Re: ps command does not show walsender's connected db - Mailing list pgsql-hackers

From bt22nakamorit
Subject Re: ps command does not show walsender's connected db
Date
Msg-id 83fa58f8f16ee6157d1120a06551d4bb@oss.nttdata.com
Whole thread Raw
In response to Re: ps command does not show walsender's connected db  (Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>)
Responses Re: ps command does not show walsender's connected db  (Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>)
List pgsql-hackers
2022-10-09 18:30 Bharath Rupireddy wrote:
> -    if (!am_walsender)
> +    if (!am_walsender || am_db_walsender)
>          appendStringInfo(&ps_data, "%s ", port->database_name);
> 
> Can the appendStringInfo be just unconditional? That is more readable
> IMO. We want the database_name to be appended whenever it isn't null.
I agree that the patch makes the code less neat.

> The only case we expect database_name to be null is for walsenders
> serving streaming replication standbys and even when database_name is
> null, nothing gets appended, it's just the appendStringInfo() call
> gets wasted, but that's okay.
appendStringInfo will append extra space after null (since "%s "), so 
the ps entry will look less neat in that case.
How about we check whether port->database_name is null or not, instead 
of making it unconditional?
It will look like this.
-    if (!am_walsender)
+    if (port->database_name != NULL)
         appendStringInfo(&ps_data, "%s ", port->database_name);

Thank you for your response,
Tatsuhiro Nakamori



pgsql-hackers by date:

Previous
From: Andy Fan
Date:
Subject: Re: Unify "In" Sublink to EXIST Sublink for better optimize opportunity
Next
From: Richard Guo
Date:
Subject: Unnecessary lateral dependencies implied by PHVs