Re: proposal: change behavior on collation version mismatch - Mailing list pgsql-hackers

From Jeff Davis
Subject Re: proposal: change behavior on collation version mismatch
Date
Msg-id f3198556e779ac887438deb7980a7eb923da479c.camel@j-davis.com
Whole thread Raw
In response to Re: proposal: change behavior on collation version mismatch  (Laurenz Albe <laurenz.albe@cybertec.at>)
List pgsql-hackers
On Mon, 2023-11-27 at 20:19 +0100, Laurenz Albe wrote:
> I forgot to add that the problem will remain a problem until the
> day we start keeping our own copy of the ICU library in the source
> tree...

Another option is for packagers to keep specific ICU versions around
for an extended time, and make it possible for Postgres to link to the
right one more flexibly (e.g. tie at initdb time, or some kind of
multi-lib system).

Even if ICU is available, we still have the problem of defaults. initdb
defaults to libc, and so does CREATE COLLATION (even if the database
collation is ICU). So it will be a long time before it's used widely
enough to consider the problem solved.

And even after all of that, ICU is not perfect, and our support for it
still has various rough edges.

Regards,
    Jeff Davis




pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Missing docs on AT TIME ZONE precedence?
Next
From: Tom Lane
Date:
Subject: Re: Partial aggregates pushdown