node-infrastructure

The PostgreSQL Node type system uses tagged-union structures managed by a NodeTag enum, enabling internal code to handle diverse parse, plan, and execution tree nodes through generated copy, equal, out, and read function families.

To add a new PostgreSQL Node type, you extend the tagged-union structures, update the NodeTag enum, and run the gen_node_support.pl code generator to automatically produce the required copy, equal, out, and read functions.

The best way to implement PostgreSQL tree walkers and mutators is by following established patterns for expression_tree_walker and expression_tree_mutator, ensuring safe navigation and modification of complex parse, plan, and execution trees.

You do not need to manually write these functions for new PostgreSQL struct fields; the gen_node_support.pl code generator automatically creates the copy, equal, out, and read function families to maintain source tree consistency.

Modifying PostgreSQL struct fields without updating the NodeTag enum or rerunning gen_node_support.pl causes serialization bugs because the generated out and read function families become desynchronized from the actual tagged-union structure.

Yes, gen_node_support.pl maintains PostgreSQL expression tree traversal integrity by generating consistent function families that expression_tree_walker and expression_tree_mutator rely on to safely navigate tagged-union structures.