Postgres Pro
Downloads
Home   >   mailing lists

Thread: "make check" and pg_hba.conf

"make check" and pg_hba.conf

From
Jeff Janes
Date:
Is there a way to get "make check" to install a custom pg_hba.conf for its temporary installation? Something like pre-prending the line:

local  all  password_user  md5

To the default pg_hba.conf?


Thanks,

Jeff

Re: "make check" and pg_hba.conf

From
Tom Lane
Date:
Jeff Janes <jeff.janes@gmail.com> writes:
> Is there a way to get "make check" to install a custom pg_hba.conf for its
> temporary installation? Something like pre-prending the line:
> local  all  password_user  md5
> To the default pg_hba.conf?

No.  Why would you want that?  External connections to the test DB seem
like exactly what you *don't* want, for reproducibility's sake.

Also, this seems like it's overlapping with the use case for making
an installation and using make installcheck.
        regards, tom lane

Re: "make check" and pg_hba.conf

From
Jeff Janes
Date:
On Tue, Oct 18, 2016 at 2:28 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Jeff Janes <jeff.janes@gmail.com> writes:
> Is there a way to get "make check" to install a custom pg_hba.conf for its
> temporary installation? Something like pre-prending the line:
> local  all  password_user  md5
> To the default pg_hba.conf?

No.  Why would you want that?  External connections to the test DB seem
like exactly what you *don't* want, for reproducibility's sake.

I would want to do that so that the code dealing with password-based logins doesn't go completely untested by "make check", like it currently is.

I don't see how connecting to an obscure unix socket with a password is more "external" than connecting to the same obscure linux socket without a password.

Cheers,

Jeff

Re: "make check" and pg_hba.conf

From
Tom Lane
Date:
Jeff Janes <jeff.janes@gmail.com> writes:
> On Tue, Oct 18, 2016 at 2:28 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> No.  Why would you want that?  External connections to the test DB seem
>> like exactly what you *don't* want, for reproducibility's sake.

> I would want to do that so that the code dealing with password-based logins
> doesn't go completely untested by "make check", like it currently is.

Well, unless you create a user with a password and use that password to
log in, you're not performing much of a test.  Getting the hba entry
in there seems to me to be about the least of the worries for making
a usable test case.
        regards, tom lane

Re: "make check" and pg_hba.conf

From
Peter Eisentraut
Date:
On 10/18/16 6:31 PM, Jeff Janes wrote:
> I would want to do that so that the code dealing with password-based
> logins doesn't go completely untested by "make check", like it currently is.

Write a TAP test for it.

-- 
Peter Eisentraut              http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services