This is pretty much work in progress, but includes tests. Just run 'make test' to get an idea of what is supported and 'make perftest' for performance comparisons.
I've changed the basic method to get passwords and comparable tests show anywhere from 1.5 to 3 times improvement (depending on the revision) on a dual-core with HT machine.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
That is so cool! I really like that you have added public tests, which has been very much needed (I have a private one, but can unfortunately not share it).
I just released 0.15 which contains all the small bits and bobs fixed during the last year, mostly so that you can pick it up and continue on your end.
As the progress on this project have pretty much stalled (I neither have the time or interest to put in it), I really wish you best of luck and hope that you can pick up the torch.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
It looks like there is a new PDF security revision 6 in play... If you're not working on this any more I'm happy to give rev 6 a shot, but may need your help. First thing would be from where can I get the specs on the rev 6 security?
The specifications for PDF 1.7 extension level 8 are not public as far as I am aware. I am also not aware of any open implementation of the encryption standard 6 in PDF.
I am not working on it and will not deal with it anytime soon.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Am pretty much done with the patterns related work. It now supports regex like character classes with support for ranges of repitition for each (in addition to the basic list of chars patterns).
You can now run something like this test in the Makefile:
Hi, this is great. I'm trying to find a password on an old Dell with 4x Xeon processors and will benefit immensely from the multi cpu support.
However, as per your notes only the pattern search benefits from the multi core support but I can't for the life of me formulate a pattern which simply mimics the regular "random with increasing character number". Is that something you can help me with?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi
I've forked the latest pdfcrack to enable simple patterns (e.g. [abc][abc][123][123]) and multi-core support based on OpenMP.
This works for all PDF security revisions (including Rev 5) and covers both user and owner passwords.
Pls see this: https://github.com/shreepads/pdfcrack-mp/tree/baseopenmp
This is pretty much work in progress, but includes tests. Just run 'make test' to get an idea of what is supported and 'make perftest' for performance comparisons.
I've changed the basic method to get passwords and comparable tests show anywhere from 1.5 to 3 times improvement (depending on the revision) on a dual-core with HT machine.
That is so cool! I really like that you have added public tests, which has been very much needed (I have a private one, but can unfortunately not share it).
I just released 0.15 which contains all the small bits and bobs fixed during the last year, mostly so that you can pick it up and continue on your end.
As the progress on this project have pretty much stalled (I neither have the time or interest to put in it), I really wish you best of luck and hope that you can pick up the torch.
Thanks Henning, yes the tests have been very handy.
I will update the master and merge 0.15 in... All the changes I make are in the 'baseopenmp' branch BTW...
Have you looked at this issue logged on the pdfcrack-ng clone on github?
https://github.com/x2q/pdfcrack-ng/issues/2
It looks like there is a new PDF security revision 6 in play... If you're not working on this any more I'm happy to give rev 6 a shot, but may need your help. First thing would be from where can I get the specs on the rev 6 security?
EDIT: Pls also see https://github.com/shreepads/pdfcrack-mp/issues for list of to-do items...
Last edit: Shreepad S 2015-09-26
The specifications for PDF 1.7 extension level 8 are not public as far as I am aware. I am also not aware of any open implementation of the encryption standard 6 in PDF.
I am not working on it and will not deal with it anytime soon.
Check out this page though for someone who claims to have reversed engineered it with a pseudo code implementation:
http://esec-lab.sogeti.com/posts/2011/09/14/the-undocumented-password-validation-algorithm-of-adobe-reader-x.html
I have not looked into it though.
Thanks! That seems useful... Now to find someone who can give me a test doc!
Am pretty much done with the patterns related work. It now supports regex like character classes with support for ranges of repitition for each (in addition to the basic list of chars patterns).
You can now run something like this test in the Makefile:
./pdfcrack -t 4 -e "[:upper:]{1,2}[:lower:]{2,3}[:digit:][:punct:]" ./testpdfs/TestPDF7.pdfThis will use 4 threads of execution in parallel to find a password that
Next up, am planning to add a character class [:word:] which will include words from a wordlist so you can have a pattern like this
{4,8}[:word:]{2,3}[:digit:][:punct:]Word is going to be special
Still need a test PDF file for the new PDF security revision 6...
Last edit: Shreepad S 2015-11-22
Hi, this is great. I'm trying to find a password on an old Dell with 4x Xeon processors and will benefit immensely from the multi cpu support.
However, as per your notes only the pattern search benefits from the multi core support but I can't for the life of me formulate a pattern which simply mimics the regular "random with increasing character number". Is that something you can help me with?