ArduinoJson/doc/Avoiding pitfalls.md

Avoiding common pitfalls in Arduino JSON

As StaticJsonBuffer is the corner stone of this library, you'll see that every pitfall listed here is related to a wrong understanding of the memory model.

Make sure you read [Arduino JSON memory model](Memory model.md) before going further.

1. Make StaticJsonBuffer big enough

By design, the library has no way to tell you why parseArray() or parseObject() failed.

There are basically two reasons why they may fail:

1. the JSON string is invalid,
2. the JSON string contains more values that the buffer can store.

So, if you are sure the JSON string is correct and you still can't parse it, you should slightly increase the number of token of the parser.

2. Make sure everything fits in memory

You may go into unpredictable trouble if you allocate more memory than your processor really has. It's a very common issue in embedded development.

To diagnose this, look at every big objects in you code and sum their size to check that they fit in RAM.

For example, don't do this:

char json[1024];        // 1 KB
JsonParser<512> parser;    // 514 B

because it may be too big for a processor with only 2 KB: you need free memory to store other variables and the call stack.

That is why an 8-bit processor is not able to parse long and complex JSON strings.

3. Keep the StaticJsonBuffer in memory long enough

Remember that the function of StaticJsonBuffer return references. References don't store data, they are just pointer to the actual. This will only work if the data actual is still in memory.

For example, don't do this:

JsonArray& getArray(char* json)
{
    StaticJsonBuffer<200> buffer;
    return buffer.parseArray(json); 
}

because the local variable buffer will be removed from memory when the function parseArray() returns, and the JsonArray& will point to an invalid location.

4. Don't make StaticJsonBuffer global

If you read the previous point, you may come to the idea of using a global variable for your StaticJsonBuffer. This is probably a bad idea because StaticJsonBuffer can be quite big (depending on your requirement) and would be eating a lot of memory, even when you don't use it.

There are some cases were a StaticJsonBuffer can be a global variable, but must of the time you should declare it in a local scope, in a function which unique role is to handle the JSON serialization.

5. Keep the JSON string in memory long enough

The library never make memory duplication. This has an important implication on string values, it means that the library will return pointer to chunks of the string.

For instance, let's imagine that you parse ["hello","world"], like this:

char[] json = "[\"hello\",\"world\"]";
StaticJsonBuffer<32> buffer;
JsonArray& array = buffer.parseArray(json);

const char* first = array[0];
const char* second = array[1];

In that case, both first and second are pointer to the content of the original string json. So this will only work if json is still in memory.

6. JSON string is altered

If you read carefully the previous pitfall, you may I come to the conclusion that the JSON parser modifies the JSON string.

Indeed, the parser modifies the string for two reasons:

1. it inserts \0 to terminate substrings,
2. it translate escaped characters like \n or \t.

Most of the time this wont be an issue, but it there are some corner case that can be problematic.

Let take the example above:

char[] json = "[\"hello\",\"world\"]";
StaticJsonBuffer<32> buffer;
JsonArray& array = buffer.parseArray(json);

If you replace it by:

char* json = "[\"hello\",\"world\"]";
StaticJsonBuffer<32> buffer;
JsonArray& array = buffer.parseArray(json);

Depending on your platform, you may have an exception because the parser tries to write at a location that is read-only. In the first case char json[] declares an array of char initialized to the specified string. In the second case char* json declares a pointer to a read only string, in fact it should be a const char* instead of a char*.