Adapt to override-mistake-fix pattern

[erights]

Fixes https://github.com/ljharb/object.assign/issues/79

Detect when the getter of an accessor alleges that it was originally a data property. This enables a leaky and cooperative partial emulation of the original data property, as code such as this uses this property to uphold the illusion. See https://github.com/ljharb/object.assign/issues/79 and https://github.com/Agoric/SES-shim/pull/473#issuecomment-697187652

Closes #113. Closes #114.

[erights]

See also https://github.com/Agoric/SES-shim/pull/475

[ljharb]

Good point; es-abstract works on versions of node that lack Symbols entirely. You could do typeof Symbol === 'function' && typeof Symbol.for('function') ? Symbol.for(x) : x in SES, and the same here?

[erights]

Hi @ljharb , ready for review. I will not merge though until https://github.com/Agoric/SES-shim/pull/476 is merged. Thanks.

Attn @kriskowal

[erights]

Hi @ljharb , ready for review. I will not merge though until Agoric/SES-shim#476 is merged. Thanks.

Sorry, habit. I will not merge this one period. It's a pull request.

[erights]

Hi @ljharb , we just published SES 0.10.4 to npm. which contains the override-mistake fix that this PR assumes. It also removes Function.prototype.apply from the whitelist, which was the particular override mistake fix causing trouble. Can we proceed on this? If not, any suggestions? Thanks.

Attn @kriskowal

[ljharb]

I'll add "running SES" to the tests; if they pass with or without this PR (as they should without the apply issue) then we can certainly land this PR as well (ie, once SES tests are verifying it). I'll look into it soon.

[erights]

Thanks!

[kumavis]

thanks for getting this in! @ljharb @erights