How to parse using boost if it is not json, but similar to it?

Question

I have an application that reads some parameters from a json file, but now the file has changed (it is not my decision). The file looks pretty the same, but instead of {} it has [], so if until now it looked like

{
  "A":
  {
    "a":"a",
    "b":"b"
  }
}

now it is like this:

[
  "A":
  [
    "a":"a",
    "b":"b"
  ]
]

My application is written in C++, so I have used boost to parse the json file, but now I am looking for a new parser, is there a way to do it with boost?

I have started to search, but I have also asked the question thinking that maybe someone may help me faster than me finding the answer. Thanks

You could be lazy and simply replace characters from your input, and then process it exactly as you do now. Then check if there is a noticeable decrease in performance. — Benjamin Lindley, Feb 19 '15 at 08:45
"My application is written in C++, so I have used boost to parse the json file" – no, you definitely don't *have to* use Boost. — The Paramagnetic Croissant, Feb 19 '15 at 08:56
@TheParamagneticCroissant: true, I do not have to use boost, but I've used it :) — sop, Feb 19 '15 at 09:00

score 1 · Accepted Answer · edited May 23 '17 at 12:28

It looks like Yet Another YAML/Mustache/JSON/... derivative.

Without a formal spec it's hard to actually assess what effort would be required, but here's a list of implementations of similar grammars in Boost Spirit, with varying amounts of feature completeness:

How to parse mustache with Boost.Xpressive correctly? <-- this is likely your best matching demonstration
Parse a substring as JSON using QJsonDocument (minimal subset, use something like this to transform the input to proper JSON, e.g.?)
Reading JSON file with C++ and BOOST A full featured JSON parser (with AST and escapes but no comments)

Applications of a toy JSON parser implementation:

How to parse using boost if it is not json, but similar to it?

1 Answers1

Linked