Horton

Thread

Commits

  1. Use palloc_object() and palloc_array() in more areas of the logical replication.

  1. Use allocation macros in the logical replication code

    Peter Smith <smithpb2250@gmail.com> — 2026-02-04T00:14:53Z

    Hi,

Here is a patch that modifies some allocation code to make use of the
appropriate macros.

Scope: src/backend/replication/logical/*

PSA v1

======
Kind Regards,
Peter Smith.
Fujitsu Australia

  2. Re: Use allocation macros in the logical replication code

    Chao Li <li.evan.chao@gmail.com> — 2026-02-04T01:04:57Z

    
> On Feb 4, 2026, at 08:14, Peter Smith <smithpb2250@gmail.com> wrote:
> 
> Hi,
> 
> Here is a patch that modifies some allocation code to make use of the
> appropriate macros.
> 
> Scope: src/backend/replication/logical/*
> 
> PSA v1
> 
> ======
> Kind Regards,
> Peter Smith.
> Fujitsu Australia
> <v1-0001-use-alloc-macros-in-logical-replication-code.patch>

LGTM.

BTW, I have a similar patch [1] there more than a month.

[1] https://www.postgresql.org/message-id/CAEoWx2m1Vo97Jg9%3DK7JAZ0xdkg5D%3DGkgOxZR1%3DEW7mUfy008fw%40mail.gmail.com

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/

  3. RE: Use allocation macros in the logical replication code

    Hayato Kuroda (Fujitsu) <kuroda.hayato@fujitsu.com> — 2026-02-04T02:47:16Z

    Dear Peter,

Thanks for the patch. I have one comment for the patch:
```
@@ -836,8 +836,9 @@ SnapBuildAddCommittedTxn(SnapBuild *builder, TransactionId xid)
 		elog(DEBUG1, "increasing space for committed transactions to %u",
 			 (uint32) builder->committed.xcnt_space);
 
-		builder->committed.xip = repalloc(builder->committed.xip,
-										  builder->committed.xcnt_space * sizeof(TransactionId));
+		builder->committed.xip = repalloc_array(builder->committed.xip,
+							TransactionId,
+							builder->committed.xcnt_space);
```

Number of tabs seems not to enough here. Other than that LGTM.

Best regards,
Hayato Kuroda
FUJITSU LIMITED

  4. Re: Use allocation macros in the logical replication code

    Peter Smith <smithpb2250@gmail.com> — 2026-02-04T05:40:12Z

    On Wed, Feb 4, 2026 at 1:47 PM Hayato Kuroda (Fujitsu)
<kuroda.hayato@fujitsu.com> wrote:
>
> Dear Peter,
>
> Thanks for the patch. I have one comment for the patch:
> ```
> @@ -836,8 +836,9 @@ SnapBuildAddCommittedTxn(SnapBuild *builder, TransactionId xid)
>                 elog(DEBUG1, "increasing space for committed transactions to %u",
>                          (uint32) builder->committed.xcnt_space);
>
> -               builder->committed.xip = repalloc(builder->committed.xip,
> -                                                                                 builder->committed.xcnt_space * sizeof(TransactionId));
> +               builder->committed.xip = repalloc_array(builder->committed.xip,
> +                                                       TransactionId,
> +                                                       builder->committed.xcnt_space);
> ```
>
> Number of tabs seems not to enough here. Other than that LGTM.
>

Thanks for the feedback. Now I have run pg_indent on all the patched files.

PSA v2

======
Kind Regards,
Peter Smith.
Fujitsu Australia

  5. Re: Use allocation macros in the logical replication code

    Masahiko Sawada <sawada.mshk@gmail.com> — 2026-03-03T20:52:26Z

    On Tue, Feb 3, 2026 at 9:40 PM Peter Smith <smithpb2250@gmail.com> wrote:
>
> On Wed, Feb 4, 2026 at 1:47 PM Hayato Kuroda (Fujitsu)
> <kuroda.hayato@fujitsu.com> wrote:
> >
> > Dear Peter,
> >
> > Thanks for the patch. I have one comment for the patch:
> > ```
> > @@ -836,8 +836,9 @@ SnapBuildAddCommittedTxn(SnapBuild *builder, TransactionId xid)
> >                 elog(DEBUG1, "increasing space for committed transactions to %u",
> >                          (uint32) builder->committed.xcnt_space);
> >
> > -               builder->committed.xip = repalloc(builder->committed.xip,
> > -                                                                                 builder->committed.xcnt_space * sizeof(TransactionId));
> > +               builder->committed.xip = repalloc_array(builder->committed.xip,
> > +                                                       TransactionId,
> > +                                                       builder->committed.xcnt_space);
> > ```
> >
> > Number of tabs seems not to enough here. Other than that LGTM.
> >
>
> Thanks for the feedback. Now I have run pg_indent on all the patched files.
>
> PSA v2

Thank you for the patch!

The patch looks good to me. I'll push it, barring objections.

Regards,

-- 
Masahiko Sawada
Amazon Web Services: https://aws.amazon.com

  6. Re: Use allocation macros in the logical replication code

    Chao Li <li.evan.chao@gmail.com> — 2026-03-04T01:39:07Z

    
> On Mar 4, 2026, at 04:52, Masahiko Sawada <sawada.mshk@gmail.com> wrote:
> 
> On Tue, Feb 3, 2026 at 9:40 PM Peter Smith <smithpb2250@gmail.com> wrote:
>> 
>> On Wed, Feb 4, 2026 at 1:47 PM Hayato Kuroda (Fujitsu)
>> <kuroda.hayato@fujitsu.com> wrote:
>>> 
>>> Dear Peter,
>>> 
>>> Thanks for the patch. I have one comment for the patch:
>>> ```
>>> @@ -836,8 +836,9 @@ SnapBuildAddCommittedTxn(SnapBuild *builder, TransactionId xid)
>>>                elog(DEBUG1, "increasing space for committed transactions to %u",
>>>                         (uint32) builder->committed.xcnt_space);
>>> 
>>> -               builder->committed.xip = repalloc(builder->committed.xip,
>>> -                                                                                 builder->committed.xcnt_space * sizeof(TransactionId));
>>> +               builder->committed.xip = repalloc_array(builder->committed.xip,
>>> +                                                       TransactionId,
>>> +                                                       builder->committed.xcnt_space);
>>> ```
>>> 
>>> Number of tabs seems not to enough here. Other than that LGTM.
>>> 
>> 
>> Thanks for the feedback. Now I have run pg_indent on all the patched files.
>> 
>> PSA v2
> 
> Thank you for the patch!
> 
> The patch looks good to me. I'll push it, barring objections.
> 

Hi Masahiko-san, no objection at all. When you have a chance, could you also take a look at [1]? It is a similar patch that fixes an incorrect usage of palloc_object() by switching it to palloc_array(), and also replaces a few repalloc() calls with repalloc_array().

[1] https://postgr.es/m/CAEoWx2m1Vo97Jg9=K7JAZ0xdkg5D=GkgOxZR1=EW7mUfy008fw@mail.gmail.com

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/

  7. Re: Use allocation macros in the logical replication code

    Peter Smith <smithpb2250@gmail.com> — 2026-03-09T05:22:36Z

    Thanks for pushing!

I have marked the corresponding CF entry as "Committed".

======
Kind Regards,
Peter Smith.
Fujitsu Australia.

  8. Re: Use allocation macros in the logical replication code

    Masahiko Sawada <sawada.mshk@gmail.com> — 2026-03-10T20:55:44Z

    On Sun, Mar 8, 2026 at 10:23 PM Peter Smith <smithpb2250@gmail.com> wrote:
>
> Thanks for pushing!
>
> I have marked the corresponding CF entry as "Committed".

I noticed I forgot to mention the patch has been pushed.

Thanks for the report and the patch!

Regards,

-- 
Masahiko Sawada
Amazon Web Services: https://aws.amazon.com