Editorial: inline infallible CreateDataPropertyOrThrow calls

[michaelficarra]

There's no reason to call CreateDataPropertyOrThrow when it cannot fail (marked with !). Instead, call CreateDataProperty directly.

[bakkot]

This no longer indicates to the reader that the property is guaranteed to be created, since CreateDataProperty can return false and there is nothing to indicate to the reader that it does not in these cases. I'm opposed.

[ljharb]

Indeed, CreateDataProperty can return false. This caused many spec bugs in the past when authors didn't realize that this needed to be explicitly checked, which was why they were all explicitly converted to ! CreateDataPropertyOrThrow.

In this case, your change would require an assertion that the result was true, which is currently lacking in the PR.

[michaelficarra]

Result from editor call today is that we should make a new AO for infallible CreateDataProperty and use that instead to make it more clear, but we'll wait to do that until after completion record reform is complete.

[ljharb]

@michaelficarra should this be left open until then? or closed?

[michaelficarra]

let's leave it open and do the work here