Blsct arith classes with tests #54

gogoex · 2022-08-21T09:29:03Z

For issue #42

Reviewed the original implementation of Scalar and G1Point and made necessary changes
Added tests for each exposed methods of Scalar and G1Point
Added Elements class that represents a vector of Scalar and G1Point classes
Added operator overloads to Scalar, G1Point and Elements classes to improve readability of code
Confirmed that Scalar, G1Point and Elements provide enough functionalities to implement a range proof by adding a test that implemented a range proof

…sVec size don't match

…p test

…ng functions to Elements

aguycalled · 2022-08-22T15:16:23Z

One of the tests fails for me:

% ./src/test/test_navcoin 
Running 587 test cases...
test/blsct/arith/integration_tests.cpp:30: error: in "test_integration_invert_zero": check x.Pow(zero) == x.Pow(zero.Invert()) has failed

*** 1 failure is detected in the test module "Navcoin Core Test Suite"

Added some log:

    std::cout << strprintf("%s\n", HexStr(zero.GetVch()));
    std::cout << strprintf("%s\n", HexStr(zero.Invert().GetVch()));

Output:

0000000000000000000000000000000000000000000000000000000000000000
72af98d20d0c60cebe24f97a434290039b9673f3e0b1bb7a981cf9380e9f5524

…rs::FirstNPow

…goex/navcoin into blsct-arith-classes-with-tests

gogoex · 2022-08-23T05:35:14Z

@aguycalled The test was originally created to test FirstNPow for a case where y is 0:

    // (64)
    auto hhp = hh * Scalars::FirstNPow(n, y.Invert());

but now i think the test is not necessary since y is in Z^*_p and non-zero.

Also since multiplying any element by its invese results in unit element 1 and multiplying any element by 0 results in 0, I think the inverse of 0 in Z_p is undefined. It probably just happes to return 0 on x86_64 implementation.

So I removed the test and added another test case for Scalars::FirstNPow to test if it always return alist that starts with 1 regardless of the value of base value instead.

Can you try again?

aguycalled · 2022-08-23T07:02:59Z

@aguycalled The test was originally created to test FirstNPow for a case where y is 0:
    // (64)
    auto hhp = hh * Scalars::FirstNPow(n, y.Invert());
but now i think the test is not necessary since y is in Z^*_p and non-zero.

Also since multiplying any element by its invese results in unit element 1 and multiplying any element by 0 results in 0, I think the inverse of 0 in Z_p is undefined. It probably just happes to return 0 on x86_64 implementation.

So I removed the test and added another test case for Scalars::FirstNPow to test if it always return alist that starts with 1 regardless of the value of base value instead.

Can you try again?

All the tests pass for me now 👍

% ./src/test/test_navcoin                       
Running 586 test cases...

*** No errors detected

Maybe, Scalar::Invert() should throw an exception if the value is zero.

mxaddict · 2022-08-23T07:43:11Z

I think this would be a good idea, since we are not supposed to use a zero value in Scalar right?

gogoex · 2022-08-23T07:58:03Z

sure. now calling Invert on Scalar(0) throws an exception.

…goex/navcoin into blsct-arith-classes-with-tests

aguycalled

Approving and merging. Pending solving the CI issues in a different task #61.

* Added the new classes Scalar and G1Point * Added scalar.{h,cpp} and g1point.{h,cpp} to make file * Implemented more of the operators * Implemented more functions * Added new << >> logic for bitwise ops * implemented Scalar::{Rand,GetString,Hash} * Updated the arith scalar logic * Added more G1Point arith logic, math... * Added hashMapTo * Removed Scalar::hashAndMap * Updates that I forgot to push * Updated G1Point::Rand() * Update the mulVec func in G1Point to throw and exception if gVec and sVec size don't match * Removed the test code that was accidentally added * Implemented mulVec * Removed some unused header calls * Added return statement for G1Point::Double * add g1point/scalar and their tests * ignore .vscode * add endianness parameter to G1Point::MapToG1. add g1point_hash_and_map test * add mulvec_mcl test * add mulvec tests * add scalars and g1points classes * add MapToG1(string&) * extracted common parts of Scalars and G1Points to Elements. add slicing functions to Elements * wip * wip * drop G1point::operator^ * drop the 2nd type artument from Elements * add skeleton elements tests * add elements tests * support non-zero rand generatation * wip * add Elements::FirstNPowers * add .gitignore to exclude dynamically generated files on macos. add Elements::RandomVector * wip * add Scalar::Pow and Scalar::GetBits * add range proof test (not tested) * add type parameter where required * calculate t1, t2 properly * fix typo * add Elements::RepeatN test * add comment * add integration_tests and move such tests there * introduce std::is_same in elements.h * add gg^z == gg^(ones * z) * wip * add working h^mu gg^l test * add Scalars::FirstNInvPow * add multiply inv power seq for G1Points * wip * wip * drop FirstNInvPow. add working 65_g_part_ts_only test * wip * add working range proof test * wip * make the range proof procedure a function * integrate inner product argument to range proof * wip * wip * wip * wip * drop unused variable * separate elements implementation to cpp * clean up code * wip * wip * ignore .vscode add endianness parameter to G1Point::MapToG1. add g1point_hash_and_map test add mulvec_mcl test add mulvec tests add scalars and g1points classes add MapToG1(string&) extracted common parts of Scalars and G1Points to Elements. add slicing functions to Elements wip wip drop G1point::operator^ drop the 2nd type artument from Elements add skeleton elements tests add elements tests support non-zero rand generatation wip add Elements::FirstNPowers add .gitignore to exclude dynamically generated files on macos. add Elements::RandomVector wip add Scalar::Pow and Scalar::GetBits add range proof test (not tested) add type parameter where required calculate t1, t2 properly fix typo add Elements::RepeatN test add comment add integration_tests and move such tests there introduce std::is_same in elements.h add gg^z == gg^(ones * z) wip add working h^mu gg^l test add Scalars::FirstNInvPow add multiply inv power seq for G1Points wip wip drop FirstNInvPow. add working 65_g_part_ts_only test wip add working range proof test wip make the range proof procedure a function integrate inner product argument to range proof wip drop unused variable wip wip wip separate elements implementation to cpp clean up code wip wip * squashed commits * fix typo * migrate indirect MulVec to Elements * fix build issues in clang * fix spelling errors * add newly introduced boost depdendency to lint-includes.py * remove local depdendent functions * wip * remove trailing spaces * Added BLS and MCL libs/includes to kernel build * Updated copy constructor and assignment operators * Added clean steps to clean-local for mcl and bls libs * FORCE mcl and bls libs to build using main project CC and CXX values * replace test_integration_invert_zero by additional test case of Scalars::FirstNPow * use fixture test suite to provide name for bls/arith tests * throw exception when Scalar(0) is inverted * include bls/arith headers in setup_common.h * add g1point and mcl_initializer dependency to test_util source * fix test_util.include * add blsct/arith/scalar depedency to libtest_util * Disabled LLVM use in MCL compilation, this disabled ASM * Added NM to android host configs for depends builds * Added ABI=32 for i386 builds of GMP in depends * fix undefined sanitizer error Co-authored-by: mxaddict <mxaddict@codedmaster.com>

fixes for CI builds related to PR #54

mxaddict and others added 30 commits June 30, 2022 12:34

Added the new classes Scalar and G1Point

da43f1b

Added scalar.{h,cpp} and g1point.{h,cpp} to make file

f8e8204

Implemented more of the operators

4aa3f19

Implemented more functions

10d1888

Added new << >> logic for bitwise ops

1a4f1b1

implemented Scalar::{Rand,GetString,Hash}

922a22c

Updated the arith scalar logic

e461ccb

Added more G1Point arith logic, math...

e060728

Added hashMapTo

94bf729

Removed Scalar::hashAndMap

44ca8c6

Updates that I forgot to push

ab05856

Updated G1Point::Rand()

012a87f

Update the mulVec func in G1Point to throw and exception if gVec and …

51d0097

…sVec size don't match

Removed the test code that was accidentally added

d5722fd

Implemented mulVec

4decbba

Removed some unused header calls

ac4a619

Added return statement for G1Point::Double

a492293

merge mxaddict/blsct-arith-classes

7ddbf11

add g1point/scalar and their tests

fa387e2

ignore .vscode

14ae49e

add endianness parameter to G1Point::MapToG1. add g1point_hash_and_ma…

b46fc34

…p test

add mulvec_mcl test

de36f68

add mulvec tests

7eec32c

add scalars and g1points classes

c2aa3c8

add MapToG1(string&)

43a7735

extracted common parts of Scalars and G1Points to Elements. add slici…

deb1e31

…ng functions to Elements

wip

c84245d

wip

7738b41

drop G1point::operator^

3fedf29

drop the 2nd type artument from Elements

774c7dc

mxaddict mentioned this pull request Aug 22, 2022

Blsct arith classes #32

Closed

mxaddict and others added 6 commits August 23, 2022 00:05

Added BLS and MCL libs/includes to kernel build

d0f47fa

Updated copy constructor and assignment operators

70051a5

Added clean steps to clean-local for mcl and bls libs

dbed23e

FORCE mcl and bls libs to build using main project CC and CXX values

89517db

replace test_integration_invert_zero by additional test case of Scala…

3ee4a11

…rs::FirstNPow

Merge branch 'blsct-arith-classes-with-tests' of github.com.gogoex:go…

0abb70f

…goex/navcoin into blsct-arith-classes-with-tests

use fixture test suite to provide name for bls/arith tests

6f10f38

gogoex and others added 10 commits August 23, 2022 17:13

throw exception when Scalar(0) is inverted

13cad21

include bls/arith headers in setup_common.h

bb939f8

add g1point and mcl_initializer dependency to test_util source

c4f16ce

fix test_util.include

2ff5738

add blsct/arith/scalar depedency to libtest_util

ca4d2a5

Disabled LLVM use in MCL compilation, this disabled ASM

183641c

Added NM to android host configs for depends builds

689df3d

Added ABI=32 for i386 builds of GMP in depends

a167ca9

fix undefined sanitizer error

1498a40

Merge branch 'blsct-arith-classes-with-tests' of github.com.gogoex:go…

add45d3

…goex/navcoin into blsct-arith-classes-with-tests

aguycalled closed this Aug 25, 2022

aguycalled reopened this Aug 25, 2022

aguycalled approved these changes Aug 25, 2022

View reviewed changes

aguycalled merged commit 575880a into nav-io:master Aug 25, 2022

aguycalled pushed a commit that referenced this pull request Aug 30, 2022

Merge pull request #66 from mxaddict/fix-ci-issues-after-pr-54

4770ec6

fixes for CI builds related to PR #54

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Blsct arith classes with tests #54

Blsct arith classes with tests #54

Uh oh!

gogoex commented Aug 21, 2022

Uh oh!

aguycalled commented Aug 22, 2022 •

edited

Loading

Uh oh!

gogoex commented Aug 23, 2022

Uh oh!

aguycalled commented Aug 23, 2022

Uh oh!

mxaddict commented Aug 23, 2022

Uh oh!

gogoex commented Aug 23, 2022 •

edited

Loading

Uh oh!

aguycalled left a comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Blsct arith classes with tests #54

Blsct arith classes with tests #54

Uh oh!

Conversation

gogoex commented Aug 21, 2022

Uh oh!

aguycalled commented Aug 22, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

gogoex commented Aug 23, 2022

Uh oh!

aguycalled commented Aug 23, 2022

Uh oh!

mxaddict commented Aug 23, 2022

Uh oh!

gogoex commented Aug 23, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

aguycalled left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

aguycalled commented Aug 22, 2022 •

edited

Loading

gogoex commented Aug 23, 2022 •

edited

Loading