removed recursive proving to reduce false positives

baseline.py

@@ -78,6 +78,9 @@ def benchmark_nextstep(pwd, get_tactics, send_command, search_depth=3, search_wi
             for (curr_goal, ps, tac_seq) in old_ps:
                 next_tactics = get_tactics(curr_goal, prev_lines + '\n'.join(tac_seq))
                 for next_tactic, _scr in sorted(next_tactics, key=lambda p: -p[1])[:search_width]:
+                    if prop_name in next_tactic:
+                        continue # although this in theory Can be correct, LEAN DOES NOT CORRECTLY THROW ERRORS when the theorem name is used in a proof. 
+                                 # in fact, Lean REPL will return a proofstate with empty goals and no errors! This creates false positives, so we skip these tactics.
                     #print('\n'.join(tac_seq + [next_tactic]))
                     outp, new_proofState = send_tactic(lean_repl, next_tactic, ps)
                     if outp is None: