atomspec model ordering not used in matrixcopy between multiple models

[Elaine Meng]

This may be the same problem as an atomspec ordering ticket, but I was unable to find that ticket, sorry. Please reassign/close as appropriate.

If I use "view position" to matrixcopy from multiple models to multiple models, only some changes in ordering have an effect, and the result is usually wrong. First, six commands to set up test system:

open 1plx
close #1.10-80
rainbow mol
move x -10 mod #1.1
move x 10 mod #1.2
move y 10 mod #1.3

then:
view pos #1.9,8,7 sameas #1.1,2,3

...gives the correct result. However, nothing changes if I give the opposite order of the "sameas" models in any of the following formats:

view pos #1.9,8,7 sameas #1.3,2,1
view pos #1.9,8,7 sameas #1.1,2,3
view pos #1.9,8,7 sameas #1.3,2,1
view pos #1.9,8,7 sameas #1.3#1.2#1.1
view pos #1.9#1.8#1.7 sameas #1.3#1.2#1.1

Although changing the order of the first set of models makes a difference, it is not necessarily right, because the following gives a different but incorrect result:

view pos #1.7,8,9 sameas #1.1,2,3

[Tom Goddard]

Fixed.

Model argument parsing was not preserving order with TopModelsArg.