| From: | Aleksander Alekseev <aleksander(at)timescale(dot)com> |
|---|---|
| To: | PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Cc: | Junwang Zhao <zhjwpku(at)gmail(dot)com> |
| Subject: | Re: [PATCH v1] PQputCopyEnd never returns 0, fix the inaccurate comment |
| Date: | 2023-08-28 11:48:05 |
| Message-ID: | CAJ7c6TOBaqrah+coyWEH4FvNUmdXd9=3WgALVNRPGOfjpYae2g@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hi Junwang,
> PQputCopyEnd returns 1 or -1, never 0, I guess the comment was
> copy/paste from PQputCopyData's comment, this should be fixed.
The patch LGTM but I wonder whether we should also change all the
existing calls of PQputCopyEnd() from:
```
PQputCopyEnd(...) <= 0
```
... to:
```
PQputCopyEnd(...) < 0
```
Given the circumstances, checking for equality to zero seems to be at
least strange.
On top of that, none of the PQputCopyData() callers cares whether the
function returns 0 or -1, both are treated the same way. I suspect the
function does some extra work no one asked to do and no one cares
about. Perhaps this function should be refactored too for consistency.
Thoughts?
--
Best regards,
Aleksander Alekseev
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Hayato Kuroda (Fujitsu) | 2023-08-28 11:48:14 | RE: persist logical slots to disk during shutdown checkpoint |
| Previous Message | Pavel Stehule | 2023-08-28 11:38:57 | Wrong usage of pqMsg_Close message code? |