On Mon, 2023-05-08 at 14:59 -0700, Jeff Davis wrote:
> The easiest thing to do is revert it for now, and after we sort out
> the
> memcmp() path for the ICU provider, then I can commit it again (after
> that point it would just be code cleanup and should have no
> functional
> impact).
The conversion won't be entirely dead code even after we handle the "C"
locale with memcmp(): for a locale like "C.UTF-8", it will still be
passed to the collation provider (same as with libc), and in that case,
we should still convert that to a language tag consistently across ICU
versions.
For it to be entirely dead code, we would need to convert any locale
with language "C" (e.g. "C.UTF-8") to use the memcmp() path. I'm fine
with that, but that's not what the libc provider does today, and
perhaps we should be consistent between the two. If we do leave the
code in place, we can document that specific "en-US-u-va-posix" locale
so that it's not too surprising for users.
Regards,
Jeff Davis