* Add basic block processing benches
* Start reviving state processing benches
* Fix old block builders
* Add optimization for faster pubkey add
* Tidy benches, add another
* Add extra block processing bench
* Start working on faster BLS scheme
* Add partially complete sig verify optimization
* Add .gitignore to state processing
* Add progress on faster signature verification
* Fix SignatureSet for fake_crypto
* Tidy attester slashings sig set
* Tidy bulk signature verifier
* Refactor signature sets to be cleaner
* Start threading SignatureStrategy through code
* Add (empty) test dir
* Move BenchingBlockBuilder
* Add initial block signature verification tests
* Add tests for bulk signature verification
* Start threading SignatureStrategy in block proc.
* Refactor per_block_processing errors
* Use sig set tuples instead of lists of two
* Remove dead code
* Thread VerifySignatures through per_block_processing
* Add bulk signature verification
* Introduce parallel bulk signature verification
* Expand state processing benches
* Fix additional compile errors
* Fix issue where par iter chunks is 0
* Update milagro_bls dep
* Remove debugs, code fragment in beacon chain
* Tidy, add comments to block sig verifier
* Fix various PR comments
* Add block_root option to per_block_processing
* Fix comment in block signature verifier
* Fix comments from PR review
* Remove old comment
* Fix comment
There was a discrepancy between the is_empty fields of fake signatures during testing, so I've added
a small hack to set the is_empty field of a fake signature based on the byte content.
Alternatively, we could just make it so that any fake signature is defined to be equal to any other.
