Lift

SessionVar and RequestVar are well documented in Chapter 3 of Exploring Lift so I thought I would go into SessionMemoize a little bit. Basically it's like a Map, tied to your session. You should generally use it for values you only want to calculate once, because they're expensive to calculate and don't change often. Conveniently, it offers a way to encapsulate the following sequence of operations:

1. Check whether the value is available in the map (containsKey)
2. Use the value if available
3. Calculate the value if not available
4. Store the value for next time
5. Return the calculated value

The syntax for this common operation is

mySessionMemoObject.get(key,valueToInsertIfNotFound)

The value will not be evaluated if the key is present. One thing that you should be careful of is that, unlike SessionVars, all SessionMemoize instances belonging to a session share underlying state. So these will not conflict:

object myMemo extends SessionVar[Int](1)
object myOtherMemo extends SessionVar[Int](2)
myMemo.is -> 1
myMemo.is -> 2

But these might surprise you, because the parameter of the second call will not be evaluated - the map already contains the key.

object myMemo extends SessionMemoize[Int,String]
object myOtherMemo extends SessionMemoize[Int,String]
myMemo(1,"Belongs to the first memo")
myOtherMemo(1,"Belongs to the second memo")

myOtherMemo(1) -> "Belongs to the first memo")

If the two SessionMemoize objects are differently parametrically typed, you will get a very confusing runtime exception.