Support to Ruby 1.8.7

[rogerleite]

Converted ntlm.rb to unix format.

Removed Gemfile.lock from git, explanation at http://yehudakatz.com/2010/12/16/clarifying-the-roles-of-the-gemspec-and-gemfile.

Based on 3f54039 commit, extracted EncodeUtil to define method, based on Ruby version.

I'm not sure if what i did on line rogerleite/rubyntlm@WinRb:master...support_to_ruby18#L4R41 is correct. Suggestions are welcome.

[rogerleite]

Forgot to say, httpi need this. 😄

[pmorton]

@rogerleite Great. I will compile and test. 1.8.7 😰

Should be able to merge and release today.

[pmorton]

@rogerleite I am getting the following failures. Any ideas?

https://gist.github.com/anonymous/5834186

OSX, ruby-1.8.7-p371

[rogerleite]

@pmorton this failures happened with me too. i don't know the "real" reason, after running rbenv rehash, the specs passed.

I don't know if is related, i think if rspec bin points to another ruby, may cause this failures. A which rspec can help us with this puzzle.

This weekend i'll be offline.
If still occurs failures, monday we see what to do.

[pmorton]

@rogerleite can you pull 4b217f9 into this request. Lets see what travis has to say about it.

I tried recompiling with rvm and and isolating the gemset and still no luck. For the moment if travis will pass, I will merge and release. which rspec shows that it is selecting /Users/pmorton/.rvm/gems/ruby-1.8.7-p371/bin/rspec...

[rogerleite]

https://travis-ci.org/WinRb/rubyntlm/builds/8399654

don't know what is ... :/

[pmorton]

I found the issue. There is some net/http magic that include ntlm. The test require 'net/ntlm' which leads to 1.8.7 aliasing the wrong method.

[rogerleite]

Great news! Thanks for the new release. 🤘