Binary data

Cloth provides explicit integer encoding and decoding in a byte[]. The operations do not expose host byte order or memory addresses.

byte[] bytes = [0, 0, 0, 0];
int32 value = 16909060;
value::writeBigEndian(bytes, 0);
println(bytes[0]); // 1
println(bytes[3]); // 4

int32 restored = bytes::readInt32BigEndian(0);
println(restored); // 16909060

Writing

value::writeLittleEndian(destination, offset) and value::writeBigEndian(destination, offset) return void.

The receiver must be an integer. The destination must be a non-null byte[], and the offset must be int32. The operation writes exactly the integer’s bit width divided by eight. Signed values use their fixed-width two’s-complement bit pattern.

Reading

Read operations use the byte array as receiver and take an int32 offset. Each prefix below supports both LittleEndian and BigEndian suffixes:

For example, bytes::readUint64LittleEndian(offset) returns uint64. Alias-based names such as readIntLittleEndian and readUintBigEndian do not exist. Eight-bit little-endian and big-endian operations produce the same bytes, but both spellings are available.

Evaluation and bounds

Receivers and arguments evaluate once, from left to right. The offset must be non-negative and the entire operation must fit within the array.

The runtime checks the full range before reading or writing. An invalid operation traps with integer byte range is out of bounds. An invalid write changes no bytes.

Floating-point bit access, unsafe views, and general serialization are not provided by these operations.