namada-net/namada v201.0.6

Namada 201.0.6

on GitHub

BUG FIXES

• Fix refund when timeout on unlimited channel
  (#4822)
• Fixed state loading from DB to ignore tx gas map from previous block
  in line with its removal from in-memory state after block commit.
  (#4843)
• Integrate an updated nam-bellperson that fixes issues in wasm,
  namely the usage of std::time::Instant, which has been replaced
  with wasmtimer::std::Instant. Moreover, silence a lot of the
  INFO log lines previously sent to the CLI, when generating proofs.
  (#4878)

FEATURES

• Extends the ShieldedWallet to also compile the
  history of shielded transactions if requested.
  (#4705)
• Added support for MASP frontend providers sustainability fees.
  (#4790)

IMPROVEMENTS

• Removed handling of failed atomic batches from handle_inner_tx_results that
  was never hit. (#4450)
• Querying the conversion state becomes more difficult as nodes progress
  to higher and higher epochs. For instance, on a mainnet clone running on
  accelerated epochs, client queries for conversions always timeout (even after
  modifying timeout_broadcast_tx_commit) when the node goes beyond a certain
  MASP epoch. This happens because the conversion state grows linearly with
  the MASP epoch counter. This PR attempts to address this problem by making
  the conversions RPC endpoint require that clients specify the MASP epoch
  they want conversions from. This implies two things: first the size of the
  response from the conversions RPC endpoint should now be constant (equal to
  the number of tokens in the shielded rewards program), and second a client
  requiring all conversions now has to do a separate query for each MASP epoch.
  (#4776)
• Optimize shielded sync and shielded wallet layout.
  (#4785)
• The masp now uses the bellperson backend to synthesize circuits and build zk
  proofs. Benchmarking this indicates substantial speed ups in proving time.
  (#4817)
• Updated tiny-hderive crate to replace unmaintained memzero crate with zeroize.
  (#4818)
• Closes #4714
  • When dry-running, dummy signatures are used in
    txs
  • Validation does not perform signature checks when dry-running

I have tested that the gas estimation hasn't changed between dry-running
with signatures and with dummies. It is curious however that in both cases there is a small discrepancy between dry-running and the actual tx.
(#4714)

SDK

• Updated the transaction building process to avoid using MASP conversions with
  no rewards. (#4789)
• Reworked SDK wrapping and signatures, including some breaking changes.
  More specifically:
  • The wrapper arguments have been extracted into a separate type and their
    presence signals the need to wrap the tx
  • The dump and dry-run arguments have been turned into enumerations
  • The wrapper signer data has been removed from SigningTxData and moved into
    SigningWrapperData
  • Simplified the interface of aux_signing_data
  • Removed redundant dispatcher functions
  • Prevent casting from a wrapped to a raw transaction type
  • Prevent submitting an unwrapped transaction
  • Avoided passing the MASP internal address as a transaction's owner
  • Updated the interface of build_batch
  • Removed the owner for reveal_pk
    (#4816)