This utility displays a 3 dimensional image of a magic cube -- similar to the original introduced to the world by its Hungarian inventor, Erno Rubik. The user rotates sections of the cube, and attempts to solve the puzzle, by moving all the blocks to their correct positions and orientations, resulting in a cube with a single colour on each face.
Invoke GNUbik, by typing its name gnubik
into a command line.
Or you can set up a window manager menu, or file manager application, to run
it.
Whatever the chosen method of starting GNUbik, it must have access to a display
with which it can render its images. Being an inherently graphic program, it won't do much
otherwise.
GNUbik accepts all the standard X resource overrides such as
-geometry wwxhh+xx+yy
as well as all those specific to the program itself.
See Options.
GNUbik is graphic, in both its display, and its interface. User interactions are immediately made visible on the display.
The following options may be specified on the command line at invocation. You can access some of these parameters from a menu whilst the program is running.
-z n
--size=n
-a m
--animation=m
-s
--solved
-h
--help
-v
--version
GNUbik uses the Athena widget set. Any widget's resource may be set to control menu labels, background colours etc.
In addition, the following application resources set the colours of the cube faces:
! Colours of the cube faces ! resource default ! gnubik.color0: Red gnubik.color1: Green gnubik.color2: Blue gnubik.color3: Cyan gnubik.color4: Magenta gnubik.color5: Yellow
You can change the viewing position of the cube using the mouse or the keyboard.
You can re-orientate the cube using the following keys. The <Up> and <Down> keys rotate the cube about the horizontal axis. The <Left> and <Right> keys rotate the cube about the vertical axis. Holding down the <Shift> key whilst pressing the <Left> or <Right> arrow keys rotates the cube about the z axis.
Another way to rotate the whole cube is to use the mouse. Hold down button 1, and move the mouse anywhere on the window. This way, you can rotate the cube horizontally and vertically, If your mouse has a scroll wheel, it can be used to rotate the cube about the z axis.
To manipulate the cube, you need to select a block to move, and a direction to turn it. Place the mouse cursor on a block to select it, then move the cursor against the edge of the block towards the direction you want to turn. Click Button 1 when you're ready to turn make the move. At all times the mouse cursor will point in the direction of motion.
As well as manipulating the cube by hand as described in the previous chapter, you can also run scripts that perform automatic manipulations right in front of your eyes. Currently there are scripts available which randomize the cube, and which solve 3x3x3 cubes.
When GNUbik starts up, it reads files ending in a .scm
suffix
in the .../share/gnubik/scripts
directory (system-wide files)
and in the .gnubik/scripts
directory under your home directory
(personal files). These files are scheme scripts, and when they run
they register themselves with the system so that items appear on the
main GNUbik menu under the heading Script-fu
. Making selections
from this menu runs the appropriate script.
GNUbik's script-fu is implemented with GUILE
, GNU's ubiquitous
extension language, which in turn implements scheme
, a
programming language descended from lisp
.
In simple words, a script works as follows. When GNUbik starts up, it
looks for scripts in standard places and executes each one it
finds. In turn, the scripts define some procedures, and register some
of their defined procedures with GNUbik, which associates the procedure with
an item under the Script-fu
menu.
When you select a menu item, GNUbik executes the procedure that was registered. The procedure will then request an object from GNUbik that reflects the current state of the cube. The procedure then works out a list of moves that it wants to impose on the cube, and sends this list to GNUbik. During the development of this list, the script may manipulate the cube reflection object so that it can keep track of the moves it makes. However, changes to the reflection object are never seen directly by the GNUbik program.
Once the script exits, GNUbik will cause all the movements which the script requested to take place on the on-screen cube, right in front of your eyes!
When GNUbik starts up, it scans the following directories in the order
given for files with a .scm
extension, and executes them in the
order found (not necessarily alphabetical!) The directories searched
are
.../share/gnubik/guile
.../share/gnubik/scripts
~/.gnubik/scripts
Note that the exact location of the share directory is system
dependent. The share/gnubik/guile
directory is for scripts
which the GNUbik program itself needs to function. They should not be
changed once the GNUbik package is installed in a system. The
share/gnubik/scripts
directory is the place for system
administrators to put scripts that all users can see on their GNUbik
menus. Finally, the .gnubik/scripts
directory is located under
your home directory, and may contain any scripts that you have
written, and which will only appear on the menu when you start the
program up.
When the system finds a script, that script is executed and must use the following function to register itself with an item on GNUbik's menu bar.
gnubik-register-script menu-path code | Procedure |
menu-path should be a string indicating how the script should
appear under the Script-fu menu, using / characters to
indicate sub-menus and _ characters in front of a letter to
indicate that that letter should be used as the keyboard accelerator
for the menu item. For example, the string solvers/my _solver
will create an item under the solvers sub-menu, with s
as the keyboard shortcut.
The second argument code should be a string specifying
|
This chapter describes the basic interface that GNUbik provides to the
scheme
world. This interface can be used with any cube geometry
that GNUbik supports, but is not intuitive for humans and the API may
change in the future.
If the API does change (to make it more consistent!) the version number of cube geometry in a cube object will also change, so any script which uses these functions directly should obtain a cube object from GNUbik and check the geometry version number.
In order to obtain a reflection of the current state of the cube, a script must call
gnubik-cube-state | Procedure |
This will return a scheme object, which is in fact a cons cell
whose first component is a list describing the geometry of the cube,
and the second component is a vector of six vectors of nxn items,
where n is the size of the cube.
The geometry list consists of a version number, which will change if
the definition of the cube object ever changes (scripts should check
this to ensure compatibility or else things will surely break); the
current version is 1 (one). The next item in the list is the
dimensionality of the cube, which currently is always 3. The list then
holds three more numbers describing the size of the cube in each
direction. So a current 3x3x3 cube will have the geometry list
The second component of the cube cons cell is a vector of six vectors, one for each face. If you were to look at a cube with face index 4 towards you (remember vector indeces are zero based), then the top face would have index 0 (zero), the bottom 1 (one), the left 2, the right 3, and the back 5. Got it? Stay there. On faces 0 (zero), 3, and 4, the panels are numbered from left to right and top to bottom, so that panel 0 (zero) is at top-left and panel 5 is middle-right, for example. On the other faces, the panels are in the corresponding positions underneath their opposite faces, which means in practice that the numbering is from right to left, top to bottom. For example, the second panel (index 1) on face 2 is at the top. Finally, the 'top' edge of the top panel is against face 2. The elements of the vectors for each page represent the colour that the panels of that face are showing, numbered from 0 (zero) to 5 inclusive (the correspondence between these numbers and the actual colours seen on the cube on the screen is not provided by GNUbik to the scripts). |
Whew.
If that was too sticky and you are using a 3x3x3 cube, then you should proceed to the next chapter which presents a more comfortable (and stable) interface.
Once you have a cube object, the following procedures are available to facilitate access to it.
get-colour cube face index | Procedure |
This procedure gets the colour of the cube at panel index on face, where the meaning of the latter two arguments is described above. |
set-colour! cube face index colour | Procedure |
This procedure sets the colour of the given face to colour. Note
that the change takes place only in the scheme cube
object, not in the actual cube that the user is seeing on her screen.
|
When a script has determined a set of moves which are to be made on a cube, it calls the following procedure to get GNUbik to put the moves into its internal move buffer.
gnubik-rotate-animated move-list | Procedure |
The move-list is a list of lists of three items. For any one
move (a quarter turn of a slice of the cube), the first item in the
specification is the face which lies parallel to the desired slice,
and may be 0 (zero), 1 (one) or two corresponding to the back, right
and bottom faces. The second item in the specification is the slice
number parallel to that plane; the slice next to the specified face
has number (n-1 ) where n is the size of the cube, and
the slices below are separated by (-2). For example, for a cube with
three blocks per edge (a 3x3x3 cube, say), the slices are numbered 2,
0 and -2. The final component in a move specification may be 0 (zero)
or 1 (one) depending on whether the slice should be moved
anticlockwise or clockwise, respectively.
|
Note that the above function may be called as many times as desired during the execution of a script, but that the moves will not actually take place until the script finishes.
gnubik-rotate-fast move-list | Procedure |
This performs the same function as above, except that the moves take place instantly as soon as the script finishes, instead of being animated on the screen. |
The following functions provide access to a cube object in a way which hides the gory details of the representation. This makes it easier for human beings to write code, and also guards against changes in the low-level API described in the previous chapter.
Unfortunately, it currently only works on 3x3x3 cubes.
The interface uses the common FLUBRD
notation to refer to the
various faces and panels on the cube. The letters correspond
respectively to the words front
, left
, up
,
back
, right
and down
, and where up
and
down
can be read as top
and bottom
, respectively.
The following procedure returns the colour on the panel addressed symbolically by the letters of a given string.
get-colour-symbolic cube symbol | Procedure |
The symbol is a string consisting of one, two or three letters. If one letter is supplied, the colour of the panel at the centre of the face indicated by the letter is returned. If two letters are supplied, the colour returned is of the edge panel touching the face given by the second letter. If three letters are supplied, it is the colour of the corner panel which touches the edges indicated by the second two letters. |
Note that the letters of the symbol must all be lower case. A corresponding function is supplied to make changes to the cube object.
set-colour-symbolic! cube symbol colour | Procedure |
This procedure sets the colour of the panel indicated by symbol, defined as above, to colour. |
Note that changing the colour of a cube does not cause any other changes to take effect. In particular, no rotations of the on-screen cube are initiated.
The following function allows scripts to request that faces of the cube are rotated by a quarter turn.
move-face cube symbol | Procedure |
Here, symbol is a string of one or two characters. The first
character should always be a lower case letter which indicates which
face to move. The default is to turn the face clockwise. This will
also occur if the second character supplied is a space or a
+ . Only if a second character is supplied and is a -
will the face be turned anti-clockwise.
|
Once a script has finished declaring moves in this way, GNUbik will send them to the display buffer for on-screen animation after the script calls
execute-move-buffer! | Procdure |
Cause all moves created with the move-face procedure to be applied
to the on-screen cube.
|
Note that the on-screen cube will not actually display any change until the script exits.
The move buffer can be accessed directly via the symbol
move-buffer
. It is a list of moves that the low-level procedure
gnubik-rotate-animated
expects. Accessing and manipulating it
directly before calling execute-move-buffer!
may allow a script
to perform some post-processing optimizations on the list of moves,
for example. Alternatively, if the script decides to duck out or start
again, the following procedure may be called to clear the buffer
without sending any move notifications to GNUbik.
clear-move-buffer! | Procedure |
Empty the move-buffer without sending any move requests to the GNUbik program. |
In order to eliminate the need to shutdown, startup, and then make selections from the menus every time a change is made to a script under development, you can create a script which reads another script every time it is invoked. This way any changes made to the latter script will manifest themselves immediately, speeding up the development cycle.
Suppose your development script is /path/to/dev_script
(this
can be anywhere at all in the filesystem). You can create a menu entry
Dev. test
to load and run this script by putting the following
lines into a file called ~/.gnubik/scripts/dev-test.scm
(define (run-dev-test) (load "/path/to/dev_script") (run-dev-script-proc)) (gnubik-register-script "Dev. test" "run-dev-test")
and then restarting GNUbik. Note that run-dev-script-proc
is
the name of the procedure you would normally have run directly from
the menu (i.e. it is the one you would have passed to
gnubik-register-script
).