Re: CREATE COLLATION does not sanitize ICU's BCP 47 language tags. Should it?

On 9/19/17 22:01, Peter Geoghegan wrote:
> On Tue, Sep 19, 2017 at 5:52 PM, Peter Eisentraut
> <peter(dot)eisentraut(at)2ndquadrant(dot)com> wrote:
>> On 9/18/17 18:46, Peter Geoghegan wrote:
>>> As I pointed out a couple of times already [1], we don't currently
>>> sanitize ICU's BCP 47 language tags within CREATE COLLATION.
>>
>> There is no requirement that the locale strings for ICU need to be BCP
>> 47. ICU locale names like 'de(at)collation=phonebook' are also acceptable.
>
> Right. But, we only document that BCP 47 is supported by Postgres.

The documentation is admittedly not very concrete about what ICU locale
names it accepts beyond talking about a "named collator provided by the
ICU library". The examples we provide use the BCP 47 style, but that's
just because we liked them that way.

ICU <54 doesn't even support the BCP 47 style, so we need to keep
supporting the old/other style anyway.

> And, I think that I
> see a bigger problem: we pass a string that is almost certainly a BCP
> 47 string to ucol_open() from within pg_newlocale_from_collation(). We
> do so despite the fact that ucol_open() apparently doesn't accept BCP
> 47 syntax locale strings until ICU 54 [1].

pg_import_system_collations() takes care to use the non-BCP-47 style for
such versions, so I think this is working correctly.

--
Peter Eisentraut http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services