I worked out what was going on with the MSSQL backend segfaulting on my
Vista machine when running the tests. It wasn't as fatal as I thought
initially, but it's a pretty serious problem if you want to use it in production
It turns out if you put in an incorrect (non-existing) host, instead of getting an
error either when you connect or when you open the connection, everything
seems to work, but it it then all blows up dramatically as soon as you try
execute some SQL.
The same happens if you supply a username that doesn't exist, or even get
your password wrong (ouch!).
I guess I'll be poking around the code a bit to try and see what I can come up
with for that one :)
I also ran the core tests, and got 2 failures on multi-statements, I assume that's
the known thing that's in the OpenDBX bug tracker?
This is all on Windows Vista, Pharo RC1 (10498), OpenDBX 1.4.4,
FreeTDS 0.82, SQL Server Express 2008, both OpenDBX and FreeTDS built
from source with ye olde mingw recommended on the OpenDBX site.