| From: | shveta malik <shveta(dot)malik(at)gmail(dot)com> |
|---|---|
| To: | "Zhijie Hou (Fujitsu)" <houzj(dot)fnst(at)fujitsu(dot)com> |
| Cc: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>, shveta malik <shveta(dot)malik(at)gmail(dot)com> |
| Subject: | Re: Remove stale XXX comment in logical launcher |
| Date: | 2026-09-03 06:26:03 |
| Message-ID: | CAJpy0uA+6X9J1j5on6Shh9nYGA6f70sL9SipuJXea3vo62cBrA@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Thu, Sep 3, 2026 at 9:12 AM Zhijie Hou (Fujitsu)
<houzj(dot)fnst(at)fujitsu(dot)com> wrote:
>
> Hi,
>
> I noticed one comment in the launcher that can be removed:
>
> if (!retain_dead_tuples)
> /* XXX unclear why we don't request logical decoding disable */
> ReplicationSlotDropAcquired(false);
>
> Since the conflict detection slot is a physical one, there is no logical
> decoding state to disable when dropping it.
I agree. However, the comments never mention that this is a physical
slot, making it easy to misconstrue it as a logical one. Do you think
we can add the keyword 'physical' in the comment above this Drop()
or/and above CreateConflictDetectionSlot or/and above macro
CONFLICT_DETECTION_SLOT?
> The condition it depends on is also
> self-evident from the surrounding code and the block comment above, which
> already explains why the slot is dropped. So here is a patch to remove the XXX
> comment.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Alexander Korotkov | 2026-09-03 06:26:42 | Re: gist_trgm_ops '=' operator: planner picks it over btree, ~300x slower |
| Previous Message | Nisha Moond | 2026-09-03 06:05:36 | Re: Fix resource leak in FindConflictTuple() retry path |