In our case the JSON::parse heuristics were not correctly recognizing the encoding of the response body. The simplest way to remedy this is to explicitly mark the body as being in UTF-8.
A more robust approach would be to check the response's content-type header and extract the encoding from there. I would give this a try if the copycopter server is able to set the encoding to anything else than UTF-8. Otherwise I would just leave it hardcoded as is.
force encoding of downloaded blubrs to utf-8 in case JSON::parse heur…
…istics fail to recognize it as such