ezStruct for Node.js

Easiest way to pack/unpack struct data with three steps in Node.js.

1. Provide C-style struct declarations (intact C source code) and get the context
2. Fetch the struct Schema from the context
3. Pack / Unpack data.

What's Supported

• typedef (also works for struct, enum and array)
• (multi-dimension) array
• (nested) struct
• enum
• anonymous enum/struct
• basic number types (little-endian {8,16,32}-bit integer, 32-bit float and double)
• char something[64] as string

Quickstart

Install via: npm install --save ezstruct

const ezstruct = require('ezstruct')
ezstruct.setStringEncoding('utf-8')
// install iconv-lite to handle more encodings

// create context
// ES6 template tag is supported.

var context = ezstruct `
typedef uint32_t uid_t;

typedef struct {
  uid_t id;
  int age;
  char name[32];
} student;
`

// then you can use `context.student` the schema

// (optional) modify schema. useful for char[]

context.student.name.default = "Unknown"
context.student.name.encoding = "utf-8"

// JS to binary

var raw = context.student.toBinary({
  id: 12345678,
  age: 19,
})

// binary to JS

var unpacked = context.student.fromBinary(raw)

Document

Context

Context is an object containing all data type declarations and handlers(parser/writer). By default, it contains int{8,16,32}, char[], bool and some common types.

Call ezstruct = require('ezstruct') and get the context builder function.

ezstruct(source[, context])

• source: string C-style declarations. You may get this from C header files.
• context: Context optional. The new Context can inherit types from another Context.
• (returns) Context

Once you get the Context, you may get data-type Schemas like ctx["struct tm"], or ctx.timer_t if it's declared.

ezstruct.setStringEncoding(encoding)

• encoding: string|null
• (remarks)
  1. Call this before creating an Context, then all new char[] fields of structs will be affected.
  2. Also works for multi-dimension char array.
  3. This method will not affect Contexts that are created before.
  4. Install iconv-lite package if you want to handle Encodings that Not Supported by NodeJS
  5. iconv-lite is not in dependency list. Run npm i --save iconv-lite before using.

Schema

Every schema has these properties:

• bytes: number the length of one instance

And these methods:

read(buf, offset)

• buf: Buffer where ezStruct parses data from
• offset: number the offset (in bytes) of data
• (returns) data it read

write(buf, offset, value)

• buf: Buffer where data shall be written to
• offset: number the offset (in bytes) of data
• value your data

Struct (Schema)

Struct comes from struct declarations. Plus, it provides these properties:

• struct: StructItem[] containing the sort of items.
• fields: {your_var_name: StructItem} the dict version of struct.
• your_var_name: StructItem same as fields.your_var_names. if not conflicted.

methods:

fromBinary(buf[, offset])

• buf: Buffer where ezStruct parses data from
• offset: number the offset in bytes. 0 by default.

toBinary(obj)

• obj: object the thing to be packed
• (returns) Buffer

toBinary(obj, buf, offset)

• obj: object the thing to be packed
• buf: Buffer which stores the binary
• offset: number

StructItem (sub-class from Struct)

Properties:

• name: string
• type: string like "int" or "char[128]"
• default: any (optional) default value
• encoding: string (optional) for char strings only. see ezstruct.setStringEncoding

Enum (Schema)

It is an alias of int, with extra properties:

• dict: {enum_item_name: number} a dict object
• enum_item_name: number same as dict.enum_item_name. if not conflicted.

TO-DO

• union
• big-endian
• set char encoding with comments