ruby or shell magic needed

[joerg]

Hi,
i am doing some research on password lists for a security tutorial. now i am with psychology of passwords for a quite while now. What came to my attention lately is the idea to automatically create a passphrase list from a huge text file with an ebook for example. Now to be clear, lets take an example. I need a little ruby magic or shell magic which does THIS: I want to input a text file and want the script to output the list of the passphrases from within this text. When a "." appears it should include the dot in the current passphrase, do a CR and let the next passphrase begin. Example:

Text Input Textfile:
The quick brown fox jumps over the lazy dog. Always look on the bright side of life. Hello World. Blablabla. This is a test.

Now the output file must look like this:
Tqbfjotld.
Alotbsol.
HW.
B.
Tiat.

I guess that is pretty easy - but as i am not quite familiar with that stuff - please can anyone take five minutes to figure that out? Would be highly appreciated.

best regards,
Joerg

[wzzbx]

This will do it:
#!/bin/bash

for i in `cat ${1}`
do echo -ne ${i:0:1}
if [[ ${i: -1} = "." ]]
then echo .
fi
done

[Ubscene]

Don't reinvent the wheel. wyd.pl was made on that purpose! (you can find it on /pentest/passwords/wyd/)

[joerg]

Thank you both so much - i will first check the shellscript and then have a look at wyd.pl

best regards,
Joerg

[joerg]

Maybe I am blind but i can´t locate wyd.pl at the location you pointed me to earlier. In Addition i read the wyd.pl specs on Max´s Site and can´t find a way which will do what i want. It simply extracts the words and allows some slick nice modifications but as i see it it can´t do what i asked for.
best regards,
Joerg

[whitelisted]

sed 's/\([a-zA-Z]\)[a-zA-Z]*/\1/g; s/ //g; s/\([^a-zA-Z]\)/\1\n/g' <input

[joerg]

holy....that is great, whitelisted. Could you point me a way which will include the comma sign "," in the passphrase? To clarify:
Input example: "The quick brown fox eats, poops and sleeps! Then he runs to the house."
Should result in Output example:
"
Tqbfe,pas!
Thrtth.
"

best regards,
Joerg

[whitelisted]

In your example, the first passphrase ends at an exclamation mark, not a period. How am I supposed to know which symbols mark the end of a passphrase and which ones should be embedded?

Assuming that every passphrase ends with a non-alpha character and begins with an uppercase character, you could try this:


sed 's/\([a-zA-Z]\)[a-zA-Z]*/\1/g; s/ //g; s/\([^a-zA-Z]\)\([A-Z]\)/\1\n\2/g' <input

[joerg]

yeah i could have pointed that out - sorry my fault.

thank you so much. never thought you could use sed for stuff like that. great!

[whitelisted]

No, thank you.

It's easy to get jaded about linux and to want to chuck it in and go back to a system that "just works" (tm). It's nice every now and again to solve a complex real-world problem in 30 seconds or less and using standard, out of the box tools, and to know you never could have done that in a windows environment.

Your problem made my day: it reminds me why I love open systems.