Make quorum unreachable if validator list expires (RIPD-1539) #2240
Conversation
Codecov Report
@@ Coverage Diff @@
## develop #2240 +/- ##
===========================================
- Coverage 70.11% 70.09% -0.02%
===========================================
Files 689 689
Lines 50800 50802 +2
===========================================
- Hits 35617 35611 -6
- Misses 15183 15191 +8
Continue to review full report at Codecov.
|
| @@ -377,7 +377,8 @@ ValidatorList::onConsensusStart ( | |||
| list.second.expiration <= | |||
| timeKeeper_.now().time_since_epoch().count()) | |||
| removePublisherList (list.first); | |||
| else if (! list.second.available) | |||
There was a problem hiding this comment.
Are all published lists required to have an expiration?
There was a problem hiding this comment.
| @@ -395,6 +395,9 @@ ValidatorList::removePublisherList (PublicKey const& publisherKey) | |||
| --iVal->second; | |||
| } | |||
|
|
|||
| iList->second.list.clear(); | |||
There was a problem hiding this comment.
Without this call to clear, would that mean if a published list with keys K1,K2 expired, then was republished with key K1, the keyListsings_ count for K2 would have been decremented twice? If so, is it worth adding a unit test to cover that case?
There was a problem hiding this comment.
Good observation. This added test fails without the clear
3bb20f9
|
Great test, still 👍 on the PR. |
👍
wilsonianb
added
the
Passed
wilsonianb
force-pushed
the
expire-val-list
branch
from
3bb20f9 to
b7aa316
Compare
|
Squashed and rebased on 0.80.0 |
|
Merged as 02059a2 |
No description provided.