
                    \     /  ))
                 (( |\   / \  ))
                ((( | \ |   \
                 (( |  \|    | \ \
                     \  \\  /   ) )               Khoros System 1.0
  ---------          __\|_\/   (* *)                Bug List
--------          (_/       \== \ /                                        
   ----------       \_______/    V
                      ))))))
		

		
		 	KHOROS 1.0  BUG LIST


	This is a list of known bugs in Khoros programs.  Some bugs occur
	on all architectures, while others occur only on a particular 
	architecture, or in a particular situation.  Users should be aware
	of current problems,  both so that they can be expected, and also
	so that they need not be reported to khoros-bugs@chama.unm.edu.

	General bugs (those that appear on most/all architectures) appear
	first, while the those that occur only in a particular environment
	are broken up into categories that follow.  Read the list of bugs 
	under "GENERAL", and then look below for any other problems that 
	may occur on your particular architecture / environment.  

	We have tried to test on most major architectures, but we simply
	don't have the facilities to test on them all.  Our major development
	architectures are DECMIPS and SUN4, on 8-plane workstations running
	the twm window manager, so Khoros will probably work best in a similar 
	environment; however, we have also compiled and tested on Apollo, SGI, 
	NeXt, Sun3, and Vax.  We have attempted to address problems experienced
	in Khoros Beta by users of the olwm window manager and those on 
	24-plane displays; however, problems may still be experienced here.

	In addition, ports have been done by others to CRAY, HP9000, IBM6000, 
	and I386/486 running interactive UNIX; we have incorporated their ports 
	into our code but we have no way of testing it.  If you find bugs that 
	are not listed below, please mail them to us via 
	khoros-bugs@bullwinkle.unm.edu => using the BugReport form found in
	$KHOROS_HOME/repos/BugReport.  Other correspondence may be mailed to 
	the Khoros Users Group, khoros@chama.unm.edu.  If you are not on the 
	mailing list and you would like to be, please mail your request to
	khoros-request@bullwinkle.unm.edu.   Thank you...


======================  GENERAL: ===============================

ANNOTATIONS -

*  annotations are created relative to device coordinates, not image
   coordinates (in editimage) or world coordinates (xprism).  this is
   a design flaw rather than a bug.

CANTATA -

*  does not work well with either CONCERT or JOURNAL PLAYBACK.

*  a minor bug in the refresh mechanism shows itself when working with
   procedures.  use the redraw button often.

*  there is a major bug dealing with management of connections between 
   procedures and external glyphs (ie, glyphs in other procedures or a 
   parent workspace).  suppose you have a glyph inside a procedure (an 
   internal glyph);  this glyph is connected to another glyph in the parent 
   procedure (the external glyph).  if you delete the internal glyph,
   cantata will not delete the connection between the external glyph and
   the procedure.  the only way to delete this connection is to delete the
   whole procedure.  after this has occurred for the first time, cantata
   may exhibit strange behavior, or even crash.

*  if you try to look up variable "i" which is not declared in the current
   or parent workspace, and you then declare the variable in the parent
   workspace, the local workspace will never look to the parent for that
   variable.

*  When/if cache becomes corrupted (or if cache is not available) cantata 
   will complain about not being able to flush the file.  The can
   also be a problem if you save a workspace with cache on, and
   then try to run it on a machine that has no cache.

*  if you change the input or output by hard-wiring a filename, the
   glyph will not reflect that the Data Available is true.

*  after saving/restoring a workspace, simple expressions do not have the
   current value.

*  When you use "paste" to duplicate a collection of
   glyphs the temporary file names associated with them
   are not changed.

	~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 

CONCERT - 

*  does not exit from the originating machine after all parties have 
   hit their respective "Quit" buttons; you must hit ^C on the 
   originating machine to exit.

*  when used with WARPIMAGE, deletion of tiepoints will not propagate 
   to other workstations unless that party's mouse pointer is exactly on top of 
   the same tiepoint that the first person is deleting.  This is because 
   warpimage queries the mouse before deleting a tiepoint.

*  often all users must click on error messages to make 
   them go away.

	~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 

EDITIMAGE -

*  when used with a large font: the Pixel Display widget (Print Pixels) is not 
   created according to the current font size;  you may have to manually resize 
   the pixel display in order to see all the little pixel-value boxes.  

*  If your image is smaller than the zoom window,
   the cursor will not move within the zoom window
   when you move the cursor within the image.


	     
	~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 

GHOSTWRITER -

*  if your system does not have "nroff", ghostwriter will give error messages;
   this is not a bug - simply use the [-format 0] option.

	~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 

XPRISM3 -

*  Mesh plots are not always correctly drawn when the perspective is changed.

*  When displaying multiple plots, there is not hidden surface removal 
   between plots;  this means that one plot will always appear to be on top
   of the next, when in fact the plots may intersect each other.


FILE FORMATTING -

* pbm2viff will produce a 3 band image of type no map, when converting a
  a 24 bit rgb pbm image.  pbm2viff in this case should be creating an
  image with a map type of generic rgb.

* raw2viff fails on large images (2000x2000).


======================  SILICON GRAPHICS (SGI): =========================

* When you do a journal playback of a program, where the journal playback 
  file was recorded on another architecture (with a smaller font), the 
  graphical user interface will start out fine, but then will resize itself 
  to be too small.  Labels are truncated;  displays mushed; in 
  some cases the playback sessions are ok, in others they are corrupted.

* Any program taking a VIFF image as input will not work with an image file 
  from an NFS mounted partition.


DISPLAY PROGRAMS (ANIMATE, EDITIMAGE, ETC) -

*  when you use a Shape mask, it works fine but you cannot get it to go
   away by NULLing out the parameter box.  you're stuck with it.

	~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 
EDITIMAGE -

* When you use Combine Annotations, it will sometimes combine a shadow of the 
  Image Utilities pulldown menu, in spite of the fact that the menu was gone
  and the currently displayed image was NOT obscured when you clicked on 
  "Combine" in the prompt widget.

* when you use the Pixel Print capability, the Pixel Display widget is
  sometimes is too small to fit all the little pixel-value boxes.  you must 
  manually resize the pixel display in order to see them all. (ie, pixel boxes 
  not sized according to font size)


	~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 

XPRISM3 -

*  labels on axes are not readable.

*  When the  Contours pane of the Options subform is used, often you will
   get the error message: "X3D_display_contour: not enough memory"  even
   though the contour plot seems fine.

	~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 

VISO2 -

*  seems to have a problem when printing out statistics.


======================  APOLLO: =========================

EDITIMAGE -

1) The white crosshairs on the interactive LUT do not appear;  however,
   interactive LUT still works fine, you just have to imagine the
   crosshairs.

PUTIMAGE -

1) Sometimes putimage does not display x, y, and grey levels for very
   small images.  this may also be a problem with editimage.


======================  NEXT: =========================

CANTATA -

1) unable to restore workspaces properly. connections cannot
   be restored because it is unable to find glyphs.


======================  OPEN WINDOWS (OLWM): =========================

* The X server seems to report the position
  of the cursor not at the hot spot, but at the lower right hand corner of the
  cursor.  This results in some inconsistent results in cantata, editimage,
  putimage, etc when the application must query the mouse.  For example, the
  position widget in editimage & putimage reports the pixel values of pixels
  that are really to the lower right of where the hot spot of the cursor is, 
  etc.

* There also seems to be a problem with Backing Store & Pixmaps that will crash
  the server periodically when using Khoros.  We were unable to get around this
  problem.

	~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 
ALL DISPLAY PROGRAMS (EDITIMAGE, ANIMATE, WARPIMAGE, etc) -

*  A bug in the X server will cause the server to dump when a display routine
   attempts to create a private colormap.  The bug has been fixed, but at the
   cost of user interface colors no longer being defined as the original colors.
   This implies that on any display routine, your user interface colors will
   be undefined if the programs cannot use the default colormap;  in this 
   case, your user interface colors may disappear, may be arbitrarily defined,
   or may have their colors re-defined according to the current colormap.
   

EDITIMAGE -

*  when using interactive LUT you must click twice (instead of once)
   to set the slope of the RGB curve.  Click 3 times to exit (getting the
   crosshairs back)

	~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 


======================  24-Plane Displays : =========================


ALL DISPLAY PROGRAMS -

* For read-only color display routines (putimage, iconimage, warpimage, 
  animate) if the default visual is of type TrueColor, these applications
  will use that visual rather than trying to find a DirectColor visual.
  A side effect of using a TrueColor visual is that the colors are not given
  precisely;  instead, the server will provide a closest-color representation 
  instead.  On most color displays, this is not very accurate; for example,
  a greyscale image with 256 colors may appear as a color image with 40 
  colors, where colors are grey-ish but not exactly grey.

	~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 
EDITIMAGE -

* Editimage will quit with an XProtocol error about XCreateWindow.  

	~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 
  
VIEWIMAGE -

* Viewimage will come up;  however, it will not be able to display colors;
  the only available plots will be 3D, horizon, mesh, and wireframe.  It will
  not be able to display imagery (as this requires colors) and all other plots
  will come up in black.
