2001-04-09 Martin Grabmueller <mgrabmue@cs.tu-berlin.de>

[bpt/guile.git] / doc / data-rep.texi

@c essay \input texinfo
@c essay @c -*-texinfo-*-
@c essay @c %**start of header
@c essay @setfilename data-rep.info
@c essay @settitle Data Representation in Guile
@c essay @c %**end of header

@c essay @include version.texi

@c essay @dircategory The Algorithmic Language Scheme
@c essay @direntry
@c essay * data-rep: (data-rep).  Data Representation in Guile --- how to use
                         Guile objects in your C code.
@c essay @end direntry

@c essay @setchapternewpage off

@c essay @ifinfo
@c essay Data Representation in Guile

@c essay Copyright (C) 1998, 1999, 2000 Free Software Foundation

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

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

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

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

@c essay @titlepage
@c essay @sp 10
@c essay @comment The title is printed in a large font.
@c essay @title Data Representation in Guile
@c essay @subtitle $Id: data-rep.texi,v 1.18 2001-04-02 21:53:20 ossau Exp $
@c essay @subtitle For use with Guile @value{VERSION}
@c essay @author Jim Blandy
@c essay @author Free Software Foundation
@c essay @author @email{jimb@@red-bean.com}
@c essay @c The following two commands start the copyright page.
@c essay @page
@c essay @vskip 0pt plus 1filll
@c essay @vskip 0pt plus 1filll
@c essay Copyright @copyright{} 1998 Free Software Foundation

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

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

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

@c essay @c @smallbook
@c essay @c @finalout
@c essay @headings double


@c essay @node Top, Data Representation in Scheme, (dir), (dir)
@c essay @top Data Representation in Guile

@c essay @ifinfo
@c essay This essay is meant to provide the background necessary to read and
@c essay write C code that manipulates Scheme values in a way that conforms to
@c essay libguile's interface.  If you would like to write or maintain a
@c essay Guile-based application in C or C++, this is the first information you
@c essay need.

@c essay In order to make sense of Guile's @code{SCM_} functions, or read
@c essay libguile's source code, it's essential to have a good grasp of how Guile
@c essay actually represents Scheme values.  Otherwise, a lot of the code, and
@c essay the conventions it follows, won't make very much sense.

@c essay We assume you know both C and Scheme, but we do not assume you are
@c essay familiar with Guile's C interface.
@c essay @end ifinfo


@page
@node Data Representation
@chapter Data Representation in Guile

@strong{by Jim Blandy}

[Due to the rather non-orthogonal and performance-oriented nature of the
SCM interface, you need to understand SCM internals *before* you can use
the SCM API.  That's why this chapter comes first.]

[NOTE: this is Jim Blandy's essay almost entirely unmodified.  It has to
be adapted to fit this manual smoothly.]

In order to make sense of Guile's SCM_ functions, or read libguile's
source code, it's essential to have a good grasp of how Guile actually
represents Scheme values.  Otherwise, a lot of the code, and the
conventions it follows, won't make very much sense.  This essay is meant
to provide the background necessary to read and write C code that
manipulates Scheme values in a way that is compatible with libguile.

We assume you know both C and Scheme, but we do not assume you are
familiar with Guile's implementation.

@menu
* Data Representation in Scheme::       Why things aren't just totally
                                        straightforward, in general terms.
* How Guile does it::                   How to write C code that manipulates
                                        Guile values, with an explanation
                                        of Guile's garbage collector.
* Defining New Types (Smobs)::          How to extend Guile with your own
                                        application-specific datatypes.
@end menu

@node Data Representation in Scheme
@section Data Representation in Scheme

Scheme is a latently-typed language; this means that the system cannot,
in general, determine the type of a given expression at compile time.
Types only become apparent at run time.  Variables do not have fixed
types; a variable may hold a pair at one point, an integer at the next,
and a thousand-element vector later.  Instead, values, not variables,
have fixed types.

In order to implement standard Scheme functions like @code{pair?} and
@code{string?} and provide garbage collection, the representation of
every value must contain enough information to accurately determine its
type at run time.  Often, Scheme systems also use this information to
determine whether a program has attempted to apply an operation to an
inappropriately typed value (such as taking the @code{car} of a string).

Because variables, pairs, and vectors may hold values of any type,
Scheme implementations use a uniform representation for values --- a
single type large enough to hold either a complete value or a pointer
to a complete value, along with the necessary typing information.

The following sections will present a simple typing system, and then
make some refinements to correct its major weaknesses.  However, this is
not a description of the system Guile actually uses.  It is only an
illustration of the issues Guile's system must address.  We provide all
the information one needs to work with Guile's data in @ref{How Guile
does it}.


@menu
* A Simple Representation::     
* Faster Integers::             
* Cheaper Pairs::               
* Guile Is Hairier::            
@end menu

@node A Simple Representation
@subsection A Simple Representation

The simplest way to meet the above requirements in C would be to
represent each value as a pointer to a structure containing a type
indicator, followed by a union carrying the real value.  Assuming that
@code{SCM} is the name of our universal type, we can write:

@example
enum type @{ integer, pair, string, vector, ... @};

typedef struct value *SCM;

struct value @{
  enum type type;
  union @{
    int integer;
    struct @{ SCM car, cdr; @} pair;
    struct @{ int length; char *elts; @} string;
    struct @{ int length; SCM  *elts; @} vector;
    ...
  @} value;
@};
@end example
with the ellipses replaced with code for the remaining Scheme types.

This representation is sufficient to implement all of Scheme's
semantics.  If @var{x} is an @code{SCM} value:
@itemize @bullet
@item
  To test if @var{x} is an integer, we can write @code{@var{x}->type == integer}.
@item
  To find its value, we can write @code{@var{x}->value.integer}.
@item
  To test if @var{x} is a vector, we can write @code{@var{x}->type == vector}.
@item
  If we know @var{x} is a vector, we can write
  @code{@var{x}->value.vector.elts[0]} to refer to its first element.
@item
  If we know @var{x} is a pair, we can write
  @code{@var{x}->value.pair.car} to extract its car.
@end itemize


@node Faster Integers
@subsection Faster Integers

Unfortunately, the above representation has a serious disadvantage.  In
order to return an integer, an expression must allocate a @code{struct
value}, initialize it to represent that integer, and return a pointer to
it.  Furthermore, fetching an integer's value requires a memory
reference, which is much slower than a register reference on most
processors.  Since integers are extremely common, this representation is
too costly, in both time and space.  Integers should be very cheap to
create and manipulate.

One possible solution comes from the observation that, on many
architectures, structures must be aligned on a four-byte boundary.
(Whether or not the machine actually requires it, we can write our own
allocator for @code{struct value} objects that assures this is true.)
In this case, the lower two bits of the structure's address are known to
be zero.

This gives us the room we need to provide an improved representation
for integers.  We make the following rules:
@itemize @bullet
@item
If the lower two bits of an @code{SCM} value are zero, then the SCM
value is a pointer to a @code{struct value}, and everything proceeds as
before.
@item
Otherwise, the @code{SCM} value represents an integer, whose value
appears in its upper bits.
@end itemize

Here is C code implementing this convention:
@example
enum type @{ pair, string, vector, ... @};

typedef struct value *SCM;

struct value @{
  enum type type;
  union @{
    struct @{ SCM car, cdr; @} pair;
    struct @{ int length; char *elts; @} string;
    struct @{ int length; SCM  *elts; @} vector;
    ...
  @} value;
@};

#define POINTER_P(x) (((int) (x) & 3) == 0)
#define INTEGER_P(x) (! POINTER_P (x))

#define GET_INTEGER(x)  ((int) (x) >> 2)
#define MAKE_INTEGER(x) ((SCM) (((x) << 2) | 1))
@end example

Notice that @code{integer} no longer appears as an element of @code{enum
type}, and the union has lost its @code{integer} member.  Instead, we
use the @code{POINTER_P} and @code{INTEGER_P} macros to make a coarse
classification of values into integers and non-integers, and do further
type testing as before.

Here's how we would answer the questions posed above (again, assume
@var{x} is an @code{SCM} value):
@itemize @bullet
@item
  To test if @var{x} is an integer, we can write @code{INTEGER_P (@var{x})}.
@item
  To find its value, we can write @code{GET_INTEGER (@var{x})}.
@item
  To test if @var{x} is a vector, we can write:
@example
  @code{POINTER_P (@var{x}) && @var{x}->type == vector}
@end example
  Given the new representation, we must make sure @var{x} is truly a
  pointer before we dereference it to determine its complete type.
@item
  If we know @var{x} is a vector, we can write
  @code{@var{x}->value.vector.elts[0]} to refer to its first element, as
  before.
@item
  If we know @var{x} is a pair, we can write
  @code{@var{x}->value.pair.car} to extract its car, just as before.
@end itemize

This representation allows us to operate more efficiently on integers
than the first.  For example, if @var{x} and @var{y} are known to be
integers, we can compute their sum as follows:
@example
MAKE_INTEGER (GET_INTEGER (@var{x}) + GET_INTEGER (@var{y}))
@end example
Now, integer math requires no allocation or memory references.  Most
real Scheme systems actually use an even more efficient representation,
but this essay isn't about bit-twiddling.  (Hint: what if pointers had
@code{01} in their least significant bits, and integers had @code{00}?)


@node Cheaper Pairs
@subsection Cheaper Pairs

However, there is yet another issue to confront.  Most Scheme heaps
contain more pairs than any other type of object; Jonathan Rees says
that pairs occupy 45% of the heap in his Scheme implementation, Scheme
48.  However, our representation above spends three @code{SCM}-sized
words per pair --- one for the type, and two for the @sc{car} and
@sc{cdr}.  Is there any way to represent pairs using only two words?

Let us refine the convention we established earlier.  Let us assert
that:
@itemize @bullet
@item
  If the bottom two bits of an @code{SCM} value are @code{#b00}, then
  it is a pointer, as before.
@item
  If the bottom two bits are @code{#b01}, then the upper bits are an
  integer.  This is a bit more restrictive than before.
@item
  If the bottom two bits are @code{#b10}, then the value, with the bottom
  two bits masked out, is the address of a pair.
@end itemize

Here is the new C code:
@example
enum type @{ string, vector, ... @};

typedef struct value *SCM;

struct value @{
  enum type type;
  union @{
    struct @{ int length; char *elts; @} string;
    struct @{ int length; SCM  *elts; @} vector;
    ...
  @} value;
@};

struct pair @{
  SCM car, cdr;
@};

#define POINTER_P(x) (((int) (x) & 3) == 0)

#define INTEGER_P(x)  (((int) (x) & 3) == 1)
#define GET_INTEGER(x)  ((int) (x) >> 2)
#define MAKE_INTEGER(x) ((SCM) (((x) << 2) | 1))

#define PAIR_P(x) (((int) (x) & 3) == 2)
#define GET_PAIR(x) ((struct pair *) ((int) (x) & ~3))
@end example

Notice that @code{enum type} and @code{struct value} now only contain
provisions for vectors and strings; both integers and pairs have become
special cases.  The code above also assumes that an @code{int} is large
enough to hold a pointer, which isn't generally true.


Our list of examples is now as follows:
@itemize @bullet
@item
  To test if @var{x} is an integer, we can write @code{INTEGER_P
  (@var{x})}; this is as before.
@item
  To find its value, we can write @code{GET_INTEGER (@var{x})}, as
  before.
@item
  To test if @var{x} is a vector, we can write:
@example
  @code{POINTER_P (@var{x}) && @var{x}->type == vector}
@end example
  We must still make sure that @var{x} is a pointer to a @code{struct
  value} before dereferencing it to find its type.
@item
  If we know @var{x} is a vector, we can write
  @code{@var{x}->value.vector.elts[0]} to refer to its first element, as
  before.
@item
  We can write @code{PAIR_P (@var{x})} to determine if @var{x} is a
  pair, and then write @code{GET_PAIR (@var{x})->car} to refer to its
  car.
@end itemize

This change in representation reduces our heap size by 15%.  It also
makes it cheaper to decide if a value is a pair, because no memory
references are necessary; it suffices to check the bottom two bits of
the @code{SCM} value.  This may be significant when traversing lists, a
common activity in a Scheme system.

Again, most real Scheme systems use a slighty different implementation;
for example, if GET_PAIR subtracts off the low bits of @code{x}, instead
of masking them off, the optimizer will often be able to combine that
subtraction with the addition of the offset of the structure member we
are referencing, making a modified pointer as fast to use as an
unmodified pointer.


@node Guile Is Hairier
@subsection Guile Is Hairier

We originally started with a very simple typing system --- each object
has a field that indicates its type.  Then, for the sake of efficiency
in both time and space, we moved some of the typing information directly
into the @code{SCM} value, and left the rest in the @code{struct value}.
Guile itself employs a more complex hierarchy, storing finer and finer
gradations of type information in different places, depending on the
object's coarser type.

In the author's opinion, Guile could be simplified greatly without
significant loss of efficiency, but the simplified system would still be
more complex than what we've presented above.


@node How Guile does it
@section How Guile does it

Here we present the specifics of how Guile represents its data.  We
don't go into complete detail; an exhaustive description of Guile's
system would be boring, and we do not wish to encourage people to write
code which depends on its details anyway.  We do, however, present
everything one need know to use Guile's data.


@menu
* General Rules::               
* Conservative GC::          
* Immediates vs Non-immediates::  
* Immediate Datatypes::         
* Non-immediate Datatypes::     
* Signalling Type Errors::      
@end menu

@node General Rules
@subsection General Rules

Any code which operates on Guile datatypes must @code{#include} the
header file @code{<libguile.h>}.  This file contains a definition for
the @code{SCM} typedef (Guile's universal type, as in the examples
above), and definitions and declarations for a host of macros and
functions that operate on @code{SCM} values.

All identifiers declared by @code{<libguile.h>} begin with @code{scm_}
or @code{SCM_}.

@c [[I wish this were true, but I don't think it is at the moment. -JimB]]
@c Macros do not evaluate their arguments more than once, unless documented
@c to do so.

The functions described here generally check the types of their
@code{SCM} arguments, and signal an error if their arguments are of an
inappropriate type.  Macros generally do not, unless that is their
specified purpose.  You must verify their argument types beforehand, as
necessary.

Macros and functions that return a boolean value have names ending in
@code{P} or @code{_p} (for ``predicate'').  Those that return a negated
boolean value have names starting with @code{SCM_N}.  For example,
@code{SCM_IMP (@var{x})} is a predicate which returns non-zero iff
@var{x} is an immediate value (an @code{IM}).  @code{SCM_NCONSP
(@var{x})} is a predicate which returns non-zero iff @var{x} is
@emph{not} a pair object (a @code{CONS}).


@node Conservative GC
@subsection Conservative Garbage Collection

Aside from the latent typing, the major source of constraints on a
Scheme implementation's data representation is the garbage collector.
The collector must be able to traverse every live object in the heap, to
determine which objects are not live.

There are many ways to implement this, but Guile uses an algorithm
called @dfn{mark and sweep}.  The collector scans the system's global
variables and the local variables on the stack to determine which
objects are immediately accessible by the C code.  It then scans those
objects to find the objects they point to, @i{et cetera}.  The collector
sets a @dfn{mark bit} on each object it finds, so each object is
traversed only once.  This process is called @dfn{tracing}.

When the collector can find no unmarked objects pointed to by marked
objects, it assumes that any objects that are still unmarked will never
be used by the program (since there is no path of dereferences from any
global or local variable that reaches them) and deallocates them.

In the above paragraphs, we did not specify how the garbage collector
finds the global and local variables; as usual, there are many different
approaches.  Frequently, the programmer must maintain a list of pointers
to all global variables that refer to the heap, and another list
(adjusted upon entry to and exit from each function) of local variables,
for the collector's benefit.

The list of global variables is usually not too difficult to maintain,
since global variables are relatively rare.  However, an explicitly
maintained list of local variables (in the author's personal experience)
is a nightmare to maintain.  Thus, Guile uses a technique called
@dfn{conservative garbage collection}, to make the local variable list
unnecessary.

The trick to conservative collection is to treat the stack as an
ordinary range of memory, and assume that @emph{every} word on the stack
is a pointer into the heap.  Thus, the collector marks all objects whose
addresses appear anywhere in the stack, without knowing for sure how
that word is meant to be interpreted.

Obviously, such a system will occasionally retain objects that are
actually garbage, and should be freed.  In practice, this is not a
problem.  The alternative, an explicitly maintained list of local
variable addresses, is effectively much less reliable, due to programmer
error.

To accommodate this technique, data must be represented so that the
collector can accurately determine whether a given stack word is a
pointer or not.  Guile does this as follows:
@itemize @bullet

@item
Every heap object has a two-word header, called a @dfn{cell}.  Some
objects, like pairs, fit entirely in a cell's two words; others may
store pointers to additional memory in either of the words.  For
example, strings and vectors store their length in the first word, and a
pointer to their elements in the second.

@item
Guile allocates whole arrays of cells at a time, called @dfn{heap
segments}.  These segments are always allocated so that the cells they
contain fall on eight-byte boundaries, or whatever is appropriate for
the machine's word size.  Guile keeps all cells in a heap segment
initialized, whether or not they are currently in use.

@item
Guile maintains a sorted table of heap segments.

@end itemize

Thus, given any random word @var{w} fetched from the stack, Guile's
garbage collector can consult the table to see if @var{w} falls within a
known heap segment, and check @var{w}'s alignment.  If both tests pass,
the collector knows that @var{w} is a valid pointer to a cell,
intentional or not, and proceeds to trace the cell.

Note that heap segments do not contain all the data Guile uses; cells
for objects like vectors and strings contain pointers to other memory
areas.  However, since those pointers are internal, and not shared among
many pieces of code, it is enough for the collector to find the cell,
and then use the cell's type to find more pointers to trace.


@node Immediates vs Non-immediates
@subsection Immediates vs Non-immediates

Guile classifies Scheme objects into two kinds: those that fit entirely
within an @code{SCM}, and those that require heap storage.

The former class are called @dfn{immediates}.  The class of immediates
includes small integers, characters, boolean values, the empty list, the
mysterious end-of-file object, and some others.

The remaining types are called, not suprisingly, @dfn{non-immediates}.
They include pairs, procedures, strings, vectors, and all other data
types in Guile.

@deftypefn Macro int SCM_IMP (SCM @var{x})
Return non-zero iff @var{x} is an immediate object.
@end deftypefn

@deftypefn Macro int SCM_NIMP (SCM @var{x})
Return non-zero iff @var{x} is a non-immediate object.  This is the
exact complement of @code{SCM_IMP}, above.
@end deftypefn

Note that, as of Guile 1.4, it is no longer necessary to use the
@code{SCM_NIMP} macro before calling a finer-grained predicate to
determine @var{x}'s type, such as @code{SCM_CONSP} or
@code{SCM_VECTORP}.  The definitions of all Guile type predicates
now include a call to @code{SCM_NIMP} where necessary.


@node Immediate Datatypes
@subsection Immediate Datatypes

The following datatypes are immediate values; that is, they fit entirely
within an @code{SCM} value.  The @code{SCM_IMP} and @code{SCM_NIMP}
macros will distinguish these from non-immediates; see @ref{Immediates
vs Non-immediates} for an explanation of the distinction.

Note that the type predicates for immediate values work correctly on any
@code{SCM} value; you do not need to call @code{SCM_IMP} first, to
establish that a value is immediate.  This differs from the
non-immediate type predicates, which work correctly only on
non-immediate values; you must be sure the value is @code{SCM_NIMP}
before applying them.


@menu
* Integer Data::                    
* Character Data::                  
* Boolean Data::                    
* Unique Values::               
@end menu

@node Integer Data
@subsubsection Integers

Here are functions for operating on small integers, that fit within an
@code{SCM}.  Such integers are called @dfn{immediate numbers}, or
@dfn{INUMs}.  In general, INUMs occupy all but two bits of an
@code{SCM}.

Bignums and floating-point numbers are non-immediate objects, and have
their own, separate accessors.  The functions here will not work on
them.  This is not as much of a problem as you might think, however,
because the system never constructs bignums that could fit in an INUM,
and never uses floating point values for exact integers.

@deftypefn Macro int SCM_INUMP (SCM @var{x})
Return non-zero iff @var{x} is a small integer value.
@end deftypefn

@deftypefn Macro int SCM_NINUMP (SCM @var{x})
The complement of SCM_INUMP.
@end deftypefn

@deftypefn Macro int SCM_INUM (SCM @var{x})
Return the value of @var{x} as an ordinary, C integer.  If @var{x}
is not an INUM, the result is undefined.
@end deftypefn

@deftypefn Macro SCM SCM_MAKINUM (int @var{i})
Given a C integer @var{i}, return its representation as an @code{SCM}.
This function does not check for overflow.
@end deftypefn


@node Character Data
@subsubsection Characters

Here are functions for operating on characters.

@deftypefn Macro int SCM_CHARP (SCM @var{x})
Return non-zero iff @var{x} is a character value.
@end deftypefn

@deftypefn Macro {unsigned int} SCM_CHAR (SCM @var{x})
Return the value of @code{x} as a C character.  If @var{x} is not a
Scheme character, the result is undefined.
@end deftypefn

@deftypefn Macro SCM SCM_MAKE_CHAR (int @var{c})
Given a C character @var{c}, return its representation as a Scheme
character value.
@end deftypefn


@node Boolean Data
@subsubsection Booleans

Here are functions and macros for operating on booleans.

@deftypefn Macro SCM SCM_BOOL_T
@deftypefnx Macro SCM SCM_BOOL_F
The Scheme true and false values.
@end deftypefn

@deftypefn Macro int SCM_NFALSEP (@var{x})
Convert the Scheme boolean value to a C boolean.  Since every object in
Scheme except @code{#f} is true, this amounts to comparing @var{x} to
@code{#f}; hence the name.
@c Noel feels a chill here.
@end deftypefn

@deftypefn Macro SCM SCM_BOOL_NOT (@var{x})
Return the boolean inverse of @var{x}.  If @var{x} is not a
Scheme boolean, the result is undefined.
@end deftypefn


@node Unique Values
@subsubsection Unique Values

The immediate values that are neither small integers, characters, nor
booleans are all unique values --- that is, datatypes with only one
instance.

@deftypefn Macro SCM SCM_EOL
The Scheme empty list object, or ``End Of List'' object, usually written
in Scheme as @code{'()}.
@end deftypefn

@deftypefn Macro SCM SCM_EOF_VAL
The Scheme end-of-file value.  It has no standard written
representation, for obvious reasons.
@end deftypefn

@deftypefn Macro SCM SCM_UNSPECIFIED
The value returned by expressions which the Scheme standard says return
an ``unspecified'' value.

This is sort of a weirdly literal way to take things, but the standard
read-eval-print loop prints nothing when the expression returns this
value, so it's not a bad idea to return this when you can't think of
anything else helpful.
@end deftypefn

@deftypefn Macro SCM SCM_UNDEFINED
The ``undefined'' value.  Its most important property is that is not
equal to any valid Scheme value.  This is put to various internal uses
by C code interacting with Guile.

For example, when you write a C function that is callable from Scheme
and which takes optional arguments, the interpreter passes
@code{SCM_UNDEFINED} for any arguments you did not receive.

We also use this to mark unbound variables.
@end deftypefn

@deftypefn Macro int SCM_UNBNDP (SCM @var{x})
Return true if @var{x} is @code{SCM_UNDEFINED}.  Apply this to a
symbol's value to see if it has a binding as a global variable.
@end deftypefn


@node Non-immediate Datatypes
@subsection Non-immediate Datatypes 

A non-immediate datatype is one which lives in the heap, either because
it cannot fit entirely within a @code{SCM} word, or because it denotes a
specific storage location (in the nomenclature of the Revised^4 Report
on Scheme).

The @code{SCM_IMP} and @code{SCM_NIMP} macros will distinguish these
from immediates; see @ref{Immediates vs Non-immediates}.

Given a cell, Guile distinguishes between pairs and other non-immediate
types by storing special @dfn{tag} values in a non-pair cell's car, that
cannot appear in normal pairs.  A cell with a non-tag value in its car
is an ordinary pair.  The type of a cell with a tag in its car depends
on the tag; the non-immediate type predicates test this value.  If a tag
value appears elsewhere (in a vector, for example), the heap may become
corrupted.


@menu
* Non-immediate Type Predicates::  Special rules for using the type
                                        predicates described here.
* Pair Data::                       
* Vector Data::                     
* Procedures::                  
* Closures::                    
* Subrs::                       
* Port Data::                       
@end menu

@node Non-immediate Type Predicates
@subsubsection Non-immediate Type Predicates

As mentioned in @ref{Conservative GC}, all non-immediate objects
start with a @dfn{cell}, or a pair of words.  Furthermore, all type
information that distinguishes one kind of non-immediate from another is
stored in the cell.  The type information in the @code{SCM} value
indicates only that the object is a non-immediate; all finer
distinctions require one to examine the cell itself, usually with the
appropriate type predicate macro.

The type predicates for non-immediate objects generally assume that
their argument is a non-immediate value.  Thus, you must be sure that a
value is @code{SCM_NIMP} first before passing it to a non-immediate type
predicate.  Thus, the idiom for testing whether a value is a cell or not
is:
@example
SCM_NIMP (@var{x}) && SCM_CONSP (@var{x})
@end example


@node Pair Data
@subsubsection Pairs

Pairs are the essential building block of list structure in Scheme.  A
pair object has two fields, called the @dfn{car} and the @dfn{cdr}.

It is conventional for a pair's @sc{car} to contain an element of a
list, and the @sc{cdr} to point to the next pair in the list, or to
contain @code{SCM_EOL}, indicating the end of the list.  Thus, a set of
pairs chained through their @sc{cdr}s constitutes a singly-linked list.
Scheme and libguile define many functions which operate on lists
constructed in this fashion, so although lists chained through the
@sc{car}s of pairs will work fine too, they may be less convenient to
manipulate, and receive less support from the community.

Guile implements pairs by mapping the @sc{car} and @sc{cdr} of a pair
directly into the two words of the cell.


@deftypefn Macro int SCM_CONSP (SCM @var{x})
Return non-zero iff @var{x} is a Scheme pair object.
The results are undefined if @var{x} is an immediate value.
@end deftypefn

@deftypefn Macro int SCM_NCONSP (SCM @var{x})
The complement of SCM_CONSP.
@end deftypefn

@deftypefn Macro void SCM_NEWCELL (SCM @var{into})
Allocate a new cell, and set @var{into} to point to it.  This macro
expands to a statement, not an expression, and @var{into} must be an
lvalue of type SCM.

This is the most primitive way to allocate a cell; it is quite fast.

The @sc{car} of the cell initially tags it as a ``free cell''.  If the
caller intends to use it as an ordinary cons, she must store ordinary
SCM values in its @sc{car} and @sc{cdr}.

If the caller intends to use it as a header for some other type, she
must store an appropriate magic value in the cell's @sc{car}, to mark
it as a member of that type, and store whatever value in the @sc{cdr}
that type expects.  You should generally not do this, unless you are
implementing a new datatype, and thoroughly understand the code in
@code{<libguile/tags.h>}.
@end deftypefn

@deftypefun SCM scm_cons (SCM @var{car}, SCM @var{cdr})
Allocate (``CONStruct'') a new pair, with @var{car} and @var{cdr} as its
contents.
@end deftypefun


The macros below perform no typechecking.  The results are undefined if
@var{cell} is an immediate.  However, since all non-immediate Guile
objects are constructed from cells, and these macros simply return the
first element of a cell, they actually can be useful on datatypes other
than pairs.  (Of course, it is not very modular to use them outside of
the code which implements that datatype.)

@deftypefn Macro SCM SCM_CAR (SCM @var{cell})
Return the @sc{car}, or first field, of @var{cell}.
@end deftypefn

@deftypefn Macro SCM SCM_CDR (SCM @var{cell})
Return the @sc{cdr}, or second field, of @var{cell}.
@end deftypefn

@deftypefn Macro void SCM_SETCAR (SCM @var{cell}, SCM @var{x})
Set the @sc{car} of @var{cell} to @var{x}.
@end deftypefn

@deftypefn Macro void SCM_SETCDR (SCM @var{cell}, SCM @var{x})
Set the @sc{cdr} of @var{cell} to @var{x}.
@end deftypefn

@deftypefn Macro SCM SCM_CAAR (SCM @var{cell})
@deftypefnx Macro SCM SCM_CADR (SCM @var{cell})
@deftypefnx Macro SCM SCM_CDAR (SCM @var{cell}) @dots{}
@deftypefnx Macro SCM SCM_CDDDDR (SCM @var{cell})
Return the @sc{car} of the @sc{car} of @var{cell}, the @sc{car} of the
@sc{cdr} of @var{cell}, @i{et cetera}.
@end deftypefn


@node Vector Data
@subsubsection Vectors, Strings, and Symbols

Vectors, strings, and symbols have some properties in common.  They all
have a length, and they all have an array of elements.  In the case of a
vector, the elements are @code{SCM} values; in the case of a string or
symbol, the elements are characters.

All these types store their length (along with some tagging bits) in the
@sc{car} of their header cell, and store a pointer to the elements in
their @sc{cdr}.  Thus, the @code{SCM_CAR} and @code{SCM_CDR} macros
are (somewhat) meaningful when applied to these datatypes.

@deftypefn Macro int SCM_VECTORP (SCM @var{x})
Return non-zero iff @var{x} is a vector.
The results are undefined if @var{x} is an immediate value.
@end deftypefn

@deftypefn Macro int SCM_STRINGP (SCM @var{x})
Return non-zero iff @var{x} is a string.
The results are undefined if @var{x} is an immediate value.
@end deftypefn

@deftypefn Macro int SCM_SYMBOLP (SCM @var{x})
Return non-zero iff @var{x} is a symbol.
The results are undefined if @var{x} is an immediate value.
@end deftypefn

@deftypefn Macro int SCM_LENGTH (SCM @var{x})
Return the length of the object @var{x}.
The results are undefined if @var{x} is not a vector, string, or symbol.
@end deftypefn

@deftypefn Macro {SCM *} SCM_VELTS (SCM @var{x})
Return a pointer to the array of elements of the vector @var{x}.
The results are undefined if @var{x} is not a vector.
@end deftypefn

@deftypefn Macro {char *} SCM_CHARS (SCM @var{x})
Return a pointer to the characters of @var{x}.
The results are undefined if @var{x} is not a symbol or a string.
@end deftypefn

There are also a few magic values stuffed into memory before a symbol's
characters, but you don't want to know about those.  What cruft!


@node Procedures
@subsubsection Procedures

Guile provides two kinds of procedures: @dfn{closures}, which are the
result of evaluating a @code{lambda} expression, and @dfn{subrs}, which
are C functions packaged up as Scheme objects, to make them available to
Scheme programmers.

(There are actually other sorts of procedures: compiled closures, and
continuations; see the source code for details about them.)

@deftypefun SCM scm_procedure_p (SCM @var{x})
Return @code{SCM_BOOL_T} iff @var{x} is a Scheme procedure object, of
any sort.  Otherwise, return @code{SCM_BOOL_F}.
@end deftypefun


@node Closures
@subsubsection Closures

[FIXME: this needs to be further subbed, but texinfo has no subsubsub]

A closure is a procedure object, generated as the value of a
@code{lambda} expression in Scheme.  The representation of a closure is
straightforward --- it contains a pointer to the code of the lambda
expression from which it was created, and a pointer to the environment
it closes over.

In Guile, each closure also has a property list, allowing the system to
store information about the closure.  I'm not sure what this is used for
at the moment --- the debugger, maybe?

@deftypefn Macro int SCM_CLOSUREP (SCM @var{x})
Return non-zero iff @var{x} is a closure.  The results are
undefined if @var{x} is an immediate value.
@end deftypefn

@deftypefn Macro SCM SCM_PROCPROPS (SCM @var{x})
Return the property list of the closure @var{x}.  The results are
undefined if @var{x} is not a closure.
@end deftypefn

@deftypefn Macro void SCM_SETPROCPROPS (SCM @var{x}, SCM @var{p})
Set the property list of the closure @var{x} to @var{p}.  The results
are undefined if @var{x} is not a closure.
@end deftypefn

@deftypefn Macro SCM SCM_CODE (SCM @var{x})
Return the code of the closure @var{x}.  The results are undefined if
@var{x} is not a closure.

This function should probably only be used internally by the
interpreter, since the representation of the code is intimately
connected with the interpreter's implementation.
@end deftypefn

@deftypefn Macro SCM SCM_ENV (SCM @var{x})
Return the environment enclosed by @var{x}.
The results are undefined if @var{x} is not a closure.

This function should probably only be used internally by the
interpreter, since the representation of the environment is intimately
connected with the interpreter's implementation.
@end deftypefn


@node Subrs
@subsubsection Subrs

[FIXME: this needs to be further subbed, but texinfo has no subsubsub]

A subr is a pointer to a C function, packaged up as a Scheme object to
make it callable by Scheme code.  In addition to the function pointer,
the subr also contains a pointer to the name of the function, and
information about the number of arguments accepted by the C fuction, for
the sake of error checking.

There is no single type predicate macro that recognizes subrs, as
distinct from other kinds of procedures.  The closest thing is
@code{scm_procedure_p}; see @ref{Procedures}.

@deftypefn Macro {char *} SCM_SNAME (@var{x})
Return the name of the subr @var{x}.  The results are undefined if
@var{x} is not a subr.
@end deftypefn

@deftypefun SCM scm_make_gsubr (char *@var{name}, int @var{req}, int @var{opt}, int @var{rest}, SCM (*@var{function})())
Create a new subr object named @var{name}, based on the C function
@var{function}, make it visible to Scheme the value of as a global
variable named @var{name}, and return the subr object.

The subr object accepts @var{req} required arguments, @var{opt} optional
arguments, and a @var{rest} argument iff @var{rest} is non-zero.  The C
function @var{function} should accept @code{@var{req} + @var{opt}}
arguments, or @code{@var{req} + @var{opt} + 1} arguments if @code{rest}
is non-zero.

When a subr object is applied, it must be applied to at least @var{req}
arguments, or else Guile signals an error.  @var{function} receives the
subr's first @var{req} arguments as its first @var{req} arguments.  If
there are fewer than @var{opt} arguments remaining, then @var{function}
receives the value @code{SCM_UNDEFINED} for any missing optional
arguments.  If @var{rst} is non-zero, then any arguments after the first
@code{@var{req} + @var{opt}} are packaged up as a list as passed as
@var{function}'s last argument.

Note that subrs can actually only accept a predefined set of
combinations of required, optional, and rest arguments.  For example, a
subr can take one required argument, or one required and one optional
argument, but a subr can't take one required and two optional arguments.
It's bizarre, but that's the way the interpreter was written.  If the
arguments to @code{scm_make_gsubr} do not fit one of the predefined
patterns, then @code{scm_make_gsubr} will return a compiled closure
object instead of a subr object.
@end deftypefun


@node Port Data
@subsubsection Ports

Haven't written this yet, 'cos I don't understand ports yet.


@node Signalling Type Errors
@subsection Signalling Type Errors

Every function visible at the Scheme level should aggressively check the
types of its arguments, to avoid misinterpreting a value, and perhaps
causing a segmentation fault.  Guile provides some macros to make this
easier.

@deftypefn Macro void SCM_ASSERT (int @var{test}, SCM @var{obj}, int @var{position}, char *@var{subr})
If @var{test} is zero, signal an error, attributed to the subroutine
named @var{subr}, operating on the value @var{obj}.  The @var{position}
value determines exactly what sort of error to signal.

If @var{position} is a string, @code{SCM_ASSERT} raises a
``miscellaneous'' error whose message is that string.

Otherwise, @var{position} should be one of the values defined below.
@end deftypefn

@deftypefn Macro int SCM_ARG1
@deftypefnx Macro int SCM_ARG2
@deftypefnx Macro int SCM_ARG3
@deftypefnx Macro int SCM_ARG4
@deftypefnx Macro int SCM_ARG5
Signal a ``wrong type argument'' error.  When used as the @var{position}
argument of @code{SCM_ASSERT}, @code{SCM_ARG@var{n}} claims that
@var{obj} has the wrong type for the @var{n}'th argument of @var{subr}.

The only way to complain about the type of an argument after the fifth
is to use @code{SCM_ARGn}, defined below, which doesn't specify which
argument is wrong.  You could pass your own error message to
@code{SCM_ASSERT} as the @var{position}, but then the error signalled is
a ``miscellaneous'' error, not a ``wrong type argument'' error.  This
seems kludgy to me.
@comment Any function with more than two arguments is wrong --- Perlis
@comment Despite Perlis, I agree.  Why not have two Macros, one with
@comment a string error message, and the other with an integer position
@comment that only claims a type error in an argument?
@comment --- Keith Wright
@end deftypefn

@deftypefn Macro int SCM_ARGn
As above, but does not specify which argument's type is incorrect.
@end deftypefn

@deftypefn Macro int SCM_WNA
Signal an error complaining that the function received the wrong number
of arguments.

Interestingly, the message is attributed to the function named by
@var{obj}, not @var{subr}, so @var{obj} must be a Scheme string object
naming the function.  Usually, Guile catches these errors before ever
invoking the subr, so we don't run into these problems.
@end deftypefn


@node Defining New Types (Smobs)
@section Defining New Types (Smobs)

@dfn{Smobs} are Guile's mechanism for adding new non-immediate types to
the system.@footnote{The term ``smob'' was coined by Aubrey Jaffer, who
says it comes from ``small object'', referring to the fact that only the
@sc{cdr} and part of the @sc{car} of a smob's cell are available for
use.}  To define a new smob type, the programmer provides Guile with
some essential information about the type --- how to print it, how to
garbage collect it, and so on --- and Guile returns a fresh type tag for
use in the @sc{car} of new cells.  The programmer can then use
@code{scm_make_gsubr} to make a set of C functions that create and
operate on these objects visible to Scheme code.

(You can find a complete version of the example code used in this
section in the Guile distribution, in @file{doc/example-smob}.  That
directory includes a makefile and a suitable @code{main} function, so
you can build a complete interactive Guile shell, extended with the
datatypes described here.)

@menu
* Describing a New Type::       
* Creating Instances::          
* Typechecking::                
* Garbage Collecting Smobs::    
* A Common Mistake In Allocating Smobs::  
* Garbage Collecting Simple Smobs::  
* A Complete Example::          
@end menu

@node Describing a New Type
@subsection Describing a New Type

To define a new type, the programmer must write four functions to
manage instances of the type:

@table @code
@item mark
Guile will apply this function to each instance of the new type it
encounters during garbage collection.  This function is responsible for
telling the collector about any other non-immediate objects the object
refers to.  The default smob mark function is to not mark any data.
@xref{Garbage Collecting Smobs}, for more details.

@item free
Guile will apply this function to each instance of the new type it could
not find any live pointers to.  The function should release all
resources held by the object and return the number of bytes released.
This is analagous to the Java finalization method-- it is invoked at
an unspecified time (when garbage collection occurs) after the object
is dead.
The default free function frees the smob data (if the size of the struct
passed to @code{scm_make_smob_type} or @code{scm_make_smob_type_mfpe} is
non-zero) using @code{scm_must_free} and returns the size of that
struct.  @xref{Garbage Collecting Smobs}, for more details.

@item print
@c GJB:FIXME:: @var{exp} and @var{port} need to refer to a prototype of 
@c the print function.... where is that, or where should it go?
Guile will apply this function to each instance of the new type to print
the value, as for @code{display} or @code{write}.  The function should
write a printed representation of @var{exp} on @var{port}, in accordance
with the parameters in @var{pstate}.  (For more information on print
states, see @ref{Port Data}.)  The default print function prints @code{#<NAME ADDRESS>} 
where @code{NAME} is the first argument passed to @code{scm_make_smob_type} or 
@code{scm_make_smob_type_mfpe}.

@item equalp
If Scheme code asks the @code{equal?} function to compare two instances
of the same smob type, Guile calls this function.  It should return
@code{SCM_BOOL_T} if @var{a} and @var{b} should be considered
@code{equal?}, or @code{SCM_BOOL_F} otherwise.  If @code{equalp} is
@code{NULL}, @code{equal?} will assume that two instances of this type are
never @code{equal?} unless they are @code{eq?}.

@end table

To actually register the new smob type, call @code{scm_make_smob_type}:

@deftypefun long scm_make_smob_type (const char *name, scm_sizet size)
This function implements the standard way of adding a new smob type,
named @var{name}, with instance size @var{size}, to the system.  The
return value is a tag that is used in creating instances of the type.
If @var{size} is 0, then no memory will be allocated when instances of
the smob are created, and nothing will be freed by the default free
function.  Default values are provided for mark, free, print, and,
equalp, as described above.  If you want to customize any of these
functions, the call to @code{scm_make_smob_type} should be immediately
followed by calls to one or several of @code{scm_set_smob_mark},
@code{scm_set_smob_free}, @code{scm_set_smob_print}, and/or
@code{scm_set_smob_equalp}.
@end deftypefun

Each of the below @code{scm_set_smob_XXX} functions registers a smob
special function for a given type.  Each function is intended to be used
only zero or one time per type, and the call should be placed
immediately following the call to @code{scm_make_smob_type}.

@deftypefun void scm_set_smob_mark (long tc, SCM (*mark) (SCM))
This function sets the smob marking procedure for the smob type specified by
the tag @var{tc}. @var{tc} is the tag returned by @code{scm_make_smob_type}.
@end deftypefun

@deftypefun void scm_set_smob_free (long tc, scm_sizet (*free) (SCM))
This function sets the smob freeing procedure for the smob type specified by
the tag @var{tc}. @var{tc} is the tag returned by @code{scm_make_smob_type}.
@end deftypefun

@deftypefun void scm_set_smob_print (long tc, int (*print) (SCM,SCM,scm_print_state*))
This function sets the smob printing procedure for the smob type specified by
the tag @var{tc}. @var{tc} is the tag returned by @code{scm_make_smob_type}.
@end deftypefun

@deftypefun void scm_set_smob_equalp (long tc, SCM (*equalp) (SCM,SCM))
This function sets the smob equality-testing predicate for the smob type specified by
the tag @var{tc}. @var{tc} is the tag returned by @code{scm_make_smob_type}.
@end deftypefun

Instead of using @code{scm_make_smob_type} and calling each of the
individual @code{scm_set_smob_XXX} functions to register each special
function independently, you can use @code{scm_make_smob_type_mfpe} to
register all of the special functions at once as you create the smob
type@footnote{Warning: There is an ongoing discussion among the developers which
may result in deprecating @code{scm_make_smob_type_mfpe} in next release
of Guile.}:

@deftypefun long scm_make_smob_type_mfpe(const char *name, scm_sizet size, SCM (*mark) (SCM), scm_sizet (*free) (SCM), int (*print) (SCM, SCM, scm_print_state*), SCM (*equalp) (SCM, SCM))
This function invokes @code{scm_make_smob_type} on its first two arguments
to add a new smob type named @var{name}, with instance size @var{size} to the system.  
It also registers the @var{mark}, @var{free}, @var{print}, @var{equalp} smob
special functions for that new type.  Any of these parameters can be @code{NULL}
to have that special function use the default behaviour for guile.
The return value is a tag that is used in creating instances of the type.  If @var{size}
is 0, then no memory will be allocated when instances of the smob are created, and
nothing will be freed by the default free function.
@end deftypefun

For example, here is how one might declare and register a new type
representing eight-bit grayscale images:
@example
#include <libguile.h>

long image_tag;

void
init_image_type ()
@{
  image_tag = scm_make_smob_type_mfpe ("image",sizeof(struct image),
				       mark_image, free_image, print_image, NULL);
@}
@end example


@node Creating Instances
@subsection Creating Instances

Like other non-immediate types, smobs start with a cell whose @sc{car}
contains typing information, and whose @code{cdr} is free for any use.  For smobs,
the @code{cdr} stores a pointer to the internal C structure holding the 
smob-specific data.
To create an instance of a smob type following these standards, you should
use @code{SCM_NEWSMOB}:

@deftypefn Macro void SCM_NEWSMOB(SCM value,long tag,void *data)
Make @var{value} contain a smob instance of the type with tag @var{tag}
and smob data @var{data}.  @var{value} must be previously declared
as C type @code{SCM}.
@end deftypefn

Since it is often the case (e.g., in smob constructors) that you will
create a smob instance and return it, there is also a slightly specialized
macro for this situation:

@deftypefn Macro fn_returns SCM_RETURN_NEWSMOB(long tab, void *data)
This macro expands to a block of code that creates a smob instance of
the type with tag @var{tag} and smob data @var{data}, and returns
that @code{SCM} value.  It should be the last piece of code in 
a block.
@end deftypefn

Guile provides the following functions for managing memory, which are
often helpful when implementing smobs:

@deftypefun {char *} scm_must_malloc (long @var{len}, char *@var{what})
Allocate @var{len} bytes of memory, using @code{malloc}, and return a
pointer to them.

If there is not enough memory available, invoke the garbage collector,
and try once more.  If there is still not enough, signal an error,
reporting that we could not allocate @var{what}.

This function also helps maintain statistics about the size of the heap.
@end deftypefun

@deftypefun {char *} scm_must_realloc (char *@var{addr}, long @var{olen}, long @var{len}, char *@var{what})
Resize (and possibly relocate) the block of memory at @var{addr}, to
have a size of @var{len} bytes, by calling @code{realloc}.  Return a
pointer to the new block.

If there is not enough memory available, invoke the garbage collector,
and try once more.  If there is still not enough, signal an error,
reporting that we could not allocate @var{what}.

The value @var{olen} should be the old size of the block of memory at
@var{addr}; it is only used for keeping statistics on the size of the
heap.
@end deftypefun

@deftypefun void scm_must_free (char *@var{addr})
Free the block of memory at @var{addr}, using @code{free}.  If
@var{addr} is zero, signal an error, complaining of an attempt to free
something that is already free.

This does no record-keeping; instead, the smob's @code{free} function
must take care of that.

This function isn't usually sufficiently different from the usual
@code{free} function to be worth using.
@end deftypefun


Continuing the above example, if the global variable @code{image_tag}
contains a tag returned by @code{scm_newsmob}, here is how we could
construct a smob whose @sc{cdr} contains a pointer to a freshly
allocated @code{struct image}:

@example
struct image @{
  int width, height;
  char *pixels;

  /* The name of this image */
  SCM name;

  /* A function to call when this image is
     modified, e.g., to update the screen,
     or SCM_BOOL_F if no action necessary */
  SCM update_func;
@};

SCM
make_image (SCM name, SCM s_width, SCM s_height)
@{
  struct image *image;
  int width, height;

  SCM_ASSERT (SCM_NIMP (name) && SCM_STRINGP (name), name,
              SCM_ARG1, "make-image");
  SCM_ASSERT (SCM_INUMP (s_width),  s_width,  SCM_ARG2, "make-image");
  SCM_ASSERT (SCM_INUMP (s_height), s_height, SCM_ARG3, "make-image");

  width = SCM_INUM (s_width);
  height = SCM_INUM (s_height);
  
  image = (struct image *) scm_must_malloc (sizeof (struct image), "image");
  image->width = width;
  image->height = height;
  image->pixels = scm_must_malloc (width * height, "image pixels");
  image->name = name;
  image->update_func = SCM_BOOL_F;

  SCM_RETURN_NEWSMOB (image_tag, image);
@}
@end example


@node Typechecking
@subsection Typechecking

Functions that operate on smobs should aggressively check the types of
their arguments, to avoid misinterpreting some other datatype as a smob,
and perhaps causing a segmentation fault.  Fortunately, this is pretty
simple to do.  The function need only verify that its argument is a
non-immediate, whose @sc{car} is the type tag returned by
@code{scm_newsmob}.

For example, here is a simple function that operates on an image smob,
and checks the type of its argument.  We also present an expanded
version of the @code{init_image_type} function, to make
@code{clear_image} and the image constructor function @code{make_image}
visible to Scheme code.
@example
SCM
clear_image (SCM image_smob)
@{
  int area;
  struct image *image;

  SCM_ASSERT (SCM_SMOB_PREDICATE (image_tag, image_smob),
              image_smob, SCM_ARG1, "clear-image");

  image = (struct image *) SCM_SMOB_DATA (image_smob);
  area = image->width * image->height;
  memset (image->pixels, 0, area);

  /* Invoke the image's update function.  */
  if (image->update_func != SCM_BOOL_F)
    scm_apply (image->update_func, SCM_EOL, SCM_EOL);

  return SCM_UNSPECIFIED;
@}


void
init_image_type ()
@{
  image_tag = scm_newsmob (&image_funs);

  scm_make_gsubr ("make-image", 3, 0, 0, make_image);
  scm_make_gsubr ("clear-image", 1, 0, 0, clear_image);
@}
@end example

Note that checking types is a little more complicated during garbage
collection; see the description of @code{SCM_GCTYP16} in @ref{Garbage
Collecting Smobs}.

@c GJB:FIXME:: should talk about guile-snarf somewhere!

@node Garbage Collecting Smobs
@subsection Garbage Collecting Smobs

Once a smob has been released to the tender mercies of the Scheme
system, it must be prepared to survive garbage collection.  Guile calls
the @code{mark} and @code{free} functions of the @code{scm_smobfuns}
structure to manage this.

As described before (@pxref{Conservative GC}), every object in the
Scheme system has a @dfn{mark bit}, which the garbage collector uses to
tell live objects from dead ones.  When collection starts, every
object's mark bit is clear.  The collector traces pointers through the
heap, starting from objects known to be live, and sets the mark bit on
each object it encounters.  When it can find no more unmarked objects,
the collector walks all objects, live and dead, frees those whose mark
bits are still clear, and clears the mark bit on the others.

The two main portions of the collection are called the @dfn{mark phase},
during which the collector marks live objects, and the @dfn{sweep
phase}, during which the collector frees all unmarked objects.

The mark bit of a smob lives in its @sc{car}, along with the smob's type
tag.  When the collector encounters a smob, it sets the smob's mark bit,
and uses the smob's type tag to find the appropriate @code{mark}
function for that smob: the one listed in that smob's
@code{scm_smobfuns} structure.  It then calls the @code{mark} function,
passing it the smob as its only argument.

The @code{mark} function is responsible for marking any other Scheme
objects the smob refers to.  If it does not do so, the objects' mark
bits will still be clear when the collector begins to sweep, and the
collector will free them.  If this occurs, it will probably break, or at
least confuse, any code operating on the smob; the smob's @code{SCM}
values will have become dangling references.

To mark an arbitrary Scheme object, the @code{mark} function may call
this function:

@deftypefun void scm_gc_mark (SCM @var{x})
Mark the object @var{x}, and recurse on any objects @var{x} refers to.
If @var{x}'s mark bit is already set, return immediately.
@end deftypefun

Thus, here is how we might write the @code{mark} function for the image
smob type discussed above:
@example
@group
SCM
mark_image (SCM image_smob)
@{
  /* Mark the image's name and update function.  */
  struct image *image = (struct image *) SCM_SMOB_DATA (image_smob);

  scm_gc_mark (image->name);
  scm_gc_mark (image->update_func);

  return SCM_BOOL_F;
@}
@end group
@end example

Note that, even though the image's @code{update_func} could be an
arbitrarily complex structure (representing a procedure and any values
enclosed in its environment), @code{scm_gc_mark} will recurse as
necessary to mark all its components.  Because @code{scm_gc_mark} sets
an object's mark bit before it recurses, it is not confused by
circular structures.

As an optimization, the collector will mark whatever value is returned
by the @code{mark} function; this helps limit depth of recursion during
the mark phase.  Thus, the code above could also be written as:
@example
@group
SCM
mark_image (SCM image_smob)
@{
  /* Mark the image's name and update function.  */
  struct image *image = (struct image *) SCM_SMOB_DATA (image_smob);

  scm_gc_mark (image->name);
  return image->update_func;
@}
@end group
@end example


Finally, when the collector encounters an unmarked smob during the sweep
phase, it uses the smob's tag to find the appropriate @code{free}
function for the smob.  It then calls the function, passing it the smob
as its only argument.

The @code{free} function must release any resources used by the smob.
However, it need not free objects managed by the collector; the
collector will take care of them.  The return type of the @code{free}
function should be @code{scm_sizet}, an unsigned integral type; the
@code{free} function should return the number of bytes released, to help
the collector maintain statistics on the size of the heap.

Here is how we might write the @code{free} function for the image smob
type:
@example
scm_sizet
free_image (SCM image_smob)
@{
  struct image *image = (struct image *) SCM_SMOB_DATA (image_smob);
  scm_sizet size = image->width * image->height + sizeof (*image);

  free (image->pixels);
  free (image);

  return size;
@}
@end example

During the sweep phase, the garbage collector will clear the mark bits
on all live objects.  The code which implements a smob need not do this
itself.

There is no way for smob code to be notified when collection is
complete.

Note that, since a smob's mark bit lives in its @sc{car}, along with the
smob's type tag, the technique for checking the type of a smob described
in @ref{Typechecking} will not necessarily work during GC.  If you need
to find out whether a given object is a particular smob type during GC,
use the following macro:

@deftypefn Macro void SCM_GCTYP16 (SCM @var{x})
Return the type bits of the smob @var{x}, with the mark bit clear.

Use this macro instead of @code{SCM_CAR} to check the type of a smob
during GC.  Usually, only code called by the smob's @code{mark} function
need worry about this.
@end deftypefn

It is usually a good idea to minimize the amount of processing done
during garbage collection; keep @code{mark} and @code{free} functions
very simple.  Since collections occur at unpredictable times, it is easy
for any unusual activity to interfere with normal code.


@node A Common Mistake In Allocating Smobs, Garbage Collecting Simple Smobs, Garbage Collecting Smobs, Defining New Types (Smobs)
@subsection A Common Mistake In Allocating Smobs

When constructing new objects, you must be careful that the garbage
collector can always find any new objects you allocate.  For example,
suppose we wrote the @code{make_image} function this way:

@example
SCM
make_image (SCM name, SCM s_width, SCM s_height)
@{
  struct image *image;
  SCM image_smob;
  int width, height;

  SCM_ASSERT (SCM_NIMP (name) && SCM_STRINGP (name), name,
              SCM_ARG1, "make-image");
  SCM_ASSERT (SCM_INUMP (s_width),  s_width,  SCM_ARG2, "make-image");
  SCM_ASSERT (SCM_INUMP (s_height), s_height, SCM_ARG3, "make-image");

  width = SCM_INUM (s_width);
  height = SCM_INUM (s_height);
  
  image = (struct image *) scm_must_malloc (sizeof (struct image), "image");
  image->width = width;
  image->height = height;
  image->pixels = scm_must_malloc (width * height, "image pixels");

  /* THESE TWO LINES HAVE CHANGED: */
  image->name = scm_string_copy (name);
  image->update_func = scm_make_gsubr (@dots{});

  SCM_NEWCELL (image_smob);
  SCM_SETCDR (image_smob, image);
  SCM_SETCAR (image_smob, image_tag);

  return image_smob;
@}
@end example

This code is incorrect.  The calls to @code{scm_string_copy} and
@code{scm_make_gsubr} allocate fresh objects.  Allocating any new object
may cause the garbage collector to run.  If @code{scm_make_gsubr}
invokes a collection, the garbage collector has no way to discover that
@code{image->name} points to the new string object; the @code{image}
structure is not yet part of any Scheme object, so the garbage collector
will not traverse it.  Since the garbage collector cannot find any
references to the new string object, it will free it, leaving
@code{image} pointing to a dead object.

A correct implementation might say, instead:
@example
  image->name = SCM_BOOL_F;
  image->update_func = SCM_BOOL_F;

  SCM_NEWCELL (image_smob);
  SCM_SETCDR (image_smob, image);
  SCM_SETCAR (image_smob, image_tag);

  image->name = scm_string_copy (name);
  image->update_func = scm_make_gsubr (@dots{});

  return image_smob;
@end example

Now, by the time we allocate the new string and function objects,
@code{image_smob} points to @code{image}.  If the garbage collector
scans the stack, it will find a reference to @code{image_smob} and
traverse @code{image}, so any objects @code{image} points to will be
preserved.


@node Garbage Collecting Simple Smobs, A Complete Example, A Common Mistake In Allocating Smobs, Defining New Types (Smobs)
@subsection Garbage Collecting Simple Smobs

It is often useful to define very simple smob types --- smobs which have
no data to mark, other than the cell itself, or smobs whose @sc{cdr} is
simply an ordinary Scheme object, to be marked recursively.  Guile
provides some functions to handle these common cases; you can use these
functions as your smob type's @code{mark} function, if your smob's
structure is simple enough.

If the smob refers to no other Scheme objects, then no action is
necessary; the garbage collector has already marked the smob cell
itself.  In that case, you can use zero as your mark function.

@deftypefun SCM scm_markcdr (SCM @var{x})
Mark the references in the smob @var{x}, assuming that @var{x}'s
@sc{cdr} contains an ordinary Scheme object, and @var{x} refers to no
other objects.  This function simply returns @var{x}'s @sc{cdr}.
@end deftypefun

@deftypefun scm_sizet scm_free0 (SCM @var{x})
Do nothing; return zero.  This function is appropriate for smobs that
use either zero or @code{scm_markcdr} as their marking functions, and
refer to no heap storage, including memory managed by @code{malloc},
other than the smob's header cell.
@end deftypefun


@node A Complete Example
@subsection A Complete Example

Here is the complete text of the implementation of the image datatype,
as presented in the sections above.  We also provide a definition for
the smob's @code{print} function, and make some objects and functions
static, to clarify exactly what the surrounding code is using.

As mentioned above, you can find this code in the Guile distribution, in
@file{doc/example-smob}.  That directory includes a makefile and a
suitable @code{main} function, so you can build a complete interactive
Guile shell, extended with the datatypes described here.)

@example
/* file "image-type.c" */

#include <stdlib.h>
#include <libguile.h>

static long image_tag;

struct image @{
  int width, height;
  char *pixels;

  /* The name of this image */
  SCM name;

  /* A function to call when this image is
     modified, e.g., to update the screen,
     or SCM_BOOL_F if no action necessary */
  SCM update_func;
@};

static SCM
make_image (SCM name, SCM s_width, SCM s_height)
@{
  struct image *image;
  SCM image_smob;
  int width, height;

  SCM_ASSERT (SCM_NIMP (name) && SCM_STRINGP (name), name,
              SCM_ARG1, "make-image");
  SCM_ASSERT (SCM_INUMP (s_width),  s_width,  SCM_ARG2, "make-image");
  SCM_ASSERT (SCM_INUMP (s_height), s_height, SCM_ARG3, "make-image");

  width = SCM_INUM (s_width);
  height = SCM_INUM (s_height);
  
  image = (struct image *) scm_must_malloc (sizeof (struct image), "image");
  image->width = width;
  image->height = height;
  image->pixels = scm_must_malloc (width * height, "image pixels");
  image->name = name;
  image->update_func = SCM_BOOL_F;

  SCM_NEWSMOB (image_smob, image_tag, image);

  return image_smob;
@}

static SCM
clear_image (SCM image_smob)
@{
  int area;
  struct image *image;

  SCM_ASSERT (SCM_SMOB_PREDICATE (image_tag, image_smob),
              image_smob, SCM_ARG1, "clear-image");

  image = (struct image *) SCM_SMOB_DATA (image_smob);
  area = image->width * image->height;
  memset (image->pixels, 0, area);

  /* Invoke the image's update function.  */
  if (image->update_func != SCM_BOOL_F)
    scm_apply (image->update_func, SCM_EOL, SCM_EOL);

  return SCM_UNSPECIFIED;
@}

static SCM
mark_image (SCM image_smob)
@{
  struct image *image = (struct image *) SCM_SMOB_DATA (image_smob);

  scm_gc_mark (image->name);
  return image->update_func;
@}

static scm_sizet
free_image (SCM image_smob)
@{
  struct image *image = (struct image *) SCM_SMOB_DATA (image_smob);
  scm_sizet size = image->width * image->height + sizeof (struct image);

  free (image->pixels);
  free (image);

  return size;
@}

static int
print_image (SCM image_smob, SCM port, scm_print_state *pstate)
@{
  struct image *image = (struct image *) SCM_SMOB_DATA (image_smob);

  scm_puts ("#<image ", port);
  scm_display (image->name, port);
  scm_puts (">", port);

  /* non-zero means success */
  return 1;
@}

static scm_smobfuns image_funs = @{
  mark_image, free_image, print_image, 0
@};

void
init_image_type ()
@{
  image_tag = scm_newsmob (&image_funs);

  scm_make_gsubr ("clear-image", 1, 0, 0, clear_image);
  scm_make_gsubr ("make-image", 3, 0, 0, make_image);
@}
@end example

Here is a sample build and interaction with the code from the
@file{example-smob} directory, on the author's machine:

@example
zwingli:example-smob$ make CC=gcc
gcc `guile-config compile`   -c image-type.c -o image-type.o
gcc `guile-config compile`   -c myguile.c -o myguile.o
gcc image-type.o myguile.o `guile-config link` -o myguile
zwingli:example-smob$ ./myguile
guile> make-image
#<primitive-procedure make-image>
guile> (define i (make-image "Whistler's Mother" 100 100))
guile> i
#<image Whistler's Mother>
guile> (clear-image i)
guile> (clear-image 4)
ERROR: In procedure clear-image in expression (clear-image 4):
ERROR: Wrong type argument in position 1: 4
ABORT: (wrong-type-arg)
 
Type "(backtrace)" to get more information.
guile> 
@end example

@c essay @bye