March 2012 - alexandria-devel - mailman3.common-lisp.net

[alexandria-devel] latest alexandria doesn't build on ECL due to non-standard LOOP
by Anton Vodonosov 18 Jan '18

18 Jan '18

Hello. Alexandria in the latest quicklisp has a regression comparing to quicklisp 2012-02-08. It doesn't build on ECL because on non-standard LOOP construct (it is tolerated by most Lisps, but not by ECL; see here: http://comments.gmane.org/gmane.lisp.ecl.general/3344) The compiler complains about copy-stream method in io.lisp. Build logs: https://cl-test-grid.appspot.com/blob?key=AMIfv97xRRWJgLOa5TvS78VriRK9mQ82m… Best regards, - Anton

2 1

[alexandria-devel] Question about FORMAT-SYMBOL
by Luís Oliveira 26 Jan '13

26 Jan '13

Hello, I often try to support Allegro's modern mode by doing things like (format-symbol t "~A-~A" :foo :bar). We just came across an issue on #quicklisp where a user had set his *PRINT-CASE* to :DOWNCASE and that caused FORMAT-SYMBOL to intern lower-case symbols, thus breaking the idiom I described. Should fix my code to use SYMBOLICATE instead or should FORMAT-SYMBOL use WITH-STANDARD-IO-SYNTAX? Any opinions? -- Luís Oliveira http://r42.eu/~luis

5 8

[alexandria-devel] broken LOOP forms
by Greg Pfeil 30 Mar '12

30 Mar '12

Alexandria (QL version) doesn't compile on ECL and warns on CLISP because invalid LOOP forms are used in (at least) COPY-STREAM. EG: (loop :while (< input-position start) :for bytes-read = (read-sequence buffer input :end (min (length buffer) (- start input-position))) :do (progn (when (zerop bytes-read) (error "Could not read enough bytes from the input to fulfill the START requirement in ~S" 'copy-stream)) (incf input-position bytes-read))) And I _think_ according to http://www.lispworks.com/documentation/HyperSpec/Body/m_loop.htm, the :WHILE expression can't precede the :FOR expression.

2 1

[alexandria-devel] destructuring-case
by Jasper 30 Mar '12

30 Mar '12

Hello all, The `destructuring-case` macro imo is a little different than I would think it would be upon just hearing the name. The thing is, i'd have made the clauses to basically the `destructuring-bind` form, except that a clause is executed if the 'nested lists-shape' matches, and all non-symbols, keywords match. I'd add `quote` as special, something that also must match exactly, and maybe `lambda`/`function` to indicate a function that must return true.(which now i think of it is an idea for the next section.) (shameless plug)Maybe the reason is that i wrote a `destructuring-regex` https://github.com/o-jasper/j-basic/blob/master/src/destructuring-regex.lisp of which the idea is that the variables can simply be written on the matches/space between the matches.(Further each lisp object can add a method giving the regular expression for it, and also provide a parser) It uses the regex package though, is cl-ppcre more commonly used as dependency? But basically, i have a `regex-case` for it. Them being regexes and either matching or not make it obvious how that should work! Looking at the git log it was added Sep23 2010, so i am probably too late to change `destructuring-case`? I dont know, i see there are no official releases, and it might sneakily be that many people are using the link from the cliki page!(from 2008) - Ending a hopefully not too inane post, Jasper

2 1

[alexandria-devel] #'shuffle docstring is misleading
by Orivej Desh 30 Mar '12

30 Mar '12

#'shuffle docstring states that “Permuted sequence may share storage with the original one.” Actually it destructively modifies the sequence. In the case of vectors the two might be the same, but in the case of lists I would say that something like (sort (copy-list l) …) shares storage with the original list but (sort l …) does more than that.

2 1

[alexandria-devel] reverting patch
by Tamas Papp 30 Mar '12

30 Mar '12

Hi, I have been trying to reach Nikodemus Siivola for a while but didn't get a response, so I thought I would write here hoping that someone can help me. Nikodemus admitted my "generic MEAN and MEDIAN, new function DISPLACE-ARRAY" patch to Alexandria 11 days ago. This was meant to be an initial patch, followed by others, with the final goal of making mean, median and variance generic. The first patch dealt with MEAN and MEDIAN. I would like to ask that this patch is removed from Alexandria for the time being, for the following two reasons: 1. The generic variance I wrote would not support the :BIASED keyword, in fact, only "unbiased" variance makes sense in certain situations. Since :BIASED T is the default, this would break existing code in a silent manner, which is not desirable. 2. The new function DISPLACE-ARRAY is badly designed: the keyword defaults handle what turns out to be a rarely used corner case for me. I would rather keep the old functionality, already present in CL-NUM-UTILS. I am sorry for the confusion. I think it is best if I experiment within my own libraries for a while before re-addressing these issues. I wonder if someone could revert the patch, preferably soon, so that the next release of Quicklisp would contain Alexandria without my code. Best, Tamas

2 1

[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