| From: | Chapman Flack <chap(at)anastigmatix(dot)net> |
|---|---|
| To: | Peter Eisentraut <peter(dot)eisentraut(at)enterprisedb(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Documentation about PL transforms |
| Date: | 2022-02-21 21:09:52 |
| Message-ID: | 6213FFA0.1080202@anastigmatix.net |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 02/07/22 15:14, Chapman Flack wrote:
> I'll work on some doc patches.
It seems a bit of an impedance mismatch that there is a get_func_trftypes
producing a C Oid[] (with its length returned separately), while
get_transform_fromsql and get_transform_tosql both expect a List.
There is only one in-core caller of get_func_trftypes (it is
print_function_trftypes in ruleutils.c), while both in-core PLs that
currently support transforms are reduced to rolling their own List by
grabbing the protrftypes Datum and hitting it with oid_array_to_list.
Would it be reasonable to deprecate get_func_trftypes and instead
provide a get_call_trftypes (following the naming of get_call_result_type
where _call_ encompasses functions and procedures) that returns a List,
and use that consistently?
Regards,
-Chap
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Chris Cleveland | 2022-02-21 21:16:31 | External data files possible? |
| Previous Message | Simon Riggs | 2022-02-21 21:04:19 | Re: Reducing power consumption on idle servers |