Fixed #43 : BnfNodeParser: I can recognize when multiple level of ISA

Fixed #44 : BnfNodeParser: I must simplify results when multiple levels of ISA
Fixed #45 : Dynamic variables cannot be parsed at restart
Fixed #46 : Concepts variables values are transformed into list by default
Fixed #47 : SheerkaAdmin. Add min, max, mean time when restoring files

src/evaluators/MultipleSuccessEvaluator.py

@@ -10,6 +10,7 @@ class MultipleSuccessEvaluator(AllReturnValuesEvaluator):
     So we cannot decide whether it's a MultipleSameSuccess or not
     All parser in error will be discarded
     Cannot match if there is at least one evaluator in error
+    Cannot match if there is at least one successful parser
     """
     NAME = "MultipleSuccess"
 
@@ -25,28 +26,50 @@ class MultipleSuccessEvaluator(AllReturnValuesEvaluator):
         nb_evaluators_in_success = 0
         to_process = False
 
+        debugger = context.get_debugger(MultipleSuccessEvaluator.NAME, "matches")
+        debugger.debug_entering(return_value=return_values)
+
         for ret in return_values:
             if ret.status and ret.who.startswith(BaseParser.PREFIX):
+                reason = "a successful parser found"
+                debugger.debug_log(f"Failed because {reason}. ret={ret}")
                 return False
+
             elif ret.who.startswith(BaseEvaluator.PREFIX) and not ret.status:
+                reason = "a failed evaluator found"
+                debugger.debug_log(f"Failed because {reason}. ret={ret}")
                 return False
+
             elif ret.status and context.sheerka.isinstance(ret.body, BuiltinConcepts.REDUCE_REQUESTED):
                 to_process = True
                 self.eaten.append(ret)
+                debugger.debug_log(f"BuiltinConcepts.REDUCE_REQUESTED is found. {to_process=}")
+
             elif ret.status and ret.who.startswith(BaseEvaluator.PREFIX):
                 if self.already_seen(context, ret):
+                    reason = "of duplicate return value"
+                    debugger.debug_log(f"Failed because {reason}. ret={ret}")
                     return False
                 nb_evaluators_in_success += 1
                 self.successful_return_values.append(ret)
                 self.eaten.append(ret)
+                debugger.debug_log(f"Eating and keeping {ret=}. {nb_evaluators_in_success=}")
+
             elif not ret.status and ret.who.startswith(BaseParser.PREFIX):
                 self.eaten.append(ret)
+                debugger.debug_log(f"Eating {ret=}.")
+
             # else:
             # other concepts. We do not care if there are successful or not
             # They won't be part of result nor part of the parent
             # --> So they will be handled by other evaluators
 
-        return to_process and nb_evaluators_in_success > 1
+        res = to_process and nb_evaluators_in_success > 1
+        if not res:
+            reason = "not to_process and nb_evaluators_in_success > 1"
+            debugger.debug_log(f"Failed because '{reason}', {to_process=}, {nb_evaluators_in_success=}")
+
+        return res
 
     def eval(self, context, return_values):
         context.log(f"{len(self.successful_return_values)} successful return values, {len(self.eaten)} item(s) eaten",