(core) Replace compute stack and frames with _current_node and _current_row_id

Summary:
This is an attempt to optimise Engine._use_node. It doesn't seem to actually improve overall performance significantly, but it shouldn't make it worse, and I think it's an improvement to the code.

It turns out that there's no need to track a stack of compute frames any more. The only time we get close to nested evaluation, we set allow_evaluation=False to prevent it actually happening. So there's only one 'frame' during actual evaluation, which means we can get rid of the concept of frames entirely. This allows simplifying the code and letting the computer do less work in general.

Test Plan: this

Reviewers: dsagal

Reviewed By: dsagal

Subscribers: dsagal

Differential Revision: https://phab.getgrist.com/D3310

sandbox/grist/functions/info.py

@@ -577,7 +577,7 @@ def _prepare_record_dict(record, dates_as_iso=False, expand_refs=0):
   table_id = record._table.table_id
   docmodel = record._table._engine.docmodel
   columns = docmodel.get_table_rec(table_id).columns
-  frame = record._table._engine.get_current_frame()
+  current_node = record._table._engine._current_node
 
   result = {'id': int(record)}
   errors = {}
@@ -590,7 +590,7 @@ def _prepare_record_dict(record, dates_as_iso=False, expand_refs=0):
     # Avoid trying to access the cell being evaluated, since cycles get detected even if the
     # CircularRef exception is caught. TODO This is hacky, and imperfect. If another column
     # references a column containing the RECORD(rec) call, CircularRefError will still happen.
-    if frame and frame.node == (table_id, col_id):
+    if current_node == (table_id, col_id):
       continue
 
     try: