On Fri, Jan 26, 2018 at 11:56 AM, Stephen Frost <sfrost@snowman.net> wrote:
>> I think you've chosen a terrible design and ought to throw the whole
>> thing away and start over.
>
> I'll all for throwing away the existing test once we've got something
> that covers at least what it does (ideally more, of course).
I'm for throwing this away now. It's a nuisance for other people to
maintain, and as Tom's reply makes clear (and it matches my
suspicions), they are maintaining it without really knowing whether
the updates are making are *correct*, just knowing that they *make the
tests pass*. It's nice to make things turn green on the code coverage
report, but if we're not really verifying that the results are
correct, we're just kidding ourselves. We'd get the same amount of
green on the code coverage report by running the pg_dump commands and
sending the output to /dev/null, and it would be a lot less work to
keep up to date.
I'm glad this helped you find some bugs. It is only worth keeping if
it prevents other hackers from introducing bugs in the future. I
doubt that it will have that effect.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
