Correctly set BerIdentifiedString values to UTF-8

[cmdrclueless]

The existing code did not set the actual internal value to UTF-8 when the encoding is correct. This patch attempts to resolve that problem by passing a UTF-8 encoded string to the super class when possible.

[jch]

The change makes sense. Could you add a test for this?

I think it might be more readable to split this into two statements. It threw me off a bit to see super begin.

encoded_args = begin
  args.respond_to?(:encode) ? args.encode('UTF-8') : args
rescue
  args
end

super(encoded_args)

Thoughts?

[cmdrclueless]

@jch I would think the existing test handle the expectations on strings returned from the LDAP server. The problem is the original line

self.encode('UTF-8') if self.respond_to?(:encoding) rescue self

didn't really do anything. I think the intent of the code was to modify the string internally, so the correct call would have been self.encode!. Is there a particular test that you're looking for?

As to the the change you propose, it doesn't really matter to me. If you prefer what you've proposed I can change it. Either one is valid.

[jch]

I'm not sure whether the intention is to modify the string in place. It's weird to me that this inherits from String rather than composing it, but looking at String::new, the doc does say it strings a new String object. I agree with you that the current code doesn't make sense though because that value isn't used or saved anywhere. Where are you running into an error with this?

[cmdrclueless]

I ran into this error when dealing with Chinese names in Active Directory.

[cmdrclueless]

@jch, is there something more you're looking for before you a decision is made on this request?

[jch]

@cmdrclueless I'm sorry for the slow response. Work has been busy, and the other maintainer is out of town. I think it's a good idea, I just want to test it more thoroughly myself before merging.

[cmdrclueless]

@jch, no worries. I just wanted to make sure I had not dropped the ball on my side.

[jch]

Did an audit through the codebase and the only callsite that initializes BerIdentifiedString is

ruby-net-ldap/lib/net/ber/ber_parser.rb

Line 40 in 4abe573

s = Net::BER::BerIdentifiedString.new(data || "")

. Seems like a reasonable fix. Thanks for the pull!