[LTS] Skip precommit for CI and add missing hint

[Growl1234]

No description provided.

[Growl1234]

Anyone knows how to deal with this test failure, which appearently has nothing to do with this PR?

[ RUN      ] 204_NO_KP_NC
[----------] non-collinear Fe Anti-ferromagnetic and lcao basis calculation
[Warning:   ]  etotref cal=-6436.21478610 ref=-6436.21478589 deviation=0.00000021
[Warning:   ]  etotperatomref cal=-3218.10739305 ref=-3218.10739294 deviation=0.00000011

[Growl1234]

Gentle ping :)
It seems this PR has not received any attention for more than a week...

[mohanchen]

Gentle ping :) It seems this PR has not received any attention for more than a week...

Thx for the reminder. The integration tests and unit tests failed, could you take a look?

[Growl1234]

[WARNING   ]  etotref cal=-6436.21478610 ref=-6436.21478589 deviation=0.00000021
[WARNING   ]  etotperatomref cal=-3218.10739305 ref=-3218.10739294 deviation=0.00000011

I suspect, this is more likely a case where the test issue was previously hidden because the CI workflow failed before the build stage, and it only became visible this time because the previous failure point was skipped.

Given this test result, however, I would personally prefer to relax the tolerance...

[Growl1234]

Strangely, I could not reproduce the deviation on my computer with current LTS branch...

Would it be acceptble to relax the tolerance to 3e-07 only for 204_NAO_KP_NC?

[Growl1234]

And it seems some other tests will have error on my machine (with gcc-mkl toolchain, except the PAW case because it's not enabled on my build...): test.log
but such non-reproduceable case never happened with the tests in develop branch...