module

Mica::Core::ByteScan

Byte-run scanning and copying for the hot paths of format scanners and trees (text runs, attribute values, names, stored values).

This is the one place in Mica that touches bytes without a bounds check on every access: each method checks its whole window against the slice once, before it reads or writes, and never leaves that window. Everything else goes through checked Slice indexing.

Class methods

copy_into(target : Bytes, offset : Int32, source : Bytes) : Nil

Copies all of source into target starting at offset, without building a subslice. Raises IndexError unless 0 <= offset <= target.size - source.size, and when target is read-only.

target = Bytes.new(4)
Mica::Core::ByteScan.copy_into(target, 1, "ab".to_slice)
target # => Bytes[0, 97, 98, 0]

Source

matches?(bytes : Bytes, offset : Int32, expected : Bytes) : Bool

Returns true when the expected.size bytes of bytes at offset equal expected. Raises IndexError unless 0 <= offset <= bytes.size - expected.size. Compares in place, which beats building a slice and calling memcmp for short names.

Mica::Core::ByteScan.matches?("<a></a>".to_slice, 5, "a".to_slice) # => true

Source

skip_while(bytes : Bytes, from : Int32, stop : Int32, & : UInt8 -> Bool) : Int32

Returns the index of the first byte in bytes[from...stop] for which the block returns false, or stop when it returns true for all of them. Raises IndexError unless 0 <= from <= stop <= bytes.size.

Mica::Core::ByteScan.skip_while("abc;d".to_slice, 0, 5) { |byte| byte != ';'.ord } # => 3

Source