helium / erlang-hbbft

83%

master: 83%

LAST BUILD ON BRANCH adt/acs-return-and-send

branch: adt/acs-return-and-send

CHANGE BRANCH

x

Reset

• adt/acs-return-and-send
• adt/acs-not-return
• adt/append-to-txn-buffer
• adt/bba-count-term-against-bval
• adt/bba-value-cache
• adt/better-stats
• adt/decrypt-performance-fixes
• adt/defer
• adt/fix-term-message-delivery
• adt/get-set-stamp-fun
• adt/handle-failed-decryption
• adt/inputs-and-stamps-are-binaries
• adt/message-ordering
• adt/next-round
• adt/one-more-acs-completion-check
• adt/stamp-fun
• adt/temporal-hacking
• adt/temporal-hacking-rebased
• adt/term-message
• adt/travis-test
• adt/txn-filter-fun
• fakecast
• master
• pevm/inherit-buf-round
• rg/bba-bug
• rg/common-tests
• rg/fix-dealer
• rg/hbbft-python-issue-59
• rg/helium-119-type-specs
• rg/helium-92-serialize-deserialize
• rg/limit-txns
• rg/otp21
• rg/protocol-audit
• rg/relcast-test
• rg/serialize-states
• rg/status-check

Build # 308

Build Type

Pull #29

travis-ci

Committed by web-flow

Commit Message

Allow ACS to return a result and messages to send simultaneously

After observing cases where all the required BBA and ACS instances
completed, but ACS itself did not return. I believe this is caused by
the final BBA completing, returning the final 1 that puts us over the
N-F threshold. However, the current code path assumes that the final BBA
won't be the final one to return 1, and also assumes that ACS can send a
message, or return a result, but not both.

This change allows ACS to return a result AND send messages, and updates
HBBFT itself to handle that return value. This along with the prior
change to ACS to check for a return in the RBC handling code, should
catch the missing termination conditions.

Pull Request Pull Request #29: Allow ACS to return a result and messages to send simultaneously

Run Details

7 of 7 new or added lines in 2 files covered. (100.0%)

371 of 449 relevant lines covered (82.63%)

36947.57 hits per line