Complete PostgreSQL CREATE TYPE Support

[LucaCappelletti94]

Adds full parsing for all PostgreSQL CREATE TYPE statement variants. Previously only composite and enum types were supported.

What's New

Range Types - Parse range type definitions with 6 configuration options

CREATE TYPE int4range AS RANGE (SUBTYPE = int4, CANONICAL = fn);

Base Types - Parse low-level type definitions with 19 options (INPUT/OUTPUT functions, ALIGNMENT, STORAGE, etc.)

CREATE TYPE complex (INPUT = in_fn, OUTPUT = out_fn, INTERNALLENGTH = 16);

Simple Declarations - Parse type shells without representation

CREATE TYPE typename;

The AST extends UserDefinedTypeRepresentation with three new variants (Range, SqlDefinition, None) and adds five supporting enums (UserDefinedTypeRangeOption, UserDefinedTypeSqlDefinitionOption, Alignment, UserDefinedTypeStorage, UserDefinedTypeInternalLength) with comprehensive rustdoc. The parser refactors parse_create_type() into modular functions (parse_create_type_composite(), parse_create_type_range(), and base type option parsers) that handle all PostgreSQL type configuration syntax. Tests provide comprehensive coverage with explicit AST verification for all variants and edge cases. Added 17 PostgreSQL-specific keywords, necessary to parse the keywords used in the type options (ALIGNMENT, CANONICAL, INTERNALLENGTH, PASSEDBYVALUE, RECEIVE, SEND, STORAGE, SUBTYPE, TYPMOD_IN, etc.).

Closes issue #2092

[iffyio]

Thanks @LucaCappelletti94! Changes look good to me overall, left some comments

[iffyio]

Could we include a link to the postgres docs for the new variants?

[LucaCappelletti94]

I wanted to, but they do not have a linkable subsection so the link would be to the same CREATE TYPE page. The sub-headers do not have the anchor link unfortunately.

[iffyio]

ah I think in that case linking to the postgres create type page for each variant would suffice, most important that its clear where each variant comes from especially if the overlying statement/enum is shared between multiple dialects. Basically similar to what the Enum { ... } variant above documents?

[LucaCappelletti94]

Ok, doing that.

[LucaCappelletti94]

Added

[iffyio]

can we add tests for SUBTYPE_OPCLASS,SUBTYPE_DIFF and any other variants that are missing coverage?

[LucaCappelletti94]

Oke

[LucaCappelletti94]

Tentatively done, I think all are covered now.

[iffyio]

LGTM! Thanks @LucaCappelletti94!