Handle keep-alive behavior to close the connection #201
Conversation
|
@leszekhanusz I fixed the virtualenv so it's not longer modifying lot of things. Note: from https://github.com/graphql-python/gql/blame/master/CONTRIBUTING.md#L34 Now I'm still getting the line length errors... I'm wondering which tool should be managing this. EDIT: I pushed a manual adjustment for lines length... sounds hacky and looks a bit weird. I'm curious what you use to make this automatic? The |
Codecov Report
@@ Coverage Diff @@
## master #201 +/- ##
=========================================
Coverage 100.00% 100.00%
=========================================
Files 16 16
Lines 985 1008 +23
=========================================
+ Hits 985 1008 +23
Continue to review full report at Codecov.
|
modifications: * clean_close = False * keep_alive_timeout is now Optional[int] default None * calling self._fail directly from the _check_ws_liveness coro * no need to cancel the _receive_data_loop coro, it will stop itself once the websocket will close
|
@sneko Could you please check if the refactor is working correctly for you ? |
Hi @leszekhanusz ,
Here the PR for the keep-alive behavior #200, tested with real servers sending
kamessages.I read the CONTRIBUTING.md but struggled with the
make check, I installed them but it changes all my files (including yours) like if parameters were not taken in account. Moreoverisorttells me some lines are >88chars but no other tools automatically adjust them (I thought it would?). I guess I missed something... I did the venv creation +make dev-setup.Tip for others: in my app code I forgot to
await ws_transport.wait_close()making the reconnection after a keep-alive failure failing because some stuff were not clean in the meantime (asynciogiving the priority to other tasks including mines trying a reconnect).Tell me if I need to adjust something :)