The current code uses identifiers which are reserved by both C++ standard (since as long as I can remember) and C11.
They are trivial to fix but some may be considered an API change, in case C users used the non-typedef-ed identifiers.
The header guards are probably very safe to fix though, at least I would consider it an obvious misuse to test for them from the outside.
Example of warning:
./json.h:99:16: warning: identifier '_json_object_entry' is reserved because it starts with '_' at global scope [-Wreserved-identifier]
typedef struct _json_object_entry
^
I can send a PR to change this, even in an API compatible way (using macros... like, if legacy define old symbols, otherwise use a standard-abiding symbol scheme) not nice but can be useful in that kind of cases), but I guess you might prefer to choose your own preffix/suffix styles.
The current code uses identifiers which are reserved by both C++ standard (since as long as I can remember) and C11.
They are trivial to fix but some may be considered an API change, in case C users used the non-typedef-ed identifiers.
The header guards are probably very safe to fix though, at least I would consider it an obvious misuse to test for them from the outside.
Example of warning:
I can send a PR to change this, even in an API compatible way (using macros... like, if legacy define old symbols, otherwise use a standard-abiding symbol scheme) not nice but can be useful in that kind of cases), but I guess you might prefer to choose your own preffix/suffix styles.