Child-pays-for-parent mempool #2814

pull petertodd wants to merge 3 commits into bitcoin:master from petertodd:greedy-child-pays-for-parent changing 6 files +566 −69

petertodd commented at 10:29 AM on July 4, 2013: contributor

Calculates total uncommitted fees and transaction size required to include the transaction in a block for all transactions in the mempool. Cost is O(1) for mempool.accept() and limited to O(n) when a new block changes what transactions are in the mempool, either due to them being mined, conflicts, or reinstated by a re-org.

I haven't changed the mining code, and don't keep track of sigops yet. This isn't useful without mining code and mempool expiration so it's not ready to be actually pulled yet. (mainly I want to see what the pulltester thinks)

Note that this is a safe implementation, not a 100% accurate one, for reasons mentioned in CMemPoolTx::calcPrioritySums() In addition the memoized approach can't directly handle the case where a transaction with a lot of outputs is made worthwhile to mine because of multiple children.

Child-pays-for-parent mempool

Calculates total uncommitted fees and transaction size required to
include the transaction in a block for all transactions in the mempool.
Cost is O(1) for mempool.accept() and maximum O(n) when a new block
changes what transactions are in the mempool, either due to them being
mined, conflicts, or a re-org.

No changes to CreateNewBlock() yet.

dac0995713

n 3e1aabd50c
n2 584c596341
BitcoinPullTester commented at 5:40 AM on July 20, 2013: none
Automatic sanity-testing: FAILED BUILD/TEST, see http://jenkins.bluematt.me/pull-tester/584c5963412effab498fdcb42280fb3265759a00 for binaries and test log.

This could happen for one of several reasons:
1. It chanages paths in makefile.linux-mingw or otherwise changes build scripts in a way that made them incompatible with the automated testing scripts (please tweak those patches in contrib/test-scripts)
2. It adds/modifies tests which test network rules (thanks for doing that), which conflicts with a patch applied at test time
3. It does not build on either Linux i386 or Win32 (via MinGW cross compile)
4. The test suite fails on either Linux i386 or Win32
5. The block test-cases failed (lookup the first bNN identifier which failed in https://github.com/TheBlueMatt/test-scripts/blob/master/FullBlockTestGenerator.java)
If you believe this to be in error, please ping BlueMatt on freenode or TheBlueMatt here.

This test script verifies pulls every time they are updated. It, however, dies sometimes and fails to test properly. If you are waiting on a test, please check timestamps to verify that the test.log is moving at http://jenkins.bluematt.me/pull-tester/current/ Contact BlueMatt on freenode if something looks broken.
luke-jr commented at 12:46 AM on August 30, 2013: member

bitcoin-qt.ljr20130721: src/main.h:1204: double CMemPoolTx::FeesPerKB() const: Assertion `nSumTxSize >= 0' failed.
petertodd commented at 1:06 AM on August 30, 2013: contributor

@luke-jr FWIW sipa has told me that his headers first patch changes a lot of the code that this mempool patch touches, so I'm going to wait until he's done that before working on this further.
gavinandresen commented at 2:03 AM on October 21, 2013: contributor

Closing; please open a new pull request after headers-first has been merged. I'm trying to clean up the open pull requests to "things ready for testing and merging right now."
gavinandresen closed this on Oct 21, 2013
Bushstar referenced this in commit bfc288afb4 on Apr 5, 2019
DrahtBot locked this on Sep 8, 2021

Contributors