Sync up with r21-4-14-chise-0_21-17.

[chise/xemacs-chise.git] / info / cl.info-5

This is ../info/cl.info, produced by makeinfo version 4.0 from cl.texi.

INFO-DIR-SECTION XEmacs Editor
START-INFO-DIR-ENTRY
* Common Lisp: (cl).            GNU Emacs Common Lisp emulation package.
END-INFO-DIR-ENTRY

   This file documents the GNU Emacs Common Lisp emulation package.

   Copyright (C) 1993 Free Software Foundation, Inc.

   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 also
that the section entitled "GNU General Public License" is included
exactly as in the original, and 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 the section entitled "GNU General Public License"
may be included in a translation approved by the author instead of in
the original English.

\1f
File: cl.info,  Node: Structures,  Next: Assertions,  Prev: Hash Tables,  Up: Top

Structures
**********

The Common Lisp "structure" mechanism provides a general way to define
data types similar to C's `struct' types.  A structure is a Lisp object
containing some number of "slots", each of which can hold any Lisp data
object.  Functions are provided for accessing and setting the slots,
creating or copying structure objects, and recognizing objects of a
particular structure type.

   In true Common Lisp, each structure type is a new type distinct from
all existing Lisp types.  Since the underlying Emacs Lisp system
provides no way to create new distinct types, this package implements
structures as vectors (or lists upon request) with a special "tag"
symbol to identify them.

 - Special Form: defstruct name slots...
     The `defstruct' form defines a new structure type called NAME,
     with the specified SLOTS.  (The SLOTS may begin with a string
     which documents the structure type.)  In the simplest case, NAME
     and each of the SLOTS are symbols.  For example,

          (defstruct person name age sex)

     defines a struct type called `person' which contains three slots.
     Given a `person' object P, you can access those slots by calling
     `(person-name P)', `(person-age P)', and `(person-sex P)'.  You
     can also change these slots by using `setf' on any of these place
     forms:

          (incf (person-age birthday-boy))

     You can create a new `person' by calling `make-person', which
     takes keyword arguments `:name', `:age', and `:sex' to specify the
     initial values of these slots in the new object.  (Omitting any of
     these arguments leaves the corresponding slot "undefined,"
     according to the Common Lisp standard; in Emacs Lisp, such
     uninitialized slots are filled with `nil'.)

     Given a `person', `(copy-person P)' makes a new object of the same
     type whose slots are `eq' to those of P.

     Given any Lisp object X, `(person-p X)' returns true if X looks
     like a `person', false otherwise.  (Again, in Common Lisp this
     predicate would be exact; in Emacs Lisp the best it can do is
     verify that X is a vector of the correct length which starts with
     the correct tag symbol.)

     Accessors like `person-name' normally check their arguments
     (effectively using `person-p') and signal an error if the argument
     is the wrong type.  This check is affected by `(optimize (safety
     ...))' declarations.  Safety level 1, the default, uses a somewhat
     optimized check that will detect all incorrect arguments, but may
     use an uninformative error message (e.g., "expected a vector"
     instead of "expected a `person'").  Safety level 0 omits all
     checks except as provided by the underlying `aref' call; safety
     levels 2 and 3 do rigorous checking that will always print a
     descriptive error message for incorrect inputs.  *Note
     Declarations::.

          (setq dave (make-person :name "Dave" :sex 'male))
               => [cl-struct-person "Dave" nil male]
          (setq other (copy-person dave))
               => [cl-struct-person "Dave" nil male]
          (eq dave other)
               => nil
          (eq (person-name dave) (person-name other))
               => t
          (person-p dave)
               => t
          (person-p [1 2 3 4])
               => nil
          (person-p "Bogus")
               => nil
          (person-p '[cl-struct-person counterfeit person object])
               => t

     In general, NAME is either a name symbol or a list of a name
     symbol followed by any number of "struct options"; each SLOT is
     either a slot symbol or a list of the form `(SLOT-NAME
     DEFAULT-VALUE SLOT-OPTIONS...)'.  The DEFAULT-VALUE is a Lisp form
     which is evaluated any time an instance of the structure type is
     created without specifying that slot's value.

     Common Lisp defines several slot options, but the only one
     implemented in this package is `:read-only'.  A non-`nil' value
     for this option means the slot should not be `setf'-able; the
     slot's value is determined when the object is created and does not
     change afterward.

          (defstruct person
            (name nil :read-only t)
            age
            (sex 'unknown))

     Any slot options other than `:read-only' are ignored.

     For obscure historical reasons, structure options take a different
     form than slot options.  A structure option is either a keyword
     symbol, or a list beginning with a keyword symbol possibly followed
     by arguments.  (By contrast, slot options are key-value pairs not
     enclosed in lists.)

          (defstruct (person (:constructor create-person)
                             (:type list)
                             :named)
            name age sex)

     The following structure options are recognized.

    `:conc-name'
          The argument is a symbol whose print name is used as the
          prefix for the names of slot accessor functions.  The default
          is the name of the struct type followed by a hyphen.  The
          option `(:conc-name p-)' would change this prefix to `p-'.
          Specifying `nil' as an argument means no prefix, so that the
          slot names themselves are used to name the accessor functions.

    `:constructor'
          In the simple case, this option takes one argument which is an
          alternate name to use for the constructor function.  The
          default is `make-NAME', e.g., `make-person'.  The above
          example changes this to `create-person'.  Specifying `nil' as
          an argument means that no standard constructor should be
          generated at all.

          In the full form of this option, the constructor name is
          followed by an arbitrary argument list.  *Note Program
          Structure::, for a description of the format of Common Lisp
          argument lists.  All options, such as `&rest' and `&key', are
          supported.  The argument names should match the slot names;
          each slot is initialized from the corresponding argument.
          Slots whose names do not appear in the argument list are
          initialized based on the DEFAULT-VALUE in their slot
          descriptor.  Also, `&optional' and `&key' arguments which
          don't specify defaults take their defaults from the slot
          descriptor.  It is legal to include arguments which don't
          correspond to slot names; these are useful if they are
          referred to in the defaults for optional, keyword, or `&aux'
          arguments which _do_ correspond to slots.

          You can specify any number of full-format `:constructor'
          options on a structure.  The default constructor is still
          generated as well unless you disable it with a simple-format
          `:constructor' option.

               (defstruct
                (person
                 (:constructor nil)   ; no default constructor
                 (:constructor new-person (name sex &optional (age 0)))
                 (:constructor new-hound (&key (name "Rover")
                                               (dog-years 0)
                                          &aux (age (* 7 dog-years))
                                               (sex 'canine))))
                name age sex)

          The first constructor here takes its arguments positionally
          rather than by keyword.  (In official Common Lisp
          terminology, constructors that work By Order of Arguments
          instead of by keyword are called "BOA constructors."  No, I'm
          not making this up.)  For example, `(new-person "Jane"
          'female)' generates a person whose slots are `"Jane"', 0, and
          `female', respectively.

          The second constructor takes two keyword arguments, `:name',
          which initializes the `name' slot and defaults to `"Rover"',
          and `:dog-years', which does not itself correspond to a slot
          but which is used to initialize the `age' slot.  The `sex'
          slot is forced to the symbol `canine' with no syntax for
          overriding it.

    `:copier'
          The argument is an alternate name for the copier function for
          this type.  The default is `copy-NAME'.  `nil' means not to
          generate a copier function.  (In this implementation, all
          copier functions are simply synonyms for `copy-sequence'.)

    `:predicate'
          The argument is an alternate name for the predicate which
          recognizes objects of this type.  The default is `NAME-p'.
          `nil' means not to generate a predicate function.  (If the
          `:type' option is used without the `:named' option, no
          predicate is ever generated.)

          In true Common Lisp, `typep' is always able to recognize a
          structure object even if `:predicate' was used.  In this
          package, `typep' simply looks for a function called
          `TYPENAME-p', so it will work for structure types only if
          they used the default predicate name.

    `:include'
          This option implements a very limited form of C++-style
          inheritance.  The argument is the name of another structure
          type previously created with `defstruct'.  The effect is to
          cause the new structure type to inherit all of the included
          structure's slots (plus, of course, any new slots described
          by this struct's slot descriptors).  The new structure is
          considered a "specialization" of the included one.  In fact,
          the predicate and slot accessors for the included type will
          also accept objects of the new type.

          If there are extra arguments to the `:include' option after
          the included-structure name, these options are treated as
          replacement slot descriptors for slots in the included
          structure, possibly with modified default values.  Borrowing
          an example from Steele:

               (defstruct person name (age 0) sex)
                    => person
               (defstruct (astronaut (:include person (age 45)))
                 helmet-size
                 (favorite-beverage 'tang))
                    => astronaut
               
               (setq joe (make-person :name "Joe"))
                    => [cl-struct-person "Joe" 0 nil]
               (setq buzz (make-astronaut :name "Buzz"))
                    => [cl-struct-astronaut "Buzz" 45 nil nil tang]
               
               (list (person-p joe) (person-p buzz))
                    => (t t)
               (list (astronaut-p joe) (astronaut-p buzz))
                    => (nil t)
               
               (person-name buzz)
                    => "Buzz"
               (astronaut-name joe)
                    => error: "astronaut-name accessing a non-astronaut"

          Thus, if `astronaut' is a specialization of `person', then
          every `astronaut' is also a `person' (but not the other way
          around).  Every `astronaut' includes all the slots of a
          `person', plus extra slots that are specific to astronauts.
          Operations that work on people (like `person-name') work on
          astronauts just like other people.

    `:print-function'
          In full Common Lisp, this option allows you to specify a
          function which is called to print an instance of the
          structure type.  The Emacs Lisp system offers no hooks into
          the Lisp printer which would allow for such a feature, so
          this package simply ignores `:print-function'.

    `:type'
          The argument should be one of the symbols `vector' or `list'.
          This tells which underlying Lisp data type should be used to
          implement the new structure type.  Vectors are used by
          default, but `(:type list)' will cause structure objects to
          be stored as lists instead.

          The vector representation for structure objects has the
          advantage that all structure slots can be accessed quickly,
          although creating vectors is a bit slower in Emacs Lisp.
          Lists are easier to create, but take a relatively long time
          accessing the later slots.

    `:named'
          This option, which takes no arguments, causes a
          characteristic "tag" symbol to be stored at the front of the
          structure object.  Using `:type' without also using `:named'
          will result in a structure type stored as plain vectors or
          lists with no identifying features.

          The default, if you don't specify `:type' explicitly, is to
          use named vectors.  Therefore, `:named' is only useful in
          conjunction with `:type'.

               (defstruct (person1) name age sex)
               (defstruct (person2 (:type list) :named) name age sex)
               (defstruct (person3 (:type list)) name age sex)
               
               (setq p1 (make-person1))
                    => [cl-struct-person1 nil nil nil]
               (setq p2 (make-person2))
                    => (person2 nil nil nil)
               (setq p3 (make-person3))
                    => (nil nil nil)
               
               (person1-p p1)
                    => t
               (person2-p p2)
                    => t
               (person3-p p3)
                    => error: function person3-p undefined

          Since unnamed structures don't have tags, `defstruct' is not
          able to make a useful predicate for recognizing them.  Also,
          accessors like `person3-name' will be generated but they will
          not be able to do any type checking.  The `person3-name'
          function, for example, will simply be a synonym for `car' in
          this case.  By contrast, `person2-name' is able to verify
          that its argument is indeed a `person2' object before
          proceeding.

    `:initial-offset'
          The argument must be a nonnegative integer.  It specifies a
          number of slots to be left "empty" at the front of the
          structure.  If the structure is named, the tag appears at the
          specified position in the list or vector; otherwise, the first
          slot appears at that position.  Earlier positions are filled
          with `nil' by the constructors and ignored otherwise.  If the
          type `:include's another type, then `:initial-offset'
          specifies a number of slots to be skipped between the last
          slot of the included type and the first new slot.

   Except as noted, the `defstruct' facility of this package is
entirely compatible with that of Common Lisp.

\1f
File: cl.info,  Node: Assertions,  Next: Efficiency Concerns,  Prev: Structures,  Up: Top

Assertions and Errors
*********************

This section describes two macros that test "assertions", i.e.,
conditions which must be true if the program is operating correctly.
Assertions never add to the behavior of a Lisp program; they simply
make "sanity checks" to make sure everything is as it should be.

   If the optimization property `speed' has been set to 3, and `safety'
is less than 3, then the byte-compiler will optimize away the following
assertions.  Because assertions might be optimized away, it is a bad
idea for them to include side-effects.

 - Special Form: assert test-form [show-args string args...]
     This form verifies that TEST-FORM is true (i.e., evaluates to a
     non-`nil' value).  If so, it returns `nil'.  If the test is not
     satisfied, `assert' signals an error.

     A default error message will be supplied which includes TEST-FORM.
     You can specify a different error message by including a STRING
     argument plus optional extra arguments.  Those arguments are simply
     passed to `error' to signal the error.

     If the optional second argument SHOW-ARGS is `t' instead of `nil',
     then the error message (with or without STRING) will also include
     all non-constant arguments of the top-level FORM.  For example:

          (assert (> x 10) t "x is too small: %d")

     This usage of SHOW-ARGS is a change to Common Lisp.  In true
     Common Lisp, the second argument gives a list of PLACES which can
     be `setf''d by the user before continuing from the error.

 - Special Form: check-type place type &optional string
     This form verifies that PLACE evaluates to a value of type TYPE.
     If so, it returns `nil'.  If not, `check-type' signals a
     continuable `wrong-type-argument' error.  The default error
     message lists the erroneous value along with TYPE and PLACE
     themselves.  If STRING is specified, it is included in the error
     message in place of TYPE.  For example:

          (check-type x (integer 1 *) "a positive integer")

     *Note Type Predicates::, for a description of the type specifiers
     that may be used for TYPE.

     Note that as in Common Lisp, the first argument to `check-type'
     should be a PLACE suitable for use by `setf', because `check-type'
     signals a continuable error that allows the user to modify PLACE,
     most simply by returning a value from the debugger.

   The following error-related macro is also defined:

 - Special Form: ignore-errors forms...
     This executes FORMS exactly like a `progn', except that errors are
     ignored during the FORMS.  More precisely, if an error is
     signalled then `ignore-errors' immediately aborts execution of the
     FORMS and returns `nil'.  If the FORMS complete successfully,
     `ignore-errors' returns the result of the last FORM.

\1f
File: cl.info,  Node: Efficiency Concerns,  Next: Common Lisp Compatibility,  Prev: Assertions,  Up: Top

Efficiency Concerns
*******************

Macros
======

Many of the advanced features of this package, such as `defun*',
`loop', and `setf', are implemented as Lisp macros.  In byte-compiled
code, these complex notations will be expanded into equivalent Lisp
code which is simple and efficient.  For example, the forms

     (incf i n)
     (push x (car p))

are expanded at compile-time to the Lisp forms

     (setq i (+ i n))
     (setcar p (cons x (car p)))

which are the most efficient ways of doing these respective operations
in Lisp.  Thus, there is no performance penalty for using the more
readable `incf' and `push' forms in your compiled code.

   _Interpreted_ code, on the other hand, must expand these macros
every time they are executed.  For this reason it is strongly
recommended that code making heavy use of macros be compiled.  (The
features labelled "Special Form" instead of "Function" in this manual
are macros.)  A loop using `incf' a hundred times will execute
considerably faster if compiled, and will also garbage-collect less
because the macro expansion will not have to be generated, used, and
thrown away a hundred times.

   You can find out how a macro expands by using the `cl-prettyexpand'
function.

 - Function: cl-prettyexpand form &optional full
     This function takes a single Lisp form as an argument and inserts
     a nicely formatted copy of it in the current buffer (which must be
     in Lisp mode so that indentation works properly).  It also expands
     all Lisp macros which appear in the form.  The easiest way to use
     this function is to go to the `*scratch*' buffer and type, say,

          (cl-prettyexpand '(loop for x below 10 collect x))

     and type `C-x C-e' immediately after the closing parenthesis; the
     expansion

          (block nil
            (let* ((x 0)
                   (G1004 nil))
              (while (< x 10)
                (setq G1004 (cons x G1004))
                (setq x (+ x 1)))
              (nreverse G1004)))

     will be inserted into the buffer.  (The `block' macro is expanded
     differently in the interpreter and compiler, so `cl-prettyexpand'
     just leaves it alone.  The temporary variable `G1004' was created
     by `gensym'.)

     If the optional argument FULL is true, then _all_ macros are
     expanded, including `block', `eval-when', and compiler macros.
     Expansion is done as if FORM were a top-level form in a file being
     compiled.  For example,

          (cl-prettyexpand '(pushnew 'x list))
               -| (setq list (adjoin 'x list))
          (cl-prettyexpand '(pushnew 'x list) t)
               -| (setq list (if (memq 'x list) list (cons 'x list)))
          (cl-prettyexpand '(caddr (member* 'a list)) t)
               -| (car (cdr (cdr (memq 'a list))))

     Note that `adjoin', `caddr', and `member*' all have built-in
     compiler macros to optimize them in common cases.


Error Checking
==============

Common Lisp compliance has in general not been sacrificed for the sake
of efficiency.  A few exceptions have been made for cases where
substantial gains were possible at the expense of marginal
incompatibility.  One example is the use of `memq' (which is treated
very efficiently by the byte-compiler) to scan for keyword arguments;
this can become confused in rare cases when keyword symbols are used as
both keywords and data values at once.  This is extremely unlikely to
occur in practical code, and the use of `memq' allows functions with
keyword arguments to be nearly as fast as functions that use
`&optional' arguments.

   The Common Lisp standard (as embodied in Steele's book) uses the
phrase "it is an error if" to indicate a situation which is not
supposed to arise in complying programs; implementations are strongly
encouraged but not required to signal an error in these situations.
This package sometimes omits such error checking in the interest of
compactness and efficiency.  For example, `do' variable specifiers are
supposed to be lists of one, two, or three forms; extra forms are
ignored by this package rather than signalling a syntax error.  The
`endp' function is simply a synonym for `null' in this package.
Functions taking keyword arguments will accept an odd number of
arguments, treating the trailing keyword as if it were followed by the
value `nil'.

   Argument lists (as processed by `defun*' and friends) _are_ checked
rigorously except for the minor point just mentioned; in particular,
keyword arguments are checked for validity, and `&allow-other-keys' and
`:allow-other-keys' are fully implemented.  Keyword validity checking
is slightly time consuming (though not too bad in byte-compiled code);
you can use `&allow-other-keys' to omit this check.  Functions defined
in this package such as `find' and `member*' do check their keyword
arguments for validity.


Optimizing Compiler
===================

The byte-compiler that comes with Emacs 18 normally fails to expand
macros that appear in top-level positions in the file (i.e., outside of
`defun's or other enclosing forms).  This would have disastrous
consequences to programs that used such top-level macros as `defun*',
`eval-when', and `defstruct'.  To work around this problem, the "CL"
package patches the Emacs 18 compiler to expand top-level macros.  This
patch will apply to your own macros, too, if they are used in a
top-level context.  The patch will not harm versions of the Emacs 18
compiler which have already had a similar patch applied, nor will it
affect the optimizing Emacs 19 byte-compiler written by Jamie Zawinski
and Hallvard Furuseth.  The patch is applied to the byte compiler's
code in Emacs' memory, _not_ to the `bytecomp.elc' file stored on disk.

   The Emacs 19 compiler (for Emacs 18) is available from various Emacs
Lisp archive sites such as `archive.cis.ohio-state.edu'.  Its use is
highly recommended; many of the Common Lisp macros emit code which can
be improved by optimization.  In particular, `block's (whether explicit
or implicit in constructs like `defun*' and `loop') carry a fair
run-time penalty; the optimizing compiler removes `block's which are
not actually referenced by `return' or `return-from' inside the block.

\1f
File: cl.info,  Node: Common Lisp Compatibility,  Next: Old CL Compatibility,  Prev: Efficiency Concerns,  Up: Top

Common Lisp Compatibility
*************************

Following is a list of all known incompatibilities between this package
and Common Lisp as documented in Steele (2nd edition).

   Certain function names, such as `member', `assoc', and `floor', were
already taken by (incompatible) Emacs Lisp functions; this package
appends `*' to the names of its Common Lisp versions of these functions.

   The word `defun*' is required instead of `defun' in order to use
extended Common Lisp argument lists in a function.  Likewise,
`defmacro*' and `function*' are versions of those forms which
understand full-featured argument lists.  The `&whole' keyword does not
work in `defmacro' argument lists (except inside recursive argument
lists).

   In order to allow an efficient implementation, keyword arguments use
a slightly cheesy parser which may be confused if a keyword symbol is
passed as the _value_ of another keyword argument.  (Specifically,
`(memq :KEYWORD REST-OF-ARGUMENTS)' is used to scan for `:KEYWORD'
among the supplied keyword arguments.)

   The `eql' and `equal' predicates do not distinguish between IEEE
floating-point plus and minus zero.  The `equalp' predicate has several
differences with Common Lisp; *note Predicates::.

   The `setf' mechanism is entirely compatible, except that
setf-methods return a list of five values rather than five values
directly.  Also, the new "`setf' function" concept (typified by `(defun
(setf foo) ...)') is not implemented.

   The `do-all-symbols' form is the same as `do-symbols' with no
OBARRAY argument.  In Common Lisp, this form would iterate over all
symbols in all packages.  Since Emacs obarrays are not a first-class
package mechanism, there is no way for `do-all-symbols' to locate any
but the default obarray.

   The `loop' macro is complete except that `loop-finish' and type
specifiers are unimplemented.

   The multiple-value return facility treats lists as multiple values,
since Emacs Lisp cannot support multiple return values directly.  The
macros will be compatible with Common Lisp if `values' or `values-list'
is always used to return to a `multiple-value-bind' or other
multiple-value receiver; if `values' is used without
`multiple-value-...' or vice-versa the effect will be different from
Common Lisp.

   Many Common Lisp declarations are ignored, and others match the
Common Lisp standard in concept but not in detail.  For example, local
`special' declarations, which are purely advisory in Emacs Lisp, do not
rigorously obey the scoping rules set down in Steele's book.

   The variable `*gensym-counter*' starts out with a pseudo-random
value rather than with zero.  This is to cope with the fact that
generated symbols become interned when they are written to and loaded
back from a file.

   The `defstruct' facility is compatible, except that structures are
of type `:type vector :named' by default rather than some special,
distinct type.  Also, the `:type' slot option is ignored.

   The second argument of `check-type' is treated differently.

\1f
File: cl.info,  Node: Old CL Compatibility,  Next: Porting Common Lisp,  Prev: Common Lisp Compatibility,  Up: Top

Old CL Compatibility
********************

Following is a list of all known incompatibilities between this package
and the older Quiroz `cl.el' package.

   This package's emulation of multiple return values in functions is
incompatible with that of the older package.  That package attempted to
come as close as possible to true Common Lisp multiple return values;
unfortunately, it could not be 100% reliable and so was prone to
occasional surprises if used freely.  This package uses a simpler
method, namely replacing multiple values with lists of values, which is
more predictable though more noticeably different from Common Lisp.

   The `defkeyword' form and `keywordp' function are not implemented in
this package.

   The `member', `floor', `ceiling', `truncate', `round', `mod', and
`rem' functions are suffixed by `*' in this package to avoid collision
with existing functions in Emacs 18 or Emacs 19.  The older package
simply redefined these functions, overwriting the built-in meanings and
causing serious portability problems with Emacs 19.  (Some more recent
versions of the Quiroz package changed the names to `cl-member', etc.;
this package defines the latter names as aliases for `member*', etc.)

   Certain functions in the old package which were buggy or inconsistent
with the Common Lisp standard are incompatible with the conforming
versions in this package.  For example, `eql' and `member' were
synonyms for `eq' and `memq' in that package, `setf' failed to preserve
correct order of evaluation of its arguments, etc.

   Finally, unlike the older package, this package is careful to prefix
all of its internal names with `cl-'.  Except for a few functions which
are explicitly defined as additional features (such as `floatp-safe'
and `letf'), this package does not export any non-`cl-' symbols which
are not also part of Common Lisp.


The `cl-compat' package
=======================

The "CL" package includes emulations of some features of the old
`cl.el', in the form of a compatibility package `cl-compat'.  To use
it, put `(require 'cl-compat)' in your program.

   The old package defined a number of internal routines without `cl-'
prefixes or other annotations.  Call to these routines may have crept
into existing Lisp code.  `cl-compat' provides emulations of the
following internal routines: `pair-with-newsyms', `zip-lists',
`unzip-lists', `reassemble-arglists', `duplicate-symbols-p',
`safe-idiv'.

   Some `setf' forms translated into calls to internal functions that
user code might call directly.  The functions `setnth', `setnthcdr',
and `setelt' fall in this category; they are defined by `cl-compat',
but the best fix is to change to use `setf' properly.

   The `cl-compat' file defines the keyword functions `keywordp',
`keyword-of', and `defkeyword', which are not defined by the new "CL"
package because the use of keywords as data is discouraged.

   The `build-klist' mechanism for parsing keyword arguments is
emulated by `cl-compat'; the `with-keyword-args' macro is not, however,
and in any case it's best to change to use the more natural keyword
argument processing offered by `defun*'.

   Multiple return values are treated differently by the two Common
Lisp packages.  The old package's method was more compatible with true
Common Lisp, though it used heuristics that caused it to report
spurious multiple return values in certain cases.  The `cl-compat'
package defines a set of multiple-value macros that are compatible with
the old CL package; again, they are heuristic in nature, but they are
guaranteed to work in any case where the old package's macros worked.
To avoid name collision with the "official" multiple-value facilities,
the ones in `cl-compat' have capitalized names:  `Values',
`Values-list', `Multiple-value-bind', etc.

   The functions `cl-floor', `cl-ceiling', `cl-truncate', and
`cl-round' are defined by `cl-compat' to use the old-style
multiple-value mechanism, just as they did in the old package.  The
newer `floor*' and friends return their two results in a list rather
than as multiple values.  Note that older versions of the old package
used the unadorned names `floor', `ceiling', etc.; `cl-compat' cannot
use these names because they conflict with Emacs 19 built-ins.

\1f
File: cl.info,  Node: Porting Common Lisp,  Next: Function Index,  Prev: Old CL Compatibility,  Up: Top

Porting Common Lisp
*******************

This package is meant to be used as an extension to Emacs Lisp, not as
an Emacs implementation of true Common Lisp.  Some of the remaining
differences between Emacs Lisp and Common Lisp make it difficult to
port large Common Lisp applications to Emacs.  For one, some of the
features in this package are not fully compliant with ANSI or Steele;
*note Common Lisp Compatibility::.  But there are also quite a few
features that this package does not provide at all.  Here are some
major omissions that you will want watch out for when bringing Common
Lisp code into Emacs.

   * Case-insensitivity.  Symbols in Common Lisp are case-insensitive
     by default.  Some programs refer to a function or variable as
     `foo' in one place and `Foo' or `FOO' in another.  Emacs Lisp will
     treat these as three distinct symbols.

     Some Common Lisp code is written in all upper-case.  While Emacs
     is happy to let the program's own functions and variables use this
     convention, calls to Lisp builtins like `if' and `defun' will have
     to be changed to lower-case.

   * Lexical scoping.  In Common Lisp, function arguments and `let'
     bindings apply only to references physically within their bodies
     (or within macro expansions in their bodies).  Emacs Lisp, by
     contrast, uses "dynamic scoping" wherein a binding to a variable
     is visible even inside functions called from the body.

     Variables in Common Lisp can be made dynamically scoped by
     declaring them `special' or using `defvar'.  In Emacs Lisp it is
     as if all variables were declared `special'.

     Often you can use code that was written for lexical scoping even
     in a dynamically scoped Lisp, but not always.  Here is an example
     of a Common Lisp code fragment that would fail in Emacs Lisp:

          (defun map-odd-elements (func list)
            (loop for x in list
                  for flag = t then (not flag)
                  collect (if flag x (funcall func x))))
          
          (defun add-odd-elements (list x)
            (map-odd-elements (function (lambda (a) (+ a x))) list))

     In Common Lisp, the two functions' usages of `x' are completely
     independent.  In Emacs Lisp, the binding to `x' made by
     `add-odd-elements' will have been hidden by the binding in
     `map-odd-elements' by the time the `(+ a x)' function is called.

     (This package avoids such problems in its own mapping functions by
     using names like `cl-x' instead of `x' internally; as long as you
     don't use the `cl-' prefix for your own variables no collision can
     occur.)

     *Note Lexical Bindings::, for a description of the `lexical-let'
     form which establishes a Common Lisp-style lexical binding, and
     some examples of how it differs from Emacs' regular `let'.

   * Common Lisp allows the shorthand `#'x' to stand for `(function
     x)', just as `'x' stands for `(quote x)'.  In Common Lisp, one
     traditionally uses `#'' notation when referring to the name of a
     function.  In Emacs Lisp, it works just as well to use a regular
     quote:

          (loop for x in y by #'cddr collect (mapcar #'plusp x)) ; Common Lisp
          (loop for x in y by 'cddr collect (mapcar 'plusp x))   ; Emacs Lisp

     When `#'' introduces a `lambda' form, it is best to write out
     `(function ...)' longhand in Emacs Lisp.  You can use a regular
     quote, but then the byte-compiler won't know that the `lambda'
     expression is code that can be compiled.

          (mapcar #'(lambda (x) (* x 2)) list)            ; Common Lisp
          (mapcar (function (lambda (x) (* x 2))) list)   ; Emacs Lisp

     XEmacs supports `#'' notation starting with version 19.8.

   * Reader macros.  Common Lisp includes a second type of macro that
     works at the level of individual characters.  For example, Common
     Lisp implements the quote notation by a reader macro called `'',
     whereas Emacs Lisp's parser just treats quote as a special case.
     Some Lisp packages use reader macros to create special syntaxes
     for themselves, which the Emacs parser is incapable of reading.

   * Other syntactic features.  Common Lisp provides a number of
     notations beginning with `#' that the Emacs Lisp parser won't
     understand.  For example, `#| ... |#' is an alternate comment
     notation, and `#+lucid (foo)' tells the parser to ignore the
     `(foo)' except in Lucid Common Lisp.

     The number prefixes `#b', `#o', and `#x', however, are supported
     by the Emacs Lisp parser to represent numbers in binary, octal,
     and hexadecimal notation (or radix), just like in Common Lisp.

   * Packages.  In Common Lisp, symbols are divided into "packages".
     Symbols that are Lisp built-ins are typically stored in one
     package; symbols that are vendor extensions are put in another,
     and each application program would have a package for its own
     symbols.  Certain symbols are "exported" by a package and others
     are internal; certain packages "use" or import the exported symbols
     of other packages.  To access symbols that would not normally be
     visible due to this importing and exporting, Common Lisp provides
     a syntax like `package:symbol' or `package::symbol'.

     Emacs Lisp has a single namespace for all interned symbols, and
     then uses a naming convention of putting a prefix like `cl-' in
     front of the name.  Some Emacs packages adopt the Common Lisp-like
     convention of using `cl:' or `cl::' as the prefix.  However, the
     Emacs parser does not understand colons and just treats them as
     part of the symbol name.  Thus, while `mapcar' and `lisp:mapcar'
     may refer to the same symbol in Common Lisp, they are totally
     distinct in Emacs Lisp.  Common Lisp programs which refer to a
     symbol by the full name sometimes and the short name other times
     will not port cleanly to Emacs.

     Emacs Lisp does have a concept of "obarrays," which are
     package-like collections of symbols, but this feature is not
     strong enough to be used as a true package mechanism.

   * Keywords.  The notation `:test-not' in Common Lisp really is a
     shorthand for `keyword:test-not'; keywords are just symbols in a
     built-in `keyword' package with the special property that all its
     symbols are automatically self-evaluating.  Common Lisp programs
     often use keywords liberally to avoid having to use quotes.

     In Emacs Lisp a keyword is just a symbol whose name begins with a
     colon; since the Emacs parser does not treat them specially, they
     have to be explicitly made self-evaluating by a statement like
     `(setq :test-not ':test-not)'.  This package arranges to execute
     such a statement whenever `defun*' or some other form sees a
     keyword being used as an argument.  Common Lisp code that assumes
     that a symbol `:mumble' will be self-evaluating even though it was
     never introduced by a `defun*' will have to be fixed.

   * The `format' function is quite different between Common Lisp and
     Emacs Lisp.  It takes an additional "destination" argument before
     the format string.  A destination of `nil' means to format to a
     string as in Emacs Lisp; a destination of `t' means to write to
     the terminal (similar to `message' in Emacs).  Also, format
     control strings are utterly different; `~' is used instead of `%'
     to introduce format codes, and the set of available codes is much
     richer.  There are no notations like `\n' for string literals;
     instead, `format' is used with the "newline" format code, `~%'.
     More advanced formatting codes provide such features as paragraph
     filling, case conversion, and even loops and conditionals.

     While it would have been possible to implement most of Common Lisp
     `format' in this package (under the name `format*', of course), it
     was not deemed worthwhile.  It would have required a huge amount
     of code to implement even a decent subset of `format*', yet the
     functionality it would provide over Emacs Lisp's `format' would
     rarely be useful.

   * Vector constants use square brackets in Emacs Lisp, but `#(a b c)'
     notation in Common Lisp.  To further complicate matters, Emacs 19
     introduces its own `#(' notation for something entirely
     different--strings with properties.

   * Characters are distinct from integers in Common Lisp.  The
     notation for character constants is also different:  `#\A' instead
     of `?A'.  Also, `string=' and `string-equal' are synonyms in Emacs
     Lisp whereas the latter is case-insensitive in Common Lisp.

   * Data types.  Some Common Lisp data types do not exist in Emacs
     Lisp.  Rational numbers and complex numbers are not present, nor
     are large integers (all integers are "fixnums").  All arrays are
     one-dimensional.  There are no readtables or pathnames; streams
     are a set of existing data types rather than a new data type of
     their own.  Hash tables, random-states, structures, and packages
     (obarrays) are built from Lisp vectors or lists rather than being
     distinct types.

   * The Common Lisp Object System (CLOS) is not implemented, nor is
     the Common Lisp Condition System.

   * Common Lisp features that are completely redundant with Emacs Lisp
     features of a different name generally have not been implemented.
     For example, Common Lisp writes `defconstant' where Emacs Lisp
     uses `defconst'.  Similarly, `make-list' takes its arguments in
     different ways in the two Lisps but does exactly the same thing,
     so this package has not bothered to implement a Common Lisp-style
     `make-list'.

   * A few more notable Common Lisp features not included in this
     package:  `compiler-let', `tagbody', `prog', `ldb/dpb',
     `parse-integer', `cerror'.

   * Recursion.  While recursion works in Emacs Lisp just like it does
     in Common Lisp, various details of the Emacs Lisp system and
     compiler make recursion much less efficient than it is in most
     Lisps.  Some schools of thought prefer to use recursion in Lisp
     over other techniques; they would sum a list of numbers using
     something like

          (defun sum-list (list)
            (if list
                (+ (car list) (sum-list (cdr list)))
              0))

     where a more iteratively-minded programmer might write one of
     these forms:

          (let ((total 0)) (dolist (x my-list) (incf total x)) total)
          (loop for x in my-list sum x)

     While this would be mainly a stylistic choice in most Common Lisps,
     in Emacs Lisp you should be aware that the iterative forms are
     much faster than recursion.  Also, Lisp programmers will want to
     note that the current Emacs Lisp compiler does not optimize tail
     recursion.