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

[mcclim-devel] Debugger take III
by Peter Mechlenborg 16 May '05

16 May '05

Hi I have made some new adjustments to the debugger. I would like to get commit access, but until I've been granted that, you can get the modified debugger the usual place: www.daimi.au.dk/~metch/lisp/ Changes: - Added a packages.lisp file for the package. - Added an asd file. You should note that that it currently doesn't depend on swank because I had some troubles making it work. Therefore you have load Slime manually. - Keyboard shortcuts: 0 - 9 chooses the current restart (like Slime). Up and down moves the pointer between the stack frames and local variables. Space opens a stack frame (enter in Slime). Enter starts Clouseau (I finally get the name :-)) on a local variable. The biggest problem now, is that it is slow. I normally use sbcl, but I tried Mcclim on cmucl and it seemed that it was quite a bit faster. I think this has something to do with garbage collection, but that could be way off. Do any of you know what causes the big differences between sbcl cmucl? Bye, and remember to have fun! -- Peter

3 2

[mcclim-devel] Context menu
by Duncan Rose 15 May '05

15 May '05

I think I'm missing something to do with defining commands, and right-click context menus. I have process objects that are presented and a number of commands defined for this type. I assumed (presumably wrongly :-) that these commands would be displayed automatically when the presentation is right-clicked over. An example command definition follows: (define-glimpse-command (com-kill-process :name t :menu nil) ((obj 'ccl::process ; need to generalize this... :prompt "process" :gesture nil)) (clim-sys:destroy-process obj) (com-show-processes)) [aside; I suspect it could be useful to define a PROCESS type in CLIM-SYS] Do commands need to have a specific ':menu' entry to appear? -Duncan

2 1

[mcclim-devel] pathology in updating output with format-graph-from-roots
by rpgoldman＠real-time.com 13 May '05

13 May '05

[discussed on #lisp, to no avail...] If you look at http://rpgoldman.real-time.com/goofy-clim.png you will see what happens when one of my panes runs a display-function that calls format-graph-from-roots inside updating-output. The display does not all get updated, leaving funny gray borders around the active area of two panes. What's interesting is that this pathology affects the interactor pane at the bottom of the layout, as well as the application pane at the top. Before I made the display function for the top pane, the interactor pane used to function correctly. Now it erases its contents after every command, and has the gray border problem. I almost wonder if my adding :incremental-redisplay t to the application frames could have somehow bled over into the options used in initializing the interactor pane.... Finally, even with updating-output, using format-graph-from-roots does not correctly relize that it needs a horizontal scrollbar as well as a vertical one, when it gets a new graph in it. Some partial solutions: 1. Make application-panes with :scrollbars :both. I think this is actually a McCLIM bug. The spec states that the default value for application panes should be T (and it is), but also that T should be the same as :both (which doesn't seem to be an available option, according to the spec). So, somewhere we're treating :both and t differently, but I think they should be the same. 2. Use the bounding rectangle for the graph-output-record that format-graph-from-roots returns to set the :min-height and :min-width of the enclosing window, for example: (defun graph-mdp (stream &optional (mdp (taems-mdp *application-frame*))) (when mdp (let ((normal-ink +foreground-ink+) (arrow-ink *graph-edge-ink*) (text-style *graph-text-style*)) (declare (special normal-ink arrow-ink text-style)) (let* ((graph-output-record (with-drawing-options (stream :text-style text-style) (format-graph-from-roots (list (taems:initial-state mdp)) ;; printer #'draw-mdp-node ;; inferior-producer #'taems:mdp-successors :stream stream :arc-drawer #'clim-internals::arrow-arc-drawer ))) (rect (bounding-rectangle graph-output-record))) (change-space-requirements stream :min-width (bounding-rectangle-width rect) :min-height (bounding-rectangle-height rect)) graph-output-record)))) [this function will be invoked wrapped in updating-output.] That seems to get the scrollbars to be oriented to the right height and width. I know that Andy has said that he doesn't like the idea of having format-graph-from-roots automagically call change-space-requirements, for the reason that it could cause too much wasted computation. As a programmer, though, having a function that is this high-level, but then requires low-level messing with output records and the fairly cryptic and ill-documented change-space-requirements seems very odd. I suspect that my use of change-space-requirements is just plain wrong. My case up above is trivial, because my pane will ONLY have the graph in it; heaven help me if I ever want to stick a graph onto a scrolling output history pane or something... There ought to be (and probably is) some way of saying to the pane "hey, please make room for this big bolus of fresh, piping-hot pixels," but I don't understand the protocol well enough to say what it is... But neither of the above fixes the weird gray stuff, which seems to be some flavor of updating error, and likely too deep in the guts of the system for the likes of me. Any suggestions for how I should start?

2 1

[mcclim-devel] patch for graph-formatting.lisp
by rpgoldman＠real-time.com 13 May '05

13 May '05

I believe that there is a bug in graph-formatting.lisp where a graph, if it gets > 10 deep, will cause adjust-array to be called to create new array entries initialized to nil instead of to zero, meaning that comparisons cause a crash. Here's a patch. Will someone familiar with this code please vet it and apply it if appropriate? Thanks, Robert Index: graph-formatting.lisp =================================================================== RCS file: /project/mcclim/cvsroot/mcclim/graph-formatting.lisp,v retrieving revision 1.14 diff -u -F^(def -r1.14 graph-formatting.lisp --- graph-formatting.lisp 23 Apr 2005 20:02:01 -0000 1.14 +++ graph-formatting.lisp 13 May 2005 02:33:33 -0000 @@ -312,7 +312,7 @@ (defmethod layout-graph-nodes ((graph-ou (walk (node depth) (unless (graph-node-minor-size node) (when (>= depth (length generation-sizes)) - (setf generation-sizes (adjust-array generation-sizes (ceiling (* depth 1.2))))) + (setf generation-sizes (adjust-array generation-sizes (ceiling (* depth 1.2) :initial-element 0)))) (setf (aref generation-sizes depth) (max (aref generation-sizes depth) (node-major-dimension node))) (setf (graph-node-minor-size node) 0)

1 0

[mcclim-devel] question about layout-graph-nodes
by rpgoldman＠real-time.com 13 May '05

13 May '05

I managed to persuade format-graph-from-roots to crash, and it did so inside the WALK local function in layout-graph-nodes. Alas, this isn't too easy to debug, even with my beloved Allegro, because of all of the closures.... But I was wondering about the following piece of layout-graph-nodes. Perhaps I'm missing something, but there seems to be a call to MAX whose return is never harvested. (labels (... (walk (node depth) (unless (graph-node-minor-size node) (when (>= depth (length generation-sizes)) (setf generation-sizes (adjust-array generation-sizes (ceiling (* depth 1.2))))) (setf (aref generation-sizes depth) (max (aref generation-sizes depth) (node-major-dimension node))) (setf (graph-node-minor-size node) 0) ** (max (node-minor-dimension node) (setf (graph-node-minor-size node) (let ((sum 0) (n 0)) (map nil (lambda (child) (let ((x (walk child (+ depth 1)))) (when x (incf sum x) (incf n)))) (graph-node-children node)) (+ sum (* (max 0 (- n 1)) within-generation-separation)))))))) (map nil #'(lambda (x) (walk x 0)) root-nodes) ...) I confess that I am not a big user of MAP, nor do I fully grok what's going on here, but it seems like the MAX on the starred line does not serve any purpose. That suggests to me that the code mgiht be garbled somehow.... Any suggestions? For reference, I attach the full defmethod at the end of the message. BTW, the error occurring somehwere in here is a 'NIL' is not of expected type REAL error, which suggests to me that one of the comparisons has gone bad... I think it's because at generation 10, the array gets resized, and the initial value is NIL instead of 0... Thanks, R (defmethod layout-graph-nodes ((graph-output-record tree-graph-output-record) stream arc-drawer arc-drawing-options) ;; work in progress! --GB 2002-08-14 (declare (ignore arc-drawer arc-drawing-options)) (with-slots (orientation center-nodes generation-separation within-generation-separation root-nodes) graph-output-record (check-type orientation (member :horizontal :vertical)) ;xxx move to init.-inst. ;; here major dimension is the dimension in which we grow the ;; tree. (let ((within-generation-separation (parse-space stream within-generation-separation (case orientation (:horizontal :vertical) (:vertical :horizontal)))) (generation-separation (parse-space stream generation-separation orientation))) (let ((generation-sizes (make-array 10 :adjustable t :initial-element 0))) (labels ((node-major-dimension (node) (if (eq orientation :vertical) (bounding-rectangle-height node) (bounding-rectangle-width node))) (node-minor-dimension (node) (if (eq orientation :vertical) (bounding-rectangle-width node) (bounding-rectangle-height node))) (walk (node depth) (unless (graph-node-minor-size node) (when (>= depth (length generation-sizes)) (setf generation-sizes (adjust-array generation-sizes (ceiling (* depth 1.2))))) (setf (aref generation-sizes depth) (max (aref generation-sizes depth) (node-major-dimension node))) (setf (graph-node-minor-size node) 0) (max (node-minor-dimension node) (setf (graph-node-minor-size node) (let ((sum 0) (n 0)) (map nil (lambda (child) (let ((x (walk child (+ depth 1)))) (when x (incf sum x) (incf n)))) (graph-node-children node)) (+ sum (* (max 0 (- n 1)) within-generation-separation)))))))) (map nil #'(lambda (x) (walk x 0)) root-nodes) (let ((hash (make-hash-table :test #'eq))) (labels ((foo (node majors u0 v0) (cond ((gethash node hash) v0) (t (setf (gethash node hash) t) (let ((d (- (node-minor-dimension node) (graph-node-minor-size node)))) (let ((v (+ v0 (/ (min 0 d) -2)))) (setf (output-record-position node) (if (eq orientation :vertical) (transform-position (medium-transformation stream) v u0) (transform-position (medium-transformation stream) u0 v))) (add-output-record node graph-output-record)) ;; (let ((u (+ u0 (car majors))) (v (+ v0 (max 0 (/ d 2)))) (firstp t)) (map nil (lambda (q) (unless (gethash q hash) (if firstp (setf firstp nil) (incf v within-generation-separation)) (setf v (foo q (cdr majors) u v)))) (graph-node-children node))) ;; (+ v0 (max (node-minor-dimension node) (graph-node-minor-size node)))))))) ;; (let ((majors (mapcar (lambda (x) (+ x generation-separation)) (coerce generation-sizes 'list)))) (let ((u (+ 0 (car majors))) (v 0)) (maplist (lambda (rest) (setf v (foo (car rest) majors u v)) (unless (null rest) (incf v within-generation-separation))) (graph-root-nodes graph-output-record)))))))))))

1 0

Re: [mcclim-devel] question about CLIM spec
by C Y 11 May '05

11 May '05

Should we at least keep a backup in case something unforseen happens to his machine? CY --- Robert Strandh <strandh(a)labri.fr> wrote: > C Y writes: > > > > Wasn't the plan to move the annotated spec to common-lisp.net? Or > did > > I miss an update on that? > > I discussed that with Gilbert a few days ago, and now it seems that > he > prefers keeping it on his home machine. > > -- > Robert Strandh > > --------------------------------------------------------------------- > Greenspun's Tenth Rule of Programming: any sufficiently complicated C > or Fortran program contains an ad hoc informally-specified bug-ridden > slow implementation of half of Common Lisp. > --------------------------------------------------------------------- > Yahoo! Mail Stay connected, organized, and protected. Take the tour: http://tour.mail.yahoo.com/mailtour.html

2 1

Re: [mcclim-devel] question about CLIM spec
by C Y 11 May '05

11 May '05

Wasn't the plan to move the annotated spec to common-lisp.net? Or did I miss an update on that? CY --- "Robert P. Goldman" <rpgoldman(a)sift.info> wrote: > >>>>> "RS" == Robert Strandh <strandh(a)labri.fr> writes: > > RS> rpgoldman(a)real-time.com writes: > >> > >> Question: have I misread the specification, or is this a > deficiency > >> in it? > > RS> The latter is very likely. You can do us a favor and go to > the > RS> annotatable spec at > > RS> http://bauhh.dyndns.org:8000/clim-spec/index.html > > RS> and enter your comments at the appropriate place. > > OK, I had a preliminary whack at it. I hope the comments are useful. > I am hardly an expert on this subject! > > R > _______________________________________________ > mcclim-devel mailing list > mcclim-devel(a)common-lisp.net > http://common-lisp.net/cgi-bin/mailman/listinfo/mcclim-devel > Yahoo! Mail Stay connected, organized, and protected. Take the tour: http://tour.mail.yahoo.com/mailtour.html

2 1

[mcclim-devel] question about CLIM spec
by rpgoldman＠real-time.com 11 May '05

11 May '05

I was looking over Christophe's dendrogram sample code, and was surprised to find the following in the comments: "...this one is technically non-conforming, in that FORMAT-GRAPH-FROM-ROOTS isn't specified as &KEY &ALLOW-OTHER-KEYS" I went back to the spec and yes, AFAICT, the format-graph-from-roots function is specified as NOT being extendible with new keyword arguments. With all due respect, this seems kooky, since FORMAT-GRAPH-FROM-ROOTS' interface seems poorly structured to do much of anything beyond trees or graphs that are essentially tree-like. The interface seems poor if you want to have edges of different flavors, if you want to control the assignment of depths, etc. Now, of course, you can do a pretty arbitrary amount of customization by developing new graph-output-records. BUT... the spec as written (to the best of my understanding) doesn't provide any way to customize these new output record classes with initargs. Not only doesn't format-graph-from-roots as specified allow for extending its argument set, it also doesn't specify (again, as far as I can tell), how to pass initargs to the new class. The McCLIM implementation takes what seems to me to be a very sensible approach of passing extra keyword arguments from format-graph-from-roots into the make-instance for the GRAPH-OUTPUT-RECORD, but this is an extension to the specification. Question: have I misread the specification, or is this a deficiency in it? Cheers, Robert

4 4

[mcclim-devel] possible buglet in highlighting shapes
by rpgoldman＠real-time.com 10 May '05

10 May '05

I have built a grapher-based tool, that allows me to distinguish between multiple types of node by using shapes. When I highlight rectangular nodes, all is well. But I have "roundtangle" nodes, and when I pass the mouse over them, instead of seeing a box around them, or a highlight outline around the roundtangle, McCLIM gives me THREE rectangles, one around the body of the node, and one each around the semi-circles on the end. Actually, even the rectangle highlighting seems to me a little odd, since the highlight outline is drawn INSIDE the node, instead of around its outline. Here's the code that draws the nodes: (defun draw-taems-node (object stream) (declare (special normal-ink text-style)) (with-drawing-options (stream :ink normal-ink :text-style text-style) (with-output-as-presentation (stream object 'taems-node) (if (typep object 'taems:tmethod) (surrounding-output-with-border (stream :shape :rectangle) (format stream "~D: ~A" (musliner:name object) (taems:taems-name object))) (surrounding-output-with-border (stream :shape :oval) (format stream "~D: ~A" (musliner:name object) (taems:taems-name object))))))) Alas, I am too stupid to figure out how to grab a screendump from X to illustrate this; any suggestions? I wasn't sure how to get xwd to wait until I had a node highlighted before writing its output.... Thanks, R

12 38

[mcclim-devel] new graph type: first drawf
by rpgoldman＠real-time.com 06 May '05

06 May '05

I have developed a tentative first version of a new graph type, which is a tree with cross-edges. I.e., you have a graph which is basically a tree, but where you are allowed to have cross edges. The cross edges are NOT the same as graph edges, and in particular, don't cause additions to the PARENT/CHILD relationships, nor (for this reason) do they cause depths to be incremented. Note that these cross-edges do NOT have to be only one one level of the tree --- they can cross levels. To build such a graph you need to add two more arguments to format-graph-from-roots: :CROSS-ARC-DRAWER a function that takes all the arguments accepted by a normal arc-drawer, but also an edge-type keyword argument, which it is free to ignore. :CROSS-ARC-PRODUCER a function that takes a graph-node as argument, like inferior-producer, but that returns two values: a list of destination nodes and (optionally) a list of type-designators, that can be passed to the cross-arc-drawer, as the value of the :edge-type keyword argument. I have posted the source for this to the following URL, hoping that it will attract comments and suggestions: http://rpgoldman.real-time.com/lisp/tree-with-cross-edges.lisp I'd love to see this integrated into McCLIM some day! KNOWN BUGS: 1. Currently only works on :horizontal orientation (I don't have the arc routing for vertical yet). 2. Only draws straight lines for cross edges. It would be nice if the cross edges would swoop around intervening nodes, but I haven't read enough of the CLIM spec to figure out how to do this yet. 3. If you have a depth-bounded tree, probably something very ugly will happen! I'll try to assemble some screendumps and post them, but for now it's weekend... Best, R

1 0