On Sun, Dec 15, 2024 at 6:05 PM Noah Misch <noah(at)leadboat(dot)com> wrote:
> On Sun, Dec 15, 2024 at 04:50:57PM +1300, Thomas Munro wrote:
> > For the all-ASCII test, why not us pg_is_ascii()?
>
> pg_is_ascii() takes a char[], but this has a wchar_t[]. Maybe I should wrap
> that part in a pg_wcs_is_ascii()? (We can't use GetCommandLineA(), because
> "best fit" may have converted non-ASCII to ASCII.)
Right, sorry, brain fade. No issue with the open-coding here if you
want an early exit (I'd be tempted to just not have it for simplicity
but no strong opinion).