I'm not seeing a problem with the original formulation.

The expression x3d[i] is equivalent to *(x3d + i); in that expression, x3d is subject to the array-to-pointer conversion...

It's certainly true that in the expression "*(x3d + i)", x3d is subject to the array-to-pointer conversion. A small improvement might be to add a cross-reference to [expr.sub] after "equivalent to blah", to get the caveats visible.

I'm not seeing a problem with the original formulation.

I think you're right, at least with the current wording. If we say that it's equivalent to *(x3d + i), then array-to-pointer conversion really takes places.

If CWG2548 had been accepted, it would definitely be wrong though, as the wording would be

The result of the expression E1[E2] is identical (by definition) to that of *((E1)+(E2))

If only the result, but not the remaining semantics are equivalent, then array-to-pointer conversion doesn't really take place.

Even though you're right right now, I still don't favor the current wording. The note is intended to explain multi-dimensional subscript math, and this side tangent on array-to-pointer conversion isn't making it easier to understand. Depending on how [expr.sub] changes (and there have been a fair amount of historical/proposed changes), this side tangent becomes less relevant or simply wrong.

In short, the status quo of this pull request is not what we want.

I'm closing this because [dcl.array] already has a reference to [expr.sub], so neither the proposed changes nor the possible small alternative can be completed here.