On Windows CMake is usually used to create a Visual Studio solution file by running it inside the Developer Command Prompt for Visual Studio, for exact steps follow the official documentation from CMake and Microsoft and use the online search engine of your choice. The descriptions of the the options above still generally apply, although not all of them work on Windows.
#### Makefile
If you don't have CMake available, but still have GNU make. You can use the makefile to build cJSON:
@ -123,269 +144,340 @@ If you installed it via CMake or the Makefile, you can include cJSON like this:
#include<cjson/cJSON.h>
```
### Some JSON:
### Data Structure
```json
cJSON represents JSON data using the `cJSON` struct data type:
```c
/* The cJSON structure: */
typedef struct cJSON
{
"name": "Jack (\"Bee\") Nimble",
"format": {
"type": "rect",
"width": 1920,
"height": 1080,
"interlace": false,
"frame rate": 24
}
}
struct cJSON *next;
struct cJSON *prev;
struct cJSON *child;
int type;
char *valuestring;
/* writing to valueint is DEPRECATED, use cJSON_SetNumberValue instead */
int valueint;
double valuedouble;
char *string;
} cJSON;
```
Assume that you got this from a file, a webserver, or magic JSON elves, whatever,
you have a `char *` to it. Everything is a `cJSON` struct.
Get it parsed:
An item of this type represents a JSON value. The type is stored in `type` as a bit-flag (**this means that you cannot find out the type by just comparing the value of `type`**).
```c
cJSON * root = cJSON_Parse(my_json_string);
```
To check the type of an item, use the corresponding `cJSON_Is...` function. It does a `NULL` check followed by a type check and returns a boolean value if the item is of this type.
This is an object. We're in C. We don't have objects. But we do have structs.
What's the framerate?
The type can be one of the following:
* `cJSON_Invalid` (check with `cJSON_IsInvalid`): Represents an invalid item that doesn't contain any value. You automatically have this type if you set the item to all zero bytes.
* `cJSON_False` (check with `cJSON_IsFalse`): Represents a `false` boolean value. You can also check for boolean values in general with `cJSON_IsBool`.
* `cJSON_True` (check with `cJSON_IsTrue`): Represents a `true` boolean value. You can also check for boolean values in general with `cJSON_IsBool`.
* `cJSON_NULL` (check with `cJSON_IsNull`): Represents a `null` value.
* `cJSON_Number` (check with `cJSON_IsNumber`): Represents a number value. The value is stored as a double in `valuedouble` and also in `valueint`. If the number is outside of the range of an integer, `INT_MAX` or `INT_MIN` are used for `valueint`.
* `cJSON_String` (check with `cJSON_IsString`): Represents a string value. It is stored in the form of a zero terminated string in `valuestring`.
* `cJSON_Array` (check with `cJSON_IsArray`): Represent an array value. This is implemented by pointing `child` to a linked list of `cJSON` items that represent the values in the array. The elements are linked together using `next` and `prev`, where the first element has `prev == NULL` and the last element `next == NULL`.
* `cJSON_Object` (check with `cJSON_IsObject`): Represents an object value. Objects are stored same way as an array, the only difference is that the items in the object store their keys in `string`.
* `cJSON_Raw` (check with `cJSON_IsRaw`): Represents any kind of JSON that is stored as a zero terminated array of characters in `valuestring`. This can be used, for example, to avoid printing the same static JSON over and over again to save performance. cJSON will never create this type when parsing. Also note that cJSON doesn't check if it is valid JSON.
* `cJSON_IsReference`: Specifies that the item that `child` points to and/or `valuestring` is not owned by this item, it is only a reference. So `cJSON_Delete` and other functions will only deallocate this item, not it's children/valuestring.
* `cJSON_StringIsConst`: This means that `string` points to a constant string. This means that `cJSON_Delete` and other functions will not try to deallocate `string`.
For every value type there is a `cJSON_Create...` function that can be used to create an item of that type.
All of these will allocate a `cJSON` struct that can later be deleted with `cJSON_Delete`.
Note that you have to delete them at some point, otherwise you will get a memory leak.
**Important**: If you have added an item to an array or an object already, you **mustn't** delete it with `cJSON_Delete`. Adding it to an array or object transfers its ownership so that when that array or object is deleted, it gets deleted as well.
Back to disk?
#### Basic types
* **null** is created with `cJSON_CreateNull`
* **booleans** are created with `cJSON_CreateTrue`, `cJSON_CreateFalse` or `cJSON_CreateBool`
* **numbers** are created with `cJSON_CreateNumber`. This will set both `valuedouble` and `valueint`. If the number is outside of the range of an integer, `INT_MAX` or `INT_MIN` are used for `valueint`
* **strings** are created with `cJSON_CreateString` (copies the string) or with `cJSON_CreateStringReference` (directly points to the string. This means that `valuestring` won't be deleted by `cJSON_Delete` and you are responsible for it's lifetime, useful for constants)
```c
char *rendered = cJSON_Print(root);
```
#### Arrays
Finished? Delete the root (this takes care of everything else).
You can create an empty array with `cJSON_CreateArray`. `cJSON_CreateArrayReference` can be used to create an array that doesn't "own" its content, so its content doesn't get deleted by `cJSON_Delete`.
```c
cJSON_Delete(root);
```
To add items to an array, use `cJSON_AddItemToArray` to append items to the end.
Using `cJSON_AddItemReferenceToArray` an element can be added as a reference to another item, array or string. This means that `cJSON_Delete` will not delete that items `child` or `valuestring` properties, so no double frees are occuring if they are already used elsewhere.
To insert items in the middle, use `cJSON_InsertItemInArray`. It will insert an item at the given 0 based index and shift all the existing items to the right.
That's AUTO mode. If you're going to use Auto mode, you really ought to check pointers
before you dereference them. If you want to see how you'd build this struct in code?
If you want to take an item out of an array at a given index and continue using it, use `cJSON_DetachItemFromArray`, it will return the detached item, so be sure to assign it to a pointer, otherwise you will have a memory leak.
Deleting items is done with `cJSON_DeleteItemFromArray`. It works like `cJSON_DetachItemFromArray`, but deletes the detached item via `cJSON_Delete`.
Hopefully we can agree that's not a lot of code? There's no overhead, no unnecessary setup.
Look at `test.c` for a bunch of nice examples, mostly all ripped off the [json.org](http://json.org) site, and
a few from elsewhere.
You can also replace an item in an array in place. Either with `cJSON_ReplaceItemInArray` using an index or with `cJSON_ReplaceItemViaPointer` given a pointer to an element. `cJSON_ReplaceItemViaPointer` will return `0` if it fails. What this does internally is to detach the old item, delete it and insert the new item in its place.
What about manual mode? First up you need some detail.
Let's cover how the `cJSON` objects represent the JSON data.
cJSON doesn't distinguish arrays from objects in handling; just type.
Each `cJSON` has, potentially, a child, siblings, value, a name.
To get the size of an array, use `cJSON_GetArraySize`. Use `cJSON_GetArrayItem` to get an element at a given index.
* The `root` object has: *Object* Type and a Child
* The Child has name "name", with value "Jack ("Bee") Nimble", and a sibling:
* Sibling has type *Object*, name "format", and a child.
* That child has type *String*, name "type", value "rect", and a sibling:
* Sibling has type *Number*, name "width", value 1920, and a sibling:
* Sibling has type *Number*, name "height", value 1080, and a sibling:
* Sibling has type *False*, name "interlace", and a sibling:
* Sibling has type *Number*, name "frame rate", value 24
Because an array is stored as a linked list, iterating it via index is inefficient (`O(n²)`), so you can iterate over an array using the `cJSON_ArrayForEach` macro in `O(n)` time complexity.
### Here's the structure:
#### Objects
```c
typedef struct cJSON {
struct cJSON *next,*prev;
struct cJSON *child;
You can create an empty object with `cJSON_CreateObject`. `cJSON_CreateObjectReference` can be used to create an object that doesn't "own" its content, so its content doesn't get deleted by `cJSON_Delete`.
int type;
To add items to an object, use `cJSON_AddItemToObject`. Use `cJSON_AddItemToObjectCS` to add an item to an object with a name that is a constant or reference (key of the item, `string` in the `cJSON` struct), so that it doesn't get freed by `cJSON_Delete`.
Using `cJSON_AddItemReferenceToArray` an element can be added as a reference to another object, array or string. This means that `cJSON_Delete` will not delete that items `child` or `valuestring` properties, so no double frees are occuring if they are already used elsewhere.
char *valuestring;
int valueint; /* writing to valueint is DEPRECATED, please use cJSON_SetNumberValue instead */
double valuedouble;
If you want to take an item out of an object, use `cJSON_DetachItemFromObjectCaseSensitive`, it will return the detached item, so be sure to assign it to a pointer, otherwise you will have a memory leak.
char *string;
} cJSON;
Deleting items is done with `cJSON_DeleteItemFromObjectCaseSensitive`. It works like `cJSON_DetachItemFromObjectCaseSensitive` followed by `cJSON_Delete`.
You can also replace an item in an object in place. Either with `cJSON_ReplaceItemInObjectCaseSensitive` using a key or with `cJSON_ReplaceItemViaPointer` given a pointer to an element. `cJSON_ReplaceItemViaPointer` will return `0` if it fails. What this does internally is to detach the old item, delete it and insert the new item in its place.
To get the size of an object, you can use `cJSON_GetArraySize`, this works because internally objects are stored as arrays.
If you want to access an item in an object, use `cJSON_GetObjectItemCaseSensitive`.
To iterate over an object, you can use the `cJSON_ArrayForEach` macro the same way as for arrays.
cJSON also provides convenient helper functions for quickly creating a new item and adding it to an object, like `cJSON_AddNullToObject`. They return a pointer to the new item or `NULL` if they failed.
### Parsing JSON
Given some JSON in a zero terminated string, you can parse it with `cJSON_Parse`.
```c
cJSON *json = cJSON_Parse(string);
```
By default all values are 0 unless set by virtue of being meaningful.
It will parse the JSON and allocate a tree of `cJSON` items that represents it. Once it returns, you are fully responsible for deallocating it after use with `cJSON_Delete`.
The allocator used by `cJSON_Parse` is `malloc` and `free` by default but can be changed (globally) with `cJSON_InitHooks`.
`next`/`prev` is a doubly linked list of siblings. `next` takes you to your sibling,
`prev` takes you back from your sibling to you.
Only objects and arrays have a `child`, and it's the head of the doubly linked list.
A `child` entry will have `prev == 0`, but next potentially points on. The last sibling has `next == 0`.
The type expresses *Null*/*True*/*False*/*Number*/*String*/*Array*/*Object*, all of which are `#defined` in
`cJSON.h`.
If an error occurs a pointer to the position of the error in the input string can be accessed using `cJSON_GetErrorPtr`. Note though that this can produce race conditions in multithreading scenarios, in that case it is better to use `cJSON_ParseWithOpts` with `return_parse_end`.
By default, characters in the input string that follow the parsed JSON will not be considered as an error.
A *Number* has `valueint` and `valuedouble`. `valueint` is a relict of the past, so always use `valuedouble`.
If you want more options, use `cJSON_ParseWithOpts(const char *value, const char **return_parse_end, cJSON_bool require_null_terminated)`.
`return_parse_end` returns a pointer to the end of the JSON in the input string or the position that an error occurs at (thereby replacing `cJSON_GetErrorPtr` in a thread safe way). `require_null_terminated`, if set to `1` will make it an error if the input string contains data after the JSON.
Any entry which is in the linked list which is the child of an object will have a `string`
which is the "name" of the entry. When I said "name" in the above example, that's `string`.
`string` is the JSON name for the 'variable name' if you will.
### Printing JSON
Now you can trivially walk the lists, recursively, and parse as you please.
You can invoke `cJSON_Parse` to get cJSON to parse for you, and then you can take
the root object, and traverse the structure (which is, formally, an N-tree),
and tokenise as you please. If you wanted to build a callback style parser, this is how
you'd do it (just an example, since these things are very specific):
Given a tree of `cJSON` items, you can print them as a string using `cJSON_Print`.
It will allocate a string and print a JSON representation of the tree into it. Once it returns, you are fully responsible for deallocating it after use with your allocator. (usually `free`, depends on what has been set with `cJSON_InitHooks`).
`cJSON_Print` will print with whitespace for formatting. If you want to print without formatting, use `cJSON_PrintUnformatted`.
If you have a rough idea of how big your resulting string will be, you can use `cJSON_PrintBuffered(const cJSON *item, int prebuffer, cJSON_bool fmt)`. `fmt` is a boolean to turn formatting with whitespace on and off. `prebuffer` specifies the first buffer size to use for printing. `cJSON_Print` currently uses 256 bytes for it's first buffer size. Once printing runs out of space, a new buffer is allocated and the old gets copied over before printing is continued.
These dynamic buffer allocations can be completely avoided by using `cJSON_PrintPreallocated(cJSON *item, char *buffer, const int length, const cJSON_bool format)`. It takes a buffer to a pointer to print to and it's length. If the length is reached, printing will fail and it returns `0`. In case of success, `1` is returned. Note that you should provide 5 bytes more than is actually needed, because cJSON is not 100% accurate in estimating if the provided memory is enough.
### Example
In this example we want to build and parse the following JSON:
if (!cJSON_IsNumber(width) || !cJSON_IsNumber(height))
{
prev->next = objects[i];
objects[i]->prev = prev;
status = 0;
goto end;
}
prev = objects[i];
if ((width->valuedouble == 1920) && (height->valuedouble == 1080))
{
status = 1;
goto end;
}
}
return root;
end:
cJSON_Delete(monitor_json);
return status;
}
```
and simply: `Create_array_of_anything(objects, 24);`
cJSON doesn't make any assumptions about what order you create things in.
You can attach the objects, as above, and later add children to each
of those objects.
As soon as you call `cJSON_Print`, it renders the structure to text.
The `test.c` code shows how to handle a bunch of typical cases. If you uncomment
the code, it'll load, parse and print a bunch of test files, also from [json.org](http://json.org),
which are more complex than I'd care to try and stash into a `const char array[]`.
Note that there are no NULL checks except for the result of `cJSON_Parse` because `cJSON_GetObjectItemCaseSensitive` checks for `NULL` inputs already, so a `NULL` value is just propagated and `cJSON_IsNumber` and `cJSON_IsString` return `0` if the input is `NULL`.
### Caveats
@ -426,7 +518,12 @@ However it is thread safe under the following conditions:
When cJSON was originally created, it didn't follow the JSON standard and didn't make a distinction between uppercase and lowercase letters. If you want the correct, standard compliant, behavior, you need to use the `CaseSensitive` functions where available.
#### Duplicate Object Members
cJSON supports parsing and printing JSON that contains objects that have multiple members with the same name. `cJSON_GetObjectItemCaseSensitive` however will always only return the first one.
# Enjoy cJSON!
- Dave Gamble, Aug 2009
- [cJSON contributors](CONTRIBUTORS.md)
- Dave Gamble (original author)
- Max Bruckner (current maintainer)
- and the other [cJSON contributors](CONTRIBUTORS.md)
Max Bruckner
7 years ago
GitHub