Safe encode input strings, when they can be flagged with the wrong encoding.

string.rb

class String

  def force_valid_encoding
    find_encoding(Encoding.list.to_enum)
  end

  def safe_encode(name)
    force_valid_encoding
    encode(name, :undef => :replace, :invalid => :replace, :replace => "")
  end

  private

  def find_encoding(encodings)
    if valid_encoding?
      self
    else
      force_next_encoding(encodings)
    end
  end

  def force_next_encoding(encodings)
    force_encoding(encodings.next)
    find_encoding(encodings)
  end
end

string_spec.rb

require "spec_helper"

describe String do

  describe "#force_valid_encoding" do

    context "when it is flagged with a valid encoding" do

      let(:string) do
        "\xff\xdb\xff\xdf".force_encoding("UTF-16")
      end

      let(:forced) do
        string.force_valid_encoding
      end

      it "retains the valid encoding" do
        forced.encoding.name.should == "UTF-16"
      end
    end

    context "when it is flagged with another valid encoding" do

      let(:string) do
        "\xff\xdb\xff\xdf".force_encoding("ASCII-8BIT")
      end

      let(:forced) do
        string.force_valid_encoding
      end

      it "retains the other encoding if valid" do
        forced.encoding.name.should == "ASCII-8BIT"
      end
    end

    context "when it is flagged with an invalid encoding" do

      let(:string) do
        "\xff\xdb\xff\xdf".force_encoding("UTF-8")
      end

      let(:forced) do
        string.force_valid_encoding
      end

      it "forces the first the valid encoding" do
        forced.encoding.name.should == "ASCII-8BIT"
      end
    end
  end

  describe "#safe_encode" do

    context "when coding with invalid characters" do

      let(:string) do
        "\xff\xdb\xff\xdf".force_encoding("ASCII-8BIT")
      end

      let(:encoded) do
        string.safe_encode("UTF-8")
      end

      it "replaces invalid sequences and characters with an empty string" do
        encoded.should be_empty
      end
    end

    context "when encoding with valid characters" do

      let(:string) do
        "\x00\x00".force_encoding("UTF-16LE")
      end

      let(:encoded) do
        string.safe_encode("UTF-8")
      end

      it "keeps the string intact" do
        encoded.should == "\u0000"
      end
    end
  end
end