Why, implement code to turn text upside-down in your favorite editor, of course!

Thank you.

The McCLIM listener in a cocoa window, running on the latest Clozure CL beta.

It took some time, since on OS X 10.5 CGFloat can be either single-float or double-float, depending on the processor. At the moment, I just rewired the hardwired (coerce … ’single-float) calls and numeric constants in the beagle backend. Also, it crashes on the all-important “Show Class Subclasses t” benchmark. Still, very nice.

Parenscript advertises itself as being easy to integrate with javascript libraries such as Prototype, but the manual gives no examples on how to do this. Furthermore, the Parenscript manual uses aserve / htmlgen for its examples, while much of the action these days is with hunchentoot. So, without further ado, a pointless example integrating hunchentoot, cl-who, parenscript and script.aculo.us. Load the code after loading hunchentoot, cl-who and parenscript into your running Lisp.

;;; Just the standard intro
(defpackage :js-hax
  (:use :common-lisp :hunchentoot :cl-who :parenscript))
(in-package :js-hax)
(start-server :port 8080)

;;; Set the path to where you unpacked script.aculo.us
;;; create-folder-dispatcher-and-handler publishes subdirectories:
;;; http://.../js/lib/prototype.js
;;; http://.../js/src/scriptaculous.js
(push (create-folder-dispatcher-and-handler
       "/js/" #p"/home/someone/source/scriptaculous-js-1.8.0/")
      *dispatch-table*)

(defmacro js-script (&rest body)
  "Utility macro for including ParenScript into the HTML
   notation of cl-who"
  `(with-html-output (*standard-output*)
     (:script :type "text/javascript"
              (format t "~%// <!--[CDATA[~%~A~%//]]-->~%" (ps ,@body)))))

;;; Slider example, taken from the scriptaculous wiki
(defun foo ()
  (with-html-output-to-string (*standard-output* nil :prologue t)
    (:html
     (:head (:title "script.aculo.us Tests")
            (:script :language "JavaScript" :type "text/javascript"
                     :src "/js/lib/prototype.js" "stupid firefox")
            (:script :language "JavaScript" :type "text/javascript"
                     :src "/js/src/scriptaculous.js" "stupid firefox"))
     (:body
      (:h1 "script.aculo.us vs. ParenScript")
      (:p "The humble slider.")
      (:p
       (:div :id "track1"
             :style "width:200px;background-color:#aaa;height:5px;"
             (:div
              :id "handle1"
              :style "width:5px;height:10px;background-color:#f00;cursor:move;"))
       (:div :id "debug1" :style "padding-top: 5px;"))
      (js-script
       ;; Create the Slider object
       (new (*control.*slider
             ;; Hook it up to the UI elements ...
             "handle1" "track1"
             ;; ... and assign behavior
             (create on-slide (lambda (v)
                                ;; Prototype-style DOM accessors:
                                ;; $('debug1').innerHTML = ...
                                (setf (slot-value ($ "debug1")
                                                  'inner-h-t-m-l)
                                      (+ "slide: " v)))
                     on-change (lambda (v)
                                 (setf (slot-value ($ "debug1")
                                                   'inner-h-t-m-l)
                                       (+ "changed! " v)))))))))))

(push (create-prefix-dispatcher "/foo" 'foo) *dispatch-table*)

All in all quite painless. The only point where I had to sit and think a bit was the translation of the prototype-style DOM accessors to parenscript notation.

Edit: Oh, and the js-script code above is missing the CDATA bits which were eaten by the web. Check page 6 of the parenscript documentation for the missing pieces.

Edit2: Thanks to Aankhen for pointing out that the CDATA wasn’t et after all, just hiding. I’ll go wrestle with this blog now to fix the stupid smartquotes.

Leopard sandboxes are flexible and interesting. They’re apparently compiled from Scheme programs (sandbox-compilerd embeds TinyScheme) that live in /usr/share/sandbox. You can break sandbox-compilerd open in TextEdit and read the compiled-in Scheme code; they’ve got a lot of the bases covered, including obscure stuff like SYSV IPC, the BSD sysctl interface, and signals. 

http://www.matasano.com/log/981/a-roundup-of-leopard-security-features/

I’m not angry that I will not get the books, of course – if I had not heard about that special offer at all, I would not have anything to be angry about either. But when I said “vanished”, I mean it: Amazon dropped two out of three books from my order, without notifying me, and would have sent me a full-price book overseas; I would have paid at least twice the store price. And since I did not save their order confirmation email, I cannot even prove that the entire thing happened …

[Update] Well, better late than never – I got a notification email explaining that it was all a dream that the books had been displayed with an incorrect price.  I guess they simply ran out of stock, since the people who ordered before the madness started did get their LiSP at Small Prices.  Still, a notification email plus 10$ gift certificate is close to the right thing.

The only inconsistency I discovered was with Undo: Aquamacs uses redo.el, which removes the emacs undo function and installs more familiar, but less powerful undo/redo functions in its place. Since redo.el clobbers the original undo function, I couldn’t adjust the keybindings myself without hacking Aquamacs distributed files, so I sent a whiney bug report.

24 hours and a short email back-and-forth later (choice quote: “You are the first to complain, but that doesn’t mean you’re not right.”), the nightly builds don’t clobber undo any more, C-/ and C-_ use emacs undo and undo/redo is available via Cmd-Z/Cmd-S-Z and the pull-down menu.

This is a good example of what I understand to be the general design philosophy of Aquamacs: offer emacs functions where an emacs user expects to find them, offer Mac-like functionality where a Mac user looks.  And its developer, David Reitter, is very responsive and friendly indeed.