I'd say this is a separate bug. The fact that the tests are included via `../../../` in the manifest is likely the catalyst for this bug. I don't know exactly what's going wrong, but there's a mismatch between manifests and dirs and this is tripping up the [logic in Taskgraph](https://searchfox.org/mozilla-central/rev/0e3d9bfae6fdaa6cb29cbce3f25471d5708aedc3/taskcluster/gecko_taskgraph/transforms/test/chunk.py#64) that is supposed to ensure test chunks run actual tests that don't get skipped. Possible solutions (in order of my preference): 1. Avoid `../../` in the manifest (dupe tests?) 2. Have updatebot pass in manifests rather than paths (unsure if will work or if acceptable?) 3. Try to fix this in test resolving logic (unsure how difficult this is) We could also make the test harness go green if this happens, but then we'd miss problems in the resolving logic.
Bug 1772754 Comment 5 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
I'd say this is a separate bug. The fact that the tests are included via `../../../` in the manifest is likely the catalyst for this bug. I don't know exactly what's going wrong, but there's a mismatch between manifests and dirs and this is tripping up the [logic in Taskgraph](https://searchfox.org/mozilla-central/rev/0e3d9bfae6fdaa6cb29cbce3f25471d5708aedc3/taskcluster/gecko_taskgraph/transforms/test/chunk.py#64) that is supposed to ensure test chunks run actual tests that don't get skipped. Possible solutions (in order of my preference): 1. Avoid `../../` in the manifest (dupe tests?) 2. Have updatebot pass in manifests rather than paths (unsure if will work or if acceptable?) 3. Try to fix this in test resolving logic (unsure how difficult this is) 4. Make the test harness go green if this happens (but then we'd miss problems in the resolving logic)