Updates #2

AndrewQuijano · 2024-03-17T16:57:38Z

Add subtract from ciphertext for easy usage of y - [x]
Fixed Joye and Salehi comparison
Automate work on creating JAR file releases
Use Log4J logging instead of printing to standard out

…will probably make an Alice debug version sometime? In the meantime I need to fix Joye comparison

…gradlew works. Not sure why it isnt consistent on passing

…nded, phew had a scare. But I did clean up log message of protocol 1 in general.

… using DGK ciphertext about getting X >= Y compared to Veugen! Nice!

…tract ciphertext

…rate a release too. Although I think this should be restricted for only the main branch.

…ould be nice to have 1.0 match what I got after my REU, and 2.0 match all the work I did for Level-Site PPDT/VICOR

…in too

…ork, I will slowly work on getting this to work

…l bits

…otocol3 for Bob? Whatever, fixed now, but with this, that should mean if I fix just alice/bob, I get every other version for free

…ing early exit, need to poke if this means my encrypted_xor works for x < y and x > y. Also, figure should be consistent on for loop logic to use xor array length

…ified Protocol 3

…check

… bits, hence it always worked. Next step, check the encrypted XOR method

Fix timing attack

…parity, but I need to drill deeper within comparison protocol 1

…d to replicate

… because I need to know t-bits for both x and y.

…imes, I hope I finally killed this lasy bug on Joye and Salehi

…the fix

…ould be bug free

… truly fixed by getting the inputs that might be causing crashes

AndrewQuijano added 8 commits March 16, 2024 21:31

Made the switch to using Log4J to log

10158a6

Removing print to stdout, just find it under logs/application.log. I …

d3203ae

…will probably make an Alice debug version sometime? In the meantime I need to fix Joye comparison

Cleaned up how I am logging the name of alice/bob under testing, but …

3f4ff26

…gradlew works. Not sure why it isnt consistent on passing

logging last stack trace

e453e99

OK, I can confirm I have the Joye and Salehi protocol working as inte…

c05e455

…nded, phew had a scare. But I did clean up log message of protocol 1 in general.

Joye and Salehi works! More importantly, it is ALWAYS consistent even…

e1951e2

… using DGK ciphertext about getting X >= Y compared to Veugen! Nice!

So fixed typo on name for 40, and based on alans feedback I added sub…

809a239

…tract ciphertext

Going to poke around about releasing JAR file

eaa3ca1

AndrewQuijano requested a review from ogbautista March 17, 2024 16:58

AndrewQuijano added 5 commits March 17, 2024 13:17

Create tag and release

16da7a0

Disable build check just to check release works

6c04d75

Attempting with this

b151250

Hate doing a hate on version naming, but lets go

b7267c1

Reintroduce automated testing for library. Also, for now it will gene…

f3aced8

…rate a release too. Although I think this should be restricted for only the main branch.

AndrewQuijano self-assigned this Mar 17, 2024

AndrewQuijano and others added 15 commits March 22, 2024 15:56

Update action versions

35b82c1

Update to avoid using set-output as that is EOL

b3f4e22

Trying to fix how tag is made

b926906

maybe needed curly brace?

4c63b0c

going to copy Panda

308f383

Fix release

2d3d2e5

Want this to be Version 1

d87ce5a

Revive testing, I think I got the whole versioning working, I think w…

858a401

…ould be nice to have 1.0 match what I got after my REU, and 2.0 match all the work I did for Level-Site PPDT/VICOR

I need to make sure I ONLY create the release if the tests pass on ma…

1ebf663

…in too

Using serial tag.

bf8d3e7

Standardize how alice will always get the XOR

b150821

Fix XOR, timing attack if bits not equal is fixed now

8260471

For lots of other XOR operations I also had to modify the index

9c9bcdd

do not make version until test passes

e5e31bd

OK, so I broke a lot more than I realized, but basic alice seems to w…

f959cf7

…ork, I will slowly work on getting this to work

Andrew Quijano and others added 29 commits April 2, 2024 16:54

Joye is alive and DO NOT mess with the bit function

d2258f8

Revived Veugen, will create branch to fix the timing attack on unequa…

26a58e7

…l bits

How tf did I not notice the code duplication I deleted in Modified Pr…

cf61cb8

…otocol3 for Bob? Whatever, fixed now, but with this, that should mean if I fix just alice/bob, I get every other version for free

I noticed encrypted_equals check works just fine somehow without need…

c211e0b

…ing early exit, need to poke if this means my encrypted_xor works for x < y and x > y. Also, figure should be consistent on for loop logic to use xor array length

Make sure Alice_joye entirely uses xor length

817cc88

Apply XOR logic to Protocol 3, I am not brave enough to try it on Mod…

2d250a0

…ified Protocol 3

Decided to be brave and update for all to use XOR length

dfd09d5

I still want to know why encrypted_equals does NOT need the bit size …

c47d791

…check

The private equality check somehow ensured x and y had same number of…

c713eef

… bits, hence it always worked. Next step, check the encrypted XOR method

Timing attack fix applied to original alice. Will try fixing Joye next

854aeeb

Joye and Salehi is fixed too for XOR

353d4e6

bit alreday does padding

2b22413

Modified Protocol3 is corrected. Fixing Protocol 3 next

c157686

Merge pull request #1 from AndrewQuijano/timing

7daca58

Fix timing attack

Should leverage warn

323765c

I will allow encrypted negative, but warn users in logger

c53279d

Update test concat, seems XOR works well even with bigger integer des…

a6e102b

…parity, but I need to drill deeper within comparison protocol 1

Refactoring to allow for integration testing directly

353cea0

Edge case if both x and y is 0 bits long so XOR is 0 bits long

6977161

Got a unit test thing to help debug Joye, still turning out to be har…

3fe76dc

…d to replicate

more linting fixes

e42b4e4

Interesting, seems like I need to change how I use hamming weight now…

e13e210

… because I need to know t-bits for both x and y.

Updating rest of testing code. Going to re-run this test a few more t…

31da1f5

…imes, I hope I finally killed this lasy bug on Joye and Salehi

Seems like the issue was getting stuck about readObject, think I got …

4f4a44b

…the fix

BRUH, really, I missed THAT it should NOT be x? UGH. But, now this sh…

ef884a0

…ould be bug free

Reduce logging level

8861696

Want exception to be thrown if an assert fails

e9b5e7f

I want to debug why PPDT still seems to act up, confirm Protocol 1 is…

a89bc37

… truly fixed by getting the inputs that might be causing crashes

Add timing attack protections

a4e6dfb

AndrewQuijano merged commit 67d18a3 into adwise-fiu:main Apr 19, 2024
2 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Updates #2

Updates #2

AndrewQuijano commented Mar 17, 2024

Updates #2

Updates #2

Conversation

AndrewQuijano commented Mar 17, 2024