defining setf-expanders in Common Lisp - common-lisp

Here's the thing: I don't "get" setf-expanders and would like to learn how they work.
I need to learn how they work because I've got a problem which seems like a typical example for why you should learn setf-expanders, the problem is as follows:
(defparameter some-array (make-array 10))
(defun arr-index (index-string)
(aref some-array (parse-integer index-string))
(setf (arr-index "2") 7) ;; Error: undefined function (setf arr-index)
How do I write a proper setf-expander for ARR-INDEX?

(defun (setf arr-index) (new-value index-string)
(setf (aref some-array (parse-integer index-string))
new-value))
In Common Lisp a function name can not only be a symbol, but also a list of two symbols with SETF as the first symbol. See above. DEFUN thus can define SETF functions. The name of the function is (setf arr-index).
A setf function can be used in a place form: CLHS: Other compound forms as places.
The new value is the first argument then.
CL-USER 15 > some-array
#(NIL NIL NIL NIL NIL NIL NIL NIL NIL NIL)
CL-USER 16 > (setf (arr-index "2") 7)
7
CL-USER 17 > some-array
#(NIL NIL 7 NIL NIL NIL NIL NIL NIL NIL)

Rainer's answer is spot on. Before ANSI Common Lisp, it was necessary to use defsetf to define an expander for simple places that could be set with a simple function call. setf functions like (setf arr-index) came into the language with CLOS and simplify a lot of things. In particular, setf functions can be generic.

Related

Controlling the printing of special cons forms (e.g printing (function +) as #'+ etc)

I want some reader macros to print as as shortened expression that the macro understands. Lets say I want to extend the #' macro to take #'~[rest-of-symbol] and turn that into (complement #'rest-of-symbol).
What controls how that is printed? On SBCL, for instance, '(function +) prints as #'+. How do i make '(complement #'listp) print as #~listp?
My first thought was
(defmethod print-object :around ((obj cons) stream)
;; if #'~fn-name / (complement (function fn-name))
;; => fn-name otherwise NIL
(let ((fn-name
(ignore-errors
(destructuring-bind (complement (function fn-name))
obj
(when (and (eq complement 'complement)
(eq function 'function))
fn-name)))))
(if fn-name
(format stream "#'~~~S" fn-name)
(call-next-method))))
This works insofar as (print-object '(complement #'evenp) *standard-output*) prints it the way I want, but the REPL doesn't. Also (print-object '#'+ *standard-output*) prints it as (function +) so the REPL isn't using print-object. With defining the print-object method for user defined classes the REPL always picks up on the new definition.
This is my first post and I'm sorry I can't get the code to format properly. If someone can put a link on how to do that I would appreciate it.
Evaluation time
You are mixing code with data in your example:
(function +)
Is a special form that evaluates to a function object, which admits a shorter syntax:
#'+
But when you are writing:
'(function +)
or
'(complement fn)
Then in both cases you are writing quoted, literal lists, which evaluates to themselves (namely a list starting with symbol function or complement, followed respectively by symbol + and fn).
However, you want the code to be evaluated at runtime to actual function objects; if you type this in the REPL:
(complement #'alpha-char-p)
The result is a value that is printed as follows:
#<FUNCTION (LAMBDA (&REST SB-IMPL::ARGUMENTS) :IN COMPLEMENT) {101AAC8D9B}>
You have an actual function object that you can funcall. In other words, by the time you reach print-object, you no longer have access to source code, you are manipulating data at runtime which happens to be functions. So you cannot use destructuring-bind to get the complement symbol that was present in the source code.
What you need to do instead is to attach metadata to your function. There is a way to do that in Common Lisp by defining a new type of function, thanks to the Meta-Object Protocol.
Funcallable objects
I'm relying on Closer-MOP for all the symbols prefixed with c2cl: below. I define a new class of functions, annotated-fn, which is a function with addditional data:
(defclass annotated-fn (c2cl:funcallable-standard-object)
((data :initform :data :initarg :data :reader annotated-fn-data))
(:metaclass c2cl:funcallable-standard-class))
Notice that this class is a funcallable-standard-object (like the usual functions), and its metaclass is funcallable-standard-class. Such an object has an additional implicit slot that is a function to call.
More precisely, you have to call c2cl:set-funcallable-instance-function to set a function associated with the object, and when later you use funcall or apply with the object, then the wrapped function is called instead. So you can transparently use this class of functions wherever you usually use a function. It just has additional slots (here data).
For example, here is how I instantiate it, with a function to wrap additional data:
(defun annotate-fn (function data)
(let ((object (make-instance 'annotated-fn :data data)))
(prog1 object
(c2cl:set-funcallable-instance-function object function))))
Let's try it:
(describe
(annotate-fn (constantly 3)
'(:category :constantly)))
#<ANNOTATED-FN {1006275C7B}>
[funcallable-instance]
Lambda-list: UNKNOWN
Derived type: FUNCTION
Documentation:
T
Source file: SYS:SRC;CODE;FUNUTILS.LISP
Slots with :INSTANCE allocation:
DATA = (:CATEGORY :CONSTANTLY)
You can also use this object like any other function.
Now, your reader macros can expand into calls to annotate-fn, and add any kind of additional metadata you need to the function.
Reader macro
For our example, imagine you define a reader macros for constant functions:
(set-macro-character #\[ 'read-constantly t)
(set-macro-character #\] (get-macro-character #\) nil))
(defun read-constantly (stream char)
(declare (ignore char))
(let* ((list (read-delimited-list #\] stream t))
(value (if (rest list) list (first list)))
(var (gensym)))
`(let ((,var ,value))
(annotate-fn (constantly ,var)
(list :category :constantly
:constant ,var)))))
Using this syntax:
> [(+ 8 5)]
=> #<ANNOTATED-FN ...>
By the way, the syntax I defined also allows the following:
> [+ 8 5]
Pretty-printing
Let's define a generic function that prints an annotated function given its :category field:
(defgeneric print-for-category (category data object stream))
(defmethod print-object ((o annotated-fn) s)
(let* ((data (annotated-fn-data o))
(category (getf data :category)))
(print-for-category category data o s)))
Then, we can specialize it for :constantly, and here we assume also that the data associated with the function contains a :constant field:
(defmethod print-for-category ((_ (eql :constantly)) data o s)
(format s "[~s]" (getf data :constant)))
For example:
(let ((value (+ 8 6)))
(annotate-fn (constantly value)
`(:constant ,value
:category :constantly)))
This above is printed as:
[14]
Which would be the same as your hypothetical reader macro.
To do this you need to understand the pretty printer. I have understood it in the past but no longer do completely. It dispatches on type and the trick for things like this is that you can specify very specific types for trees of conses, although doing so is verbose.
Here is an example which is almost certainly not completely correct, but does achieve what you want in this case:
(defparameter *ppd* (copy-pprint-dispatch))
(defun pprint-complement-function (s form)
;; This is the thing that the pretty printer will call. It can
;; assume that the form it wants to print is already correct.
(destructuring-bind (complement (function name)) form
(declare (ignore complement function))
(format s "#'~~~W" name)))
;;; Now set this in the table with a suitable hairy type specification
;;;
(set-pprint-dispatch '(cons (eql complement)
(cons (cons (eql function)
(cons t null))
null))
'pprint-complement-function
0
*ppd*)
And now
> (let ((*print-pprint-dispatch* *ppd*))
(pprint '(complement (function foo)))
(pprint '((complement (function foo)) (function foo))))
#'~foo
(#'~foo #'foo)
You can make the awful nested cons type specifier easier by defining this (which, perhaps, should be the compound type specifier for list except you can't do that):
(deftype list-of-types (&rest types)
(labels ((lot (tt)
(if (null tt)
'null
`(cons ,(first tt) ,(lot (rest tt))))))
(lot types)))
And then
(set-pprint-dispatch '(list-of-types (eql complement)
(list-of-types (eql function)
*))
'pprint-complement-function
0
*ppd*)
is perhaps easier to read.

Defining a type on a list that begins with a particular symbol

I am trying to use generic functions' ability to specify behaviour based on the first argument of a list.
In other words, I want the list (atypelist 1 2 3) and the list (btypelist 1 2 3) to have their individual behaviour when passed to foo. So far, this is what I came up with:
(deftype atypelist (lst)
`(eq (car ,lst) 'atypelist))
(deftype btypelist (lst)
`(eq (car ,lst) 'btypelist))
(defmethod foo ((lst atypelist))
(format nil "success atypelist: ~S" lst))
(defmethod foo ((lst btypelist))
(format nil "success btypelist: ~S" lst))
However, when I call (typep (list 'atypelist 1 2 3) 'atypelist) I get the following error:
error while parsing arguments to DEFTYPE ATYPELIST:
too few elements in
()
to satisfy lambda list
(LST):
exactly 1 expected, but got 0
I am guessing the error is in my definition of atypelist.
Questions:
Is there a better way to get the functionality I am looking for?
If yes - what is the way?
If not - how to properly define a type on a list/cons that has a particular symbol in the car?
Before I start: what you want to do can't work, and is confused in two ways.
Firstly deftype defines a type in terms of other type specifiers: the body of a deftype form must expand into a type specifier, not an expression, as yours does. And deftype's arguments are not the thing you want to check the type for, they are parts of the type specification.
In this case you want to specify that the thing is a cons, and that its car is eql to something. Fortunately there are specializing type specifiers for both of these things, and you end up with something like this:
(deftype cons-with-specified-car (x)
`(cons (eql ,x) t))
And now
> (typep '(1) '(cons-with-specified-car 1))
t
> (typep '(a) '(cons-with-specified-car a))
t
> (typep '() '(cons-with-specified-car a))
nil
And if you want:
(deftype cons-with-a ()
'(cons-with-specified-car a))
and now
> (typep '(a) 'cons-with-a)
t
Secondly none of this will work because this it not how CLOS works. CLOS dispatches on classes not types, and you have merely defined a type, not a class: your method definitions simply cannot work, since classes cannot be parametrized in this way like types can.
Some ways you might achieve what you want.
If what you want to do is to dispatch on the first element of a list, then the obvious approach, if you want to use CLOS, is to use a two-level approach where you first dispatch on the class of the thing (cons is a class), and then use eql specializers to pick out the things you want.
(defgeneric select (it)
(:method ((it cons))
(select* (car it) it))
(:method (it)
nil))
(defgeneric select* (key it)
(:method (key it)
(format t "~&unknown key ~S in ~S~%" key it)))
(defmethod select* ((key (eql 'a)) it)
(format t "~&~S begins with a~%" it))
However in a case like this, unless you very much want the extensibility that CLOS gets you (which is a good reason to use CLOS here), I'd just use typecase. You could do this using the type defined above:
(defun select (it)
(typecase it
((cons-with-specified-car a)
'(cons a))
(cons
'cons)
(t
nil)))
or, probably simpler, just use what the deftype expands into:
(defun select (it)
(typecase it
((cons (eql a) t)
'(cons a))
(cons
'cons)
(t
nil)))
Finally probably what anyone doing this would actually write (again, assuming you do not want the extensibility CLOS gets you) is:
(defun select (it)
(typecase it
(cons
(case (car it)
...))
(t
...)))
Here is a possible solution, using the type specifier satisfies:
CL-USER> (defun is-atypelist (list)
(eq (car list) 'atypelist))
IS-ATYPELIST
CL-USER> (defun is-btypelist (list)
(eq (car list) 'btypelist))
IS-BTYPELIST
CL-USER> (deftype atypelist ()
`(satisfies is-atypelist))
ATYPELIST
CL-USER> (deftype btypelist ()
`(satisfies is-btypelist))
BTYPELIST
CL-USER> (typep (list 'atypelist 1 2 3) 'atypelist)
T
CL-USER> (typep (list 'atypelist 1 2 3) 'btypelist)
NIL
Note that this does not define a class, but a type, if this is what you need.
Is there a better way to get the functionality I am looking for?
1. Wrap your lists in container types
(defclass lst () ((items :initarg :items :reader items)))
(defclass alst (lst) ())
(defclass blst (lst) ())
It may be a little bit more cumbersome to work with but this is pretty much straightforward and not too suprising.
2. Douple-dispatch
(defgeneric foo (val))
(defgeneric foo/tag (tag val))
For example:
(defmethod foo ((c cons))
(destructuring-bind (tag . list) c
(foo/tag tag list)))
3. Define a custom method combination
It should be possible to hack the meta-object protocol dispatch mechanism to dispatch on the first item of a list. I wouldn't recommend it however.
4. Use a different dispatch mechanism
Use a completely different dispatching mechanism outside of CLOS, like pprint-dispatch does. For example you may want to use trivia or optima pattern-matching libraries, or cl-algebraic-data-type. This may be more useful if you are dealing with trees of symbols.

Using a quoted list of slots for Peter Seibel's defclass macro

I am using the defclass macro from Practical Common Lisp, which takes as argument a list of symbols.
I would like to change the macro in order to let it accept a quoted list of symbols. This has the benefit of having thenm defined as constants that can be used in other convenience functions, e.g. here. I confused myself in trying to get this done.
My use case is following:
(defconstant state-slots '(unit motion mode moc-offset woc-pos woc-inc feed spindle))
;; would like to use the quoted list here:
(defclass-by-slots gc-state (unit ; :mm(=G21) :inch(=G20)
motor ; nil :on
motion ; :jog(=G0) :lin(=G1) :cw(=G2) :ccw(=G3)
mode ; :abs(=G90) :inc(=G91)
moc-offset ; woc-zero(xyz, mm) from moc-zero
woc-pos ; woc-pos(xyz, mm) from woc-zero
woc-inc
feed
spindle))
;; can use quoted slot list when using a convenience function, e.g:
(defun format-by-slots (o slots &optional str-type)
(let* ((f (lambda (s$) (eval (format-slot o s$))))
(str-type (string-upcase str-type))
(r (concatenate
'string
(format nil "~A (~A)" o (class-of o))
(reduce (lambda (s1 s2) (concatenate 'string s1 s2))
(loop for s in slots
when (funcall f s) collect it)
:from-end t :initial-value (format nil "~%")))))
(if str-type
(ppcre:regex-replace-all
(format nil "^#<~A \\{(\\d|[A-F])+\\}> " str-type)
r
(format nil "#<~A {...}> " str-type))
r)))
I am using this for several classes defined by different slots.
The nuisance is that I cannot have defined the slots uniformly for the type definition and the convenience functions which is source of annoying errors.
Solution based on [Rainer Joswig's answer] (https://stackoverflow.com/a/61154538/2336738):
(defmacro def-my-class (name supers slots-symbol)
"The value of a symbol of slots-symbol is used as the
list of slots."
`(defclass ,name ,supers
,(if (and (symbolp slots-symbol)
(symbol-value slots-symbol)
(listp (symbol-value slots-symbol)))
(mapcar #'slot->defclass-slot (symbol-value slots-symbol))
(error "~a is not a symbol which names a list of slot names" slots-symbol))))
Symbol values at compile time
If you want to use the value of a symbol at compile time, then you need to make sure that it is defined. Two usual ways to do that:
define the symbol in a file and load it before compiling another file where it is used
use EVAL-WHEN to execute a symbol definition (defvar, defparameter,defconstant` ...) at compile time
Read-time evaluation
One possibility to use a symbol value during compilation is to use read-time evaluation. You would need to make sure that the constant value of +state-slots+ is defined during compilation:
(eval-when (:compile-toplevel :load-toplevel :execute)
(defconstant +state-slots+
'(unit motion mode moc-offset woc-pos woc-inc feed spindle)))
(defclass foo ()
#.+state-slots+)
Custom Macro
If the value of +state-slots+ is defined at compile time, then we can also use it in a macro:
(defmacro def-my-class (name supers slots-symbol)
"The value of a symbol of slots-symbol is used as the
list of slots."
`(defclass ,name ,supers
,(if (and (symbolp slots-symbol)
(symbol-value slots-symbol)
(listp (symbol-value slots-symbol)))
(symbol-value slots-symbol)
(error "~a is not a symbol which names a list of slot names"))))
(def-my-class foo () +state-slots+)

common-lisp higher-order logical or function

I often have a truth-list like the following '(nil nil nil t nil t nil nil nil) and I would like to call (reduce #'or truth-list)
However, this does not work and I found that or is a macro. Is there a quick and easy way I can get this to work like a function? What I have been doing is passing: (lambda (p q) (or p q)) as my function, but since I have come across this so many times, I bet there is a better way.
Thanks for all the help!
Not really. You actually need the function. Just define a BINARY-OR function. If you use it often, then just add it to your code.
Alternatives:
(some #'identity '(nil nil nil t nil t nil nil nil))
or
(loop for i in '(nil nil nil t nil t nil nil nil) thereis i)
Bonus: both above forms will stop at the first true value. The reduce variant won't.

Common lisp: Redefine an existing function within a scope?

In Common Lisp, is it possible to redefine an already defined function within a certain scope? For example, given a function A that calls a function B. Can I temporarily redefine B during a call to A?
I'm looking for something along the lines of a let block, but that can redefine functions.
Within a given lexical scope, yes. Use FLET or LABELS. Any function defined with FLET will be unable to call functions defined in the same lexical scope, if you want that (for, say, self-recursive of a group of mutually recursive functions), you will need to use LABELS.
Note that both FLET and LABELS only establish lexical shadowing, should not be used to shadow functions from the COMMON-LISP package and will not dynamically change what function is called from outside the lexical scope the form establishes.
Local functions can be introduced with FLET and LABELS.
If you want to redefine/shadow an existing function using dynamic scope, this is a macro I've been using for a while.
(defmacro! with-shadow ((fname fun) &body body)
"Shadow the function named fname with fun
Any call to fname within body will use fun, instead of the default function for fname.
This macro is intentionally unhygienic:
fun-orig is the anaphor, and can be used in body to access the shadowed function"
`(let ((fun-orig))
(cond ((fboundp ',fname)
(setf fun-orig (symbol-function ',fname))
(setf (symbol-function ',fname) ,fun)
(unwind-protect (progn ,#body)
(setf (symbol-function ',fname) fun-orig)))
(t
(setf (symbol-function ',fname) ,fun)
(unwind-protect (progn ,#body)
(fmakunbound ',fname))))))
Usage:
Clozure Common Lisp Version 1.9-r15759 (DarwinX8664) Port: 4005 Pid: 4728
; SWANK 2012-03-06
CL-USER>
(defun print-using-another-fname (x)
(print x))
PRINT-USING-ANOTHER-FNAME
CL-USER>
(let ((*warn-if-redefine-kernel* nil))
(with-shadow (print (lambda (x)
(funcall fun-orig (+ x 5))))
(print-using-another-fname 10)))
15
15
CL-USER>
(print 10)
10
10
CL-USER>
Note that it relies on Doug Hoyte's defmacro! macro, available in Let Over Lambda.
Also as written, it's anaphoric (fun-orig is available within the body). If you want it completely hygienic, just change the fun-orig's to ,g!fun-orig's.
I most often redefine functions when writing unit tests. Mocking functions within the scope of a particular unit test is helpful, and sometimes that needs to be done with dynamic (not lexical) scope.
You can simulate dynamic-binding for funs like this:
(defmacro setvfun (symbol function)
`(progn
(setf ,symbol ,function)
(setf (symbol-function ',symbol) (lambda (&rest args) (apply (symbol-value ',symbol) args)))))
and then ,for example, with
(setvfun some-fun (lambda() (format t "initial-definition~%")))
(defun test-the-fun (&rest args) (apply #'some-fun args))
(defun test ()
(test-the-fun)
(flet ((some-fun () (format t "Lexically REDEFINED (if you see this, something is very wrong)~%")))
(test-the-fun))
(let ((some-fun (lambda (x) (format t "Dynamically REDEFINED with args: ~a~%" x))))
(declare (special some-fun))
(test-the-fun "Hello"))
(test-the-fun))
you get:
REPL> (test)
==>initial-definition
==>initial-definition
==>Dynamically REDEFINED with args: Hello
==>initial-definition

Resources