On 2022-06-14 Tu 12:20, Tom Lane wrote:
> Andrew Dunstan <andrew@dunslane.net> writes:
>> The first makes the argument for $node->config_data() optional. If it's
>> not supplied, pg_config is called without an argument and the whole
>> result is returned. Currently, if you try that you get back a nasty and
>> cryptic error.
> No opinion about whether that's useful.
>
>> The second changes the new GUCs TAP test to check against the installed
>> postgresql.conf.sample rather than the one in the original source
>> location. There are probably arguments both ways, but if we ever decided
>> to postprocess the file before installation, this would do the right thing.
> Seems like a good idea, especially since it also makes the test code
> shorter and more robust(-looking).
>
> Looking at the patch itself,
>
> +my $share_dir = $node->config_data('--sharedir');
> +chomp $share_dir;
> +$share_dir =~ s/^SHAREDIR = //;
> +my $sample_file = "$share_dir/postgresql.conf.sample";
>
> I kind of wonder why config_data() isn't doing the chomp itself;
> what caller would not want that? Pulling off the variable name
> might be helpful too, since it's hard to conceive of a use-case
> where you don't also need that.
It already chomps the output, and pg_config doesn't output "SETTING = "
if given an option argument, so we could just remove those two lines -
they are remnants of an earlier version. I'll do it that way.
cheers
andrew
--
Andrew Dunstan
EDB: https://www.enterprisedb.com