CLIM mail archive
[Prev][Next][Index][Thread]

Strange accepting-values behavior (and more)

    Date: Wed, 29 Jul 1992 12:55 EDT
    From: Randy Coulman <coulman@skdad.usask.ca>


    I'm using Franz CLIM 1.1 on a Sparc 1+ with Allegro CL 4.1.  I have a
    couple of questions/comments:

    1.)  I am working on a type of graph editor using CLIM.  I have defined
    a new gesture which I call :part.  I also have a command com-attach-part
    which takes one argument, the part to attach.  The command can either be
    invoked via menu or by using the :part gesture on an appropriate object.
    The command is designed to attach the selected object to the current
    focus object in the graph.  The command pops up an accepting values
    dialog (:own-window t) to allow the user to customize the interaction
    between the part and the focus object.  Two of the accepts in the dialog
    look like this:

    (setq rel
      (accept '(member = < > <= >= /=)
	      :stream stream
	      :default rel ;; initially '=
	      :provide-default t
	      :insert-default t
	      :prompt "Number of occurrences must be"))
    (princ " " stream)
    (setq comp
      (accept 'number
	      :stream stream
	      :default comp ;; initially 1
	      :provide-default t
	      :insert-default t
	      :prompt nil
	      :display-default nil))

    Two things:

    If the command is activated via the menu, the relevant section of the
    dialog looks like this:

    Number of occurrences must be: =
				   <
				   >
				   <=
				   >=
				   /= 1

The window used for the own-window dialog is gotten from the same
resource as the window used for the menu, and the narrower width window
from the menu is affecting the dialog.  I believe that this is now
fixed in CLIM 2.0

    If, instead, the gesture is used to invoke the command, the dialog looks
    like this:

    Number of occurrences must be: = < > <= >= /= 1

    which is what I want all the time.  Why are the two different?

    Second, the = is not initially highlighted.  If I select something else,
    that gets highlighted.  If I then re-select =, it is not highlighted.
    Any ideas?  Or is it that the bolding of the = sign just doesn't show up
    on my display?

The bolding of the = sign is not showing up.

    2.)  This is more of a comment.  I use define-border-type to define a
    new border type for nodes in my graph.  Here is the code:

    (clim:define-border-type :ellipse (stream record left top right bottom)
      (declare (ignore record))
      (let ((cx (/ (+ left right) 2))
	    (cy (/ (+ top bottom) 2))
	    (rx (+ 8 (/ (- right left) 2)))   ;; the 8 is to ensure that the
	    (ry (+ 8 (/ (- bottom top) 2))) ) ;; ellipse surrounds the text
	(clim:draw-ellipse* stream cx cy rx 0 0 ry :filled nil))
      )

    When this code is compiled, the declare form is ignored and I still get
    a warning message about not using record.  I'm not sure if this is a
    general CLIM issue, or if it is specifically related to Franz's version.
    It doesn't cause a crash, but I pride myself on writing "warning-free"
    code and this warning message violates that.

Just omit RECORD from the arglist.
References:
Main Index | Thread Index