Re: Is it time to kill support for very old servers?

On September 18, 2017 4:15:31 AM PDT, Michael Paquier <michael(dot)paquier(at)gmail(dot)com> wrote:
>On Mon, Sep 18, 2017 at 8:09 PM, Andres Freund <andres(at)anarazel(dot)de>
>wrote:
>>
>>
>> On September 18, 2017 4:08:21 AM PDT, Michael Paquier
><michael(dot)paquier(at)gmail(dot)com> wrote:
>>>On Mon, Sep 18, 2017 at 7:54 PM, Andres Freund <andres(at)anarazel(dot)de>
>>>wrote:
>>>>>It seems to me that you are looking more for a connection parameter
>>>>>here.
>>>>
>>>> I'm not seeing a meaningful distinction here? Env vars and
>connection
>>>parameters are handled using the same framework in libpq. And using
>>>the env var in the test would be better, because you'd only set one
>>>value - hard to do within our non TAP tests (i.e. in an existing
>psql,
>>>started by pg regress) otherwise.
>>>
>>>Or both? I don't really understand why an environment variable is
>>>better than a connection string. For the TAP tests, you could just
>set
>>>the base of the connection string once and you are done as well. See
>>>the SSL tests for example.
>>
>> Did you read what I wrote?
>
>Sorry, I missed the "non" with "TAP" tests. Having a connection
>parameter would still be low-cost in maintenance, so if you add that
>at the same time I won't complain.

Private:

And now you missed the "same infrastructure" part.
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.