On Mon, Mar 1, 2021 at 5:17 PM Hubert Tong <hubert.reinterpretcast@gmail.com> wrote:

On Mon, Mar 1, 2021 at 11:27 AM Steve Downey via SG16 <sg16@lists.isocpp.org> wrote:

Could we perhaps make use of the encoding used by the "C" locale to talk about how the "encoding of the execution character set" is meant to be interpreted? http://eel.is/c++draft/lex.ccon#2
Execution encoding isn't currently used in the standard as that exact phrase, although lex.ccon does come close, as does http://eel.is/c++draft/tab:lex.string.literal

The sentence above is using "encoding of the execution character set" for its position in the status quo of the working draft, right? That is, we should read it as saying that the "literal encoding" can be taken as the locale-specific encoding used in the C locale. In practice, that's not true (e.g., literals encoded as UTF-8 on systems with a C locale using US-ASCII). What is probably true is that the encoding difference is not observable if only characters from the basic execution character set are used. I think it is safe to say that there are seriously many scripts/programs that (ab)use text processing facilities via the property that "C" locales basically treat bytes as characters.

Yes, I'm trying to make sure we're agreeing what the status quo is that we're changing. Execution encoding isn't currently defined (although everyone thinks it is), so maybe we can just avoid defining it.

There's literal encoding and what locale provides for the associated encodings for char and wchar. That literal encoding may be different than that at runtime, and also from the "C" locale, is a large part of what we're trying to describe?