Lily Wang

Thanks -- can you let me know how that affects #3941? Should it (continue being) on hold?

I wouldn't mind getting this done for 2.0, but to put doctest in the CI suite we'll need it to actually pass first. I'll treat this as the overall issue...

@IAlibay not totally sure we can do that without an append mode -- the writers require the number of atoms to initialize and I think XYZWriter might run a check...

#2375 implemented the following, where `type(ag) == AtomGroup`: * `Universe.select_atoms(ag)` returns `Universe.atoms.intersection(ag)`, i.e. copy of `ag` with any duplicates and ordering lost. * `AtomGroup.select_atoms(ag)` returns `AtomGroup.intersection(ag)`. * `Universe.select_atoms(ag, updating=True)` returns...

My objections lay mainly along having to contort code to fit operations that pint doesn't support, similar to what @mattwthompson's mentioned. Given that we return a copy of a numpy...

> Should this be 3-2-1-0, instead of 3-1-2-0? Just to comment on this -- that's right, it should be 3, 2, 1, 0 @wenyan4work. It was a minor bug, fixed...

> exactly according to the SMIRKS atom labeling, without doing any sorting IIRC ``Molecule/Topology.chemical_environment_matches`` does this, but the parameter handlers used in creating openmm systems and labelling molecules then store...

> For parameter assignment, I'm pretty sure that's wrong. I don't think this would have caused much trouble, but given that some of our parameters are (for reasons I don't...

``` >>> gsd = mda.Universe(GSD) >>> gsd.atoms.resnames array([0, 1, 2, ..., 647, 647, 647], dtype=object) >>> gsd.select_atoms("resname 0") >>> gsd.select_atoms("resname 2") ``` ^ A consequence of integer resnames.

> docstring of MDAnalysis.lib.qcprot:12: ERROR: Unexpected indentation Hmm, yes, I'd say go ahead and see the change works but maybe raise a new (low-priority) issue to check out what's going...