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

Re: mcclim: redisplay-frame-panes where application-pane contains a gadget fails
by Daniel Kochmański 11 Nov '22

11 Nov '22

Hey Paul, sorry for taking so much time to respond. I'm answering your emails in batch. > [two down one to go] The issue was that (draw-rectangle* t 0 0 10 10 :filled nil) has the extent [-.5 -.5 10.5 10.5], I've explained that the line thickness when filled is nil is normally 1 (or 2 depending on the backend), so it extends beyond the rectangle dimensions in question. You've concured that LispWorks makes the bounding rectangle [0 0 11 11] (proving that they take into account the line thickness), but I wonder why they are doing that? If we draw a rectangle with line thickness 4 then the bounding rectangle is [0 0 14 14] and not [-2 -2 12 12]? Does it mean that drawing a border around the object may produce a result where the object is not in the center, because it is "too close to the edge"? Or does it mean that /after all drawing/ they move all output so it fits inside the sheet region defined (by them) as [0 0 width height]? McCLIM allows defining sheets with a region that is not aligned with 0X and 0Y, so perhaps it is their limitation of the implementation? > [follow up to previous convo] > > Please give the attached code a try. (clim-user::tryit) > Keep left clicking in the pane. Expected behavior? >From the file in question I think that you've mixed some other program after the initial output (starting from the function ~test-draw-line~). Either way the behavior I'm observing is expected. You take the sheet dimensions to draw on a sheet and then the sheet dimensions are computed based on the output (bar the fact that we seem to come beyond specified max-width - this is most likely a bug). > [with-radio-box] Thank you for the kind words. CAPI backend sounds nice. I'm working on SDL2 backend on the other hand (in parallel with other things, i.e today I've finished an initial port of ECL to emscripten, so it is possible to run CL in the web browser (see turtleware.eu/static/ecl/ecl.html, open the js console). Thanks for the code - it had an issue that "selected" toggle boxes did not have a desired background. I've implemented this feature by hooking the background directly in the local function MAKE-PANE (while still respecting the background if the programmer explicitly supplied one). See the recent commit if you are curious, here is the code for a simple test: ``` (define-application-frame with-boxes () () (:panes (app :application) (rad (with-radio-box (:background +light-blue+ :type :some-of) "Daniel" "Ma" (radio-box-current-selection "Kota") (radio-box-current-selection "bata") (make-pane :toggle-button :label "make-pane 1") (radio-box-current-selection (make-pane :toggle-button :label "make-pane 2")) (make-pane :toggle-button :label "make-pane 1" :background +light-pink+) (radio-box-current-selection (make-pane :toggle-button :label "make-pane 2" :background +light-pink+)))))) (find-application-frame 'with-boxes) ``` > [mcclim: redisplay-frame-panes where application-pane contains a gadget > fails] Thanks for the report. Notice that if we define a redisplay command, then the problem does not happen (when the command is activated) - that means that the issue is caused by the fact that we try to redisplay the frame asynchronously ``` (define-v-frame-command (com-redisplay :menu t) () (with-application-frame (frame) (redisplay-frame-panes frame))) ``` and indeed the issue is caused by the fact that the pane realizer is not bound. REDISPLAY-FRAME-PANES works fine if we bind it: ``` (let ((fm (frame-manager *frame*))) (with-look-and-feel-realization (fm *frame*) (redisplay-frame-panes *frame*))) ``` That said asynchronous access is not safe unless you either call EXECUTE-FRAME-COMMAND (safe!) and/or dispatch an event and handle it (safe and low level). I hope that this is a satisfying answer. Best regards, Daniel -- Daniel Kochmański ;; aka jackdaniel | Przemyśl, Poland TurtleWare - Daniel Kochmański | www.turtleware.eu "Be the change that you wish to see in the world." - Mahatma Gandhi ------- Original Message ------- On Thursday, November 10th, 2022 at 19:38, Paul Werkowski <pw(a)snoopy.qozzy.com> wrote: > LispWorks 8.0.1 Windows: > McCLIM current source: > Example attached > > Paul > > CLIM-USER 5 > (redisplay-frame-panes frame) > > > Error: No applicable methods for #<STANDARD-GENERIC-FUNCTION MAKE-PANE-1 > 4160C3ACB4> with args (NIL #<CLIM-INTERNALS::DEFAULT-APPLICATION > > 405047874B> LABEL-PANE :LABEL "A Label" :BACKGROUND > > #<CLIM-INTERNALS::NAMED-COLOR "green">) > > 1 (continue) Call #<STANDARD-GENERIC-FUNCTION MAKE-PANE-1 4160C3ACB4> > > again > 2 Clear the output history of the pane and reattempt forceful redisplay. > 3 Clear the output history of the pane, but don't redisplay. > 4 Skip this redisplay. > 5 (abort) Return to top loop level 0. > > Type :b for backtrace or :c <option number> to proceed. > > Type :bug-form "<subject>" for a bug report template or :? for other > > options. > > CLIM-USER 6 : 1 >

1 0

updating-output issues
by Paul Werkowski 01 Feb '22

01 Feb '22

I've been revisiting a clim project from 15 years ago which was to provide clim graphics to Dan Corkhill's very nice GBBopen blackboard system. I have an initial version now working nicely on Lispworks clim-2.0 on Windows. It is close to running with McCLIM as well. I have two issues, updating-output appears to work in that the :cache-test is respected but where Lispworks clim calls clear-output-record when a displayed object moves or is deleted McCLIM does not do so. I have attached a simple demo that shows that problem. Just click right or left anywhere in the pane to move the green disk left or right to see the problem. The other issue is that redisplay-frame-pane (not used in the demo) always causes an infinite recursion that finally results in a blown stack. Part of the problem is with window-clear which in addition to using medium-clear-area, etc to clear the screen also sets window size to zero and then calls compose-space. I have an :around method that avoids that problem but the redisplay recursion persists. My workaround for that is to just call the display function directly but that does not allow updating-output to work correctly. Paul

2 3

scigraph datasets mixin behavior
by John Phillips 01 Oct '21

01 Oct '21

Hello, I have recently started trying to use the scigraph package in McClim for a simulation course I'm taking. I spent more time than I care to admit this morning trying to get a scatter plot dataset to show up -- I kept getting the graph with tick marks but no data. Finally after reading the code I found out that the hidden-datasets member of the graph-datasets-mixin object will by default initialize to be a copy of the datasets member. For example if I do something like this: (setf *my-graph* (make-instance 'graph:graph :datasets (list *my-graph-data*) :hidden-datasets nil)) And then this: (graph::hidden-datasets *my-graph*) I get: (#<GRAPH-DATA GRAPH-DATA-15>) In other words, all the datasets are hidden by default. Is this the way the class is meant to behave? If so, I feel like I must be doing something wrong in the way I'm constructing my graph -- although it does work (I manually set the hidden-datasets member to nil after construction). Is there anywhere where I can find example usages of this package? There don't seem to be scigraph examples in the Examples/ folder, the things that say 'graph' usually seem to be manual graph creations. Thanks, John

1 0

[new irc server]
by Daniel Kochmański 23 May '21

23 May '21

Hey, Due to the recent events revolving freenode IRC network I've moved the IRC channelto #clim @ libera.chat. The old channel is obsolete. Best regards, Daniel -- Daniel Kochmański ;; aka jackdaniel | Przemyśl, Poland TurtleWare - Daniel Kochmański | www.turtleware.eu "Be the change that you wish to see in the world." - Mahatma Gandhi

1 0

make-translator-fun
by Paul Werkowski 13 May '21

13 May '21

(defun make-translator-fun (args body) (cond ((null args) (warn "OBJECT parameter is obligatory (adding ignored parameter)") (let ((object-arg (gensym "OBJECT-ARG"))) `(lambda (,object-arg &key &allow-other-keys) (declare (ignore ,object-arg)) ,@body))) (t `(lambda (,(car args) &key ,@(cdr args) &allow-other-keys) (declare (ignorable ,(car args))) ,@body)))) I guess the above does the right thing if ARGS is NIL but won't catch the case of ARGS incorrectly given as, for instance, (x y window) instead of (object x y window). Paul

3 3

define-presentation-to-command-translator
by Paul Werkowski 13 May '21

13 May '21

Trying to run McCLIM on windows-10/LispWorks 64-bit. This issue regards define-presentation-to-command-translator. The marked lines below show the problem. Apparently the McCLIM code resulted in replacing the DEV-X argument I wanted with a DEVICE-EVENT object I did not want. I then had to use a non exported function from clim-internals to retrieve the missing value. The &rest arg I put in is NIL in both cases. I have not figured out how that happened. (defun %set-a-cursor #+clim-2.0 (chart cursor dev-x dev-y pane &rest stuff) ;<<<< #+mcclim (chart cursor event dev-y pane &rest stuff) ;<<<< (let* (#+mcclim (dev-x (climi::device-event-x event)) (x (untransform-position (medium-transformation pane) dev-x dev-y))) (with-slots (lcursor rcursor) chart (let ((newx (ecase cursor (rcursor (max (x-cursor-position lcursor) (ceiling x))) (lcursor (min (x-cursor-position rcursor) (floor x)))))) ;; args to com-set-cursor (list (slot-value chart cursor) newx pane))))) (define-presentation-to-command-translator xlate-set-rcursor (clim:blank-area com-set-cursor plotter :documentation "Set Right Cursor" :gesture :menu :tester ((object window) (declare (ignore object)) (typep window 'basic-chart-pane))) (x y window) (%set-a-cursor (chart-pane-chart window) 'rcursor x y window))

2 1

scigraph suffering from bitrot?
by John Morrison 03 May '21

03 May '21

Hi; Before I strap on my hip-waders, is my intuition that climi::pixmap type has been refactored correct? Please advise If so, that's where I'll start looking - and if you suggest a backend-independent type I should be using, I will be most grateful. I also checked scigraph-evo, and its demos fail in precisely the same way as bundled Apps/Scigraph's demos. Thanks in advance for pointing the way, -jm p.s., I was looking for a real pan-and-zoom XY canvas - if anybody has better suggestions, I'm all ears.

1 0

default-icons.lisp
by Paul Werkowski 29 Apr '21

29 Apr '21

(make-pathname :directory '(:relative "data/icons") "data/icons" is not a valid directory component unless the underlying OS allows "/" character in names. At least on Windows is should be '(:relative "data" "icons"). Paul

2 1

Current McCLIM does not load
by zbyszek 29 Apr '21

29 Apr '21

This is the message, both from quickload, and loading version taken from github debugger invoked on a ASDF/FIND-COMPONENT:MISSING-COMPONENT in thread #<THREAD "main thread" RUNNING {10005605B3}>: Component "clim" not found Type HELP for debugger help, or (SB-EXT:EXIT) to exit from SBCL. restarts (invokable by number or by possibly-abbreviated name): 0: [RETRY ] Retry ASDF operation. 1: [CLEAR-CONFIGURATION-AND-RETRY] Retry ASDF operation after resetting the configuration. 2: Retry ASDF operation. 3: Retry ASDF operation after resetting the configuration. 4: [ABORT ] Give up on "clim" 5: Give up on "mcclim" 6: Exit debugger, returning to top level. ((:METHOD ASDF/OPERATE:OPERATE (SYMBOL T)) ASDF/LISP-ACTION:LOAD-OP "clim" :VERBOSE NIL) [fast-method] error finding frame source: Bogus form-number: the source file has probably changed too much to cope with. source: NIL 0] What is wrong with it? Best regards, Zbyszek Jurkiewicz

2 1

Regression in compose-xx-with-transformation
by Paul Werkowski 23 Mar '21

23 Mar '21

A recent change in master branch seems to have reversed the argument order in the subject definitions. I've only checked compose-scaling-with-transformation which is again wrong causing my graphics to be written way off screen. Paul

2 3