Run CI-testing on GHA/Windows

[amotl]

Dear Mathias,

while working on crate/crate-python#388, we faced some deadlock issues when tearing down the testlayer on GHA/Windows. So, we wanted to take the chance to also test the testlayer incorporated into cr8.

Please consider this as a work in progress / draft unless further notice.

With kind regards,
Andreas.

[amotl]

Apparently, the deadlock issues on testlayer teardown also happen here.

======================================================================
ERROR: D:\a\cr8\cr8\tests\..\README.rst
Doctest: README.rst
----------------------------------------------------------------------
Traceback (most recent call last):
  File "C:\hostedtoolcache\windows\Python\3.8.6\x64\lib\doctest.py", line 2178, in tearDown
    self._dt_tearDown(test)
  File "D:\a\cr8\cr8\tests\test_integration.py", line 38, in teardown
    node.stop()
  File "D:\a\cr8\cr8\cr8\run_crate.py", line 395, in stop
    self.process.communicate(timeout=10)
  File "C:\hostedtoolcache\windows\Python\3.8.6\x64\lib\subprocess.py", line 1024, in communicate
    stdout, stderr = self._communicate(input, endtime, timeout)
  File "C:\hostedtoolcache\windows\Python\3.8.6\x64\lib\subprocess.py", line 1397, in _communicate
    raise TimeoutExpired(self.args, orig_timeout)
subprocess.TimeoutExpired: Command '['C:\\Users\\runneradmin\\.cache\\cr8\\crates\\crate-4.3.1\\bin\\crate.bat', '-Cdiscovery.initial_state_timeout=0', '-Cnetwork.host=127.0.0.1', '-Cudc.enabled=false', '-Ccluster.name=cr8-tests', '-Chttp.port=44200-44250', '-Cpath.data=C:\\Users\\RUNNER~1\\AppData\\Local\\Temp\\tmp2yvi7ihz']' timed out after 10 seconds

-- https://github.com/mfussenegger/cr8/pull/337/checks?check_run_id=1395324258#step:6:237

[amotl]

Another thing I would like to share in this context is that

echo "select name from sys.cluster" | cr8 timeit --hosts http://127.0.0.1:44200

will croak with

SqlException: SQLParseException[line 1:1: mismatched input '"select name from sys.cluster"' expecting {'SELECT', 'DEALLOCATE', 'CREATE', 'ALTER', 'KILL', 'BEGIN', 'COMMIT', 'ANALYZE', 'DISCARD', 'EXPLAIN', 'SHOW', 'OPTIMIZE', 'REFRESH', 'RESTORE', 'DROP', 'INSERT', 'VALUES', 'DELETE', 'UPDATE', 'SET', 'RESET', 'COPY', 'GRANT', 'DENY', 'REVOKE'}] occurred using: {"stmt": "\"select name from sys.cluster\""}

-- https://github.com/mfussenegger/cr8/pull/337/checks?check_run_id=1395743511#step:6:173

This happens even when running that on bash after 94ad6b2.

Apparently, the quotes are taken verbatim when echoing a string through a pipe. I believe [1] is discussing this very issue and others are experiencing similar pains, see also rundeck/rundeck#602 and rundeck/rundeck#1406.

[1] https://stackoverflow.com/questions/804646/how-do-you-strip-quotes-out-of-an-echoed-string-in-a-windows-batch-file

[amotl]

We have been able to dedicate some more minutes of research into the deadlocking issue on layer teardown when running on Windows.

Firstly, we have been able to discover very valuable resources at [1,2,3], which give many general insights into signal handling for child processes on Windows.

However, in the context of subprocess.Popen, we also just found [4,5,6], where [5] specifically states that the CREATE_NEW_PROCESS_GROUP creationflags parameter to subprocess.Popen is required for using os.kill() on the subprocess. We will try that within one of the next iterations.

[1] https://stackoverflow.com/questions/813086/can-i-send-a-ctrl-c-sigint-to-an-application-on-windows
[2] https://stackoverflow.com/questions/3342941/kill-child-process-when-parent-process-is-killed/4657392#4657392
[3] https://stanislavs.org/stopping-command-line-applications-programatically-with-ctrl-c-events-from-net/
[4] https://docs.python.org/3/library/subprocess.html#subprocess.Popen.send_signal
[5] https://docs.python.org/3/library/subprocess.html#subprocess.CREATE_NEW_PROCESS_GROUP
[6] https://docs.python.org/3/library/asyncio-subprocess.html#asyncio.asyncio.subprocess.Process.send_signal