pack-d 0.2.0

pack-d

License

Licensed under MIT License. See LICENSE file.

About

Pack-D is small binary IO helper written in D Programming Language based on Python's struct module. It provides simple interface for reading and writing binary encoded data.

Installation

Manual

Download source/binary/pack.d and add it to your project.

Using DUB

Add pack-d dependency to your package.json file:

"dependencies": {
	"pack-d": ">=0.1.0"
}

Example

import binary.pack;
import std.stdio;

void main()
{
    int a, b, c;
    ubyte[] bytes;
    
    /// Packing 3 integers to binary
    bytes = pack(20, 30, 40);
    
    /// Unpack 3 integers from bytes to a, b and c
    bytes.unpackTo(a, b, c);
    writefln("%d %d %d", a, b, c); // 20 30 40
    
    /// Pack 2 shorts and a string
    bytes = pack!`hhs`(42, 18, "hello, world!");
    writeln(bytes.unpack!`hhs`); /// Tuple!(short, short, string)(42, 18, "hello, world!")
    
    /// Pack ushort, uint and ulong (big endian)
    bytes = pack!`>HIL`(42, 80, 150);
    /// Unpack ushort, skip 4 bytes and unpack ulong
    writeln(bytes.unpack!`>H4xL`); /// Tuple!(ushort, ulong)(42, 150)
}

Format reference

Most Pack-D functions use a format string to define types of values. Format strings can be ommited and value types are inferred, although it is strongly recommended to specify it whenever possible.

Format strings used in general are very similar to format strings used by Python's struct module.

Available modifier characters

Available type specifiers

TIP: Common rule for (almost) all type specifiers is that all lowercase letters represent signed types and

uppercase letters represent unsigned types.

Types with size 4/8 (p and P) depend on local machine architecture. On 32 bit architectures they occupy 4 bytes, on 64 bit architectures they occupy 8 bytes.

Additionaly all type specifiers can be preceded by number of occurences. For example, pack!"cc"('a', 'b') is equivalent to pack!"2c"('a', 'b'). Note that behaviour is different with strings: if type specifier is preceded by a number and parameter is an array, n characters are packed. For example: pack!"5c"("Hello World") will pack only first 5 characters.

When number appears in unpack format string, returned type is an static array. For example, type of unpack!"6h" is short[6].

Character x have slighty different meaning depending if used in pack or unpack functions. When passed to pack, null byte is added to output. When passed to unpack, one byte is skipped from input data.

Examples:

• 2h 2 signed shorts (native endian)
• <2I 2 insigned integers (little endian)
• i4xL signed integer, 4 null bytes and unsigned long
• Sx or s null terminated string

Quick API reference

• pack([string format])(T... params)

Packs specified parameters according to format. Passing inconvertible parameter and type specifier, results in static assert failure. All packed data is returned as ubyte[].

• pack([string format])(File file, T... params)

Works exacly like previous one, except that all packed data is written to file.

• unpackTo([string format])([ref] Range range, T... params)
unpackTo([string format])(File range, T... params)

Unpacks data from range or file and writes it to params. Range is taken by refernce whenever possible (auto ref), which means passed array of bytes is modified. To prevent that, pass yourarray.save as first parameter.

NOTE: Specified Range must be a valid input range of ubyte element type.

• unpack(string format)([ref] Range range)
unpack(string format)(File file)

Works exacly like previous one, except that all data is returned as tuple. In this overload format is required.

• unpacker(string format)( Range range)
unpacker(string format)(File range)
  

Returns instance of Unpacker struct. Useful when there's repeating binary encoded data.

   ubyte[] bytes = pack!`<hshs`(1, "one", 2, "two");
   auto unpacker = unpacker!`<hs`(bytes);
   
   foreach(num, str; unpacker)
   {
       writeln(num, " ", str); // Prints 1 one\n 2 two
   }