+1, I think this is a good idea in principal, and I'm game for giving it a
go.
The problem in practice is (I suspect that) the build is going to fail a
lot, making more work for committers as we go (rather than building up
problems for later).
The only thing that annoys me about it is that if a developer does their job
correctly (create a failing test, make it pass, run the whole build, then
check-in), they're still going to see failures on other dialects sometimes.
Which can be disheartening - but perhaps it's still the lesser of two evils.
-----Original Message-----
From: Patrick Earl
Sent: Thursday, July 21, 2011 11:47 PM
To: nhibernate-development
Subject: [nhibernate-development] Alternate Database Builds
What do you guys think about making the alternate database builds fail
if any newly added tests fail? For myself, I would prefer it if
people were not adding tests that don't work properly across multiple
databases. While we don't have time to fix every old test for every
database right now, it makes sense to me that we shouldn't be adding
new code that is already broken.
Patrick Earl
