> Maybe it doesn't matter for the purposes of SD-6, but for P0220R1
> the shared_ptr changes for arrays were approved in Jacksonville, but
> are not in the CD due to editorial conflicts. P0414R0 aims to
> correct that.
>
> Similarly, P0067R3 was approved in Oulu, but isn't in the CD because
> the editor noticed inconsistencies in the wording which mean it
> needs to be revised and go through LWG again.
>
> If those features do make it back into the WP in time for C++17 the
> date 201603 might not make sense for their macros. On the other hand
> maybe it does still make sense, as the substance of the proposals
> will still be what was voted on in Jacksonville and Oulu
> respectively.

Hmm. Thanks for pointing these out.

For P0067R3, which wasn't applied to the WD at all, the interesting
question will be when and how R4 (or successor) makes it into the WD.
If it's voted in at a subsequent meeting, we should probably use the date of
that vote. But if the changes relative to R3 are considered to be editorial,
there might conceivably not be another WG21 vote, in which case we should
use the Oulu date. (But in either case, are we really going to try adding it
to C++17 after the CD goes out?)

All the same considerations apply to P0220R1, but it's even a little more
complicated because it was only part of the paper.

For the time being, I'm inclined to omit the row for P0067; I can always
resurrect it later.

For P0220, because of the complications of the row-spanning in the table,
I'd kind of rather not try to branch-predict. I think I'll touch base with
Herb about his expectations.

Clark