Re: pgsql-server/src/test/regress expected/random. ...

Tom Lane wrote:
> momjian@svr1.postgresql.org (Bruce Momjian) writes:
> >     Improve random regression tests to fail less frequently.
>
> If you're gonna do that you should fix the discussion in the admin
> guide.  It presently says that failures in random should be expected
> every five to ten trials, and even hints that you should be concerned
> if you never see a failure.

Oh, I didn't know we had a mention of random there too.  New text is:

    <title>The <quote>random</quote> test</title>

    <para>
     There is at least one case in the <literal>random</literal> test
     script that is intended to produce random results. This causes
     random to fail the regression test occasionally.  Typing
<programlisting>
diff results/random.out expected/random.out
</programlisting>
     should produce only one or a few lines of differences.  You need
     not worry unless the random test always fails in repeated
     attempts.
    </para>

I wanted to reduce the frequency of random failure by doing more random
tests, and reduce the amount of confusion when it fails.  I see no way
of preventing it from ever failing.

--
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073