Question

I make a call to google’s dictionary api like this:

var json = new WebClient().DownloadString(string.Format(@"http://www.google.com/dictionary/json?callback=dict_api.callbacks.id100&q={0}&sl=en&tl=en", "bar"));

However I get a response that this code fails to parse correctly:

json = json.Replace("dict_api.callbacks.id100(", "").Replace(",200,null)", "");
JObject o = JObject.Parse(json);

The parse dies at encountering this:

"entries":[{"type":"example","terms":[{"type":"text","text":"\x3cem\x3ebars\x3c/em\x3e of sunlight shafting through the broken windows","language":"en"}]}]}

The

\x3cem\x3ebars\x

stuff kills the parse

Is there some way to handle this JSONP response with JSON.NET?

The answer by aquinas to another “Parse JSONP” question shows nice regex x = Regex.Replace(x, @"^.+?\(|\)$", ""); to handle with JSONP part (may need to tweak regex for this case), so main part here is how to deal with hex-encoded characters.

Answer 1

Reference: How to decode HTML encoded character embedded in a json string

JSON specs for strings do not allow hexadecimal ASCII escape-sequences, but only Unicode escape-sequences, which is why the escape sequence is unrecognized and which is why using \u0027 instead should work … now you could blindly replace \x with \u00 (this should perfectly work on valid JSON, although some comments may get damaged in theory, but who cares … :D)

So change your code to this will fix it:

        var json = new WebClient().DownloadString(string.Format(@"http://www.google.com/dictionary/json?callback=dict_api.callbacks.id100&q={0}&sl=en&tl=en", "bar"));

        json = json
                .Replace("dict_api.callbacks.id100(", "")
                .Replace(",200,null)", "")
                .Replace("\\x","\\u00");

        JObject o = JObject.Parse(json);