Re: [PATCH] PostgresNode.pm enhancements, pg_lsn helper, and some more recovery tests

On Thu, Jan 5, 2017 at 1:58 AM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
> On 4 January 2017 at 08:16, Craig Ringer <craig(at)2ndquadrant(dot)com> wrote:
>> When committed I will update the decoding on standby series to omit
>> these pre-requisite patches.
>
> Committed, thanks.

I was planning a round of reviews of those patches, but you were faster than me.

So now looking at what has been committed.

The perldoc documentation is broken for the new routines. The commits
have added patterns like that:
=pod $node->routine_name
But what needs to be done is to use =pod and =item, like that:
=pod
=item $node->routine_name

+Look up xlog positions on the server:
This has better be "*WAL* positions". There is another reference with
xlog (see recent threads about renaming those functions for example).

+* insert position (master only, error on replica)
+* write position (master only, error on replica)
+* flush position
+* receive position (always undef on master)
+* replay position
Replay position is always undefined on primary, let's document it in
the description of the routine. And flush position generates an error
for a standby.

The documentation of $node->lsn is generated like that, with a very
unfriendly list of modes:
$node->lsn(mode)
Look up xlog positions on the server:

* insert position (master only, error on replica) * write position
(master only, error on replica) * flush position * receive position
(always undef on master) * replay position
A trick that I have found here is to add a space before the '*'.

It may be a good idea to run perltidy on top of that to be honest...

Attached is a patch to fix all those small issues.
--
Michael