Comments for malb::blog

Comment on SAT Solvers for Sage by martinralbrecht

martinralbrecht — Mon, 27 May 2013 09:29:37 +0000

Not at present (as far as I remember) but “it’s easy: implement it and send us a patch”

Comment on SAT Solvers for Sage by Ruslan Kiianchuk (@zoresvit)

Ruslan Kiianchuk (@zoresvit) — Sun, 26 May 2013 16:51:13 +0000

Hello. Thanks for the tutorial!
Is there a way to make cryptominisat run several threads using the Sage interface?

Comment on M4RI 20121224 by tom denton

tom denton — Mon, 25 Mar 2013 07:50:33 +0000

Looks great! And thanks for looking into the compilation issue for those of us building on dinky ARM machines!

Comment on SAT Solvers for Sage by James McLaughlin

James McLaughlin — Sun, 20 Jan 2013 23:36:01 +0000

Okay, I’ve tried a few more times, and the explosion of error messages does indeed appear to be completely at random – so I was wrong about sage.sat.solvers import CryptoMiniSat being needed.

I’m guessing the line I was supposed to pick up on in the documentation was “If neither P, K nor C are provided, a random pair (P, K) will be generated.”, which would have indicated to me that this had a chance of causing the “inversion on the zero element”.

Comment on SAT Solvers for Sage by martinralbrecht

martinralbrecht — Sun, 20 Jan 2013 23:03:53 +0000

Have you read the docs I pointed to? sr.polynomial_system() constructs a random system of equations which has a certain possibility of being unsatisfiable because it contains a zero inversion.

Comment on SAT Solvers for Sage by James McLaughlin

James McLaughlin — Sun, 20 Jan 2013 22:29:50 +0000

All I can say is that
F,s = mq.SR(1,2,4,4,gf2=True,polybori=True).polynomial_system()
only works for me if I’ve previously typed
from sage.sat.solvers import CryptoMiniSat

Have you tried quitting Sage completely, restarting it and entering the code in “a converter for Boolean Polynomials” one line at a time?

Comment on SAT Solvers for Sage by martinralbrecht

martinralbrecht — Sun, 20 Jan 2013 21:59:28 +0000

No, see http://www.sagemath.org/doc/reference/sage/crypto/mq/sr.html#sage.crypto.mq.sr.SR_generic.sub_byte

Comment on SAT Solvers for Sage by James McLaughlin

James McLaughlin — Sun, 20 Jan 2013 21:51:05 +0000

Ah – further experimentation seems to suggest that
from sage.sat.solvers import CryptoMiniSat
needs to come BEFORE
F,s = mq.SR(1,2,4,4,gf2=True,polybori=True).polynomial_system()
in the source. The reason the line was working before was that I’d typed it in while trying one of the source code examples earlier on the page.

Comment on SAT Solvers for Sage by James McLaughlin

James McLaughlin — Sun, 20 Jan 2013 21:45:20 +0000

Just trying to retest – and a line of code which worked before has mysteriously started giving error messages:

sage: F,s = mq.SR(1,2,4,4,gf2=True,polybori=True).polynomial_system()
—————————————————————————
ZeroDivisionError Traceback (most recent call last)

/home/linuxuser/Sage/sage-5.5-linux-64bit-ubuntu_12.04.1_lts-x86_64-Linux/ in ()

/home/linuxuser/Sage/sage-5.5-linux-64bit-ubuntu_12.04.1_lts-x86_64-Linux/local/lib/python2.7/site-packages/sage/crypto/mq/sr.pyc in polynomial_system(self, P, K, C)
2094
2095 if ciphertext is None:
-> 2096 ciphertext = self(plaintext, key)
2097 elif ciphertext is False:
2098 raise TypeError, “type %s of C not understood”%(type(ciphertext))

/home/linuxuser/Sage/sage-5.5-linux-64bit-ubuntu_12.04.1_lts-x86_64-Linux/local/lib/python2.7/site-packages/sage/crypto/mq/sr.pyc in __call__(self, P, K)
1356 P = AddRoundKey(P, K)
1357
-> 1358 P = SubBytes(P)
1359 if get_verbose() >= 2:
1360 print “R[%02d].s_box %s”%(self.n, self.hex_str_vector(P))

/home/linuxuser/Sage/sage-5.5-linux-64bit-ubuntu_12.04.1_lts-x86_64-Linux/local/lib/python2.7/site-packages/sage/crypto/mq/sr.pyc in sub_bytes(self, d)
669 “”"
670 d = self.state_array(d)
–> 671 return Matrix(self.base_ring(), d.nrows(), d.ncols(), [self.sub_byte(b) for b in d.list()])
672
673 def sub_byte(self, b):

/home/linuxuser/Sage/sage-5.5-linux-64bit-ubuntu_12.04.1_lts-x86_64-Linux/local/lib/python2.7/site-packages/sage/crypto/mq/sr.pyc in sub_byte(self, b)
712 if not b:
713 if not self._allow_zero_inversions:
–> 714 raise ZeroDivisionError, “A zero inversion occurred during an encryption or key schedule.”
715 else:
716 return self.sbox_constant()

ZeroDivisionError: A zero inversion occurred during an encryption or key schedule.
——————————————————————————–
I’ve actually uninstalled Sage and then reinstalled both it and CryptoMiniSat trying to fix this, but to no avail.

Comment on SAT Solvers for Sage by martinralbrecht

martinralbrecht — Sun, 20 Jan 2013 20:29:55 +0000

I fixed the source code for the first issue you reported. The second works fine for me.