When parsing numbers I just looked for all valid symbols that could occur within a number, and assumed that json_get_number_size had verified the number was correct. This was wrong, because the way json.h is setup it will parse a number, and once it gets to the first character that it doesn't know about will bail successfully and then set about parsing (this allows us to support comments for instance).
When parsing numbers I just looked for all valid symbols that could occur within a number, and assumed that json_get_number_size had verified the number was correct. This was wrong, because the way json.h is setup it will parse a number, and once it gets to the first character that it doesn't know about will bail successfully and then set about parsing (this allows us to support comments for instance).