Re: postgres_fdw: Obsolete comments in GetConnection()

From: Etsuro Fujita <etsuro(dot)fujita(at)gmail(dot)com>
To: Bharath Rupireddy <bharath(dot)rupireddyforpostgres(at)gmail(dot)com>
Cc: PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: postgres_fdw: Obsolete comments in GetConnection()
Date: 2021-10-06 11:26:34
Message-ID: CAPmGK175QfrLx38A0wa+BwmaogEs-xkpbw78h+zWzKWAo5nm9g@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi Bharath,

On Wed, Oct 6, 2021 at 6:37 PM Bharath Rupireddy
<bharath(dot)rupireddyforpostgres(at)gmail(dot)com> wrote:
> +1 for rewording the comments. Here are my thoughts on the patch:
>
> 1) Just to be consistent(we are using this word in the error message,
> and in other comments around there), how about
> + * Determine whether to try to reestablish the connection.
> instead of
> + * Determine whether to try to remake the connection later.

Actually, we use the word “remake” as well in comments in
connection.c: e.g., “If the connection needs to be *remade* due to
invalidation, disconnect as soon as we're out of all transactions.” in
GetConnection(). But I don’t have a strong opinion about that, so
I’ll change the word as proposed.

> 2) Just to be consistent, how about
> + * cases where we're starting new transaction (not subtransaction),
> if a broken connection is
> instead of
> + * cases where we're out of all transactions, if a broken connection is

Actually, I modified the comment to match existing comments like the
one mentioned above. I think the patch would actually be more
consistent.

> 3) IMO we don't need the word "later" here because we are immediately
> reestablishing the connection, if it is decided to do so.
> + * Determine whether to try to remake the connection later.

Ok, I’ll drop the word “later”.

Thanks!

Best regards,
Etsuro Fujita

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Andrew Dunstan 2021-10-06 11:33:22 Re: More business with $Test::Builder::Level in the TAP tests
Previous Message Bharath Rupireddy 2021-10-06 11:01:33 Re: can we add subscription TAP test option "vcregress subscriptioncheck" for MSVC builds?