Nightly tests adjust run test_memory_leak in test move high_link to nightly

[unkcpz]

• aiida_computer_ssh fixture set safe_interval -> 1.0s
• The test_memary_leak test for SSH timeout because the long intervals of 4 steps.
• 25ccef9 test whether move high_link to nightly will decrease coverage, as suggested in Consider moving slow tests to nightly test suite #6526

I make safe_interval short for pytest ssh computer fixture, so it can used for running memory leak test not just in nightly.

@GeigerJ2 you may already have it in your PR. the calcjob run in this test can reproduce the 4 times transport request which you may interested to have a look.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 78.00%. Comparing base (7a460c0) to head (83625ca).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #6701      +/-   ##
==========================================
+ Coverage   77.99%   78.00%   +0.01%     
==========================================
  Files         563      563              
  Lines       41766    41766              
==========================================
+ Hits        32573    32574       +1     
+ Misses       9193     9192       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[unkcpz]

Although not failed with pytest-xdist, I think this kind of DB test should run with aiida_profile_clean to start from a clean profile. If it can be moved to nightly then the slow down because of aiida_profile_clean is less noticeable.

[GeigerJ2]

Do you have any timing for this? Currently, you move it to nightly, but are not using aiida_profile_clean?

[unkcpz]

Good catch, I think this one also need to use aiida_profile_clean.

[danielhollas]

btw: To fix the lockfile issue, you could have just run uv lock. This would reduce the diff here, since it would not upgrade the diff packages, it would only remove the aiida-core version from the lockfile.

[unkcpz]

btw: To fix the lockfile issue, you could have just run uv lock. This would reduce the diff here, since it would not upgrade the diff packages, it would only remove the aiida-core version from the lockfile.

Didn't know, I update with --upgrade in #6698 and merged. Sorry.

[danielhollas]

Didn't know, I update with --upgrade in #6698 and merged. Sorry.

No worries, I am also new to this lockfile business. I'll try to find some time next week and write up our experience so far on wiki (and perhaps move the wiki article to an internal doc in the repo as we discussed).

[unkcpz]

I'll try to find some time next week and write up our experience so far on wiki (and perhaps move the wiki article to an internal doc in the repo as we discussed).

That would be super nice. uv is a great stuff, I switch all package management when working with engine part to uv, now every change is expected and easy to switch between python versions. I can also contribute to dev wiki on how to set and test in different python versions.