| A service interface that maps attributes between two flows. Used by the
subflow state to map attributes between a parent flow and its sub flow.
An attribute mapper may map attributes of a parent flow down to a child flow
as input when the child is spawned as a subflow. In addition, a
mapper may map output attributes of a subflow into a resuming parent flow as
output when the child session ends and control is returned to the
parent flow.
For example, say you have the following parent flow session:
Parent Flow Session
-------------------
-> flow = myFlow
-> flowScope = [map-> attribute1=value1, attribute2=value2, attribute3=value3]
For the "Parent Flow Session" above, there are 3 attributes in flow scope
("attribute1", "attribute2" and "attribute3", respectively). Any of these
three attributes may be mapped as input down to child subflows when those
subflows are spawned. An implementation of this interface performs the actual
mapping, encapsulating knowledge of which attributes should be
mapped, and how they will be mapped (for example, will the same
attribute names be used between flows or not?).
For example:
Flow Attribute Mapper Configuration
-----------------------------------
-> inputMappings = [map-> flowScope.attribute1->attribute1, flowScope.attribute3->attribute4]
-> outputMappings = [map-> attribute4->flowScope.attribute3]
The above example "Flow Attribute Mapper" specifies
inputMappings that define which parent attributes to map as
input to the child. In this case, two attributes in flow scope of the parent
are mapped, "attribute1" and "attribute3". "attribute1" is mapped with the
name "attribute1" (given the same name in both flows), while "attribute3" is
mapped to "attribute4", given a different name that is local to the child
flow.
Likewise, when a child flow ends the outputMappings define
which output attributes to map into the parent. In this case the subflow
output attribute "attribute4" will be mapped up to the parent as "attribute3",
updating the value of "attribute3" in the parent's flow scope. Note: only
output attributes exposed by the end state of the ending subflow are eligible
for mapping.
A FlowAttributeMapper is typically implemented using 2 distinct
org.springframework.binding.mapping.AttributeMapper implementations:
one responsible for input mapping and one taking care of output mapping.
Note: because FlowAttributeMappers are singletons, take care not to store
and/or modify caller-specific state in a unsafe manner. The
FlowAttributeMapper methods run in an independently executing thread on each
invocation so make sure you deal only with local data or internal,
thread-safe services.
See Also: org.springframework.webflow.engine.SubflowState
See Also: org.springframework.binding.mapping.AttributeMapper
author:
Keith Donald
author:
Erwin Vervaet |