Matching behavior of SharedArrayBuffers in ES spec

[binji]

The current proposal very closely matches the behavior of SharedArrayBuffers in the ECMAScript spec. The goal is to make it easier to match the behavior of existing JS engines. Is this a good goal for v1, or is there more we should do?

[lars-t-hansen]

Matching the ES spec makes it easy to reason about. A superset of operations is not really a problem, but an "extended subset" type of situation should be a warning sign. IMO.

[binji]

This issue is phrased a bit vaguely. The question that was actually polled was:
"Poll: include same atomic and threading primitives as ES except is_lock_free"

This was described as being the minimum; we later agreed to add additional functionality, including i64 atomics, etc. Those features are covered in separate issues.