Re: Add standard collation UNICODE

On 04.03.23 19:29, Jeff Davis wrote:
> I do like your approach though because, if someone is using a standard
> collation, I think "not built with ICU" (feature not supported) is a
> better error than "collation doesn't exist". It also effectively
> reserves the name "unicode".

By the way, speaking of reserving names, I don't remember the reason for
this bit in initdb.c:

/*
* Add SQL-standard names. We don't want to pin these, so they don't go
* in pg_collation.h. But add them before reading system collations, so
* that they win if libc defines a locale with the same name.
*/

Why don't we want them pinned?

If we add them instead as entries into pg_collation.dat, it seems to
work for me.

Another question: What is our current thinking on using BCP 47 names?
The documentation says for example

"""
The first example selects the ICU locale using a “language tag” per BCP
47. The second example uses the traditional ICU-specific locale syntax.
The first style is preferred going forward, but it is not supported by
older ICU versions.
"""

My patch uses 'und' [BCP 47 style], which appears to be in conflict with
that statement.

But we have had some discussions on how correct that statement is, but I
don't remember the outcome.