K matrix broken in WHIZARD 2

[Juergen Reuter]

The new version of the colorizer is symmetrizing the external legs. This symmetrization breaks my hack I did for WHIZARD1 using a preferred choice for the external "hang-up" leg to distinguish s-channels from t- and u-channels. I'm not sure whether there is at all a solution for this problem within the new colorizer and targets/fusion complex of WHIZARD2. I do hope so. At the moment there is a double counting, and also the isospin eigenchannels are not completely selected, as they depend also on the time axis.

[Juergen Reuter]

As there is a running diploma thesis connected to this problem, I rank this a bit higher.

[Juergen Reuter]

TO and JRR found the culprit: the crossing restrictions are applied to the uncolored DAG, but not to the colorized one. One needs to insert the crossing cross check there, too. TO will do that, and then I start the comparison of the K matrix amplitudes between W1 and W2.

[ohl]

Fixed the colorizer part in r3060, but I'm not to happy with the 'magic' integers 0, 1, 2, 3 in Vector4_K_Matrix_jr.

[Juergen Reuter]

Reproduced the plots from our paper with Ana. Also when inserted into a fermionic 2->4 amplitude results coincide. Re-enabled the SM_km part of the models self test. Closing.