Lisp and software architecture
in [Lisp]
|
Prev: cosx - sinx + 3x2
Next: the Modernization of Emacs
From: bradb on 11 Mar 2006 19:35 Ah, I wasn't aware you could get a tarball like that. Which is funny because I've used that view cvs interface a bit before, I guess you very quickly filter stuff out :) Brad
From: Tayssir John Gabbour on 11 Mar 2006 21:21 bradb wrote: > No luck with the demos, sorry. Now the function CV is undefined. Does anyone else have problems with DEF-C-OUTPUT? Its OLD-VALUE always reporting NIL to me... but that doesn't make sense; I expect it to hold whatever the last slot value was. Adapted from cells-test/hello-world.test (I tried both latest CVS Cells and the tarball, with SBCL PPC Darwin 0.9.9): (use-package :cells) (use-package :utils-kt) (defmodel computer () ((happen :cell :ephemeral :initform (c-in nil) :accessor happen) (location :cell t :initform (c? (case (^happen) (:leave :away) (:arrive :at-home) (t .cache))) ;; ie, unchanged :accessor location) (response :cell :ephemeral :initform nil :initarg :response :accessor response))) (macrolet ((fmt (name) `(format t "~%~A: ~15A (previously ~:[an unbound ~A~;~A~])" ,name new-value old-value-boundp old-value))) (def-c-output response () (fmt "response")) (def-c-output happen () (fmt "happen ")) (def-c-output location () (fmt "location"))) (defun hello-world () (let ((dell (make-be 'computer :response (c? (bwhen (h (happen self)) (if (eql (^location) :at-home) (case h (:knock-knock "who's there?") (:world "hello, world.")) "<silence>")))))) (dotimes (n 2) (setf (happen dell) :knock-knock)) (setf (happen dell) :arrive) (setf (happen dell) :knock-knock) (setf (happen dell) :world) (values))) CL-USER> (hello-world) happen : NIL (previously an unbound NIL) location: NIL (previously NIL) response: NIL (previously NIL) happen : KNOCK-KNOCK (previously NIL) response: <silence> (previously NIL) happen : KNOCK-KNOCK (previously NIL) response: <silence> (previously NIL) happen : ARRIVE (previously NIL) location: AT-HOME (previously NIL) happen : KNOCK-KNOCK (previously NIL) response: who's there? (previously NIL) happen : WORLD (previously NIL) response: hello, world. (previously NIL) ; No value
From: Ken Tilton on 11 Mar 2006 22:33 Tayssir John Gabbour wrote: > bradb wrote: > >>No luck with the demos, sorry. Now the function CV is undefined. > > > Does anyone else have problems with DEF-C-OUTPUT? Its OLD-VALUE always > reporting NIL to me... That just happens to be the case, partly because it is a silly example, mostly because most of the slots are ephemeral, which are actually not the norm in real world programming. The silly example has the location start out as neither :at-home nor :away, but instead nil. Hell, I'll go so far as to call that a bug since that is not a sensible location. I should probably have it be :at-home if nothing has yet transpired, but the example wants the system to be away for the first few knocks. And so it goes. :) if you add (setf (happen dell) :leave) you should see the location output go from :at-home to :away. As for ephemerals, oddly enough, Cells being good at modelling long-lived state created a problem: how does one model events, which happen but then are no more? Ask on the Cells list and I will expand further on why events must be handled differently, but ephemerals are how I hacked events into the paradigm: slots specified as ephemeral revert to nil immediately after propagating any non-nil value they assume, but not as a state change seen by the model. > but that doesn't make sense; I expect it to hold > whatever the last slot value was. Yep. Ephemerals are the oddball in this regard. ken > > Adapted from cells-test/hello-world.test (I tried both latest CVS Cells > and the tarball, with SBCL PPC Darwin 0.9.9): > > > (use-package :cells) > (use-package :utils-kt) > > (defmodel computer () > ((happen :cell :ephemeral :initform (c-in nil) :accessor happen) > (location :cell t > :initform (c? (case (^happen) > (:leave :away) > (:arrive :at-home) > (t .cache))) ;; ie, unchanged > :accessor location) > (response :cell :ephemeral > :initform nil > :initarg :response > :accessor response))) > > (macrolet ((fmt (name) > `(format t > "~%~A: ~15A (previously ~:[an unbound ~A~;~A~])" > ,name new-value old-value-boundp old-value))) > (def-c-output response () (fmt "response")) > (def-c-output happen () (fmt "happen ")) > (def-c-output location () (fmt "location"))) > > (defun hello-world () > (let ((dell (make-be 'computer > :response (c? (bwhen (h (happen self)) > (if (eql (^location) :at-home) > (case h > (:knock-knock "who's > there?") > (:world "hello, world.")) > "<silence>")))))) > (dotimes (n 2) > (setf (happen dell) :knock-knock)) > (setf (happen dell) :arrive) > (setf (happen dell) :knock-knock) > (setf (happen dell) :world) > (values))) > > > > CL-USER> (hello-world) > > happen : NIL (previously an unbound NIL) > location: NIL (previously NIL) > response: NIL (previously NIL) > happen : KNOCK-KNOCK (previously NIL) > response: <silence> (previously NIL) > happen : KNOCK-KNOCK (previously NIL) > response: <silence> (previously NIL) > happen : ARRIVE (previously NIL) > location: AT-HOME (previously NIL) > happen : KNOCK-KNOCK (previously NIL) > response: who's there? (previously NIL) > happen : WORLD (previously NIL) > response: hello, world. (previously NIL) > ; No value >
From: Frank Buss on 11 Mar 2006 22:41 Ken Tilton wrote: > Actually, I think c-l.net has the advantage over SourceForge here with > nightly tarballs, so no one has to study the CVS /language/ just to get > a little source. SourceForge has nightly tarballs, too: http://sourceforge.net/docs/E04/#nightly_tarballs But it is different from c-l.net: It is a backup of the repository itself, with all versions included. The same is possible with rsync read-only access to the subversion repository. But I don't see the advantage to download the head revision of all files in CVS or SVN. If you are a developer and you want the latest source, use checkout and then only bandwith saving updates. If you can't install a SVN or CVS client, you are not a developer :-) If you are a more normal user, then use official releases, only. > Of course if you want to /track/ an active project one, well, I use > TortoiseCVS so I /still/ do not need to learn CVS. There is not much to learn for CVS or SVN, if you just want to track an active project: an initial checkout (e.g. "svn co https://svn.sourceforge.net/svnroot/lispbuilder lispbuilder" for the Common Lisp Application Builder project at http://sf.net/projects/lispbuilder ) and a "svn up" (started in the project directory, no other options are needed) everytime you want to be up-to-date is sufficient. Same with cvs, only the path for the "co" command is different. -- Frank Buss, fb(a)frank-buss.de http://www.frank-buss.de, http://www.it4-systems.de
From: Tayssir John Gabbour on 12 Mar 2006 00:53
Ken Tilton wrote: > Tayssir John Gabbour wrote: > > Does anyone else have problems with DEF-C-OUTPUT? Its OLD-VALUE always > > reporting NIL to me... > > That just happens to be the case, partly because it is a silly example, > mostly because most of the slots are ephemeral, which are actually not > the norm in real world programming. Aaah, yes that makes sense. My understanding is that with discrete processes, we basically have a choice between cells being T or :ephemeral ("states" and "events"). And :delta seems to be absent from your sourcetree, which makes sense too. Do you think it would be an improvement to have :cell T by default, and if you want to cause an event, you'd do something other than setting the slot-value? Like: (setf (happens obj) :new-state) ;stays :new-state until changed (eventf (happens obj) :new-event) ;reverts to initform immediately Incidentally, if anyone could clarify a couple quick questions I'd be grateful: * So, cells only update other cells in their own object instance (not even cells allocated as :class)? Which means a cell in one object has to explicitly set another object's c-in cell? * What other notable vars are there, like .cache and the ^slotnames? Thanks, Tayssir |