You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The goal of this issue would be to improve the following
Currently each integration test relies on separate docker containers, meaning there's some time being lost starting/stopping (and building) containers for each test individually.
this doesn't seem strictly required, as most or even all present integration tests can be run on the same set of containers
moreover, sharing the same set of containers means tests could be run concurrently (they're run sequentially atm)
A corollary to 1 is that adding multiple tests which call set_test_fixture into the same file is not reliable as there are docker container port races ongoing
one way to resolve it now is by being explicit that those tests should be executed serially
ideally though this is resolved alongside 1, i.e. by sharing docker infra between the tests
Presumably at some point there might be tests that actually change something which other tests shouldn't observe, so the ideal solution would allow for having a separate set of containers for those tests alone, and shared ones for all the rest.
The text was updated successfully, but these errors were encountered:
The goal of this issue would be to improve the following
set_test_fixture
into the same file is not reliable as there are docker container port races ongoingtests
folder is compiled into a separate binary and executed serially: https://corrode.dev/blog/tips-for-faster-rust-compile-times/#combine-all-integration-tests-into-a-single-binaryPresumably at some point there might be tests that actually change something which other tests shouldn't observe, so the ideal solution would allow for having a separate set of containers for those tests alone, and shared ones for all the rest.
The text was updated successfully, but these errors were encountered: