sort by topology does poor job for big versus small tree

[westram]

When sorting a bigger according to a smaller (e.g. useful after inserting species into an existing to see where they have been dropped) sorting is very poor.

The algorithm assumes a position values of 0.5 for unknown species. Doing that forces unknown species

• downwards in the upper half and
• upwards in the lower half.

Instead it should use some magic value (with meaning unknown) only forcing these species towards -center if no other information is present.

Pointer:

• feature has been introduced with #445

[westram]

Example of misbehavior

When input-
is sorted according to the following smaller sort-
the result- is

Sort order of colored species in result- matches their order in sort-. so far so good.

Problems (marked species):

• the subtree containing CloTyro3/4 (and 4 more)
  • gets "random" order, because all species are "unknown" in sort- → original order of input- should be kept for such subtrees
  • whole subtree is "pulled" downwards, towards the -center (swapping position with its brother-subtree containing CloButyr/2)
• CelBiazo "pulls" upwards, towards the -center → should stay below its brother-subtree (as in input-, regardless of new order of brother-subtree

[westram]

by [12019]

Fixed result- is