majerle / lwjson Goto Github PK

This is to be used by user prior input token set allocation.

Using this feature, user can check if input is valid and can understand how many tokens are necessary for successful JSON parsing.

prv_skip_blank has the following comment:

Skip all characters that are considered *blank* as per RFC4627

However, \f has never been considered a whitespace character in RFC4627.

Also applies to streaming mode.

To be able to generate JSON string output from input tokens.

• Must be able to generate all token types as output string
• Float generation support must be minimal. Very basic with fixed precision
• Minimalistic output with optional beautify support (TBC)

should this macro be adding one in the return value?

LwJSON returns error if there is a comment between key and value, as:

{"key":/* Comment */"value"}

long long integers are not correctly supported due to how prv_parse_number is using lwjson_real_t for parsing integers. Even if lwjson_real_t is defined as a double, this only allows storing up to 53 bit integers (assuming you are using a system which supports IEEE 754 double precision floating point numbers). I propose adding a second variable handle the value as an integer. I will follow up with a pull request.

To think supporting extended JSON5 features:

https://json5.org/

Use LwMEM library for dynamic token generation if enabled by user. Not necessary in default mode.

When unicode characters are part of json string, length is wrongly calculated.

Hi @MaJerle,

would be nice to add a parser function to manage an input stream data. What do you think about?

Best regards,

Giulio Dalla Vecchia

Will you have JSON-RPC 2.0 ?