Closed toebsen closed 11 months ago
OK, looks good -- I think the other transports are not affected of this pesky problem.
@christoph2 Thanks for merging. It seems that the deploy to pypi failed: https://github.com/christoph2/pyxcp/actions/runs/6509156789/job/17680708294. Can you please take a look?
OK, now available as 0.21.6 on PyPI.
The small problem is/was the lack of documentation -- If you submit a PR bumpversion patch
is required, otherwise the upload_pypi action is faced with an already existing version on PyPI and fails.
I'm currently working on a larger update, hopefully including documentation (definitely including high-level measurement functions)...
When trying to connect e.g. via UDP to a XCP slave which is not responding, the default error handler will try to reconnect infinite times. In some scenarios this is not sensible. Therefore, I've disabled the error handler.
This unfortunately leads to a
RuntimeError: threads can only be started once
fromself._packet_listener.start()
in thetransport.connect()
I've fixed this for all the relevant transports, but only verified that it works with the ethernet transport. Possible repro script:
With the proposed changes it is possible to call the connect multiple times, when the error handler is disabled.
Types of changes