std::unordered_map::erase( const_iterator pos ) returns an iterator to the element following the removed one. Use that to optimize (probably minor-performance-wise, and definitely code-structure-wise) the implementation of CCoinsViewCache::BatchWrite().
Some people keep thinking that MAX_BLOCK_BASE_SIZE is a separate
size limit from the weight limit when it fact it is superfluous,
and used in early tests before the witness data has been
validated or just to compute worst case sizes. The size checks
that use it would not behave any differently consensus wise
if they were eliminated completely.
Its correct value is not independently settable but is a function
of the weight limit and weight formula.
This patch just eliminates it and uses the scale factor as
required to compute the worse case constants.
It also moves the weight factor out of primitives into consensus,
which is a more logical place for it.
This removes the possibility for GetCoin/HaveCoin/HaveCoinInCache to return
true while the respective coin is spent. By doing it across all calls, some
extra checks can be eliminated.
coins_tests is modified to call HaveCoin sometimes before and sometimes
after AccessCoin. A further change is needed because the semantics for
GetCoin slightly changed, causing a pruned entry in the parent cache to not
be pulled into the child in FetchCoin.
Previously it was possible for HaveCoinInCache to return true for a spent
coin. It is more clear to keep the semantics the same. HaveCoinInCache is
used for two reasons:
- tracking coins we may want to uncache, in which case it is unlikely there
would be spent coins we could uncache (not dirty)
- optimistically checking whether we have already included a tx in the
blockchain, in which case a spent coin is not a reliable indicator that we have.
This patch makes several related changes:
* Changes the CCoinsView virtual methods (GetCoins, HaveCoins, ...)
to be COutPoint/Coin-based rather than txid/CCoins-based.
* Changes the chainstate db to a new incompatible format that is also
COutPoint/Coin based.
* Implements reconstruction code for hash_serialized_2.
* Adapts the coins_tests unit tests (thanks to Russell Yanofsky).
A side effect of the new CCoinsView model is that we can no longer
use the (unreliable) test for transaction outputs in the UTXO set
to determine whether we already have a particular transaction.
The new functions are:
* CCoinsViewCache::AddCoin: Add a single COutPoint/Coin pair.
* CCoinsViewCache::SpendCoin: Remove a single COutPoint.
* AddCoins: utility function that invokes CCoinsViewCache::AddCoin for
each output in a CTransaction.
* AccessByTxid: utility function that searches for any output with
a given txid.
* CCoinsViewCache::AccessCoin: retrieve the Coin for a COutPoint.
* CCoinsViewCache::HaveCoins: check whether a non-empty Coin exists
for a given COutPoint.
The AddCoin and SpendCoin methods will eventually replace ModifyCoins
and ModifyNewCoins, AddCoins will replace CCoins::FromTx, and the new
AccessCoins and HaveCoins functions will replace their per-txid
counterparts.
Note that AccessCoin for now returns a copy of the Coin object. In a
later commit it will be change to returning a const reference (which
keeps working in all call sites).
We were marking coins FRESH before being sure they were not overwriting dirty undo data. This condition was never reached in existing code because undo data was always flushed before UpdateCoins was called with new transactions, but could have been exposed in an otherwise safe refactor.
Clarify in the comments the assumptions made in ModifyNewCoins.
Add ability to undo transactions to UpdateCoins unit test.
Thanks to Russ Yanofsky for suggestion on how to make logic clearer and fixing up the ccoins_modify_new test cases.
Add a method Cursor() to CCoinsView that returns a cursor which can be
used to iterate over the whole UTXO set.
- rpc: Change gettxoutsetinfo to use new Cursor method
- txdb: Remove GetStats method - Now that GetStats is implemented in
terms of Cursor, remove it.
Compute the value of inputs that already are in the chain at time of mempool entry and only increase priority due to aging for those inputs. This effectively changes the CTxMemPoolEntry's GetPriority calculation from an upper bound to a lower bound.
It's possible coins with the same hash exist when you create a duplicate coinbase, so previously we were reading from the database to make sure we had the old coins cached so if we were to spend the new ones, the old ones would also be spent. This pull instead just marks the new coins as not fresh if they are from a coinbase, so if they are spent they will be written all the way down to the database anyway overwriting any duplicates.
Previously it would break if you flushed a parent cache while there was a child cache referring to it. This change will allow the flushing of parent caches.
When processing a new transaction, in addition to spending the Coins of its txin's it creates a new Coins for its outputs. The existing ModifyCoins function will first make sure this Coins does not already exist. It can not exist due to BIP 30, but because of that the lookup can't be cached and always has to go to the database. Since we are creating the coins to match the new tx anyway, there is no point in checking if they exist first anyway. However this should not be used for coinbase tx's in order to preserve the historical behavior of overwriting the two existing duplicate tx pairs.
Otherwise, if CCoinsViewCache::ModifyCoins throws an exception in between
setting hasModifier and constructing the CCoinsModifier, the cache ends up
in an inconsistent state, resulting in an assert failure in the next
modification.
Bug discovered by Wladimir J. van der Laan.
- Update comments in checkpoints to be doxygen compatible
- Update comments in checkqueue to be doxygen compatible
- Update coins to be doxygen compatible
- Fix comment typo in crypter.h
- Update licenses/copyright dates
Closes#5325#5184#5183#5182
Instead of storing CCoins entries directly in CCoinsMap, store a CCoinsCacheEntry
which additionally keeps track of whether a particular entry is:
* dirty: potentially different from its parent view.
* fresh: the parent view is known to not have a non-pruned version.
This allows us to skip non-dirty cache entries when pushing batches of changes up,
and to remove CCoins entries about transactions that are fully spent before the
parent cache learns about them.
All direct modifications are now done through ModifyCoins, and BatchWrite is
used for pushing batches of queued modifications up, so we don't need the
low-level SetCoins and SetBestBlock anymore in the top-level CCoinsView class.
Replace the mutable non-copying GetCoins method with a ModifyCoins, which
returns an encapsulated iterator, so we can keep track of concurrent
modifications (as iterators can be invalidated by those) and run cleanup
code after a modification is finished.
This also removes the overloading of the 'GetCoins' name.
The efficient version of CCoinsViewCache::GetCoins only works for known-to-exist
cache entries, requiring a separate HaveCoins call beforehand. This is
inefficient as both perform a hashtable lookup.
Replace the non-mutable GetCoins with AccessCoins, which returns a potentially-NULL
pointer. This also decreases the overloading of GetCoins.
Also replace some copying (inefficient) GetCoins calls with equivalent AccessCoins,
decreasing the copying.
Akira Takizawa