Avoid deriving CanonicalSerialize/CanonicalDeserialize for relevant structs and overriding default implementations with concrete ones to get the expected behaviour (e.g. deserialize_uncompressed calls the deserialize(compressed) if the function is not overridden, but that's not what we may want) (please review this carefully, we want to make sure that no (unexpected) default implementation is called);
Making sure that canonical_serialize_without_metadata() will act like the previous ToBytes/FromBytes (preparing for a future PR in which we are going to remove them): meaning all data is saved in uncompressed form and no info regarding length or types of data structures are saved (please review this carefully too, as we currently use the to_bytes! in Fiat Shamir sponges);
Add extra ci jobs for asm and target-feature;
Added missing assertions on poly degrees;
Refactored index() function in circuit_specific_setup()and get_index_info() functions;
Refactored IndexInfo.num_variables in two new fields: num_witness and num_inputs;
Added new negative test case where proof fails in hard part;
canonical_serialize_without_metadata()will act like the previous ToBytes/FromBytes (preparing for a future PR in which we are going to remove them): meaning all data is saved in uncompressed form and no info regarding length or types of data structures are saved (please review this carefully too, as we currently use the to_bytes! in Fiat Shamir sponges);index()function incircuit_specific_setup()andget_index_info()functions;IndexInfo.num_variablesin two new fields:num_witnessandnum_inputs;