class RSpec::Support::EncodedString

def matching_encoding(string)

e.g. the UTF-16LE emoji: 128169.chr
RangeError: out of char range
Raised by byte <-> char conversions

e.g. "\x80".force_encoding('ASCII-8BIT').encode('Emacs-Mule')
or a converter path cannot be found
e.g. 'abc'.force_encoding('UTF-8').encode('foo')
when a named encoding does not correspond with a known converter
Encoding::ConverterNotFoundError:
Raised by transcoding methods:

that doesn't respond to #to_str
when calling force_encoding on an object
Encoding.find(:"UTF-8")
when a symbol is passed as an encoding
TypeError
e.g."\x80".split("\n")
when operating on a string with invalid bytes
ArgumentError
# => ''
vs "\x80".encode('UTF-8','US-ASCII', invalid: :replace, replace: '')
e.g. "\x80".encode('UTF-8','US-ASCII')
either the source or target encoding
when the string being transcoded contains a byte invalid for
Encoding::InvalidByteSequenceError:
e.g. valid_unicode_string.encode(utf8_encoding) << ascii_string
e.g. utf_16le_emoji_string.split("\n")
when Encoding.compatible?(str1, str2) is nil
Encoding::CompatibilityError
# => ''
vs "\x80".encode('UTF-8','ASCII-8BIT', undef: :replace, replace: '')
e.g. "\x80".encode('UTF-8','ASCII-8BIT')
if the String contains characters invalid for the target encoding
when a transcoding operation fails
Encoding::UndefinedConversionError:
Raised by Encoding and String methods:

Encoding Exceptions: 
def matching_encoding(string)
  string = remove_invalid_bytes(string)
  string.encode(@encoding)
rescue Encoding::UndefinedConversionError, Encoding::InvalidByteSequenceError
  # Originally defined as a constant to avoid unneeded allocations, this hash must
  # be defined inline (without {}) to avoid warnings on Ruby 2.7
  #
  # In MRI 2.1 'invalid: :replace' changed to also replace an invalid byte sequence
  # see https://github.com/ruby/ruby/blob/v2_1_0/NEWS#L176
  # https://www.ruby-forum.com/topic/6861247
  # https://twitter.com/nalsh/status/553413844685438976
  #
  # For example, given:
  #   "\x80".force_encoding("Emacs-Mule").encode(:invalid => :replace).bytes.to_a
  #
  # On MRI 2.1 or above: 63  # '?'
  # else               : 128 # "\x80"
  #
  string.encode(@encoding, :invalid => :replace, :undef => :replace, :replace => REPLACE)
rescue Encoding::ConverterNotFoundError
  # Originally defined as a constant to avoid unneeded allocations, this hash must
  # be defined inline (without {}) to avoid warnings on Ruby 2.7
  string.dup.force_encoding(@encoding).encode(:invalid => :replace, :replace => REPLACE)
end