xlaby - X Labyrinth, the aMAZEment

xlaby [ options ]

X Labyrinth is a labyrinth game under X11 that is played directly with the mouse pointer: the walls block the pointer's movement on the screen.

The goal of the game is to retrieve the four colored squares: to retrieve a square, it is sufficient to move the pointer over it, and it will disappear. However, to make things more infuriating, the squares have to be taken in the following order: red, yellow, green and blue. When the blue square is obtained, the game is won.

Unless otherwise specified (see OPTIONS below), it is possible to quit the game at any moment by pressing Q or escape and to suspend the game by pressing Z or tab (same key to resume).

If the maze window is suddenly obscured or changes size or position, it is possible to restore it to the correct position, size and visibility by pressing the space bar.

It is possible to chose between three types of maze building algorithms (see OPTIONS below): the "tree", "flood" and "chain" maze.

Exit code is 1 in case of error, 42 if the maze was successfully solved, and 0 otherwise. This can be useful in shell scripts for example to test the user's patience.

-b, --blind

Run in blind mode. You don't get to see the walls of the maze, which makes it slightly harder.

--build, --slowbuild

Show the maze as it is being built. This is the best possible explanation of how the maze building algorithms work.

-c, --chain

Use the chain maze building algorithm: the maze is built by a self-avoiding random walk that starts from start every time it is blocked.

--copying

Display copying information and exit. See the file COPYING for more information.

-d, --discover

Run in semi-blind mode. The walls become visible only when you bump into them.

-display name

Run on the specified display. Overrides the DISPLAY environment variable.

--grabkbd

Grab the keyboard as long as the game is being played. This inhibits window manager hotkeys for example; frequently used in conjunction with --noquit

-f, --flood

Use the flood maze building algorithm: the maze is built by a self-avoiding random walk that starts from the last possible place every time it is blocked.

-h, --help

Print help information and exit.

-i, --info

Print game instructions and exit.

--noplay

Do not play the game (only display the generated maze).

--noquit

Forbid quitting the game (i.e. disables the Q and Z keys). Beware, this can be dangerous, especially when used in conjunction with --grabkbd.

-p, --play

Play the game. This option is compulsory (so that people won't accidentally start xlaby and then not know how to quit).

--pos X Y

Define the X and Y position of the maze window. If this parameter is omitted (or invalid values are given), window position will be chosen at random.

-q, --quantum

Run in quantum mode: every time the pointer changes cell, it has a certain probability of doing a quantum leap to another position in the maze. It is uncertain whether that actually makes things easier or more difficult.

-s, --size number

Set the maze size. number is a number between 0 (very small) and 9 (huge). Note that the maze will always fit in the screen: if it is too large, the cell size will be reduced, and if that is still not sufficient, the maze dimensions will also be reduced. Note that xlaby will not run if the dimension of the root window is less than 40x40 or so :-(

-t, --tree

Use the tree maze generating algorithm: the maze is built by growing trees of walls from the edges and the center. It is very easy to solve, and also quite slow for large sizes.

-v, --version

Print version number and exit.

-w, --warranty

Print absence of warranty and exit. See the file COPYING for more information.

Note that multiple options cannot be concatenated, so xlaby -pt is not valid: use xlaby -p -t instead.

Also note that standard X toolkit options are not recognized, as the program uses low-level X.

DISPLAY

The default display name.

None!

Option syntax is very non-standard. Having to use the --play option is a pain.

X resources are not recognized. The colors cannot be specified. Only the position of the window can be given, and it does not interact properly with the window manager (only I do not trust WMs).

There are probably cases which I overlooked and in which the window is not fully visible. Unfortunately, as far as I know, the Xlib does not furnish any easy way of making absolutely sure a window is entirely visible.

I guess it must look awful on a monochrome display.

The tree maze is outrageously slow for large maze sizes.

The game is about unplayable when run across a network as the mouse cursor may freeze for a long time on a cell boundary.

There is no high score table. So the name of the first person who completes the size 9 chain maze in blind mode shall not go down in History (still, if you do, send me a mail).

The blue square (randomly placed) may overlap another one. Actually, this is more a feature than a bug.

There is no way to change the number, colors, or position of the squares.

The man page does not describe all the bugs (otherwise this section would be the longest by far).

Peter Horvai (peter.horvai@ens.fr) for the initial idea and chain maze algorithm.

David Madore (david.madore@ens.fr) for most of the programming, the tree maze algorithm, and this man page.

Special thanks to Mladen Dimitrov for the flood maze algorithm.

GNU public license. See the file COPYING for more information.

maze(1)