Skip to content

Releases: ethereum/go-verkle

v0.2.2

30 Oct 14:28
7dd8079
Compare
Choose a tag to compare

What's Changed

  • revert deactivation + import fixes from 454 by @gballet in #461

Full Changelog: v0.2.1...v0.2.2

v0.2.1

25 Oct 16:36
af95552
Compare
Choose a tag to compare

What's Changed

  • disable broken serialization optimizations by @jsign in #460

Full Changelog: v0.2.0...v0.2.1

v0.2.0

25 Oct 14:27
a80b12b
Compare
Choose a tag to compare

What's Changed

New Contributors

Full Changelog: v0.1.0...v0.2.0

v0.1.0

12 Sep 18:54
354ce60
Compare
Choose a tag to compare

The following list of changes is pretty long since we'll start creating releases with vX.Y.Z format.

Most notable changes are marked in bold, but there are a lot of:

  • Performance improvements.
  • Breaking APIs.
  • Proof format changes: proofs now contains pre-state and post-state.
  • Changes in precomputed tables representation: no file is required.

Future tagged versions will contain a more digestible set of referenced PRs.

What's Changed

  • README: fix precomp file referencing by @jsign in #336
  • Enable batch serialization by @jsign in #333
  • Batch Point->Fr transformation in leaf node value updating & proof items generation by @jsign in #337
  • update (Des|S)erializeProof to the electra proof format by @gballet in #324
  • mod: update go-ipa with optimized sqrt by @jsign in #339
  • Cache EmptyCode hash point in polynomial commitment by @jsign in #340
  • remove all references to InsertOrdered by @gballet in #341
  • tree: fix global pointer mutation by @jsign in #342
  • add a helper to dump a tree to JSON by @gballet in #344
  • Overlay tree migration explorations by @jsign in #343
  • use UnknownNode in InternalNode to turn it into a stateless node by @gballet in #345
  • deleting from a missing node is a noop by @gballet in #352
  • Improve test coverage by @jsign in #350
  • fix: set nodes known to be absent to Empty, to differentiate from unknown nodes by @gballet in #353
  • tree: simplify GetProofItems for LeafNode by @jsign in #355
  • mod: update to faster go-ipa by @jsign in #360
  • implement proper node deletion in tree by @gballet in #357
  • GetProofItems now get the proof values from the tree by @gballet in #361
  • remove panics in StemFromBytes and FromLEBytes by @gballet in #366
  • Remove all references to children in InternalNode & use uncompressed serialization by @gballet in #354
  • mod: update go-ipa by @jsign in #375
  • fix: after keyvals param removal, GetProofItem returns one nil value per stem by @gballet in #376
  • revert the hashing of nodes in commit by @gballet in #380
  • Remove needing a file persisting precomputed tables by @jsign in #369
  • Make GetProofItems capable of resolving hashed nodes by @gballet in #377
  • fix: check resolver func is present in GetProofItems by @gballet in #385
  • Parallelize Commit() and converted leaf node insertions by @jsign in #379
  • Update go-ipa & other nits by @jsign in #386
  • Update to cleaned up go-ipa APIs by @jsign in #387
  • mod: go-ipa panic removal by @jsign in #389
  • proof: use safe deserialization method by @jsign in #390
  • create proofs with post state by @gballet in #384

Full Changelog: banderwagonv3...v0.1.0

Banderwagon precomputed lagrange points v3

17 Mar 05:14
35cc2d3
Compare
Choose a tag to compare

After a rework of the precomputed Lagrange points by @jsign the performance has improved, at the cost of a larger file.

Precomputed Lagrange points for Banderwagon v2

30 Jun 17:23
4e4c3f6
Compare
Choose a tag to compare

The curve was changed from Bandersnatch to Banderwagon. This release provides the pre-computed Lagrange points. Drop the precomp file in the directory that your program is running in, in order to avoid the file generation on the first run.

Precomputed Lagrange points for Banderwagon

26 May 09:51
8b21c55
Compare
Choose a tag to compare

The curve was changed from Bandersnatch to Banderwagon. This release provides the pre-computed Lagrange points. Drop the precomp file in the directory that your program is running in, in order to avoid the file generation on the first run.