Conversation
|
@aabadie I really don't understand that pydantic backtrace, and the tests. |
I understand it as Pydantic validator doesn't like some user_env set in jobmodel: Could the following diff help? diff --git a/murdock/job.py b/murdock/job.py
index 5095a19..d296b46 100644
--- a/murdock/job.py
+++ b/murdock/job.py
@@ -26,7 +26,7 @@ class MurdockJob:
config: MurdockSettings = MurdockSettings(),
trigger: Optional[str] = "api",
triggered_by: Optional[str] = None,
- user_env: Optional[dict] = None,
+ user_env: Optional[dict[str,str]] = None,
):
self.trigger: Optional[str] = trigger
self.triggered_by: Optional[str] = triggered_byBTW the project is a bit outdated and needs a refresh. I can't even run tox on my machine :) |
Let's see. :))
I have the same problem ... But this needs some love, I don't think the instance at ci.riot-os.org will survive a restart. 😨 |
It runs in a Docker container so normally if the Python version in the image is 3.10 (or 3.9) and the tests are passing here. It should work. But it's indeed better to make a maintenance pass one day. |
Co-authored-by: Alexandre Abadie <alexandre.abadie@inria.fr>
IIUC as pydantic is not pinned, a recent update requires some changes. These are what I needed to get RIOT's ci-staging back up.
edit I won't have time to work on this (e.g., fix the test cases). Please take over. (maybe just pin the working pydantic?)