release-26.1: sql: fix SHOW CREATE FUNCTION with column casting to UDT

[blathers-crl]

Backport 1/1 commits from #159642 on behalf of @ZhouXing19.

---

Fixes #152233
Fixes #159145
Fixes #87475

Previously, if a function contains cast from a column to UDT in the body,
the SHOW CREATE TABLE statement for this function will error out. This
commit is to fix it.

Release note (bug fix): enable SHOW CREATE TABLE for function created with
columns casting to user defined types.

---

Release justification: fix a bug for a common usage of UDF

[blathers-crl]

Thanks for opening a backport.

Before merging, please confirm that the change does not break backwards compatibility and otherwise complies with the backport policy. Include a brief release justification in the PR description explaining why the backport is appropriate. All backports must be reviewed by the TL for the owning area. While the stricter LTS policy does not yet apply, please exercise judgment and consider gating non-critical changes behind a disabled-by-default feature flag when appropriate.

[cockroach-teamcity]

This change is 

[michae2]

@michae2 reviewed 5 files and all commit messages, and made 1 comment.
Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @ZhouXing19).

---

pkg/sql/show_create_clauses.go line 425 at r1 (raw file):

		}

		dEnum, ok := typedExpr.(*tree.DEnum)

Could this also be a DTuple (for a user-defined composite type)?

(I'm trying to understand how the cast_to_composite testcase passes if we don't handle composite types here...)

[ZhouXing19]

pkg/sql/show_create_clauses.go line 425 at r1 (raw file):

Previously, michae2 (Michael Erickson) wrote…

Could this also be a DTuple (for a user-defined composite type)?

(I'm trying to understand how the cast_to_composite testcase passes if we don't handle composite types here...)

Sorry, I realized I wrote a silent bug here.

For column_name::composite_type casts, the current code would actually hit a "column doesn't exist" error in the earlier expr.TypeCheck() call because we lack table descriptor context. So we didn't even reach here, but already exited in the previous error check. But since we use simple visitor to walk the statement, this error is silently swallowed and we just stopped walking the remaining statement, so we just returned the expr as-is - that's why the cast_to_composite test still passed.

This enum-specific logic is only needed for const::enum situations where we need to deserialize binary enum data back to human-readable strings. For composite types, the original expr is already of human-readable string, so no need for this step.

Fixed this in PR #160424 by explicitly handling the UndefinedColumn error and adding proper support for non-enum UDT types.