Silence warning #647
Silence warning #647
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #647 +/- ##
=======================================
Coverage 83.46% 83.47%
=======================================
Files 17 17
Lines 3556 3558 +2
=======================================
+ Hits 2968 2970 +2
Misses 588 588
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
| @@ -1302,6 +1302,10 @@ def _observe_code_setup(self, _=None): | |||
| for key, value in self.code_setup.items(): | |||
| if hasattr(self, key): | |||
| if key == "default_calc_job_plugin": | |||
| if "None" in value: | |||
There was a problem hiding this comment.
What is the type of value? Somewhat surprising if it would be list for tuple?
There was a problem hiding this comment.
A string. It's an entry point. Like quantumespresso.pw. But if you don't associate a default calcjob plugin, when you select the Quantum ESPRESSO code from the code selector in the resource setup widget, the value somehow ends up being quantumespresso.None. So I check if the string None is in the entry point string. But yes, need to investigate why this happens.
There was a problem hiding this comment.
Ah, okay, that definitely needs an explanatory comment in the code, because "None" in value looks suuuper weird 😄
Referenced the issue, where the matter is described in detail. Added a reference in the TODO comment. @danielhollas let me know if I can merge this 🙏 |
edan-bainglass
force-pushed
the
HACK/silence-warning
branch
from
9b30bf7 to
d082f0a
Compare
|
@danielhollas Wow! I'm at a loss as to how changing the comment now causes a test to fail 🤦🏻♂️ |
There was a problem hiding this comment.
Sure, go ahead! I am a bit concerned this hacky fix will bite us in the future unless we investigate it properly, but it's fine for now. :-)
@danielhollas Wow! I'm at a loss as to how changing the comment now causes a test to fail 🤦🏻♂️
Looks like a test flake, tests are passing after I rerun them. Could you open an issue about a flaky test_cod_widget, including the errors from the CI? That test is accessing an external database, we should mock that out.
Thanks!
Indeed, very flaky, as I also re-ran the tests before making my previous comment, but they still failed 😢 Okay, opening an issue. |
Recent developments in aiidalab-home of an external notebook to manage codes run into an issue with the AWB computational resources widgets when used without a default calcjob plugin. This PR silences a suspected false warning for now, to be properly investigated in a future PR.
Handling issue #648