- mcclim-devel - mailman3.common-lisp.net

Implementing new backends?
by Fabrice Leal 12 Jul '16

12 Jul '16

I've read that there's been some work documenting mcclim's internals, i was wondering how hard can it be to implement a new backend? Like implementing a sdl2 backend, for instance (I've been fantasizing of making a xconq like clone using only sbcl + sdl2) -- --- Fabrice Leal

2 1

Crowd-funding McCLIM development maybe?
by Robert Strandh 07 Jun '16

07 Jun '16

Hello everybody, I am investigating the possibility of crowd-funding some work on McCLIM. The idea is to pay Daniel Kochmański to work on it around 10 hours per week for the foreseeable future. I can pitch in some of the cost for this myself, and I will probably pitch in all of it if nobody is interested in contributing. But the money I pitch in comes from my personal budget so it would be preferable to get some help. If some people would be willing to contribute around 50€ per month, it would not take many contributors to have a near-permanent effort. Smaller contributions would of course also be welcome. The work would initially consist of cleaning up the code as is already happening. Future work would consist of improving performance, adding new backends, creating a set of gadgets with a consistent look, fixing the current problems with mirrored sheets, improving the documentation, etc. But not necessarily in that order. If you are one of the people who would be interested in contributing, please send me an email with an approximate amount that you would be willing to contribute. I don't need a firm pledge. For now, I am just investigating whether this is a realistic idea or not. -- Robert Strandh

1 0

[website][development]
by Daniel Kochmański 31 May '16

31 May '16

Hey, we're currently working on documenting the McCLIM internals and cleaning up the codebase. There is also a new refreshed website: https://common-lisp.net/project/mcclim/ Just writing to let you know that the development is all active etc. Best regards, Daniel -- Daniel Kochmański ;; aka jackdaniel | Poznań, Poland TurtleWare - Daniel Kochmański | www.turtleware.eu "Be the change that you wish to see in the world." - Mahatma Gandhi

7 15

[irc channel]
by Daniel Kochmański 31 May '16

31 May '16

Hey, a few people suggested lately that it would be nice to have a separate irc channel for a development discussion, help requests etc. I've created #clim channel on the freenode as a general channel (not only limited to McCLIM but to any topics related to the spec). Best regards, Daniel -- Daniel Kochmański ;; aka jackdaniel | Poznań, Poland TurtleWare - Daniel Kochmański | www.turtleware.eu "Be the change that you wish to see in the world." - Mahatma Gandhi

1 0

command-table s
by jtowler＠soncom.com 30 Dec '15

30 Dec '15

Hello, I am working on the LKB as previosly mentioned. Q: Do command-tables have to be a list per the specs of CLIM? The lkb has code that uses it as a symbol but it is not a list, and I would like to understand how to make it work and not just compile. With sbcl plus some threads, I got it to compile. Checking, I will read sources and manuals again. Thanks in advance for any help. http://lingo.stanford.edu/ LKB is the source site of the project as I best understand their organization. I applied to the delph-in develop mailing list to ask them about their code as I try it out. Just making sure I am not making beginner's or other foolish mistakes. John Towler

1 0

LKB with McCLIM
by jtowler＠soncom.com 22 Dec '15

22 Dec '15

Hello, I sent a post not being subscribed about the LKB. I just wanted to let folk know that the LKB from http://lingo.stanford.edu/, written from an Allegro CL platform compiles with McCLIM and some fixes/kludges on sbcl-1.3.1 completely with all of the code cited as Allegro dependent working without porting CLM (Motif from ftp.x.org/R5contrib) Clisp-2.49 is being worked on in parallel and interleaving the two. The developers list it as being open source, CL, requiring cltl2 (assumed now), Motif, and CLIM, all of which Allegro CL gives them. I got this to this significant point, and I wanted people concerned about McCLIM and grammar development environments to be informed. I found that there is a chunk of stuff about code involving threading and multiprocessing mixed into the Allegro based code. I used bordeaux-threads and looked at clocc/src/port/proc.lisp and I will need to rework and make sure the intended stuff runs like Un*x processes. sbcl has something now or in process. I intend to make it run in the direction chosen by clocc, to have it abstracted away from the particular CL. This will add to my experience, not a problem. My original problem about command-tables in McCLIM was solved by using defvar ... nil and at one point giving () as an argument to :command-table in the top-level code stuff. I don't know enough to be able to discuss it, but I made chunks of progress, and I wanted to thank the moderator for posting my query, and for all of the people who work/ed on McCLIM for their indirect contribution. I will be going through it again and fixing the thread/process/function uses for CLIM objects, to get it running, and then I will go back and organize the resulting changes and make patch files etc. Odds are my fixes to get it to compile and load messed up a chunk of CLIM stuff. tty only I discovered several years ago works with sbcl, and clisp. I run NetBSD-6.1.4 on a Dell D610 laptop for the most part. I just wanted to get to working with grammars of natural language. The project has a self-made open source license, and my changes will be open source to fit with this and my beliefs about that topic. Thanks again. Sincerely, John R. Towler jtowler(a)soncom.com NetBSD-6.1.4, sbcl-1.3.1 (now out), clisp-2.49 with NetBSD pkgsrc patches to be more current with their developing code base. cmucl for NetBSD i86 stops on lisp pathname errors and because it doesn't just compile, I moved back to sbcl, clisp. Sorry about the verbosity.

1 0

mcclim, CLIM, commands.lisp, clisp
by jtowler＠soncom.com 19 Dec '15

19 Dec '15

Tim Moore or whoever knows better. Mr. Moore is mentioned in the comments to mcclim/commands.lisp. I am not subscribed to the list, this may well bounce, and I shall try to check with gmane to see how this could be dealt with. This is from the Lkb (http://lingo.stanford.edu/) compilation which was developed with CLIM on Allegro CL and linux and Motif(r). I run NetBSD (6.1.4 currently) on a Dell D610 laptop and I am using clisp-2.49 with patches to be more current from NetBSD's pkgsrc. This is adjusting the lkb.system file to run the compilation throught the Allegro specific CLIM based code in a directory ACL_specific. - Compiling source file ; "/usr/local/home/jtowler/lisp/delph-in/Lkb2012/src/ACL_specific/topmenu.lsp" ;; Compiling file /usr/local/home/jtowler/lisp/delph-in/Lkb2012/src/ACL_specific/topmenu.lsp ... *** - CAR: LKB-TOP-COMMAND-TABLE is not a list The following restarts are available: ABORT :R1 Abort main loop I read a little documentation and source code. I have not written tons of lisp code so these may be beginner's issues. The Lkb is open source but I am not sure I can read Allegro manuals. Not a problem, just more to work around. mcclim/command.lisp was helpful as was the mcclim-paper.pdf from a Lisp conference or something that is around. The source at issue is the following: The problem is the error is thrown when some function I can't find is running through lkb-top-command-table but the standard commands.lisp taken code is not making lkb-top-command-table a list. Attempts to stick in a defvar or a setq or change the order of the defmethod and defgeneric forms, all just kludges to make something work, failed. Defing them to always true stubs shows the error problem with car. This is in (defgeneric construct-menu *), and (defmethod ccpnstruct-menu *) below in the included topmenu.lsp file. How can I with a better mcclim or better command-table code make this work to get to the next fix for the Allegro specific code. This in the end runs a Grammar Development Environment for natural language syntax and semantics using Head Driven Phrase Structure Grammar. I would like to run it without Allegro CL but with an open source CLIM. Thanks in advance for any suggestions. John R. Towler jtowler(a)soncom.com -------------------------topmenu.lsp--LICENSE is an open source license------ ;;; Copyright (c) 1991-2001 John Carroll, Ann Copestake, Robert Malouf, Stephan Oepen ;;; see LICENSE for conditions ;;; MCL port ;;; split old toplevel.lsp into toplevel.lsp which should be generic CL ;;; and this file which has the commands to create the actual menu ;;; ACL port - redefine menu commands ;;; split file again - menus.lsp is independent between ACL and MCL ;;; Note - this file now must be read in before any of the other ;;; CLIM files which associate menus etc with *lkb-top-frame* (in-package :clim-user) (eval-when (compile load eval) (merge-text-styles '(:sans-serif nil nil) *default-text-style*) (export '(*lkb-top-frame* *lkb-top-stream* *last-directory* set-up-lkb-interaction enable-type-interactions disable-type-interactions))) (defvar *lkb-menu-type* :core) (defvar *lkb-top-frame* nil) (defvar *lkb-top-stream* nil) (defvar *lkb-top-process* nil) (defvar *last-directory* nil) (defvar *complete-lisp-close* nil) ;;; Top level menus etc (defvar *lkb-menu-disabled-list* nil "Kludge because of MCL bug!!!!") (defvar *lkb-menu-grammar-file-list* nil) (defvar *lkb-menu-mrs-list* nil) (defvar *lkb-menu* nil) ;; Classes for abstract menus, with methods for turning them into real ;; clim (or emacs?) menus (defclass menu-thing () ((menu-title :initarg :menu-title :type string :accessor menu-title) (available-p :initarg :available-p :accessor available-p))) (defclass menu (menu-thing) ((menu-items :initarg :menu-items :accessor menu-items))) (defclass menu-item (menu-thing) ((menu-value :initarg :value :accessor menu-value))) ;; Create a leaf menu item (defun make-menu-item (&key name value available-p) (let ((menu (make-instance 'menu-item :menu-title name :value value :available-p available-p))) (unless (or (eql available-p :always) (and (eql available-p :grammar) lkb::*current-grammar-load-file*)) (push (intern (concatenate 'string "COM-" name)) *lkb-menu-disabled-list*)) (when (eql available-p :grammar) (push (intern (concatenate 'string "COM-" name)) *lkb-menu-grammar-file-list*)) (when (eql available-p :mrs) (push (intern (concatenate 'string "COM-" name)) *lkb-menu-mrs-list*)) menu)) ;; Create a sub-menu (defun make-lkb-submenu-item (&key menu-title menu-items available-p) (let ((menu (make-instance 'menu :menu-title menu-title :menu-items menu-items :available-p available-p))) (unless (eql available-p :always) (push (intern (concatenate 'string "MENU-" menu-title)) *lkb-menu-disabled-list*)) (when (eql available-p :mrs) (push (intern (concatenate 'string "MENU-" menu-title)) *lkb-menu-mrs-list*)) menu)) ;; Process menu description (defgeneric construct-menu (menu &optional rest)) (defmethod construct-menu ((menu menu) &optional table) (let ((new-table (make-command-table (intern (concatenate 'string "MENU-" (menu-title menu))) :errorp nil))) (push new-table (command-table-inherit-from (find-command-table 'lkb-top-command-table))) (add-menu-item-to-command-table table (menu-title menu) :menu new-table :errorp nil) (mapc #'(lambda (submenu) (construct-menu submenu new-table)) (menu-items menu)))) (defmethod construct-menu ((menu menu-item) &optional table) (let ((name (intern (concatenate 'string "COM-" (menu-title menu))))) (eval `(define-command (,name :menu ,(menu-title menu) :command-table ,table) () (handler-case (funcall (quote ,(menu-value menu))) #+:allegro (excl:interrupt-signal (condition) (format t "~%Interrupted")) ;; placeholder - we need a way ;; of generating an interrupt which will ;; affect these processes (storage-condition (condition) (format t "~%Memory allocation problem: ~A~%" condition)) (error (condition) (format t "~%Error: ~A~%" condition)) (serious-condition (condition) (format t "~%Something nasty: ~A~%" condition))))))) #| (defun construct-menu (menu) (apply #'concatenate 'string (nconc (list "(" (prin1-to-string (slot-value menu 'menu-title)) " ") (mapcar #'construct-menu-1 (slot-value menu 'menu-items)) (list ")")))) (defun construct-menu-1 (menu) (if (lkb-menu-item-p menu) (apply #'concatenate 'string (nconc (list "(" (prin1-to-string (lkb-menu-item-menu-title menu)) " ") (mapcar #'construct-menu-1 (lkb-menu-item-menu-items menu)) (list ")"))) (format nil "[ ~S ~A t ]" (first menu) (string-downcase (string (third menu)))))) |# ;; Create lkb interaction frame (defun expand-lkb-menu nil (setf *lkb-menu-type* :big) (set-up-lkb-interaction)) (defun shrink-lkb-menu nil (setf *lkb-menu-type* :core) (set-up-lkb-interaction)) (defun set-up-lkb-interaction (&optional system-type) (unless system-type (setf system-type (or *lkb-menu-type* :core))) ;; remove any old commands (setf *lkb-menu-disabled-list* nil) (setf *lkb-menu-mrs-list* nil) (ecase system-type (:core (create-mini-lkb-system-menu)) (:big (create-big-lkb-system-menu))) #| (:full (create-lkb-system-menu)) (:yadu (create-yadu-system-menu))) |# (unless (EXT:getenv "LKB_GUI_EXTERNAL") (set-up-clim-interaction))) (defun set-up-clim-interaction () ;; Flush old commands (let ((table (find-command-table 'lkb-top-command-table :errorp nil))) (when table (let ((menu-items nil)) ;; Note - removing items inside the mapping function ;; does not always work - hence the fudge (map-over-command-table-menu-items #'(lambda (name char item) (declare (ignore char item)) (pushnew name menu-items)) table) (dolist (name menu-items) (remove-menu-item-from-command-table table name))) (map-over-command-table-commands #'(lambda (name) (unless (eql name 'COM-CLOSE-TO-REPLACE) (remove-command-from-command-table name table))) table :inherited nil))) (setf (command-table-inherit-from (find-command-table 'lkb-top-command-table)) nil) ;; make sure we have a way out (setf (command-table-inherit-from (find-command-table 'lkb-top-command-table)) (list (make-command-table 'menu-quit :errorp nil))) (define-command (com-quit :menu "Click to confirm quit" :command-table menu-quit) () (setq *complete-lisp-close* t) (frame-exit *application-frame*)) (add-menu-item-to-command-table 'lkb-top-command-table "Quit" :menu 'menu-quit :errorp t) (define-command (com-close-to-replace :command-table lkb-top-command-table) () (frame-exit *application-frame*)) ;; add correct menu items (dolist (submenu (menu-items *lkb-menu*)) (construct-menu submenu 'lkb-top-command-table)) ;; go to it (start-lkb-frame)) ;; Top-level CLIM frame (define-application-frame lkb-top () (standard-application-frame) (:panes (display (outlining (:thickness 1 :record-p t) (spacing (:thickness 1 :record-p t) (scrolling (:scroll-bars :both :record-p t) (make-pane 'application-pane :name "lkb-pane" :text-cursor nil :end-of-line-action :allow :borders nil :background +white+ :foreground +black+ :draw t :record-p t :display-time t)))))) (:layouts (default display)) (:geometry :width 550 :height 200) (:command-table lkb-top-command-table)) (defun start-lkb-frame () (let ((old-frame *lkb-top-frame*)) (setf *lkb-top-process* (mp:join-thread "start-lkb-frame" #'run-lkb-top-menu #+:allegro excl::*initial-terminal-io* #-:allegro *terminal-io*)) ;; note - if this is being called from a command in the old frame it's ;; important this is the last action ... (when old-frame (execute-frame-command old-frame '(com-close-to-replace))))) (defun run-lkb-top-menu (background-stream) ;; define this function so that stuff can be called on exit from LKB (let ((frame (make-application-frame 'lkb-top))) (dolist (command *lkb-menu-disabled-list*) (setf (command-enabled command frame) nil)) (setf *lkb-top-frame* frame) (setf *lkb-top-stream* (get-frame-pane *lkb-top-frame* 'display)) ;; crude way of seeing whether this is being called when we already have a ;; grammar (when lkb::*current-grammar-load-file* (enable-type-interactions)) (setf lkb::*lkb-background-stream* background-stream) (unwind-protect (run-frame-top-level frame) (when *complete-lisp-close* ;; ;; with the latest set of CLIM patches, it appears we need to rebind the ;; standard streams to avoid an `operation on closed stream' error(), ;; while shutting down the Lisp. not quite sure why, but alas. ;; (8-feb-08; oe) #+:allegro (let* ((stream excl:*initial-terminal-io*) (*standard-output* stream) (*debug-io* stream) (*terminal-io* stream) (*standard-input* stream) (*error-output* stream) (*query-io* stream) (*trace-output* stream)) (excl:exit 0 :no-unwind t :quiet t)) #+:lispworks (lw:quit :ignore-errors-p t) #-(or :allegro :lispworks) (error "no known mechanism to shutdown Lisp (see `topmenu.lsp'"))))) #| (defun user-exit-lkb-frame (frame) ;; Check if user really wants to do this. By default, exit Lisp as ;; well. For stand-alone application, always exit Lisp as well. (if (lep:lep-is-running) (let ((result (lkb::ask-user-for-multiple-choice "Really exit?" 'Lisp 'LKB 'Cancel))) (cond ((eq result 'lkb) (frame-exit frame)) ((eq result 'lisp) (setf *complete-lisp-close* t) (frame-exit frame)) (t nil))) (when (lkb::lkb-y-or-n-p "Really exit the system?") (setf *complete-lisp-close* t) (frame-exit frame)))) (defun restart-lkb-function nil (lkb::read-psort-index-file) (setf *last-directory* nil) (set-up-lkb-interaction)) |# (defun restart-lkb-window nil (setf *last-directory* nil) (set-up-lkb-interaction)) #+:allegro (defun dump-lkb nil (if lkb::*current-grammar-load-file* (progn (lkb::lkb-beep) (format t "~%Dump system will not work after a grammar has been loaded")) (let ((image-location (lkb::ask-user-for-new-pathname (format nil "File for image (local file advised)")))) (when image-location ;;; apparently 5.0 requires that the file be .dxl ;;; this lets the user give another type - since they may know more ;;; than I do, but issues a warning message #+(and :allegro (version>= 5 0)) (let ((image-type (pathname-type image-location))) (unless image-type (setf image-location (merge-pathnames image-location (make-pathname :type "dxl")))) (when image-type (unless (equal image-type "dxl") (format t "~%Warning - image type was ~A when dxl was expected" image-type)))) (setf excl:*restart-init-function* #'restart-lkb-window) (excl:dumplisp :name image-location) (lkb::lkb-beep) (format t "~%Image saved~%") nil)))) (defun enable-type-interactions nil ;; it may only work from within the application frame (dolist (command *lkb-menu-disabled-list*) (if (or lkb::*mrs-loaded* (not (member command *lkb-menu-mrs-list*))) (setf (command-enabled command *lkb-top-frame*) t)))) (defun disable-type-interactions nil (when clim-user::*lkb-top-frame* ;; this is called when a type file is being redefined it may only ;; work from within the application frame (dolist (command *lkb-menu-disabled-list*) (unless (member command *lkb-menu-grammar-file-list*) (setf (command-enabled command *lkb-top-frame*) nil))))) (defun enable-grammar-reload-interactions nil (dolist (command *lkb-menu-grammar-file-list*) (setf (command-enabled command *lkb-top-frame*) t))) (defun enable-mrs-interactions nil (when lkb::*mrs-loaded* (dolist (command *lkb-menu-mrs-list*) (setf (command-enabled command *lkb-top-frame*) t)))) ;;; functions called from top level menu which are time ;;; consuming (defun parse-sentences-batch nil ;;; for MCL this can just be parse-sentences (mp:join-thread "Batch parse" #'lkb::parse-sentences)) ;; Direct output to LKB window, if present (defun invoke-with-output-to-top (body) (unwind-protect (let ((*standard-output* *lkb-top-stream*) ;; ;; _fix_me_ ;; we believe that debug output from the CLIM patches may cause a ;; force-output() on *debug-io* to raise an error(), when running ;; in a background process. (13-feb-08; oe) ;; #-:logon (*debug-io* *lkb-top-stream*) ;; (*terminal-io* *lkb-top-stream*) (*standard-input* *lkb-top-stream*) (*error-output* *lkb-top-stream*) (*query-io* *lkb-top-stream*) (*trace-output* *lkb-top-stream*)) (when (not (eq mp:current-thread *lkb-top-process*)) (mp:process-add-arrest-reason *lkb-top-process* :output)) (setf (stream-recording-p *standard-output*) t) (funcall body)) (mp:process-revoke-arrest-reason *lkb-top-process* :output)))

1 0

memory leaks while drawing images
by José Ronquillo 23 Jun '15

23 Jun '15

Hello. Please, can anybody tell me the correct way to display an image on an application pane? I tried the approach written in the "image-viewer" example, but when I run that example, the program leaks memory. Every time that the function draw-pattern* was called the memory reported by (room) increased and eventually SBCL runs out of memory so the game is over. I create the pattern with make-pattern-from-bitmap-file and then I display it with draw-pattern*. The code that shows up the problem is attached. I also include res.txt which collect the output of (progn (sb-ext:gc :full t) (room)) that I evaluated after each execution of the command that displayed the image. Thanks in advance for any suggestion.

1 0

drag-and-drop-translator exclude reflexivity
by Kilian Sprotte 08 Apr '15

08 Apr '15

Hi, do you know how it is possible to extend the following definition of a drag-and-drop-translator so that it is not possible to link items with themselves? (define-drag-and-drop-translator connect-items (item command item command-table) (object destination-object) `(com-connect-items ,object ,destination-object)) Regards, Kilian

1 0

Bug in McCLIM ASD file
by John Morrison 08 Apr '15

08 Apr '15

Hey All; Hope somebody out there is listening. FYI, in file mcclim.asd, in the defsystem for :clim, there is a dependency on :goatee-core (good) but also a dependency on file presentation-history within goatee (bad). AFAICT short version is that this screws up under later versions of ASDF2 and all versions of ASDF3 - creating a cycle in depedencies that causes McCLIM to be reloaded unnecessarily. This was particularly troublesome in my prototype quicklisp GUI<https://bitbucket.org/symbolicsimulation/com.symsim.oss.ql-gui>, which borked whenever anything based upon McCLIM got loaded... Eliding the dependency seems to do the right thing. I have done it in my fork of McCLIM <https://github.com/jmorrison/mcclim> (bleeding-edge branch), but I am obviously git-challenged, have not adhered to git best practices, and have not issued any pull requests... Would like to contribute more fully, and being indoctrinated in correct git workflow by whomever is leading the effort... Please advise. -jm

2 1