PDD 19: Parrot Intermediate Representation (PIR)


This document outlines the architecture and core syntax of Parrot Intermediate Representation (PIR).


PIR is a stable, middle-level language intended both as a target for the generated output from high-level language compilers, and for human use developing core features and extensions for Parrot.

Basic Syntax

A valid PIR program consists of a sequence of statements, directives, comments and empty lines.


A statement starts with an optional label, contains an instruction, and is terminated by a newline (<NL>). Each statement must be on its own line.

  [label:] [instruction] <NL>

An instruction may be either a low-level opcode or a higher-level PIR operation, such as a subroutine call, a method call, a directive, or PIR syntactic sugar.


A directive provides information for the PIR compiler that is outside the normal flow of executable statements. Directives are all prefixed with a ".", as in .local or .sub.


Comments start with # and last until the following newline. PIR also allows comments in Pod format. Comments, Pod content, and empty lines are ignored.


Identifiers start with a letter or underscore, then may contain additionally letters, digits, and underscores. Identifiers don't have any limit on length at the moment, but some sane-but-generous length limit may be imposed in the future (256 chars, 1024 chars?). The following examples are all valid identifiers.


Opcode names are not reserved words in PIR, and may be used as variable names. For example, you can define a local variable named print. Note that currently, by using an opcode name as a local variable name, the variable will hide the opcode name, effectively making the opcode unusable. In the future this will be resolved.

The PIR language is designed to have as few reserved keywords as possible. Currently, in contrast to opcode names, PIR keywords are reserved, and cannot be used as identifiers. Some opcode names are, in fact, PIR keywords, which therefore cannot be used as identifiers. This, too, will be resolved in a future re-implementation of the PIR compiler.

The following are PIR keywords, and cannot currently be used as identifiers:

 goto      if       int         null
 num       pmc      string      unless


A label declaration consists of a label name followed by a colon. A label name conforms to the standard requirements for identifiers. A label declaration may occur at the start of a statement, or stand alone on a line, but always within a subroutine.

A reference to a label consists of only the label name, and is generally used as an argument to an instruction or directive.

A PIR label is accessible only in the subroutine where it's defined. A label name must be unique within a subroutine, but it can be reused in other subroutines.

  goto label1
     # ...

Registers and Variables

There are two ways of referencing Parrot's registers. The first is through named local variables declared with .local.

  .local pmc foo

The type of a named variable can be int, num, string or pmc, corresponding to the types of registers. No other types are used.

The second way of referencing a register is through a register variable $In, $Sn, $Nn, or $Pn. The capital letter indicates the type of the register (integer, string, number, or PMC). n consists of digit(s) only. There is no limit on the size of n. There is no direct correspondence between the value of n and the position of the register in the register set, $P42 may be stored in the zeroth PMC register, if it is the only register in the subroutine.


Constants may be used in place of registers or variables. A constant is not allowed on the left side of an assignment, or in any other context where the variable would be modified.

'single-quoted string constant'
Are delimited by single-quotes ('). They are taken to be ASCII encoded. No escape sequences are processed.
"double-quoted string constants"
Are delimited by double-quotes ("). A " inside a string must be escaped by \. The default format for a double-quoted string constant is 7-bit ASCII, other character sets and encodings must be marked explicitly using a format flag.
<<"heredoc", <<'heredoc'
Heredocs work like single or double quoted strings. All lines up to the terminating delimiter are slurped into the string. The delimiter has to be on its own line, at the beginning of the line and with no trailing whitespace.Assignment of a heredoc:
  $S0 = <<"EOS"
A heredoc as an argument:
  .local pmc function, arg
  # ...

  function(<<"END_OF_HERE", arg)


Although currently not possible, a future implementation of the PIR language will allow you to use multiple heredocs within a single statement or directive:
format:"string constant"
Like above with a format attached to the string. Valid formats are currently: ascii (the default), binary, iso-8859-1, utf8, utf16, ucs2, and ucs4.The format is attached to the string constant, and adopted by any string container the constant is assigned to.The standard escape sequences are honored within strings with an alternate format, so you can include a particular Unicode character as either a literal sequence of bytes, or as an escape sequence.

String escape sequences

Inside double-quoted strings the following escape sequences are processed.

  \xhh        1..2 hex digits
  \ooo        1..3 oct digits
  \cX         control char X
  \x{h..h}    1..8 hex digits
  \uhhhh      4 hex digits
  \Uhhhhhhhh  8 hex digits
  \a, \b, \t, \n, \v, \f, \r, \e, \\, \"
numeric constants
Both integers (42) and numbers (3.14159) may appear as constants. 0x and 0b denote hex and binary constants respectively.


.local <type> <identifier>
Define a local name identifier within a subroutine with the given type. You can define multiple identifiers of the same type by separating them with commas:
  .local int i, j
.lex <string constant>, <reg>
Declare a lexical variable that is an alias for a PMC register. For example the following two snippets have an identical effect:
    .lex '$a', $P0
    $P1 = new 'Integer'
    $P0 = $P1
    .lex '$a', $P0
    $P1 = new 'Integer'
    store_lex '$a', $P1
    And these two snippets also have an identical effect:
    .lex '$a', $P0
    $P1 = new 'Integer'
    $P1 = $P0
    .lex '$a', $P0
    $P1 = new 'Integer'
    $P1 = find_lex '$a'
.const <type> <identifier> = <const>
Define a constant named identifier of type type and assign value const to it. The type must be int, num, string or a string constant indicating the PMC type. This allows you to create PMC constants representing subroutines; the value of the constant in that case is the name of the subroutine. If the referred subroutine has an :immediate modifier and it returns a value, then that value is stored instead of the subroutine..const declarations representing subroutines can only be written within a .sub. The constant is stored in the constant table of the current bytecode file.
.globalconst <type> <identifier> = <const>
As .const above, but the defined constant is globally accessible. .globalconst may only be used within a .sub.
  .sub <identifier> [:<modifier> ...]
  .sub <quoted string> [:<modifier> ...]
Define a subroutine. All code in a PIR source file must be defined in a subroutine. See the section "Subroutine modifiers" for available modifiers. Optional modifiers are a list separated by spaces.The name of the sub may be either a bare identifier or a quoted string constant. Bare identifiers must be valid PIR identifiers (see Identifiers above), but string sub names can contain any characters, including characters from different character sets (see Constants above).Always paired with .end.
End a subroutine. Always paired with .sub.
.namespace [ <identifier> ; <identifier> ]
   .namespace [ <key>? ]

   key: <identifier> [';' <identifier>]*
Defines the namespace from this point onwards. By default the program is not in any namespace. If you specify more than one, separated by semicolons, it creates nested namespaces, by storing the inner namespace object in the outer namespace's global pad.You can specify the root namespace by using empty brackets, such as:
    .namespace [ ]
The brackets are not optional, although the key inside them is.
.loadlib 'lib_name'
Load the given library at compile time, that is, as soon that line is parsed. See also the loadlib opcode, which does the same at run time.A library loaded this way is also available at runtime, as if it has been loaded again in :load, so there is no need to call loadlib at runtime.
.HLL <hll_name>
Define the HLL namespace from that point on in the file. Takes one string constant, the name of the HLL. By default, the HLL namespace is 'parrot'.
.line <integer>
Set the current PIR line number to the value specified. This is useful in case the PIR code is generated from some source PIR files, and error messages should print the source file's line number, not the line number of the generated file. Note that line numbers increment per line of PIR; if you are trying to store High Level Language debug information, you should instead be using the .annotate directive.
.file <quoted_string>
Set the current PIR file name to the value specified. This is useful in case the PIR code is generated from some source PIR files, and error messages should print the source file's name, not the name of the generated file.
.annotate <key>, <value>
Makes an entry in the bytecode annotations table. This is used to store high level language debug information. Examples:
  .annotate "file", "aardvark.p6"
  .annotate "line", 5
  .annotate "column", 24
An annotation stays in effect until the next annotation with the same key or the end of the current file (that is, if you use a tool such as pbc_merge to link multiple bytecode files, then annotations will not spill over from one mergee's bytecode to another).One annotation covers many PIR instructions. If the result of compiling one line of HLL code is 15 lines of PIR, you only need to emit one annotation before the first of those 15 lines to set the line number.
  .annotate "line", 42
The key must always be a quoted string. The value may be an integer, a number or a quoted string. Note that integer values are stored most compactly; should you instead of the above annotate directive emit:
  .annotate "line", "42"
then instead "42" is stored as a string, taking up more space in the resulting bytecode file.

Subroutine modifiers

Define "main" entry point to start execution. If multiple subroutines are marked as :main, the last marked subroutine is used. Only the first file loaded or compiled counts; subs marked as :main are ignored by the load_bytecode op. If no :main modifier is specified, execution starts at the first subroutine in the file.
Run this subroutine when loaded by the load_bytecode op (i.e. neither in the initial program file nor compiled from memory). This is complementary to what :init does (below); to get both behaviours, use :init :load. If multiple subs have the :load pragma, the subs are run in source code order.
Run the subroutine when the program is run directly (that is, not loaded as a module), including when it is compiled from memory. This is complementary to what :load does (above); to get both behaviours, use :init :load.
Do not install this subroutine in the namespace. Allows the subroutine name to be reused.
:multi(type1, type2...)
Engage in multiple dispatch with the listed types. See docs/pdds/pdd27_multi_dispatch.pod for more information on the multiple dispatch system.When used in combination with :method (below), the first type (type1) refers to the type of the invocant (self).
Execute this subroutine immediately after being compiled, which is analogous to BEGIN in Perl 5.In addition, if the sub returns a PMC value, that value replaces the sub in the constant table of the bytecode file. This makes it possible to build constants at compile time, provided that (a) the generated constant can be computed at compile time (i.e. doesn't depend on the runtime environment), and (b) the constant value is of a PMC class that supports saving in a bytecode file.{{ TODO: need a freeze/thaw reference }}.For instance, after compilation of the sub 'init', that sub is executed immediately (hence the :immediate modifier). Instead of storing the sub 'init' in the constants table, the value returned by 'init' is stored, which in this example is a FixedIntegerArrray.
    .sub main :main
      .const "Sub" initsub = "init"

    .sub init :immediate
      .local pmc array
      array = new 'FixedIntegerArray'
      array = 256 # set size to 256

      # code to initialize array
      .return (array)
Execute immediately after being compiled, but only if the subroutine is in the initial file (i.e. not in PIR compiled as result of a load_bytecode instruction from another file).As an example, suppose file main.pir contains:
    .sub main
        load_bytecode 'foo.pir'
and the file foo.pir contains:
    .sub foo :immediate
        $I0 = 4

    .sub bar :postcomp
        $I0 = 3
Executing foo.pir will run both foo and bar. On the other hand, executing main.pir will run only foo. If foo.pir is compiled to bytecode, only foo will be run, and loading foo.pbc will not run either foo or bar.
  .sub bar :method
    # ...

  .sub bar :method('foo')
    # ...
The marked .sub is a method, added as a method in the class that corresponds to the current namespace, and not stored in the namespace. In the method body, the object PMC can be referred to with self.If a string argument is given to :method the method is stored with that name instead of the .sub name.
  .sub bar :vtable
    # ...

  .sub bar :vtable('foo')
    # ...
The marked .sub overrides a vtable function, and is not stored in the namespace. By default, it overrides a vtable function with the same name as the .sub name. To override a different vtable function, use :vtable('...'). For example, to have a .sub named ToString also be the vtable function get_string), use :vtable('get_string').When the :vtable modifier is set, the object PMC can be referred to with self, as with the :method modifier.
The marked .sub is lexically nested within the sub known by subname.
:subid( <string_constant> )
Specifies a unique string identifier for the subroutine. This is useful for referring to a particular subroutine with :outer, even though several subroutines in the file may have the same name (because they are multi, or in different namespaces).
:instanceof( <string_constant> )
The :instanceof pragma is an experimental pragma that creates a sub as a PMC type other than 'Sub'. However, as currently implemented it doesn't work well with :outer or existing PMC types such as Closure, Coroutine, etc.
:nsentry( <string_constant> )
Specify the name by which the subroutine is stored in the namespace. The default name by which a subroutine is stored in the namespace (if this modifier is missing), is the subroutine's name as given after the .sub directive. This modifier allows to override this.

Directives used for Parrot calling conventions.

.begin_call and .end_call
Directives to start and end a subroutine invocation, respectively.
.begin_return and .end_return
Directives to start and end a statement to return values.
.begin_yield and .end_yield
Directives to start and end a statement to yield values.
Takes either 2 arguments: the sub and the return continuation, or the sub only. For the latter case an invokecc gets emitted. Providing an explicit return continuation is more efficient, if its created outside of a loop and the call is done inside a loop.
Directive to specify the object for a method call. Use it in combination with .call to call methods.
.set_return <var> [:<modifier>]*
Between .begin_return and .end_return, specify one or more of the return value(s) of the current subroutine. Available modifiers: :flat, :named.
.set_yield <var> [:<modifier>]*
Between .begin_yield and .end_yield, specify one or more of the yield value(s) of the current subroutine. Available modifiers: :flat, :named.
.set_arg <var> [:<modifier>]*
Between .begin_call and .call, specify an argument to be passed. Available modifiers: :flat, :named.
.get_result <var> [:<modifier>]*
Between .call and .end_call, specify where one or more return value(s) should be stored. Available modifiers: :slurpy, :named, :optional, and :opt_flag.

Directives for subroutine parameters

.param <type> <identifier> [:<modifier>]*
At the top of a subroutine, declare a local variable, in the manner of .local, into which parameter(s) of the current subroutine should be stored. Available modifiers: :slurpy, :named, :optional, :opt_flag.

Parameter Passing and Getting Flags

See PDD03 for a description of the meaning of the flag bits SLURPY, OPTIONAL, OPT_FLAG, and FLAT, which correspond to the calling convention modifiers :slurpy, :optional, :opt_flag, and :flat.

Catching Exceptions

Using the push_eh op you can install an exception handler. If an exception is thrown, Parrot will execute the installed exception handler. In order to retrieve the thrown exception, use the .get_results directive. This directive always takes one argument: an exception object.

   push_eh handler
   # ...
   .local pmc exception
   .get_results (exception)
   # ...

This is syntactic sugar for the get_results op, but any modifiers set on the targets will be handled automatically by the PIR compiler. The .get_results directive must be the first instruction of the exception handler; only declarations (.lex, .local) may come first.

To resume execution after handling the exception, just invoke the continuation stored in the exception.

   .local pmc exception, continuation
   # ...
   # ...
   continuation = exception['resume']
   # ...

See PDD23 for accessing the various attributes of the exception object.

Syntactic Sugar

Any PASM opcode is a valid PIR instruction. In addition, PIR defines some syntactic shortcuts. These are provided for ease of use by humans producing and maintaining PIR code.

goto <identifier>
branch to identifier (label or subroutine name).Examples:
  goto END
if <var> goto <identifier>
If var evaluates as true, jump to the named identifier.
unless <var> goto <identifier>
Unless var evaluates as true, jump to the named identifier.
if null <var> goto <identifier>
If var evaluates as null, jump to the named identifier.
unless null <var> goto <identifier>
Unless var evaluates as null, jump to the named identifier.
if <var1> <relop> <var2> goto <identifier>
The relop can be: <, <=, ==, != >= >. which translate to the PASM opcodes lt, le, eq, ne, ge or gt. If var1 relop var2 evaluates as true, jump to the named identifier.
unless <var1> <relop> <var2> goto <identifier>
The relop can be: <, <=, ==, != >= >. Unless var1 relop var2 evaluates as true, jump to the named identifier.
<var1> = <var2>
Assign a value.
<var1> = <unary> <var2>
Unary operations ! (NOT), - (negation) and ~ (bitwise NOT).
<var1> = <var2> <binary> <var3>
Binary arithmetic operations + (addition), - (subtraction), * (multiplication), / (division), % (modulus) and ** (exponent). Binary . is concatenation and only valid for string arguments.<< and >> are arithmetic shifts left and right. >>> is the logical shift right.Binary logic operations && (AND), || (OR) and ~~ (XOR).Binary bitwise operations & (bitwise AND), | (bitwise OR) and ~ (bitwise XOR).Binary relational operations <, <=, ==, != >= >.
<var1> <op>= <var2>
This is equivalent to <var1> = <var1> <op> <var2>. Where op is called an assignment operator and can be any of the following binary operators described earlier: +, -, *, /, %, ., &, |, ~, <<, >> or >>>.
<var> = <var> [ <var> ]
A keyed set operation for PMCs to retrieve a value from an aggregate. This maps to:
  set <var>, <var> [ <var> ]
<var> [ <var> ] = <var>
A keyed set operation to set a value in an aggregate. This maps to:
  set <var> [ <var> ], <var>
<var> = <opcode> <arguments>
Many opcodes can use this PIR syntactic sugar. The first argument for the opcode is placed before the =, and all remaining arguments go after the opcode name. For example:
  new $P0, 'Type'
  $P0 = new 'Type'
Note that this only works for opcodes that have a leading OUT parameter. [this restriction unimplemented: TT #906]
([<var1> [:<mod1> ...], ...]) = <var2>([<arg1> [:<mod2> ...], ...])
This is short for:
  .set_arg <arg1> <modifier2>
  .call <var2>
  .get_result <var1> <modifier1>
<var> = <var>([arg [:<modifier> ...], ...])
<var>([arg [:<modifier> ...], ...])
<var>."_method"([arg [:<modifier> ...], ...])
<var>.<var>([arg [:<modifier> ...], ...])
Function or method call. These notations are shorthand for a longer PCC function call. var can denote a global subroutine, a local identifier or a reg.
.return ([<var> [:<modifier> ...], ...])
Return from the current subroutine with zero or more values.
.tailcall <var>(args)
.tailcall <var>.'somemethod'(args)
.tailcall <var>.<var>(args)
Tail call: call a function or method and return from the sub with the function or method call return values.Internally, the call stack doesn't increase because of a tail call, so you can write recursive functions and not have stack overflows.Whitespace surrounding the dot ('.') that separates the object from the method is not allowed.

Assignment and Morphing

The = syntactic sugar in PIR, when used in the simple case of:

  <var1> = <var2>

directly corresponds to the set opcode. So, two low-level arguments (int, num, or string registers, variables, or constants) are a direct C assignment, or a C-level conversion (int cast, float cast, a string copy, or a call to one of the conversion functions like string_to_num).

Assigning a PMC argument to a low-level argument calls the get_integer, get_number, or get_string vtable function on the PMC. Assigning a low-level argument to a PMC argument calls the set_integer_native, set_number_native, or set_string_native vtable function on the PMC (assign to value semantics). Two PMC arguments are a direct C assignment (assign to container semantics).

For assign to value semantics for two PMC arguments use assign, which calls the assign_pmc vtable function.


This section describes the macro layer of the PIR language. The macro layer of the PIR compiler handles the following directives:

The macro layer is completely implemented in the lexical analysis phase. The parser does not know anything about what happens in the lexical analysis phase.

When the .include directive is encountered, the specified file is opened and the following tokens that are requested by the parser are read from that file.

A macro expansion is a dot-prefixed identifier. For instance, if a macro was defined as shown below:

 .macro foo(bar)
   # ...

this macro can be expanded by writing .foo(42). The body of the macro will be inserted at the point where the macro expansion is written.

A .macro_const expansion is more or less the same as a .macro expansion, except that a constant expansion cannot take any arguments, and the substitution of a .macro_const contains no newlines, so it can be used within a line of code.

Macro parameter list

The parameter list for a macro is specified in parentheses after the name of the macro. Macro parameters are not typed.

 .macro foo(bar, baz, buz)
   # ...

The number of arguments in the call to a macro must match the number of parameters in the macro's parameter list. Macros do not perform multidispatch, so you can't have two macros with the same name but different parameters. Calling a macro with the wrong number of arguments gives the user an error.

If a macro defines no parameter list, parentheses are optional on both the definition and the call. This means that a macro defined as:

 .macro foo
   # ...

can be expanded by writing either .foo or .foo(). And a macro definition written as:

 .macro foo()
   # ...

can also be expanded by writing either .foo or .foo().

Note: IMCC requires you to write parentheses if the macro was declared with (empty) parentheses. Likewise, when no parentheses were written (implying an empty parameter list), no parentheses may be used in the expansion.

Unique local labels

Within the macro body, the user can declare a unique label identifier using the value of a macro parameter, like so:

  .macro foo(a)
    # ...
 .label $a:
    # ...

Unique local variables (not yet implemented)

Note: This is not yet implemented in IMCC.

Within the macro body, the user can declare a local variable with a unique name.

  .macro foo()
    # ...
  .macro_local int b
    # ...
  .b = 42
  print .b # prints the value of the unique variable (42)
    # ...

The .macro_local directive declares a local variable with a unique name in the macro. When the macro .foo() is called, the resulting code that is given to the parser will read as follows:

  .sub main
    .local int local__foo__b__2
      # ...
    local__foo__b__2 = 42
    print local__foo__b__2


The user can also declare a local variable with a unique name set to the symbolic value of one of the macro parameters.

  .macro foo(b)
    # ...
  .macro_local int $b
    # ...
  .$b = 42
  print .$b # prints the value of the unique variable (42)
  print .b  # prints the value of parameter "b", which is
            # also the name of the variable.
  #  ...

So, the special $ character indicates whether the symbol is interpreted as just the value of the parameter, or that the variable by that name is meant. Obviously, the value of b should be a string.

The automatic name munging on .macro_local variables allows for using multiple macros, like so:

  .macro foo(a)
  .macro_local int $a

  .macro bar(b)
  .macro_local int $b

  .sub main

This will result in code for the parser as follows:

  .sub main
    .local int local__foo__x__2
    .local int local__bar__x__4

Each expansion is associated with a unique number; for labels declared with .macro_label and locals declared with .macro_local expansions, this means that multiple expansions of a macro will not result in conflicting label or local names.

Ordinary local variables

Defining a non-unique variable can still be done, using the normal syntax:

  .macro foo(b)
  .local int b
  .macro_local int $b

When invoking the macro foo as follows:

  .macro foo(b)


there will be two variables: b and x. When the macro is invoked twice:

  .sub main

the resulting code that is given to the parser will read as follows:

  .sub main
    .local int b
    .local int local__foo__x
    .local int b
    .local int local__foo__y

Obviously, this will result in an error, as the variable b is defined twice. If you intend the macro to create unique variables names, use .macro_local instead of .local to take advantage of the name munging.


Subroutine Definition

A simple subroutine, marked with :main, indicating it's the entry point in the file. Other sub modifiers include :load, :init, etc.

    .sub sub_label :main
      .param int a
      .param int b
      .param int c

      # ...
      .local pmc xy

Subroutine Call

Invocation of a subroutine. In this case a continuation subroutine is created.

    .const "Sub" $P0 = "sub_label"
    $P1 = new 'Continuation'
    set_addr $P1, ret_addr
    # ...
    .local int x
    .local num y
    .local string z
      .set_arg x
      .set_arg y
      .set_arg z
      .call $P0, $P1    # r = _sub_label(x, y, z)
      .local int r      # optional - new result var
      .get_result r

Subroutine Call Syntactic Sugar

Below there are three different ways to invoke the subroutine sub_label. The first retrieves a single return value, the second retrieves 3 return values, whereas the last does not save any return values.

  .local int r0, r1, r2
  r0 = sub_label($I0, $I1, $I2)
  (r0, r1, r2) = sub_label($I0, $I1, $I2)
  sub_label($I0, $I1, $I2)

This also works for NCI calls, as the subroutine PMC will be a NCI sub, and on invocation will do the Right Thing.

Instead of the label a subroutine object can be used too:


  .namespace [ "Foo" ]

  .sub _sub_label :method [,Subpragma, ...]
    .param int a
    .param int b
    .param int c
    # ...
    # ...
    .set_return xy

The variable "self" automatically refers to the invocating object, if the subroutine declaration contains "method".

Calling Methods

The syntax is very similar to subroutine calls. The call is done with .call immediately preceded by .invocant:

The return continuation is optional. The method can be a string constant or a string variable.

Returning and Yielding

  .return ( a, b )      # return the values of a and b

  .return ()            # return no value

  .tailcall func_call()   # tail call function

  .tailcall o."meth"()    # tail method call

Similarly, one can yield using the .yield directive

  .yield ( a, b )      # yield with the values of a and b

  .yield ()            # yield with no value


There are multiple implementations of PIR, each of which will meet this specification for the syntax. Currently there are the following implementations: