I just wondered - what is the policy on unit test failures in the
ActiveRecord trunk? Actually I’ve got two questions…
is it normal for the trunk to fail unit tests, or is it always an
oversight when this happens?
if you notice test failures, what should you do to inform people?
Specifically, I’m testing a version of a patch that someone else has
modified. I just checked out the trunk and ran the unit tests on SQL
Server to get a baseline before I applied the patch, but got 6
failures and 4 errors. I’m guessing that relatively few Rails
developers use SQL Server so the tests are not run against it as
often (we still use it at work for some stuff so I have access to a
copy). How do you get round the problem of different developers
using different systems?
Or is the Rails team already hiding a big datacentre somewhere,
running SQL Server, DB2 and Oracle enterprise editions checking that
nothing ever breaks…