Should ranges validate their subtypes' domains after canonicalization?

Hi Hackers,

While reviewing
https://www.postgresql.org/message-id/CACJufxGoAmN_0iJ%3DhjTG0vGpOSOyy-vYyfE%2B-q0AWxrq2_p5XQ%40mail.gmail.com,
I noticed this inconsistency in how ranges enforce domains over their
subtypes:

```
[v19devel:5432][426675] postgres=# create type int4_d_range;
CREATE TYPE
[v19devel:5432][426675] postgres=# create or replace function
int4_d_range_canonical(int4_d_range) returns int4_d_range language
internal immutable parallel safe strict as
$function$int4range_canonical$function$;
NOTICE: argument type int4_d_range is only a shell
LINE 1: create or replace function int4_d_range_canonical(int4_d_ran...
^
NOTICE: return type int4_d_range is only a shell
CREATE FUNCTION
[v19devel:5432][426675] postgres=# create type int4_d_range as range
(subtype = int4_d, canonical = int4_d_range_canonical);
CREATE TYPE
[v19devel:5432][426675] postgres=# select '[9,10)'::int4_d_range;
ERROR: value for domain int4_d violates check constraint "int4_d_check"
LINE 1: select '[9,10)'::int4_d_range;
^
[v19devel:5432][426675] postgres=# select '[9,9]'::int4_d_range;
int4_d_range
--------------
[9,10)
(1 row)
```

The same inconsistency happens without a canonical function, but it
seems okay there, based on this line from the docs:

> If a canonicalization function is not specified, then ranges with different formatting will always be treated as unequal, even though they might represent the same set of values in reality.

But should a range validate its domain after canonicalizing? It seems
like it is happening before.

Yours,

--
Paul ~{:-)
pj(at)illuminatedcomputing(dot)com