Hey.

Agreed, it's a byte thing - actually, have you considered std::byte ?

You showed that it is trivial to programmatically add a null terminator, which seems sufficient to cover all use cases

BUT

We need a (completely orthogonal) feature which should have no bearing on your paper: std::trust_me_these_bytes_are_utf8

On Wed, 6 Nov 2019 at 13:12, JeanHeyd Meneide <phdofthehouse@gmail.com> wrote:

Dear SG16,

     I presented #embed_str to EWG. They said I need to take a trip through SG16.

    The semantics of #embed_str are that the contents of the file are loaded up as individual entries in a regular string literal. For example, if the contents of a file were

foo

Then the array that backs the string literal would be loaded as

[f | o | o | \0]
[0 | 1 | 2 | 3]

     A better name for it might be #embed_null_terminated`. I don't think this has much to do with Unicode at the end of the day, because it deals with code units. There would be no embed_u8str or embed_u16str or embed_u32str, because there's no guarantee the contents of the file would be valid UTFX and I am not about to get into the mess that is "source resource encoding" and "compile-time resource encoding" conversions.

     If #embed_str is too suggestive of text, I will be more than happy to put an axe through it. Do let me know how to proceed.

Sincerely,
JeanHeyd Meneide
_______________________________________________
SG16 Unicode mailing list
Unicode@isocpp.open-std.org
http://www.open-std.org/mailman/listinfo/unicode