Re: COPY command character set

As Tom says, this doesn't really address my original issue, which was
not that I read the material on encoding and misunderstood it, but that
I didn't even see that material because it was mixed in with a bunch of
other notes on all sorts of random subjects.

To address this issue in the documentation at large, I would like to see
every command that has I18N/L10N-related behavior have a separate
sub-head for the explanation of that behavior. That way, anyone who
needs to know about that aspect (which should be everyone), just has to
look for the sub-head to be sure they have found what they need to know.
Whilst I know we can't do that for every single cross-command topic, it
seems to me that I18N/L10N is sufficiently important to users of pg that
it merits this treatment.

FWIW, I think error handling/behavior also merits its own sub-heads
throughout. And there should be links within the pages to sub-heads (cf.
DB2's online doc).

Of course, all of this would be a substantial project.

Note that for the COPY command the I18N/L10N material covers both
DateStyle and encoding.

In respect of Bruce's proposed changes, I prefer the original wording
(for the same reasons as Tom), but with the addition of the mention of
the server - "... read from or written to a file directly by the
server".

--
Peter Headland
Architect
Actuate Corporation

-----Original Message-----
From: Bruce Momjian [mailto:bruce(at)momjian(dot)us]
Sent: Monday, February 22, 2010 22:01
To: Tom Lane
Cc: Peter Headland; Adrian Klaver; pgsql-general(at)postgresql(dot)org
Subject: Re: [GENERAL] COPY command character set

Tom Lane wrote:
> Bruce Momjian <bruce(at)momjian(dot)us> writes:
> > I have updated the documentation to be more direct about COPY
encoding
> > behavior. Patch attached and applied.
>
> Uh, why exactly do you find that better? "Processes data" seems a lot
> vaguer to me than the previous wording. I certainly don't think that
> this does much to address Peter's original complaint.

I thought the problem was that we said "input", then "output" and then
got to the point about the server, and I thought the reader just stopped
reading that far, so I tried to shorten it so the idea was sooner, and I
mentioned "server" at the end. It might not be better, but I tried.

We don't want to highlight the input/output, we want to highlight that
all input and output are controlled by the client encoding.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com
PG East: http://www.enterprisedb.com/community/nav-pg-east-2010.do
+ If your life is a hard drive, Christ can be your backup. +