Jump to content

Parrot intermediate representation

From Wikipedia, the free encyclopedia
This is an old revision of this page, as edited by Scientus (talk | contribs) at 01:22, 8 May 2009. The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

The Parrot intermediate representation or PIR, previously called Intermediate code (IMC), is one of the two assembly languages for the Parrot virtual machine. The other is Parrot assembly language or PASM. Compared to PASM, PIR exists at a slightly higher level of abstraction, and provides temporary registers and named registers, simplifying code generation.

While parrot is still evolving, it is currently being used in many different capacities, and has undergone several releases.

Overview

PIR provides a set of abstractions that allow the programmer to ignore certain redundancies in the Parrot bytecode and quickly write code that adheres to the complexities of Parrot, such as the calling conventions.

Abstractions

PIR provides both type abstraction and polymorphism to some degree. For example, the "+" operator can be used with int, num or both:

.local int a
.local num b
a = 1
b = 1.1
.local num c
c = a + b

Calling conventions

The calling conventions in Parrot are fairly complex, but all of that complexity can be hidden by using PIR directives:

.sub foo
 .param int a
 .param int b
 .local int tmp
 tmp = a + b
 .return (tmp)
.end

Each of these directives that begins with a "." expands out to the required Parrot bytecode, but does not directly represent any fundamental Parrot operation.

Example

The hello world program in PIR is

.sub hello :main
  print "Hello world!\n"
.end

If the program is saved as hello.pir, it can be compiled and executed with this command: parrot hello.pir