Background
Follow-up from PR #9699 (thread). The new tests in proofreading_split_mapping.spec.ts thoroughly cover the return values of splitAgglomerateInMapping, but they don't verify the end-to-end behavior when a foreign proofreading update action is polled in: that the correct segment ids (not agglomerate ids) are requested for reload and that meshes are refreshed correctly. This is the bug class fixed in 587dde5.
Unit-e2e tests ("e2e" without a real backend)
Extend our existing proofreading test setup (BackendMock + AgglomerateMapping, as in proofreading_single_user/proofreading_multi_user) to cover, with high control:
Test-infra cleanup
Playwright (additional safety net)
Background
Follow-up from PR #9699 (thread). The new tests in
proofreading_split_mapping.spec.tsthoroughly cover the return values ofsplitAgglomerateInMapping, but they don't verify the end-to-end behavior when a foreign proofreading update action is polled in: that the correct segment ids (not agglomerate ids) are requested for reload and that meshes are refreshed correctly. This is the bug class fixed in 587dde5.Unit-e2e tests ("e2e" without a real backend)
Extend our existing proofreading test setup (
BackendMock+AgglomerateMapping, as inproofreading_single_user/proofreading_multi_user) to cover, with high control:splitAgglomerateupdate action is correctly handledTest-infra cleanup
proofreading_split_mapping.spec.tsoff the hand-written mapping fixtures (error-prone) and ontoBackendMock/AgglomerateMappingPlaywright (additional safety net)