O'Mega's rearranging of particles invalidates flavor summation

[kilian]

Looking at the cause for #397, I find that WHIZARD can sum over flavors only if their mass is identical. While O'Mega keeps the ordering in the initial state, it rearranges the final state. Take, for instance

process bbu   = b:B, u => b:B, u

O'Mega transforms this into

!       1: bbar u -> bbar u
!       2: b u -> u b

Now the first particle in the final state has mass zero and mb at the same time, since in the flavor summation algorithm the contributing flavor states share a common phase space. Currently, this is not checked (it should be, however), and the numerical results are nonsense.

Note that this does not exclude summing over quarks and gluons, since only masses count here.

O'Mega has to sort the final state in order to eliminate duplicates, but it should undo the sorting before generating output.

[Juergen Reuter]

What is the status of that guy???

[ohl]

Should be fixed as of r3216, but I leave it open for some testing.

[Juergen Reuter]

Started some testing, the example above looks fine now, and also the LEP_higgs example from the test suite (which triggers the daily Jenkins failure) works now. So up to now everything looks good. Waiting for comments and approvals from CS and WK.

[kilian]

Also see it working now, thanks Thorsten! Closing.