Testing the "Resizable ArrayBuffer" proposal

[jugglinmike]

https://github.com/tc39/proposal-resizablearraybuffer

This proposal advanced to stage 3 on 2021-05-25

[jugglinmike]

I've begun writing these tests

[jugglinmike]

I've submitted these tests for review, split up across 11 pull requests in order to facilitate incremental review (the pull requests are listed above).

There's one design decision which effected multiple patches, so to keep things organized, I'll describe it here:

Hosts are permitted to fail any resize operation at their own discretion. This is fairly straightforward to accommodate in direct tests (e.g. those for ArrayBuffer.prototype.resize), but it's more difficult to address for the features which integrate.

The approach I've settled on involves dynamically assigning assertion expectation values within try/catch statements. This is most jarring in those tests which concern abrupt completions: they use Test262Error and assert.throws to verify that a given invocation does not throw. In all cases, the design is intended to maximize coverage (by validating something for the "host-defined failure" case) while limiting the amount of branching in the test body (which is generally an invitation for test bugs and reader confusion).

This is all a little unorthodox, but I feel it's warranted by the proposal's unique behavior.

/cc @syg