Re: Order changes in PG16 since ICU introduction

On 05.06.23 19:54, Jeff Davis wrote:
>
> New patch series attached. I plan to commit 0001 and 0002 soon, unless
> there are objections.
>
> 0001 causes the "C" and "POSIX" locales to be treated with
> memcmp/pg_ascii semantics in ICU, just like in libc. We also considered
> a new "none" provider, but it's more invasive, and we can always
> reconsider that in the v17 cycle.
>
> 0002 introduces an upgrade check for users who have explicitly
> requested provider=icu and iculocale=C on older versions, and rejects
> upgrading from v15 in that case to avoid index corruption. Having such
> a collation is almost certainly a mistake by the user, because the
> collator would not give the expected memcmp semantics.

I'm dubious about these two.

0003 seems like the correct direction. In createdb.c, the change you
add makes sense, but you should also remove the existing use of the
locale variable:

- if (locale)
- {
- if (!lc_ctype)
- lc_ctype = locale;
- if (!lc_collate)
- lc_collate = locale;
- }
-