Message-ID: <381eb298-92b3-9599-abe6-f04c9a151248@uni-muenster.de>
Date: Thu, 14 Sep 2023 13:33:04 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.15.0
Subject: Re: [PATCH] Add inline comments to the pg_hba_file_rules view
To: Michael Paquier <michael@paquier.xyz>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>,
 David Zhang <david.zhang@highgo.ca>
References: <4d623899-36ac-71b5-311d-2a4672d75736@uni-muenster.de>
 <ZP5ELFdn28Hn1Hfi@paquier.xyz>
Content-Language: en-US
From: Jim Jones <jim.jones@uni-muenster.de>
In-Reply-To: <ZP5ELFdn28Hn1Hfi@paquier.xyz>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <https://www.postgresql.org/message-id/381eb298-92b3-9599-abe6-f04c9a151248%40uni-muenster.de>
Precedence: bulk

Hi

On 11.09.23 00:33, Michael Paquier wrote:
> Well, it looks like what I wrote a couple of days ago was perhaps
> confusing:
> https://www.postgresql.org/message-id/ZPHAiNp%2ByKMsa/vc%40paquier.xyz
> https://www.postgresql.org/message-id/ZPE8A7EnUH+ax5kw@paquier.xyz
>
> This patch touches hbafuncs.c and the system view pg_hba_file_rules,
> but I don't think this stuff should touch any of these code paths.
> That's what I meant in my second message: the SQL portion should be
> usable for all types of configuration files, even pg_ident.conf and
> postgresql.conf, and not only pg_hba.conf.  A new SQL function
> returning a SRF made of the comments extracted and the line numbers
> can be joined with all the system views of the configuration files,
> like sourcefile and sourceline in pg_settings, etc.
> --
> Michael

Thanks for the feedback.

I indeed misunderstood what you meant in the other thread, as you 
explicitly only mentioned hba.c.

The change to hbafunc.c was mostly a function call and a new column to 
the view:


comment = GetInlineComment(hba->rawline);
if(comment)
    values[index++] = CStringGetTextDatum(comment);
else
    nulls[index++] = true;


Just to make sure I got what you have in mind: you suggest to read the 
pg_hba.conf a second time via a new (generic) function like 
pg_read_file() that returns line numbers and their contents (+comments), 
and the results of this new function would be joined pg_hba_file_rules 
in SQL. Is that correct?

Thanks