command "surface surf-model close" broken

[Elaine Meng]

Previously one could specify a surface model to close by its model number. Now there is an error message that the specification does not contain atoms. (CC'd Greg because it might be his doing) Example:

open 2gbp
surf ligand
surf #1.2 close

Log (lack of space in error message is another bug, albeit minor):
surface #1.2 close
No atoms specifiedby #1.2

[Elaine Meng]

One possibility is to disallow that usage, since "close #1.2" would work. However, people may be confused if they can use (sticking to the example above) "surf ligand close" but not "surf #1.2 close" when they can see in the Model Panel that the surface is #1.2.

[Tom Goddard]

This broke 15 months ago due to a change Conrad made to atomspec parsing where specifying a molecular surface no longer implies the atoms of the surface are specified. But I vaguely recall running into some bad behavior when specifying a molecular surface implied its atoms were specified -- don't remember the command or scenario. I think it is probably safer to just make surface close specifically handle this situation instead of having a molecular surface model spec always imply the atoms are specified.

I'm puzzled now why we use "surface #1.2 close" instead of the subcommand syntax "surface close #1.2". Also surface show and hide don't use subcommands. Is that just because the surface command predates are use of subcommands? Should I change surface close/show/hide to use subcommands requiring syntax "surface close #1.2"?

[Tom Goddard]

Fixed