A tool to find valid words in a WordCrack board.
After cloning, just run make CONF=Release. You will find the executable under dist/Release/<architecurte>/wordcracksolver-st.
You need to put a dict.txt and a config file alongside it.
- The
dict.txtmust contain a list of valid words. There is already a file for spanish words (which is not complete and contains some words not considered valid by WordCrack --mainly some forms of verb conjugation) - The
configfile has the boards letters, and the special tiles (double letter, double word, triple letter, triple word). Detail of theconfigfile:
letters= 16 letters, that compose the board: ABCDEFGHIJKLMNOP represent the board:
| A | B | C | D |
| E | F | G | H |
| I | J | K | L |
| M | N | O | P |
- Use the letter
qfor representingQuand#forÑ minLenandmaxLen: the minimum and maximun path length to compute. A word with N letters has N-1 path length. Thus,minLen 1means "words of 2 letters", andmaxLen 15means "words of 16 letters". In practice, paths lengths greater than 11 or 12 do not find any extra word, and add a lot of time of computation. Recomended values areminLen=4andmaxLen=10dl,tl,dwandtw: DoubleLetter, TripleLetter, DoubleWord and TripleWord modifiers. The upper left tile of the Board (Ain the example above) has coordinates "00",Bhas "01",Chas "02", ...,Ehas "10" and so on, up toPthat has "33". Enter the coordinates for each modifier using "," to separate. Example:
dl = 00, 11
tl =
dw =
tw = 31
This means that A and F have the DoubleLetter modifier, and that N has the TripleWord modifier.
-
The project can be opened with Netbeans.
-
There are 2 configs, as usual,
DebugandRelease. -
The
masterbranch is a single-thread version, that computes all the possible paths every time. There are 2 other branches: -
mtwith multi-threading (though I did not find it specially faster. I must be missing something) and -
path_preload: all the possible paths are already pre-computed. This saves some time, but uses more memory.