Re: [HACKERS] timeouts in PostgresNode::psql - Mailing list pgsql-hackers

From Craig Ringer
Subject Re: [HACKERS] timeouts in PostgresNode::psql
Date
Msg-id CAMsr+YGX0NEiviy1FyPUiQd0FiDXBJSg_mwi4Y+oEDAPfXAhvQ@mail.gmail.com
Whole thread Raw
In response to Re: [HACKERS] timeouts in PostgresNode::psql  (Alvaro Herrera <alvherre@2ndquadrant.com>)
List pgsql-hackers
On 28 February 2017 at 20:39, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:

> Lately I've been wondering about backpatching the whole TAP test
> infrastructure, all the way back.  As we notice bugs, it's really useful
> to use newly added tests in all branches; but currently PostgresNode
> doesn't work with old branches, particularly since the '-w' switch was
> removed from pg_ctl invokations in PostgresNode->start and ->restart
> methods -- (the test just fail without any indication of what is going
> on).

Yeah. I actually did prepare a backpatch of PostgresNode to 9.5 (with
room to go back further), and it's pretty simple if you don't care
about breaking all existing TAP tests ;)

The irritating bit is that the various TAP stuff introduced earlier
for src/bin and so on uses its own node management plus some helpers
in TestLib. Those helpers went away when we introduced PostgresNode
and they aren't easily emulated because the tests make assumptions
about the directory structure that don't fit well with PostgresNode's
way of doing things.

It's fixable, it just wasn't worth the hassle for what I needed given
that I saw the backport as having minimal chance of getting into core.

I've pushed what I did to
https://github.com/2ndQuadrant/postgres/tree/dev/backport-96-tap-to-95
in case anyone finds it useful.

-- Craig Ringer                   http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training & Services



pgsql-hackers by date:

Previous
From: Yugo Nagata
Date:
Subject: Re: [HACKERS] [POC] hash partitioning
Next
From: Robert Haas
Date:
Subject: Re: [HACKERS] Write Ahead Logging for Hash Indexes