Re: ICU for global collation

On 27.01.22 09:10, Peter Eisentraut wrote:
> On 21.01.22 17:13, Julien Rouhaud wrote:
>> On Fri, Jan 21, 2022 at 03:24:02PM +0100, Peter Eisentraut wrote:
>>> On 21.01.22 14:51, Julien Rouhaud wrote:
>>>> Is that change intended?  There isn't any usage of the
>>>> collversionstr before
>>>> the possible error when actual_versionstr is missing.
>>>
>>> I wanted to move it closer to the SysCacheGetAttr() where the "datum"
>>> value
>>> is obtained.  It seemed weird to get the datum, then do other things,
>>> then
>>> decode the datum.
>>
>> Oh ok.  It won't make much difference performance-wise, so no objection.
>
> I have committed this and will provide follow-up patches in the next few
> days.

Here is the main patch rebased on the various changes that have been
committed in the meantime. There is still some work to be done on the
user interfaces of initdb, createdb, etc.

I have split out the database-level collation version tracking into a
separate patch [0]. I think we should get that one in first and then
refresh this one.

[0]:
https://www.postgresql.org/message-id/flat/f0ff3190-29a3-5b39-a179-fa32eee57db6%40enterprisedb.com