I am wondering what the best way to read a unicode file into char8_t character strings would be
before Corentin’s paper lands on casting to unicode characters lands (P2626?).

The best strategy I can see is:

1) Read a file into a `char` buffer, as most OS file system APIs traffic in `char *`
2) Validate that the buffer contents are valid UTF-8
3) call `start_lifeimte_as_aray<char8_t>` with the char buffer
4) leak the trivial array object, but carefully manage the lifetime of the character buffer

Does that sound reasonable? Do we have better ideas using just standard components today?

I am also wondering if `start_lifetime_as_array` is the missing component from when
Corentin wrote his original paper?

Of course, the main problem with my plan is that no-one implements start_lifeimte_as_aray yet!
Do we have any good answers while waiting for the standard libraries to catch up?

AlisdairM