This adds up the last missing cooldown feature. This prevents a malicious actor from updating a proof with a better one in a short time frame (dfefault 1h).
Ref T1854.
Depends on D10788.
Details
Details
- Reviewers
tyler-smith - Group Reviewers
Restricted Project - Maniphest Tasks
- Restricted Maniphest Task
- Commits
- rABC2e1c1bc1d747: [avalanche] Add a peer replacement cooldown
ninja all check-all
Diff Detail
Diff Detail
- Repository
- rABC Bitcoin ABC
- Branch
- avalanche_peer_replacement_cooldown
- Lint
Lint Passed - Unit
No Test Coverage - Build Status
Buildable 17840 Build 35506: Build Diff lint-circular-dependencies · build-without-wallet · build-diff · build-debug · build-clang-tidy · build-clang Build 35505: arc lint + arc unit
Event Timeline
| src/net_processing.cpp | ||
|---|---|---|
| 4482 ↗ | (On Diff #31752) | It seems safer to me to have this logic wrapped into a function instead of pairing the right string ("-avalanchepeerreplacementcooldown") to the right default var (AVALANCHE_DEFAULT_PEER_REPLACEMENT_COOLDOWN) every time you might want to use this arg. |
| src/net_processing.cpp | ||
|---|---|---|
| 4482 ↗ | (On Diff #31752) | Agreed but at this point there is a single call site. If I need to add more it will be a good idea to encapsulate this. |
Comment Actions
The logic and implementation are good. The negative case added to the tests appears sufficient to check the new behavior.
| src/net_processing.cpp | ||
|---|---|---|
| 4482 ↗ | (On Diff #31752) | Sure. Really this is a wider observation; this pattern is used a lot and it seems error-prone. Perhaps in another diff we can hash out a new standard pattern. |