Minor adjustments for the tests and codestyle

[dsaveliev]

During the #1062 a couple of minor problems was found:

• GetVirtualTransactionSize must have a sigOpCost as an argument. Right now this test passes just by coincidence.
• In p2p_fingerprint.py we need to take unspent coins before the substituted subchain.
• Clang formatting

Signed-off-by: Dmitry Saveliev [email protected]

[cmihai]

GetVirtualTransactionSize must have a sigOpCost as an argument. Right now this test passes just by coincidence.

I think it's not coincidental that the test passes, since the scripts used in the test have a sigop cost of 0 anyway. It doesn't hurt to add the parameter, but Bitcoin doesn't have it, so maybe we should keep everything as it is?

[castarco]

Although this PR is quite small, I think we should separate the fix to the p2p_fingerprint test from the style changes.

[dsaveliev]

I think it's not coincidental that the test passes, since the scripts used in the test have a sigop cost of 0 anyway. It doesn't hurt to add the parameter, but Bitcoin doesn't have it, so maybe we should keep everything as it is?

Not quite, TestMemPoolEntryHelper is initialized with sigOpCost = 4:

unit-e/src/test/test_unite.h

Line 122 in c85d170

spendsCoinbase(false), sigOpCost(4) { }

which goes further to CTxMemPoolEntry

unit-e/src/test/test_unite.cpp

Line 169 in c85d170

spendsCoinbase, sigOpCost, lp);

which finally affects indexed_transaction_set

unit-e/src/txmempool.cpp

Line 100 in c85d170

mapTx.modify(cit, update_ancestor_state(updateIt->GetTxSize(), updateIt->GetModifiedFee(), 1, updateIt->GetSigOpCost()));

and finally affects the TX order.
So, here we have a green test just because of fee based on nWeight, which match to the value, calculated inside of indexed_transaction_set:

unit-e/src/policy/policy.cpp

Line 265 in 365cf0d

return (std::max(nWeight, nSigOpCost * nBytesPerSigOp) + WITNESS_SCALE_FACTOR - 1) / WITNESS_SCALE_FACTOR;

[castarco]

utACK b42886c

[cmihai]

utACK b42886c