Closed ramya-rao-a closed 3 years ago
cc @richardpark-msft, @HarshaNalluru
I looked into why the CI is failing for this PR. It looks like the new tests don't set reconnect: false
on the connections they create. When I ran the tests locally I found there were 2 timers hanging around related to rhea's built in reconnect logic. When I set reconnect
to false for all the added tests those timers went away.
I didn't dig into exactly which tests were still allowing reconnects to occur, maybe there are some that don't explicitly close the connection so rhea attempts to reconnect.
yay! Thanks for the hint @chradek! Tests are green now
Please add changelog
Description
Brief description of the changes made in the PR. This helps in making better changelog
Few design discussions:
AbortSignalLike
interface is redefined in this PR instead of relying on an Azure scoped package like@azure/abort-controller
which is used in the testsAbortError
class due to potential conflicts with upstream usage. Instead, using same oldError
class with custom name. This is definitely open for debate as to how we should approach this.Reference to any github issues
48