Add support for TimestampAnnotation#99
Conversation
91da59c to
3edb24c
Compare
|
This looks pretty straightforward - can you add a test or two just to make sure it's working as intended? |
|
Sure, where should I add the tests? There are a lot of moving parts to this testing infrastructure! I assume I should test both packing and unpacking a TimestampAnnotation. |
|
I should add that I can successfully run the tests locally with |
|
I'd recommend adding a TimestampAnnotation to the pack in |
|
Thanks. Do you have a way to achieve a quicker edit/test iteration loop? Waiting for the entire |
|
my solution is also hacky: manually edit the pytest command in |
|
omero-test-infra is definitely for bots and not mere mortals, but there are some proto-docs on developer mode -- https://github.com/ome/omero-test-infra#developer-mode But essentially it's just docker with call backs. If you don't want to keep copying the edits & tests into the docker, add them as a mount into your image, then edit on the non-docker side and pytest within the container. |
|
My solution so far is to first run this once to establish a persistent set of containers: .omero/docker dev start_up
.omero/docker dev install
.omero/docker dev run py common
.omero/docker dev run py setupThen run this every time I've made changes I want to test: .omero/docker dev install && .omero/docker dev run py check && .omero/docker dev run --user cli buildAlthough the only way to run specific limited tests with this approach is to edit I can also run However I'm running into a bigger issue now: These tests might not be idempotent or fully independent of each other. I'm experiencing unexpected failures when running certain individual tests by themselves or when running the same test repeatedly. |
Assumptions, e.g., on the ID numbers? The only thing I can imagine is for the suite to first load the lowest ID and build from there, but that's going to be very finicky. |
|
No, weird stuff like the implicit omero session not being available if I run an arbitrary test: But it works if I also include the first test in this suite in the same run: |
cea4dc2 to
091a35c
Compare
|
Ah, ok. That seems like it should be more straight-forward, like a module setup is needed. (Launched workflows) |
No description provided.