Python-based regression tests #3763

These are in a "good enough to get other people involved" state. Improvements welcome, but I think we should pull this before nitpicking it too much. See skeleton.py for the "do-almost-nothing" regression test.

So future changes to these files should be submitted to https://github.com/jgarzik/python-bitcoinrpc first or should python-bitcoinrpc.git be deprecated? And maybe this answer could be included in this README either way.

@kanzure : python-bitcoinrpc changes should be submitted upstream. I updated the README to make that clear.

Automatic sanity-testing: PASSED, see http://jenkins.bluematt.me/pull-tester/e1c0e4b2fdffcbafbb71089dcc6b501669ff51c3 for binaries and test log. 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.

FWIW python-bitcoinrpc is considered depreciated by jgarzik; it's now incorporated into python-bitcoinlib, which in turn I've been maintaining: https://github.com/petertodd/python-bitcoinlib

@petertodd It seems to be that we don't want to include the complete python-bitcoinlib. And it's useful to have the tests self-contained instead of asking the user to install dependencies.The only part we need is the RPC proxy to be able to run tests against bitcoind.

The other option would be to use a Python wrapper around bitcoin-cli in the test harness, but using the RPC directly seems somehow more elegant to me.

@laanwj Yeah, it is a order of magnitude more code and will just keep getting bigger. It's what I do with the dust-b-gone repo though - computers are fast and disks big.