Re: [PATCH 02/14] Add support for a generic wal reading facility dubbed XLogReader

Heikki Linnakangas wrote:

> I'm disappointed to see that there has been no progress on this
> patch since last commitfest. I thought we agreed on the approach I
> championed for here:
> http://archives.postgresql.org/pgsql-hackers/2012-09/msg00636.php.
> There wasn't much work left to finish that, I believe.
>
> Are you going to continue working on this?

Here's what I have right now.  It's your patch, plus some tweaks such as
changing the timing for allocating readRecordBuf; I also added a struct
to contain XLogReadPage's private data, instead of using global
variables.  (The main conclusion I get from this, is that it's
relatively easy to split out reading of XLog out of xlog.c; there are
some global variables still remaining, but AFAICS that should be
relatively simple to fix).

There is no consensus on the way to handle error reporting.  Tom
suggests having the hypothetical client-side code redefine ereport()
somehow; as far as I can see that means we would have to reimplement
errstart, errfinish, etc.  That doesn't sound all that nice to me.

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