On Tue, Jan 10, 2017 at 07:22:17AM +0900, Michael Paquier wrote:
> On Tue, Jan 10, 2017 at 4:32 AM, Bruce Momjian <bruce@momjian.us> wrote:
> > On Sun, Jan 8, 2017 at 03:27:59PM +0900, Michael Paquier wrote:
> >> >> + subdirectory of both the current directory and the
> >> >> + <envar>PGDATA</envar> directory.
> >> >> This could say "of both the current directory and *then* the PGDATA
> >> >> directory" to outline the order of the actions taken by the check
> >> >> routine. Just my 2c on the matter.
> >> >
> >> > Ah, I see your point. I ended up rewording the text to be more explicit
> >> > about the supplied argument and "searching". Updated patch attached.
> >>
> >> - Directory in which to find log segment files. The default is to search
> >> - for them in the <literal>pg_wal</literal> subdirectory of the current
> >> - directory.
> >> + Specifies a directory in which to find log segment files.
> >> + In addition, searches are performed in the current directory,
> >> + and the <literal>pg_wal</literal> subdirectory of both the current
> >> + directory and the <envar>PGDATA</envar> directory.
> >> Hm. This still misses the point that the lookup at PGDATA is done only
> >> if no directory is defined. What do you think about the patch attached
> >> that lists the five possible patterns: 2 for the case where a
> >> directory is defined, and 3 for the default case.
> >
> > Understood. Here is an updated patch which incorporates your
> > suggestions.
>
> Thanks. This looks good to me.
Slightly modified version applied and backpatched to 9.6 (use pg_xlog).
Thanks for your help.
--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient Roman grave inscription +
