Class LazyPlanningAdornments

java.lang.Object
tools.refinery.interpreter.localsearch.matcher.integration.LazyPlanningAdornments
All Implemented Interfaces:
IAdornmentProvider

public class LazyPlanningAdornments extends Object implements IAdornmentProvider
This adornment provider does not trigger the preparation of any plans. Actual query plans will be computed on demand, when the first actual match request is made with a given adornment.

Caution: this is a safe default adornment provider for GenericLocalSearchResultProvider only; do not use for the EMF-specific LS backend.

The benefits is in execution time: query planning costs for adornments are postponed until first usage or even entirely avoided (when adornment is never used in practice). However, query evaluation time may become less predictable, as the first matcher call (with a given adornment) will include the planning cost. For benchmarking or other purposes where this is not desirable, use an adornment provider that demands plan precomputation for all necessary adornments.

Since:
2.1