LEB (little endian base) 128 encoding and decoding

Leb128 (little endian base 128) is a variable-length data format (variable number of bytes) which can be divided into unsinged leb128 and signed leb128.

It can be used to store any large integer in a small number of bytes.

I. Encoding unsigned leb128

The calculation method is like this (dwarf-2.0.0.pdf Appendix 4 99 pages)

do{  byte = low order 7 bits of value;  value >>= 7;  if (value != 0) /* more bytes to come */    set high order bit of byte;  emit byte;} while (value != 0);

Input: Unsigned int A = 12857 (0x3239)

Output: Unsigned leb128 result

Logic:

1>. Convert 0x3239 to binary

A
= 0000 0000 0000 0000 0011 0010
1001

2>. Capture 7 bits (gray back
Scene) to form a new byte

Byte0 = 0x39

3>. Add seven zeros for the original data.

A = 0000 000 0 0000 0000 0000 0000
0100

4>. If (! = 0), byte | = 0x80

Byte0 = 0x39 + 0x80 = 0xb9

5>. Repeat 3>, 4> until a = 0

Byte1
= 0x64

A
= 0000 000 0 0000 00 00 0000 0000 0000

The final result is B9.
64

Ii. Decoding
Unsigned leb128

Algorithm (dwarf-2.0.0.pdf Appendix 4 99 pages)

1>. Take the next byte:

Bytex = 0xb9

2>. Take it as low as 7 bits and place it in the position of the nth 7 bits.

Result | = (bytex & 0x7f) <(7 * X );

3>. If (! (Bytex | 0x80), continue 1>, otherwise it ends.

Byte0 = 0xb9;

Result = 0x39;

Byte0 & 0x80 = 1

Byte1 = 0x64

Result = 0x3239

Iii. Features of leb128

1. The number of bytes occupied by a leb128 data is unknown.

2. Whether the maximum bit of a leb128 byte is 0 indicates whether the byte stream has ended.

3. A leb128 may be converted to an unsigned int through codec.

Iv. References:

1.dwarf-2.0.0.pdf section 7.6 and
Http://dwarfstd.org/Appendix 4/

2. libdwarf dwarf_leb.c contains the leb128 decoding function.

References:

Blog of micklongen: http://blog.chinaunix.net/uid-20704646-id-95934.html

Http://en.wikipedia.org/wiki/LEB128