Encode true as xFF

[mtodd]

According to the spec (http://tools.ietf.org/html/rfc4511#section-5.1), we're currently encoding true incorrectly as \x01\x01\x01 instead of \x01\x01\xFF.

Fixes #31.

Wouldn't mind finding a good integration test for this, but nothing obvious or direct comes to mind.

cc @jch @schaary

[mtodd]

Ruby 1.8.7 has been retired for well over a year. Moving ahead, we should assume 1.9+ only.

Once this is merged and released, it will mark the first release that does not support ~1.8.7. We should make that explicit by setting appropriate Gemspec flags.

[jch]

ruby 1.9.3+ requirement enforced in #145

[jch]

I think this is sufficiently covered by the search integration test since the default size limit is '0'.to_ber. If your encoding was wrong, then search would break. Yes, the test is implicit, but it's such a common value it'd be pretty obvious if it broke.

[mtodd]

@jch size isn't a boolean, though, since this PR only applies to the BER encoding of the true value. This would apply for the "criticality" flag of a search control, and few other places (not much boolean usage in the spec from what I've seen).

[mtodd]

Luckily for us, I think all we have to do is use a non-existing sort control and set the criticality to true, then assert that the response is "unavailable critical extension" (result code 12).

[jch]

Oops, my brain just autocompleted boolean to 0 and 1, but ya what you said makes sense.

[mtodd]

@jch ready for your review.

[jch]

🚢