[ET Trac] [Einstein Toolkit] #1073: point_classification should not have storage and be removed

[Einstein Toolkit]

#1073: point_classification should not have storage and be removed
--------------------+-------------------------------------------------------
 Reporter:  rhaas   |       Owner:  eschnett
     Type:  task    |      Status:  new     
 Priority:  minor   |   Milestone:          
Component:  Carpet  |     Version:          
 Keywords:          |  
--------------------+-------------------------------------------------------
 ClassifyPoints and point_class should be removed from Carpet since right
 now they always set the class to 1.

 There is interest in having the functionality that the routine and grid
 function provide, namely to know at which MoL substep we need to have
 valid data at a given point (eg. for con2prim and con2prim failures as
 well as possibly skipping computations in buffer regions that are no
 longer valid). This could be added to CarpetEvolutionMask quite
 straightforwardly, but might require some redefinition of the mask values,
 eg. 0 for valid in all MoL substeps, 1 valid until and incl substep 1,
 ..., MoL_Num_Intermediate_Steps-1 valid only during the first substep,
 MoL_Num_Intermediate_Steps for restricted points. This would mean eg an
 analysis routine can ignore anything >0 while an evolution routine would
 have to check with MoL.

-- 
Ticket URL: <https://trac.einsteintoolkit.org/ticket/1073>
Einstein Toolkit <http://einsteintoolkit.org>
The Einstein Toolkit