* Copy over boilerplate and begin implementing TestDAA.
* Implement a fairly reliable method of hashing at a certain hashrate.
* Convert the DAA test to an application.
* Start kaspad and make sure that hashrate throttling works with that as well.
* Finish implementing testConstantHashRate.
* Tidied up a bit.
* Convert TestDAA back into a go test.
* Reorganize TestDAA to be more like a traditional test.
* Add sudden hashrate drop/jump tests.
* Simplify targetHashNanosecondsFunction.
* Improve progress logs.
* Add more tests.
* Remove the no-longer relevant `hashes` part of targetHashNanosecondsFunction.
* Implement a constant hashrate increase test.
* Implement a constant hashrate decrease test.
* Give the correct run duration to the constant hashrate decrease test.
* Add cooldowns to exponential functions.
* Add run.sh to the DAA test.
* Add a README.
* Add `daa` to run-slow.sh.
* Make go lint happy.
* Fix the README's title.
* Keep running tests even if one of them failed on high block rate deviation.
* Fix hashrate peak/valley tests.
* Preallocate arrays for hash and mining durations.
* Add more statistics to the "mined block" log.
* Make sure runDAATest stops when it's suppposed to.
* Add a newline after "5 minute cooldown."
* Fix variable names.
* Rename totalElapsedTime to tatalElapsedDuration.
* In measureMachineHashNanoseconds, generate a random nonce only once.
* In runDAATest, add "DAA" to the start/finish log.
* Remove --logdir from kaspadRunCommand.
* In runDAATest, enlarge the nonce range to the entirety of uint64.
* Explain what targetHashNanosecondsFunction is.
* Move RunKaspadForTesting into common.
* Rename runForDuration to loopForDuration.
* Make go lint happy.
* Extract fetchBlockForMining to a separate function.
* Extract waitUntilTargetHashDurationHadElapsed to a separate function.
* Extract pushHashDuration and pushMiningDuration to separate functions.
* Extract logMinedBlockStatsAndUpdateStatFields to a separate function.
* Extract submitMinedBlock to a separate function.
* Extract tryNonceForMiningAndIncrementNonce to a separate function.
* Add comments.
* Use a rolling average instead of appending to an array for performance/accuracy.
* Change a word in a comment.
* Explain why we wait for five minutes at the end of the exponential increase/decrease tests.
Co-authored-by: Svarog <feanorr@gmail.com>
* First limit the candidates size to 3*csm.maxBlockParents before taking the bottom csm.maxBlockParents/2
* Change log level of printing all tips to Tracef
* Add many-tips to run-fast.sh and run-slow.sh
* Fix preallocation size
* Assign intermediate variables
* Use selectedParent instead of selectedTip for non-selectedTip blocks in restoreSingleBlockStatus
* Cache the selectedParent for re-use in a resolveSingleBlockStatus chain
* Implement and use reverseUTXOSet
* Reverse blocks in correct order
* Support resolveBlockStatus without separate stagingAreas for usage of testConsensus
* Handle the case where the tip of the resolved block is not the next selectedTip
* Unify isResolveTip
* Some minor fixes and cleanup
* Add full finality window re-org test to stability-slow
* rename: useSeparateStagingAreasPerBlock -> useSeparateStagingAreaPerBlock
* Better logs in resolveSingleBlockStatus
* A few retouches to reverseUTXODiffs
* TEMPORARY COMMIT: EXTRAT ALL DIFFFROMS TO SEPARATE METHODS
* TEMPORARY COMMIT: REMOVE DIFFICULTY CHECKS IN DEVNET
* Don't pre-allocate in utxo-algebra, since the numbers are not known ahead-of-time
* Add some logs to reverseUTXODiffs
* Revert "TEMPORARY COMMIT: REMOVE DIFFICULTY CHECKS IN DEVNET"
This reverts commit c0af9dc6ade78a914c970e11bc63c34605565f57.
* Revert "TEMPORARY COMMIT: EXTRAT ALL DIFFFROMS TO SEPARATE METHODS"
This reverts commit 4fcca1b48c3a1183598833a355b9bfaf169edba1.
* Remove redundant paranthesis
* Revise some logs messages
* Rename:oneBlockBeforeCurrentUTXOSet -> lastResolvedBlockUTXOSet
* Don't break if the block was resolved as invalid
* rename unverifiedBlocks to recentlyVerifiedBlcks in reverseUTXODiffs
* Add errors.New to the panic, for a stack trace
* Reverse the UTXODiffs after the main block has been commited
* Use the correct value for previousUTXODiff
* Add test for ReverseUTXODiff
* Fix some names and comments
* Update TestReverseUTXODiffs to use consensus.Config
* Fix comments mentioning 'oneBlockBeforeTip'
* Copy some boilerplate from the other stability tests.
* Fix a copy+paste error in run.sh.
* Copy over some stability test boilerplate go code.
* Run kaspad in the background.
* Catch panics and initialize the RPC client.
* Mine enough blocks to fund filling up the mempool.
* Extract coinbase transactions out of the generated blocks.
* Tidy up a bit.
* Implement submitting transactions.
* Lower the amount of outputs in each transaction.
* Verify that the mempool size has the expected amount of transactions.
* Pregenerate enough funds before submitting the first transaction so that block creation doesn't interfere with the test.
* Empty mempool out by continuously adding blocks to the DAG.
* Handle orphan transactions when overfilling the mempool.
* Increase mempoolSizeLimit to 1m.
* Fix a comment.
* Fix a comment.
* Add mempool-limits to run-slow.sh.
* Rename generateTransactionsWithLotsOfOutputs to generateTransactionsWithMultipleOutputs.
* Rename generateCoinbaseTransaction to mineBlockAndGetCoinbaseTransaction.
* Make generateFundingCoinbaseTransactions return an object instead of store a global variable.
* Convert mempool-limits into a Go test.
* Convert panics to t.Fatalfs.
* Fix a comment.
* Increase mempoolSizeLimit to 1m.
* Run TestMempoolLimits only if RUN_STABILITY_TESTS is set.
* Move the run of mempool-limits in run-slow.sh.
* Add a comment above fundingCoinbaseTransactions.
* Make a couple of stylistic changes.
* Use transactionhelper.CoinbaseTransactionIndex instead of hardcoding 0.
* Make uninteresting errors print %+v instead of %s.
Co-authored-by: Svarog <feanorr@gmail.com>
