You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

1145 lines
41 KiB

state: add more tests for block validation (#3674) * Expose priv validators for use in testing * Generalize block header validation test past height 1 * Remove ineffectual assignment * Remove redundant SaveState call * Reorder comment for clarity * Use the block executor ApplyBlock function instead of implementing a stripped-down version of it * Remove commented-out code * Remove unnecessary test The required tests already appear to be implemented (implicitly) through the TestValidateBlockHeader test. * Allow for catching of specific error types during TestValidateBlockCommit * Make return error testable * Clean up and add TestValidateBlockCommit code * Fix formatting * Extract function to create a new mock test app * Update comment for clarity * Fix comment * Add skeleton code for evidence-related test * Allow for addressing priv val by address * Generalize test beyond a single validator * Generalize TestValidateBlockEvidence past first height * Reorder code to clearly separate tests and utility code * Use a common constant for stop height for testing in state/validation_test.go * Refactor errors to resemble existing conventions * Fix formatting * Extract common helper functions Having the tests littered with helper functions makes them less easily readable imho, so I've pulled them out into a separate file. This also makes it easier to see what helper functions are available during testing, so we minimize the chance of duplication when writing new tests. * Remove unused parameter * Remove unused parameters * Add field keys * Remove unused height constant * Fix typo * Fix incorrect return error * Add field keys * Use separate package for tests This refactors all of the state package's tests into a state_test package, so as to keep any usage of the state package's internal methods explicit. Any internal methods/constants used by tests are now explicitly exported in state/export_test.go * Refactor: extract helper function to make, validate, execute and commit a block * Rename state function to makeState * Remove redundant constant for number of validators * Refactor mock evidence registration into TestMain * Remove extraneous nVals variable * Replace function-level TODOs with file-level TODO and explanation * Remove extraneous comment * Fix linting issues brought up by GolangCI (pulled in from latest merge from develop)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
state: add more tests for block validation (#3674) * Expose priv validators for use in testing * Generalize block header validation test past height 1 * Remove ineffectual assignment * Remove redundant SaveState call * Reorder comment for clarity * Use the block executor ApplyBlock function instead of implementing a stripped-down version of it * Remove commented-out code * Remove unnecessary test The required tests already appear to be implemented (implicitly) through the TestValidateBlockHeader test. * Allow for catching of specific error types during TestValidateBlockCommit * Make return error testable * Clean up and add TestValidateBlockCommit code * Fix formatting * Extract function to create a new mock test app * Update comment for clarity * Fix comment * Add skeleton code for evidence-related test * Allow for addressing priv val by address * Generalize test beyond a single validator * Generalize TestValidateBlockEvidence past first height * Reorder code to clearly separate tests and utility code * Use a common constant for stop height for testing in state/validation_test.go * Refactor errors to resemble existing conventions * Fix formatting * Extract common helper functions Having the tests littered with helper functions makes them less easily readable imho, so I've pulled them out into a separate file. This also makes it easier to see what helper functions are available during testing, so we minimize the chance of duplication when writing new tests. * Remove unused parameter * Remove unused parameters * Add field keys * Remove unused height constant * Fix typo * Fix incorrect return error * Add field keys * Use separate package for tests This refactors all of the state package's tests into a state_test package, so as to keep any usage of the state package's internal methods explicit. Any internal methods/constants used by tests are now explicitly exported in state/export_test.go * Refactor: extract helper function to make, validate, execute and commit a block * Rename state function to makeState * Remove redundant constant for number of validators * Refactor mock evidence registration into TestMain * Remove extraneous nVals variable * Replace function-level TODOs with file-level TODO and explanation * Remove extraneous comment * Fix linting issues brought up by GolangCI (pulled in from latest merge from develop)
6 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
state: add more tests for block validation (#3674) * Expose priv validators for use in testing * Generalize block header validation test past height 1 * Remove ineffectual assignment * Remove redundant SaveState call * Reorder comment for clarity * Use the block executor ApplyBlock function instead of implementing a stripped-down version of it * Remove commented-out code * Remove unnecessary test The required tests already appear to be implemented (implicitly) through the TestValidateBlockHeader test. * Allow for catching of specific error types during TestValidateBlockCommit * Make return error testable * Clean up and add TestValidateBlockCommit code * Fix formatting * Extract function to create a new mock test app * Update comment for clarity * Fix comment * Add skeleton code for evidence-related test * Allow for addressing priv val by address * Generalize test beyond a single validator * Generalize TestValidateBlockEvidence past first height * Reorder code to clearly separate tests and utility code * Use a common constant for stop height for testing in state/validation_test.go * Refactor errors to resemble existing conventions * Fix formatting * Extract common helper functions Having the tests littered with helper functions makes them less easily readable imho, so I've pulled them out into a separate file. This also makes it easier to see what helper functions are available during testing, so we minimize the chance of duplication when writing new tests. * Remove unused parameter * Remove unused parameters * Add field keys * Remove unused height constant * Fix typo * Fix incorrect return error * Add field keys * Use separate package for tests This refactors all of the state package's tests into a state_test package, so as to keep any usage of the state package's internal methods explicit. Any internal methods/constants used by tests are now explicitly exported in state/export_test.go * Refactor: extract helper function to make, validate, execute and commit a block * Rename state function to makeState * Remove redundant constant for number of validators * Refactor mock evidence registration into TestMain * Remove extraneous nVals variable * Replace function-level TODOs with file-level TODO and explanation * Remove extraneous comment * Fix linting issues brought up by GolangCI (pulled in from latest merge from develop)
6 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Refactor tagging events using list of lists (#3643) ## PR This PR introduces a fundamental breaking change to the structure of ABCI response and tx tags and the way they're processed. Namely, the SDK can support more complex and aggregated events for distribution and slashing. In addition, block responses can include duplicate keys in events. Implement new Event type. An event has a type and a list of KV pairs (ie. list-of-lists). Typical events may look like: "rewards": [{"amount": "5000uatom", "validator": "...", "recipient": "..."}] "sender": [{"address": "...", "balance": "100uatom"}] The events are indexed by {even.type}.{even.attribute[i].key}/.... In this case a client would subscribe or query for rewards.recipient='...' ABCI response types and related types now include Events []Event instead of Tags []cmn.KVPair. PubSub logic now publishes/matches against map[string][]string instead of map[string]string to support duplicate keys in response events (from #1385). A match is successful if the value is found in the slice of strings. closes: #1859 closes: #2905 ## Commits: * Implement Event ABCI type and updates responses to use events * Update messages_test.go * Update kvstore.go * Update event_bus.go * Update subscription.go * Update pubsub.go * Update kvstore.go * Update query logic to handle slice of strings in events * Update Empty#Matches and unit tests * Update pubsub logic * Update EventBus#Publish * Update kv tx indexer * Update godocs * Update ResultEvent to use slice of strings; update RPC * Update more tests * Update abci.md * Check for key in validateAndStringifyEvents * Fix KV indexer to skip empty keys * Fix linting errors * Update CHANGELOG_PENDING.md * Update docs/spec/abci/abci.md Co-Authored-By: Federico Kunze <31522760+fedekunze@users.noreply.github.com> * Update abci/types/types.proto Co-Authored-By: Ethan Buchman <ethan@coinculture.info> * Update docs/spec/abci/abci.md Co-Authored-By: Ethan Buchman <ethan@coinculture.info> * Update libs/pubsub/query/query.go Co-Authored-By: Ethan Buchman <ethan@coinculture.info> * Update match function to match if ANY value matches * Implement TestSubscribeDuplicateKeys * Update TestMatches to include multi-key test cases * Update events.go * Update Query interface godoc * Update match godoc * Add godoc for matchValue * DRY-up tx indexing * Return error from PublishWithEvents in EventBus#Publish * Update PublishEventNewBlockHeader to return an error * Fix build * Update events doc in ABCI * Update ABCI events godoc * Implement TestEventBusPublishEventTxDuplicateKeys * Update TestSubscribeDuplicateKeys to be table-driven * Remove mod file * Remove markdown from events godoc * Implement TestTxSearchDeprecatedIndexing test
6 years ago
abci: Refactor tagging events using list of lists (#3643) ## PR This PR introduces a fundamental breaking change to the structure of ABCI response and tx tags and the way they're processed. Namely, the SDK can support more complex and aggregated events for distribution and slashing. In addition, block responses can include duplicate keys in events. Implement new Event type. An event has a type and a list of KV pairs (ie. list-of-lists). Typical events may look like: "rewards": [{"amount": "5000uatom", "validator": "...", "recipient": "..."}] "sender": [{"address": "...", "balance": "100uatom"}] The events are indexed by {even.type}.{even.attribute[i].key}/.... In this case a client would subscribe or query for rewards.recipient='...' ABCI response types and related types now include Events []Event instead of Tags []cmn.KVPair. PubSub logic now publishes/matches against map[string][]string instead of map[string]string to support duplicate keys in response events (from #1385). A match is successful if the value is found in the slice of strings. closes: #1859 closes: #2905 ## Commits: * Implement Event ABCI type and updates responses to use events * Update messages_test.go * Update kvstore.go * Update event_bus.go * Update subscription.go * Update pubsub.go * Update kvstore.go * Update query logic to handle slice of strings in events * Update Empty#Matches and unit tests * Update pubsub logic * Update EventBus#Publish * Update kv tx indexer * Update godocs * Update ResultEvent to use slice of strings; update RPC * Update more tests * Update abci.md * Check for key in validateAndStringifyEvents * Fix KV indexer to skip empty keys * Fix linting errors * Update CHANGELOG_PENDING.md * Update docs/spec/abci/abci.md Co-Authored-By: Federico Kunze <31522760+fedekunze@users.noreply.github.com> * Update abci/types/types.proto Co-Authored-By: Ethan Buchman <ethan@coinculture.info> * Update docs/spec/abci/abci.md Co-Authored-By: Ethan Buchman <ethan@coinculture.info> * Update libs/pubsub/query/query.go Co-Authored-By: Ethan Buchman <ethan@coinculture.info> * Update match function to match if ANY value matches * Implement TestSubscribeDuplicateKeys * Update TestMatches to include multi-key test cases * Update events.go * Update Query interface godoc * Update match godoc * Add godoc for matchValue * DRY-up tx indexing * Return error from PublishWithEvents in EventBus#Publish * Update PublishEventNewBlockHeader to return an error * Fix build * Update events doc in ABCI * Update ABCI events godoc * Implement TestEventBusPublishEventTxDuplicateKeys * Update TestSubscribeDuplicateKeys to be table-driven * Remove mod file * Remove markdown from events godoc * Implement TestTxSearchDeprecatedIndexing test
6 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
lint: Enable Golint (#4212) * Fix many golint errors * Fix golint errors in the 'lite' package * Don't export Pool.store * Fix typo * Revert unwanted changes * Fix errors in counter package * Fix linter errors in kvstore package * Fix linter error in example package * Fix error in tests package * Fix linter errors in v2 package * Fix linter errors in consensus package * Fix linter errors in evidence package * Fix linter error in fail package * Fix linter errors in query package * Fix linter errors in core package * Fix linter errors in node package * Fix linter errors in mempool package * Fix linter error in conn package * Fix linter errors in pex package * Rename PEXReactor export to Reactor * Fix linter errors in trust package * Fix linter errors in upnp package * Fix linter errors in p2p package * Fix linter errors in proxy package * Fix linter errors in mock_test package * Fix linter error in client_test package * Fix linter errors in coretypes package * Fix linter errors in coregrpc package * Fix linter errors in rpcserver package * Fix linter errors in rpctypes package * Fix linter errors in rpctest package * Fix linter error in json2wal script * Fix linter error in wal2json script * Fix linter errors in kv package * Fix linter error in state package * Fix linter error in grpc_client * Fix linter errors in types package * Fix linter error in version package * Fix remaining errors * Address review comments * Fix broken tests * Reconcile package coregrpc * Fix golangci bot error * Fix new golint errors * Fix broken reference * Enable golint linter * minor changes to bring golint into line * fix failing test * fix pex reactor naming * address PR comments
5 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
cleanup: Reduce and normalize import path aliasing. (#6975) The code in the Tendermint repository makes heavy use of import aliasing. This is made necessary by our extensive reuse of common base package names, and by repetition of similar names across different subdirectories. Unfortunately we have not been very consistent about which packages we alias in various circumstances, and the aliases we use vary. In the spirit of the advice in the style guide and https://github.com/golang/go/wiki/CodeReviewComments#imports, his change makes an effort to clean up and normalize import aliasing. This change makes no API or behavioral changes. It is a pure cleanup intended o help make the code more readable to developers (including myself) trying to understand what is being imported where. Only unexported names have been modified, and the changes were generated and applied mechanically with gofmt -r and comby, respecting the lexical and syntactic rules of Go. Even so, I did not fix every inconsistency. Where the changes would be too disruptive, I left it alone. The principles I followed in this cleanup are: - Remove aliases that restate the package name. - Remove aliases where the base package name is unambiguous. - Move overly-terse abbreviations from the import to the usage site. - Fix lexical issues (remove underscores, remove capitalization). - Fix import groupings to more closely match the style guide. - Group blank (side-effecting) imports and ensure they are commented. - Add aliases to multiple imports with the same base package name.
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
Normalize priorities to not exceed total voting power (#3049) * more proposer priority tests - test that we don't reset to zero when updating / adding - test that same power validators alternate * add another test to track / simulate similar behaviour as in #2960 * address some of Chris' review comments * address some more of Chris' review comments * temporarily pushing branch with the following changes: The total power might change if: - a validator is added - a validator is removed - a validator is updated Decrement the accums (of all validators) directly after any of these events (by the inverse of the change) * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * Fix 2960 by re-normalizing / scaling priorities to be in bounds of total power, additionally: - remove heap where it doesn't make sense - avg. only at the end of IncrementProposerPriority instead of each iteration - update (and slightly improve) TestAveragingInIncrementProposerPriorityWithVotingPower to reflect above changes * fix tests * add comment * update changelog pending & some minor changes * comment about division will floor the result & fix typo * Update TestLargeGenesisValidator: - remove TODO and increase large genesis validator's voting power accordingly * move changelog entry to P2P Protocol * Ceil instead of flooring when dividing & update test * quickly fix failing TestProposerPriorityDoesNotGetResetToZero: - divide by Ceil((maxPriority - minPriority) / 2*totalVotingPower) * fix typo: rename getValWitMostPriority -> getValWithMostPriority * test proposer frequencies * return absolute value for diff. keep testing * use for loop for div * cleanup, more tests * spellcheck * get rid of using floats: manually ceil where necessary * Remove float, simplify, fix tests to match chris's proof (#3157)
6 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
abci: Synchronize FinalizeBlock with the updated specification (#7983) This change set implements the most recent version of `FinalizeBlock`. # What does this change actually contain? * This change set is rather large but fear not! The majority of the files touched and changes are renaming `ResponseDeliverTx` to `ExecTxResult`. This should be a pretty inoffensive change since they're effectively the same type but with a different name. * The `execBlockOnProxyApp` was totally removed since it served as just a wrapper around the logic that is now mostly encapsulated within `FinalizeBlock` * The `updateState` helper function has been made a public method on `State`. It was being exposed as a shim through the testing infrastructure, so this seemed innocuous. * Tests already existed to ensure that the application received the `ByzantineValidators` and the `ValidatorUpdates`, but one was fixed up to ensure that `LastCommitInfo` was being sent across. * Tests were removed from the `psql` indexer that seemed to search for an event in the indexer that was not being created. # Questions for reviewers * We store this [ABCIResponses](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/proto/tendermint/state/types.pb.go#L37) type in the data base as the block results. This type has changed since v0.35 to contain the `FinalizeBlock` response. I'm wondering if we need to do any shimming to keep the old data retrieveable? * Similarly, this change is exposed via the RPC through [ResultBlockResults](https://github.com/tendermint/tendermint/blob/5721a13ab1f4479f9807f449f0bf5c536b9a05f2/rpc/coretypes/responses.go#L69) changing. Should we somehow shim or notify for this change? closes: #7658
3 years ago
  1. package state_test
  2. import (
  3. "bytes"
  4. "context"
  5. "fmt"
  6. "math"
  7. "math/big"
  8. mrand "math/rand"
  9. "os"
  10. "testing"
  11. "github.com/stretchr/testify/assert"
  12. "github.com/stretchr/testify/require"
  13. dbm "github.com/tendermint/tm-db"
  14. abci "github.com/tendermint/tendermint/abci/types"
  15. "github.com/tendermint/tendermint/config"
  16. "github.com/tendermint/tendermint/crypto/ed25519"
  17. "github.com/tendermint/tendermint/crypto/encoding"
  18. sm "github.com/tendermint/tendermint/internal/state"
  19. statefactory "github.com/tendermint/tendermint/internal/state/test/factory"
  20. tmstate "github.com/tendermint/tendermint/proto/tendermint/state"
  21. "github.com/tendermint/tendermint/types"
  22. )
  23. // setupTestCase does setup common to all test cases.
  24. func setupTestCase(t *testing.T) (func(t *testing.T), dbm.DB, sm.State) {
  25. cfg, err := config.ResetTestRoot(t.TempDir(), "state_")
  26. require.NoError(t, err)
  27. dbType := dbm.BackendType(cfg.DBBackend)
  28. stateDB, err := dbm.NewDB("state", dbType, cfg.DBDir())
  29. require.NoError(t, err)
  30. stateStore := sm.NewStore(stateDB)
  31. state, err := stateStore.Load()
  32. require.NoError(t, err)
  33. require.Empty(t, state)
  34. state, err = sm.MakeGenesisStateFromFile(cfg.GenesisFile())
  35. assert.NoError(t, err)
  36. assert.NotNil(t, state)
  37. err = stateStore.Save(state)
  38. require.NoError(t, err)
  39. tearDown := func(t *testing.T) { os.RemoveAll(cfg.RootDir) }
  40. return tearDown, stateDB, state
  41. }
  42. // TestStateCopy tests the correct copying behavior of State.
  43. func TestStateCopy(t *testing.T) {
  44. tearDown, _, state := setupTestCase(t)
  45. defer tearDown(t)
  46. stateCopy := state.Copy()
  47. assert.True(t, state.Equals(stateCopy),
  48. "expected state and its copy to be identical.\ngot: %v\nexpected: %v",
  49. stateCopy, state)
  50. stateCopy.LastBlockHeight++
  51. stateCopy.LastValidators = state.Validators
  52. assert.False(t, state.Equals(stateCopy), "expected states to be different. got same %v", state)
  53. }
  54. // TestMakeGenesisStateNilValidators tests state's consistency when genesis file's validators field is nil.
  55. func TestMakeGenesisStateNilValidators(t *testing.T) {
  56. doc := types.GenesisDoc{
  57. ChainID: "dummy",
  58. Validators: nil,
  59. }
  60. require.Nil(t, doc.ValidateAndComplete())
  61. state, err := sm.MakeGenesisState(&doc)
  62. require.NoError(t, err)
  63. require.Equal(t, 0, len(state.Validators.Validators))
  64. require.Equal(t, 0, len(state.NextValidators.Validators))
  65. }
  66. // TestStateSaveLoad tests saving and loading State from a db.
  67. func TestStateSaveLoad(t *testing.T) {
  68. tearDown, stateDB, state := setupTestCase(t)
  69. defer tearDown(t)
  70. stateStore := sm.NewStore(stateDB)
  71. state.LastBlockHeight++
  72. state.LastValidators = state.Validators
  73. err := stateStore.Save(state)
  74. require.NoError(t, err)
  75. loadedState, err := stateStore.Load()
  76. require.NoError(t, err)
  77. assert.True(t, state.Equals(loadedState),
  78. "expected state and its copy to be identical.\ngot: %v\nexpected: %v",
  79. loadedState, state)
  80. }
  81. // TestABCIResponsesSaveLoad tests saving and loading ABCIResponses.
  82. func TestABCIResponsesSaveLoad1(t *testing.T) {
  83. tearDown, stateDB, state := setupTestCase(t)
  84. defer tearDown(t)
  85. stateStore := sm.NewStore(stateDB)
  86. state.LastBlockHeight++
  87. // Build mock responses.
  88. block, err := statefactory.MakeBlock(state, 2, new(types.Commit))
  89. require.NoError(t, err)
  90. abciResponses := new(tmstate.ABCIResponses)
  91. dtxs := make([]*abci.ExecTxResult, 2)
  92. abciResponses.FinalizeBlock = new(abci.ResponseFinalizeBlock)
  93. abciResponses.FinalizeBlock.TxResults = dtxs
  94. abciResponses.FinalizeBlock.TxResults[0] = &abci.ExecTxResult{Data: []byte("foo"), Events: nil}
  95. abciResponses.FinalizeBlock.TxResults[1] = &abci.ExecTxResult{Data: []byte("bar"), Log: "ok", Events: nil}
  96. pbpk, err := encoding.PubKeyToProto(ed25519.GenPrivKey().PubKey())
  97. require.NoError(t, err)
  98. abciResponses.FinalizeBlock.ValidatorUpdates = []abci.ValidatorUpdate{{PubKey: pbpk, Power: 10}}
  99. err = stateStore.SaveABCIResponses(block.Height, abciResponses)
  100. require.NoError(t, err)
  101. loadedABCIResponses, err := stateStore.LoadABCIResponses(block.Height)
  102. require.NoError(t, err)
  103. assert.Equal(t, abciResponses, loadedABCIResponses,
  104. "ABCIResponses don't match:\ngot: %v\nexpected: %v\n",
  105. loadedABCIResponses, abciResponses)
  106. }
  107. // TestResultsSaveLoad tests saving and loading ABCI results.
  108. func TestABCIResponsesSaveLoad2(t *testing.T) {
  109. tearDown, stateDB, _ := setupTestCase(t)
  110. defer tearDown(t)
  111. stateStore := sm.NewStore(stateDB)
  112. cases := [...]struct {
  113. // Height is implied to equal index+2,
  114. // as block 1 is created from genesis.
  115. added []*abci.ExecTxResult
  116. expected []*abci.ExecTxResult
  117. }{
  118. 0: {
  119. nil,
  120. nil,
  121. },
  122. 1: {
  123. []*abci.ExecTxResult{
  124. {Code: 32, Data: []byte("Hello"), Log: "Huh?"},
  125. },
  126. []*abci.ExecTxResult{
  127. {Code: 32, Data: []byte("Hello")},
  128. },
  129. },
  130. 2: {
  131. []*abci.ExecTxResult{
  132. {Code: 383},
  133. {
  134. Data: []byte("Gotcha!"),
  135. Events: []abci.Event{
  136. {Type: "type1", Attributes: []abci.EventAttribute{{Key: "a", Value: "1"}}},
  137. {Type: "type2", Attributes: []abci.EventAttribute{{Key: "build", Value: "stuff"}}},
  138. },
  139. },
  140. },
  141. []*abci.ExecTxResult{
  142. {Code: 383, Data: nil},
  143. {Code: 0, Data: []byte("Gotcha!"), Events: []abci.Event{
  144. {Type: "type1", Attributes: []abci.EventAttribute{{Key: "a", Value: "1"}}},
  145. {Type: "type2", Attributes: []abci.EventAttribute{{Key: "build", Value: "stuff"}}},
  146. }},
  147. },
  148. },
  149. 3: {
  150. nil,
  151. nil,
  152. },
  153. 4: {
  154. []*abci.ExecTxResult{nil},
  155. nil,
  156. },
  157. }
  158. // Query all before, this should return error.
  159. for i := range cases {
  160. h := int64(i + 1)
  161. res, err := stateStore.LoadABCIResponses(h)
  162. assert.Error(t, err, "%d: %#v", i, res)
  163. }
  164. // Add all cases.
  165. for i, tc := range cases {
  166. h := int64(i + 1) // last block height, one below what we save
  167. responses := &tmstate.ABCIResponses{
  168. FinalizeBlock: &abci.ResponseFinalizeBlock{
  169. TxResults: tc.added,
  170. },
  171. }
  172. err := stateStore.SaveABCIResponses(h, responses)
  173. require.NoError(t, err)
  174. }
  175. // Query all before, should return expected value.
  176. for i, tc := range cases {
  177. h := int64(i + 1)
  178. res, err := stateStore.LoadABCIResponses(h)
  179. if assert.NoError(t, err, "%d", i) {
  180. t.Log(res)
  181. responses := &tmstate.ABCIResponses{
  182. FinalizeBlock: &abci.ResponseFinalizeBlock{
  183. TxResults: tc.expected,
  184. },
  185. }
  186. sm.ABCIResponsesResultsHash(res)
  187. sm.ABCIResponsesResultsHash(responses)
  188. assert.Equal(t, sm.ABCIResponsesResultsHash(responses), sm.ABCIResponsesResultsHash(res), "%d", i)
  189. }
  190. }
  191. }
  192. // TestValidatorSimpleSaveLoad tests saving and loading validators.
  193. func TestValidatorSimpleSaveLoad(t *testing.T) {
  194. tearDown, stateDB, state := setupTestCase(t)
  195. defer tearDown(t)
  196. statestore := sm.NewStore(stateDB)
  197. // Can't load anything for height 0.
  198. _, err := statestore.LoadValidators(0)
  199. assert.IsType(t, sm.ErrNoValSetForHeight{}, err, "expected err at height 0")
  200. // Should be able to load for height 1.
  201. v, err := statestore.LoadValidators(1)
  202. require.NoError(t, err, "expected no err at height 1")
  203. assert.Equal(t, v.Hash(), state.Validators.Hash(), "expected validator hashes to match")
  204. // Should be able to load for height 2.
  205. v, err = statestore.LoadValidators(2)
  206. require.NoError(t, err, "expected no err at height 2")
  207. assert.Equal(t, v.Hash(), state.NextValidators.Hash(), "expected validator hashes to match")
  208. // Increment height, save; should be able to load for next & next next height.
  209. state.LastBlockHeight++
  210. nextHeight := state.LastBlockHeight + 1
  211. err = statestore.Save(state)
  212. require.NoError(t, err)
  213. vp0, err := statestore.LoadValidators(nextHeight + 0)
  214. assert.NoError(t, err)
  215. vp1, err := statestore.LoadValidators(nextHeight + 1)
  216. assert.NoError(t, err)
  217. assert.Equal(t, vp0.Hash(), state.Validators.Hash(), "expected validator hashes to match")
  218. assert.Equal(t, vp1.Hash(), state.NextValidators.Hash(), "expected next validator hashes to match")
  219. }
  220. // TestValidatorChangesSaveLoad tests saving and loading a validator set with changes.
  221. func TestOneValidatorChangesSaveLoad(t *testing.T) {
  222. tearDown, stateDB, state := setupTestCase(t)
  223. defer tearDown(t)
  224. stateStore := sm.NewStore(stateDB)
  225. // Change vals at these heights.
  226. changeHeights := []int64{1, 2, 4, 5, 10, 15, 16, 17, 20}
  227. N := len(changeHeights)
  228. // Build the validator history by running updateState
  229. // with the right validator set for each height.
  230. highestHeight := changeHeights[N-1] + 5
  231. changeIndex := 0
  232. _, val := state.Validators.GetByIndex(0)
  233. power := val.VotingPower
  234. var err error
  235. var validatorUpdates []*types.Validator
  236. for i := int64(1); i < highestHeight; i++ {
  237. // When we get to a change height, use the next pubkey.
  238. if changeIndex < len(changeHeights) && i == changeHeights[changeIndex] {
  239. changeIndex++
  240. power++
  241. }
  242. header, blockID, responses := makeHeaderPartsResponsesValPowerChange(t, state, power)
  243. validatorUpdates, err = types.PB2TM.ValidatorUpdates(responses.FinalizeBlock.ValidatorUpdates)
  244. require.NoError(t, err)
  245. state, err = state.Update(blockID, &header, sm.ABCIResponsesResultsHash(responses), responses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  246. require.NoError(t, err)
  247. err := stateStore.Save(state)
  248. require.NoError(t, err)
  249. }
  250. // On each height change, increment the power by one.
  251. testCases := make([]int64, highestHeight)
  252. changeIndex = 0
  253. power = val.VotingPower
  254. for i := int64(1); i < highestHeight+1; i++ {
  255. // We get to the height after a change height use the next pubkey (note
  256. // our counter starts at 0 this time).
  257. if changeIndex < len(changeHeights) && i == changeHeights[changeIndex]+1 {
  258. changeIndex++
  259. power++
  260. }
  261. testCases[i-1] = power
  262. }
  263. for i, power := range testCases {
  264. v, err := stateStore.LoadValidators(int64(i + 1 + 1)) // +1 because vset changes delayed by 1 block.
  265. assert.NoError(t, err, fmt.Sprintf("expected no err at height %d", i))
  266. assert.Equal(t, v.Size(), 1, "validator set size is greater than 1: %d", v.Size())
  267. _, val := v.GetByIndex(0)
  268. assert.Equal(t, val.VotingPower, power, fmt.Sprintf(`unexpected powerat
  269. height %d`, i))
  270. }
  271. }
  272. func TestProposerFrequency(t *testing.T) {
  273. ctx, cancel := context.WithCancel(context.Background())
  274. defer cancel()
  275. // some explicit test cases
  276. testCases := []struct {
  277. powers []int64
  278. }{
  279. // 2 vals
  280. {[]int64{1, 1}},
  281. {[]int64{1, 2}},
  282. {[]int64{1, 100}},
  283. {[]int64{5, 5}},
  284. {[]int64{5, 100}},
  285. {[]int64{50, 50}},
  286. {[]int64{50, 100}},
  287. {[]int64{1, 1000}},
  288. // 3 vals
  289. {[]int64{1, 1, 1}},
  290. {[]int64{1, 2, 3}},
  291. {[]int64{1, 2, 3}},
  292. {[]int64{1, 1, 10}},
  293. {[]int64{1, 1, 100}},
  294. {[]int64{1, 10, 100}},
  295. {[]int64{1, 1, 1000}},
  296. {[]int64{1, 10, 1000}},
  297. {[]int64{1, 100, 1000}},
  298. // 4 vals
  299. {[]int64{1, 1, 1, 1}},
  300. {[]int64{1, 2, 3, 4}},
  301. {[]int64{1, 1, 1, 10}},
  302. {[]int64{1, 1, 1, 100}},
  303. {[]int64{1, 1, 1, 1000}},
  304. {[]int64{1, 1, 10, 100}},
  305. {[]int64{1, 1, 10, 1000}},
  306. {[]int64{1, 1, 100, 1000}},
  307. {[]int64{1, 10, 100, 1000}},
  308. }
  309. for caseNum, testCase := range testCases {
  310. // run each case 5 times to sample different
  311. // initial priorities
  312. for i := 0; i < 5; i++ {
  313. valSet := genValSetWithPowers(testCase.powers)
  314. testProposerFreq(t, caseNum, valSet)
  315. }
  316. }
  317. // some random test cases with up to 100 validators
  318. maxVals := 100
  319. maxPower := 1000
  320. nTestCases := 5
  321. for i := 0; i < nTestCases; i++ {
  322. N := mrand.Int()%maxVals + 1
  323. vals := make([]*types.Validator, N)
  324. totalVotePower := int64(0)
  325. for j := 0; j < N; j++ {
  326. // make sure votePower > 0
  327. votePower := int64(mrand.Int()%maxPower) + 1
  328. totalVotePower += votePower
  329. privVal := types.NewMockPV()
  330. pubKey, err := privVal.GetPubKey(ctx)
  331. require.NoError(t, err)
  332. val := types.NewValidator(pubKey, votePower)
  333. val.ProposerPriority = mrand.Int63()
  334. vals[j] = val
  335. }
  336. valSet := types.NewValidatorSet(vals)
  337. valSet.RescalePriorities(totalVotePower)
  338. testProposerFreq(t, i, valSet)
  339. }
  340. }
  341. // new val set with given powers and random initial priorities
  342. func genValSetWithPowers(powers []int64) *types.ValidatorSet {
  343. size := len(powers)
  344. vals := make([]*types.Validator, size)
  345. totalVotePower := int64(0)
  346. for i := 0; i < size; i++ {
  347. totalVotePower += powers[i]
  348. val := types.NewValidator(ed25519.GenPrivKey().PubKey(), powers[i])
  349. val.ProposerPriority = mrand.Int63()
  350. vals[i] = val
  351. }
  352. valSet := types.NewValidatorSet(vals)
  353. valSet.RescalePriorities(totalVotePower)
  354. return valSet
  355. }
  356. // test a proposer appears as frequently as expected
  357. func testProposerFreq(t *testing.T, caseNum int, valSet *types.ValidatorSet) {
  358. N := valSet.Size()
  359. totalPower := valSet.TotalVotingPower()
  360. // run the proposer selection and track frequencies
  361. runMult := 1
  362. runs := int(totalPower) * runMult
  363. freqs := make([]int, N)
  364. for i := 0; i < runs; i++ {
  365. prop := valSet.GetProposer()
  366. idx, _ := valSet.GetByAddress(prop.Address)
  367. freqs[idx]++
  368. valSet.IncrementProposerPriority(1)
  369. }
  370. // assert frequencies match expected (max off by 1)
  371. for i, freq := range freqs {
  372. _, val := valSet.GetByIndex(int32(i))
  373. expectFreq := int(val.VotingPower) * runMult
  374. gotFreq := freq
  375. abs := int(math.Abs(float64(expectFreq - gotFreq)))
  376. // max bound on expected vs seen freq was proven
  377. // to be 1 for the 2 validator case in
  378. // https://github.com/cwgoes/tm-proposer-idris
  379. // and inferred to generalize to N-1
  380. bound := N - 1
  381. require.True(
  382. t,
  383. abs <= bound,
  384. fmt.Sprintf("Case %d val %d (%d): got %d, expected %d", caseNum, i, N, gotFreq, expectFreq),
  385. )
  386. }
  387. }
  388. // TestProposerPriorityDoesNotGetResetToZero assert that we preserve accum when calling updateState
  389. // see https://github.com/tendermint/tendermint/issues/2718
  390. func TestProposerPriorityDoesNotGetResetToZero(t *testing.T) {
  391. tearDown, _, state := setupTestCase(t)
  392. defer tearDown(t)
  393. val1VotingPower := int64(10)
  394. val1PubKey := ed25519.GenPrivKey().PubKey()
  395. val1 := &types.Validator{Address: val1PubKey.Address(), PubKey: val1PubKey, VotingPower: val1VotingPower}
  396. state.Validators = types.NewValidatorSet([]*types.Validator{val1})
  397. state.NextValidators = state.Validators
  398. // NewValidatorSet calls IncrementProposerPriority but uses on a copy of val1
  399. assert.EqualValues(t, 0, val1.ProposerPriority)
  400. block, err := statefactory.MakeBlock(state, state.LastBlockHeight+1, new(types.Commit))
  401. require.NoError(t, err)
  402. bps, err := block.MakePartSet(testPartSize)
  403. require.NoError(t, err)
  404. blockID := types.BlockID{Hash: block.Hash(), PartSetHeader: bps.Header()}
  405. abciResponses := &tmstate.ABCIResponses{
  406. FinalizeBlock: &abci.ResponseFinalizeBlock{
  407. ValidatorUpdates: nil,
  408. },
  409. }
  410. validatorUpdates, err := types.PB2TM.ValidatorUpdates(abciResponses.FinalizeBlock.ValidatorUpdates)
  411. require.NoError(t, err)
  412. updatedState, err := state.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  413. assert.NoError(t, err)
  414. curTotal := val1VotingPower
  415. // one increment step and one validator: 0 + power - total_power == 0
  416. assert.Equal(t, 0+val1VotingPower-curTotal, updatedState.NextValidators.Validators[0].ProposerPriority)
  417. // add a validator
  418. val2PubKey := ed25519.GenPrivKey().PubKey()
  419. val2VotingPower := int64(100)
  420. fvp, err := encoding.PubKeyToProto(val2PubKey)
  421. require.NoError(t, err)
  422. updateAddVal := abci.ValidatorUpdate{PubKey: fvp, Power: val2VotingPower}
  423. validatorUpdates, err = types.PB2TM.ValidatorUpdates([]abci.ValidatorUpdate{updateAddVal})
  424. assert.NoError(t, err)
  425. updatedState2, err := updatedState.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  426. assert.NoError(t, err)
  427. require.Equal(t, len(updatedState2.NextValidators.Validators), 2)
  428. _, updatedVal1 := updatedState2.NextValidators.GetByAddress(val1PubKey.Address())
  429. _, addedVal2 := updatedState2.NextValidators.GetByAddress(val2PubKey.Address())
  430. // adding a validator should not lead to a ProposerPriority equal to zero (unless the combination of averaging and
  431. // incrementing would cause so; which is not the case here)
  432. // Steps from adding new validator:
  433. // 0 - val1 prio is 0, TVP after add:
  434. wantVal1Prio := int64(0)
  435. totalPowerAfter := val1VotingPower + val2VotingPower
  436. // 1. Add - Val2 should be initially added with (-123) =>
  437. wantVal2Prio := -(totalPowerAfter + (totalPowerAfter >> 3))
  438. // 2. Scale - noop
  439. // 3. Center - with avg, resulting val2:-61, val1:62
  440. avg := big.NewInt(0).Add(big.NewInt(wantVal1Prio), big.NewInt(wantVal2Prio))
  441. avg.Div(avg, big.NewInt(2))
  442. wantVal2Prio -= avg.Int64() // -61
  443. wantVal1Prio -= avg.Int64() // 62
  444. // 4. Steps from IncrementProposerPriority
  445. wantVal1Prio += val1VotingPower // 72
  446. wantVal2Prio += val2VotingPower // 39
  447. wantVal1Prio -= totalPowerAfter // -38 as val1 is proposer
  448. assert.Equal(t, wantVal1Prio, updatedVal1.ProposerPriority)
  449. assert.Equal(t, wantVal2Prio, addedVal2.ProposerPriority)
  450. // Updating a validator does not reset the ProposerPriority to zero:
  451. // 1. Add - Val2 VotingPower change to 1 =>
  452. updatedVotingPowVal2 := int64(1)
  453. updateVal := abci.ValidatorUpdate{PubKey: fvp, Power: updatedVotingPowVal2}
  454. validatorUpdates, err = types.PB2TM.ValidatorUpdates([]abci.ValidatorUpdate{updateVal})
  455. assert.NoError(t, err)
  456. // this will cause the diff of priorities (77)
  457. // to be larger than threshold == 2*totalVotingPower (22):
  458. updatedState3, err := updatedState2.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  459. assert.NoError(t, err)
  460. require.Equal(t, len(updatedState3.NextValidators.Validators), 2)
  461. _, prevVal1 := updatedState3.Validators.GetByAddress(val1PubKey.Address())
  462. _, prevVal2 := updatedState3.Validators.GetByAddress(val2PubKey.Address())
  463. _, updatedVal1 = updatedState3.NextValidators.GetByAddress(val1PubKey.Address())
  464. _, updatedVal2 := updatedState3.NextValidators.GetByAddress(val2PubKey.Address())
  465. // 2. Scale
  466. // old prios: v1(10):-38, v2(1):39
  467. wantVal1Prio = prevVal1.ProposerPriority
  468. wantVal2Prio = prevVal2.ProposerPriority
  469. // scale to diffMax = 22 = 2 * tvp, diff=39-(-38)=77
  470. // new totalPower
  471. totalPower := updatedVal1.VotingPower + updatedVal2.VotingPower
  472. dist := wantVal2Prio - wantVal1Prio
  473. // ratio := (dist + 2*totalPower - 1) / 2*totalPower = 98/22 = 4
  474. ratio := (dist + 2*totalPower - 1) / (2 * totalPower)
  475. // v1(10):-38/4, v2(1):39/4
  476. wantVal1Prio /= ratio // -9
  477. wantVal2Prio /= ratio // 9
  478. // 3. Center - noop
  479. // 4. IncrementProposerPriority() ->
  480. // v1(10):-9+10, v2(1):9+1 -> v2 proposer so subsract tvp(11)
  481. // v1(10):1, v2(1):-1
  482. wantVal2Prio += updatedVal2.VotingPower // 10 -> prop
  483. wantVal1Prio += updatedVal1.VotingPower // 1
  484. wantVal2Prio -= totalPower // -1
  485. assert.Equal(t, wantVal2Prio, updatedVal2.ProposerPriority)
  486. assert.Equal(t, wantVal1Prio, updatedVal1.ProposerPriority)
  487. }
  488. func TestProposerPriorityProposerAlternates(t *testing.T) {
  489. // Regression test that would fail if the inner workings of
  490. // IncrementProposerPriority change.
  491. // Additionally, make sure that same power validators alternate if both
  492. // have the same voting power (and the 2nd was added later).
  493. tearDown, _, state := setupTestCase(t)
  494. defer tearDown(t)
  495. val1VotingPower := int64(10)
  496. val1PubKey := ed25519.GenPrivKey().PubKey()
  497. val1 := &types.Validator{Address: val1PubKey.Address(), PubKey: val1PubKey, VotingPower: val1VotingPower}
  498. // reset state validators to above validator
  499. state.Validators = types.NewValidatorSet([]*types.Validator{val1})
  500. state.NextValidators = state.Validators
  501. // we only have one validator:
  502. assert.Equal(t, val1PubKey.Address(), state.Validators.Proposer.Address)
  503. block, err := statefactory.MakeBlock(state, state.LastBlockHeight+1, new(types.Commit))
  504. require.NoError(t, err)
  505. bps, err := block.MakePartSet(testPartSize)
  506. require.NoError(t, err)
  507. blockID := types.BlockID{Hash: block.Hash(), PartSetHeader: bps.Header()}
  508. // no updates:
  509. abciResponses := &tmstate.ABCIResponses{
  510. FinalizeBlock: &abci.ResponseFinalizeBlock{
  511. ValidatorUpdates: nil,
  512. },
  513. }
  514. validatorUpdates, err := types.PB2TM.ValidatorUpdates(abciResponses.FinalizeBlock.ValidatorUpdates)
  515. require.NoError(t, err)
  516. updatedState, err := state.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  517. assert.NoError(t, err)
  518. // 0 + 10 (initial prio) - 10 (avg) - 10 (mostest - total) = -10
  519. totalPower := val1VotingPower
  520. wantVal1Prio := 0 + val1VotingPower - totalPower
  521. assert.Equal(t, wantVal1Prio, updatedState.NextValidators.Validators[0].ProposerPriority)
  522. assert.Equal(t, val1PubKey.Address(), updatedState.NextValidators.Proposer.Address)
  523. // add a validator with the same voting power as the first
  524. val2PubKey := ed25519.GenPrivKey().PubKey()
  525. fvp, err := encoding.PubKeyToProto(val2PubKey)
  526. require.NoError(t, err)
  527. updateAddVal := abci.ValidatorUpdate{PubKey: fvp, Power: val1VotingPower}
  528. validatorUpdates, err = types.PB2TM.ValidatorUpdates([]abci.ValidatorUpdate{updateAddVal})
  529. assert.NoError(t, err)
  530. updatedState2, err := updatedState.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  531. assert.NoError(t, err)
  532. require.Equal(t, len(updatedState2.NextValidators.Validators), 2)
  533. assert.Equal(t, updatedState2.Validators, updatedState.NextValidators)
  534. // val1 will still be proposer as val2 just got added:
  535. assert.Equal(t, val1PubKey.Address(), updatedState.NextValidators.Proposer.Address)
  536. assert.Equal(t, updatedState2.Validators.Proposer.Address, updatedState2.NextValidators.Proposer.Address)
  537. assert.Equal(t, updatedState2.Validators.Proposer.Address, val1PubKey.Address())
  538. assert.Equal(t, updatedState2.NextValidators.Proposer.Address, val1PubKey.Address())
  539. _, updatedVal1 := updatedState2.NextValidators.GetByAddress(val1PubKey.Address())
  540. _, oldVal1 := updatedState2.Validators.GetByAddress(val1PubKey.Address())
  541. _, updatedVal2 := updatedState2.NextValidators.GetByAddress(val2PubKey.Address())
  542. // 1. Add
  543. val2VotingPower := val1VotingPower
  544. totalPower = val1VotingPower + val2VotingPower // 20
  545. v2PrioWhenAddedVal2 := -(totalPower + (totalPower >> 3)) // -22
  546. // 2. Scale - noop
  547. // 3. Center
  548. avgSum := big.NewInt(0).Add(big.NewInt(v2PrioWhenAddedVal2), big.NewInt(oldVal1.ProposerPriority))
  549. avg := avgSum.Div(avgSum, big.NewInt(2)) // -11
  550. expectedVal2Prio := v2PrioWhenAddedVal2 - avg.Int64() // -11
  551. expectedVal1Prio := oldVal1.ProposerPriority - avg.Int64() // 11
  552. // 4. Increment
  553. expectedVal2Prio += val2VotingPower // -11 + 10 = -1
  554. expectedVal1Prio += val1VotingPower // 11 + 10 == 21
  555. expectedVal1Prio -= totalPower // 1, val1 proposer
  556. assert.EqualValues(t, expectedVal1Prio, updatedVal1.ProposerPriority)
  557. assert.EqualValues(
  558. t,
  559. expectedVal2Prio,
  560. updatedVal2.ProposerPriority,
  561. "unexpected proposer priority for validator: %v",
  562. updatedVal2,
  563. )
  564. validatorUpdates, err = types.PB2TM.ValidatorUpdates(abciResponses.FinalizeBlock.ValidatorUpdates)
  565. require.NoError(t, err)
  566. updatedState3, err := updatedState2.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  567. assert.NoError(t, err)
  568. assert.Equal(t, updatedState3.Validators.Proposer.Address, updatedState3.NextValidators.Proposer.Address)
  569. assert.Equal(t, updatedState3.Validators, updatedState2.NextValidators)
  570. _, updatedVal1 = updatedState3.NextValidators.GetByAddress(val1PubKey.Address())
  571. _, updatedVal2 = updatedState3.NextValidators.GetByAddress(val2PubKey.Address())
  572. // val1 will still be proposer:
  573. assert.Equal(t, val1PubKey.Address(), updatedState3.NextValidators.Proposer.Address)
  574. // check if expected proposer prio is matched:
  575. // Increment
  576. expectedVal2Prio2 := expectedVal2Prio + val2VotingPower // -1 + 10 = 9
  577. expectedVal1Prio2 := expectedVal1Prio + val1VotingPower // 1 + 10 == 11
  578. expectedVal1Prio2 -= totalPower // -9, val1 proposer
  579. assert.EqualValues(
  580. t,
  581. expectedVal1Prio2,
  582. updatedVal1.ProposerPriority,
  583. "unexpected proposer priority for validator: %v",
  584. updatedVal2,
  585. )
  586. assert.EqualValues(
  587. t,
  588. expectedVal2Prio2,
  589. updatedVal2.ProposerPriority,
  590. "unexpected proposer priority for validator: %v",
  591. updatedVal2,
  592. )
  593. // no changes in voting power and both validators have same voting power
  594. // -> proposers should alternate:
  595. oldState := updatedState3
  596. abciResponses = &tmstate.ABCIResponses{
  597. FinalizeBlock: &abci.ResponseFinalizeBlock{
  598. ValidatorUpdates: nil,
  599. },
  600. }
  601. validatorUpdates, err = types.PB2TM.ValidatorUpdates(abciResponses.FinalizeBlock.ValidatorUpdates)
  602. require.NoError(t, err)
  603. oldState, err = oldState.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  604. assert.NoError(t, err)
  605. expectedVal1Prio2 = 1
  606. expectedVal2Prio2 = -1
  607. expectedVal1Prio = -9
  608. expectedVal2Prio = 9
  609. for i := 0; i < 1000; i++ {
  610. // no validator updates:
  611. abciResponses := &tmstate.ABCIResponses{
  612. FinalizeBlock: &abci.ResponseFinalizeBlock{
  613. ValidatorUpdates: nil,
  614. },
  615. }
  616. validatorUpdates, err = types.PB2TM.ValidatorUpdates(abciResponses.FinalizeBlock.ValidatorUpdates)
  617. require.NoError(t, err)
  618. updatedState, err := oldState.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  619. assert.NoError(t, err)
  620. // alternate (and cyclic priorities):
  621. assert.NotEqual(
  622. t,
  623. updatedState.Validators.Proposer.Address,
  624. updatedState.NextValidators.Proposer.Address,
  625. "iter: %v",
  626. i,
  627. )
  628. assert.Equal(t, oldState.Validators.Proposer.Address, updatedState.NextValidators.Proposer.Address, "iter: %v", i)
  629. _, updatedVal1 = updatedState.NextValidators.GetByAddress(val1PubKey.Address())
  630. _, updatedVal2 = updatedState.NextValidators.GetByAddress(val2PubKey.Address())
  631. if i%2 == 0 {
  632. assert.Equal(t, updatedState.Validators.Proposer.Address, val2PubKey.Address())
  633. assert.Equal(t, expectedVal1Prio, updatedVal1.ProposerPriority) // -19
  634. assert.Equal(t, expectedVal2Prio, updatedVal2.ProposerPriority) // 0
  635. } else {
  636. assert.Equal(t, updatedState.Validators.Proposer.Address, val1PubKey.Address())
  637. assert.Equal(t, expectedVal1Prio2, updatedVal1.ProposerPriority) // -9
  638. assert.Equal(t, expectedVal2Prio2, updatedVal2.ProposerPriority) // -10
  639. }
  640. // update for next iteration:
  641. oldState = updatedState
  642. }
  643. }
  644. func TestLargeGenesisValidator(t *testing.T) {
  645. tearDown, _, state := setupTestCase(t)
  646. defer tearDown(t)
  647. genesisVotingPower := types.MaxTotalVotingPower / 1000
  648. genesisPubKey := ed25519.GenPrivKey().PubKey()
  649. // fmt.Println("genesis addr: ", genesisPubKey.Address())
  650. genesisVal := &types.Validator{
  651. Address: genesisPubKey.Address(),
  652. PubKey: genesisPubKey,
  653. VotingPower: genesisVotingPower,
  654. }
  655. // reset state validators to above validator
  656. state.Validators = types.NewValidatorSet([]*types.Validator{genesisVal})
  657. state.NextValidators = state.Validators
  658. require.True(t, len(state.Validators.Validators) == 1)
  659. // update state a few times with no validator updates
  660. // asserts that the single validator's ProposerPrio stays the same
  661. oldState := state
  662. for i := 0; i < 10; i++ {
  663. // no updates:
  664. abciResponses := &tmstate.ABCIResponses{
  665. FinalizeBlock: &abci.ResponseFinalizeBlock{
  666. ValidatorUpdates: nil,
  667. },
  668. }
  669. validatorUpdates, err := types.PB2TM.ValidatorUpdates(abciResponses.FinalizeBlock.ValidatorUpdates)
  670. require.NoError(t, err)
  671. block, err := statefactory.MakeBlock(oldState, oldState.LastBlockHeight+1, new(types.Commit))
  672. require.NoError(t, err)
  673. bps, err := block.MakePartSet(testPartSize)
  674. require.NoError(t, err)
  675. blockID := types.BlockID{Hash: block.Hash(), PartSetHeader: bps.Header()}
  676. updatedState, err := oldState.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  677. require.NoError(t, err)
  678. // no changes in voting power (ProposerPrio += VotingPower == Voting in 1st round; than shiftByAvg == 0,
  679. // than -Total == -Voting)
  680. // -> no change in ProposerPrio (stays zero):
  681. assert.EqualValues(t, oldState.NextValidators, updatedState.NextValidators)
  682. assert.EqualValues(t, 0, updatedState.NextValidators.Proposer.ProposerPriority)
  683. oldState = updatedState
  684. }
  685. // add another validator, do a few iterations (create blocks),
  686. // add more validators with same voting power as the 2nd
  687. // let the genesis validator "unbond",
  688. // see how long it takes until the effect wears off and both begin to alternate
  689. // see: https://github.com/tendermint/tendermint/issues/2960
  690. firstAddedValPubKey := ed25519.GenPrivKey().PubKey()
  691. firstAddedValVotingPower := int64(10)
  692. fvp, err := encoding.PubKeyToProto(firstAddedValPubKey)
  693. require.NoError(t, err)
  694. firstAddedVal := abci.ValidatorUpdate{PubKey: fvp, Power: firstAddedValVotingPower}
  695. validatorUpdates, err := types.PB2TM.ValidatorUpdates([]abci.ValidatorUpdate{firstAddedVal})
  696. assert.NoError(t, err)
  697. abciResponses := &tmstate.ABCIResponses{
  698. FinalizeBlock: &abci.ResponseFinalizeBlock{
  699. ValidatorUpdates: []abci.ValidatorUpdate{firstAddedVal},
  700. },
  701. }
  702. block, err := statefactory.MakeBlock(oldState, oldState.LastBlockHeight+1, new(types.Commit))
  703. require.NoError(t, err)
  704. bps, err := block.MakePartSet(testPartSize)
  705. require.NoError(t, err)
  706. blockID := types.BlockID{Hash: block.Hash(), PartSetHeader: bps.Header()}
  707. updatedState, err := oldState.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  708. require.NoError(t, err)
  709. lastState := updatedState
  710. for i := 0; i < 200; i++ {
  711. // no updates:
  712. abciResponses := &tmstate.ABCIResponses{
  713. FinalizeBlock: &abci.ResponseFinalizeBlock{
  714. ValidatorUpdates: nil,
  715. },
  716. }
  717. validatorUpdates, err := types.PB2TM.ValidatorUpdates(abciResponses.FinalizeBlock.ValidatorUpdates)
  718. require.NoError(t, err)
  719. block, err := statefactory.MakeBlock(lastState, lastState.LastBlockHeight+1, new(types.Commit))
  720. require.NoError(t, err)
  721. bps, err = block.MakePartSet(testPartSize)
  722. require.NoError(t, err)
  723. blockID := types.BlockID{Hash: block.Hash(), PartSetHeader: bps.Header()}
  724. updatedStateInner, err := lastState.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  725. require.NoError(t, err)
  726. lastState = updatedStateInner
  727. }
  728. // set state to last state of above iteration
  729. state = lastState
  730. // set oldState to state before above iteration
  731. oldState = updatedState
  732. _, oldGenesisVal := oldState.NextValidators.GetByAddress(genesisVal.Address)
  733. _, newGenesisVal := state.NextValidators.GetByAddress(genesisVal.Address)
  734. _, addedOldVal := oldState.NextValidators.GetByAddress(firstAddedValPubKey.Address())
  735. _, addedNewVal := state.NextValidators.GetByAddress(firstAddedValPubKey.Address())
  736. // expect large negative proposer priority for both (genesis validator decreased, 2nd validator increased):
  737. assert.True(t, oldGenesisVal.ProposerPriority > newGenesisVal.ProposerPriority)
  738. assert.True(t, addedOldVal.ProposerPriority < addedNewVal.ProposerPriority)
  739. // add 10 validators with the same voting power as the one added directly after genesis:
  740. for i := 0; i < 10; i++ {
  741. addedPubKey := ed25519.GenPrivKey().PubKey()
  742. ap, err := encoding.PubKeyToProto(addedPubKey)
  743. require.NoError(t, err)
  744. addedVal := abci.ValidatorUpdate{PubKey: ap, Power: firstAddedValVotingPower}
  745. validatorUpdates, err := types.PB2TM.ValidatorUpdates([]abci.ValidatorUpdate{addedVal})
  746. assert.NoError(t, err)
  747. abciResponses := &tmstate.ABCIResponses{
  748. FinalizeBlock: &abci.ResponseFinalizeBlock{
  749. ValidatorUpdates: []abci.ValidatorUpdate{addedVal},
  750. },
  751. }
  752. block, err := statefactory.MakeBlock(oldState, oldState.LastBlockHeight+1, new(types.Commit))
  753. require.NoError(t, err)
  754. bps, err := block.MakePartSet(testPartSize)
  755. require.NoError(t, err)
  756. blockID := types.BlockID{Hash: block.Hash(), PartSetHeader: bps.Header()}
  757. state, err = state.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  758. require.NoError(t, err)
  759. }
  760. require.Equal(t, 10+2, len(state.NextValidators.Validators))
  761. // remove genesis validator:
  762. gp, err := encoding.PubKeyToProto(genesisPubKey)
  763. require.NoError(t, err)
  764. removeGenesisVal := abci.ValidatorUpdate{PubKey: gp, Power: 0}
  765. abciResponses = &tmstate.ABCIResponses{
  766. FinalizeBlock: &abci.ResponseFinalizeBlock{
  767. ValidatorUpdates: []abci.ValidatorUpdate{removeGenesisVal},
  768. },
  769. }
  770. block, err = statefactory.MakeBlock(oldState, oldState.LastBlockHeight+1, new(types.Commit))
  771. require.NoError(t, err)
  772. bps, err = block.MakePartSet(testPartSize)
  773. require.NoError(t, err)
  774. blockID = types.BlockID{Hash: block.Hash(), PartSetHeader: bps.Header()}
  775. validatorUpdates, err = types.PB2TM.ValidatorUpdates(abciResponses.FinalizeBlock.ValidatorUpdates)
  776. require.NoError(t, err)
  777. updatedState, err = state.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  778. require.NoError(t, err)
  779. // only the first added val (not the genesis val) should be left
  780. assert.Equal(t, 11, len(updatedState.NextValidators.Validators))
  781. // call update state until the effect for the 3rd added validator
  782. // being proposer for a long time after the genesis validator left wears off:
  783. curState := updatedState
  784. count := 0
  785. isProposerUnchanged := true
  786. for isProposerUnchanged {
  787. abciResponses := &tmstate.ABCIResponses{
  788. FinalizeBlock: &abci.ResponseFinalizeBlock{
  789. ValidatorUpdates: nil,
  790. },
  791. }
  792. validatorUpdates, err = types.PB2TM.ValidatorUpdates(abciResponses.FinalizeBlock.ValidatorUpdates)
  793. require.NoError(t, err)
  794. block, err = statefactory.MakeBlock(curState, curState.LastBlockHeight+1, new(types.Commit))
  795. require.NoError(t, err)
  796. bps, err := block.MakePartSet(testPartSize)
  797. require.NoError(t, err)
  798. blockID = types.BlockID{Hash: block.Hash(), PartSetHeader: bps.Header()}
  799. curState, err = curState.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  800. require.NoError(t, err)
  801. if !bytes.Equal(curState.Validators.Proposer.Address, curState.NextValidators.Proposer.Address) {
  802. isProposerUnchanged = false
  803. }
  804. count++
  805. }
  806. updatedState = curState
  807. // the proposer changes after this number of blocks
  808. firstProposerChangeExpectedAfter := 1
  809. assert.Equal(t, firstProposerChangeExpectedAfter, count)
  810. // store proposers here to see if we see them again in the same order:
  811. numVals := len(updatedState.Validators.Validators)
  812. proposers := make([]*types.Validator, numVals)
  813. for i := 0; i < 100; i++ {
  814. // no updates:
  815. abciResponses := &tmstate.ABCIResponses{
  816. FinalizeBlock: &abci.ResponseFinalizeBlock{
  817. ValidatorUpdates: nil,
  818. },
  819. }
  820. validatorUpdates, err := types.PB2TM.ValidatorUpdates(abciResponses.FinalizeBlock.ValidatorUpdates)
  821. require.NoError(t, err)
  822. block, err := statefactory.MakeBlock(updatedState, updatedState.LastBlockHeight+1, new(types.Commit))
  823. require.NoError(t, err)
  824. bps, err := block.MakePartSet(testPartSize)
  825. require.NoError(t, err)
  826. blockID := types.BlockID{Hash: block.Hash(), PartSetHeader: bps.Header()}
  827. updatedState, err = updatedState.Update(blockID, &block.Header, sm.ABCIResponsesResultsHash(abciResponses), abciResponses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  828. require.NoError(t, err)
  829. if i > numVals { // expect proposers to cycle through after the first iteration (of numVals blocks):
  830. if proposers[i%numVals] == nil {
  831. proposers[i%numVals] = updatedState.NextValidators.Proposer
  832. } else {
  833. assert.Equal(t, proposers[i%numVals], updatedState.NextValidators.Proposer)
  834. }
  835. }
  836. }
  837. }
  838. func TestStoreLoadValidatorsIncrementsProposerPriority(t *testing.T) {
  839. const valSetSize = 2
  840. tearDown, stateDB, state := setupTestCase(t)
  841. t.Cleanup(func() { tearDown(t) })
  842. stateStore := sm.NewStore(stateDB)
  843. state.Validators = genValSet(valSetSize)
  844. state.NextValidators = state.Validators.CopyIncrementProposerPriority(1)
  845. err := stateStore.Save(state)
  846. require.NoError(t, err)
  847. nextHeight := state.LastBlockHeight + 1
  848. v0, err := stateStore.LoadValidators(nextHeight)
  849. assert.NoError(t, err)
  850. acc0 := v0.Validators[0].ProposerPriority
  851. v1, err := stateStore.LoadValidators(nextHeight + 1)
  852. assert.NoError(t, err)
  853. acc1 := v1.Validators[0].ProposerPriority
  854. assert.NotEqual(t, acc1, acc0, "expected ProposerPriority value to change between heights")
  855. }
  856. // TestValidatorChangesSaveLoad tests saving and loading a validator set with
  857. // changes.
  858. func TestManyValidatorChangesSaveLoad(t *testing.T) {
  859. const valSetSize = 7
  860. tearDown, stateDB, state := setupTestCase(t)
  861. defer tearDown(t)
  862. stateStore := sm.NewStore(stateDB)
  863. require.Equal(t, int64(0), state.LastBlockHeight)
  864. state.Validators = genValSet(valSetSize)
  865. state.NextValidators = state.Validators.CopyIncrementProposerPriority(1)
  866. err := stateStore.Save(state)
  867. require.NoError(t, err)
  868. _, valOld := state.Validators.GetByIndex(0)
  869. var pubkeyOld = valOld.PubKey
  870. pubkey := ed25519.GenPrivKey().PubKey()
  871. // Swap the first validator with a new one (validator set size stays the same).
  872. header, blockID, responses := makeHeaderPartsResponsesValPubKeyChange(t, state, pubkey)
  873. // Save state etc.
  874. var validatorUpdates []*types.Validator
  875. validatorUpdates, err = types.PB2TM.ValidatorUpdates(responses.FinalizeBlock.ValidatorUpdates)
  876. require.NoError(t, err)
  877. state, err = state.Update(blockID, &header, sm.ABCIResponsesResultsHash(responses), responses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  878. require.NoError(t, err)
  879. nextHeight := state.LastBlockHeight + 1
  880. err = stateStore.Save(state)
  881. require.NoError(t, err)
  882. // Load nextheight, it should be the oldpubkey.
  883. v0, err := stateStore.LoadValidators(nextHeight)
  884. assert.NoError(t, err)
  885. assert.Equal(t, valSetSize, v0.Size())
  886. index, val := v0.GetByAddress(pubkeyOld.Address())
  887. assert.NotNil(t, val)
  888. if index < 0 {
  889. t.Fatal("expected to find old validator")
  890. }
  891. // Load nextheight+1, it should be the new pubkey.
  892. v1, err := stateStore.LoadValidators(nextHeight + 1)
  893. assert.NoError(t, err)
  894. assert.Equal(t, valSetSize, v1.Size())
  895. index, val = v1.GetByAddress(pubkey.Address())
  896. assert.NotNil(t, val)
  897. if index < 0 {
  898. t.Fatal("expected to find newly added validator")
  899. }
  900. }
  901. func TestStateMakeBlock(t *testing.T) {
  902. tearDown, _, state := setupTestCase(t)
  903. defer tearDown(t)
  904. proposerAddress := state.Validators.GetProposer().Address
  905. stateVersion := state.Version.Consensus
  906. block, err := statefactory.MakeBlock(state, 2, new(types.Commit))
  907. require.NoError(t, err)
  908. // test we set some fields
  909. assert.Equal(t, stateVersion, block.Version)
  910. assert.Equal(t, proposerAddress, block.ProposerAddress)
  911. }
  912. // TestConsensusParamsChangesSaveLoad tests saving and loading consensus params
  913. // with changes.
  914. func TestConsensusParamsChangesSaveLoad(t *testing.T) {
  915. tearDown, stateDB, state := setupTestCase(t)
  916. defer tearDown(t)
  917. stateStore := sm.NewStore(stateDB)
  918. // Change vals at these heights.
  919. changeHeights := []int64{1, 2, 4, 5, 10, 15, 16, 17, 20}
  920. N := len(changeHeights)
  921. // Each valset is just one validator.
  922. // create list of them.
  923. params := make([]types.ConsensusParams, N+1)
  924. params[0] = state.ConsensusParams
  925. for i := 1; i < N+1; i++ {
  926. params[i] = *types.DefaultConsensusParams()
  927. params[i].Block.MaxBytes += int64(i)
  928. }
  929. // Build the params history by running updateState
  930. // with the right params set for each height.
  931. highestHeight := changeHeights[N-1] + 5
  932. changeIndex := 0
  933. cp := params[changeIndex]
  934. var err error
  935. var validatorUpdates []*types.Validator
  936. for i := int64(1); i < highestHeight; i++ {
  937. // When we get to a change height, use the next params.
  938. if changeIndex < len(changeHeights) && i == changeHeights[changeIndex] {
  939. changeIndex++
  940. cp = params[changeIndex]
  941. }
  942. header, blockID, responses := makeHeaderPartsResponsesParams(t, state, &cp)
  943. validatorUpdates, err = types.PB2TM.ValidatorUpdates(responses.FinalizeBlock.ValidatorUpdates)
  944. require.NoError(t, err)
  945. state, err = state.Update(blockID, &header, sm.ABCIResponsesResultsHash(responses), responses.FinalizeBlock.ConsensusParamUpdates, validatorUpdates)
  946. require.NoError(t, err)
  947. err := stateStore.Save(state)
  948. require.NoError(t, err)
  949. }
  950. // Make all the test cases by using the same params until after the change.
  951. testCases := make([]paramsChangeTestCase, highestHeight)
  952. changeIndex = 0
  953. cp = params[changeIndex]
  954. for i := int64(1); i < highestHeight+1; i++ {
  955. // We get to the height after a change height use the next pubkey (note
  956. // our counter starts at 0 this time).
  957. if changeIndex < len(changeHeights) && i == changeHeights[changeIndex]+1 {
  958. changeIndex++
  959. cp = params[changeIndex]
  960. }
  961. testCases[i-1] = paramsChangeTestCase{i, cp}
  962. }
  963. for _, testCase := range testCases {
  964. p, err := stateStore.LoadConsensusParams(testCase.height)
  965. assert.NoError(t, err, fmt.Sprintf("expected no err at height %d", testCase.height))
  966. assert.EqualValues(t, testCase.params, p, fmt.Sprintf(`unexpected consensus params at
  967. height %d`, testCase.height))
  968. }
  969. }
  970. func TestStateProto(t *testing.T) {
  971. tearDown, _, state := setupTestCase(t)
  972. defer tearDown(t)
  973. tc := []struct {
  974. testName string
  975. state *sm.State
  976. expPass1 bool
  977. expPass2 bool
  978. }{
  979. {"empty state", &sm.State{}, true, false},
  980. {"nil failure state", nil, false, false},
  981. {"success state", &state, true, true},
  982. }
  983. for _, tt := range tc {
  984. tt := tt
  985. pbs, err := tt.state.ToProto()
  986. if !tt.expPass1 {
  987. assert.Error(t, err)
  988. } else {
  989. assert.NoError(t, err, tt.testName)
  990. }
  991. smt, err := sm.FromProto(pbs)
  992. if tt.expPass2 {
  993. require.NoError(t, err, tt.testName)
  994. require.Equal(t, tt.state, smt, tt.testName)
  995. } else {
  996. require.Error(t, err, tt.testName)
  997. }
  998. }
  999. }