Feature: Use hartree-like sum to unify the definition of drho

[WHUweiqingzhou]

Linked Issue

Fix #3225 after PR #3226

What's changed?

Before: use simple sum in mix_rho_recip_new
Now: reuse hartree-like sum to cooperate with get_drho()

Any changes of core modules? (ignore if not applicable)

calculated results may be changed.

Please Notes:
I find some old ref-data in current CI refers to wrong magnetic configuration. For example:
in 104_PW_AF_magnetic, in result.ref, we have etotref -5866.174728188967
This energy refers to non-magnetic configuration instead of FM.
Same wrong results appear in 104_PW_FM_magnetic, and 127_PW_15_PK_AF.
To correct these errors, I will change INPUT and make corresponding replacement of result.ref.

[Qianruipku]

Please check new warnings and update results.

[WHUweiqingzhou]

Please check new warnings and update results.
OK. I will do this.