Avoid allocating a Set every time requestIdCompatible is called
#8522
Conversation
Signed-off-by: Fabio Di Fabio <fabio.difabio@consensys.net>
| .contains(code); | ||
| EthProtocolMessages.RECEIPTS); | ||
| REQUEST_ID_MESSAGES = | ||
| new BitSet(requestIdMessages.stream().mapToInt(i -> i).max().getAsInt() + 1); |
There was a problem hiding this comment.
is this more optimal than just getting the size of the set?
There was a problem hiding this comment.
We can make it even constant, like 64 which is the size of a long (64 bits), if we want to decouple it from requestIdMessages.
There was a problem hiding this comment.
the requestIdMessage is the index of the bitset, so the bitset should be large enough for the message with the largest index, since bitset starts from 0 then the optimal size is max index + 1, not sure what do you mean by decoupling?
Then I plan to make EthProtocolMessages an enum, so we can directly use EnumSet that is clearer.
| .contains(code); | ||
| EthProtocolMessages.RECEIPTS); | ||
| REQUEST_ID_MESSAGES = | ||
| new BitSet(requestIdMessages.stream().mapToInt(i -> i).max().getAsInt() + 1); |
There was a problem hiding this comment.
We can make it even constant, like 64 which is the size of a long (64 bits), if we want to decouple it from requestIdMessages.
PR description
EthProtocol.requestIdCompatibleis called for every p2p message, and in the previous implementation it was allocating a newSetevery time, but since the data in the set does not change, it can be created once, moreover to avoid (un)boxing from and to int and hashes, using aBitSetoptimize the lookup.Fixed Issue(s)
Thanks for sending a pull request! Have you done the following?
doc-change-requiredlabel to this PR if updates are required.Locally, you can run these tests to catch failures early:
./gradlew spotlessApply./gradlew build./gradlew acceptanceTest./gradlew integrationTest./gradlew ethereum:referenceTests:referenceTests