Comment 2 for bug 1179660

I've linked this bug to Evergreen as well, since CDBI subclasses defined there have historically been prone to not being fully synchronized with the schema, and removing the autoload will convert those mistakes into (run-time, alas) fatal errors. I do think that they *should* be fatal errors, but we should fix the ones we know about first.

I've also removed the milestone targeting for the moment -- if we can get away with releasing it just for OpenSRF 2.2.x, that would be my preference. Regardless, related Evergreen fixes (if any) will probably have to be backported to all supported Evergreen release branches, since we don't say that older Evergreen version series *must* use older OpenSRF versions.