diff --git a/.changelog/5003.bugfix.md b/.changelog/5003.bugfix.md deleted file mode 100644 index 72b26347180..00000000000 --- a/.changelog/5003.bugfix.md +++ /dev/null @@ -1,4 +0,0 @@ -go/runtime/txpool: republish sooner if republish limit is reached - -This fixes a case where a part of a batch of transaction would take a long -time to be published in case there are no new transactions being submitted. diff --git a/.changelog/5005.bugfix.1.md b/.changelog/5005.bugfix.1.md deleted file mode 100644 index c3db957acbc..00000000000 --- a/.changelog/5005.bugfix.1.md +++ /dev/null @@ -1,5 +0,0 @@ -go/worker/common: Reorder state determination checks - -Otherwise the shown state would be misleading, e.g. showing that it is -waiting for runtime host being provisioned while it is actually blocked -in initialization like storage sync. diff --git a/.changelog/5005.bugfix.2.md b/.changelog/5005.bugfix.2.md deleted file mode 100644 index 552a10a62b9..00000000000 --- a/.changelog/5005.bugfix.2.md +++ /dev/null @@ -1,7 +0,0 @@ -go/worker/storage: Fix case when checkpoint sync disabled but forced - -If checkpoint sync is disabled but sync has been forced (e.g. because -the state at genesis is non-empty), we must request to sync the -checkpoint at genesis as otherwise we will jump to a later state which -may not be desired given that checkpoint sync has been explicitly -disabled via config. diff --git a/.changelog/5005.bugfix.3.md b/.changelog/5005.bugfix.3.md deleted file mode 100644 index f0c9924acf2..00000000000 --- a/.changelog/5005.bugfix.3.md +++ /dev/null @@ -1 +0,0 @@ -go/storage/mkvs/checkpoint: Exclude initial version when pruning diff --git a/.changelog/5005.bugfix.4.md b/.changelog/5005.bugfix.4.md deleted file mode 100644 index 4b95b23f27b..00000000000 --- a/.changelog/5005.bugfix.4.md +++ /dev/null @@ -1 +0,0 @@ -go/p2p/rpc: Fix multi call dispatch to different peers diff --git a/.changelog/5007.bugfix.1.md b/.changelog/5007.bugfix.1.md deleted file mode 100644 index 996b2c084f1..00000000000 --- a/.changelog/5007.bugfix.1.md +++ /dev/null @@ -1,4 +0,0 @@ -go/p2p/rpc: Fix peer grading when context is canceled - -When method `CallMulti` finishes early, the requests in progress are canceled -and unfairly recorded as failed. diff --git a/.changelog/5007.bugfix.2.md b/.changelog/5007.bugfix.2.md deleted file mode 100644 index 5adda41fb19..00000000000 --- a/.changelog/5007.bugfix.2.md +++ /dev/null @@ -1,4 +0,0 @@ -go/p2p/rpc: Fix memory leak when RPC multi call finishes early - -When method `CallMulti` finishes early, the result channel is never cleared. -Therefore, the channel never closes and leaves one go routine hanging. diff --git a/.changelog/5008.bugfix.md b/.changelog/5008.bugfix.md deleted file mode 100644 index cbbbced8212..00000000000 --- a/.changelog/5008.bugfix.md +++ /dev/null @@ -1,5 +0,0 @@ -go/common/workerpool: Fix memory leak when workerpool is stopped early - -When workerpool si stopped, the job channel might still contain jobs which -haven't been processed. Therefore, the channel never closes and leaves one -go routine hanging. diff --git a/.changelog/5011.feature.md b/.changelog/5011.feature.md deleted file mode 100644 index e182e815046..00000000000 --- a/.changelog/5011.feature.md +++ /dev/null @@ -1 +0,0 @@ -go/consensus/tendermint/apps/staking: Reduce DelegationsTo scanning diff --git a/.punch_version.py b/.punch_version.py index b2246e26790..22876920183 100644 --- a/.punch_version.py +++ b/.punch_version.py @@ -1,3 +1,3 @@ year = '22' minor = 2 -micro = 0 +micro = 1 diff --git a/CHANGELOG.md b/CHANGELOG.md index ef88868cc30..bb5fd5a710d 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -12,6 +12,68 @@ The format is inspired by [Keep a Changelog]. +## 22.2.1 (2022-11-05) + +| Protocol | Version | +|:------------------|:---------:| +| Consensus | 6.0.0 | +| Runtime Host | 5.1.0 | +| Runtime Committee | 4.0.0 | + +### Features + +- go/consensus/tendermint/apps/staking: Reduce DelegationsTo scanning + ([#5011](https://github.com/oasisprotocol/oasis-core/issues/5011)) + +### Bug Fixes + +- go/runtime/txpool: republish sooner if republish limit is reached + ([#5003](https://github.com/oasisprotocol/oasis-core/issues/5003)) + + This fixes a case where a part of a batch of transaction would take a long + time to be published in case there are no new transactions being submitted. + +- go/worker/common: Reorder state determination checks + ([#5005](https://github.com/oasisprotocol/oasis-core/issues/5005)) + + Otherwise the shown state would be misleading, e.g. showing that it is + waiting for runtime host being provisioned while it is actually blocked + in initialization like storage sync. + +- go/p2p/rpc: Fix multi call dispatch to different peers + ([#5005](https://github.com/oasisprotocol/oasis-core/issues/5005)) + +- go/storage/mkvs/checkpoint: Exclude initial version when pruning + ([#5005](https://github.com/oasisprotocol/oasis-core/issues/5005)) + +- go/worker/storage: Fix case when checkpoint sync disabled but forced + ([#5005](https://github.com/oasisprotocol/oasis-core/issues/5005)) + + If checkpoint sync is disabled but sync has been forced (e.g. because + the state at genesis is non-empty), we must request to sync the + checkpoint at genesis as otherwise we will jump to a later state which + may not be desired given that checkpoint sync has been explicitly + disabled via config. + +- go/p2p/rpc: Fix peer grading when context is canceled + ([#5007](https://github.com/oasisprotocol/oasis-core/issues/5007)) + + When method `CallMulti` finishes early, the requests in progress are canceled + and unfairly recorded as failed. + +- go/p2p/rpc: Fix memory leak when RPC multi call finishes early + ([#5007](https://github.com/oasisprotocol/oasis-core/issues/5007)) + + When method `CallMulti` finishes early, the result channel is never cleared. + Therefore, the channel never closes and leaves one go routine hanging. + +- go/common/workerpool: Fix memory leak when workerpool is stopped early + ([#5008](https://github.com/oasisprotocol/oasis-core/issues/5008)) + + When workerpool si stopped, the job channel might still contain jobs which + haven't been processed. Therefore, the channel never closes and leaves one + go routine hanging. + ## 22.2 (2022-10-13) | Protocol | Version |