679cdd20ce
make sure to exit c method when trusted setup is not loaded
2022-11-24 12:10:56 +00:00
3136083139
using jni instead of swig (still needs work)
2022-11-24 09:14:12 +00:00
6cc5cb07e0
Try -fPIC
2022-11-22 12:56:05 -08:00
bcc138560a
Update interface for verify_kzg_proof
2022-11-19 09:35:55 +00:00
9611bfde9a
Add const
2022-11-19 09:31:36 +00:00
8820b7a659
Remove n == 1 special case
2022-11-18 22:32:41 +00:00
27f772402e
Make compute_powers easier to read
2022-11-18 22:20:18 +00:00
deb29eb59f
Fixes
...
- Switch ordering of domain separator
- When copying the polynomial into the bytes array, use the `j` value to position the pointer
2022-11-18 00:16:10 +00:00
084fd2209e
Add n=0 special case for compute_aggregate_kzg_proof
2022-11-08 23:11:19 +00:00
fec7acb87e
Handle n < 2 cases better in compute_aggregated_poly_and_commitment
2022-11-08 20:15:48 +00:00
27f0d7fb25
Remove unused Polynomial typedef from interface
2022-11-04 23:29:33 +00:00
d007bf79d0
Do not change the default behavior of src `make`
2022-11-04 13:44:04 -07:00
1d20797874
Regen dist
2022-11-04 12:06:06 -07:00
062e4e0f30
Simplify, given updated API
2022-11-04 11:56:41 -07:00
ac65930b15
Use BYTES_PER_FIELD_ELEMENT
2022-11-04 11:44:57 -07:00
8ca4fd9e83
Merge branch '4844_3038' into dgcoffman/nodejs-bindings
2022-11-04 11:41:17 -07:00
99a990ff32
Add some of the bytes conversions back to the interface
...
These might still be needed for the G1 elements even though they're not
needed for blobs now.
Also make FIAT_SHAMIR_PROTOCOL_DOMAIN static to avoid redefinitions.
2022-11-04 09:05:39 +00:00
f819fe1f20
Patch blst submodule sha
2022-11-04 00:13:02 -07:00
be0f51bd7e
make blst now works
2022-11-03 23:38:17 -07:00
a29b1580bd
Make task for blst
2022-11-03 23:23:27 -07:00
1514d5b826
computeAggregateKzgProof + verifyAggregateKzgProof test passes
2022-11-03 16:20:33 -07:00
b8151db670
Cleanup
2022-11-03 15:13:49 -07:00
672346f017
Closer to working
2022-11-03 14:39:20 -07:00
a5ca06450f
Some cleanup
2022-11-03 13:31:02 -07:00
fcd7fbd9e7
Checkpoint
2022-11-03 12:57:59 -07:00
2b0f6a5c09
Update src for merged-3038 specs
...
bindings/tests not updated
2022-11-03 17:16:57 +00:00
726fd38a5e
Revert unnecessary platform-specific makefile change
2022-11-02 13:32:15 -07:00
d478234491
Revert changes that were only to make SWIG work
2022-11-02 13:30:19 -07:00
b169d4ab6e
Remove SWIG
2022-11-02 13:27:00 -07:00
3b0d68036b
Fix bug in compute_powers
2022-11-02 15:51:06 +00:00
002fc3fa54
Tweaks
2022-11-01 21:58:21 -07:00
203033613f
I learned about extern C
2022-11-01 21:39:04 -07:00
6ec86af547
Generate bindings using SWIG and node-gyp
2022-11-01 15:00:00 -07:00
270ba19030
Add back trusted_setup.txt
...
was lost in the move
2022-11-01 14:36:23 +00:00
434c9ae9b9
Remove unused code, remove min- prefix
2022-11-01 14:24:02 +00:00
30864e643f
Remove stride from eval_poly_l
...
To match b611c51874
2022-10-03 20:49:45 +01:00
aab3df643b
Use pointers for scalars too
2022-10-02 12:09:01 +01:00
2281cce711
Use array of pointers for vector_lincomb
...
This seems more likely to avoid copying field elements when interfacing
from high-level languages (we still have to store the pointers though).
If this is the right move, the same could be done for
blob_to_kzg_commitment, and maybe others that read sequences.
2022-10-02 11:29:23 +01:00
b284636afd
Remove 4844 version from src (use min-src)
2022-10-02 11:12:38 +01:00
7dcd70b95e
Use non-contiguous array for vector_lincomb
2022-10-01 16:23:41 +01:00
96e3fc9bd8
Add blob_to_kzg_commitment to the 4844 interface
2022-09-27 13:51:08 +01:00
48903d567f
Use bit-reversal permutation
2022-09-27 13:37:46 +01:00
efef81bdf1
Fix segfault
2022-09-21 08:08:02 +01:00
7e1dc4ddef
Add test of load_trusted_setup_test
...
not working yet: segfaults after the test runs
2022-09-20 23:43:38 +01:00
2958210f67
Trusted setup for testing
2022-09-20 23:04:18 +01:00
e18c68c64c
Add possible implementation of load_trusted_setup (untested)
2022-09-20 21:58:36 +01:00
a7eea8b533
Start on a 4844 minimal interface
...
The object files to implement this are not correct in the Makefile.
The files probably need to be split more carefully to separate 4844
things from the rest.
2022-09-19 23:18:42 +01:00
63fc842d6d
Make y const in check_proof_single
2022-09-19 23:17:28 +01:00
cc049a05ca
Fix some whitespace issues
2022-09-19 19:49:11 +01:00
c0aa4eb2cc
Add fr_vector_lincomb function
2022-09-19 19:47:16 +01:00
8e5209c491
Work example up to evaluating a poly_l
2022-09-19 15:40:17 +01:00
67c6a61a18
Fix compiler warnings in zero_poly.c
2022-09-18 12:20:39 +01:00
928e9360c0
Add docs for compute_proof_single_l
2022-09-18 12:16:35 +01:00
56f40fdfcf
Fix documentation for compute_proof_multi
2022-09-18 12:14:54 +01:00
25be720f62
Add some documentation
2022-09-18 09:51:55 +01:00
8cdc4e62e3
Remove some trailing whitespace
2022-09-18 08:59:32 +01:00
80b256ad5f
Use batch inversion for Lagrange commitment
2022-09-17 23:25:21 +01:00
a2a4e57622
Lagrange proof bench
2022-09-17 18:34:59 +01:00
f7053514b3
Free temporary arrays
2022-09-17 18:21:56 +01:00
dcf68949b3
Optimized eval_poly_l with batch inversion
2022-09-17 18:00:54 +01:00
4c8787d87b
Fix fr_batch_inv
2022-09-17 17:06:54 +01:00
9c36ca4f3d
WIP: Montgomery batch inversion
2022-09-17 16:24:14 +01:00
3aff3eb727
Implement proof for Lagrange at a root of unity
2022-09-17 13:05:12 +01:00
c4537ccdae
Make y a parameter of compute_proof_single_l
2022-09-17 11:00:13 +01:00
3d286c6050
Add proof_single_l to tests
2022-09-17 10:59:39 +01:00
94c7bb56cf
Add another test at a different poly_l and root
2022-09-17 10:53:25 +01:00
10bff7d5c8
Fix eval_poly_l for evaluation at a root
...
Move the TODO for the special formula to compute_proof_single_l
2022-09-17 10:49:04 +01:00
f2454c284c
Add barycentric benchmark to Makefile
2022-09-17 09:27:41 +01:00
30f5911e14
Benchmark for barycentric formula
2022-09-17 09:20:08 +01:00
279eb5b514
Add TODO for barycentric formula special case
2022-09-17 07:23:00 +01:00
12c3cb7645
Remove tmp2
...
This is possible assuming the fr_ ops don't need to read their inputs
after writing to their output.
2022-09-16 15:16:12 +01:00
8eda02ab72
Remove tmp3 and unnecessary op
2022-09-16 15:16:12 +01:00
509e75b1d5
Add new_poly_l to interface
2022-09-16 15:16:11 +01:00
f7aa56679b
Remove unused variable
2022-09-16 15:16:11 +01:00
c8b18c3a22
Free allocated coeffs, avoid leak
2022-09-16 15:16:11 +01:00
0861520a7e
Test for proofs computed in Lagrange form
2022-09-16 15:16:11 +01:00
7f1a25bbc5
First pass at compute_proof_single_l
2022-09-16 15:16:11 +01:00
15f2b9eef9
Add evaluation of Lagrange poly
2022-09-16 15:16:11 +01:00
56e0769ce4
Add Lagrange form, with a test
2022-09-16 15:16:11 +01:00
6b5140af66
Fixed test for check_proof_multi
2022-09-16 15:15:58 +01:00
9bd8c3fdf8
Change coset_scale in test proof_multi
2022-09-15 13:35:02 +01:00
f2f20141db
Stop adding one to secrets_len
2022-09-15 13:33:17 +01:00
fa80d94168
Add newlines at ends of some files
2022-09-15 13:32:08 +01:00
63612c1119
Fix toeplitz_coeffs_stride() for minimal sized poly ( #18 )
...
Fix an edge case in FKMulti when the chunk length was the same as the polynomial size, resulting in writing outside the return array in `toeplitz_coeffs_stride()`.
Co-authored-by: Ben Edgington <ben@benjaminion.xyz>
2021-08-20 16:07:16 +01:00
350a961b90
Fix condition check ( #17 )
2021-08-11 11:13:08 +01:00
950d53c325
Nashatyrev feature/add g2 add dbl ( #16 )
...
* Add g2_add_or_dbl() function
* Add unit test
Co-authored-by: Anton Nashatyrev <anton.nashatyrev@gmail.com>
2021-08-11 10:06:32 +01:00
b21d13684b
Consolidate header files ( #14 )
...
* Consolidate header files
User should now need only include c_kzg.h and bls12_381.h.
* Update README
2021-07-09 13:35:19 +01:00
ca1acf61d6
Fix up docs (include static functions)
2021-07-09 09:54:37 +01:00
1f2c946ce8
Merge pull request #12 from Nashatyrev/fix/double-include-guard
...
Add double header include guard for kzg_proofs.h
2021-07-09 06:26:15 +01:00
e8b140a88b
Comment end of #if
2021-07-09 06:23:35 +01:00
13d523a079
Merge pull request #13 from benjaminion/testing_refactor
...
Refactor unit testing
2021-07-09 06:19:35 +01:00
1712cd83bb
Add newline at the end of file
2021-07-08 17:59:32 +03:00
0c8c174a31
Add double header include guard for kzg_proofs.h
2021-07-08 17:58:15 +03:00
f6b1d0ff57
Refactor unit testing
2021-07-03 13:55:54 +01:00
144d81b2c1
Fix up documentation
2021-07-03 11:45:39 +01:00
27bde1aff7
Fix initial stride in das_fft_extension
2021-07-03 10:50:56 +01:00
958dc1aee7
Minor tweaks and add a unit test
2021-07-03 09:37:20 +01:00
fc641a2197
Add fr_to_uint64s() function
2021-07-02 14:48:36 +03:00
ba81164330
Implements fast division for long divisors ( #7 )
2021-06-28 12:42:14 +01:00
244bfe8740
Add ASSERT macro
2021-06-24 16:22:37 +01:00
2cf8782e91
Add length check for polynomial commitment
2021-06-23 11:51:50 +01:00
37021b9653
Use Blst's multiscalar multiplication for commitments
2021-06-22 19:52:44 +01:00
7f8b81717d
Add KZG_CFLAGS env var to compiler options for passing additional compiler options ( #6 )
...
* Add KZG_CFLAGS env var to compiler options for passing additional compiler options
2021-06-14 13:00:17 +01:00
32becfefcf
Change the primitive root from 5 to 7 ( #5 )
2021-06-11 14:22:47 +01:00
9e4cfa4325
Move specific include dir to Makefile from .h
2021-06-10 18:52:36 +03:00
0924f243b4
Simplify reduction loop
2021-04-28 19:56:22 +01:00
f25ada5ea9
Micro-optimisation
2021-04-20 12:28:30 +01:00
e7b4e9f06d
Rename shift_poly to scale_poly
2021-03-08 12:06:25 +00:00
e9537b29a3
Fix benchmark
2021-03-06 13:17:07 +00:00
42013d27a0
Fix missing initialisation
2021-03-06 13:16:32 +00:00
5edef551f4
Test for division by zero
2021-03-06 13:15:57 +00:00
7d3d422005
Improve comments
2021-03-04 15:10:28 +00:00
9d1b622f21
Make zero poly work for large numbers of missing indices
...
Previously, like the Go code, calculating the zero polynomial would fail
for very lare numbers of missing indices. For example, 253 missing with
a domain size of 256 - this is where the number of partials flips from 4
to 5 and more working space is needed.
With this commit, the zero polynomial can be calculated right up to all
but one of the indices missing. The case with all indices missing
doesn't work as the return data is too large, but the solution is known
to be `x^width - 1` in case we need to know.
2021-03-03 21:51:32 +00:00
f09d1a70b2
Handle many missing (step 3)
2021-03-03 17:43:31 +00:00
8fbea3b3ef
Handle many missing (step 2)
2021-03-03 11:48:56 +00:00
80af76581d
Handle many missing (step 1)
2021-03-03 08:44:49 +00:00
e2cbccdc9d
Start improving the docs
2021-03-01 11:39:18 +00:00
a28860ae39
Benchmarks for recover and zero_poly
2021-02-28 17:01:00 +00:00
79a9419791
Reduce memory usage
2021-02-28 17:00:05 +00:00
52e620467a
Minor tweaks
2021-02-28 13:09:59 +00:00
2c22bb9dae
Implement recovery from samples
2021-02-27 15:19:46 +00:00
345a16bf8a
Remove in-place FFTs
2021-02-26 14:44:31 +00:00
a697775c4f
Implement zero polynomial - passes tests, but could be tidier
2021-02-26 14:38:04 +00:00
5cc7ea488c
Implement in-place FFTs for field elements
2021-02-22 21:36:43 +00:00
e2c74624d7
Add functions to allocate multi-dimensional arrays
2021-02-22 17:03:16 +00:00
3d6fd31610
Improve tests
2021-02-21 12:53:39 +00:00
159fc0024d
Collect various utility functions into a separate file
2021-02-20 16:03:04 +00:00
7fd0899972
Cast `void *` to something before doing arithmetic on it
2021-02-20 15:26:16 +00:00
5bd6f1f6bc
Implement DAS extension
2021-02-17 12:25:03 +00:00
39e190984f
Change ASSERT to CHECK and always fail with C_KZG_BADARGS
2021-02-17 12:22:35 +00:00
f557f32ed1
Improve G1 multiplication time with Blst
2021-02-17 09:06:17 +00:00
c10e1f40f1
Tidy up include files
2021-02-16 19:42:27 +00:00
66c50a49b1
Groundwork for being able to use alternative BLS libraries
2021-02-16 18:13:20 +00:00
aec19b5eee
Implement FK20 multi proofs
2021-02-15 20:48:16 +00:00
cfe5fa6e49
Revamp error handling and memory management
2021-02-14 14:20:03 +00:00
d22bbcca62
Introduce new_ prefix for functions that allocate memory
2021-02-14 08:28:37 +00:00
06afe430a7
Initialise FK20 multi proofs
2021-02-13 10:27:36 +00:00
9dc2611ac9
Improve docs
2021-02-12 19:49:43 +00:00
19f58f25fa
Documentation and tidy up
2021-02-12 16:55:42 +00:00
9111f47f13
Allocate space in new_kzg_settings for the secrets
2021-02-12 12:11:48 +00:00
ed33a391f5
Documentation and tidy up
2021-02-12 12:01:11 +00:00
ef4be2309d
Move field element division to blst_util
2021-02-11 20:01:57 +00:00
167a39cc4c
Correct argument names
2021-02-11 19:41:10 +00:00
8497b7bc4b
Documentation
2021-02-11 19:36:35 +00:00
7b023e10c4
The beginnings of documentation, yay
2021-02-11 11:43:08 +00:00
1a196dd748
Simplify by removing the unnecessary extended polynomial
2021-02-11 07:43:49 +00:00
5dc78d40f2
Move some function arguments around
2021-02-10 13:02:22 +00:00
f476a0f6ce
Add FK20 single proofs
2021-02-10 11:55:38 +00:00
8986f1cde3
Fix strided FFT
2021-02-10 11:15:35 +00:00
ea18b23bf5
Fix bug with strided FFT
2021-02-10 10:39:50 +00:00
Stefan Bratanov
dancoffman
Ramana Kumar
kevaundray
Dankrad Feist
Anton Nashatyrev
Ben Edgington
Anton Nashatyrev