path: root/doc/guile-vm.texi

\input texinfo  @c -*-texinfo-*-
@c %**start of header
@setfilename guile-vm.info
@settitle Guile VM Specification
@footnotestyle end
@setchapternewpage odd
@c %**end of header

@set EDITION 0.6
@set VERSION 0.6
@set UPDATED 2005-04-26

@c Macro for instruction definitions.
@macro insn{}
Instruction
@end macro

@c For Scheme procedure definitions.
@macro scmproc{}
Scheme Procedure
@end macro

@c Scheme records.
@macro scmrec{}
Record
@end macro

@ifinfo
@dircategory Scheme Programming
@direntry
* Guile VM: (guile-vm).         Guile's Virtual Machine.
@end direntry

This file documents Guile VM.

Copyright @copyright{} 2000 Keisuke Nishida
Copyright @copyright{} 2005 Ludovic Court`es

Permission is granted to make and distribute verbatim copies of this
manual provided the copyright notice and this permission notice are
preserved on all copies.

@ignore
Permission is granted to process this file through TeX and print the
results, provided the printed document carries a copying permission
notice identical to this one except for the removal of this paragraph
(this paragraph not being relevant to the printed manual).

@end ignore
Permission is granted to copy and distribute modified versions of this
manual under the conditions for verbatim copying, provided that the
entire resulting derived work is distributed under the terms of a
permission notice identical to this one.

Permission is granted to copy and distribute translations of this manual
into another language, under the above conditions for modified versions,
except that this permission notice may be stated in a translation
approved by the Free Software Foundation.
@end ifinfo

@titlepage
@title Guile VM Specification
@subtitle for Guile VM @value{VERSION}
@author Keisuke Nishida

@page
@vskip 0pt plus 1filll
Edition @value{EDITION} @*
Updated for Guile VM @value{VERSION} @*
@value{UPDATED} @*

Copyright @copyright{} 2000 Keisuke Nishida
Copyright @copyright{} 2005 Ludovic Court`es

Permission is granted to make and distribute verbatim copies of this
manual provided the copyright notice and this permission notice are
preserved on all copies.

Permission is granted to copy and distribute modified versions of this
manual under the conditions for verbatim copying, provided that the
entire resulting derived work is distributed under the terms of a
permission notice identical to this one.

Permission is granted to copy and distribute translations of this manual
into another language, under the above conditions for modified versions,
except that this permission notice may be stated in a translation
approved by the Free Software Foundation.
@end titlepage

@contents

@c *********************************************************************
@node Top, Introduction, (dir), (dir)
@top Guile VM Specification

This document would like to correspond to Guile VM @value{VERSION}.
However, be warned that important parts still correspond to version
0.0 and are not valid anymore.

@menu
* Introduction::                
* Variable Management::         
* Instruction Set::             
* The Compiler::                
* Concept Index::               
* Function and Instruction Index::  
* Command and Variable Index::  

@detailmenu
 --- The Detailed Node Listing ---

Instruction Set

* Environment Control Instructions::  
* Branch Instructions::         
* Subprogram Control Instructions::  
* Data Control Instructions::   

The Compiler

* Overview::                    
* The Language Front-Ends::     
* GHIL::                        
* Compiling Scheme Code::       
* GLIL::                        
* The Assembler::               

@end detailmenu
@end menu

@c *********************************************************************
@node Introduction, Variable Management, Top, Top
@chapter What is Guile VM?

A Guile VM has a set of registers and its own stack memory.  Guile may
have more than one VM's.  Each VM may execute at most one program at a
time.  Guile VM is a CISC system so designed as to execute Scheme and
other languages efficiently.

@unnumberedsubsec Registers

@itemize
@item pc - Program counter    ;; ip (instruction poiner) is better?
@item sp - Stack pointer
@item bp - Base pointer
@item ac - Accumulator
@end itemize

@unnumberedsubsec Engine

A VM may have one of three engines: reckless, regular, or debugging.
Reckless engine is fastest but dangerous.  Regular engine is normally
fail-safe and reasonably fast.  Debugging engine is safest and
functional but very slow.

@unnumberedsubsec Memory

Stack is the only memory that each VM owns.  The other memory is shared
memory that is shared among every VM and other part of Guile.

@unnumberedsubsec Program

A VM program consists of a bytecode that is executed and an environment
in which execution is done.  Each program is allocated in the shared
memory and may be executed by any VM.  A program may call other programs
within a VM.

@unnumberedsubsec Instruction

Guile VM has dozens of system instructions and (possibly) hundreds of
functional instructions.  Some Scheme procedures such as cons and car
are implemented as VM's builtin functions, which are very efficient.
Other procedures defined outside of the VM are also considered as VM's
functional features, since they do not change the state of VM.
Procedures defined within the VM are called subprograms.

Most instructions deal with the accumulator (ac).  The VM stores all
results from functions in ac, instead of pushing them into the stack.
I'm not sure whether this is a good thing or not.

@node Variable Management, Instruction Set, Introduction, Top
@chapter Variable Management

FIXME:  This chapter needs to be reviewed so that it matches reality.
A more up-to-date description of the mechanisms described in this
section is given in @ref{Instruction Set}.

A program may have access to local variables, external variables, and
top-level variables.

@section Local/external variables

A stack is logically divided into several blocks during execution.  A
"block" is such a unit that maintains local variables and dynamic chain.
A "frame" is an upper level unit that maintains subprogram calls.

@example
             Stack
  dynamic |          |  |        |
    chain +==========+  -        =
        | |local vars|  |        |
        `-|block data|  | block  |
         /|frame data|  |        |
        | +----------+  -        |
        | |local vars|  |        | frame
        `-|block data|  |        |
         /+----------+  -        |
        | |local vars|  |        |
        `-|block data|  |        |
         /+==========+  -        =
        | |local vars|  |        |
        `-|block data|  |        |
         /|frame data|  |        |
        | +----------+  -        |
        | |          |  |        |
@end example

The first block of each frame may look like this:

@example
       Address  Data
       -------  ----
       xxx0028  Local variable 2
       xxx0024  Local variable 1
  bp ->xxx0020  Local variable 0
       xxx001c  Local link       (block data)
       xxx0018  External link    (block data)
       xxx0014  Stack pointer    (block data)
       xxx0010  Return address   (frame data)
       xxx000c  Parent program   (frame data)
@end example

The base pointer (bp) always points to the lowest address of local
variables of the recent block.  Local variables are referred as "bp[n]".
The local link field has a pointer to the dynamic parent of the block.
The parent's variables are referred as "bp[-1][n]", and grandparent's
are "bp[-1][-1][n]".  Thus, any local variable is represented by its
depth and offset from the current bp.

A variable may be "external", which is allocated in the shared memory.
The external link field of a block has a pointer to such a variable set,
which I call "fragment" (what should I call?).  A fragment has a set of
variables and its own chain.

@example
    local                    external
    chain|     |              chain
       | +-----+     .--------, |
       `-|block|--+->|external|-'
        /+-----+  |  `--------'\,
       `-|block|--'             |
        /+-----+     .--------, |
       `-|block|---->|external|-'
         +-----+     `--------'
         |     |
@end example

An external variable is referred as "bp[-2]->variables[n]" or
"bp[-2]->link->...->variables[n]".  This is also represented by a pair
of depth and offset.  At any point of execution, the value of bp
determines the current local link and external link, and thus the
current environment of a program.

Other data fields are described later.

@section Top-level variables

Guile VM uses the same top-level variables as the regular Guile.  A
program may have direct access to vcells.  Currently this is done by
calling scm_intern0, but a program is possible to have any top-level
environment defined by the current module.

@section Scheme and VM variable

Let's think about the following Scheme code as an example:

@example
  (define (foo a)
    (lambda (b) (list foo a b)))
@end example

In the lambda expression, "foo" is a top-level variable, "a" is an
external variable, and "b" is a local variable.

When a VM executes foo, it allocates a block for "a".  Since "a" may be
externally referred from the closure, the VM creates a fragment with a
copy of "a" in it.  When the VM evaluates the lambda expression, it
creates a subprogram (closure), associating the fragment with the
subprogram as its external environment.  When the closure is executed,
its environment will look like this:

@example
      block          Top-level: foo
  +-------------+
  |local var: b |       fragment
  +-------------+     .-----------,
  |external link|---->|variable: a|
  +-------------+     `-----------'
@end example

The fragment remains as long as the closure exists.

@section Addressing mode

Guile VM has five addressing modes:

@itemize
@item Real address
@item Local position
@item External position
@item Top-level location
@item Constant object
@end itemize

Real address points to the address in the real program and is only used
with the program counter (pc).

Local position and external position are represented as a pair of depth
and offset from bp, as described above.  These are base relative
addresses, and the real address may vary during execution.

Top-level location is represented as a Guile's vcell.  This location is
determined at loading time, so the use of this address is efficient.

Constant object is not an address but gives an instruction an Scheme
object directly.

[ We'll also need dynamic scope addressing to support Emacs Lisp? ]


Overall procedure:

@enumerate
@item A source program is compiled into a bytecode.
@item A bytecode is given an environment and becomes a program.
@item A VM starts execution, creating a frame for it.
@item Whenever a program calls a subprogram, a new frame is created for it.
@item When a program finishes execution, it returns a value, and the VM
      continues execution of the parent program.
@item When all programs terminated, the VM returns the final value and stops.
@end enumerate


@node Instruction Set, The Compiler, Variable Management, Top
@chapter Instruction Set

The Guile VM instruction set is roughly divided two groups: system
instructions and functional instructions.  System instructions control
the execution of programs, while functional instructions provide many
useful calculations.

@menu
* Environment Control Instructions::  
* Branch Instructions::         
* Subprogram Control Instructions::  
* Data Control Instructions::   
@end menu

@node Environment Control Instructions, Branch Instructions, Instruction Set, Instruction Set
@section Environment Control Instructions

@deffn @insn{} link binding-name
Look up @var{binding-name} (a string) in the current environment and
push the corresponding variable object onto the stack.  If
@var{binding-name} is not bound yet, then create a new binding and
push its variable object.
@end deffn

@deffn @insn{} variable-ref
Dereference the variable object which is on top of the stack and
replace it by the value of the variable it represents.
@end deffn

@deffn @insn{} variable-set
Set the value of the variable on top of the stack (at @code{sp[0]}) to
the object located immediately before (at @code{sp[-1]}).
@end deffn

As an example, let us look at what a simple function call looks like:

@example
(+ 2 3)
@end example

This call yields the following sequence of instructions:

@example
(link "+")      ;; lookup binding "+"
(variable-ref)  ;; dereference it
(make-int8 2)   ;; push immediate value `2'
(make-int8 3)   ;; push immediate value `3'
(tail-call 2)   ;; call the proc at sp[-3] with two args
@end example

@deffn @insn{} local-ref offset
Push onto the stack the value of the local variable located at
@var{offset} within the current stack frame.
@end deffn

@deffn @insn{} local-set offset
Pop the Scheme object located on top of the stack and make it the new
value of the local variable located at @var{offset} within the current
stack frame.
@end deffn

@deffn @insn{} external-ref offset
Push the value of the closure variable located at position
@var{offset} within the program's list of external variables.
@end deffn

@deffn @insn{} external-set offset
Pop the Scheme object located on top of the stack and make it the new
value of the closure variable located at @var{offset} within the
program's list of external variables.
@end deffn

@deffn @insn{} make-closure
Pop the program object from the stack and assign it the current
closure variable list as its closure.  Push the result program
object.
@end deffn

Let's illustrate this:

@example
(let ((x 2))
  (lambda ()
    (let ((x++ (+ 1 x)))
      (set! x x++)
      x++)))
@end example

The resulting program has one external (closure) variable, i.e. its
@var{nexts} is set to 1 (@pxref{Subprogram Control Instructions}).
This yields the following code:

@example
   ;; the traditional program prologue with NLOCS = 0 and NEXTS = 1

   0    (make-int8 2)
   2    (external-set 0)
   4    (make-int8 4)
   6    (link "+")     ;; lookup `+'
   9    (vector 1)     ;; create the external variable vector for
                       ;; later use by `object-ref' and `object-set'
        ...
  40    (load-program ##34#)
  59    (make-closure) ;; assign the current closure to the program
                       ;; just pushed by `load-program'
  60    (return)
@end example

The program loaded here by @var{load-program} contains the following
sequence of instructions:

@example
   0    (object-ref 0)     ;; push the variable for `+'
   2    (variable-ref)     ;; dereference `+'
   3    (make-int8:1)      ;; push 1
   4    (external-ref 0)   ;; push the value of `x'
   6    (call 2)           ;; call `+' and push the result
   8    (local-set 0)      ;; make it the new value of `x++'
  10    (local-ref 0)      ;; push the value of `x++'
  12    (external-set 0)   ;; make it the new value of `x'
  14    (local-ref 0)      ;; push the value of `x++'
  16    (return)           ;; return it
@end example

At this point, you should know pretty much everything about the three
types of variables a program may need to access.


@node Branch Instructions, Subprogram Control Instructions, Environment Control Instructions, Instruction Set
@section Branch Instructions

All the conditional branch instructions described below work in the
same way:

@itemize
@item They take the Scheme object located on the stack and use it as
the branch condition;
@item If the condition if false, then program execution continues with
the next instruction;
@item If the condition is true, then the instruction pointer is
increased by the offset passed as an argument to the branch
instruction;
@item Finally, when the instruction finished, the condition object is
removed from the stack.
@end itemize

Note that the offset passed to the instruction is encoded on two 8-bit
integers which are then combined by the VM as one 16-bit integer.

@deffn @insn{} br offset
Jump to @var{offset}.
@end deffn

@deffn @insn{} br-if offset
Jump to @var{offset} if the condition on the stack is not false.
@end deffn

@deffn @insn{} br-if-not offset
Jump to @var{offset} if the condition on the stack is false.
@end deffn

@deffn @insn{} br-if-eq offset
Jump to @var{offset} if the two objects located on the stack are
equal in the sense of @var{eq?}.  Note that, for this instruction, the
stack pointer is decremented by two Scheme objects instead of only
one.
@end deffn

@deffn @insn{} br-if-not-eq offset
Same as @var{br-if-eq} for non-equal objects.
@end deffn

@deffn @insn{} br-if-null offset
Jump to @var{offset} if the object on the stack is @code{'()}.
@end deffn

@deffn @insn{} br-if-not-null offset
Jump to @var{offset} if the object on the stack is not @code{'()}.
@end deffn


@node Subprogram Control Instructions, Data Control Instructions, Branch Instructions, Instruction Set
@section Subprogram Control Instructions

Programs (read: ``compiled procedure'') may refer to external
bindings, like variables or functions defined outside the program
itself, in the environment in which it will evaluate at run-time.  In
a sense, a program's environment and its bindings are an implicit
parameter of every program.

@cindex object table
In order to handle such bindings, each program has an @dfn{object
table} associated to it.  This table (actually a Scheme vector)
contains all constant objects referenced by the program.  The object
table of a program is initialized right before a program is loaded
with @var{load-program}.

Variable objects are one such type of constant object: when a global
binding is defined, a variable object is associated to it and that
object will remain constant over time, even if the value bound to it
changes.  Therefore, external bindings only need to be looked up once
when the program is loaded.  References to the corresponding external
variables from within the program are then performed via the
@var{object-ref} instruction and are almost as fast as local variable
references.

Let us consider the following program (procedure) which references
external bindings @code{frob} and @var{%magic}:

@example
(lambda (x)
  (frob x %magic))
@end example

This yields the following assembly code:

@example
(make-int8 64)   ;; number of args, vars, etc. (see below)
(link "frob")
(link "%magic")
(vector 2)       ;; object table (external bindings)
...
(load-program #u8(20 0 23 21 0 20 1 23 36 2))
(return)
@end example

All the instructions occurring before @var{load-program} (some were
omitted for simplicity) form a @dfn{prologue} which, among other
things, pushed an object table (a vector) that contains the variable
objects for the variables bound to @var{frob} and @var{%magic}.  This
vector and other data pushed onto the stack are then popped by the
@var{load-program} instruction.

Besides, the @var{load-program} instruction takes one explicit
argument which is the bytecode of the program itself.  Disassembled,
this bytecode looks like:

@example
(object-ref 0)  ;; push the variable object of `frob'
(variable-ref)  ;; dereference it
(local-ref 0)   ;; push the value of `x'
(object-ref 1)  ;; push the variable object of `%magic'
(variable-ref)  ;; dereference it
(tail-call 2)   ;; call `frob' with two parameters
@end example

This clearly shows that there is little difference between references
to local variables and references to externally bound variables since
lookup of externally bound variables if performed only once before the
program is run.

@deffn @insn{} load-program bytecode
Load the program whose bytecode is @var{bytecode} (a u8vector), pop
its meta-information from the stack, and push a corresponding program
object onto the stack.  The program's meta-information may consist of
(in the order in which it should be pushed onto the stack):

@itemize
@item optionally, a pair representing meta-data (see the
@var{program-meta} procedure); [FIXME: explain their meaning]
@item optionally, a vector which is the program's object table (a
program that does not reference external bindings does not need an
object table);
@item either one immediate integer or four immediate integers
representing respectively the number of arguments taken by the
function (@var{nargs}), the number of @dfn{rest arguments}
(@var{nrest}, 0 or 1), the number of local variables (@var{nlocs}) and
the number of external variables (@var{nexts}) (@pxref{Environment
Control Instructions}).
@end itemize

@end deffn

@deffn @insn{} object-ref offset
Push the variable object for the external variable located at
@var{offset} within the program's object table.
@end deffn

@deffn @insn{} return
Free the program's frame.
@end deffn

@deffn @insn{} call nargs
Call the procedure, continuation or program located at
@code{sp[-nargs]} with the @var{nargs} arguments located from
@code{sp[0]} to @code{sp[-nargs + 1]}.  The
procedure/continuation/program and its arguments are dropped from the
stack and the result is pushed.  When calling a program, the
@code{call} instruction reserves room for its local variables on the
stack, and initializes its list of closure variables and its vector of
externally bound variables.
@end deffn

@deffn @insn{} tail-call nargs
Same as @code{call} except that, for tail-recursive calls to a
program, the current stack frame is re-used, as required by RnRS.
This instruction is otherwise similar to @code{call}.
@end deffn


@node Data Control Instructions,  , Subprogram Control Instructions, Instruction Set
@section Data Control Instructions

@deffn @insn{} make-int8 value
Push @var{value}, an 8-bit integer, onto the stack.
@end deffn

@deffn @insn{} make-int8:0
Push the immediate value @code{0} onto the stack.
@end deffn

@deffn @insn{} make-int8:1
Push the immediate value @code{1} onto the stack.
@end deffn

@deffn @insn{} make-false
Push @code{#f} onto the stack.
@end deffn

@deffn @insn{} make-true
Push @code{#t} onto the stack.
@end deffn

@itemize
@item %push
@item %pushi
@item %pushl, %pushl:0:0, %pushl:0:1, %pushl:0:2, %pushl:0:3
@item %pushe, %pushe:0:0, %pushe:0:1, %pushe:0:2, %pushe:0:3
@item %pusht
@end itemize

@itemize
@item %loadi
@item %loadl, %loadl:0:0, %loadl:0:1, %loadl:0:2, %loadl:0:3
@item %loade, %loade:0:0, %loade:0:1, %loade:0:2, %loade:0:3
@item %loadt
@end itemize

@itemize
@item %savei
@item %savel, %savel:0:0, %savel:0:1, %savel:0:2, %savel:0:3
@item %savee, %savee:0:0, %savee:0:1, %savee:0:2, %savee:0:3
@item %savet
@end itemize

@section Flow control instructions

@itemize
@item %br-if
@item %br-if-not
@item %jump
@end itemize

@section Function call instructions

@itemize
@item %func, %func0, %func1, %func2
@end itemize

@section Scheme built-in functions

@itemize
@item cons
@item car
@item cdr
@end itemize

@section Mathematical buitin functions

@itemize
@item 1+
@item 1-
@item add, add2
@item sub, sub2, minus
@item mul2
@item div2
@item lt2
@item gt2
@item le2
@item ge2
@item num-eq2
@end itemize



@node The Compiler, Concept Index, Instruction Set, Top
@chapter The Compiler

This section describes Guile-VM's compiler and the compilation process
to produce bytecode executable by the VM itself (@pxref{Instruction
Set}).

@menu
* Overview::                    
* The Language Front-Ends::     
* GHIL::                        
* Compiling Scheme Code::       
* GLIL::                        
* The Assembler::               
@end menu

@node Overview, The Language Front-Ends, The Compiler, The Compiler
@section Overview

Compilation in Guile-VM is a three-stage process:

@cindex intermediate language
@cindex assembler
@cindex compiler
@cindex GHIL
@cindex GLIL
@cindex bytecode

@enumerate
@item the source programming language (e.g. R5RS Scheme) is read and
translated into GHIL, @dfn{Guile's High-Level Intermediate Language};
@item GHIL code is then translated into a lower-level intermediate
language call GLIL, @dfn{Guile's Low-Level Intermediate Language};
@item finally, GLIL is @dfn{assembled} into the VM's assembly language
(@pxref{Instruction Set}) and bytecode.
@end enumerate

The use of two separate intermediate languages eases the
implementation of front-ends since the gap between high-level
languages like Scheme and GHIL is relatively small.

@vindex guilec
From an end-user viewpoint, compiling a Guile program into bytecode
can be done either by using the @command{guilec} command-line tool, or
by using the @code{compile-file} procedure exported by the
@code{(system base compile)} module.

@deffn @scmproc{} compile-file file . opts
Compile Scheme source code from file @var{file} using compilation
options @var{opts}.  The resulting file, a Guile object file, will be
name according the application of the @code{compiled-file-name}
procedure to @var{file}.  The possible values for @var{opts} are the
same as for the @code{compile-in} procedure (see below, @pxref{The Language
Front-Ends}).
@end deffn

@deffn @scmproc{} compiled-file-name file
Given source file name @var{file} (a string), return a string that
denotes the name of the Guile object file corresponding to
@var{file}.  By default, the file name returned is @var{file} minus
its extension and plus the @code{.go} file extension.
@end deffn

@cindex self-hosting
It is worth noting, as you might have already guessed, that Guile-VM's
compiler is written in Guile Scheme and is @dfn{self-hosted}: it can
compile itself.

@node The Language Front-Ends, GHIL, Overview, The Compiler
@section The Language Front-Ends

Guile-VM comes with a number of @dfn{language front-ends}, that is,
code that can read a given high-level programming language like R5RS
Scheme, and translate it into a lower-level representation suitable to
the compiler.

Each language front-end provides a @dfn{specification} and a
@dfn{translator} to GHIL.  Both of them come in the @code{language}
module hierarchy.  As an example, the front-end for Scheme is located
in the @code{(language scheme spec)} and @code{(language scheme
translate)} modules.  Language front-ends can then be retrieved using
the @code{lookup-language} procedure of the @code{(system base
language)} module.

@deftp @scmrec{} <language> name title version reader printer read-file expander translator evaluator environment
Denotes a language front-end specification a various methods used by
the compiler to handle source written in that language.  Of particular
interest is the @code{translator} slot (@pxref{GHIL}).
@end deftp

@deffn @scmproc{} lookup-language lang
Look for a language front-end named @var{lang}, a symbol (e.g,
@code{scheme}), and return the @code{<language>} record describing it
if found.  If @var{lang} does not denote a language front-end, an
error is raised.  Note that this procedure assumes that language
@var{lang} exists if there exist a @code{(language @var{lang} spec)}
module.
@end deffn

The @code{(system base compile)} module defines a procedure similar to
@code{compile-file} but that is not limited to the Scheme language:

@deffn @scmproc{} compile-in expr env lang . opts
Compile expression @var{expr}, which is written in language @var{lang}
(a @code{<language>} object), using compilation options @var{opts},
and return bytecode as produced by the assembler (@pxref{The
Assembler}).

Options @var{opts} may contain the following keywords:

@table @code
@item :e
compilation will stop after the code expansion phase.
@item :t
compilation will stop after the code translation phase, i.e. after
code in the source language @var{lang} has been translated into GHIL
(@pxref{GHIL}).
@item :c
compilation will stop after the compilation phase and before the
assembly phase, i.e. once GHIL has been translated into GLIL
(@pxref{GLIL}).
@end table

Additionally, @var{opts} may contain any option understood by the
GHIL-to-GLIL compiler described in @xref{GLIL}.
@end deffn


@node GHIL, Compiling Scheme Code, The Language Front-Ends, The Compiler
@section Guile's High-Level Intermediate Language

GHIL has constructs almost equivalent to those found in Scheme.
However, unlike Scheme, it is meant to be read only by the compiler
itself.  Therefore, a sequence of GHIL code is only a sequence of GHIL
@emph{objects} (records), as opposed to symbols, each of which
represents a particular language feature.  These records are all
defined in the @code{(system il ghil)} module and are named
@code{<ghil-*>}.

Each GHIL record has at least two fields: one containing the
environment (Guile module) in which it is considered, and one
containing its location [FIXME: currently seems to be unused].  Below
is a list of the main GHIL object types and their fields:

@example
;; Objects
(<ghil-void> env loc)
(<ghil-quote> env loc obj)
(<ghil-quasiquote> env loc exp)
(<ghil-unquote> env loc exp)
(<ghil-unquote-splicing> env loc exp)
;; Variables
(<ghil-ref> env loc var)
(<ghil-set> env loc var val)
(<ghil-define> env loc var val)
;; Controls
(<ghil-if> env loc test then else)
(<ghil-and> env loc exps)
(<ghil-or> env loc exps)
(<ghil-begin> env loc exps)
(<ghil-bind> env loc vars vals body)
(<ghil-lambda> env loc vars rest body)
(<ghil-call> env loc proc args)
(<ghil-inline> env loc inline args)
@end example

As can be seen from this examples, the constructs in GHIL are pretty
close to the fundamental primitives of Scheme.

It is the role of front-end language translators (@pxref{The Language
Front-Ends}) to produce a sequence of GHIL objects from the
human-readable, source programming language.  The next section
describes the translator for the Scheme language.

@node Compiling Scheme Code, GLIL, GHIL, The Compiler
@section Compiling Scheme Code

The language object for Scheme, as returned by @code{(lookup-language
'scheme)} (@pxref{The Language Front-Ends}), defines a translator
procedure that returns a sequence of GHIL objects given Scheme code.
Before actually performing this operation, the Scheme translator
expands macros in the original source code.

The macros that may be expanded can come from different sources:

@itemize
@item core Guile macros, such as @code{false-if-exception};
@item macros defined in modules used by the module being compiled,
e.g., @code{receive} in @code{(ice-9 receive)};
@item macros defined within the module being compiled.
@end itemize

@cindex macro
@cindex syntax transformer
@findex define-macro
@findex defmacro
The main complexity in handling macros at compilation time is that
Guile's macros are first-class objects.  For instance, when using
@code{define-macro}, one actually defines a @emph{procedure} that
returns code; of course, unlike a ``regular'' procedure, it is
executed when an S-exp is @dfn{memoized} by the evaluator, i.e.,
before the actual evaluation takes place.  Worse, it is possible to
turn a procedure into a macro, or @dfn{syntax transformer}, thus
removing, to some extent, the boundary between the macro expansion and
evaluation phases, @inforef{Internal Macros, , guile}.

[FIXME: explain limitations, etc.]


@node GLIL, The Assembler, Compiling Scheme Code, The Compiler
@section Guile's Low-Level Intermediate Language

A GHIL instruction sequence can be compiled into GLIL using the
@code{compile} procedure exported by the @code{(system il compile)}
module.  During this translation process, various optimizations may
also be performed.

The module @code{(system il glil)} defines record types representing
various low-level abstractions.  Compared to GHIL, the flow control
primitives in GLIL are much more low-level:  only @code{<glil-label>},
@code{<glil-branch>} and @code{<glil-call>} are available, no
@code{lambda}, @code{if}, etc.


@deffn @scmproc{} compile ghil environment . opts
Compile @var{ghil}, a GHIL instruction sequence, within
environment/module @var{environment}, and return the resulting GLIL
instruction sequence.  The option list @var{opts} may be either the
empty list or a list containing the @code{:O} keyword in which case
@code{compile} will first go through an optimization stage of
@var{ghil}.

Note that the @code{:O} option may be passed at a higher-level to the
@code{compile-file} and @code{compile-in} procedures (@pxref{The
Language Front-Ends}).
@end deffn

@deffn @scmproc{} pprint-glil glil . port
Print @var{glil}, a GLIL sequence instructions, in a human-readable
form.  If @var{port} is passed, it will be used as the output port.
@end deffn


Let's consider the following Scheme expression:

@example
(lambda (x) (+ x 1))
@end example

The corresponding (unoptimized) GLIL code, as shown by
@code{pprint-glil}, looks like this:

@example
(@@asm (0 0 0 0)
  (@@asm (1 0 0 0)           ;; expect one arg.
    (@@bind (x argument 0))  ;; debugging info
    (module-ref #f +)       ;; lookup `+'
    (argument-ref 0)        ;; push the argument onto
                            ;; the stack
    (const 1)               ;; push `1'
    (tail-call 2)           ;; call `+', with 2 args,
                            ;; using the same stack frame
    (@@source 15 33))        ;; additional debugging info
  (return 0))
@end example

This is not unlike the VM's assembly language described in
@ref{Instruction Set}.

@node The Assembler,  , GLIL, The Compiler
@section The Assembler

@findex code->bytes

The final compilation step consists in converting the GLIL instruction
sequence into VM bytecode.  This is what the @code{assemble} procedure
defined in the @code{(system vm assemble)} module is for.  It relies
on the @code{code->bytes} procedure of the @code{(system vm conv)}
module to convert instructions (represented as lists whose @code{car}
is a symbol naming the instruction, e.g. @code{object-ref},
@pxref{Instruction Set}) into binary code, or @dfn{bytecode}.
Bytecode itself is represented using SRFI-4 byte vectors,
@inforef{SRFI-4, SRFI-4 homogeneous numeric vectors, guile}.


@deffn @scmproc{} assemble glil environment . opts
Return a binary representation of @var{glil} (bytecode), either in the
form of an SRFI-4 @code{u8vector} or a @code{<bytespec>} object.
[FIXME:  Why is that?]
@end deffn



@c *********************************************************************
@node Concept Index, Function and Instruction Index, The Compiler, Top
@unnumbered Concept Index
@printindex cp

@node Function and Instruction Index, Command and Variable Index, Concept Index, Top
@unnumbered Function and Instruction Index
@printindex fn

@node Command and Variable Index,  , Function and Instruction Index, Top
@unnumbered Command and Variable Index
@printindex vr

@bye

@c Local Variables:
@c ispell-local-dictionary: "american";
@c End:

@c  LocalWords:  bytecode