Thanks to awemany for pointing this out.
This replaces #10172 which appears to be abandoned, but uses the constants as requested on that PR.
Thanks to awemany for pointing this out.
504 | @@ -504,7 +505,7 @@ static bool AcceptToMemoryPoolWorker(const CChainParams& chainparams, CTxMemPool 505 | { 506 | for (const CTxIn &_txin : ptxConflicting->vin) 507 | { 508 | - if (_txin.nSequence < std::numeric_limits<unsigned int>::max()-1) 509 | + if (_txin.nSequence <= MAX_BIP125_RBF_SEQUENCE)
MAX_BIP125_RBF_SEQUENCE is 0xfffffffd
The old value was 0xfffffffe (on supported platforms)
Is this intended?
Edit: oh, <=, ok never mind
19 | @@ -20,6 +20,7 @@ 20 | #include "init.h" 21 | #include "policy/fees.h" 22 | #include "policy/policy.h" 23 | +#include "policy/rbf.h"
Do we want validation to depend on those policy headers? It seems the wrong way around.
ATMP does policy-based decisions. I think the real answer here is ATMP belongs outside of validation.cpp, but dear god thats a big refactor (I think).
I don't make the standards, I just follow them. (didn't think much about it because of the above two lines). :)
Well it would be wrong to rely on a policy constant if ATMP was a consensus critical function, but you're right.
utACK 095b9174645f5b855dd5946c99cea4f4ffb5a034
utACK 095b917