Tom Lane wrote:
Bruno Wolff III <[EMAIL PROTECTED]> writes:
Is a test going to get added to the regression tests to catch similar
regressions in the future?

I've been thinking about that.  It seems that the regression tests have
fairly poor coverage of use of typmod-bearing data types in general;
most of our tests of complicated queries tend to use "simple" datatypes
like int or text.  I don't have any immediate thoughts what to do about
that --- massive expansion of the tests doesn't seem justified --- but
this isn't the first bug we've hit in this area.  It's just a bit more
embarrassing than most :-(

I think at least the most simple cases should be added. At the very least a test that would have caught this issue. This is really the first time that I had to pull a minor release and go back to a previous version. ;-)

As far as I understand, it's as simple as this (untested):

        c DECIMAL(5,2) NOT NULL,
        CHECK (c >= 0)
INSERT INTO tab ('0');


Or at least:
UPDATE tab SET c='0';

Best Regards
Michael Paesold

---------------------------(end of broadcast)---------------------------
TIP 6: explain analyze is your friend

Reply via email to