20 Commits

Author SHA1 Message Date
Ethan Buchman
a22c962e28
TMHASH is 32 bytes. Closes #1990 (#2732)
* tmhash is fully 32 bytes. closes #1990

* AddressSize

* fix tests

* fix max sizes
2018-10-31 12:42:05 -04:00
Dev Ojha
56d7160606 Add ValidatorPubkeyTypes as a consensus param (#2636)
* Add ValidatorPubkeyTypes as a consensus param

Ref #2414

* update spec

* address anton's comment

* Switch to Validator and Validator Params

* Correct changelog entry

* Address bucky's comments!

* forgot to update changelog

* fix typo

* fix Params naming
2018-10-30 11:36:53 -04:00
Ethan Buchman
455d34134c
ADR-016: Add versions to Block and State (#2644)
* types: add Version to Header

* abci: add Version to Header

* state: add Version to State

* node: check software and state protocol versions match

* update changelog

* docs/spec: update for versions

* state: more tests

* remove TODOs

* remove empty test
2018-10-17 15:30:53 -04:00
Anton Kaliaev
e1538bf67e state: require block.Time of the fist block to be genesis time (#2594)
* require block.Time of the fist block to be genesis time

Refs #2587:

```
We only start validating block.Time when Height > 1, because there is no
commit to compute the median timestamp from for the first block. This
means a faulty proposer could make the first block with whatever time
they want.

Instead, we should require the timestamp of block 1 to match the genesis
time.

I discovered this while refactoring the ValidateBlock tests to be
table-driven while working on tests for #2560.
```

* do not accept blocks with negative height

* update changelog and spec

* nanos precision for test genesis time

* Fix failing test (#2607)
2018-10-12 01:03:58 -04:00
Ethan Buchman
6ec52a9233
types: cap evidence in block validation (#2560)
* cap evidence in block validation

* state: use table-driven test for ValidateBlockHeader

* state: test evidence cap

* fixes from review
2018-10-09 13:31:21 -04:00
Zarko Milosevic
7b88172f41 Implement BFT time (#2203)
* Implement BFT time

* set LastValidators when creating state in state helper

for heights >= 2
2018-08-31 19:33:51 -04:00
Dev Ojha
2756be5a59 libs: Remove usage of custom Fmt, in favor of fmt.Sprintf (#2199)
* libs: Remove usage of custom Fmt, in favor of fmt.Sprintf

Closes #2193

* Fix bug that was masked by custom Fmt!
2018-08-10 09:25:57 +04:00
Ethan Buchman
e1062a657f fixes for ProposerAddress
- state.MakeBlock takes a proposerAddr
- validateBlock only checks that the ProposerAddress is in the validator
  set
- fix raceyness from bad proposer test:
  - use privValidator to get the proposer address (instead of racy
    state)
  - note we had to remove the test that checked the correct proposer was
    included for higher rounds because we don't have a good way to test
    this with multiple consensus states and not using the
    privValidator.Address while calling createProposalBlock was a hack!
2018-08-05 15:19:21 -04:00
Anton Kaliaev
bec9d5cba9 add proposer address to block's Header
Refs #1134

Validation:

- ignored in block.ValidateBasic since it's stateful information
- checked in blockExec.ValidateBlock
2018-08-05 15:16:49 -04:00
Ethan Buchman
5453aa6169 Merge branch 'develop' into jae/literefactor4 2018-07-02 14:57:30 -04:00
Ethan Buchman
d55243f0e6 fix import paths 2018-07-01 22:36:49 -04:00
Jae Kwon
a5b7ea93c4 Delay validator set changes by 1 block. 2018-06-25 16:59:00 -07:00
Ethan Buchman
71556c62eb fixes from rebase 2018-06-05 22:14:37 -07:00
Ethan Buchman
3d2c4fd309 update Evidence type - requires pubkey and valset to verify and convert to abci.Evidence 2018-06-05 22:04:26 -07:00
Ethan Buchman
edb851280a state: b -> block 2018-06-04 21:20:23 -07:00
Ethan Buchman
dd62f06994 state: s -> state 2018-06-04 21:20:23 -07:00
Ethan Buchman
444db4c242 metalinter 2017-12-28 23:15:54 -05:00
Ethan Buchman
cb845ebff5 fix EvidencePool and VerifyEvidence 2017-12-28 23:15:54 -05:00
Ethan Buchman
6112578d07 ValidateBlock is a method on blockExec 2017-12-28 23:15:54 -05:00
Ethan Buchman
f82b7e2a13 state: re-order funcs. fix tests 2017-12-28 23:15:54 -05:00