class RBS::WASM::Deserializer

Rebuilds RBS::AST objects from the binary buffer produced by rbs_serialize_node (src/serialize.c), driven by the generated SerializationSchema. This is the pure-Ruby counterpart of the C extension’s ast_translation.c, used when the parser runs inside WebAssembly (JRuby).

All locations are reconstructed through the public RBS::Location API, so the same code works whether RBS::Location is backed by the C extension (CRuby) or by a pure-Ruby implementation (JRuby).

Rebuilds RBS::AST objects from the binary buffer produced by rbs_serialize_node (src/serialize.c), driven by SerializationSchema.

Public Class Methods

deserialize (bytes, buffer)

(String bytes, Buffer buffer) → untyped

Source

# File vendor/bundle/ruby/4.0.0/gems/rbs-4.1.1/lib/rbs/wasm/deserializer.rb, line 18
def self.deserialize(bytes, buffer)
  new(bytes, buffer).read_node
end

Deserialize a buffer produced for a whole signature, returning ‘[directives, declarations]` to match RBS::Parser._parse_signature.

Deserialize a buffer produced for a whole signature, returning [directives, declarations] to match RBS::Parser._parse_signature.

deserialize_node_list (bytes, buffer)

(String bytes, Buffer buffer) → Array[untyped]

Source

# File vendor/bundle/ruby/4.0.0/gems/rbs-4.1.1/lib/rbs/wasm/deserializer.rb, line 24
def self.deserialize_node_list(bytes, buffer)
  new(bytes, buffer).read_node_list
end

Deserialize a bare node list (rbs_serialize_node_list), e.g. the result of RBS::Parser._parse_type_params.

Deserialize a bare node list (RBS::Parser._parse_type_params).

deserialize_tokens (bytes, buffer)

(String bytes, Buffer buffer) → Array[[ Symbol, Location ]]

Source

# File vendor/bundle/ruby/4.0.0/gems/rbs-4.1.1/lib/rbs/wasm/deserializer.rb, line 30
def self.deserialize_tokens(bytes, buffer)
  new(bytes, buffer).read_tokens
end

Deserialize the token stream produced by rbs_wasm_lex into the

type, location: pairs RBS::Parser._lex returns.

Deserialize the token stream from rbs_wasm_lex (RBS::Parser._lex).

new (bytes, buffer)

(String bytes, Buffer buffer) → void

Source

# File vendor/bundle/ruby/4.0.0/gems/rbs-4.1.1/lib/rbs/wasm/deserializer.rb, line 34
def initialize(bytes, buffer)
  @bytes = bytes
  @buffer = buffer
  # Symbols and rbs_string fields (comments, annotations) inherit the
  # source encoding, matching ast_translation.c. String/Integer literal
  # nodes are always UTF-8 (see read_node).
  @encoding = buffer.content.encoding
  @pos = 0
  @class_cache = {} #: Hash[String, untyped]
end

Public Instance Methods

read_node ()

() → untyped

Source

# File vendor/bundle/ruby/4.0.0/gems/rbs-4.1.1/lib/rbs/wasm/deserializer.rb, line 45
def read_node
  tag = read_u8
  return nil if tag == 0
  return read_string(@encoding).to_sym if tag == SerializationSchema::SYMBOL_TAG

  entry = SerializationSchema::SCHEMA[tag] or raise "Unknown node tag: #{tag}"

  case entry[0]
  when :node then read_struct(entry)
  when :bool then read_u8 != 0
  when :integer then read_string(Encoding::UTF_8).to_i
  when :string_value then read_string(Encoding::UTF_8)
  when :record_field then [read_node, read_u8 != 0]
  when :signature then [read_node_list, read_node_list]
  when :namespace then RBS::Namespace[read_node_list, read_u8 != 0]
  when :type_name then RBS::TypeName[read_node, read_node]
  else raise "Unknown schema entry kind: #{entry[0].inspect}"
  end
end

Reads the next node and returns the reconstructed Ruby value.

read_node_list ()

() → Array[untyped]

Source

# File vendor/bundle/ruby/4.0.0/gems/rbs-4.1.1/lib/rbs/wasm/deserializer.rb, line 65
def read_node_list
  Array.new(read_count) { read_node }
end

read_tokens ()

() → Array[[ Symbol, Location ]]

Source

# File vendor/bundle/ruby/4.0.0/gems/rbs-4.1.1/lib/rbs/wasm/deserializer.rb, line 71
def read_tokens
  tokens = [] #: Array[[ Symbol, Location ]]
  until @pos >= @bytes.bytesize
    type = read_string(Encoding::UTF_8).to_sym
    start_char = read_i32
    end_char = read_i32
    tokens << [type, RBS::Location.new(@buffer, start_char, end_char)]
  end
  tokens
end

The lex stream has no leading count: read records until the buffer is exhausted. Each is a token type name followed by its character range.