- alexandria-devel - mailman3.common-lisp.net

[alexandria-devel] [PATCH] fix docstring typos in numbers.lisp
by Tobias C. Rittweiler 11 Mar '08

11 Mar '08

--- old-alexandria/numbers.lisp 2008-03-10 17:42:17.000000000 +0100 +++ new-alexandria/numbers.lisp 2008-03-10 17:42:17.000000000 +0100 @@ -44,7 +44,7 @@ (defun iota (n &key (start 0) (step 1)) "Return a list of n numbers, starting from START (with numeric contagion from STEP applied), each consequtive number being the sum of the previous one -and STEP. START defaults to 0 and STEP to 0. +and STEP. START defaults to 0 and STEP to 1. Examples: @@ -62,13 +62,14 @@ (defun map-iota (function n &key (start 0) (step 1)) "Calls FUNCTION with N numbers, starting from START (with numeric contagion from STEP applied), each consequtive number being the sum of the previous one -and STEP. START defaults to 0 and STEP to 0. Returns N. +and STEP. START defaults to 0 and STEP to 1. Returns N. Examples: - (iota 4) => (0 1 2 3 4) - (iota 3 :start 1 :step 1.0) => (1.0 2.0 3.0) - (iota 3 :start -1 :step -1/2) => (-1 -3/2 -2) + (map-iota #'print 3 :start 1 :step 1.0) => 3 + ;;; 1.0 + ;;; 2.0 + ;;; 3.0 " (declare (type (integer 0) n) (number start step)) (loop repeat n

2 1

[alexandria-devel] [PATCH] improvement wrt. signalling errors on improper sequences
by Tobias C. Rittweiler 11 Mar '08

11 Mar '08

The attached patch improves error signalling when passing improper lists to some of the sequence function where the docstring said that they'd signal an error. Additionally, introduce the macro DO-PROPER-LIST, and the function PROPER-LIST-LENGTH -- which are used to remove some code duplication. -T.

1 1

[alexandria-devel] ENSUREF
by Tobias C. Rittweiler 10 Mar '08

10 Mar '08

ENSUREF can be used to initialize &key and &optional key args more conveniently. Another name would be ORF, which looks funny to my eyes. Another name would be INITIALIZEF, but this might open questions regarding its relation to the object initialization machinery. -T. (defmacro ensuref (place thing &rest more &environment env) "If PLACE evaluates to NIL, set it to THING. Otherwise, let it unchanged. Examples: (let ((a nil) (b :foo) (c (list 1 2 nil 4))) (ensuref a :a b :b (third c) :c) (values a b c)) ==> :A, :FOO, (1 2 :C 4) Notes: (ENSUREF PLACE THING) is approximately equivalent to (SETF PLACE (OR PLACE THING)) except that ENSUREF ensures (no pun indented) that PLACE is evaluated only once, and except that PLACE isn't set at all if it evaluated to NIL. (E.g. as a result, auxiliary methods of (setf foo) generic functions aren't run in that case.) " (multiple-value-bind (gvars vals gstorevars setter getter) (get-setf-expansion place env) (when (second gstorevars) (error "ENSUREF does not support setting multiple values ~ via the (VALUES ...) place.")) (if (null more) (let ((gstorevar (first gstorevars)) (gtmp (gensym "CURVAL+"))) `(let ,(mapcar #'list gvars vals) (let ((,gtmp ,getter)) (if ,gtmp ,gtmp (let ((,gstorevar ,thing)) ,setter))))) `(progn (ensuref ,place ,thing) (ensuref ,@more)))))

1 0

[alexandria-devel] [PATCH] DEFINE-CONSTANT
by Tobias C. Rittweiler 10 Mar '08

10 Mar '08

The DEFINE-CONSTANT that comes with Alexandria causes "deleting unreachable code" style-warnings in SBCL; I couldn't understand what caused them, so Nikodemus as group-local SBCL expert may want to take a look at it; either SBCL is way smarter than me, or there's a bug hiding somewhere. Anyway, my patch moves most of the macroexpansion of DEFINE-CONSTANT to its own function which is prettier anyway. This also makes the warnings go away. -T.

2 1

[alexandria-devel] DX declaration in COMPOSE &c
by Tobias C. Rittweiler 10 Mar '08

10 Mar '08

;;; Isn't the DYNAMIC-EXTENT declaration wrong on ;;; ;;; (flet ((id (&rest args) args)) ;;; (compose #'id #'id)) ;;; ;;; ? (defun compose (function &rest more-functions) "Returns a function composed of FUNCTION and MORE-FUNCTIONS that applies its arguments to to each in turn, starting from the rightmost of MORE-FUNCTIONS, and then calling the next one with the primary value of the last." (declare (optimize (speed 3) (safety 1) (debug 1))) (reduce (lambda (f g) (lambda (&rest arguments) (declare (dynamic-extent arguments)) (funcall f (apply g arguments)))) more-functions :initial-value function)) Similiar for MULTIPLE-VALUE-COMPOSE, CURRY, and RCURRY. -T.

1 0

[alexandria-devel] arnesi:with-collector
by Kilian Sprotte 08 Mar '08

08 Mar '08

Hi, I am just wondering whether the above might be a good candidate to be included with alexandria? Best, Kilian PS: Its here http://common-lisp.net/project/bese/docs/arnesi/html/Reducing_0020and_0020C…

1 0

[alexandria-devel] [PATCH] remove-from-plistf
by Michael Weber 08 Mar '08

08 Mar '08

Hi, remove-from-plistf is show as having arglist (at least with SBCL, similarly with OpenMCL): (remove-from-plistf place plist &rest keys) which is wrong. Attached patch fixes it. Cheers, Michael

2 1

[alexandria-devel] [PATCH] fix featurep
by Stelian Ionescu 01 Mar '08

01 Mar '08

The attached patch makes FEATUREP accept any compound test specifier, not just the keywords :AND, :OR and :NOT . -- Stelian Ionescu a.k.a. fe[nl]ix Quidquid latine dictum sit, altum videtur.

2 1

[alexandria-devel] REMOVEF and DELETEF
by Chun Tian (binghe) 23 Feb '08

23 Feb '08

Hi, Alexandria Developers I saw this two modify macro in sequences.lisp: (define-modify-macro removef (item &rest remove-keywords) (lambda (seq item &rest keyword-arguments) (apply #'remove item seq keyword-arguments)) "Modify-macro for REMOVE. Sets place designated by the first argument to the result of calling REMOVE with ITEM, place, and the REMOVE- KEYWORDS.") (define-modify-macro deletef (item &rest remove-keywords) (lambda (seq item &rest keyword-arguments) (apply #'delete item seq keyword-arguments)) "Modify-macro for DELETE. Sets place designated by the first argument to the result of calling DELETE with ITEM, place, and the REMOVE- KEYWORDS.") It seems that the use of DEFINE-MODIFY-MACRO doesn't match the ANSI CL standard. HyperSpec[1] says the 3rd argument of DEFINE-MODIFY-MACRO must be a symbol, not a lambda expression. A new version of LispWorks (still under test) will cannot compile these two definitions: ANSI CL requires the argument to be a symbol, so I think ARNESI is wrong. We changed this in LW 5.1. --- Martin Simmons from LispWorks, Ltd. Hope you can fix this, thanks. Chun TIAN (binghe) [1] http://www.lispworks.com/documentation/HyperSpec/Body/m_defi_2.htm#define-m…

2 2

[alexandria-devel] [PATCH] add MEMQ
by Stelian Ionescu 11 Feb '08

11 Feb '08

Hello, the attached patch adds MEMQ. I think that MEMQ is popular and useful enough that it should be part of Alexandria. -- Stelian Ionescu a.k.a. fe[nl]ix Quidquid latine dictum sit, altum videtur.

3 3