Hey guys!

I just wanted to post the sample code for dictionary based rainbow tables.
I had it finished for quite some time, but then forgot about it. :p
Its a rainbowtable generator which takes a dictionary file as a source, and then does word mangling rules on the words to generate new words based on the options. (password => password12, password => PaSSword etc)
It supports adding numbers in front of the word and in the end, full case permutation, title case, last character case mutation and 1337 permutations.
The 1337 permutations should probably be looked at. I have a feeling its not distribution the characters evenly.
Feel free to try it out. I attached both binary and source

To get help for rtgendict, just type rtgendict and it will show you the options.


Edit: Files updated to reflect the optimization changes

Nice!

I think I'll need to look at this while not drunk but ohh well.

wow... CanLeetPerm and GetLeetPerm are very slow. Try using a map (an array with 256 elements that have the mapped char in it). wow i just noticed that's messed up, fixing code.

It "looks" like it is evenly distribution. Granted I haven't immersed myself into the code but doing the binary search thing will basically make it evenly distributed unless there is a collision in the dictionary, but this will be a rare occurrence.

btw: "pow((double)2, (double)nChar)" is the same as doing "1 << nChar" but much slower since pow deals with doubles and << is a bit shift on integers.

Hi Sc00bz.

I really appreciate your feedback. I will try to implement your suggestions to speedup the code

I've updated the source with Sc00bz suggestions and reattached new source and binary to the first post

if i have a wordlist with 100.000 words (let's say no permutation is used) and i want my rainbow table to have 100% "success " (to cover all the words in the wordlist)
the chain count to be 655360 (10Mb) what would be the chain length?
and how to compute that value ?
thx

the dictionary based table must be sorted :-w ... include please rtsort in the archive...

Hi,
I have some problemens bye compiling under linux.

Openssl and libcurl are installt. I use g++ 4.2.3
I heart that itoa is not C/C++ Standartconform and are only avalibel under windows, is this right?
Or missed I something to include.
Sincerly goodvirus

Edit: I just looked around the code and found a lot of goto statements. Is this realy usefull? My programming professor would kill me...

wouldn't be nice if we could use 2 other files like : prefix.txt and suffix.txt to use with the words ? not only numbers
eg :
.com .net .im .org for suffix.txt
www. www2. http// .. etc for prefix.txt or anything else ?

the optimization found by sc00bz isn't working
rtgendict md5 mil-dic.txt 0 2400 6553600 test caseperm,6;1337perm,6;numprefix,4;numsuffix,4
tried this on the second upload of rdict and it creates lots of errors
on the first everything is working ok

You can use the standard rtsort from rainbowcrack.

The goto statements is from the original rainbowcrack code. I didn't make those.
To compile on Linux you need to remove the oracle hash routine. (Just comment out the stuff inside the hashing function in HashAlgorithm.cpp)



Maybe i will add it in a future release. My main thoughts of making the dictionary tables has been to perform full case mutation/leet mutation in a quick way.



Uhm.. im pretty sure i tested out the code before i released it.. strange.

I just tested it again, and its working fine here..
I have no idea whats wrong at your side..