Re: Removing useless \. at the end of copy in pgbench

Hello Peter,

>> """
>> Pgbench is expected to work with all PostgreSQL supported versions at
>> the time it is released. Some options may work only with newer servers. It
>> may work with older version down to 8.2.
>> """
>
> It is generally expected (nowadays) that client programs work
> independent of the server version, unless restrictions are specifically
> documented (e.g., pg_dump) or there is some obvious feature dependency
> (e.g., unlogged tables).

Yep, that is somehow what I said in the mail.

> So the above paragraph does not add any useful information.

It tells that it will not work from 8.1 and below.

> Also, I don't find any reason why 8.2 is the cutoff, and saying that it
> may work down to 8.2 (implying that it may not) is content-free.

The "may" is because I could *not* test: I could not run a 8.2 on my
laptop, "initdb" fails on:

creating template1 database in <...>/src/test/regress/./tmp_check/data/base/1 ... ok
initializing pg_authid ... FATAL: wrong number of index expressions
STATEMENT: CREATE TRIGGER pg_sync_pg_database AFTER INSERT OR UPDATE OR DELETE ON pg_database FOR EACH STATEMENT EXECUTE PROCEDURE flatfile_update_trigger();

So I did not feel confident in saying that it would work for 8.2. That is
just me being precise:-)

As explained in the previous mail, pgbench "CREATE TABLE" always uses
"FILLFACTOR" which was introduced in 8.2, so it is sure to fail before
that.

What about:

"""
Pgbench requires a PostgreSQL version 8.2 or above server.
"""

Some information is provided...

I understood that Tom found that an explicit compatibility note would be
welcome, so I provided one. I'm also fine with saying nothing.

--
Fabien.