import "android.googlesource.com/platform/tools/gpu/binary/vle"
Package vle implements binary.Reader and binary.Writer using a variable length encoding format.
Boolean values are encoded as single bytes, where 0 represents false and non- zero represents true.
8 bit values are encoded as single bytes. 16, 32 and 64 bit unsigned integers are encoded into a one or more bytes. The number of sequential ones starting from the most-significant bit of the first encoded byte describe the number of additional bytes that make up the unsigned integer. If the run of ones does not fill the byte, then the run is terminated with a zero bit. The unsigned integer value is then big-endian encoded into the remainder of the bits from the first byte and any additional bytes.
For example, the 16-bit number 0xABC would be encoded as follows:
MSB LSB MSB LSB
╔═══╤═══╤═══╤═══╤═══╤═══╤═══╤═══╗ ╔═══╤═══╤═══╤═══╤═══╤═══╤═══╤═══╗
║C₀ │C₁ │D₁₃│D₁₂│D₁₁│D₁₀│D₉ │D₈ ║ ║D₇ │D₆ │D₅ │D₄ │D₃ │D₂ │D₁ │D₀ ║
║ │ │ │ │ │ │ │ ║ ║ │ │ │ │ │ │ │ ║
║ 1 │ 0 │ 0 │ 0 │ 1 │ 0 │ 1 │ 0 ║ ║ 1 │ 0 │ 1 │ 1 │ 1 │ 1 │ 0 │ 0 ║
╚═══╧═══╧═══╧═══╧═══╧═══╧═══╧═══╝ ╚═══╧═══╧═══╧═══╧═══╧═══╧═══╧═══╝
Byte₀ Byte₁
C has a run of 1, meaning there is one extra byte of data. D holds the unsigned integer value 0xABC with bits 0000 1010 1011 1100.
Signed integers are converted to unsigned integers by interleaving negative then positive numbers before being encoded as unsigned integers (as described above). For example the signed numbers [±0, -1, +1, -2, +2] are first transformed to the unsigned integers [0, 1, 2, 3, 4] before being encoded. This is done so that small negative and small positive numbers are encoded with fewer bytes.
32 bit floating-point numbers are first converted to a 32 bit unsigned integer using math.Float32bits and then byte-reversed before being encoded as a unsigned integer.
64 bit floating-point numbers are first converted to a 64 bit unsigned integer using math.Float64bits and then byte-reversed before being encoded as a unsigned integer.
The bytes of floating-point numbers are reversed so that floating pointer numbers with simple fractional parts are encoded with fewer bytes, as these are considered the common case.
String encodes a 32 bit unsigned integer representing the length of the string in bytes followed by the string data encoded in UTF-8:
count uint32 // in bytes data0, data1, data2... byte // string in UTF-8
func Reader(r io.Reader) binary.Reader
Reader creates a binary.Reader that reads from the provided io.Reader.
func Writer(w io.Writer) binary.Writer
Writer creates a binary.Writer that writes to the supplied io.Writer.