Re: [HACKERS] Patch to address concerns about ICU collcollatestability in v10 (Was: CREATE COLLATION does not sanitize ICU's BCP 47language tags. Should it?)

On Mon, Sep 25, 2017 at 11:42 AM, Peter Eisentraut
<peter.eisentraut@2ndquadrant.com> wrote:
> On 9/25/17 00:24, Peter Geoghegan wrote:
>> * Creates root collation as root-x-icu (collcollate "root"), not
>> und-x-icu. "und" means undefined language.
>
> I'm curious about this point.  "und" is defined in BCP 47.  I don't see
> "root" defined anywhere.  ICU converts the root collation to "und",
> AFAIK, so it seems to agree with the current naming.

In my patch, "root" is a string that is passed to get a language tag.
That's technically in the old format.

I think that this is another ICU vs. UCA/CLDR thing (this causes much
confusion). Note that "root" is mentioned in the ICU locale explorer,
for example: https://ssl.icu-project.org/icu-bin/locexp

Note also that ucol_open() comments/docs say this:
* @param loc The locale containing the required collation rules.*            Special values for locales can be passed
in-*            if NULL is passed for the locale, the default locale*            collation rules will be used. If empty
string("") or*            "root" are passed, UCA rules will be used.
 

I went with "root" because that produces a meaningful/useful display
name for pg_collation, and seems to be widely used elsewhere.

-- 
Peter Geoghegan


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers