[Bug] Query fails to execute when exceeding the PostgreSQL 63 byte limit on identifiers

[therealbrad]

Description and expected behavior
Executing a deep-nested query exceeds the 63 byte limit on identifiers in PostgreSQL.

Failed to execute query: Error: Failed to execute query: error: table name 
"RepositoryCases$template$caseFields$caseField$fieldOptions$sub$" 
specified more than once

Query that triggered it (ZenStack auto-generated hook useFindFirstRepositoryCases):

const { data } = useFindFirstRepositoryCases({
  where: { id: caseId },
  include: {
    template: {
      include: {
        caseFields: {
          include: {
            caseField: {
              include: {
                fieldOptions: {
                  include: {
                    fieldOption: true  // 5 levels deep
                  }
                }
              }
            }
          }
        }
      }
    }
  }
});

REST API endpoint: GET /api/model/repositoryCases/findFirst?q={...}

Server-side (ZenStack v3 RPCApiHandler): The handler receives this query and translates it to Kysely SQL, generating aliases like:

RepositoryCases$template$caseFields$caseField$fieldOptions$fieldOption (73 chars)

Why it fails: The generated alias RepositoryCases$template$caseFields$caseField$fieldOptions$fieldOption is 73 characters. PostgreSQL truncates at 63 chars (NAMEDATALEN-1). When ZenStack also generates a $sub variant, both truncate to the same 63 chars, causing the "table name specified more than once" error.

Schema models involved:

RepositoryCases -> template (Templates) -> caseFields (TemplateCaseAssignment) -> caseField (CaseFields) -> fieldOptions (CaseFieldAssignment) -> fieldOption (FieldOptions)

[ymc9]

Fixed in 3.4.0