A standalone Extensible Data Notation (EDN) parser in plain C
A data notation that is so flexible it almost seems like it came from Hell, deserves a parser that is made like somebody goes down to Hell, guns blazing, and taking no heap-allocation prisoners. But this works as well as I could make it for the data that I have tested and the interpretation that I made of the documents I could find. It is incomplete but in working order for what I am doing with it; I will finish it to potentially full compatibility in the future -- or you can do it.
I built it because I needed it, not because this is how I like to spend my time, so if it is faulty, fix it. The philosophy is simple:
-
"state machine" (that is what parsing should be as per Niklaus Wirth's book)
-
keep reading data as if it is a long, read-only string
-
detect what each token encountered is (do not tokenize too soon)
-
allocate a generic struct (I use a "union") for every type of token
-
each token holds its own data; pointers to memory blocks (like in strings)
-
a "top level" object exists (no phaquing way around it)
To understand how to use it, locate the function that "unrolls" the "toplevel" object, and it will all make sense. If you want to know what I did, just turn on the very verbose debugging preprocessor directives in the Makefile. I have injected a string of data in the driver which I used for developing this.
I wish you luck; you will need it.
IN 2024/01/21