Maybe it would be best just to add a comment to the complete_array and complete_set functions to make clear that the implementation assumes that map_to_asu has been called previously?

Richard

On 12 July 2012 23:23, Pavel Afonine <pafonine@lbl.gov> wrote:
Yes, Ralf's concern was performance. If you hide away map_to_asu in a function that you or someone else runs repetitively then one can run into performance surprises. So it might be better to call map_to_asu consciously, when it's really needed.

Pavel


On 7/12/12 11:11 PM, Peter Zwart wrote:
I floated this once to Ralf and his objection was performance. I guess
it also depends on the situation if you want this to happen.

P


On 12 July 2012 23:06, Luc Bourhis <luc_j_bourhis@mac.com> wrote:
Hi Peter and Nat,

On Thu, Jul 12, 2012 at 8:45 AM, Petrus Zwart <phzwart@lbl.gov> wrote:
In general, I use a 'map_to_asu' whenever possible, this avoids some
problems you illustrate, but isn't elegant.
+1

This is pretty essential for much of our code, assuming you're working
with merged/symmetry-unique data.
It would make sense to automagically map to asu in complete_set and complete_array in order to avoid surprises, wouldn't it? Would you see any adverse effect?

Best wishes,

Luc Bourhis

_______________________________________________
cctbxbb mailing list
cctbxbb@phenix-online.org
http://phenix-online.org/mailman/listinfo/cctbxbb




_______________________________________________
cctbxbb mailing list
cctbxbb@phenix-online.org
http://phenix-online.org/mailman/listinfo/cctbxbb