I'm retrieving some stringifyed JSON via TCP in node.js an want to parse it. So my approach is similiar to this. I shortend and simplified it, so you don't have to know the surrounding logic.
socket.on("data", function(data) {
console.log(data.toString()); // Shows the original stringifyed version
console.log(JSON.parse(data.toString())); // Doesn't work
});
The complete (beautified) JSON is this. As you can see, there are no errors.
{
"result": "success",
"source": "chat",
"success": {
"message": "test",
"time": 1331770513,
"player": "silvinci"
}
}
But JSON.parse(data.toString())
always throws me this dumb error:
{"result":"success","source":"console","success":{"time":1331762264,"line":"20
^
SyntaxError: Unexpected token {
at Object.parse (native)
at Socket.<anonymous> (/home/node/api.js:152:35) // irrelevant from here on
at Socket.emit (events.js:67:17)
at TCP.onread (net.js:347:14)
So I thougt: "What could be wrong with the JSON-String. Let's try it directly. Should not work." Surprise, Surprise! It worked. Why does it work when I directly enter the String?
Thanks to @Felix Kling I found my bug. It's very important to filter unescaped characters, especially outside the stringified JSON. I didn't and overlooked an invisible linebreak right after the stringified JSON.
This is the fix:
socket.on("data", function(data) {
console.log(data.toString()); // Shows the original stringified version
console.log(JSON.parse(data.toString().slice(0, -4))); // Trim the sequence "\r\n" off the end of the string
});
Please note, that this only works for me, as I have a very specialized case. The server is always responding in lines of JSON, that are terminated with an \r\n
- not the whitespace characters, but literally backslash r and backslash n.
Your code may (or probably) fail due to other errors. But checking the server's response is a good starting point when you get parsing errors.
As @Zack correctly pointed out, this is a more general fix for removing unintended whitespace:
JSON.parse(data.toString().trim());