Age | Commit message (Collapse) | Author |
|
This reverts the previous commit. Obviously, if the sigma(I) value is
negative, stuff goes wrong. Need a better weighting scheme here.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
connected panels
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
There are already things which alter this structure during processing,
and probably more to come. Taking a copy avoids problems down the line
and fixes a serious bug when using coffset and a fixed clen.
|
|
|
|
Otherwise, the later code adds the new RGs to a shallow copy of the old
list (i.e. just a copy of the pointer). Hilarity ensues.
|
|
|
|
|
|
|