String.encode(Encoding, :xml => :text) double encodes ampersand

[pepijnve]

Tested with 1.0-RC11.

This test passes on MRI. On Truffle it produces ±.

describe "String.encode" do
  it 'should behave like MRI' do
    expect("\u00B1".encode(Encoding::US_ASCII, :xml => :text)).to(eq("±"))
  end
end

[nirvdrum]

Thanks for the test. This looks like something that should be added to the Ruby Spec Suite. Unfortunately, String#encode with the xml option isn't completely implemented and we fail several other related specs, as seen in our tags file: https://github.com/oracle/truffleruby/blob/5975d2aa48ea24e560baa81abe62cffc89cb9b3c/spec/tags/core/string/encode_tags.txt

If you don't mind me asking, what's the use case? We need to implement it, regardless. I'm just curious because I don't think I've ever seen it used.

[pepijnve]

@nirvdrum see asciidoctor/asciimath#10
I'm using it as a really simple way to prepare a string for embedding in other XML content.

The input string is MathML containing various mathematical symbols which are represented as their Unicode codepoint. In order to make the final MathML string as safe and simple to use as possible I opted to XML escape anything that's outside 7-bit ASCII. String.encode was the most straightforward way to achieve that.

[nirvdrum]

@pepijnve I've fixed this issue in 5667e6c and it will ship in the next TruffleRuby release (1.0.0 RC 12). Thanks again for the detailed bug report!

[nirvdrum]

@pepijnve TruffleRuby 1.0.0 RC 12 is now available with this bug fixed. Please try out the new release at your convenience and let me know if you run into any problems. Thanks again for filing an issue!