Hosts are permitted to define an Error.prototype.cause property, so we should remove the assertion for its absence.

[legendecas]

Thanks for the suggestions! Well, I didn't notice that there are statements about host extensions on built-in objects. I was noticing v8 is putting a cause with undefined value on Error.prototype which is not expected. If spec allows such extension, I'd close this PR and just file a bug to v8.

[ljharb]

I added tc39/agendas@964d2d3 to the next meeting to discuss tightening restrictions on hosts, so that helpful tests like this can be permitted.

[leobalter]

@ljharb can you share what do you suggest/recommend as the next steps for the restrictions?

The issue on this one is that ECMAScript allows extensions. This is imperative as globals add names for the platform it is contained, and other names are not exactly defined by ECMAScript. Otherwise, it would be easy to enumerate the names expected in Error.prototype, but if the host adds anything (not limited to cause), how is that an error?

[ljharb]

Allowing new globals, as well as property names on any object that are not specified in 262 or 402, is fine.

My hope is that we can forbid extensions on any specified objects whose names are specified elsewhere in 262 or 402, which would include “cause” on any language builtins other than the specified ones.