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

[alexandria-devel] ALEXANDRIA:COPY-STREAM broken
by Zach Beane 07 Mar '12

07 Mar '12

ALEXANDRIA:COPY-STREAM currently calls (< start end) even when END is NIL. This breaks COPY-FILE unconditionally and COPY-STREAM whenever END (or, less likely, START) is NIL. Zach

2 1

[alexandria-devel] PARSE-BODY printing its argument
by Gustavo Henrique Milaré 02 Mar '12

02 Mar '12

Hi, For some reason, whenever I use the function PARSE-BODY, the argument BODY that is passed to it is printed (tested in SBCL and Clisp). Clisp: [4]> (alexandria.0.dev:parse-body '((declare (optimize speed)) (+ x y))) (DECLARE (OPTIMIZE SPEED)) (+ X Y) ((+ X Y)) ; ((DECLARE (OPTIMIZE SPEED))) ; NIL SBCL: * (parse-body '((declare (optimize speed)) (+ x y))) (DECLARE (OPTIMIZE SPEED)) (+ X Y) ((+ X Y)) ((DECLARE (OPTIMIZE SPEED))) NIL I'm using Alexandria of Quicklisp.

2 1

Re: [alexandria-devel] 9 new unexpected test failures on CLISP
by Anton Vodonosov 14 Feb '12

14 Feb '12

Good it's fixed 14.02.2012, 08:41, "James M. Lawrence" <llmjjmll(a)gmail.com>: > I am unable to post to alexandria-devel due to spam rejections. (Even > my message to the admin was rejected.) So FYI here's what I tried to > send -- > > ---------- Forwarded message ---------- > From: James M. Lawrence <llmjjmll(a)gmail.com> > Date: Mon, Feb 13, 2012 at 10:48 PM > Subject: Re: [alexandria-devel] 9 new unexpected test failures on CLISP > To: alexandria-devel(a)common-lisp.net > > On Mon, Feb 13, 2012 at 12:05 PM, Anton Vodonosov <avodonosov(a)yandex.ru> wrote: > >> We will see. Opened a CLISP ticket: >> https://sourceforge.net/tracker/?func=detail&aid=3487338&group_id=1355&atid… > > It's now fixed: > http://clisp.hg.sourceforge.net/hgweb/clisp/clisp/rev/5bcad87b2d40 > > I ran alexandria-tests with a patched clisp -- no unexpected failures.

1 0

[alexandria-devel] 9 new unexpected test failures on CLISP
by Anton Vodonosov 13 Feb '12

13 Feb '12

Hello. The alexandria version in the recent quicklisp has 9 unexpected failures on CLISP: ALEXANDRIA-TESTS::MEAN.1, ALEXANDRIA-TESTS::MEAN.2, ALEXANDRIA-TESTS::MEAN.3, ALEXANDRIA-TESTS::MEAN.4, ALEXANDRIA-TESTS::MEDIAN.1, ALEXANDRIA-TESTS::MEDIAN.2, ALEXANDRIA-TESTS::MEDIAN.3, ALEXANDRIA-TESTS::VARIANCE.1, ALEXANDRIA-TESTS::STANDARD-DEVIATION.1. The library test log: https://cl-test-grid.appspot.com/blob?key=AMIfv960NH7TxHvJO5_RSARVN0mmN_31J…

2 2

[alexandria-devel] patch for generic variance and standard-deviation
by Tamas Papp 23 Jan '12

23 Jan '12

2 1

[alexandria-devel] generic mean, variance, median, standard-deviation
by Tamas Papp 21 Jan '12

21 Jan '12

Hi, Would it be possible to make MEAN, VARIANCE, MEDIAN, and STANDARD-DEVIATION generic? The concepts are meaningful for objects other than sequences --- including arrays and random-variables (see cl-random), sequences pre-sorted for easy quantile calculation, etc. Currently I am shadowing these and redefining them as generic in some of my libraries. I am happy to implement the change and send a patch. Best, Tamas

2 4

[alexandria-devel] [PATCH] lazy-evaluate the default value in ENSURE-GETHASH
by Ryan Davis 20 Jan '12

20 Jan '12

I'm using a hash-table as a cache, and had a problem with ensure-gethash. Some example code / output: CL-USER> (let ((h (make-hash-table))) (alexandria:ensure-gethash :x h (progn (format T "expensive computation~%") 1)) (alexandria:ensure-gethash :x h (progn (format T "expensive computation~%") 1))) expensive computation expensive computation 1 T CL-USER> This patch (created using git format-patch origin/master) converts ENSURE-GETHASH to a macro that only evaluates the default if we get to the setf branch. Has a check to create and call a lambda if the default value looks complicated (using CONSTANTP). After applying my patch, here's the output: CL-USER> (let ((h (make-hash-table))) (alexandria:ensure-gethash :x h (progn (format T "expensive computation~%") 1)) (alexandria:ensure-gethash :x h (progn (format T "expensive computation~%") 1))) expensive computation 1 T CL-USER> Thoughts? Thanks, -- Ryan Davis Acceleration.net Director of Programming Services 2831 NW 41st street, suite B Gainesville, FL 32606 Office: 352-335-6500 x 124 Fax: 352-335-6506

4 5

[alexandria-devel] [PATCH] Fixes to make Alexandria work with :invert readtable case
by Jyrki Jaakkola 20 Jan '12

20 Jan '12

Hi, I have included a patch to make Alexandria work with :invert readtable case. Fixed also the tests to work with :invert readtable case. Best regards, Jyrki Jaakkola

2 1

[alexandria-devel] Bug in shuffle (sequences.lisp)
by Jianshi Huang 09 Nov '11

09 Nov '11

(loop for i from (- end 1) downto start do (rotatef (aref sequence i) (aref sequence (random (+ i 1))))) should be (loop for i from start below end do (rotatef (aref sequence i) (aref sequence (+ i (random (- end i)))))) otherwise the element selected won't be bound by [start, end), similar bug in the sequence typed version Cheers, -- 黄　澗石 (Jianshi Huang) http://huangjs.net/

2 2

[alexandria-devel] A fix for `curry'
by James M. Lawrence 03 Nov '11

03 Nov '11

Failing test: ========== diff --git a/tests.lisp b/tests.lisp index a4a8e55..8aa6730 100644 --- a/tests.lisp +++ b/tests.lisp @@ -508,6 +508,16 @@ (funcall fun 2))) 4) +(deftest curry.4 + (let ((curried (let ((x 1)) + (curry (progn + (incf x) + (lambda (y z) (* x y z))) + 3)))) + (list (funcall curried 7) + (funcall curried 7))) + (42 42)) + (deftest rcurry.1 (let ((r (rcurry '/ 2))) (funcall r 8)) ========== Simplest fix: ========== diff --git a/functions.lisp b/functions.lisp index 15032be..a2eb1d5 100644 --- a/functions.lisp +++ b/functions.lisp @@ -121,10 +121,12 @@ it is called with to FUNCTION." (define-compiler-macro curry (function &rest arguments) (let ((curries (make-gensym-list (length arguments) "CURRY"))) - `(let ,(mapcar #'list curries arguments) - (declare (optimize (speed 3) (safety 1) (debug 1))) - (lambda (&rest more) - (apply ,function ,@curries more))))) + (with-gensyms (fn) + `(let ((,fn (ensure-function ,function)) + ,@(mapcar #'list curries arguments)) + (declare (optimize (speed 3) (safety 1) (debug 1))) + (lambda (&rest more) + (apply ,fn ,@curries more)))))) (defun rcurry (function &rest arguments) "Returns a function that applies the arguments it is called ========== Alternate fix: Since CURRY has some optimizations, we could follow suit by avoiding an unnecessary binding. I think a function designator which looks like (function ...) or (quote ...) needs no binding, and that those are the only binding-free possibilities. I have not proved this, however. ========== diff --git a/functions.lisp b/functions.lisp index 15032be..5096456 100644 --- a/functions.lisp +++ b/functions.lisp @@ -119,13 +119,21 @@ it is called with to FUNCTION." ;; Using M-V-C we don't need to append the arguments. (multiple-value-call fn (values-list arguments) (values-list more))))) -(define-compiler-macro curry (function &rest arguments) +(defmacro curry-helper (function &rest arguments) (let ((curries (make-gensym-list (length arguments) "CURRY"))) `(let ,(mapcar #'list curries arguments) (declare (optimize (speed 3) (safety 1) (debug 1))) (lambda (&rest more) (apply ,function ,@curries more))))) +(define-compiler-macro curry (function &rest arguments) + (if (and (consp function) + (member (first function) '(function quote))) + `(curry-helper ,function ,@arguments) + (with-gensyms (fn) + `(let ((,fn (ensure-function ,function))) + (curry-helper ,fn ,@arguments))))) + (defun rcurry (function &rest arguments) "Returns a function that applies the arguments it is called with and ARGUMENTS to FUNCTION." ==========

2 2