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

Initial set of (mostly doc) bugs

I just got mail from Franz and they said to use Bug-CLIM@ILA.COM so here goes...

The following problems are based upon the version of CLIM that we got through
Franz.  The documentation is dated 9/26/90 and is for Release 0.9 of Version I.

Documentation bugs (no response desired unless otherwise indicated):
Page 13 of the Users manual (near the bottom) has
	(accept '(alist-member :alist ...))
while page 171 of the documentation has
1	alist-member alist &key test
0(rather than    1alist-member &key alist test0  )

Page 171 of the documentation has two 1WITH-TEXT-STYLE0s with different
argument lists.

Page 23 of the user's guide has 1+fill+0 rather than 1*fill*0.

All throughout the documentation the server path is given as something like
	1(:X11 :host localhost" :screen 0 :display 0)"
0(eg, see pages 20, 108 (twice))

Page 32 of the docs.  Even-scaling transformation seems to be defined
as a rigid scaling.  Did you mean "by the factor in x and the same factor in y"?

Page 56.  "will usually 1be 0buffered"  (add the "be").

Page 84.  "and triggers are describe1d0 more completely below"  (add the "d")

Page 123.  1define-menu-group name &rest clauses
0However it appears (by example) that it should be
	   1define-menu-group name clauses

0Page 126.  1define-sheet-class0 is defined but the example under it uses
	1def-sheet-class


0Page 147.  Under 1make-pane0 is a reference to 1defframe 0which doesn't seem to
be specified in the docs.

Page 150.  1menu-choose0:  Where it says "2atom 0is displayed using1 princ0...",
it ought to say something about2 printer0.
Your description of what gets passed to 2printer0 needs to be clarified.  It
is not clear whether it is the2 atom0, the2 object0 (for a non-atomic item),
or the 2item 0itself (atom or non-atomic).  I would guess the item, but
the "object" argument to 2printer0 is confusing.  (The :2printer0 example for
2presentation-type0 doesn't help either!)

Page 115.  1add-command-to-command-table0 "2command-nmae0" -> "2command-name0"


Other problems:

There is no documented way (that I could find) to specify a submenu of a menu.
I found that if I said :MENU as the type to a menu-group item, then it
seemed to work.  (I first tried :menu-group.)

1add-menu-group-entry0 and 1remove-menu-group-entry0 take a
menu-group (not menu-group name) as an arg but there doesn't seem to be a
way of getting a handle on the menu-group itself.  I have found that
1define-menu-group0 does return the menu-group as a value but that is
not exactly the way I'd expect you meant for me to get a handle on it.

Furthermore, I find that when I did modify a menu-group (on a submenu),
the menu that got presented to me never changed.   It didn't even change
when I launched (I thought) a new instance of the application  (after
I stop the previous frame).  In looking at the structure of the frame,
I saw something like :reusable t on a property list.  That isn't documented
anywhere that I saw so I couldn't figure out what to do.

By the way, when debugging in a launched frame, I couldn't find an easy
way to grab the value of the frame.  Launch-frame isn't documented to return
any particular value.  Nor is it documented to set up any kind of variable
that can be queried for the frame that you are running under.

In general, it would be nice to include in the docs, the possible values
of particular args.  Some places you do (eg text-styles but somewhere did
mention :VERY-LARGE but that isn't mentioned on page 46).  Other places you
don't ("type" in define-menu-group).  I had problems in figuring out, for
example, what kind of pane to specify to get :nlines to work.  (I gave up
as I had more pressing things.)

I haven't figured out under what conditions (if any) I can get different
gestures to work on menu items.  Right now I only know how to get it to
read any old click and not do something different on, say, a right or
left click.

In different spots, I've seen claim that the only (clim-top-level) can be
used as the top-level of an application (eg, see DEFINE-APPLICATION-FRAME
in the API).

BUILD-COMMAND was used in the CAD-DEMO example but isn't documented in the API.
(I kinda question its use.)

When I pop up the submenu of an application's top-level menu, it seems to
come up at a particular spot relative to the displayed frame irregardless
of whether the submenu fits on the screen.  At least with my window manager,
this window doesn't seem to be movable so I had to move my entire frame
just to get the submenu to fit on the screen.  (Maybe this is a problem with
my window manager or, more likely, my ability to use X.)


In any case, I'm just starting with this but expect to beat on it hard for
the next few weeks (when I find the time).  I'm mostly reporting all this
stuff for your info.  About the only things I'd really like back is info
on whether I have to write my own menu stuff to recognize different
mouse clicks (left vs right) and any hints you might have on highlighting
menus.


(Have you tried running the Cad-Demo with the Symbolics as the X-server
(8.0.1)?  It chokes on the arcs.  Another Symbolics X problem I've found
is that xon-type commands die in the C code (but that has nothing to do
with CLIM).)

0,,
Main Index | Thread Index