Hi Jens and Hubert,

Thank you for your helpful feedback. I will revise the paper for the September 2021 mailing to include these improvements to the wording.

Best wishes,

               Peter

> -----Original Message-----
> From: Jens Maurer <Jens.Maurer_at_[hidden]>
> Sent: 26 August 2021 06:46
>
> On 26/08/2021 03.42, Hubert Tong via SG16 wrote:
> > On Fri, Aug 20, 2021 at 12:02 PM JF Bastien via Ext
> <ext_at_[hidden] <mailto:ext_at_lists.isocpp.org>> wrote:
> >
> > I was asked in today's SG16 telecon to comment on the wording in P2362
> ahead of the EWG telecon tomorrow (that I unfortunately will be missing).
> >
> > For context, all changes in the paper are to [lex.ccon].
> >
> > In paragraph 2, the following additional change would be appropriate
> (since there is only one applicable base kind now):
> >
> > their respective base kinds => the base kind
> >
> > For the change to bullet 3.2.2, the discussion in the paper does not
> propose a change to the status quo regarding numeric escapes. The proposed
> change to this bullet should be omitted. That is, L'\x80000000' remains
> valid when wchar_t has a signed 32-bit underlying type.
> >
> > The rest of the changes appear sufficient and in line with my
> understanding of the consensus out of SG16.
>
> I would prefer to keep "ordinary" in the non-encodable and multicharacter
> literals.
> In my view, the "base kind" is the "ordinary character literal", and we
> slap
> adjectives on top of that. It is confusing if the term of the base kind
> changes.