@copilot the TestAddAuthor.test_append test is failing, please investigate.
Originally posted by @nathanjmcdougall in #1591 (comment)
When thinking about whether to consider a test worth running during an agent session, we need heuristics. In this case, the agent didn't realize that modifying the TOMLFileManger could affect higher-level functions and the TestAddAuthor test was failing once the agent finished its session, i.e. it stopped too early.
The more general principle here is to consider the layers of abstraction affected by a change, e.g. by inspecting the layered architecture in .importlinter config. Aspects in a higher layer can always be affected by a lower layer and the ramifications should be investigated in terms of running more tests.
@copilot the
TestAddAuthor.test_appendtest is failing, please investigate.Originally posted by @nathanjmcdougall in #1591 (comment)
When thinking about whether to consider a test worth running during an agent session, we need heuristics. In this case, the agent didn't realize that modifying the TOMLFileManger could affect higher-level functions and the
TestAddAuthortest was failing once the agent finished its session, i.e. it stopped too early.The more general principle here is to consider the layers of abstraction affected by a change, e.g. by inspecting the layered architecture in
.importlinterconfig. Aspects in a higher layer can always be affected by a lower layer and the ramifications should be investigated in terms of running more tests.