Stephen Gelman

For sure. There is a bug that exists at least with ruby 2.2.1 running with RVM on Debian Jessie that causes the following error when running `:CommandT`: ``` Error detected...

That's unfortunate. I'll see what I can do to fix it.

@wincent I pushed a new commit that fixes this as far as I can tell. I didn't squash the commits but I am happy to if you'd prefer that.

Actually, this now doesn't fix the problem for some reason. ☹️ Back to the drawing board!

@wincent OK, _now_ I believe it's fixed. Turns out Gem::Ext::BuildError isn't defined until `rubygems/ext` is autoloaded so I needed to `require` it before I could check if it's defined.

For what it's worth, I believe that e7100031be7e21a89044b88062b5a1979a5e557e is what broke it. Ruby stops evaluating the constants once it matches one so in the original code when `LoadError` is hit...

Yeah, that would work. Do you think it is at least cleaner to do something like the following? ```ruby begin # ... rescue Exception => e EXCEPTIONS = [LoadError] EXCEPTIONS

OK, change is made. Did you want me to squash these commits?

Does this look ok to merge now?

FYI I mitigated this somewhat in #172 by adding the ability to bind as the user logging in instead of a service account. I know it doesn't solve your initial...