Comments on: MP's fabulous hash function, and its family http://trilema.com/2017/mps-fabulous-hash-function-and-its-family/ Moving targets for a fast crowd. Mon, 06 Apr 2026 03:37:06 +0000 http://polimedia.us hourly 1 By: PeterL http://trilema.com/2017/mps-fabulous-hash-function-and-its-family/#comment-122686 PeterL Fri, 18 Aug 2017 17:48:32 +0000 http://trilema.com/?p=70955#comment-122686 http://archive.is/cJNp3 Updated again, this time it should work! Line 126 had the wrong logic after line: def cyclecheck(m, oldbits, R, S, L): - if R.val(m) == 1 and S.length() > 1 and S.val(-1) == 1: + if R.val(m) == 1 and S.length() > 1 and S.val(-1) == 0: http://archive.is/cJNp3

Updated again, this time it should work!

Line 126 had the wrong logic after line:

def cyclecheck(m, oldbits, R, S, L):

- if R.val(m) == 1 and S.length() > 1 and S.val(-1) == 1:
+ if R.val(m) == 1 and S.length() > 1 and S.val(-1) == 0:

]]> By: Mircea Popescu http://trilema.com/2017/mps-fabulous-hash-function-and-its-family/#comment-122676 Mircea Popescu Thu, 17 Aug 2017 20:19:18 +0000 http://trilema.com/?p=70955#comment-122676 Cool deal. Cool deal.

]]> By: PeterL http://trilema.com/2017/mps-fabulous-hash-function-and-its-family/#comment-122671 PeterL Thu, 17 Aug 2017 17:56:10 +0000 http://trilema.com/?p=70955#comment-122671 An updated version of the reverse hash with a couple bug fixes: http://archive.is/Wl893 The bugs: 1: the "show" function in the Register class was missing the comparison with the "_inverted" variable, so it would be wrong in any case where there was an odd number of inversions. I think I left that out when I originally wrote the mpfhf function because the R never gets inverted, and originally the function only returned the R value instead of this version which returns both the R and S values. 2: I added a check on the branches that involve expanding to make sure S would never be shrunk below a length of 0. 3: There is a special case where the rewind does not happen if we are on step 0, I added that to the "check" recursive function. Further testing is welcome. An updated version of the reverse hash with a couple bug fixes:

http://archive.is/Wl893

The bugs:
1: the "show" function in the Register class was missing the comparison with the "_inverted" variable, so it would be wrong in any case where there was an odd number of inversions. I think I left that out when I originally wrote the mpfhf function because the R never gets inverted, and originally the function only returned the R value instead of this version which returns both the R and S values.

2: I added a check on the branches that involve expanding to make sure S would never be shrunk below a length of 0.

3: There is a special case where the rewind does not happen if we are on step 0, I added that to the "check" recursive function.

Further testing is welcome.

]]> By: Mircea Popescu http://trilema.com/2017/mps-fabulous-hash-function-and-its-family/#comment-122637 Mircea Popescu Tue, 15 Aug 2017 20:34:49 +0000 http://trilema.com/?p=70955#comment-122637 To quote his lordship vulpes, noice! To quote his lordship vulpes, noice!

]]> By: PeterL http://trilema.com/2017/mps-fabulous-hash-function-and-its-family/#comment-122631 PeterL Tue, 15 Aug 2017 19:14:12 +0000 http://trilema.com/?p=70955#comment-122631 A python implementation of the mpfhf and a reverse hash function to go with it. http://archive.is/jmFMO A python implementation of the mpfhf and a reverse hash function to go with it.

http://archive.is/jmFMO

]]>