Re: Make SIGHUP less painful if pg_hba.conf is not readable

Selena Deckelmann wrote:
> Tom Lane wrote:
>> Selena Deckelmann <selena(at)endpoint(dot)com> writes:
>>> From the comment in hba.c, it appears that the desired behavior is to
>>> have the system ignore the failure,
>> I'm not sure how you could possibly read that comment that way.
>
> Right. Sorry, poor choice of words. I meant "don't die on reload",
> essentially.

The comment is wrong, the patch is correct.

>> It might be sane to distinguish initial load from reload, but I think
>> the behavior is correct as-is for initial load.
>
> Agreed.

The patch solves this perfectly fine. The caller takes care of sending a
FATAL error if it's in startup mode.

>> Also, if we are going to do something like this, we should make sure
>> it's consistent across all the config files.
>
> Ok. I can do that. I'll check with some other people before I send
> another patch, and I'll go through the rest of the config file loads
> tomorrow.

From what I can tell, it is already consistent (once this fix is
applied). Permissions wrong on postgresql.conf already sends a warning
and not a FATAL. Same for pg_ident.conf.

So. I've updated the comment, and applied your patch. Thanks!

(You also added a trailing space on the if line - might want to check if
you can get your editor to warn about that. Or "git diff" will..)

//Magnus