Explicitly unittest the canonical prunable txout format #2336

Test every invalid op code, and add an explicit test for Gavin's proposed format for provably unspendable/always prunable transactions using the 0xFF invalid opcode. Specifically any scriptPubKey of the following form:

0xFF {zero or more ops}

Because the scriptSig and scriptPubKey are processed separately one can't even use IF/ENDIF trickery to defeat this. I also like how it says "INVALID" quite clearly as the first opcode. I've tried this out on testnet with no issues.

Unless anyone has a reason why this standard is a bad idea I'm also going to raise the issue with P2Pool; currently they are the main producer of non-prunable txouts that could be made prunable this way.

No code to actually do anything with this standard is included in this patch; that can be done later.

Automatic sanity-testing: WARNING, see http://jenkins.bluematt.me/pull-tester/b7cf0454ad6d831965d8ace7209b54726c166aec for binaries and test log.

This pull decreases total test coverage, please add unit tests to test all new code and help us add test cases for existing code. Coverage report can be found at http://jenkins.bluematt.me/pull-tester/b7cf0454ad6d831965d8ace7209b54726c166aec/bitcoin/src/total.coverage/

@BitcoinPullTester LOL. Should he write a unit test for the unit test?

"Yo dawg..."

Of course, I did test that those unittest's fail if you invert the logic...

Changed to RETURN from 0xff as per IRC discussion.

Automatic sanity-testing: WARNING, see http://jenkins.bluematt.me/pull-tester/091f18419011bfee6242270e52237f524ca23c8e for binaries and test log.

This pull decreases total test coverage, please add unit tests to test all new code and help us add test cases for existing code. Coverage report can be found at http://jenkins.bluematt.me/pull-tester/091f18419011bfee6242270e52237f524ca23c8e/bitcoin/src/total.coverage/