On Tue, Sep 8, 2020 at 1:57 PM Peter Brett via SG16 <email@example.com> wrote:
Having reviewed the paper, I’m struggling to understand how most of those concerns are pertinent to discussing P2178R1 proposal 1.This is possibly about being clear in terms of proactively dealing with assumptions that non-committee members may have when they hear the "UTF-8" source is "completely okay" for C++. Tom's questions would point out all sorts of caveats. For example, the compiler might say that UTF-8 source is supported with a flag that causes all file processing for that invocation to need UTF-8 source. This is going to cause problems for header inclusion.
Hubert's perception is exactly right. I'm raising these
questions because I believe more analysis is needed before we
proceed in any specific direction. Proposal 1 lacks sufficient
analysis to inform direction other than to say, "we want to
support UTF-8". That is ok; proposal 1 clearly wasn't intended as
a final proposal as written. It is clear to me that we have
consensus for support of UTF-8, but there are devils lurking in
the details that we have yet to exorcise. The discussion tomorrow
will focus on those devils.
That is possible. I am ok with not getting to P2194 tomorrow, or only getting a start on it, if the UTF-8 discussion is productive. If it becomes unproductive, we'll switch.
I’m not adverse to talking about them, because they are important and need to be addressed at some point, but it feels like giving them the attention that they deserve would not leave time for discussing P2194R0.
Please could we consider scheduling a discussion of these points for another meeting when your draft paper is ready to discuss in detail?
My goal is that the discussion will help to inform further
development of that draft paper and to attract collaborators.
Just as P2194 is the evolution of P2178 proposal 9, I hope that
draft will become the evolution of proposal 1. That implies that
it must reflect consensus opinions as well as dissenting ones and
offer a choice of options to present to EWG.
SG16 mailing list