json.Unescape has issues with UNC paths

[richard-mauri]

I have a test string that has a value with an escaped UNC path in it . It validates ok according to: https://jsonlint.com/

"{\"dataPath\":\"\\\\bnk11977fs\\bnk11977\\xyz11146682\\xyzdata\\\"}"

When you pass it through json.Unescape the result is invalid json

See: playground

Highlights from the stdout are:

raw sample0 "\"{\\\"dataPath\\\":\\\"\\\\\\\\bnk11977fs\\\\bnk11977\\\\be11146682\\\\encompassdata\\\\\\\"}\""

unescape for sample0 was invalidated "{\"dataPath\":\"\\\\bnk11977fs\\bnk11977\\be11146682\\encompassdata\\\"}" ; error=json: invalid character 'e' in string escape code: "\\bnk11977fs\bnk11977\be1114668...

You may ask, "If the original is valid json, then why do you need to unescape it anyway?"

Answer: The raw escaped form comes from reading a logging entry. A log processor service (lambda) needs to create output without all the escapes, else another downstream reader (logstash ingest) see's the dcument event as a big ass string and not a nice nested object.

So the goal is to be able to have the lambda processor write the line something like the following. yes the UNC and escapes are a bitch

And this is a simple example. Imagine a raw json with 100 deply nested fields. You certinly can't just blind Replacement of backslash-quote with quote !

{"dataPath":"\\bnk11977fs\bnk11977\be11146682\encompassdata\"}