rubyx/lib/vool/send_statement.rb

149 lines
5.2 KiB
Ruby
Raw Normal View History

2017-04-01 20:28:57 +02:00
module Vool
2018-03-10 14:17:36 +01:00
# Sending in a dynamic language is off course not as simple as just calling.
# The function that needs to be called depends after all on the receiver,
# and no guarantees can be made on what that is.
#
# It helps to know that usually (>99%) the class of the receiver does not change.
# Our stategy then is to cache the functions and only dynamically determine it in
# case of a miss (the 1%, and first invocation)
#
# As cache key we must use the type of the object (which is the first word of _every_ object)
# as that is constant, and function implementations depend on the type (not class)
2017-04-02 11:59:07 +02:00
class SendStatement < Statement
2018-06-26 19:28:27 +02:00
attr_reader :name , :receiver , :arguments , :block
def initialize(name , receiver , arguments )
@name , @receiver , @arguments = name , receiver , arguments
@arguments ||= []
2017-04-02 17:25:30 +02:00
end
2017-04-08 11:10:42 +02:00
2018-06-28 19:15:24 +02:00
def block
return nil if arguments.empty?
bl = arguments.last
bl.is_a?(BlockStatement) ? bl : nil
end
def add_block( block )
@arguments << block
2018-06-26 19:28:27 +02:00
end
def normalize
statements = Statements.new([])
arguments = []
2018-04-27 20:55:41 +02:00
@arguments.each_with_index do |arg , index |
normalize_arg(arg , arguments , statements)
end
if statements.empty?
return SendStatement.new(@name, @receiver , @arguments)
else
statements << SendStatement.new(@name, @receiver , arguments)
return statements
end
end
2018-04-27 20:55:41 +02:00
def normalize_arg(arg , arguments , statements)
if arg.respond_to?(:slot_definition) and !arg.is_a?(SendStatement)
arguments << arg
return
end
assign = LocalAssignment.new( "tmp_#{arg.object_id}".to_sym, arg)
statements << assign
2018-04-27 20:55:41 +02:00
arguments << LocalVariable.new(assign.name)
end
def to_s
"#{receiver}.#{name}(#{arguments.join(', ')})"
end
def each(&block)
block.call(self)
block.call(@receiver)
@arguments.each do |arg|
block.call(arg)
end
self.block.each(&block) if self.block
2017-04-08 11:10:42 +02:00
end
2018-03-18 18:06:36 +01:00
# lazy init this, to keep the dependency (which goes to parfait and booting) at bay
def dynamic_call
@dynamic ||= Mom::DynamicCall.new()
end
# A Send breaks down to 2 steps:
# - Setting up the next message, with receiver, arguments, and (importantly) return address
2017-09-14 15:07:02 +02:00
# - a CachedCall , or a SimpleCall, depending on wether the receiver type can be determined
def to_mom( compiler )
@receiver = SelfExpression.new(compiler.method.self_type) if @receiver.is_a?(SelfExpression)
2017-09-10 21:54:56 +02:00
if(@receiver.ct_type)
simple_call(compiler)
2017-09-10 21:54:56 +02:00
else
cached_call(compiler)
2017-09-10 21:54:56 +02:00
end
end
# When used as right hand side, this tells what data to move to get the result into
# a varaible. It is (off course) the return value of the message
def slot_definition(compiler)
Mom::SlotDefinition.new(:message ,[ :return_value])
end
def message_setup(compiler,called_method)
setup = Mom::MessageSetup.new( called_method )
mom_receive = @receiver.slot_definition(compiler)
2017-09-10 21:54:56 +02:00
arg_target = [:message , :next_message , :arguments]
args = []
2018-05-15 15:25:55 +02:00
@arguments.each_with_index do |arg , index| # +1 because of type
args << Mom::SlotLoad.new( arg_target + [index + 1] , arg.slot_definition(compiler))
end
setup << Mom::ArgumentTransfer.new( mom_receive , args )
end
def simple_call(compiler)
type = @receiver.ct_type
called_method = type.resolve_method(@name)
raise "No method #{@name} for #{type}" unless called_method
message_setup(compiler,called_method) << Mom::SimpleCall.new(called_method)
end
2017-09-14 15:07:02 +02:00
# this breaks cleanly into two parts:
# - check the cached type and if neccessary update
# - call the cached method
def cached_call(compiler)
cache_check(compiler) << call_cached_method(compiler)
end
2017-09-14 15:07:02 +02:00
# check that current type is the cached type
# if not, change and find method for the type (simple_call to resolve_method)
# conceptually easy in ruby, but we have to compile that "easy" ruby
def cache_check(compiler)
ok = Mom::Label.new("cache_ok_#{self.object_id}")
check = build_condition(ok, compiler) # if cached_type != current_type
check << Mom::SlotLoad.new([dynamic_call.cache_entry, :cached_type] , receiver_type_definition(compiler))
check << Mom::ResolveMethod.new( @name , dynamic_call.cache_entry )
check << ok
2017-09-14 15:07:02 +02:00
end
# to call the method (that we know now to be in the cache), we move the method
# to reg1, do the setup (very similar to static) and call
def call_cached_method(compiler)
message_setup(compiler,dynamic_call.cache_entry) << dynamic_call
2017-09-14 15:07:02 +02:00
end
2017-12-05 20:46:37 +01:00
2018-07-03 21:18:19 +02:00
def to_s(depth = 0)
sen = "#{receiver}.#{name}(#{@arguments.collect{|a| a.to_s}.join(', ')})"
at_depth(depth , sen)
end
2017-09-14 17:42:01 +02:00
private
def receiver_type_definition(compiler)
defi = @receiver.slot_definition(compiler)
defi.slots << :type
defi
end
def build_condition(ok_label, compiler)
2018-03-18 18:06:36 +01:00
cached_type = Mom::SlotDefinition.new(dynamic_call.cache_entry , [:cached_type])
current_type = receiver_type_definition(compiler)
Mom::NotSameCheck.new(cached_type , current_type, ok_label)
2017-09-14 17:42:01 +02:00
end
2017-04-01 20:28:57 +02:00
end
end