We can avoid generating multiple versions of the slot_getattr. Once we
deform any of the attributes in the tuples, we make it a virtual tuple.
At code generation time, we know exactly how many need to be deformed
and can in fact go ahead deform all the way. This way we don't need to
worry about the case when slot_getattr is called on a virtual tuple with
attnum > nvalid - that is deformation is partially complete.

To enable this, we need to collect information from all the code
generators that depend on SlotGetAttrCodegen before it is generated. We
maintain a static map (keyed on the manager and the slot) to instances of
SlotGetAttrCodegen. We also introduce a InitDependencies phase that
happens before the GenerateCode phase, when dependants of
SlotGetAttrCodegen can retrieve instances from the static map.

  * Skip generating ExecEvalExpr ScanNodes when ProjInfo is a Var list
  * Skip generating SlotGetAttr in AggNodes

ExprTreeGeneratorInfo object keeps track of metadata information
needed to generate code for expression trees.
Refactor ExprTreeGenerators to always use econtext->ecxt_scantuple to
generate the functions. We also have an extra slot_check block to fall
back if it is called with a different slot - which shouldn't happen for
Agg and Scan nodes.

This closes #874

Verify that input expression is of type VAR <= CONST
Codegen expression evaluation
Fix wrong return value type
Support constants and variables of Date type
Add comments and fix code style