First, you need a recent installation (>= 0.13 should be fine) of Bitcoin Core, get it from here if you need to.
If you use a self-compiled version, note you do not need GUI support but you do of course need wallet support.
You do not need a synced blockchain, either mainnet or testnet, since these tests only use regtest mode.
You should already have installed CoinSwapCS as explained in the main README file
for this repo.
The test does not currently support the coinswapcs.cfg file in any place
other than ~/.CoinSwapCS/coinswapcs.cfg, so you should also move this file
someplace else (changing the name to e.g. coinswapcs.mine is enough).
In the virtualenv you set up, install pytest:
pip install pytest
You should then be able to navigate to test/ in the CoinSwapCS directory
and then run:
./run_all.sh "/path/to/bitcoind" [--keep] [TEST1 TEST2 TEST3...]
--keep - Do not clean up after a run, leaving all test files in the
temporary directory.
TEST.. - Test names in a list, or nothing to run all tests.
For example:
./run_all.sh $HOME/software/bitcoin/bin/bitcoind --keep cooperative
This will execute about 20 test cases (currently), and should take 10-15
minutes in total. Test files are kept in a random directory with the path:
/dev/shm/cstest[0-9]*
If you want to try individual test cases with py.test, simply run the command
with the correct runtype flag, e.g. the case where both sides behave normally
would be run as:
py.test --btcroot=/path/to/bitcoin/bin/ --btcpwd=123456abcdef --btcconf=/path/to/bitcoin.conf --runtype=cooperative -s
(The individual test cases are listed in test_coinswap.py in the variables
alice_classes and carol_classes.)
If the tests fail, please report this as an issue on this repo. Thanks.
===
(A technical detail: the use of a shell script wrapping pytest looks a bit stupid
if you're familiar with these things; and arguably it is. The reason is that it
is not possible to run the twisted reactor more than once in one program execution.
It is possible to get round this with the twisted.trial module, but it's all a
bit of a mess, so I decided not to bother with it. This bash script will error out
on any pytest failure based on exit codes, so it's good enough, arguably.)