Re: pg_hba_lookup function to get all matching pg_hba.conf entries

On Tue, Feb 2, 2016 at 8:57 AM, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> wrote:
> Haribabu Kommi wrote:
>
>> Hi, Do you have any further comments on the patch that needs to be
>> taken care?
>
> I do. I think the jsonb functions you added should be added to one of
> the json .c files instead, since they seem of general applicability.

moved these functions into jsonb_util.c file.

> But actually, I don't think you have ever replied to the question of why
> are you using JSON in the first place; isn't a normal array suitable?

It was discussed and told to use JSON for options instead of array in [1],
because of that reason I changed.

> The callback stuff is not documented in check_hba() at all. Can you
> please add an explanation just above the function, so that people trying
> to use it know what can the callback be used for? Also a few lines
> above the callback itself would be good.

Added some details in explaining the call back function.

>Also, the third argument of
> check_hba() is a translatable message so you should enclose it in _() so
> that it is picked up for translation. The "skipped"/"matched" words
> (and maybe others?) need to be marked similarly.

Changed mode column (skipped/matched) and reason for mismatch details
are enclosed in _() for translation. Do you find any other details needs to be
enclosed?

> That "Failed" in the errmsg in pg_hba_lookup() should be lowercase.

corrected.

> Moving to next CF.

Thanks. updated patch is attached with the above corrections.
This patch needs to be applied on top discard_hba_and_ident_cxt patch
that is posted earlier.

[1] - http://www.postgresql.org/message-id/5547DB0A.2020904@gmx.net

Regards,
Hari Babu
Fujitsu Australia