gmane.comp.lang.racket.user

Re: Why are some Racket functions such a 'member' designedas partially boolean?

Matthias Felleisen — 2012-05-25T16:57:08

In some Racket languages they are: 



Note the '?'. 

In untyped languages, conditionals count every value other than #f as a true value. So there isn't a real need to return #t. Better still, in some cases it is much better to return an additional result value -- in member's case it would have been '(2 3) above -- because this value is useful and avoids a recompilation. 

Example: 
  
 (define (take-step game-state)
   (if (move-available game-state)
       (step state (determine-best-move state))
       (give-up)))

It all likelihood there is a tremendous overlap between move-available? and determine-best-move so why not rewrite it like this:

 (define (take-step game-state)
   (define good-move (determine-best-move state))
   (if good-move (step state good-move) (give-up)))







On May 25, 2012, at 12:29 PM, Don Green wrote:


____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: complex sort or how to sort lines

Pierpaolo Bernardi — 2012-05-25T16:35:02

(define (sort-line line)
  (sort line
        (λ (a b)
          (or (null? a)
              (and (not (null? b))
                   (string<? (first a) (first b)))))))

____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Why are some Racket functions such a 'member' designed aspartially boolean?

Don Green — 2012-05-25T16:29:40

Why are some Racket functions such a 'member' not designed as booleans?
I would expect 'member' to be a boolean.
I am inclined to write my own version of 'member' that is boolean.
Before I do, I am wondering if the bright person who designed the Racket
function: 'member', did so for some very good reason that is not obvious to
me.
I notice that there are a number of other Racket functions like this, where
they return #t or the value of the argument passed to the function, not #t
or #f.
I am just looking for the concept so that I might be convinced that I
should use the racket function as it is, rather than writing a boolean
version.
Thanks
____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: complex sort or how to sort lines

Matthias Felleisen — 2012-05-25T16:29:10


On May 25, 2012, at 12:16 PM, Don Green wrote:






#lang racket

(require 2htdp/batch-io rackunit)

;; --- answer to your question: 
(define (answer file)
  (sort (read-words/line file) string<=? #:key first))

;; --- test set up
(define content
  #<< eos
This is a silly sample file. 
Do you think it is long enough? 
Or should I add a line? 
 eos
  )

(define expected
  #<< eos
Do you think it is long enough?
Or should I add a line?
This is a silly sample file.

 eos
  )

(define file (path->string (make-temporary-file "tst~a")))
(with-output-to-file file
  #:exists 'replace
  (lambda () 
    (for ((ch content)) (display ch))))

;; --- test
(check-equal? (with-output-to-string
               (lambda ()
                 (for ((ln (answer file)))
                   (displayln (string-join ln)))))
              expected)
 
____________________
  Racket Users list:
  http://lists.racket-lang.org/users

complex sort or how to sort lines

Don Green — 2012-05-25T16:16:31

Given the output of Racket's read-words/line function, which produces a
list of lists. Each word is in a list, each group of words on a line is in
a list, all within a list.

Any ideas on how to sort on the car (first element) of each list that
corresponds to a line?

In other words, I want to sort lines based on the first word on a line.  I
want to keep the lines intact.
Thanks.
____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: Second display monitor issue

Kieron Hardy — 2012-05-25T15:26:58

Thanks for the info Robby - dunno how I missed finding those files ... oh
wait yes I do - Windows Search - doh!

On Thu, May 24, 2012 at 6:32 PM, Robby Findler
<robby-5YcgHWA4rVnfxgxxWyf0drK6bRJNSFrb< at >public.gmane.org>wrote:

____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: [racket-dev] `string-split'

Pierpaolo Bernardi — 2012-05-25T11:55:09


This order looks much better because you use a literal string in the
function call, which is an artificial example.

In real code the string would be the value of a variable or the return
value of a function call, and your example would not look much
different whatever the order of the arguments.

And which of the two arguments is more interesting, varies from case
to case, IMO.

Cheers
P.

____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: Serializable struts and the quoted module path

Matthew Flatt — 2012-05-25T11:53:53

The intent is that you use a collection, so that the serialized form
would use a collection path instead of an absolute path.

That is, you'd create a new collection for your library, possibly using
`raco link' to register it, and then refer to the library (that
contains the serializable structure-type declaration) through a
collection-based path instead of a file path.

To make `serialize' more flexible, it might make sense for us to change
`serialize' so that it's sensitive to a parameter such as
`current-write-relative-directory'.

At Thu, 24 May 2012 18:14:20 -0600, Michael Wilber wrote:
____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: Bindings for libgsl (GNU Scientific Library)

Jens Axel Søgaard — 2012-05-25T07:56:22

Thanks for testing this. It was very helpful.

Using your program I got the library working on Racket 5.2.1.6 but
not on 5.3.0.6.

In mzgsl/low-level/gsl-lib.rkt I had to make sure CBLAS was loaded
in GLOBAL mode.

That is, in https://github.com/noelwelsh/mzgsl/blob/master/low-level/gsl-lib.ss
I changed

    (define libgslcblas (ffi-lib cblaslibpath))

to

    (define libgslcblas (ffi-lib cblaslibpath #:global? #t))

Now I get this transcript using your test program:

$ /Applications/Racket\ Full\ v5.2.1.6/bin/plt-r6rs proof.sps
#(2.0 2.0 2.0)
2.0
3.0
2.0

$ /Applications/Racket\ Full\ v5.3.0.6/bin/plt-r6rs proof.sps
Segmentation fault: 11

Note that prior to this I reinstalled GSL with this command:

    sudo port install gsl +universal

whether this made any change I do not know.

/Jens Axel
____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: Bindings for libgsl (GNU Scientific Library)

Marco Maggi — 2012-05-25T05:36:13



Cannot help with  OS X, but on my  GNU+Linux 32-bit with GSL
1.15 and Racket 5.1.2 the following works fine:

  $ git clone http://github.com/noelwelsh/mzgsl.git
  $ export PLTCOLLECTS=$PLTCOLLECTS:$PWD
  $ cat proof.sps
  #!r6rs
  (import (rnrs)
    (mzgsl gslvector))
  (define p (make-gslvector 3))
  (gslvector-fill! p 2)
  (display (gslvector->vector p))
  (newline)
  (gslvector-set! p 1 3.0)
  (do ((i 0 (+ 1 i)))
      ((= i 3))
    (display (gslvector-ref p i))
    (newline))

  $ plt-r6rs proof.sps
  #(2.0 2.0 2.0)
  2.0
  3.0
  2.0

HTH

Re: Plot: plot-new-window never happens?

Neil Toronto — 2012-05-25T04:17:53

I just pushed a fix that turns your so-called laziness into correct 
behavior. :) The original loop will work as you intended it.

Neil ⊥

____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: Plot: plot-new-window never happens?

Neil Toronto — 2012-05-25T03:55:54

Never mind. I was calling `super-new' before the canvas instance was 
initialized, which put the canvas in the frame, so the frame forwarded 
events to the canvas.

Neil ⊥

On 05/25/2012 12:44 PM, Neil Toronto wrote:

____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: Plot: plot-new-window never happens?

Neil Toronto — 2012-05-25T03:44:27

Looks like that mostly works. Here's a problem I ran into: it seems 
that, unless I'm careful in a way I don't understand, the children of a 
frame with a new eventspace receive callbacks before they're fully 
initialized. Here's a simple example:

#lang racket

(require racket/gui)

(define test-editor-canvas%
   (class editor-canvas%
     (init parent)
     (super-new [parent parent])

     (define/override (on-size w h)
       (printf "~v~n" should-this-be-defined?))

     (define should-this-be-defined? #f)))

(define frame
   (let ([new-es (make-eventspace)])
     (parameterize ([current-eventspace new-es])
       (define frame (new frame% [label "bob"] [width 400] [height 400]))
       (new test-editor-canvas% [parent frame])
       frame)))

(send frame show #t)


Most of the time, this prints #<undefined> instead of #f, indicating 
that the test-editor-canvas% instance isn't fully constructed before its 
`on-size' is called. If I move its construction outside the 
`current-eventspace' parameterizatio

Re: Plot: #:samples parameter of function---is this what was meant?

Neil Toronto — 2012-05-25T02:16:51

Quick reminder: your worry was spot-on, and I was going to fix the problem.

I've just pushed a fix. This test program used to sample the function 
`f' about 500,000 times:

#lang racket
(require plot plot/utils)

(define count 0)
(define (f x)
   (set! count (+ count 1))
   (sin x))

(define xs (linear-seq -4 4 501))  ; bounds for 500 function renderers
(time
  (plot (for/list ([x1  (in-list xs)]
                   [x2  (in-list (rest xs))])
          (function f x1 x2))))
count


It now samples `f' only 1000 times: twice for each renderer's endpoints.

You might need to be a bit careful rendering piecewise functions with 
many function renderers. If the pieces are drawn partially transparent 
and a bit thick, the endpoints will visibly overlap, giving the function 
a "knotted" look. Here's an example:

#lang racket
(require plot plot/utils)

(define xs (linear-seq -4 4 50))
(plot (for/list ([x1  (in-list xs)]
                  [x2  (in-list (rest xs))])
         (function sin x1 x2 #:width 5 #:alpha 1/2))

Re: Plot woes

Neil Toronto — 2012-05-25T01:55:36

I just pushed a fix for this. The following now works as you'd expect:

#lang racket

(require plot)

(plot (list (function (compose / exact->inexact) -1 0)
             (function (compose / exact->inexact) 0 1)))

(plot (list (function / -1 -0.01)
             (function / 0.01 1)))

Neil ⊥
____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: Second display monitor issue

Robby Findler — 2012-05-25T00:32:06

I've pretty much no clue how to fix it, but the code you see is in
collects/mred/private/wx/win32/frame.rkt. The display-origin function
(it is called from mred/private/wxtop.rkt).

hth,
Robby

On Thu, May 24, 2012 at 7:08 PM, Kieron Hardy <kieron.hardy< at >gmail.com> wrote:

____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: Rationale behind missing string-trim

Neil Toronto — 2012-05-25T00:29:17

Thanks, Eli. I've wanted some of those functions a few times, and had to 
write my own. Yours are most likely more correct and faster. :)

Neil ⊥
____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Serializable struts and the quoted module path

Michael Wilber — 2012-05-25T00:14:20

Hey there!

So I'm trying to use racket/serialize according to
http://docs.racket-lang.org/reference/serialization.html?q=serialize#(def._((lib._racket/private/serialize..rkt)._deserialize))

According to the third bullet point, custom serializable structure types
include a "quoted module path" pointing to a module that provides the
serialized structure's serialization info. It seems to be a full
absolute filesystem path in my tests.

Problem is, I want to send serializable structs across the network to a
machine where the module path is different.

Is there a sane way to, uh, strip off the absolute parts in that path? I
could just mess around with the serialized info myself, but... oy,
that just feels nasty.

(If anyone's curious, my ultimate goal is to send serializable lambdas
to other machines. Just like Distributed Places, but not as conceptually
confusing and constraining.)
http://blog.racket-lang.org/2009/06/serializable-closures-in-plt-scheme.html

Thanks in advance.
____________________
  Racket Us

Second display monitor issue

Kieron Hardy — 2012-05-25T00:08:07

Hi all,

On one of my client's laptops (Windows 7), the attached code detects the
second display monitor (get-display-count returns 2), but prints exactly
the same values for the location and size of the two displays. i.e. The
top-left coordinate and size values returned ((0,0)  and 1024 x 768)  are
the same for both monitors. This effectively maps the second monitor
underneath(?) the main monitor, making it inaccessible to Racket.

In this particular case the second monitor is a TV connected through the
laptop's s-video port. Windows itself sees the two displays and is
configured to extend the desktop over both of them and 'Identify monitors'
works as expected. At least one other program can see and use the TV
without problems.

Racket has no problems seeing and using a second display monitor on this
same laptop when that second monitor is a video converter box connected
through the laptop's VGA port. Also Racket has no problems seeing and using
a TV as a second display monitor when connected to my (Vista)

Re: module browser

Matthias Felleisen — 2012-05-24T21:00:26

I wonder whether an approximate tool would be just fine for now, especially if it listed things that it didn't understand. 



On May 23, 2012, at 4:04 PM, Ryan Culpepper wrote:



____________________
  Racket Users list:
  http://lists.racket-lang.org/users

Re: [racket-dev] `string-split'

Eli Barzilay — 2012-05-24T20:51:03

It's close to the order that you get in many languages where the
string is the main subject, especially in the OO cases where it
becomes str.split(";").



That's sometimes true, but changing these things is a huge hassle.
(And it's not clear that one order is better than the other.)
Note that `regexp-split' does take the regexp first -- it's more
fitting there since "regexp" is the obvious part of the name where
here we have "string" instead.



Yeah, currying is certainly cute, but it would make for a very
different language.  (As a sidenote, one aspect of currying cuteness
is that decisions on argument order can in many cases (like the above)
have an obvious choice.  Not in all cases though, otherwise haskell
wouldn't have a `flip'.)